MEASURING AND MONITORING BIODIVERSITY IN

TROPICAL AND TEMPERATE FORESTS

' 1995 by Center for International Forestry Research
All rights reserved. Published 1995.
Printed in Malaysia

ISBN: 979-8764-01-3

Center for International Forestry Research

Bogor, Indonesia

Mailing Address: PO Box 6596 JKPWB, Jakarta 10065, Indonesia

iv
Preface

This book contains 24 papers selected from among those presented at an interna-
tional symposium on Measuring and Monitoring Biodivcrsity in Tropical and Tem-
perate Forests , held in Chiang Mai, Thailand, between the 27th of August and 2nd of
September, 1994. The symposium was attended by over 240 scientists from more
than 40 countries from around the world. In addition to four days of paper presen-
tations, there was also a one-day field trip, and a continuous poster session, with
more than 35 posters, as well as computer demonstrations of software packages for
identifying and measuring biodiversity.
Biodiversity is an immense subject, and as tropical and temperate forests are home to a large
proportion of the earth s terrestrial biodiversity, it is obviously very difficult to cover the topic
comprehensively in a single volume. The papers included in this book were selected to give as
broad a coverage as possible of key topics under the general title, including Principles of Measur-
ing and Monitoring Biodiversity (8 papers), Genetic Diversity (6 papers), Species and Ecosystem
Diversity (5 papers), and Methodology (5 papers). Inevitably, forest trees are the theme of many
papers, but also included are papers dealing with diversity of arthropods, microfungi, birds and
butterflies, and gibbons, as well as many papers dealing with the entire range of biodiversity.
We would like to thank many people who helped to make the symposium and this book
possible. Among these are the sponsors of the symposium, the Royal Forest Department of
Thailand, the Commission of the European Communities, the Canadian International Devel-
opment Agency (including the ASEAN Forest Tree Seed Centre and SADCC Tree Seed Centre
Network), the Canadian Forest Service, the U.S. Forest Service, the Center for International
Forestry Research, and the International Plant Genetic Resources Institute.
We would also like to acknowledge the contribution of the entire Organizing Committee
in Thailand, and particularly the assistance of Ms. Rosita Go of CIFOR, for her work before,
during, and after the symposium, especially in arranging financial support for developing coun-
try scientists, and preparing the manuscripts for publication.
Finally, we would like to thank reviewers of the manuscripts: Md. K. Alam, S. Appanah, P.
Ashton, K. Bawa, K. Boonpragob, W. Brockelman, J. Brouard, N. Byron, K. Chaisurisri, J.
Coles, M. Collins, J. Cornelius, C. Cossalter, J. Davie, L. Duchesne, Y. El-Kassaby, R Finkeldey,
A. Gillison, M. Green, P. Hall, S. Harris, O. Hendrickson, M. Hossain, M. Ibach, H. Joly, P.
Kanowski, M. Kariuki-Larsen, R Leakey, S. Magnussen, E. McKenzie, J. McNeely, D. Meidinger,
C. Nair, F. Ng, I. Nielsen, T. Nieman, M. van Noordwijk, H. Offerman, A.S. Ouedraogo, C.

V
Palmberg Lerche, C. Pielou, B. Ponoy, R. Prabhu, S. Rai, A. Rao, W. Ratnam, U. Rosalina
Wasrin, C. Sastry, J. Sayer, R. Shivas, R Szaro, J. Turnbull, R Uma Shaanker, S. Yatabe, C.
Yeatman, and A. Young.

T.J.B. Boyle
B. Boontawee
(with Dr. Gunnar Seidenfadon), Edible and Poisonous Plants in Thailand, produced by
the Military Research and Development Centre of the Ministry of Defense, and Wild
Plants of Thailand, in the Thai Encyclopedia for Youths, initiated by H.M. King Bhumipol
Adulyadej .
Professor Tern had been working on wild flowers and wild plants from the very
beginning of his career until the end of his life. He acquired an enormous depth of
experience in botany, on his own, by trekking through all the forests and regions of
Thailand. His worldwide reputation meant that any visiting botanist or forest scientist
were always recommended to see him when they visited Thailand.
At the end of his life, he contributed actively to the debate on forest biodiversity.
He was a member of the organizing committee for the IUFRO Symposium on “Mea-
suring and Monitoring Biodiversity in Tropical and Temperate Forests”, held in Chiang
Mai, August 27 - September 2, 1994. He presented the keynote speech at the Sympo-
sium, which is included in these Proceedings, and led participants on a field trip to
examine forest biodiversity at Doi Inthanon National Park.

Vlll
Table of Contents

Preface v
Dedication vii
Keynote address 1

Chapter 1 1
Overview of the status of biodiversity in tropical and temperate forests:
Prof. Tern Smitinand

Principles of Measuring and Monitoring Biodiversity

Chapter 2 5
Biodiversity versus old-style diversity: measuring biodiversity
for conservation:
E. C. Pielou
Chapter 3 19
Measuring and monitoring forest biodiversity: a commentary:
J. Burley and I. Gauld
Chapter 4 47
Landscape characterization and biodiversity research:
V. Dale, H. Offernan, R. Frohn and R. Gardner
Chapter 5 67
Comparative paradigms for biodiversity assessment:
W. Myers, G. Patil and C. Taillie
Chapter 6 87
What do measurements of biodiversity tell us?
G. Namkoong

ix
Chapter 15 227
Gene diversity study based on isozyme analysis in teak (Tectona
grandis L.f.) provenances:
A. Kertadikara and D. Prat

Species and ecosystem diversity

Chapter 16 237
Species diversity of the second growth at Ngao Demonstration Forest,
Lampang Province:
S. Kiratiprayoon, J. Luangjame, P. Damrongtbai and M. Tarumatsawas
Chapter 17 247
Biodiversity assessment of forest tree species in Bia National Park Ghana:
A. Oteng-Yeboah
Chapter 18 257
Measuring and inventorying arthropod diversity in temperate and
tropical forests:
N. Stork
Chapter 19 271
Diversity of microfungi in the wet tropics of north Queensland:
K. D. Hyde
Chapter 20 287
Reserve size and implications for the conservation of biodiversity in the
Andaman Islands:
P. Davidar, S. Devy, T. Yoganand and T. Ganesh

Methodology
Chapter 21 303
Methods for measurement of species diversity:
J. Gajaseni and K. Boonpragob

xi
Chapter 22 309
Tools to diagnose forest integrity: an appraisal method substantiated
by SILW-STAR assessment of diversity and forest structure:
H. Koop, H. Rijksen and J. Wind
Chapter 23 335
A GIS approach to mapping spatial variation in tropical rain forest
biodiversity:
A. Grainger, S. Rose, Trisurat and W. Brockelman
Chapter 24 355
Assessing priority areas for biodiversity and protected areas networks:
C. Margules and I. Kitching
Chapter 25 365
Zapped, scanned, timed and digitized - advanced technologies for
measuring and monitoring vegetation diversity:
G. Lund, D. Evans and D. Linden

Index 383

xii
 Chapter

1
OVERVIEW OF THE STATUS OF BIODIVERSITY
IN TROPICALAND TEMPERATE FORESTS
KEYNOTE ADDRESS
Prof. Dr. Tern Smitinand

Ladies and gentlemen, since the Director General of the Royal Forest Department
has an urgent commitment elsewhere, the organising committee has asked me to
deliver the keynote address in his stead. I will try my best to make it as brief as
possible.

CONCEPT OF BIODIVERSITY
Biodiversity is a term that has recently been widely used. It is the variety and
variability among living organisms and the ecological complexes in which they
occur. It can be defined as the number of different items and their relative fre-
quency, For biological diversity, these items are organised at many levels, ranging
from DNA sequences that are the molecular basis of heredity to complete eco-
systems. Thus, the term encompasses genes, species, ecosystems and their relative
abundance.
Biodiversity can be divided into three hierarchical categories:
- Genetic diversity refers to the variation of genes within species covering dis-
tinct populations of the same species or genetic variation within a population.
- Species diversity refers to the variety of living organisms on earth.
- Ecosystem diversity refers to the variability of habitats and biotic communities
including the variety of ecological processes within ecosystems.

VALUE OF BIODIVERSITY
The diversity manifested by the countless kinds of genetic materials, varied species
and ecosystem types has enormous value. The variety of distinct micro-organisms,
plants, animals and habitats can influence the productivity and services derived
from the ecosystems.
Biodiversity provides direct economic benefits in terms of food, medicine and
industrial raw materials, and supplies the functional ingredients for natural
ecosystems to provide an array of essential services to man (photosynthesis,
regulation of absorption and breakdown of the hydrologic cycle and climate,
absorption and breakdown of pollutants and many others). Plants and animals,
like human beings, have an established right to existence, therefore we should be
concerned with their value and conserve their diversity.

LOSS OF BIODIVERSITY AND ITS CAUSES

Our biological resources are being overused, hampering their regenerative capa-
city. The crucible of extinction is already threatening our natural environment. In
order to save our remaining forests from ultimate destruction, a clear understanding
of the major causes of biodiversity losses is needed. The known manifestations of
the loss of biodiversity are: habitat loss and fragmentation, overharvesting, intro-
duced species, chemical pollution, global climatic changes, and agricultural and
forest industries. The decline of biodiversity is caused by direct and indirect factors.
It is the consequence of use or misuse of the environment by man for his own
purposes.
The loss of tropical forests can have far-reaching effects, including changes in
regional climate, especially rainfall pattern, and biological productivity; accelera-
tion of soil erosion, disruption of watershed stability, and an increase in the global
atmospheric temperature, as well as further impacts on global climate dynamics.

2
In terms of biological diversity, the destruction of primary tropical and temperate
forests threatens quite a number of species with extinction. Their inherent and
aesthetic value, and their potential agricultural, pharmaceutical, and silvicultural
values vanish with them.

GENETIC DIVERSITY
At present the research in genetic diversity of tropical and temperate forests is
carrying on at an accelerating rate, resulting in genetic conservation of economic
species.

SPECIES DIVERSITY
Plants
It is estimated that some 248,000 species of plants have been described world-wide,
with increasing numbers where the current exploration in the tropics is still being
undertaken.

Animals
It is estimated that some 1.4 million animal species have been described world-
wide and follows the same trend as plant species. The diversity of animals is higher
than that of plants, and yet many more are awaiting further study.

ECOSYSTEM DIVERSITY
Terrestrial ecosystems
Terrestrial ecosystems consist of seven main categories, namely: temperate and
tropical evergreen forests, dry and moist mixed deciduous forests, scrub forest,
savana and agricultural land.

Freshwater ecosystem
Freshwater ecosystems can be defined into three major habitats, namely: lakes,
ponds/reservoirs, flood plains.

Marine ecosystems
Marine ecosystems consist of mangrove forests, saline lakes, coral reefs, seagrass
beds, and off-shore ecosystems. These ecosystems play important roles in fisheries
as spawning grounds for marine life and in coastal protection. They are facing the
same trends as freshwater ecosystems.

FUTURE PLANS
Our understanding of the Earth’s biological diversity has significant gaps. The
lack of information hampers our ability to comprehend the magnitude of the loss
of biodiversity and to formulate sustainable alternatives to resource depletion.
To achieve good conservation and management of our natural resources, we
should know the status of our genetic and biological resources. Thus continuous
work and intensive research in the fields of genetic diversity, species diversity and
ecosystem diversity are urgently needed. It is indeed a time to instigate international
collaboration in education, technology transfer, research and last, but not least,
financial support.
This symposium will certainly provide a good opportunity to learn and exchange
ideas and expertise with experts and researchers from different countries and
contribute to international collaboration in measuring and monitoring forest
biodiversity in an effective way in the near future. Thank you.

4
 Chapter

2
BIODIVERSITY VERSUS OLD-STYLE DIVERSITY
MEASURING. BIODIVERSITY
FOR CONSERVATION

E. C. Pielou1

INTRODUCTION
This Symposium is about biodiversity. Right at the outset, it is worthwhile ponder-
ing the difference between two words frequently used by ecologists nowadays:
diversity and biodiversity. This is the distinction, as I see it, between the meanings
of the two words:
Diversity is a concept that was introduced into ecology more than 50 years ago,
and it has been a central topic in theoretical ecology ever since. It is at the heart of
several fundamental problems ecologists are still wrestling with, for example:
does competitive exclusion limit the number of related taxa that can coexist in a
given environment? Is there a limit to the similarity between coexisting species?
And, how finely can “niche-space” be subdivided?
Investigators of these and related problems are usually concerned with
communities that are fairly narrowly defined (in the taxonomic sense), and which
are confined to fairly restricted spaces. For instance, one might want to compare

1 Denman Island, British Columbia, Canada VOR 1TO.

5
communities of cavity-nesting birds in deciduous forest with those in evergreen
forest, or the communities of benthic organisms in flowing water with those in
still water. To facilitate comparisons it is desirable to measure quantitatively an
attribute of each community that can be called its diversity index. The index is
analogous to a variance, it measures the “qualitative variability” of a collection of
objects (e.g., organisms belonging to several species) in the same way that vari-
ance measures the “quantitative variability” of a collection of magnitudes (e.g., the
heights of the trees in a forest).
Biodiversity is an altogether different topic. Though diversity and biodiversity
both deal with variability in a general sense, the scales of variability concerned,
and the reasons for studying it differ enormously. Biodiversity is concerned with
the biota as a whole, and the reason for investigating it is in order to conserve it
as it is in nature, by maintaining extant species in their natural habitats.
The distinction between the two concepts can be summarized thus:
Diversity is a theoretical topic, biodiversity is part of applied ecology. Note,
however, that biodiversity is an attribute of the natural world, one that applied
ecologists are seeking to conserve. To attempt this by artificially increasing the
diversity in local patches of forest (Namkoong 1991) is an aberration.
Diversity is value-neutral, whereas biodiversity weights (or values) different
items according to their rarity. For example, consider two cavity nesting birds of
the forests of northeastern North America: the eastern bluebird, an indigenous rari-
ty, and the starling, an introduced pest that excludes other birds from many of the
available nesting cavities. A student of diversity would treat these species as
equal. A student of biodiversity, concerned to conserve species and prevent
extinctions, would attach a high value to bluebirds, and a very low, possibly nega-
tive, value to starlings.
Diversity is a property of taxonomically narrowly defined groups of organi-
sms, biodiversity is a property of the whole biosphere. For example, a student of
the diversity of shorebirds and their diets would not put several species of birds
and the organisms they eat (insects, crustaceans, and the like) all on a single list to
be treated as a unit. But a conservationist wishes to treat the whole biosphere as a
unit and finds it an unfortunate necessity, not a theoretically desirable simplifica-
tion that, for practical reasons, the biosphere has to be studied piecemeal, generally
one ecosystem at a time.

6
MEASURES OF DIVERSITY AND BIODIVERSITY: THE CONTRASTS
Species Richness
The simplest measure of both diversity and biodiversity, in spite of the great
conceptual contrast between them, is S, the number of species in the community
or ecosystem concerned. This is the species richness. Its value is arrived at merely
by counting. Admittedly, statistical problems arise when the community of inte-
rest is too large to census in its entirety; then S must be estimated by extrapolation.
For a student of biodiversity, concerned with all the organisms existing in an
ecosystem, the counting of species is wholly impossible. It is often useful to treat
some well-defined, manageable taxon, for instance all vascular plants, as a surro-
gate for “all life.” In such a case, one often wishes to do more than merely count
species. It may be desirable to weight them, or equivalently, attach values to them.
In the context of plant biodiversity, Usher and Pineda (1991) have recommended
giving greatest weight to native species growing in their characteristic habitats,
or “quality species”. Lower weights are given to native species not in their
characteristic habitats, and still lower weights (zero, perhaps) to introduced spe-
cies. Alternatively, in a region with a well-known flora, species’ weights could be
made inversely proportional to some measure of their rarity, or of the extent of
their geographic ranges. Many possible weighting schemes can be devised, and
the whole topic deserves the attention of research workers.
It is worth noting, and dismissing, the criticism sometimes made that weigh-
ting species is somehow “unscientific” (Walters and Ramsay 1992). The criticism
applies to theoretical studies of diversity, but is wholly inappropriate in the con-
text of biodiversity and its conservation.

Conservation of Species-poor Ecosys terns

All ecologists are aware of the urgent need to conserve the world’s natural
biodiversity. It should not be necessary (though unfortunately it is) to emphasize
an obvious fact: that the concern is, or should be, with biodiversity as opposed to
diversity. The “worthiness” of a tract of land from the conservation point of view
cannot be judged by considering the tract in isolation and comparing its species
richness with that of other tracts. Such an approach leads to absurd consequences,
such as the protection of tracts rich with ubiquitous species at the expense of species-

7
poor tracts where uncommon species of limited range or unusual habitat
requirements are found.
Magurran (1988) has rightly argued that in choosing lands to protect, "[a]ssessing
sites by diversity alone can be misleading.” It is worse than misleading - it is
counterproductive. Biodiversity, not diversity, is the point at issue. Colwell and
Coddington (1994) likewise emphasize the distinction, they discuss measures of
complementarity - the degree to which a tract being considered for preservation
complements other tracts instead of merely duplicating them.

Diversity Indices
Numerous diversity indices based on measures (or estimates) of the quantities of
the different species in each sample plot have been devised. The two best known
are Shannon’s Index, and various inverse functions of Simpson’s Index of
Concentration (the converse of diversity). They are wholly unsuitable for measuring
biodiversity. This becomes clear from consideration of the following points:
- Biodiversity studies need rapidly gathered data from as many plots as possi-
ble, over large areas. This requires presence/absence (binary) data, not detailed,
time-consuming measures of quantity.
- Biodiversity deals with organisms ranging in size from trees to bacteria, ma-
king comparable quantitative measurements on such different scales is
impracticable.
- Use of binary data overcomes difficulties with quantitative data, such as the
size of herbaceous plants changing rapidly with the seasons.
- In many animals (e.g., insects), population sizes fluctuate enormously from year
to year. Binary data are much less affected than quantitative data by such
quantitative variability.

SPECIES LISTS AND CHECK LISTS: THE PROBLEM OF MONITORING

Check lists
This Symposium is about “measuring and monitoring biodiversity.” In my opinion,
these activities are separate: measuring is not a necessary preliminary to monito-
ring. This is fortunate, considering the tremendous disagreements among ecolo-

8
gists on how measuring should be done. Some ecologists believe that the over-
riding criterion for a biodiversity measure is that it be quick, cheap and easy to
obtain. Certainly monitoring should be quick, cheap and easy, but measuring need
not be. Keeping monitoring separate from measuring makes our problems much
easier to solve.
In exploited forests, the components of biodiversity most at risk are those spe-
cies (of all taxa) characteristic of late successional forests, in a word, old growth.
Conservation can be called successful if it ensures the survival of old growth species.
Species associated with earlier successional stages are at much less risk, and are
rarely in need of special protection. It is nonsense to suppose that biodiversity
benefits from converting a natural forest to a patchwork of plantations of different
ages; or that removing old growth and then tinkering with the genetic diversities of
the replacement trees, as can easily be done, is an exercise in conservation
(Namkoong 1991).
Monitoring should therefore be based on an assessment of the degree to which
old growth attributes are retained in an exploited forest. This is readily done by
drawing up a check list of species and objects typical of old-growth forest. In the
context of the temperate rainforest of British Columbia, the species listed might be
a few herbaceous angiosperms and a larger number of bryophytes associated
with old growth, several lichen species, some cavity-nesting bird species, and seve-
ral amphibians (mostly salamanders). The objects would be such things as stan-
ding dead trees (“snags”), which are known to provide habitats for many specialist
species, logs in different stages of decay, upturned root discs of fallen trees, streams
with deep, shady pools overhung by tree roots, and so on.
It would be straightforward to train people to inspect forest plots and check off
the items on the check list. The items chosen should be easy to identify, and most of
them should be fairly common. Rarities should not be omitted, but they cannot be
depended upon to appear. Checking would be done before and after commercial
forest operations. It would be the responsibility of trained ecologists in charge of a
monitoring program to compile a suitable check list, assign scores to the items listed,
and decide how large a drop between “before” and “after” scores would be
permissible without penalty.

Biodiversity measures
Whereas a check list approach yields both names and a number (the “score”), a

9
biodiversity measure by itself is merely a number. It cannot be too strongly
emphasized that a comparison of biodiversity indices, however computed, is
altogether useless without lists of the species concerned.
This obvious fact is often overlooked, because species lists may not be needed
when diversity comparisons are made in theoretical work. For instance, a student
of niche widths in wood-boring insects might be interested in comparing the
wood-borers of deciduous forests in Europe with those of eucalypt forests in
Australia. Knowledge of the respective diversity indices would be relevant, since
the interest lies in the way each group of insects partitions the available micro-
habitats within and between trees. The specific identity of the insects from the two
regions would be of comparatively minor interest in this context.
In the context of conservation, however, species lists are obviously necessary.
To compare indices derived from two species lists without naming the species
makes as much sense as comparing the weights of two locked boxes without no-
ting that one contains diamonds, the other firewood.

MEASURING BIODIVERSITY
It is unreasonable to search for one single index to measure as complicated a proper-
ty of the biosphere as its biodiversity. Consider a parallel - Physicists describing a
mechanical system’s properties define and measure several separate attributes of
each component:mass, velocity, acceleration, force, work, momentum, and the
like. An ecosystem is many times more complicated than a mechanical system. To
summarize its properties quantitatively will surely require the definition and
measurement of several different aspects of the system. Attempts to single out
one measure as paramount in some way seem doomed to failure. The following
paragraphs describe two of the many measures that may be useful in particular
contexts.

Biodiversity as Species Richness

To recapitulate: comparisons between diversity indices are meaningful only when
the communities being compared are closely similar in a qualitative sense. The same
is true for any biodiversity “indices” that may be proposed, when whole ecosystems

10
are being compared. That said, the problem remains: how can biodiversity be
measured. To estimate S, the number of species, is obviously the first thing to do.
It is seldom possible to determine S from an exhaustive census: communities of
concern in conservation contexts are usually far too big for this to be possible,
and estimation by extrapolating a cumulative “collector’s curve” is unavoidable.
The topic has recently been reviewed by Colwell and Coddington (1994).

Biodiversity as Qualitative Variance

Another measure of biodiversity is the following (Pielou and Fenger 1993). It ari-
ses from considering biodiversity as analogous to the variance of a body of data.
Suppose plants are used as a surrogate for all living things, and that a species list
is made for each of a sample of plots. Construct an s x n data matrix, where n is the
number of sample plots and s the number of species observed in all plots com-
bined (note that s is the number of species sampled, and will normally be less than
S, the number in the whole population). The (i, j)th element in the matrix is 0 or 1
according to whether species i did or did not occur in plot j. Obtain the SSCP
matrix of the data (sum of squares and cross products) and compute its trace,
Tr(SSCP). This is the within-group dispersion, i.e. the sum of squared distances
between every data point and the centroid of the whole swarm of points or,
equivalently, n times the sum of the variances of the s species’ binary scores. As
has already been noted, biodiversity (and also diversity) could be described as
“qualitative variances.”
The value of Tr(SSCP) depends on s and n. To standardize the measure, bringing
it into the range [0,l], it should be divided by its maximum possible value, which
is ns/4 (proof in Appendix). Then the (estimated) index, denoted by c, becomes

c = 4 Tr(SSCP)/ns.

Estimation of C, the population value of c, is discussed below, as also is the

reason for choosing C and c as appropriate symbols.

11
further advantages. First, the only data required for its calculation are lists of the
species observed in a number of sample plots. But these are the data already collec-
ted, for estimating S. No additional observations are needed.
Second, the measure can be used at any scale. For example, if the area of interest
is small (say 10 ha or less), one might use sampling plots (quadrats) of 1 or 4 m2 and
list all species of vascular plants and bryophytes in each plot; the data would
allow estimation of a “high resolution” value of C. Given an area of 100 ha, one
might sample with plots of 16 to 25 m 2 and list either the genus of all sampled
plants, or else all woody plants identified to species. Given an area of 1000 ha, one
could aim for even lower “resolution,” by using 100 m2 plots and listing trees and
tall shrubs only; at 10000 ha one might list life forms rather than taxonomic units.
And so on. For a given geographic region, it would be worthwhile to standardize
the plot sizes and taxa to be used in estimating C for areas of different extent. The
principle is the same whatever the area, only the degree of resolution changes.
The disadvantage of C, shared by all conceivable measures of complementa-
rity, is that a large sample of plots is needed for its estimation. This is illustrated
below, in the context of an example.

AN EXAMPLE
Studies are in progress on the biodiversity of the forests of Quadra Island, in the
Strait of Georgia, British Columbia. The island is in the Coastal Western Hemlock
Zone (Meidinger and Pojar 1991). Three contrasted sites (Sites 1, 2, and 3), each of
20 ha, have so far been studied. A sample of 50 plots, each of 25 m2 was examined
at each site, and vascular plants species in each plot were listed.
The population value of S (species richness) for each site was estimated using
Chao’s method (Chao 1987 and see Colwell and Coddington 1994). The observed
numbers of species, s, and the 95% confidence intervals for S were:

Site 1: s = 60; S = 59 to 102.

Site 2: s = 58; S = 60 to 128.
Site 3: s = 46; S = 46 to 78.

13
ACKNOWLEDGEMENT
I thank Dr. Jennifer Balke of Denman Island B.C. for allowing me to use her data.

REFERENCES
Chao, A. 1987. Estimating the population size for capture-recapture data with
unequal catchability. Biometrics 43:783-791.
Colwell, R.K. and J.A. Coddington. 1994. Estimating the extent of terrestrial
biodiversity through extrapolation. Phil. Trans. Roy. Soc. Lond. B 345:101-118.
Namkoong, G. 1991. Biodiversity issues in genetics, forestry and ethics. The Forestry
Chronicle 68:438-443.
Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton: Princeton
University Press, Princeton, NJ.
Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. British Columbia
Ministry of Forests, Special Report Series, No.6.
Pielou, E.C. 1977. Mathematical ecology. Wiley, New York.
Pielou, E.C. 1984. The Interpretation of ecological data. Wiley, New York.
Pielou, E.C. and M.A. Fenger (eds.). 1993. Our Living Legacy. Royal British
Columbia Museum, Victoria, B.C.
Usher, M.B. and F.D. Pineda (eds.). 1991. Biological Diversity. Fundacion Ramon
Areces, Madrid.
Walters, C. and L.R. Ramsay (eds.). 1992. Methodology for monitoring wildlife
biodiversity in BC forests. BC Environment, Wildlife Branch, Victoria, B.C.

APPENDIX
Proof that Max[Tr(SSCP)] = ns/4

The s x n data matrix can be represented as a scatter diagram of n points in s-space.

Because the data are binary, the points must be concentrated at the comers of an s-
dimensional hypercube.

Tr(SSCP) is the sum of the distances2 of the data points from their centroid. Its

16
 Chapter

3 . . . I. I. .,. .

MEASURING AND MONITORIN G

FOREST BIODIVERSIT Y

A Commentary

J. Burley1 and I. Gauld2

INTRODUCTION
In this joint presentation we review briefly the theory of biodiversity and the general
principles and problems of its measurement and monitoring. We identify immediate
assessment needs, new research requirements and possible rapid monitoring
methods, illustrating some of these with entomological examples from tropical
forests, especially examples from collaborative research between our own
institutions and several in tropical countries with which we have inter-governmental
agreements. This should set the scene for the remainder of the symposium in which
a large number of exciting papers will address many of these scientific topics in
detail.
It is not possible in this conference to detail the many benefits obtained from
forests. One of us reviewed these at a previous symposium in this room (Burley
1993); in summary, renewable natural resources, especially forests, are conserved

1 Oxford Forestry Institute, South Parks Road, Oxford, UK.

2 Natural History Museum, Cromwell Road, London, UK.

19
for three major sets of purposes: the maintenance of ecological processes and life
support systems; the sustained use of the resources for consumptive and social
benefits; and the conservation of biodiversity in its own right for ethical, moral,
aesthetic and evolutionary reasons. Only if all three of these can be met and
integrated into the socio-political process will true sustainable development be
achieved; to meet them requires quantification of the current status and of changes
in the resources. To ensure political and public support for conservation and wise
use of such resources may require economic evaluation also (see e.g. Flint 1991)
and the various values and benefits of presenting biodiversity were summarized
by Pitelka (1993).
The focus of this symposium is “Measuring and monitoring biodiversity in tropical
and temperate forests” but in this overview we shall concentrate primarily on tropical
forests although we acknowledge that, in temperate countries, many of the rarest
and most endangered species are those that are restricted to mature forests - such
as some large wood-boring beetles and, amongst the Hymenoptera, the wood-
feeding sawflies and the guild of parasitoids associated with them. It is these
cerambycids, orussids and acaenitine ichneumonids that are amongst the most
extinction-prone of all British insects. Indeed orussids, a taxonomically isolated
and evolutionarily bizarre group of parasitic sawflies, have not been found in the
British Isles for more than 140 years (Gauld and Bolton 1988). However, only a
small proportion of the British fauna is restricted to forests and conservation ef-
forts are focused on species-rich anthropogenic habitats such as wetlands, chalk
grasslands, heathlands and highland moor areas maintained for hundreds of years
as subclimactic non-forest communities by human intervention. In many tropical
countries a very much larger proportion of the biodiversity is restricted to forests.
Both these are strong arguments for the urgent assessment of biodiversity from a
landscape or “forested lands” viewpoint.
Comparisons of all but a few groups are difficult but, as one of us (I. Gauld) has
been involved with the arthropod inventory of parts of the Costa Rica biota for a
decade, preliminary comparison can be made. In Britain (area 24.2 million hectares)
there are approximately 6500 species in the insect order Hymenoptera, and less
than 500 of these seem to be restricted to forest areas (Gauld, unpubl.). Indeed
more than one third occur in suburban gardens miles from any significant area of
forest (Owen et al. 1981). In Costa Rica (area 5.1 million hectares) the situation is
apparently rather different, for although many species have been collected in high-
ly disturbed areas, over 12,000 of the 20,000 hymenopterans have only ever been
found in afforested areas.

20
 Whilst we should not ignore temperate forests, it is tropical forests, on a global
scale, that are home for the great majority of our planet’s species of plants and
animals, and it is for this reason that interest in their conservation is so great. Loss
of extensive areas of tropical forest (currently approaching 20 million hectares
annually) will undoubtedly cause the loss of innumerable species and populations
of plants and animals and thus will impoverish the global species and genetic
diversity. This in turn will restrict the standard use of resources and the application
of the emerging gene technologies to an ever-decreasing resource base. Paradoxical-
ly, it is the immense species-richness of tropical forests that severely complicates
any attempt to measure and monitor their biodiversity.
Socioeconomic development, population pressure and an almost insatiable
demand by developed nations for wood products mean that tropical forests will
continue to diminish. As so-called wilderness areas vanish an increasing propor-
tion of the Earth’s plant and animal species will survive only in managed forests.
Successful conservation will depend on balancing many conflicting interests but it
will be important to have both quantitative information about biodiversity and
methods for monitoring changes.
Tropical forests are lost or degraded for several reasons including the following
(Evans 1982): clearance for agriculture; intensive logging; production of firewood
and charcoal; shifting cultivation; urban and industrial expansion; over-grazing
and fodder collection; accidental or deliberate burning; and war damage. In addition
to these, the constituent biodiversity may be decreased as a consequence of the
following (Burley 1994): over-exploitation; fragmentation of habitats; climate change;
pollution; introduction of exotics, some of which may become invasive; and
widespread application of artificially bred material.
Knowledge of the political issues and constraints underlying these reasons is
necessary in allocating research resources and it is vital to define clearly the objects
of biodiversity assessment. We recognize that biodiversity may be assessed for four
reasons:
(a) comprehending ecosystem structure and function (for scientific under-
standing of ecosystems and evolution and as a basis for managing resources
for their life support functions and productivity)
(b) conserving and developing germplasm for breeding and genetic
improvement of planted forests
(c) monitoring the impact of land management interventions and both natural
and anthropogenic environmental changes on biodiversity

21
(d) deciding areas of priority for conservation of biodiversity in its own right
for reasons of ethics, aesthetics, religion, culture, scientific enquiry, or future
production including “biodiversity prospecting” for foods, drugs,
pharmaceuticals, other chemical products and biological control agents.

For such conservation, comparative information between sites will be essential

for allocating priorities to sites for planning new conservation areas and, more
depressingly, for advising land-use planners which areas of forest can be converted
to other uses with least impact on biodiversity. Preliminary baseline assessment
followed by monitoring will be necessary to assess the impact of forest manage-
ment practices on biodiversity (with a view to finding an optimum practice that
results in both timber and non-timber commodity production and the conservation
of biodiversity); it will also be necessary to assess other effects on biodiversity
through time such as global climate change or pollution.

TYPES OF BIODIVERSITY AND CAUSES OF CHANGE

The first two plenary sessions of this symposium will deal with the general con-
cepts of measuring biodiversity and it will become apparent that, historically, several
schools of thought have been interested in different concepts and have developed
different indicators of biodiversity. Taxonomists have been responsible for
inventories of biota in the sense that they are responsible for formally describing
the species present. Ecologists and taxonomists have been concerned mostly with
diversity of both plants and animals and their interactions within and between
ecosystems and landscapes (e.g. alpha-, beta-, and gamma-diversity; Whittaker 1975,
Magurran 1988). For species diversity itself, ecologists, taxonomists and evolu-
tionists share interest in the number of taxonomic groups within a habitat, species-
richness, relative abundance or rarity, the degree of endemism, size and form classes
and trophic levels.
Measures and derived indices for these purposes have been based largely on
species counts or relative abundances and are “value-neutral” (i.e. they do not give
special weightings to rarity or endemicity, for example) so that they have
questionable value for conservation prioritization. The United States Nature
Conservancy, in its National Heritage Programme, while essentially undertaking
inventories of State biological resources, has identified global, regional and local
values for conservation through coarse screening based on rarity of ecosystem types
initially, later progressing to individual species. An outstanding example of the

22
development of conservation prioritization that includes taxonomic, ecological,
biogeographic and socio-economic knowledge is the ranking of all forest reserves
in Ghana by Hawthorne and Musah (1993), to be published also by the International
Union for the Conservation of Nature and Natural Resources, IUCN, in 1994; see in
addition Hawthorne (1995); the “genetic heat index” revealed subtle variation within
and across “hot-spots” and showed that subtle and fine-grained exploitation has
little or no negative influence on those aspects of biodiversity measured.
Forest geneticists and tree breeders, on the other hand, have been more con-
cerned with the infra-specific genetic variation between and within populations of
the tree species themselves. Their concern is for the conservation and use of gene
complexes, genes, alleles and DNA sequences for current or potential future use
and this often includes the need to determine natural breeding systems, including
pollinators and gene flows.
Specialist authors later in the symposium will tend to take a species or species
group approach but we stress the inter-dependence of organisms in an ecosystem
and thus emphasize the ecosystem approach. Further, we emphasize the need to
define precisely the objectives of measuring and monitoring such biodiversity
whether it is to determine responses to human intervention or environmental change,
or to evaluate resources in land use decisions.
The major, potential, environmental changes are global and local climate change
(whether natural or caused by man’s activities) and pollution of soil, air and water.
There is considerable evidence on the effect of these factors on individual plant
growth and on the genetic variation within some species; in addition there is some
evidence for their effects on the species composition of ecosystems. However, more
information is required to be collected in a planned, systematic and comparative
manner that will provide greater visibility of the data, facilitate comparable analyses,
permit widely applicable conclusions, and support the planning of remedial or
conservation efforts.
For forests there is a continuous spectrum of managerial interventions that have
known or hypothetical effects on biodiversity. These include the selective
management of natural forest with thinnings at various intensities, enrichment of
natural forest with desirable species, fragmentation of forest through partial
clearance, total clearance of forest (with natural or artificial regeneration), and the
establishment of plantations and agroforestry combinations. In addition to the
obvious effects of these on the constituent tree species, they have significant
implications for the diversity of other organisms.
Simply counting numbers of species present in an area (always assuming the

23
 taxonomy is robust) may give surprising results. Several studies have shown that
the species-richness of some groups of organisms (such as butterflies) increases
following disturbance in an area. This increase is the result of increasing
heterogeneity of the forest area, permitting the immigration of species characteris-
tic of disturbed habitats. However, the species that are characteristic of disturbed
habitats generally are widespread and opportunistic generalists that may have
potential for utilization but that are in little danger of extinction and thus may not
be of great conservation interest; it is the species characteristic of closed canopy
forests that are usually of higher conservation interest. Thus simply counting the
number of taxa present is not a useful measure of biodiversity for conservation
prioritization. A pristine forest and a seriously degraded site may have identical
numbers of species present but the composition of those species may be quite
different, and the degraded site may have lost many of the rarer, highly restricted
species. This was demonstrated by Daniels et al. (1992) in peninsular India, where
they found that in evergreen forests 77 percent of the 200 bird species present were
closed forest specialists but in teak plantations this percentage had decreased to 37
and in eucalypt plantations to 5. This type of evidence emphasizes the need to
consider functional groups as well as taxic groupings to discriminate better bet-
ween otherwise misleading or uninformative species richness alone.
Working with the leaf litter ant fauna, Belshaw and Bolton (1993) demonstrated
that there was no significant difference in species-richness between the faunas of
primary forest, secondary forest and cocoa plantation. Thus differences in the
species-richness of the ground ant faunas are not necessarily indicative of differences
in woody plant species-richness, nor are they likely to be of much use in monitoring
the effects of human impact on forest ecosystems.
In addition the ownership or tenure of forested land has a potentially important
impact on the likelihood of sustainable management and the conservation of
biodiversity. Different managers will have varying concepts of the importance and
valuation of diversity. Tropical forest management may be conducted by
governments (with or without external donor agency support), by joint ventures
between government and industrial companies, by commercial companies or
individual entrepreneurs, by community forestry activities, or by individual small-
holders. These are compared in Table 1 (from Burley 1994) on the basis of system
complexity, inputs, level of biodiversity, non-market benefits obtained and the
commitment to such benefits, time horizon, and institutional and human resource
needs.

24
PROBLEMS OF EXISTING DATA AND SAMPLING
For the assessment of biodiversity at the ecosystem and species level for conser-
vation, there are several sources of variation that must be considered in designing
sampling procedures for measurement and monitoring.

Temporal effects
Historical data. A great deal of verifiable, specimen-based data about the distribution
of forest biodiversity are available in or from museum or herbarium collections.
Although these collections and published data derived from them are potentially
valuable sources of information, care is needed in interpretation for several reasons.
First, the actual geographical range of many of the organisms is likely to be
seriously under-represented. This is because, at least in the tropics, most small
species, such as insects and other invertebrates, are described from single individuals
or small samples made at one favoured site, and few have subsequently been
recorded from elsewhere. A clear example of this can be seen by comparing two
quite closely related groups of South American parasitic wasps, the Mesosteninae
and the Ophioninae. The former is a group that has been studied in the traditional
way, by taxonomists working with small samples (Townes and Townes 1966), whilst
the latter group has been collected more systematically in very large numbers (Gauld
1988). Only 114 (21 percent) of the 543 described species of mesostenines are known
to occur in more than one country, whilst 91 (76 percent) of the 120 ophionines
occur in two or more countries. Given the general similarity of these two groups of
organisms, there is no reason to believe their species should show any great
differences in range size, and to some extent this is confirmed by the fact that only
39 percent of the ophionines had been recorded from more than one country prior
to Gauld (1988) by Townes and Townes (1966). High levels of endemicity may thus
simply reflect taxonomists’ predilections to visit certain areas.
Second, the specimens on which the information is based are not usually a
random sample of the biota of a region. Taxonomists’ “sampling” of tropical
biodiversity is highly selective, and there is a general tendency to maximise the
species-richness obtained in a collecting foray, favouring rarity or scarcity, rather
than to attempt to sample in any ecologically meaningful way. Thus common
widespread organisms or physically very similar species are usually under-
represented in collections, whilst uncommon organisms are highly sought after,
and thus exceptionally well-represented.
Third, there have been strong biases towards collecting and describing large

25
and showy organisms, such as butterflies and dragonflies whilst, much more
numerous smaller and less attractive species are both undersampled and have
escaped the attention of many taxonomists (Gaston 1991). Any attempt to assess
patterns of biodiversity from taxonomic literature must consider such biases.

Individual growth stage. The numbers, abundance and rarity of individual species
change throughout their life span from plant seed through germinant, seedling,
juvenile, mature and senescent stages (with corresponding phases for animals).
Knowledge of the changes in frequency with age and of general population dynamics
is needed ideally to determine optimum conservation management strategies for
any one species within an ecosystem; however, in real field situations, it is often
difficult to disentangle confounded effects such as inherent seral developmental
growth stage from the results of progressive environmental change. Further, the
time and resources required for such dynamic studies preclude rapid assessment
and monitoring.

Seasonal and annual scale. In addition to these effects of growth time within the life
of one organism, there is a major series of time effects on the presence and sampling
of biodiversity; these include the it-&a-annual seasonal change and the natural varia-
tion between years caused largely by differences in climate and ontogeny but also,
in some areas, by pollution. Further, human decisions on land use rotations or
multiple rotations obviously have major impacts on biodiversity over time; in most
tropical regions there are examples of the human introduction of fruit trees that
subsequently spread naturally into forests. The entire question of invasiveness is
of current concern.
Reproduction in rainforest species is nearly exclusively sexual, yet this process
is greatly influenced by life stage and by environment, flowering con-specifics and
flowering heterospecifics in ways that are little understood. Furthermore, any
measure of biodiversity at a single point in time must take into account problems of
sampling within and between years. Through a collaborative project with the Oxford
Forestry Institute (OFI) and Centro de Pesquisa Agropecuaria do Tropical Umido
(CPATU), Empresa Brasileira de Pesquisa Agropecuaria (EMBRAPA), the dynamics
of reproduction of Amazonian trees is being investigated in order to provide
information for conservation initiatives, representative sampling of intraspecific
genetic variations and sustainable management of natural forests. Analysis of floral
phenology data for 50 species over a ten year period reveals an array of flowering

27
syndromes, including: synchronous annual flowering, supra-annual flowering
(masting), asynchronous supra-annual flowering and bi-annual flowering. Clearly,
one definition of reproduction is not applicable and attempts at modelling
regeneration will need to take into account species differences in density of flower-
ing individuals, pollen dispersal distances, functional gender of individuals,
synchronicity of flowering and alternate food sources (i.e. other species flowering
at the same time) for the generalist biotic pollen vectors.
In lowland tropical habitats with rather little seasonal change in weather pat-
tern, there may be marked differences in the abundance of many insect species
(Wolda and Roubik 1986; Hanson and Gauld, in press). In habitats with pronounced
seasonality, there may be very pronounced differences in abundance of any given
species between one year and the next. This may present a particular problem in
tropical forests where some species of insects exist at levels of abundance that are
so low they are virtually undetectable by a variety of well-established sampling
procedures, then have a large population in a single year. In this situation, if
sampling is restricted to one technique, such a species may only “appear” in a habitat
in an outbreak year. A good example of this was observed recently in Central
American dry forest. Sustained light-trapping over a number of years has shown
that the ichneumonid Enicospilus madrigalae seems to be present in Santa Rosa most
years (Gauld 1988; unpubl.), but more than 80 percent of individuals collected over
a decade were taken in a single year. Although light trapping is a highly efficient
method of collecting nocturnally active wasps, flight interception traps are more
generally used to trap wasps as they are less labour-intensive to use, and require
only about one hour’s attention every fortnight instead of several hours every night.
Examination of interception trap catches in the same area reveals that E. madrigalae
was only taken in the outbreak year, and has never been found either before or
since.

Mobility. A further problem with sampling of highly mobile organisms, and one
that, of course, does not occur for sedentary species, is what does an occurrence at
a site mean? It may be part of a viable population occurring in a particular habitat,
or it may be simply a vagrant that chanced to be caught. Gauld (unpubl.) recently
found a female of the ant Atta cephalotes in a flight interception trap catch from 3100
metres on the Cerro de la Muerte in Costa Rica. Because attine (leafcutter) ants are
such a conspicuous part of the fauna of tropical America it is well known that they
almost never nest above 1500 metres, and certainly will not do so above 3000 metres.

28
Thus it is obvious that the single female encountered is a vagrant, probably blown
up by wind from lower altitudes. Such vagrancy might not be appreciated for
decades if the insect involved were one of the more cryptic, little-studied denizens
of the lowlands. It is interesting to note, that in Costa Rica, the most species-rich
site for Hymenoptera (Hanson and Gauld, in press) is the garden of a roadside
restaurant on Cerro Zurqui ! We consider it unlikely that the restaurant garden
supports such diversity; it is more likely that its high species richness is due to
large numbers of species moving up an adjacent gully that runs through a large
national park. Clearly very mobile species may not be reliable for site biodiversity
assessment but this may vary with species and sampling method.

Community seral stage. There has always been debate about the static and dynamic
aspects of conservation of resources, particularly genetic resources in which
evolutionary change is a natural phenomenon. Whatever the reasons for
conservation, natural vegetation commonly progresses through stages from clear-
ed land through pioneers to “climax” populations in the current environmental
conditions although some fluctuations in composition still occur. Most of the
associated species change as this process occurs and thus sampling methods for
monitoring must take this into account. The poster displayed by one of our
colleagues (D. Sheil) at this symposium illustrates such changes with time over half
a century in the Budongo Forest, Uganda.

Locational scale
Physical position. Different organisms obviously occur in different locations within
the ecosystem from the soil/root interface through the stem to the crown of mature
trees. The sampling procedures must clearly differ between these even for similar
groups of organisms (particularly insects). This seems to be the case for ants, a
group that is particularly suitable for biodiversity studies as they are both abundant
and diverse in most tropical habitats. They may comprise up to a third of the arboreal
arthropod biomass (Fittkau and Klinge 1973) and nearly 90 percent of the indivi-
dual insects encountered in a tropical ecosystem (see e.g. Majer 1993), and there
may be as many as 43 different species in a single tree (Wilson 1987). Sampling of
ants can be undertaken by various means which can be grouped into two methods
- extraction from litter samples using a process involving gradual drying, and
extraction from vegetation by sweep netting, insecticidal fogging (Paarmann and

29
Stork 1987), hand picking, etc. These two types of methods may yield quite different
results. As we have already mentioned, little change was observed in the composi-
tion of leaf litter ant communities in variously disturbed afforested areas in West
Africa (Belshaw and Bolton 1993) but Dejean et al. (1994), also working in West
Africa, observed considerable differences between the arboreal ant mosaics in old
secondary forest and a forest edge composed of fast growing colonizing species.

Geographic scale. The different concepts of alpha-, beta-, and gamma-diversity and of
abundance, rarity and endemicity were referred to above but, even at the ecosystem
level, these indices differ and have different implications for sampling and
conserving biodiversity at the global, regional, national and ecosystem/habitat/
patch levels. What is abundant or apparently stable at one location may be rare or
threatened at a global scale.
In Costa Rica, the Osa peninsula in the extreme southwest is notable in that it
supports small populations of lowland rainforest bird species that are otherwise
only known to occur in the Amazon basin (Stiles and Skutch 1989). Such species
may be very local, scarce and probably endangered on the Osa, and thus in Costa
Rica they may be a major focus of conservation effort, although further south in the
Amazon Basin, such species may be very common. These endangered peripheral
populations may reflect incipient speciation and they may also detract conserva-
tion effort from other areas, such as the rather less exotic high altitude fauna of the
Cordillera de Talamancas, an area of exceptional endemism straddling the Costa
Rican/Panamanian border (Gauld 1988, 1991, Hanson and Gauld, in press).

Interactions between species

Links in the food or pollinator chain. The links in the food chain range from the primary
producers and saprophytes through herbivores and predators to decomposing
micro-organisms. As individual size and/or natural range area increase, so do the
public recognition and valuation of particular species. Concurrently, the scientist
specializing in one species or related group tends to develop intensive recurrent
sampling systems to obtain adequate precision for the group but these may not be
applicable to other species. The inter-dependencies and other inter-actions between
species must be understood to obtain adequate overall sampling to support the
objective of ecosystem maintenance.

30
BIODIVERSITY INDICATORS
It is clear from the above that there is an urgent need for collaboration between
scientists who study many different disciplines and who work on different groups
of species, firstly to seek consensus on the concepts and priorities for biodiversity
assessment, and secondly to identify means of linking disparate assessments or
initiating joint studies.
In addition to direct total or sample counts of species or individuals in a given
area and an understanding of population dynamics, there are six sets of information
that facilitate or add to biodiversity indicators in support of decisions on biodiver-
sity conservation:

Species-urea relationships. These are commonly used by ecologists to determine,

describe or predict species richness within an area and by conservationists to help
estimate minimum population sizes in reserves (see e.g. Soulé 1986, Simberloff 1992).

Keystone species. These species are recognized as playing a major role in maintaining
ecosystem structure and integrity; for a classification of types of keystone species
see Bond (1993). There has been considerable debate about the concept and reality
of keystone species but there seems little doubt that in some cases a species or spe-
cies group may have a major role in the survival of the ecosystem as currently
recognized, e.g. the fruits of figs are a fundamental resource for primates and many
frugivorous birds that themselves ensure the perpetuation of the Budongo Forest
(yet for decades foresters wished to eradicate figs for silvicultural reasons).

Ecological indicator species. These species are adapted to (or react characteristically
to) changes in specific environmental factors, or their diversity appears to be
correlated with that of several or many other species. The nymphs of some groups
of aquatic arthropods (e.g. Plecoptera and Odonata) are used for river water quality
assessment in the UK. Some authors (e.g. Kremen et al. 1993) have advocated using
terrestrial arthropods as an easily sampled “indicator” group and, although some
groups of arthropods do seem to be useful indicator groups (Klein 1989, Brown
1991), one common problem with measuring or monitoring such speciose taxa is
obtaining representative samples. Samples collected over fairly short periods often
comprise very large numbers of individuals with few represented by more than
one or two specimens. Longer term sampling, such as has been undertaken in Costa
Rica as part of the fauna1 survey of Costa Rican Hymenoptera (Hanson and Gauld
1995), shows that in the short term samples of arthropods from sites can be very

31
different, but species-accumulation curves at these sites continue to rise very steep-
ly, and over a period of two years the species-composition of the sites becomes
increasingly similar. Comparison of species-accumulation curves for individual
sites and for the region show, at the regional level an asymptote being approached
after about 18 months malaise trap continuous sampling, while at any one site the
species accumulation curve continues to rise steeply, suggesting increasing sam-
pling effort at any one site will yield an ever greater proportion of a regional fauna
at that site. Thus more intensive sampling will result in the samples from the sites
having an ever more similar species composition, and apparent differences in species
composition between sites, revealed by short term sampling, are perhaps only
artefacts resulting from insufficient sampling effort.
Another strategy for measuring and monitoring biodiversity involves work-
ing with very well-known and well-characterized groups of organisms such as birds
or mammals. With such well-known groups areas of endemicity can easily be
recognized as can areas of high species richness. Obviously, for conservation
planning to be most easy and effective, it would be convenient if two postulates
were true - i.e. habitats that are species-rich for one taxon are species-rich for other
taxa; and rare species occur in species-rich habitats. These postulates have often
been assumed by conservationists working in tropical habitats but have rarely been
tested. One of the most detailed studies on this question was conducted in a
temperate country, the UK, by Prendergast et al. (1993) looking at the distribution
of birds, butterflies, dragon flies, liverworts and aquatic angiosperms. They found
little support for either proposition: species-rich areas (so-called hotspots) frequent-
ly do not coincide for different taxa, and many rare species do not occur in the
most species-rich sample areas.
Few comparable analyses have been undertaken in the tropics, although in India
it has been shown that high bird species-richness does not necessarily correlate
with high species-richness in other fairly well-known groups (although it frequent-
ly correlates well with ecosystem structural diversity). For example, Daniels et al.
(1992) demonstrated that, in the Uttara Kannada district of southern India, bird
species-richness is inversely correlated with woody plant species diversity. In this
particular case Daniels et al. were able to suggest plausible reasons for such an
apparent anomaly; they suggested that this inverse relationship may be explained
by the fact that although the peninsular Indian evergreen forests are rich in woody
plant species when compared with drier surrounding areas, they harbour an
impoverished bird fauna due to their small overall extent and great isolation from
other extensive tracts of wet forests. Thus the evergreen forests rich in plant species

32
have a smaller pool of potential colonizing species than the relatively plant species-
poor dry forests. Whatever the causes of this avian impoverishment, it is well
established that it is erroneous to assume simply that maximum bird species richness
is indicative of maximum richness in biodiversity; in such comparisons it is axiom-
atic that niche size is important. Ecologists are seeking better indicators; for example,
four Danish institutions are working on montane biota in southern Africa and South
America using DNA-based population phylogenies to supplement traditional
biogeographic methods for detailed interpretation of the diversity in selected groups
and the age of endemicity (Fjeldsa 1992).
The distribution of few tropical species is well enough known to allow
comparison of species-richness with rarity. However, two observations are perti-
nent here. First, whilst there is a general decline in species-richness in most insect
groups above about 1000 metres in tropical habitats, there is a general increase in
endemicity with altitude. Gauld and Mitchell (1981) observed that whilst less than
30 percent of a sample of ichneumonids collected in lowland forests in New
Guinea were endemic to the island, at high altitudes (above about 2000m) more
than 90 percent of the species were endemics.
Second, the intensive Hymenoptera survey of Costa Rica, carried out over the
past decade (see Hanson and Gauld, in press) has clearly shown that some of the
very rarest of all Hymenoptera occur in rather species-poor areas - areas that are
very seldom collected in because they are regarded as unproductive. The most
striking examples of this include members of the extraordinarily rare
Hymenoptera groups Bradynobaenidae, Scolebythidae, Sierolomorphidae and
brachycistidine Tiphiidae, all of which are found in species-poor, open, dry areas
during the dry season.

Taxic groups. Groups of species and/or higher taxa, as accepted by current taxono-
mic agreement, offer a straightforward method of comparing sites and ecosystems
for their diversity and conservation status. More recently methods have been
developed for assigning conservation priorities not just on species richness but by
incorporating measures of taxonomic distinctness of the species concerned (e.g.
Vane-Wright et al. 1991, Faith 1992, 1994). Areas inhabited by groups of phylo-
genetically distantly related species have been accorded higher conservation priori-
ty than similar areas occupied by an approximately equal number of phylo-
genetically very closely related species.

Functional groups. These comprise groups of species fulfilling the same function

33
and having similar morphological structure within a ecosystem; for example,
lianes may be assessed as a group without identification of individual species.

Economic valuation species. While ecologists, taxonomists and geneticists have a range
of indicators and assessment methods for biodiversity, land use or conservation
decisions are commonly based on over-riding economic valuations. These are species
of known or potential economic value per se or that occur in assemblages that have
other individually valuable species. In the Budongo Forest of Uganda Khaya and
other mahogany timbers in the Meliaceae are found in the same formations and
have similar ecological requirements. The economic weighting of such productive
species, or of habitats that have non-wood products, amenity and touristic value
(such as drug plants, gorilla forests), frequently over-ride ecological and genetic
weightings in current political and financial climates.

ASSESSMENT METHODS
Whatever the index or indicator desired, and whatever scale or sampling intensity
is sought, a range of methods may be more or less applicable and more or less
expensive:

Traditional forest inventory and vegetation analysis . Foresters have historically developed
the science of forest inventory, principally for estimates of standing volumes of
wood in forests and for recurrent measurements to indicate changes with time or
management; the numbers and densities of non-wood plant species are occasional-
ly recorded but not as a principal object of the survey nor in an internationally
comparable standard system. Systems of permanent and temporary sample plots
in forests have sometimes been established for these forestry purposes (see e.g.
Adlard 1990, Husch 1971). Substantial work in Australia by the Commonwealth
Scientific and Industrial Research Organization (CSIRO) has expanded such
traditional forest inventories into multi-taxa surveys. There has also been consi-
derable international activity recently to establish biodiversity monitoring plots
e.g. United Nations Educational, Scientific and Cultural Organization (UNESCO)
Man and the Biosphere Programme; Smithsonian Institution, Washington; and the
Food and Agriculture Organization of the United Nations (FAO).
Several standard textbooks of vegetation analysis exist and one of our Oxford
colleagues (D. Sheil) is currently preparing a guide to field assessment for conser-
vation and biodiversity research in East Africa; he will also be presenting a poster

34
in this symposium on the evaluation of long-term change in permanent plots within
the Budongo Forest of Uganda. Ideally insect and other animal assessments should
be undertaken in such plots and another Oxford colleague, A. Plumptre, has shown
that primate densities increase in Budongo after selective logging.
A large number of quantitative biodiversity indices have been developed
based on counts or measurements of standing trees or trap samples of animals
(see e.g. Magurran 1988, Pielou 1975, Solbrig 1991, or Whittaker 1975) and many of
these will be reviewed or used during this symposium. (Pielou prefers to use diversity
indices to biodiversity indices and stresses that these are not suited for monitoring.)
Many of the published indices do not specifically value rarity and may thus have
less significance for conservation; the National Heritage Program of the United
States Nature Conservancy identifies areas of vulnerable and rare species to
maximize biodiversity on a global as well as a local scale. On a different level, Reid
et al. (1993) have shown the applicability of these and other measures as indicators
of biodiversity conservation that would have political and administrative
applications (summarized in Table 2) while the Danish studies referred to above
seek to identify major locations of species endemism.

Molecular methods. The last decade has seen an increasing use of high profile mole-
cular genetic techniques to study genetic diversity, systematics and population
genetics at the DNA and protein levels. These technologies have included isozyme,
restriction fragment length polymorphism (RFLPs), randomly amplified
polymorphic DNA (RAPD), DNA fingerprinting and, more recently micro-satellites.
These, and the techniques of secondary product analysis (e.g. Hanover 1992), are
now relatively commonplace and have been tested in a wide range of taxa, providing
powerful methods for: (i) the inference of phylogenies (e.g. Harris et al. 1994); (ii)
the identification of hybridisation (e.g. Carr et al. 1986, Keim et al. 1989); (iii) the
assessment of genetic diversity (e.g. Lavin et al. 1991, Chalmers et al. 1992, Loveless
1992, Bardakci and Skibinski 1994) and (iv) the potential to explore the interaction
between genome adaptation and ecotypic variation (e.g. Rieseberg et al. 1993).

Remote sensing. A large array of technologies now exists for examination of terres-
trial resources including aerial photography and satellite imagery in various
electromagnetic wavebands. Their scales and precision differ but the locations and
changes in forest or ecosystem boundaries can be identified easily and the stand-

35
ing volume of wood in some forest types can be estimated reasonably precisely. If
coupled to appropriately detailed ground truthing, the techniques are applicable
to identifying rare communities and vulnerable remnants, for mapping vegetation
and for zonation and land use planning (e.g. Mount Elgon, Uganda). However,
none of these techniques are yet refined sufficiently to identify individual plants
unequivocally at a scale and precision that would permit biodiversity monitoring
within ecosystems.

Databases and geographic information systems. All of the historical and current data
collected by any of these technologies may now be combined into electronic
databases and portrayed by a large number of geographic information systems
that are commercially available. A session of this symposium will hear of applications
of several of these. Together with the growth in databases, there is a need to develop
analytical techniques to highlight areas of high species richness or high conser-
vation priority (see e.g. ICBP 1992, WCMC 1992). Amongst the techniques currently
being developed are the WORLDMAP programme (Vane-Wright et al. 1991,
Williams et al. 1993) currently under development at the Natural History Museum.
One of the most intensive and extensive database systems now available appears
to be the Biological and Conservation Database (BCD), of the United States Nature
Conservancy (see Carr 1988); this allows rare taxa to be identified and critical sites
to be recognized. An excellent example of a currently operational management tool
is “FROGGIE”, the database-GIS system developed by Hawthorne and Musah
(1993) for the forest reserves of Ghana; this allows forest managers to act upon
research data with all knowledge of the distribution of all species in the forest area
(and it is now being linked to other OFI software including “BRAHMS” and
"SISTEM+" for the incorporation of data from herbaria and field experiments). It
must be recognized continually that the success of such research is dependent on
the careful, repetitive and tedious recording and management of essentially simple,
primary, “key” data; standardization and agreement on parameters to measure must
not preclude the addition of other observations for specific purposes.

CURRENT NEEDS
There appear to be three main groups of requirements relating to current know-
ledge, policies and practices for biodiversity assessment and conservation. These

37
are immediate assessment needs, new research required, and evaluation or
implementation of rapid monitoring techniques; they are summarized below.
Underlying all of them is a great need for unequivocal taxonomies and for guides
to species identification (see e.g. the field guide to trees of Ghana by Hawthorne
1990).

Immediate assessment needs

Methods
- define key data
- develop standardized sampling protocols for baseline studies and later
monitoring for defined objectives
Organisms
agree internationally on focal groups of organisms (whether taxonomic,
locational or service groups) that can be assessed easily and relatively cheaply,
and that can indicate biodiversity in other organisms for stated objectives
resolve current taxonomic uncertainties in these groups and establish adequate
reference material
produce simple taxonomic/identification manuals and computer-based
identification systems
undertake baseline studies for the focal groups nationally

Data management
- develop database and information systems to accept and process information
from a wide range of historical and current sources, including knowledge of
generalist and specialist species derived from earlier ecological research
- review the holdings of herbaria and museums
- review and interpret significance of historical records of forest management
- review value of existing permanent forest sample plots and long-term ecological
monitoring plots; establish new plots or transects if required

New research requirements

- prioritize the information required for the four major objectives and at the three
main geographic levels

38
 refine and validate baseline surveys of focal groups and extend to other groups
to begin development of monitoring models
determine which groups are sensitive to environmental and managerial change;
consider them for use as indicators; check the rarity of species and ecosystems
conduct ecological studies to understand principal linkages and propagation
systems, and to determine whether keystone species exist
use molecular methods on samples from the extensive networks of international
population (provenance) trials of forest trees to examine intra-specific diversity,
provided that adequate information is available on the original source and
subsequent management; if possible these should be compared with samples
from the original natural populations to detect change following transfer to exotic
conditions. It is critical that material from the whole range be analysed (for
example, the OFI study of the isozymes of Faidherbia albida)
determine the correlation between conservation of species richness, species rarity
and intra-specific variation within given species and the correlation between
different species (in the different regions of the world)
establish regional and national databases and geographic information systems
to summarize, display, digest and interpret information on national biodiversity
for land use managers and policy makers

Rapid monitoring methods

define and justify what is to be monitored
involve local human populations and indigenous knowledge in recording of
species occurrence, distribution and use
progressively increase sampling proportion among permanent sample plots until
acceptable accuracy is achieved
use molecular sampling methods to determine intra-specific variation of focal
groups; this will require resolution of the debate over the b‘ est method
use remote sensing to detect ecosystem boundary changes and some structural
changes, plus geographic information systems to portray all levels of biodiversity
currently known
link monitoring data to forest management and to subsequent model building.

39
EXAMPLES FROM THE UK-MALAYSIA COLLABORATIVE PROJECT
Confronted by the problem of designing a project to measure the impact on
biodiversity of intervention practices one thus has to address the question “what
should be monitored?” with some caution, as it is obviously possible to register
loss, gain or no change in biodiversity depending on what group is monitored, or
to fail to obtain a sample of sufficient size to have any meaning whatsoever. This
dilemma had to be addressed when, under two recent Intergovernmental
Memoranda of Understanding, organizations in Malaysia (including the Forest
Research Institute of Malaysia and several Universities) and the United Kingdom
(the Natural History Museum, the Natural Resources Institute, the Oxford Forestry
Institute and several Universities) agreed to collaborate in undertaking pro-
grammes to study the effects of human intervention practices on biodiversity.
Experimental sites were selected with good historical records of site factors and
management, particularly the Pasoh Forest Reserve in Peninsular Malaysia.
To address the question of what to monitor, it was decided to try and encom-
pass some of the biotic variation present in the site. To help achieve this we looked
at components of this variation and recognized several including taxic, genetic,
trophic and positional variation together with some that reflected variation in
economic importance.

Taxic variation. This refers to the entire taxonomic spectrum of variation present in
the biota. It focuses on the variation of species richness within taxonomic rank (e.g.
monotypic or polytypic) and is used to weight species sets when assigning
conservation priorities. This encompasses a broad spectrum from prokaryotes to
higher plants and the charismatic megavertebrate fauna. It was felt that different
components of the taxic spectrum might respond in different ways and, because of
their different mean generational times, with different speed. We hoped that by
using a range of organisms it would prove possible to identify groups that have.
different degrees of sensitivity to a range of human interventions such as logging.
The groups included mycorrhizal fungi, epiphytic cryptogams including lichens,
climbing plants, shrubs, palms, Zingiberaceae (for their chemical interest),
Dipterocarpaceae (for their commercial wood interest), ants, birds and small
mammals; these mixed (a) taxonomic groups with functional groups and (b)
individual with community indicators.

Genetic variation. This implies the level of genetic variation within a particular species.
In practice this is both more expensive and labour intensive to monitor than some

40
other axes of variation, but it is both interesting and important to address questions
concerning the change in genetic diversity occurring in a population of forest trees
following selective logging. For example, does this select against characteristics
desirable in timber yielding trees (as these are the ones removed) or does it favour
them (as they have already seeded, are present in the understorey and grow
immediately to fill gaps created by logging)?

Trophic variation. It is well known in temperate ecosystems that perturbations may

affect organisms higher in the food chain more severely than the more numerous
organisms lower in the food chain. Thus we attempted to select organisms with
different positions in the food chain, primary producers, herbivores, carnivores,
parasites, saprophytes and so on in an attempt to embrace this spectrum and to
cover other levels of functional variation, guilds etc.

Positional variation. Many organisms are normally restricted to a certain position or

site in an ecosystem and intervention may have a greater impact on some of these.
For example, the high degree of similarity between leaf litter ant communities in
pristine habitats and in a range of progressively increasingly disturbed habitats
suggested that leaf litter communities may not be very sensitive to perturbation
(Belshaw and Bolton, 1993). However, as intervention involves removal of many
large trees it may significantly reduce the canopy, and consequently have a
deleterious effect on the canopy community of ants. By monitoring a range of
organisms occurring in the soil, leaf-litter, under bark and in the canopy it was
hoped that variation in this spectrum could be accommodated although it was
recognized that this was difficult to test.
Other axes of variation relating to different aspects of behaviour and/or habitat
requirement could have been addressed, but in the budgetary constraints of the
programme it was only possible to select groups of organisms that embraced the
variation outlined. This then is one attempt to answer the question of what should
be monitored. We hope that the discussions at this symposium will provide some
insights into the broad range of problems of measuring and monitoring forest
biodiversity.

CONCLUSION
It is somewhat difficult to preview the results of a symposium without having seen

41
all the contributed papers and without having heard all the discussions. However,
we hope that the symposium will result in some approach towards consensus on
the needs for future research at a time when financial resources are limited and
competition is increasing. Biodiversity clearly means different things to different
people and it is affected in various ways by the several purposes for which resources
are managed and at different scales from local to global.
Knowledge of taxonomy, reproductive biology and genetic systems is clearly
fundamental. We need to determine how conservation values should be assessed
and to identify what types of information are required at the different geographic
scales. We must also agree on the types of indicators that may be appropriate to
determine adequate biodiversity conservation and management (e.g. trees, vascu-
lar plants, habitat varieties, viable habitat sizes, level of habitat protection, or
existence of corridors to other sites).

REFERENCES
Adlard, P.G. 1990. Procedures for monitoring tree growth and site change. Tropical
Forestry Paper 23, Oxford Forestry Institute, University of Oxford, UK, 188p.
Bardakci, F. and D.O.F. Skibinski. 1994. Application of the RAPD technique in tilapia
fish: species and subspecies identification. Heredity 73, 117-123.
Belshaw, R. and B. Bolton. 1993. The effect of forest disturbance on the leaf litter ant
fauna in Ghana. Biodiversity and Conservation 2, 656-666.
Bond, W.J. 1993. Keystone species. In: Schultze, E-D and Mooney H.A. (eds.).
Biodiversity and ecosystem function. Ecological studies Vo1.99, Springer-Verlag,
Berlin, Germany, 237-253.
Brown, K.S. 1991. Conservation of neotropical environments: insects as indicators.
Pages: 349-404 In: Collins, N.M. and Thomas, J.A. (eds.). The Conservation of
Insects and their Habitats. Academic Press, London, UK.
Burley J. 1993. Balance between development and genetic conservation. Pages: 17-
25 In: Proceedings of the international symposium on genetic conservation and
production of tropical forest tree seed, Chiang Mai, Thailand. ASEAN-Canada
Forest Tree Seed Centre, Muak-Lek, Thailand.
Burley, J. 1994. Biodiversity in development and conservation. Invited Paper, British
Council - Overseas Development Administration Seminar on “Forest land use
options: conflicts and solutions”, Kumasi, Ghana, 15p.
Carr, K.M. 1988. The BCD system - the next generation of heritage computer systems.
Reprinted from Biodiversity Network News, The Nature Conservancy,

42
 Arlington, Virginia, USA, 6pp.
Carr, S.M., S.W. Ballinger, J.N. Derr, L.H. Blankenship and J.W. Bickham. 1986.
Mitochondrial DNA analysis of hybridisation between sympatric white-tailed
deer and mule deer in west Texas. Proc. Natl. Acad. Sci. USA 83, 9576-9580.
Chalmers, K.J., R. Waugh, J.I. Sprent, A. J. Simons and W. Powell. 1992. Detection of
genetic variation between and within populations of Gliricidia sepium and G.
maculata using RAPD markers. Heredity 69, 465-472.
Daniels, R.J.R., N.V. Joshi and M. Gadgil. 1992. On the relationship between bird
and woody plant species diversity in the Uttara Kannada district of south India.
Pages: 5311-5315 In: Proceedings of the National Academy of Sciences of the
USA, Volume 89.
Dejean, A., A. Akoa, C. Djieto-Lordon and A. Lenoir. 1994. Mosaic ant territories in
the African secondary rain forest (Hymenoptera: Formicidae). Sociobiology 23,
275-292.
Evans, J. 1982. Plantation forestry in the tropics. 2nd edition. Clarendon Press,
Oxford, UK, 403p.
Faith, D.P. 1992. Conservation evaluation and phylogenetic diversity. Biological
Conservation 61: l-10.
Faith, D.P. 1994. Phylogenetic pattern and the quantification of organismal
biodiversity. Philosophical Transactions of the Royal Society of London B, 345:
45-58.
Fittkau, E.J. and H. Klinge. 1973. On biomass and trophic structure of the Central
Amazonian rainforest ecosystem. Biotropica 5, 2-14.
Fjeldsa, J. 1992. Patterns of endemism in the tropics: testing competing models by
integrating biogeographic and molecular data. In: Conference Programme and
Abstracts, “Conservation of biodiversity in Africa”, National Museums of Kenya,
Nairobi, Kenya.
Flint, M. 1991. Biological diversity and developing countries. Issues and options.
Overseas Development Administration, London, UK, 50p.
Gaston, K.J. 1991. Body size and probability of description: the beetle fauna of
Britain. Ecological Entomology 16, 505-508.
Gauld, I.D. 1988. A survey of the Ophioninae (Hymenoptera: Ichneumonidae) of
tropical Mesoamerica with special reference to the fauna of Costa Rica. Bulletin
of the British Museum (Natural History) Entomology 57, 1-309.
Gauld, I.D. 1991. The Ichneumonidae of Costa Rica, 1. Memoirs of the American
Entomological Institute 47, 1-589.

43
Gauld, I.D. and B. Bolton. 1988. The Hymenoptera. Oxford University Press,
Oxford, UK, 332p.
Gauld, I.D. and P. Mitchell. 1981. The taxonomy, distribution and host preferences
of Indo-Papuan parasitic wasps of the subfamily Ophioninae (Hymenoptera:
Ichneumonidae). CAB, Slough, UK, 611p.
Hanover, J.W. 1992. Applications of terpene analysis in forest genetics. New Forests
6, 159-168.
Hanson, P.E. and I.D. Gauld (eds.). 1995. The Hymenoptera of Costa Rica. Oxford
University Press, Oxford. (In press).
Harris, S.A., C.E Hughes,. R. Ingram and R.J. Abbott. 1994. A phylogenetic analysis
of Leucaena (Leguminosae: Mimosoideae). Pl. Syst. Evol. 191, 1-26.
Hawthorne, W. 1990. Field guide to the forest trees of Ghana. Natural Resources
Institute, Chatham, UK, 276p.
Hawthorne, W.D. 1995. Regeneration, scale and sustainable use: holes and the sums
of parts in Ghanaian forest. In: Swaine, M.D., Alexander, I.J. and Watling, R.
(eds.). Studies in GuineoCongolean rain forests. Royal Society of Edinburgh,
Edinburgh, UK. (In press).
Hawthorne, W. and A.J. Musah. 1993. Forest protection in Ghana. Forest Inventory
and Management Project, Forestry Department, Kumasi, Ghana, 64p + 139p
appendices. (IUCN, in press.)
Husch, B. 1971. Planning a forest inventory. Forestry and Forest Products Paper 17,
FAO, Rome, Italy, 121p.
ICBP. 1992. Putting biodiversity on the map: priority areas for conservation.
International Council for Bird Preservation, Cambridge, UK, 90p.
Keim, P., K.N. Paige, T.G. Whitham and K.G. Lark. 1989. Genetic analysis of an
interspecific hybrid swarm of Populus: occurrence of unidirectional
introgression. Genetics 123, 557-565.
Klein, G.C. 1989. The effects of forest fragmentation on dung and carrion beetle
communities in central Amazonia. Ecology 70, 1715-1725.
Kremen, C., R.K. Colwell, T.L. Erwin, D.D. Murphy, R.F. Noss and M.A. Sanjayan.
1993. Terrestrial arthropod assemblages: their use in conservation planning.
Conservation Biology 7, 796-808.
Lavin, M., S. Mathews, and C. Hughes. 1991. Chloroplast DNA variation in Gliricidia
sepium (Leguminosae): Intraspecific phylogeny and tokogeny. Am. J. Bot. 78,
1576-1585.
Loveless, M.D. 1992. Isozyme variation in tropical trees: patterns of genetic
organization. New Forests 6, 67-94.
Magurran, A.E. 1988. Ecological diversity and its measurement. Croom Helm,
London, UK, 179p.
Majer, J.D. 1993. Comparison of the arboreal ant mosaic in Ghana, Brazil, Papua
New Guinea and Australia - its structure and influence on arthropod diversity.
Pages: 115-141 In: LaSalle, J. and Gauld, I. (eds.). Hymenoptera and Biodiversity,
CAB, Wallingford, UK.
Owen, J., H. Townes and M. Townes, 1981. Species diversity of Ichneumonidae
and Serphidae (Hymenoptera) in an English suburban garden. Biol. J. Linn.
Sot. London 16, 315-336.
Paarmann, W. and N.E. Stork. 1987. Canopy fogging, a method of collecting living
insects for investigations of life history strategies. Journal of Natural History
21, 563-566.
Pielou, E.C. 1975. Ecological diversity. Wiley-Interscience, London, UK, 165p.
Pitelka, L.F. 1993. Biodiversity and policy decisions. Pages 481-493, In: Schultze, E-
D. and Mooney, H.A. (eds.). Biodiversity and ecosystem function. Springer-
Verlag, Berlin, Germany.
Prendergast, J.R., R.M. Quinn, J.H. Lawton, B.C. Eversham, and D. W. Gibbons. 1993.
Rare species; the coincidence of diversity hotspots and conservation strategies.
Nature 365, 335-337.
Reid, WV., J.A. McNeely, D.B. Tunstall, D.A. Bryant and M. Winograd, 1993.
Biodiversity indicators for policy-makers. World Resources Institute,
Washington, USA, 42p.
Rieseberg, L.H., H. Choi, R. Chan and C. Spore. 1993. Genomic map of a diploid
hybrid species. Heredity 70, 285-293.
Simberloff, D. 1992. Do species-area curves predict extinction in fragmented forest?
Pages: 75 - 89 In: Whitmore, T.C. and Sayer, J.A. (eds.). Tropical deforestation
and species extinction. Chapman and Hall, London, UK.
Solbrig, O.T. 1991. Biodiversity: scientific issues and collaborative research proposals.
MAB Digest 9, UNESCO, Paris, France, 77p.
Soulé M.E. 1986. Conservation biology: the science of scarcity and diversity. Sinauer
Associates, Sunderland, Massachusetts, USA, 584p.
Stiles, F.G. and A.F. Skutch. 1989. A Guide to the Birds of Costa Rica. Cornell
University Press, Ithaca, NY, USA, 511p.
Townes, H. and M. Townes. 1966. A catalogue and reclassification of neotropic
Ichneumonidae. Memoirs of the American Entomological Institute 8, l-367.
Vane-Wright, R.I., C.J. Humphries and P.H. Williams. 1991. What to protect?

45
 Systematics and the agony of choice. Biological Conservation 55, 235-254.
WCMC. 1992. Global biodiversity. Report by World Conservation Monitoring
Centre. Chapman and Hall, London, UK, 585p.
Whittaker, R.H. 1975. Communities and ecosystems. MacMillan, New York, USA,
385p.
Williams, P.H., R.I. Vane-Wright and C. J. Humphries. 1993. Measuring biodiversity
for choosing conservation areas. Pages: 309-328 In: LaSalle J. and Gauld, I. (eds.).
Hymenoptera and biodiversity. CAB, Wallingford, UK.
Wilson, E.O. 1987. The arboreal ant fauna of Peruvian Amazon forests: a first
assessment. Biotropica 19, 245-251.
Wolda, H. and D.W. Roubik. 1986. Nocturnal bee abundance and seasonal bee
activity in a Panamanian forest. Ecology 67, 426-433.

46
 Chapter

LANDSCAPE CHARACTERIZATION AN D
BIODIVERSITY RESEARC H

Virginia H. Dale1, Holly Offerrnan2, Robert Frohn3 and

Robert H. Gardner4

INTRODUCTION
Rapid deforestation often produces landscape-level changes in forest characteristics
and structure, including area, distribution, and forest habitat types. Changes in
landscape pattern through fragmentation or aggregation of natural habitats can
alter patterns of abundance for single species and entire communities (Quinn and
Harrison 1988). Examples of single-species effects include increased predation along
the forest edge (Andrean and Angelstam 1988), the decline in the number of species
with poor dispersal mechanisms, and the spread of exotic species that have
deleterious effects (e.g., gypsy moth). A decrease in the size and number of natural

5
1 Environmental Sciences Division, Oak Ridge National Laboratory , P. 0. Box 2008, Oak Ridge,
Tennessee 378316035.
2 Geography Department, 1113 LeFrak Hall, University of Maryland, College Park, Maryland 20742.
3 Department of Geography, Remote Sensing Unit, University of California, Santa Barbara, California
93106.
4 Appalachian Environmental Laboratory, Gunter Hall, Frostburg, Maryland 21532.
5 Managed by Martin Marietta Energy Systems, Inc. for the U.S. Department of Energy, under contract
DE-AC05840R214OO.

47
habitat patches increases the probability of local extirpation and loss of diversity
of native species, whereas a decline in connectivity between habitat patches can
negatively affect regional species persistence (Fahrig and Merriam 1985). Thus,
there is empirical justification for managing entire landscapes, not just individual
habitat types, in order to ensure that native plant and animal diversity is maintained
(McGarigal and Marks 1993).
A landscape can be defined as an area composed of a mosaic of interacting
ecosystems, or patches (Forman and Godron 1986), with the heterogeneity among
the patches significantly affecting biotic and abiotic processes in the landscape
(Turner 1989). Patches comprising a landscape are usually composed of discrete
areas of relatively homogeneous environmental conditions (McGarigal and Marks
1993) and must be defined in terms of the organisms of interest. For example, in a
landscape composed of equal parts of forest and pasture, a photophilic butterfly
species would perceive the pasture areas as suitable habitat whereas a shade-tolerant
species would prefer the forest. In addition, both landscapes and patches are
dynamic and occur on a variety of spatial and temporal scales that vary as a function
of each animal’s perceptions (McGarigal and Marks 1993). For instance, a long-
lived and far-ranging bird will view its environment at broader spatial and temporal
scales than a short-lived, wingless insect (Allen and Starr 1982, Urban et al. 1987).
These differences must be incorporated and used in landscape analysis by changing
the spatial or temporal resolution of a database or simulation model.
Species with different life-history characteristics have been used in simulation
models (Gardner et al. 1993) which show that the interaction of natural and
anthropogenic disturbance with existing landscape pattern may dramatically
affect the risk of species loss. Those species which are most vulnerable are ones
that become isolated as a result of landscape fragmentation and are also restricted
to specific habitat types. These simulation results have also shown that policies for
land management that change the degree of landscape fragmentation will result in
a change in the competitive balance between species, further jeopardizing the
maintenance of native species diversity (Gardner et al. 1993).
Theoretical work in landscape ecology has provided a wealth of methods for
quantifying fragmentation and other spatial characteristics of landscapes (e.g.,
Baker and Cai 1992, Gardner and O'Neill 1991, Gustafson and Parker 1992,
Krummel et al. 1987, O’Neill et al. 1988, Plotnick et al. 1993, Loehle and Wien
1994). Recent advances in remote sensing and geographic information systems
(GIS) allow these methods to be readily applied over large areas. One of today’s
challenges is to relate quantitative measures of landscape characteristics to
changes in biodiversity of animals dependent on the landscape structure. The current
paucity of spatially-explicit ecological field data makes exploring this relationship
difficult.
The objectives of this paper are to present a brief overview of common mea-
sures of landscape characteristics, to explore the new technology available for their
calculation, to provide examples of their application and to call attention to the
need for collection of spatially-explicit field data. The paper focuses on spatial
issues related to macroscopic tropical fauna, although the ideas are, in theory,
applicable to temporal analysis and other biotic groups.

MEASURES OF LANDSCAPE CHARACTERISTICS

Landscapes can be quantified in terms of area, diversity, and pattern. Area measures
such as total area of habitat suitable for a particular species, maximum patch size,
and mean patch size are often the simplest to calculate and interpret. For instance,
a decrease in the total area of habitat available often correlates with species decline
(Wilson 1988, Saunders et al. 1991). Similarly, information on maximum patch size
may provide insight into long-term population viability because populations are
unlikely to persist in landscapes where the largest patch is smaller than that species’
home range.
Traditional diversity indices such as the Shannon Index and Simpson Index
quantify diversity rather than pattern. These indices first gained popularity as
measures of plant and animal diversity and are easily applied to landscape diver-
sity (O'Neill et al. 1988). Unfortunately, these indices convey no information about
the structure and arrangement of patches within the landscape. For instance, a
landscape composed of 90% forest and 10% pasture would yield the same diver-
sity index value as a landscape of 10% forest and 90% pasture. In addition, these
diversity indices combine patch richness and evenness information, although
these components are often more useful when considered separately. Richness
refers to the number of patch types present; because many organisms are asso-
ciated with a single type, patch richness may correlate well with species richness
(McGarigal and Marks 1993). Following this line of reasoning, Stoms and Estes
(1993) outline a remote sensing agenda for mapping and monitoring biodiversity
which focuses almost exclusively on species richness. Evenness, on the other hand,
refers to the distribution of area or abundance among patch types.
Indices which represent the spatial arrangement of landscapes have been
developed from theoretical work in landscape ecology. Because no single index

49
 can capture the full complexity of the spatial arrangement of patches, a set of in-
dices are frequently evaluated. Three of the more common indices are dominance,
contagion, and fractal dimension ( O'neill et al. 1988). Dominance, which is the
complement of evenness, provides a measure of how common one land cover is
over the landscape (Figure 1). Its value indicates the degree to which species
dependent on a single habitat can pervade the landscape (e.g., koalas dependent
on eucalyptus groves). The contagion index measures the extent to which land
covers are clumped or aggregated (Figure 2). Contagion is a useful metric for those
species which require large contiguous areas of a particular land cover (e.g., carrion
beetles unwilling to cross deforested gaps between forest patches, Klein 1989).
Fractal dimension uses perimeter-to-area calculations to provide a measure of
complexity of patch shape (Figure 3). Natural areas tend to have a more complex
shape and a higher fractal value, whereas human-altered landscapes have more
regular patch structure and a lower fractal dimension (Krummel et al. 1987). This
difference can influence the diversity of species which inhabit edges or require
multiple habitats (e.g., large herbivores requiring both forests for cover and open
fields for forage, Senft et al. 1987).

INDICES: DOMINANCE

- degree to which one land cover

type dominates the landscape.

HIGH
Figure 1: Examples of dominance.

50
 RECENT APPROACHES FOR QUANTIFYING LANDSCAPE PATTERN
Spatial indices and other landscape-level measures can be painstakingly
calculated by hand from maps but are typically calculated digitally by computer
from a grid of numeric values which represent those maps. Both field work and
aerial photography can provide mapping data, but satellite-borne sensors
automatically collect and store such data in a digital grid-cell format. This format
is ideal for quantifying spatial characteristics of landscapes or as input to geogra-
phic information systems (GIS) and computer simulation models.
Satellite remote sensing offers several other advantages over traditional field
work. First, data can be collected simultaneously over large areas. Whereas it might
take two years of field work to map the vegetation over a 1000 km area, a satellite
2

can obtain an image of the same area in a few seconds. In addition, satellites collect
data for multiple time periods and at multiple spatial and spectral resolutions
using a repeatable and non-destructive sampling method. Finally, satellite images
have a very high information content, and the prices for both images and computer
equipment are dropping rapidly. Free public domain software is available for image
analysis and the quantification of the results maps (McGarigal and Marks 1993).
These features combine to make remote sensing, and satellite imagery in particu-
lar, an important tool for ecological monitoring and quantitative assessment of
landscape pattern.
The utility of remotely sensed data is increased by integration with compu-
terized geographic information systems (GIS) and simulation models that project
changes in spatial cover under specific scenarios. GIS allows the efficient layering
of many types of data (e.g., vegetation, hydrology, elevation) by referen-cing all
data to a common denominator: geographic location. This multilayered data set
can be used to drive spatially-explicit simulation models which examine causes
and effects of changes in the spatial arrangement of each layer. Existing biodiversity
data can be stored as one of these layers, or biodiversity information can be in-
ferred from other layers (e.g., fauna1 biodiversity may be associated with struc-
tural diversity of the vegetation layer).
The theoretical and technical groundwork has been laid to allow efficient
quantification of GIS landscape layers for biodiversity research. Nevertheless, the
ties between theory, technology, and reality are tenuous at best. Dale et al. (1994)
used the Dynamic Ecological-Land Tenure Analysis (DELTA) model to explore the
implications of various land management alternatives on Amazonian biodiversity
as discussed below. This case study demonstrates how spatially-explicit ecological
data can be used to strengthen the ties between theory, technology and reality.

52
between 1979 and 1988 (Frohn et al. 1990) which opened the interior forest areas to
colonization. Colonists used slash and burn techniques to clear the forest for
agriculture, producing a dynamic mosaic of agricultural fields, pasture, regrowth,
and mature forest, with most of the clearing originating near roads. Between 1978
and 1988, 17,717 km2 of Rondonia’s forest were cleared, and an additional 1,417
km of forest were isolated from the contiguous forest into small (<l00 km 2 ) patches
2

(Skole and Tucker 1993).

Changing patterns of forest clearance and isolation can be simulated by the
Dynamic Ecological-Land Tenure Analysis (DELTA) model (Southworth et al. 1991,
Dale et al. 1993, 1994). DELTA is a stochastic spatially-explicit model which com-
bines a decision model of farmers’ land-use choices with ecological information
about changes in biomass. The model uses side-looking radar imagery, GIS, field
estimates of biomass in forests, and socio-economic data to produce statistics and
maps of the simulated changes in the area, biomass, and pattern of land-cover types.

Quantifying modelled landscapes

DELTA model simulations suggest that different scenarios of land management
result in unique land-cover patterns (Dale et al . 1994) (Figure 5). Simulating the
best, typical, and worst-case scenarios permits evaluation of the causes of specific
land-cover changes. Land-use activities that are typical for colonists in Rondonia
(Coy 1987, Dale and Pedlowski 1992, Leite and Furley 1985) involve rapid clearing
of the forest and almost complete deforestation within 18 years. The worst case
scenario (taken from the extreme of the Transamazon Highway experience as
reported by Moran 1981 and Fearnside 1980, 1984, and 1986) results in total clear-
ance in the first 10 years. On the other hand, a best case scenario can be simulated
in which forest clearance stabilizes at about 40% by year 20. The best case scenario
involves some clearing, but no burning, of the virgin forest and planting of perennial
trees. The worst and best case model projections are hypothetical, but the typical
model scenario is meant to replicate recent land management activities in central
Rondonia.
Comparing model projections to satellite imagery over recent years is a way to
verify the modeled projections. Frohn et al. (in prep.) compare the percent of forest
cleared, contagion and fractal indices from the three model scenarios to those values
obtained from classified Landsat imagery for 1978, 1980 and 1986 (Figure 6). The
clearing pattern for 1978 and 1980 are similar to the typical simulation projections
(Figure 6a). However, the model overestimates the amount of clearing for the 1986
scene. Initially, contagion is high for both the simulation and the Landsat estimate

54
(Figure 6b) because the landscape consists primarily of large contiguous patches of
forest. Contagion decreases in both estimates as the number of small forest clearings
increases and the landscape is less dominated by large patches of forest. In the
simulations, contagion increases as larger patches of cleared forest dominate the
landscape. However, this pattern has not been verified by Landsat data. The fractal
dimension (Figure 6c) values for the typical simulation and the Landsat estimate
show similar patterns, indicating that the model predicts land-scape patch
complexity similarly to that determined from remote sensing.
These comparisons show that the typical scenario simulation is consistent with
both the amounts and patterns of forest clearing for central Rondonia for the years
tested. Model estimates can therefore be used with greater confidence to predict
landscape changes in later years and the response of biodiversity to those changes.

Modelling faunal response to landscape pattern

In order to relate landscape-level changes to changes in fauna1 abundance and
distribution, spatially-explicit data were collated for 9 taxonomically diverse
groups of neotropical forest animals and summarized in Table 1 (as discussed by
Dale et al 1994).
. Examples of spatially-explicit data include the maximum gap
width between habitat patches that an animal is likely to cross; the minimum patch
area required to maintain normal behavioral patterns; the spatial distribution of
rare or patchily distributed resources vital to a particular species’ survival (e.g.,
special habitats for breeding); and the width of the “buffer zone” at a forest edge
where climatic or ecological edge effects render the area uninhabitable for a particu-
lar species. These data were collated from a literature search and studies of animal
activity subsequent to experimental manipulation of intact forests into patches of
1, 10, 100 and 1000 km2 (the manipulation and full data sets are discussed by
Bierregaard et al. 1992 and Offerman et al. in press).
The landscape scale and patch characteristics were defined based on the animals’
perception of their environment. DELTA typically runs on an area of -3000 km2
this scale represents an intermediate landscape size for the macroscopic, mobile
fauna selected. Model output data was stored in a grid with 37.5 m resolution,
because field observations of maximum gap width crossed between habitat patches
was most easily divided into multiples of 37.5. In other words, those animals that
could not cross a distance greater than 37.5 m were assigned a low gap-crossing
ability. Patches were defined simply as areas covered by forest, because the 9 selected
groups of animals were all primarily forest-dwellers.
For each model year, the area of forest habitat suitable for each animal group

57
 was measured. First, “connected” clusters of habitat cells were identified. A cluster
is connected if an animal in one cell can move to any other cell in that cluster (i.e.,
gaps between cells in a cluster are not wider than the maximum gap width that
animal is able to cross). Next, clusters with areas less than the minimum area
required by an individual or group (for those that only occur in groups) were
discarded. Further discussion of this technique can be found in Pearson et al. (in
press).
The result of this analysis is that changes in available habitat are similar for
animals that have their gap-crossing ability proportional to area requirements (Figure
7a), regardless of taxonomic affiliation (Dale et al. 1994). For instance, the model
suggests that species with large gap-crossing abilities and large area requirements
(e.g., jaguars) respond in a similar fashion as species with small gap-crossing abili-
ties and smaller area requirements (e.g., sloths). In contrast, animals with gap-
crossing ability disproportionately small in comparison to their area
requirements (e.g., scarab beetles) decline more rapidly (Figure 7b). Few animals
larger than insects seem to fall into this latter group; therefore landscape-level
analysis using simply gap-crossing ability and area requirements may provide a
swift preliminary identification of the animals most susceptible to rapid decline
and possible extirpation, assuming that fragmentation does not induce behavioral
changes.
Once sensitive species have been identified, additional spatially-explicit life
history data may be incorporated to improve the accuracy of the assessment. For
example, when possible edge effects and breeding habitat requirements are inclu-
ded in the assessment of suitable habitat available for the tropical frog (Chiasmocleis
shudikarensis), the amount of suitable habitat is decreased to 39% of the original
area defined by gap-crossing and area requirements alone (Dale et al. 1994).

Case study results

Spatially-explicit land-cover and fauna1 life-history data are both vital for assessing
the impact of landscape change on biodiversity. These data can be derived from re-
mote sensing and from in situ ecological studies, and integrated with models which
simulate the cause and effect of changes in spatial pattern. Maps produced from
the model simulations can be quantified using spatial indices; these indices can
then be used to represent the changes in land-cover patterns to which species
respond.
Combining spatial indices with species-specific ecological data provides a use-
ful method for identifying species sensitive to landscape-level habitat modifica-

59
tions. Species response to these modifications may be based on spatial-explicit
behavioral characteristics rather than taxonomic classification. The major impli-
cation of the Rondonia study is that a “balance” between gap-crossing ability and
minimum area requirements allows species to maintain themselves under varied
land cover conditions.

CONCLUSIONS AND FUTURE RESEARCH OPPORTUNITIES

The theory and technology currently exist to perform rapid, large-scale quantita-
tive analysis of biodiversity in real and modelled landscapes. Policymakers request
this type of analysis before making high-profile, million-dollar decisions (e.g., the
issue of harvesting old-growth forests of the United States’ Pacific Northwest while
protecting the spotted owl). However, the current paucity of spatially-explicit
fauna1 life-history data makes it difficult to verify the link between real-world
phenomena and the statistical phenomena seen in the landscape indices.
Policymakers require the linkage between indices and biodiversity be firmly
established before the indices can be used to define policy. The urgency of
biodiversity conservation issues, therefore, suggests first that field-based research
agendas should focus less on taxonomy and morphological description, and more
on collection of spatial data; and second, that researchers with remote sensing, GIS,
and modelling capabilities should quantify the link between measures of landscape
characteristics and the observed ecology of species occupying those landscapes.

ACKNOWLEDGMENTS
Tim Boyle and the Center for International Forestry Research (CIFOR) provided
funds for attending the International Union of Forestry and Research Organizations’
Symposium on Measuring and Monitoring Biodiversity in Tropical and Temperate
Forests from August 2- September 2, 1994 in Chiang Mai, Thailand at which this
paper was presented. ORISE provided administration of funding for Holly
Offerman. Support, advice, and suggestions from Monica Turner, Tom Ashwood,
and the Oak Ridge National Laboratory GIS staff are appreciated. Comments on
an earlier draft by Rebecca Efroymson and Andrew Schiller were helpful. This is
Environmental Sciences Division Publication Number 4347.

61
REFERENCES
Allen, T.F.H. and T.B. Starr. 1982. Hierarchy: Perspectives for Ecological Complexity.
University of Chicago Press, Chicago.
Andrean, H. and P. Angelstam. 1988. Elevation of predation rates as an edge effect
in habitat islands: experimental evidence. Ecology 69: 544-547.
Baker, W.L. and Y. Cai. 1992. The role programs for multiscale analysis of landscape
structure using the GRASS geographical information system. Landscape
Ecology, 7: 291-302.
Becker, P., J.S. Moure and F.J.A. Peralta. 1991. More about euglossine bees in
Amazonian forest fragments. Biotropica 23: 586-591.
Bierregaard, R.O. 1990. Avian communities in the understory of Amazonian forest
fragments. Pages 333-343 In: A. Keast and J. Kikkawa (eds.) Biogeography and
ecology of forest bird communities. SPB Academic Publishing, The Hague.
Bierregaard, R.O., T.E. Lovejoy, V. Kapos, A.A. dos Santos and R.W. Hutchings.
1992. The biological dynamics of tropical rainforest fragments. BioScience 42
(11): 859-866.
Coy, M. 1987. Rondonia: frente pioneira e programa Polonoroeste. O processo de
diferenciação socio economica na periferia e os limites do planejamento publico.
Tubinguen Geographhische Studien 95: 253-270.
Dale, V.H. and M.A. Pedlowski. 1992. Farming the forests. Forum for Applied
Research and Public Policy 7(4): 20-21.
Dale, V.H., R.V. O’Neill and F. Southworth. 1993. Causes and effects of land-use
change in central Rondonia, Brazil. Photogrammetric Engineering & Remote
Sensing 59: 997-1005.
Dale, V.H., R.V. O’Neill, F. Southworth, and M.A. Pedlowski. 1994. Modeling effects
of land-use change in central Rondonia, Brazil. Conservation Biology 8:196-
206.
Fahrig, L., and G. Merriam. 1985. Habitat patch connectivity and population
survival. Ecology 66: 1762-1768.
Feamside, P.M. 1980. Land use allocation of the Transamazon highway colonists
of Brazil and its relation to human carrying capacity. Pages 114-138 In: Barbira-
Scazzocchio (ed.) Land, people and planning in contemporary Amazonia.
University of Cambridge Centre of Latin American Studies Occasional Paper
No. 3, Cambridge, England.
Feamside, P.M. 1984. Land clearing behavior in small farmer settlement schemes
in the Brazilian Amazon and its relation to human carrying capacity. Pages 255-

62
 271 In: A.C. Chadwick and S.L. Sutton (eds.) Tropical rain forests: the Leeds
Symposium. Leeds Philosophical and Literary Society, Leeds, England.
Feamside, PM. 1986. Human carrying capacity of the Brazilian rainforest. Columbia
University Press, New York, New York.
Forman, R.T.T. and M. Godron. 1986. Landscape Ecology. John Wiley & Sons,
New York.
Frohn, R., V.H. Dale and B. Jimenez. 1990. The effects of colonization on deforestation
in Rondonia, Brazil. ORNL/TM-11470, Oak Ridge, Tennessee.
Frohn, R.C., K.C. McGwire, V.H. Dale and J.E. Estes. In review. Using satellite
remote sensing analysis to evaluate a socioeconomic and ecological model of
land-use change in Rondonia, Brazil. International Journal of Remote Sensing.
Gardner, R.H. and R.V. O’Neill. 1991. Pattern, process and predictability: The use
of neutral models for landscape analysis. Pages 289-307 In: M.G. Turner and
R.H. Gardner (eds.) Quantitative Methods in Landscape Ecology. Springer-
Verlag, New York.
Gardner, R.H., A.W. King and V.H. Dale. 1993. Interactions between forest
harvesting, landscape heterogeneity, and species persistence. In: D.C. LeMaster
and R.A. Sedjo (eds.) Modeling Sustainable Forest Ecosystems, Proceedings of
a 1992 workshop in Washington, D.C. Published by American Forests,
Washington, DC.
Gustafson, E.J., and G.R. Parker. 1992. Relationships between landcover proportion
and indices of landscape spatial pattern. Landscape Ecology 7: 101-110.
Klein, B.C. 1989. Effects of forest fragmentation on dung and carrion beetle
communities in central Amazonia. Ecology 70(6):1715-1725.
Krummel, J., R.H. Gardner, G. Sugihara, R.V. O’Neill and P.R. Coleman. 1987.
Landscape patterns in a disturbed environment. Oikos 48:321-324.
Leite, L.L. and P.A. Furley. 1985. Land development in the Brazilian Amazon with
particular reference to Rondonia and the Ouro Preto colonization project. Pages
119-140 In: J.Hemming(ed.) Change in the Amazon basin. Manchester University
Press, Manchester, England.
Loehle, C. and G. Wien. 1994. Landscape habitat diversity: a multiscale information
theory approach. Ecological Modelling 73:311-329.
Malcolm, J.R. 1990. Estimation of mammalian densities in continuous forest north
of Manaus. Pages 339-357 In: A. Gentry (ed.) Four neotropical rainforests. Yale
University Press, New Haven, Connecticut.

63
Malcolm, J. R. 1991. The small mammals of Amazonian forest fragments: pattern
and process. Ph.D. dissertation. University of Florida, Gainesville, Florida.
Montgomery, G.G. and M.E. Sunquist. 1975. Impact of sloths on neotropical energy
flow and nutrient cycling. Pages 69-78 In: F. B. Golley and E. Medina (eds.)
Ecological studies, vol. 2. Springer-Verlag, New York, New York.
Montgomery, G.G. and M.E. Sunquist. 1978. Habitat selection and use by two-toed
and three-toed sloths. Pages 329-359 In: G.G. Montgomery (ed.) The ecology of
arboreal folivores. Smithsonian Institution Press, Washington, D.C.
Moran, E.F. 1981. Developing the Amazon. Indiana University Press, Bloomington,
Indiana.
McGarigal, K. and B. J. Marks. 1993. FRAGSTATS: spatial pattern analysis program
for quantifying landscape structure (unpubl.) software, Dept. Forest Science,
Oregon State University.
Offerman, H.L., V.H. Dale, S.M. Pearson, R.O. Bierregaard, Jr. and R.V. O’Neill (in
press). Effects of forest fragmentation on neotropical fauna: current research
and data availability. Environmental Reviews.
O’Neill, R.V., J.R. Krummel, R.H. Gardner, G. Sugihara, B. Jackson, D.L. DeAngelis,
B.T. Milne, M.G. Turner, B. Zygmnuht, S.W. Christensen, V.H. Dale and R.L.
Graham. 1988. Indices of landscape pattern. Landscape Ecology 1:153-162.
Parker, S.B. (ed.). 1990. Grzimek’s Encyclopedia of Mammals, vol. 3,4. McGraw
Hill Publishing Co., New York, New York.
Pearson, S.M., M.G. Turner, R.H. Gardner and R.V. O’Neill (in press). An organism-
based perspective of habitat fragmentation. In: R.C. Szaro (ed.) Biodiversity in
managed landscapes: theory and practice. Oxford University Press, Oxford,
England.
Plotnick, R.E., R.H. Gardner, and R.V. O’Neill. 1993. Lacunarity indices as measures
of landscape texture. Landscape Ecology, 8: 201-211.
Powell, A.H. and G.V.N. Powell. 1987. Population dynamics of male euglossine
bees in Amazonian forest fragments. Biotropica 19(2): 176-179.
Quinn, J.F. and S. Harrison. 1988. Effects of habitat fragmentation and isolation on
species richness: evidence from biogeographic patterns. Oecologia 75: 132-140.
Rylands, A.B. and A. Keuroghlian. 1988. Primate populations in continuous forest
and forest fragments in central Amazonia. Acta Amazonica 18: 291-307.
Saunders, D.A., R.J. Hobbs and C.R. Margules. 1991. Biological consequences of
ecosystem fragmentation: A review. Conservation Biology 5: 18-32.
Schwarzkopf, L. and A.B. Rylands. 1989. Primate species richness in relation to

64
 habitat structure in Amazonian rainforest fragments. Biological Conservation
48: l-12.
Senft, R.L., M.B. Coughenour, D.W. Bailey, L.R. Rittenhouse, O.E. Sala and D.M.
Swift. 1987. Large herbivore foraging and ecological hierarchies. BioScience 37:
789-799.
Skole, D. and C.J. Tucker. 1993. Tropical deforestation and habitat fragmentation
in the Amazon: Satellite data from 1978-188. Science 260: 1905-1910.
Southworth, F., V.H. Dale and R.V. O’Neill. 1991. Contrasting patterns of land use
in Rondonia, Brazil: simulating the effects on carbon release. International Social
Sciences Journal 130: 681-698.
Stoms, D.M. and J.E. Estes. 1993. A remote sensing research agenda for mapping
and monitoring biodiversity. International Journal of Remote Sensing 14(10):
1839-1860.
Turner, M.G. 1989. Landscape ecology: The effect of pattern on process. Annual
Review of Ecology and Systematics, 20: 171-197.
Urban, D.L., R.V. O’Neill and H.H. Shugart.1987. Landscape ecology. BioScience
37: 119-427.
Wilson, E.O. (ed.). 1988. Biodiversity. National Academy Press. Washington, D.C.
Zimmerman, B.L. and R.O. Bierregaard. 1986. Relevance of the equilibrium theory
of island biogeography and species - area relations to conservation with a case
from Amazonia. Journal of Biogeography 13: 133-143.

65
 Chapter

5
COMPARATIVE PARADIGMS
FOR BIODIVERSITY ASSESSMENT

W. L. Myersl, G. P . Patil2 and C. Taillie2

INTRODUCTION
Formulation and comparison of diversity measures has such extensive precedent
as to be considered an integral domain of classical statistical ecology (Grassle et al.
1979, Hurlbert 1971, MacArthur and MacArthur 1961, Magurran 1988, Patil and
Taillie 1982, Pielou 1975, 1977). Such work, however, has mostly cast the several
measures as competitors rather than-being complementary. The case basis for
conventional diversity work lies primarily in local intensive studies, with recorded
occurrence of taxa being considered definite, and relative abundance estimates
considered as quasi-ratio information. Issues of uncertainty, such as mis-
identification and differential detection, have been largely relegated to the
background. Increasing representation of taxa with expanding area of observa-
tion has been extensively studied, but issues of appropriate plot size and
configuration have overshadowed the more fundamental implication that any

1 School of Forest Resources; Environmental Resources Research Institute, Pennsylvania, State

University, University Park, PA 16802, USA.
2 Center for Statistical Ecology and Environmental Statistics, Department of Statistics.

67
diversity determination is relative to its area basis. Temporal gap dynamics of
forested landscapes are well-documented, but not explicitly recognized as a scale
consideration in diversity assessments. Recognition of need for regional diversity
investigations, as evidenced by actual funding of operations, is relatively recent.
The classical view of diversity remains important for intensive studies of
particular ecological communities and forest stands (Gove et al. 1994, Hunter 1990,
Swindel et al. 1984, Swindel et al. 1987, Swindel et al. 1991). However, the emerging
sciences of landscape ecology and conservation biology have made evident the
logistical and economical impracticality of such intensive observational coverage
for regions on the order of square kilometers and larger (Scott et al. 1989). Such
spatial scales are necessarily encompassed by contemporary ecosystem-oriented
resource management and design of regional/national networks of biodiversity
reserves. Furthermore, species/area and minimum viable population issues become
fundamental in these latter contexts.
Turner provides an overview of these scale, observation, and measurement
issues (Chapter 7, this volume). Lund et al. (Chapter 25, this volume) likewise
provide an overview of contemporary broad-area observational technology, but
do not emphasize the important consideration that the informational outputs of
these technologies must be interpreted either manually or algorithmically relative
to diversity factors. Available evidence regarding diversity at landscape and regional
scales is thus substantially indirect. Our focus is on possible approaches to
determination of patterns having probable relevance to diversity. Patterns constitute
second-order information, subject to analysis with respect to diversity, and
refinement by further collection of additional information. Patterns likewise provide
a basis for designing acquisition strategies for obtaining further information that
addresses unresolved questions. Any particular pattern, however, often has several
potential implications.
We first consider patterns of different diversity indices that are subject to joint
interpretation as diversity profiles. We then turn to spatial patterns derived from
broad-area observational technologies and knowledge-based models.

PROFILES OF DIVERSITY INDICES

Diversity is generally described as a composite property that reflects both the number
of species (richness) in a biological community and the evenness with which
abundance is distributed among the different species. A wealth of indices have
been proposed for measuring diversity; the most popular appear to be (i) the number
SOME EXAMPLES
Patil and Taillie (1979) have employed the s(m) profile to examine the temporal
trend in avian diversity in the vicinity of Colstrip, Montana, during the 1975-1977
time period. The 1975 breeding season was taken as the baseline since several power
plants had not yet gone into operation. In all, 81 bird species were recorded over
the survey period. Observations were made along a number of short transects, but
the data were available by transect for the Western Meadowlark species only. Annual
total frequencies were available for the other species.

SAMPLE SIZE m so
SAMPLE SIZE m

Figure 1: Ninety-five percent confidence bands for the jack-knifed estimates of the avian
s(m) diversity profiles. (A) 1 0 and (B) 1 00.

The s(m) profiles are shown in Figure 1. The 95 percent confidence bands were
obtained using the jack-knifed estimates of standard error. Jack-knifing was done
by individual, although jack-knifing by transect would have been preferable if the
information had been available. The profiles show that between 1975 and 1977
there was a drop in diversity for small m (abundant species) but an increase for
large m (rare species). These changes correspond to a growing dominance of Western
Meadowlark, on the one hand, and an increase in the number of occasional species,
on the other hand. The changes are statistically significant since the confidence
bands about the profiles do not overlap. But the changes are not necessarily
attributable to start up of the power plants since the confidence bands reflect only
within-year sampling variability and do not include between-year fluctuations in
avian diversity.
Swindel et al. (1987) has used the profiles to study trends in plant diversity
following clearcutting of a Douglas-fir community in the Oregon Cascades. TWO

71
measures of abundance were considered: species plot-frequency and crown cover.
When frequency was used, clear patterns emerged from the profiles. There was a
sharp drop in diversity immediately after clearcutting, followed by a steady in-
crease to levels above those that prevailed before clearcutting. Interpretation was
more difficult with crown cover since the profiles displayed several intersections.
Gove et al. (1994) uses the profiles to examine the management issue of
maximizing the diameter-class diversity in an uneven-aged northern hardwood
stand.

EXTENSIVE SPATIALLY COMPARATIVE DIVERSITY

Progressive impoverishment of biota at regional and global scales has engendered
a growing sense of urgency for conservation of biological diversity (Wilson and
Peter, 1988) in light of which locally intensive biodiversity analysis becomes
somewhat akin to making a detailed statistical inventory of a particular stand in a
burning forest. The onslaught of destruction limits the value of the detailed
information to documentation of impending loss. While there may be virtue in
such documentation for the historical record, the more critical need is for rapid
assessment over threatened regions to guide formulation of strategy for damage
control and focus attention on critical areas. Quick acquisition of more approximate
information over extensive areas is thus in order. When the situation becomes more
stable, the approximate information can also serve to guide more rigorous
characterization at the site level.
Fundamentally, diversity consists of co-occurrence of differing entities in space
and time. The conventional approach to acquisition of diversity information involves
observing multiple entities “simultaneously” during a time period in space
partitions. An alternative is to observe different entities separately in common space
partitions, and then spatially compare (overlay) the separate observations for co-
occurrence. The latter approach has been promulgated for rapid assessment through
GIS (geographic information systems) technology by Scott et al. (1987).

Apparent Diversity
Rapid diversity assessment involves synoptic recording of ordered categorical
evidence for occurrence and/or expectation of taxa in spatial tessellations as GIS
layers. Associations of taxa with landscape features, such as habitat, are exploited
for evidential refinement and preliminary analysis. This has been characterized as

72
a “biodiversity filter” approach working from coarser to finer filters (Davis et al.
1990). Such work is inherently comparative, with a major goal of locating areas
likely to have high biodiversity for further elaboration with finer (both spatial and
categorical) filters.
In the USA this work is being conducted in a state-wide national program called
“Gap Analysis” with the purpose of determining “gaps” in the conservation safety
net of protected areas of critical habitat (Scott et al. 1993). The philosophy under-
lying Gap Analysis is essentially that protecting substantial areas of representative
habitats which support complexes of species will simultaneously forestall the onset
of endangerment for all members of the complex. The goal thus becomes one of
identifying regional occurrences of critical habitats and then determining instances
where such habitats are lacking in long-term protective management. Proponents
of Gap Analysis are emphatic that it is intended to complement, not replace, the
species-by-species approach to preserving biodiversity which is so critical to the
survival of species now nearing extinction. The main goal of Gap Analysis is to
help prevent additional species from being listed as threatened or endangered. The
U.S. Environmental Protection Agency’s EMAP (Environmental Monitoring and
Assessment Program) includes similar components under a monitoring perspective.
While there is an ongoing evolution of Gap Analysis protocols, it typically
proceeds through the following stages (Scott et al. 1991):

1. Derive and digitize map information on vegetation type distribution.

2. Verify the vegetation map information with field work.
3. Digitize existing species distribution maps for fauna.
4. Refine species distribution maps using digital map information relative to habitat
factors.
5. Verify fauna1 distributions.
6. Input data on land ownership status.
7. Digitize current management areas according to levels of protection.
8. Generate map(s) depicting species richness and compositional variations from
information of steps 1-5.
9. Generate map(s) for special-interest species such as threatened, endangered,
and sensitive plants and animals; endemic taxa; or uncommon species found in
fewer than three vegetation types. Information for such species comes largely
from existing databases.
10. Regionalize the map information on diversity from steps 8-9, and locate centres
of species richness (recognizing compositional variation).

73
11. Compare centers of richness regarding species representation and vegetation
types to determine redundancy.
12 Prioritize centers of richness in light of their contribution to state, regional, and
continental biodiversity.
13. Determine to what extent the areas of species richness and vegetation types are
in protected zones, using information from steps 6-7.
14. Identify minimum and optimum areas required for protection of predetermined
.
levels of state wide diversity.
15. Identify landscape corridors between candidate areas.

There is a good deal more parallelism and iteration in the Gap Analysis scenario
than the numbering of steps might suggest. Assembly of the basic maps typically
proceeds concurrently rather than sequentially. Both vegetation and species
distribution maps may go through “first-cut,” “second-cut,” . . . , “nth-cut” versions.
Idaho, as a prototype state, has had a rather complete second “go-round” of the
entire Gap Analysis scenario.
Gap Analysis renditions of vegetation types and fauna1 distributions are
essentially knowledge-based digital map models. Vegetation types are typically
interpreted visually from small-scale satellite imagery and/or inferred by
unsupervised digital analysis of the corresponding image data. Larger scale ima-
gery and existing local maps support the interpretive and unsupervised inferences.
Limited sample-based verification of vegetation types serves more as an indicator
of realized classification error levels than for rigorous quality control.
Fauna1 distribution maps are effectively representations of apparent habitat,
obtained by partial range deletion. Species are ascribed to vegetation types on the
basis of current ecological knowledge of habitat requirements. Vegetation types
that are thought to constitute unsuitable habitat are deleted from the best available
range maps. Negative physiographic elements are likewise deleted using common-
ly available GIS layers, such as soils, geology, topography, and hydrography.
Detailed habitat needs, such as dead snags and ephemeral ponds, are assumed to
occur in the absence of evidence to the contrary. Verification of fauna1 occurrence is
even more difficult and sparse than for vegetation. The recently completed breed-
ing bird atlas database for Pennsylvania (Brauning 1992) is among the stronger
sources of such information for verification. The breeding bird atlas database divides
the state into a grid of cells, each covering one-sixth of a 7 l/2 minute topographic
quadrangle map. In each cell, each observed species is designated in one of four
categories as: a) confirmed breeder, b) probable breeder, c) possible breeder, or

74
d) observed without indication of breeding. The distinctions are based on behaviour
exhibited by-birds as recorded in 19 activity types.
For most species of vertebrates, Gap Analysis synthesizes records of occurrence
in terms of the 635 sq. km. hexagonal grid used by the U.S. Environmental Protec-
tion Agency for its EMAP work. Gap Analysis avoids the issue of intra-stand patch
dynamics by imposing a minimum mapping area of 100 hectares. Gap Analysis
presently makes the rather heroic assumption that vertebrate diversity is indica-
tive of diversity in other major taxa. Since Gap Analysis casts its fauna1 distribu-
tions in terms of apparently suitable habitat, it seems appropriate to suggest the
term apparent diversity for the corresponding patterns of spatial co-occurrence.
Gap Analysis for Pennsylvania is proceeding generally according to the fore-
going regime, with the work being conducted in the spatial analysis laboratory of
the Office for Remote Sensing of Earth Resources within the Environmental
Resources Research Institute. Vegetation and ecological land type determination
are being conducted concurrently A first-level photo-interpretive breakdown of
the landscape into naturalistic and humanistic complexes is based on overall
vegetation and land-use patterns. Delineation of such complexes is accomplished
by on-screen digitizing. Unsupervised digital image analysis yields breakdowns
of vegetation cover within each complex. Knowledge of the ecological setting for
the complex is exploited in assigning attributes to cluster-based polygons. Future
research will contribute additional ecological detail to the attribute information for
polygons.

Ordinal Formations as Spatial Pattern Comparatives

Given the uncertain and speculative nature of Gap Analysis maps, diversity
information arising therefrom is at best ordinal. Nevertheless, such maps do
represent a synthesis of available ecological knowledge and advanced broad-area
observational technology. In and of themselves, however, Gap Analysis species
distribution maps from steps l-5 above are fundamentally just visual aids to think-
ing and dialogue concerning the species. If it is to move forward, such thinking
and dialogue inevitably become spatially comparative among locations, landscapes,
and regions.Aside from highlighting what we don’t know, the utility of Gap
Analysis thus rests on ability to extract coherent composite spatial patterns for
comparative purposes. The composition, regionalization, and comparisons in steps
l0-11 above are therefore crucial.
Humans have considerable perceptual facility for obtaining a visual sense of

75
 pattern from maps, but such perception is unfortunately very subjective and
indefinite. Different observers perceive differently, and have little ability to ex-
press a perceived sense of pattern. If discussion and debate are to transcend
perceptual differences, regionalization and pattern extraction/characterization from
maps must become more systematic. In what follows, we propose a systematic
approach to ordinal spatial structure (pattern) that is applicable to biodiversity in
the Gap Analysis context, but also extends more generally to any synoptic ordinal
spatial context.
A prerequisite to systematic extraction of meaningful spatial pattern is the
selection of mapping variables that are relevant to the issues and lead naturally to
more specific analyses. We suggest that two classes of synoptic mapping variables
are particularly relevant to the biodiversity context. One such class is species richness
with possible restriction to certain taxa, guilds, functional groups, or conservation
status. The second class is number of species, however restricted, that a particular
spatial partition (cell or polygon) has in common with at least one of its neighbour-
ing partitions. The first class addresses diversity in a general sense. The second
class is indicative of interesting compositional regionalizations. Both classes are
obviously scale dependent. The new approach we offer also provides for systematic
exploration of progressively more generalized scales working upward from the
base resolution of the map data source as a scale floor.
Our proposed approach addresses either smooth-surface or tessellated spatial
variables of ordinal, interval, or ratio strength. Tessellations may be regular (cells)
or irregular (polygons), but the tessera are viewed as flat-topped facets versus the
sloping-facet elements of a TIN (triangular irregular network) model. Since our
habitat mapping is still in progress at this time of writing, we cannot offer a full
illustration of application to Gap Analysis. Therefore, we focus instead on conveying
the concepts in terms of contrived data. The smooth-surface case provides the
simplest point of departure for explanation, although it is not necessarily simple in
terms of software.
Consider the stylized “contour map” of a spatial variable in Figure 2. To
emphasize that we are assuming only ordinal strength of information, the contour
“levels” are lettered rather than numbered. Relative to Figure 2, the letter “a” would
be the base (lowest) level. The “a” level is absent from Figure 2 in analogy to the
absence of mean sea level from most topographic maps. Our strategy is to provide
for constructing spatial “objects” that are arranged in formations which are

76
 progression and process any ringlets.
If there is a single (enclosing) exterior neighbour ring at equal or lower level,
then join the object between the rings and progress to the exterior of the neigh-
bour.
If there are any non-enclosing exterior neighbour rings at equal or higher level,
then truncate the exterior progression and process any ringlets.
If all exterior neighbour rings are at lower level, then join the object between the
rings, add non-enclosing neighbour rings to the ringlet list, and progress to the
enclosing neighbour.
If exterior progress reaches an enclosing ring that is already joined to an object,
then truncate exterior progress, attach the current object to the exterior object, and
process any ringlets.
Ringlets are processed similarly with interior replacing exterior. If inward
progress meets an existing object, that object is attached to the current object. When
all ringlets have been processed, then object number is incremented with a new
initial ring being selected. Initial rings are selected from high to low, with rings
already joined on both sides being ineligible. When the formation is complete, any
object attached to an object of the same order is merged with its parent object.
Figure 3 shows the result of the process for Figure 2 in terms of object numbers.
Objects numbered 1, 2, and 3 are first-order. Object 4 is second-order, and object 5
is third-order. Object 4 has objects 1 and 3 as children. Object 5 has objects 2 and 4
as children, with objects 1, 2, 3 and 4 all being its descendants.
Geometrically, the object structuring is motivated by “peaks and saddles”. A
complete intuitive understanding can be obtained by constructing a clay model
and then “slicing” it into objects. “Hillocks” will be sliced off first, then hills, then
foundations of entire “ranges” of hills. Depressions that “hold water” are ignored
in the slicing process.
The dashed line in Figure 3 is a course that transects elements which decide
objects. Figure 4 is a “vertical profile” along this course showing the formation of
the associated dendrogram. The solid descending arrows indicate “centers” of first-
order objects. The dashed descending arrows indicate nodal features.
With this relatively graphic introduction to the formation domain as background,
consider next the cellular equivalent which is easier to manage with respect to
software but more difficult to grasp visually. For purposes of illustration, consider
Figure 5 to be a cellular array (raster grid) of species richness. To elaborate the
formation domain of nested objects from such a cellular grid, begin at the highest
cell and assign object number 1, as long as “outward” movement would be only

78
tion, our interest extends from local concentrations, to ranges of deposits, to the
“mother lode.” Our context is one of prospecting for particular aspects of
biodiversity.
The tree structures of the formation domain place it within the purview of
mathematical graph theory for tree metrics. Formation domains for a given area at
different times can be compared quantitatively in these terms, as can formation
subdomains. Formation domains for different areas are likewise comparable in
terms of structural parameters.
With regard to signal processing, the formation objects become signals modu-
lated at different orders by their families. High-pass filters will direct attention to
the “busy” areas for further investigation and more intensive data collection to
determine whether the business constitutes signal or noise. Conversely, low-pass
filters will direct attention to the “quiet” areas having low orders of spatial variance.
It thus becomes possible to explore the spatial variance properties of the data
systematically in a manner that should serve to guide further data collection and
suggest appropriate resurvey intervals and sampling intensities for monitoring.
Such exploration requires no prior assumptions beyond that of at least ordinal
strength for the data.
Scale and spatial modulation are intimately intertwined. While it is impossible
to go below the scale floor corresponding to the spatial resolution of the data, one
can study modulation at coarser scales by pruning the formation trees downward
from the leaves toward the root. This is accomplished by suppressing objects through
assigning the parent object number to them and their descendants. Mapping of
pruned objects can serve for such generalizing exploration of the scale spectrum.
Whereas fractal dimension is only indicative of pattern over scale, the formation
domain expresses pattern as structure while defining the elements of structure.
Stated differently, fractal dimension indicates the nature of pattern whereas the
formation domain elaborates the pattern.
Because of their binary nature, formation analysis cannot be applied directly to
Gap Analysis apparent habitat maps for individual species. Mapping a habitat
suitability index for each species would, however, provide an ordinal spatial varia-
ble appropriate to formation analysis.

SUMMARY
Diversity is intrinsically relative since data are necessarily acquired in time and
space. Patterns and comparatives are thus fundamental to diversity questions.

82
Perceptual heuristics have dominated approaches to pattern and comparison with
respect to biodiversity, usually generating more heat than light. We outline
systematic approaches to pattern and comparatives that are independent of
perception. Diversity profiles address patterns of different indices. The spatially-
based formation domain approach is object-oriented as well as perceptually
objective.
Data of ordinal strength can convey spatial pattern. Both informational strength
of data and spatial resolution will, however, affect coherence of patterns and
definitiveness of interpretation. Even vague patterns at broad scales can be useful
in focusing further data collection effort. Gap Analysis and apparent diversity
exemplify such broad-scale to fine-scale progression of pattern analysis.

ACKNOWLEDGEMENT
Prepared with partial support from the Statistical Analysis and Computing Branch,
Environmental Statistics and Information Division, Office of Policy, Planning, and
Evaluation, United State Environmental Protection Agency, Washington DC under
a Cooperative Agreement number CR-821531, and the United States Department
of Interior, National Biological Survey The Contents have not been subjected to
Agency review and therefore do not necessarily reflect the views of the Agency and
no official endorsement should be inferred.

REFERENCES
Brauning, D., (ed.). 1992. Atlas of Breeding Birds in Pennsylvania. University of
Pittsburgh Press, Pittsburgh.
Davis, F.W., D.M. Stoms, J.E. Estes, J. Scepan and J.M. Scott. 1990. An information
systems approach to the preservation of biological diversity. International J.
Geographical Information Systems, 4,55-78.
Gove, J.H., G.P. Patil, B.F. Swindel and C. Taillie. 1994. Ecological diversity and
forest management. Chapter 12 In: G.P. Patil and C.R. Rao, (eds.) Handbook of
Statistics (Environmental Statistics), Vol. 12. North-Holland, New York.
Grassle, J.F., G.P. Patil, W.K. Smith and C. Taillie (eds.). 1979. Ecological Diversity
in Theory and Practice. International Co-operative Publishing House, Fairland,
MD.
Hunter, M.L., Jr. 1990. Wildlife, Forests, and Forestry: Principles of Managing

83
 Forests for Biological Diversity. Prentice-Hall, Englewood Cliffs, NJ.
Hurlbert, S.H. 1971. The nonconcept of species diversity. A critique and alternative
parameters. Ecology, 52, 577-586,
Lund, H.G., D.L. Evan and D.S. Linden. 1995. Scanned, zapped, timed, and
digitized: Advanced technologies for measuring and monitoring vegetation
diversity (Chapter 25, this volume).
MacArthur, R.H. and J.W. MacArthur. 1961. On bird species diversity. Ecology, 42,
594-598.
Magurran, A. E. 1988. Ecological Diversity and Its Measurement. Princeton
University Press, Princeton, NJ.
Patil, G.P and C. Taillie. 1979. A study of diversity profiles and orderings for a bird
community in the vicinity of Colstrip, Montana. Pages 23-48, In: G.P Patil and
M. Rosenzweig (eds.). Contemporary Quantitative Ecology and Related
Ecometrics. International Co-operative Publishing House, Fairland, MD.
Patil G.P. and C. Taillie. 1982. Diversity as a concept and its measurement. Journal
of the American Statistical Association, 77, 548-567.
Pielou, E.C. 1974. Population and Community Ecology. Gordon and Breach, New
York.
Pielou, E.C. 1975. Ecological Diversity. Wiley, New York.
Pielou, E.C. 1977. Mathematical Ecology Wiley, New York.
Scott, J.M., B. Csuti, J.J. Jacobi and J.E. Estes. 1987. Species richness: a geographic
approach to protecting future biological diversity. Bioscience, 37, 782-788.
Scott, J.M., B. Csuti, J.E. Estes and H. Anderson. 1989. Status assessment of
biodiversity protection. Conservation Biology, 3, 85-87.
Scott, J.M., B. Csuti, K. Smith, J.E. Estes and S. Caicco.1991. Gap analysis of species
richness and vegetation cover: an integrated biodiversity conservation strategy.
Pages: 282-297, In: K. Kohm (ed.). Balancing on the brink of extinction: the
endangered species act and lessons for the future. Island Press, Washington,
D.C.
Scott, J.M., F. Davis, B. Csuti, et al 1993. Gap analysis: a geographic approach to
protection of biological diversity. Wildlife Monographs, No. 123.
Swindel, B.F., L.F. Conde and J.E. Smith. 1984. Species diversity: concept,
measurement, and response to clearcutting and site preparation. Forest Ecology
and Management, 8, 11-22.
Swindel, B.F, R. Lloyd, G.P. Patil, C. Taillie and L. Conde. 1987. U.S. National

84
 Forest Management Act, forest ecosystems diversity, and diversity profiles.
Coenoses, 2, 1-8.
Swindel, B.F., J.E. Smith and R.C. Abt. 1991. Methodology for predicting species
diversity in managed forests. Forest Ecology and Management, 40, 75-85.
Turner, S.J. 1995. Scale, observation and measurement: Critical choices for
biodiversity research. (Chapter 7, this volume).
Wilson, E.O. and EM. Peter (eds.). 1988. Biodiversity. National Academy Press,
Washington, D.C.
Zahl, S. 1977. Jackknifing an index of diversity. Ecology, 58, 907-913.

85
 Chapter

6
WHAT DO MEASURES OF BIODIVERSIT Y
TELL US?

Gene Namkoong1

As stated in the background document for this meeting, there is a need to quickly
and accurately measure and monitor biodiversity on an operational scale as a first
step towards effective conservation. It is implicitly assumed in this statement that
if these measures are embodied in operational plans and manuals, the policy
objectives will be achieved if the measures are satisfied. The need to quickly develop
operational scale measures underlines a sense of urgency since forests are being
harvested and otherwise changed so rapidly that any measure that would help
target monitoring, management, or rescue efforts would be useful. Therefore,
estimates of summary statistics are reasonable candidates for use as measures and
one of the questions we have to ask is whether there are conditions which limit the
utility of those measures to guide management.
From an operational perspective, simple measures or summary statistics are
often desirable since they encapsulate considerable information in a few numbers.
They are necessary because decisions must be made before much research on
possibly relevant factors can be conducted. From a biological perspective, the
organisms and the interactions that determine the efficacy of management seldom
lend themselves to reduction to easily measured features. Biologists who under-
stand the complexity of forests are loath to admit that biological processes can be

1 U.S. Forest Service and University of British Columbia, Faculty of Forestry.

87
summarized in simple statistics. They also believe that the complexity of forests is
due to complex interactions, and are not emergent properties of simple and easily
controlled factors. The problem for biologists is that if the processes are so complex
that only vast amounts of detailed research can yield useful guidelines, or if the
processes are so simple that they can be summarized in a few simple statistics, then
either way, biology is irrelevant to conservation management. Particularly when
different measures of biodiversity at different scales are to be used and integrated,
biological detail at any one level will be suppressed, possibly to the vanishing point.
Therefore, a task for this conference is to assess the extent to which simple mea-
sures of biodiversity can tell us enough about its present biological structure and
dynamics that some management can be effectively targeted. If we conclude that
biodiversity is so complex and unknowable that no targeting is possible, then biology
will be largely irrelevant to conservation.

MEASURES OF DIVERSITY
In general, measures are intended to tell us something about things that we think
we can know. Formally, measures are often numbers that indicate size, distance,
extent, or other features of objects. The objects are usually assumed to be known
items, and these quantities describe important features of those objects or the
relationships between them. In fact, there is a whole section of mathematics that
deals with the theory of measures and their features and properties (Doob 1994).
While I will not discuss these concepts, I introduce the topic to indicate that we
are talking about abstractions of the physical world and as such, we must necessari-
ly have a concept in mind about what is a significant and abstractable feature. Thus,
when we measure weight, we think that mass is a significant feature of the object of
interest and that a single number representing one of the features of mass is useful
for comparing the object to another, or that it has something to do with other
properties of the object, such as how easy it is to lift or move.
When applied to biodiversity, we are dealing with a collective measure of
something that has many parts and different distinctive features, is evaluated in
many different value systems, and is already an abstraction. Therefore, its measure
cannot be simple. Nevertheless, we can use features of individuals, populations,
species, etc., their spatial patterns and frequency distributions, to tell us something
about those distributions and about what we think those measures imply about
things we cannot see. While not diminishing the importance of understanding and
describing the features of diversity that we can state with given probabilities, I

88
would like to focus attention on those features of biodiversity that we cannot see,
but about which we would like to make inferences. Among the features that forest
managers may find useful for delimiting areas for sampling and intervention are
the presence or absence of environmental factors associated with diversity; or
whether the taxa are randomly distributed, and diversity is uniform.
To begin with, the simplest features of biodiversity or of any feature of a biological
assemblage, are its average levels and variability over an area of concern such as a
state forest. Just a broad statement of total variability however, is seldom sufficient,
just as average size or stocking density and a variance is insufficient to tell us much
about tree structure. The measures described by Magurran and Pielou in this
conference and elsewhere (Magurran 1988, Pielou 1969) substantially refine our
ideas of total variability and allow us to compare groups or areas with respect to
the frequency distributions of organisms or features of sets of entities. Ever since
early observations of the distribution of species showed that different sites have
characteristically different curves relating species abundance to sample area, ran-
dom models of species assembly could be fitted and parameters derived to
characterize area differences. Various indices could also be derived for the number
and relative frequencies of species that reflected species richness or packing that
were relatively stable with respect to sample size. These indices could then be
estimated and, as they varied from site to site and characteristically differed among
types of organisms, inferences could be drawn about the structure of diversity.
One of the uses to which these indices have been put is to discern relationships
between diversity and factors of the environment, such as nutrient or soil quality,
distance from source populations of invaders, frequency and type of disturbance,
etc. The attempt is often to discern causal factors by relating diversity measures to
some set of physical factors, for some set of organisms. In one study conducted by
the U.S. Forest Service in a 5,000 ha. forest in the southern Appalachian Mountains,
some fifty plots were simultaneously surveyed for trees, shrubs and low vegeta-
tion, herbaceous vegetation, birds, reptiles and amphibians. In this study, we found
good agreement among various simple measures of biodiversity, including
Simpson’s and Shannon’s Indices, and species richness, and that there was
substantial variation in all of these measures among plots for all types of orga-
nisms. We also found that some plots had high diversity for more than one class of
organism and some were low for all classes.
In order to determine the extent to which certain ecological relationships with
possible causes of diversity could be discerned, correlations of diversity with several
variables were estimated. If there were site qualities that contained generally optimal

89
levels of available moisture or nutrients or high variation in those variables, then
there should be some sites that are high in diversity for all types of organisms.
High diversity might also be correlated with other measures of biomass produc-
tivity. If site quality for diversity was related to physical features of the environ-
ment, then we should be able to detect regressions with measures of independent
variables, such as the mean or variance in elevation, soil moisture, stand age, etc.,
and may then infer some causal mechanisms. If diversity was strongly affected by
biotic relationships among mutualists, competitors, symbionts, etc., then sets of
species should have similar distributional attributes. Then, if relationships were
relatively simple, a few guidelines might be inferred for sampling and managing
diversity.
We found that some generalizations could be derived. Older stands and sites
situated at lower elevations had higher diversity than comparable younger or high-
er elevation sites for several classes, but not for all. In general, the alpha diversity
was not well focused if all types of organisms were considered. We could identify
sets of plots that captured most of the alpha and beta diversity such that less than
20% of the sites contained more than 90% of the species. Also, beta diversity was
related primarily to elevation gradients and hence physical variability could account
for diversity only in large scales and only generally in the one factor. Nevertheless,
some drainage basins that contained high alpha and beta diversity in compact areas
were delineated which would be highly efficient for sampling total diversity. In
general, we could find some correlation of diversity for each organism type with
some site features, but the correspondence of diversity between classes of orga-
nisms was quite low.
From the preliminary analyses, it was immediately obvious that we could detect
no pattern of sites that were generally rich for everything. This might be expected
either because sites occupied by one species or set of species would exclude some
others by chance, by preferential migrations to alternate sites, or perhaps by
mechanisms of competitive exclusion. When we investigated correlations of species
richness or diversity with biomass or with physical features of the sites, no general
relationships were found. There were particular subsets of species within classes
that had high diversity in identifiable site types such as a positive association of
salamanders with moist sites, and a negative association of herbaceous vegetation
with trees on most sites. Patterns could be detected, but only if finer subdivisions of
site factors and species groups were used. Therefore management systems cannot
simply identify rich versus poor but must specify the kinds of richness desired.
Furthermore, we identified rare species of plants in the areas studied and found

90
that their presence or absence was not strongly correlated with either general species
diversity nor with any single environmental factor. It seems that rarity is associated
with different phenomena, some because certain sites are rare and others because
of a general and diffuse rarity. This again may not be a surprise to the botanists and
ecologists who are familiar with this vegetation, but it does imply that for these
species, even finer individual species management plans have to be considered
since there are no other obvious indicator variables that can be used. Some of these
species may well be adapted to rarity and their rarity does not in itself pose a sur-
vival problem. However, even then, their dependence on maintaining a specific
density may be critical for survival or reproduction, and detailed life histories would
be needed before a rational conservation plan could be devised.
No single index is sufficient for describing the distribution of diversity. It may
still be possible, though, to decompose the total set of species into subsets and to
generate at least a moderately simple management plan without having to research
every species and every pair of species-environment relationships. These sets may
be groups of similarly behaving species of the same class that can competitively
coexist, or may be members of a mutualistic association or food web of species in
different classes or trophic levels. It was possible to discern some general patterns
of sites and diversity for subsets of species that suggest priority areas for
management programs by using correlates of diversity indices in these areas.
However, there is enough variation from a broadly inclusive and uniformly useful
pattern, to question the development of a conservation plan based on only this
information. Further, there is no information on population dynamics, so that by
sampling and conserving only the targeted areas, it is not likely that we can main-
tain the present diversity. This is asking diversity indices to do far more than they
were intended to do. In the words of Magurr an (1988), "Diversity measures are
valuable, but are only a means to an end. That end is that ecologists should be able
to ask the questions and formulate the hypotheses to help them understand, and
sensibly manage, the natural world. ”Since it is the ecosystem that we manage that
produces diversity, we try to avoid the mistake of managing for measures of diver-
sity that may result in poor ecosystem management and the ultimate diminution of
diversity.

TAXONOMIC DIVERSITY
While questions of dynamics require other kinds of studies and surveys than can
be considered here, we can consider other refinements of general diversity indices

91
that may assist conservationists in targeting their efforts. Since all forms of
conservation management will require starting with a finite sample of forests and
managing finite populations and stands, efficiency in choosing initial samples and
targeting types of management activity is highly desirable. Some forms of structural
diversity for example, can be managed most effectively by site or silvicultural
treatments. Taxonomic diversity, on the other hand, is highly dependent on the
initial sampling and slow to respond to area management. Some of the diversity
indices mentioned above can be modified to include hierarchical levels of taxo-
nomic diversity (Pielou 1969) and hence provide more refined descriptors of the
richness of phylogenetic differentiation.
Phylogenetic information is often available and provides considerably more
information on genetic diversity than only the number and frequency of species.
When that information is available a conservation objective for diversity can be to
distinguish plots that may have equivalent distributions of species but different
phylogenetic diversity as represented by those species. Sampling for diversity may
be considered to be inefficient if we include only a small range of taxa, such as only
Pinus or only Compositae, but more efficient in some sense if more phylogenetical-
ly diverse taxa are sampled. Thus, the total number of species, or frequency weighted
numbers, are very useful as ecological indicators but may not be as useful for
conservation purposes as measures that include genera, families, orders, etc.
To construct a measure of taxonomic diversity we can consider measures that
have been used for phylogenetic analysis as indicators of evolutionary relation-
ships as well as of phenetic similarity. For purposes of this paper, I will not
discriminate between phenetic versus cladistic approaches for inferring phylogeny
but will instead assume that phenotypic, genetic, and historical evolutionary data
are informative and any or all are used to define distances between genotypes and
taxa. This is a useful, but only approximately valid, assumption that allows us to
consider that distances can be estimated between taxonomic units. If the distances
are metrics then various descriptions and properties of evolutionary trees can be
used for phylogenetic analysis and some features can be summarized into indices
of average or total similarity or dissimilarity. Several types of indices come to mind
that, as far as I know, have not been explored for application to biodiversity
conservation. They include total taxonomic branch length, number of cladistic nodes,
or the volume and dimensionality included in a convex hull of inter-taxon distances.
Recent research on phylogenetic trees and other kinds of network phylogenies (Smith
1989, Eigen et al. 1988) present interesting new possibilities for estimating
phylogenetic topologies even when evolution has not proceeded hierarchically.

92
 With other measures of taxonomic diversity available, it is still possible to
consider which environmental variables may be associated with that diversity. A
single measure of total diversity such as the volume and dimensionality measure
mentioned above, could be correlated with environmental variables. There would
also be several levels of diversity such as at the family or order level that could now
be used. In addition, since taxonomic diversity measures often have phylogenetic
distance measures associated with them, it is possible to determine whether the
matrix of phylogenetic distances or their complementary similarities are associated
with matrices of distances between the taxonomic units as measured by geographic
or environmental similarities. By means such as described by Leduc et al. (1992),
and testing for the consensus of distances (Lapointe and Legendre 1992) it is possible
to directly examine whether patterns of taxonomic diversity are related to patterns
of environmental diversity.

CONSERVING DIVERSITY
Measures of diversity can clearly help us to locate and to begin to understand the
present and possible future states of its distribution. They can be more useful if the
measures can be sufficiently refined as to inform us of their associations and possible
causative mechanisms, but we cannot expect that these relationships will be simple
or easily accommodated into forest management plans. We might be able to
understand the extent to which historical events determine how species packing
and assembly rules are the result of initial conditions, and how physical and biotic
environmental factors control the evolution of species and communities, but these
also cannot be easily translated into conservation plans. We do not derive information
on evolutionary dynamics from static samples and hence cannot discern whether
the sample is from a stable and strongly attracting equilibrium, or if we are in
unstable and transitory phases that may have more complex dynamics. For these
questions, we still, and always will need, more biology.
What the measures can do for us however, is to pinpoint our ignorance so that
we can put priorities on research, and perhaps indicate whether more research will
ever provide enough information, soon enough, to guide policy. On the one hand,
it is conceivable that the distribution and biological dynamics of most species are
so simple that forest policy and management can use one index that suffices for
sampling and conserving all biodiversity in perpetuity. On the other hand, it is
conceivable that the population and evolutionary dynamics of many species are so
complex and intertwined with other species that no general rules or conservation

93
 targets can be defined. Using currently available statistics and estimable parameters,
initial indications are that we can derive some predictions about the present loca-
tions of diversity but that those measures and predictions require at least moderate
levels of environmental and taxonomic refinement. At this point, we do not know
how much more refined we have to be in order to derive feasible targets for forest
management to capture the present levels of diversity. We also have no indication
of how much more research is needed to define management for conserving diver-
sity in the foreseeable future.
In this uncertain situation, a complete conservation program cannot be well
described, but an adaptive program can be suggested. Most obviously, a core set of
sites and areas can be delineated that would at least include a great majority of
species. In the case of the sampled forests in the southern Appalachians, more than
90% of all species of all types could be included in less than 20% of the area. Rare or
sparsely distributed species could be specially treated in designated areas that would
require only small additions to the core. Below the species level, further genetic
sampling may be required, but in most cases can probably be best done by sampling
different areas for species with genetic variation that is widely distributed. Above
the species level, the sampling would not necessarily include the same levels of
structural or community diversity, but once designated, additional areas can be
designated as increments to a core structure. If this kind of an incremental
designation can form a plan for conservation, then the measures of diversity already
available can be used to design a programme, even if they are not as refined as they
can be nor as refined as biologists would like them to be. It can also be expected
that further developments in diversity measures and their analyses will make them
even more useful.

LITERATURE CITED
Doob, J.L. 1994. Measure Theory. Springer-Verlag. New York. 210 pp.
Eigen, M., R. Winkler-Oswatitsch and A. Dress. 1988. Statistical geometry in
sequence space: A method of quantitative comparative sequence analysis. Proc.
Nat. Acad. Sci. USA. 85: 5913-5917.
Lapointe, F.J. and P. Legendre. 1992. A statistical framework to test the consensus
among additive trees (cladograms). Syst. Biol. 41: 158-171.
Leduc, A., P. Drapeau, Y. Bergeron and P. Legendre. 1992. Study of spatial
components of forest cover using partial Mantel tests and path analysis. J. Veg.

94
 Sci. 3: 69-78.
Magurran, A.E. 1988. Ecological Diversity and Its Measurement. Princeton Univ.
Press. Princeton, N.J. 179 pp.
Pielou, E.C. 1969. An Introduction to Mathematical Ecology. Wiley-Interscience. New
York. 286 pp.
Smith, J.M. 1989. Trees bundles or nets? TREE. 4: 302-304.

95
 Chapter

7
SCALE, OBSERVATION AND MEASUREMENT:
CRITICAL CHOICES FOR
BIODIVERSITY RESEARC H

Sandra J. Turner 1

INTRODUCTION
Biodiversity has taken on a meaning in the consciousness of the world’s societies
that goes far beyond the original definition known to science. Biodiversity has
become a mantra and a rallying cry for maintaining ecosystem functions which
have been identified as socially important for many reasons. Biodiversity has become
the focus of ecosystem management and a great deal of political debate.
In addition “diversity” now has considerably more meaning to scientists than it
did in the last century. The fact that species diversity is not a steadily increasing
historical trend such as. Wallace (1876) and Willis (1922 citation in Ricklefs and
Schluter 1993) suggested has been brought home to us rather dramatically with the
loss of species after species; the result of human manipulation of ecosystems. Early
scientists thought that diversity was an artefact of time and that the amount of
diversity measured was related to the age of the community which was being

1 Biology Department, University of New Mexico, Albuquerque, NM 87131, USA.

97
investigated. These assumptions may be true for some communities but impor-
tantly, science has moved well beyond this notion of diversity. Diversity is dynamic
in space and time; both upward and downward fluctuations occur. Membership in
a community is a function of the spatial and temporal dynamics of the community
in which it is embedded. As Ricklefs and Schluter (1993) suggest, it is necessary to
recognize that ecology, evolution, geography, and history are different facets of a
single set of processes and the patterns they generate. One cannot isolate one system,
of a particular dimension, from processes and structures at a smaller scale embedded
within it, or from those at a larger scale containing it.
Species diversity is the type of diversity which springs most quickly to the mind
of both scientists and lay people whenever diversity is mentioned. However, it is
not the only level of diversity important for modern investigation and for
measurement and monitoring of resources. Allele diversity, genetic diversity,
polygenic genetic diversity, species diversity, patch diversity, habitat diversity,
community diversity, landscape diversity, regional diversity: at every scale within
the biotic system diversity is under investigation. Each scale has its own describa-
ble dynamics and, in addition, many features of diversity are common to all scales.
In some cases the diversity at one scale can be shown to relate to diversity at
another scale (Franklin 1993). This relationship may provide a method for asso-
ciating diversity at a scale which is easily measured (e.g., vegetation structure
measured from remotely sensed data) with diversity at a scale which is more diffi-
cult to measure (e.g., diversity of the animal community living in that vegetation).
This paper addresses the problem of identifying the appropriate observation
and measurement criteria, and the appropriate time and space scales for
measurements. It is also about the dynamics of change - how can we find out how
diversity may be altered with time.? I am going to begin with landscape ecology
and hierarchy theory and use that as a backdrop for making these critical choices
about measurement and monitoring. My examples will be from measurement but
the application is the same for monitoring.

LANDSCAPE ECOLOGY
Landscape ecology is an integrative science that focuses on the way ecological
systems are arrayed in space and through time. Landscape ecology covers such
diverse topics as: (1) elements in a park created for recreation; (2) the functional
analysis of agricultural landscapes; (3) analysis of human flow through park and
wilderness areas; (4) analysis of beetle eye view of the world to understand fractal

98
 patterns (Wiens and Milne 1989; Johnson et al . 1992); (5) understanding what reserve
size means for maintaining species diversity (Margule et al . 1988; s Nicholls and
Margules 1993; and others); (6) how fire influences forest landscapes (Turner and
Romme 1994); and (7) the integration of human values into long term monitoring
of ecological processes, such as the U.S. Environmental Monitoring and Assessment
Program - Landscape Ecology Resource Group (EMAP-LE) seeks to accomplish.
Although landscape ecology usually involves temporal considerations, the
unifying theme is that landscape ecology always involves space. For instance,
Whittaker’s (1972) work which treated large spatial scales is an example of land-
scape ecology, even though it ignored temporal considerations. In contrast, the
work of Lotka (1925) and of Volterra (1926), cannot be considered landscape eco-
logy because, while it considered the temporal dynamic, their important work did
not consider space.
Moreover, the spatial scale at which landscape ecologists work is often
considerably broader than traditional ecology. Traditional plant ecology, for
example, has focused inordinately on scales of 1m2 and one growing season.
However, the distribution and abundance, and processes affecting a species at the
community level is partly a function of the landscape in which it lives. By
acknowledging this interaction, a landscape study adds spatial heterogeneity to a
population, community, or ecosystem study.
Expanding the spatial scale and including heterogeneity can reveal how the
distribution and abundance of species within a local community is influenced by
the larger landscape context in which it is embedded. As larger areas are consi-
dered, longer-term temporal processes become important determinants of ecolo-
gical dynamics.
Understanding phenomena within their landscape context can provide
important insight for management. The original efforts to save the Northern Spotted
Owl (S trix occiden talis) in the northwest United States focused primarily on the
total amount of reserved habitat without considering the ability of juveniles to
disperse to suitable habitat within the landscape. Dan Doak and colleagues (Simon
Moffat 1994) found that the original patchy preserves were especially vulnera-
ble to extinction and showed that the arrangement and size of habitat patches
on the landscape was as important as the overall amount of habitat. Thus, patch
level diversity at a larger scale can influence species diversity and species
distributions at local and regional scales.
The landscape component of Ecological Measurement and Assessment
Program of the United States Environmental Protection Agency (EMAP-LE) is

99
expected to tie together studies at all levels to yield a regional analysis (Figure 1).
EMAP-LE is using societal values to determine the direction and focus for monitoring
and measurement. Water “purity”, which is valued by society for health, recreational
and aesthetic reasons, is analyzed with fine scale measurements - but water quality
is the integration of landscape structure and function at the regional scale.

PRINCIPLES OF LANDSCAPE ECOLOGY

The consequences of adding broad-scale spatial heterogeneity and long-term
temporal dynamics are profound. The spatial dimension allows for integration
across studies from studies of small organisms to regional analysis. It allows for
the analysis of impacts which are at scales different from those normally incor-
porated into the system. Expanding the time scale of the temporal dynamics allows
us to integrate history, both abiotic and biotic, and the activities and cultures of
human beings into our analysis.
Three properties of landscapes form the core of the discipline. First, landscapes
in general have structure. That is, there are spatial relationships between elements.
The relationship may be between trees and gaps in a forest. At a broader scale there
is a relationship between forest and agricultural patches in a landscape mosaic.
Landscapes have spatial continuity between these elements but they are not
considered to be homogenous. Landscapes exhibit pattern and internal
heterogeneity. It is that heterogeneity which is of interest. The continuity of the
landscape suggests that it also has boundaries where landscape elements come
together. Boundaries and barriers may prevent individual species from respon-
ding to changes by shifting between habitat types or sites and that, in turn, may
result in decreased diversity (Vos and Opdam 1993). Boundaries, natural and
anthropogenic, are very important in designing research and monitoring for
biodiversity.
Second, landscapes also have function. That is, there are interactions among
spatial elements. These may be flows of energy, materials and species or even genes
between component elements of the landscape. Sometimes these ecological
functions become important in a societal context. For instance, a corridor may
provide access between interbreeding populations of important species and so
help to maintain species diversity. A wetland may function as a filter for pollutants
and maintain or increase water purity within the system. These are ecological
functions which are a service to society.
Patterning and patchiness of both structure and function can be recognized at
virtually every scale of investigation. Thus landscapes may be of any size. In fact,

100
I suggest that the notion of landscape is scale independent in the same sense as
ecosystem is scale independent. This is a widely held view (Allen and Hoekstra
1992 and others). However, some researchers, among them Richard Forman
(Forman and Godron 1986), hold that landscapes occupy a definable and limited
portion of the spectrum of ecological scales. They view landscapes as occupying
the scale that incorporates a number of ecosystems and is yet smaller than regional.
I do not view that scale as definitive. It is very useful and practical to view the
landscape concept as scale independent because it can then be easily applied to
such notions as diversity. Diversity occurs at many different spatio-temporal
scales.
Third, landscapes change. There is an alteration of structure and function of the
ecological mosaic (the elements arrayed in space) over time. Landscape change
may result in, or conversely be the result of, changes in species frequencies or
abundances or even of local extinctions. Clearly change is a natural process and
natural changes in dynamical ecological systems are integrated from the smallest
organisms in a stream to the watershed and regional levels (Figure 1). Change may
be the normal turnover in species in a forest which varies in time or in space with
the formation of gaps and the resulting successional dynamics. Alternatively,
change may be anthropogenically induced, such as by the clearing of an
agricultural field or the harvesting of timber from a forest.

HIERARCHY THEORY
Landscapes can also be profitably viewed as hierarchical in organization. The
landscape is a hierarchy of differently scaled structures and functions (Figure 2).
The structure and function at a particular level which we choose to observe, say a
forest, has a particular time and space scale. This is the "N" level. The lower level
(N-l) is composed of sub-systems of the forest. These may be patches or gaps or
soil differences, or individual species which have smaller and quicker spatio-
temporal dynamics and provide the mechanism for creation and maintenance of
the forest. The next level (N+l) is larger in space but changes more slowly. The
larger slower level may be the region which exerts control on the forest.
Hierarchy theory (O’Neill et al. 1986) applied to landscapes implies this upper
level control and lower level mechanisms and provides an epistemology for
measurement based on the scale of observation and level of analysis. For instance,
growth of a plant (N) is controlled by physiological processes which are N-l. Of
course, the level of interest (N) and the N-l, N+l levels are set by the problem
which is being investigated and may not be the same hierarchy for all problems.

102
around important interactions. This is the art. Leave out everything that is
unimportant in sorting through the complexity. Find the correct criteria to address
the problem. The hard part is to determine the correct level in the system, that
which is truly important to the observer. It may be a single species, a biotic
community or an ecosystem type. Choose the important level N. Then carefully,
remaining within the same hierarchy, identify the N+l level of control and the N-l
level of mechanisms. These drive landscape investigations whether they are field
tests, modelling, applied or theoretical.

WHAT CAN LANDSCAPE ECOLOGY DO FOR DIVERSITY RESEARCH?

What can landscape ecology and the concept of ecological hierarchies contribute to
measuring and monitoring forest biodiversity? Classical diversity theory considers
population processes and ecological interactions at a local scale in small uniform
habitats. But these small time and space scales are not sufficient to understand
diversity patterns (Schluter and Ricklefs 1993), particularly at larger scales.
Patterns of diversity, whether we are interested in genetic, species, habitat,
ecosystem or landscape diversity are arrayed in space. The measure which we
define as diversity is in part a function of the size, shape and temporal dynamics of
the area considered and of our observational protocols. Many researchers have
included spatio-temporal dynamics in diversity research and have found that the
landscape perspective often provides considerable explanatory power.
Species-area curves clearly indicate that diversity has a spatial component. More
than just the extent of space occupied by species, structure, function and change of
landscapes are also implicated.
Two decades ago Pianka (1974 and others), and more recently Franklin (1993
and others), have shown that species diversity is related to the structure of the
forest. Genetic diversity may be maintained by the structural array of refugia for
some populations according to Shepherd and Brown (1993). There is evidence that
variability in polygenic genetic diversity within salmon populations is partitioned
temporally and spatially (Gharrett and Smoker 1993). Rosenberg and Raphael (1986)
have show that abundance is associated with patch size. Rey-Benayas and Scheiner
(1993) show that biodiversity at the mosaic and community levels and landscape
complexity respond to geochemical function of the landscape at those scales. Pearson
(1993) found that variation in bird species richness and diversity was explained
solely by landscape variables, structure and function. Franklin’s (1993), work
suggests that “structural complexity offers the abundance of habitats that in turn

104
support a large array of specialized species. Directly and indirectly the structural
complexity provides biological and functional diversity”. And of course, anyone
who is familiar with the monsoon and fire driven forest of Kakadu (Woinarski 1993)
and northern Australia are fully aware that change is paramount. If you are not
there at the correct time, or many times, all measures of diversity will be suspect.
Spatio-temporal effects on diversities are the results of events ranging from those
that occur over very long evolutionary scales of time to recent events (Blonde1 and
Vigne 1993), such as the last bulldozer.
Thus landscape ecology provides the tools to describe and understand how the
diversity of species, alleles or habitats are arrayed in space and though time.
Hierarchy theory within this context suggests methodology for making decisions
about observation scales and measurement.

WHAT CAN HIERARCHY THEORY DO FOR DIVERSITY RESEARCH?

Figure 2 suggests that there is a focal level “N” to which research, measurement
and monitoring should be directed. It also suggests that the research, whether
modelling of ecological systems or field based research, need only account for three
levels as suggested by Johnson (1995). Remember that functions at one level produce
the structure of the next (i.e., physiological processes grow and maintain the trees).
Often we find that it is change or variability at the higher level which controls or
constrains the system (i.e., abiotic processes constrain forest expansion). Diversity
fits this pattern. Landscape diversity is at least in part the result of the structure of
vegetation. Population processes are the mechanism which create diverse
communities.
The level at which the inquiry begins is arbitrary (Allen et al. 1987). This decision
sets the answers to the questions of observations, scales and measurements. “N” is
defined uniquely for the question at hand.“N” may be defined by purely scientific
interest or in response to ecological crisis or by societal concerns as is happening
more frequently in North America. It may also be defined politically.
Figure 3 is one (far from unique) conception of a phenomena and measurement
criteria hierarchy. This is a vegetation hierarchy. It could have been any set of
interesting and related phenomena. More familiar are space-time diagrams in which
structures, functions, systems or other lists of ecologically interesting and related
things are arranged along a diagonal of their inherent time and space scales. By
related, we mean that there is a connection between the processes and functioning
of one level and the expression of phenomena at a higher level. In this figure

105
level of the plant is the N-l level of mechanisms and community is the N+l control
level. Among the criteria which could be chosen to describe the population level
are intraspecific competition effects, genetic diversity, and the spatial dispersion
and biomass production of the population, and dispersal or invasion rate of
propagules. Understanding of the mechanisms which create and maintain the
population would come from the lower level of the single plant (N-l) where
reproduction and growth potential are important. The limits and context of the
population will be found in the community (N+l) interactions of interspecific
competition, for instance, between various populations.
Measurement of species diversity, a criterion at the community level, is
straightforward in this context. The number and spatial extent of populations are
the N-l metrics which are important if all you want is a one time answer to the
question: how many species are there out there right now?
Understanding the dynamics of species diversity is a different question and it is
paramount for monitoring. Certainly counting populations remains important,
however it is necessary to understand how those populations are likely to change
in the future. In this context the population will be profitably understood both as
the mechanism which creates the community and as controlled by the community.
The controls upon species diversity dynamics may be very complex and involve
all the factors that have been discussed including structural and functional
components, current and past history of the region under investigation and, most
likely, the variability at which controlling functions and structures can be expected
to change. It may well also include some measure of society's value of that particular
community and the probability that anthropogenic influences will affect species
diversity.

ARE DIVERSITY INDICES AT ONE SCALE RELATED TO DIVERSITY AT

ANOTHER?
The complexity of the controls of biodiversity have sent researchers scurrying to
find surrogates which are easy to measure and which integrate all the processes
which control diversity (much as water quality is an integration of regional scale
structure and function) or which change in the same way as biodiversity but are
easier to monitor. Of particular interest to monitoring is the question of the
relationship of diversity at one level in the hierarchy of phenomena and
measurement of diversity at another scale.
Regional or landscape complexity has been implicated in patterns of diversity

107
at these large scales. Can indices of large scale diversity be used to infer anything
about diversity at different levels in the system ? At the largest scale, remotely
sensed image data can be analyzed to yield a metric of landscape diversity which
is, in fact, a measure of variance between pixel colour values for several spectra.
These are being shown to relate to large scale structure. Many researchers are using
remotely sensed data to characterise natural vegetation using colour and textural
measurement (Bijlsma 1993 and others). Using satellite images, B.T.T. Burns (Bums
et al . 1994) is able to delineate old growth forest stands from a dense matrix of
multiple use forests having multiple histories in northern Mexico. Low aerial
reconnaissance confirms his classification. These are structural components of the
landscape which are available to us from remotely sensed data.
As mentioned earlier, the structural components of a landscape do, in fact, help
to structure the distribution of species at all levels. If landscape structure can be
derived from satellite images and diversity can be derived from structure, then
indices of landscape diversity at these largest scales can be shown to relate to
species diversity. Thus, it does seem possible to make this translation across scales
in some cases. Further development will add important techniques for measure-
ment and monitoring of diversity. It remains to be seen how indices of species
diversity may relate to diversity at lower levels such as genetic diversity within a
population.

SUMMARY
Diversity is apparent at every scale in the ecological system. Landscape ecology
and hierarchy theory provide a way to understand and sort though the complexity
of the system so that investigations can focus on that information which will most
clearly answer the research question. The level or levels in the system which are
investigated will depend upon the nature of the question and the type of answer
required. A single time answer to diversity may be found by simply counting
numbers of species, but the dynamics of diversity will require the investigation of
several levels in the system. At whatever the scale of investigation, genetic diversity,
population, community or ecosystem, and with whatever techniques chosen to
conduct the investigation, the vital research that must be done is to understand the
dynamics which generate and maintain diversity. In order to do that, it is necessary
to understand and investigate the lower level mechanisms which create diversity
and the higher level controls upon diversity.

108
REFERENCES
Allen, T.F.H. and T.W. Hoekstra. 1992. Toward a unified ecology. Columbia
University Press.
Allen, T.F.H., R.V. O’Neill and T. W. Hoekstra. 1987. Interlevel relations in ecological
research and management: some working principles from hierarchy theory.
Journal of Applied Systems Analysis 14: 63-79.
Bijlsma, R.J. 1993. The characterization of natural vegetation using first-order and
texture measurements in digitized, colour-infrared photography. International
Journal of Remote Sensing 14: 1547-1562.
Blondel, J. and J. Vigne. 1993. Space, Time, and Man as Determinants of Diversity of
Birds and Mammals in the Mediterranean Region. Pages: 135-146 In: R.E.
Ricklefs, D. Schluter (eds.) Species Diversity in Ecological Communities. Univ.
of Chicago Press, Chicago.
Burns, B.T.T., M. Drees, D. Hadley and W. Laird. 1994. Proyecto De Recursos
Tarahumara. Technical Report: Biodiversity Support Program; Available from:
Native Seeds / SEARCH, Tucson, AZ 85711, USA.
Forman, R.T.T. and M. Godron. 1986. Landscape Ecology. John Wiley and Sons,
NY.
Franklin, G.F. 1993. Lessons from old-growth. Journal of Forestry 12: 11-13.
Gharrett, A.J. and W.W. Smoker. 1993. A perspective on the adaptive importance of
genetic infrastructure in salmon populations to ocean ranching in Alaska.
Fisheries Research 18: 45-58.
Johnson, A.R. (In Press). Spatiotemporal Hierarchies in Ecological Theory and
Modeling. In: Integrating geographic information systems and environmental
modeling, GIS World, Ft. Collins, CO.
Johnson, A.R., B.T. Milne and J.A. Wiens. 1992. Diffusion in a fractal landscape:
Simulations and experimental investigation of Tenebrionid beetle movements.
Ecology 73: 1968-1979.
Lotka, A.J. 1925. Elements of Physical Biology.
MacMahon, J.A., D.L. Phillips, J.V. Robinson and D.J. Schimpf. 1978. Levels of
biological organization: an organism-centered approach. Bioscience 28:700-704.
Margules, C.R., A.O. Nicholls and R.L. Pressey. 1988. Selecting networks of reserves
to maximize biological diversity. Biological Conservation 43: 63-67.
Nicholls, A.O and C.R. Margules. 1993. An upgraded reserve selection procedure.
Biological Conservation 64: 165-169.

109
O’Neill, RX, D.L. DeAngelis, J.B. Waide and T.F.H. Allen. 1986. A hierarchical
concept of ecosystems. Princeton University Press.
Pearson, S.M. 1993. The spatial extent and relative influence of landscape-level factors
on wintering bird populations. Landscape Ecology 8: 3-18.
Pianka, E.R. 1974. Evolutionary Ecology. Harper and Row, New York.
Ray-Benayas, J.M. and S.M. Scheiner. 1993. Diversity patterns of wet meadows
along geochemical gradients in central Spain. Journal of Vegetation Science 4:
103-108.
Ricklefs, R.E. and Schluter, D. 1993. Species Diversity in Ecological Communities.
Univ. of Chicago Press.
Rosenberg, K.V. and M.G. Raphael. 1986. Effects of forest fragmentation on
vertebrates in Douglas-fir forests. Pages: 263-272 In: J. Vemer, M.I. Morrison
and C.J. Ralph (eds.) Modeling habitat relationships of terrestrial vertebrates.
Univ. Wisconsin Press.
Schluter, D. and Ricklefs, R.E. 1993. Species Diversity: An Introduction to the
problem. Pages: l-10 In: R.E. Ricklefs and D. Schluter (eds.) Species Diversity in
Ecological Communities. Univ. of Chicago Press.
Shepherd, S.A. and D.L. Brown. 1993. What is an abalone stock: Implications for
the role of refugia in conservation. Canadian Journal of Fisheries and Aquatic
Sciences 50: 2001-2009.
Simon Moffat, A. 1994. Theoretical Ecology: Winning its spurs in the real world.
Science 263: 1090-1092.
Turner, M. G. and W.H. Romme. 1994. Landscape dynamics in crown fire ecosystems.
Landscape Ecology 9: 59-65.
Volterra, V 1926. Variation and fluctuations of the number of individuals in animal
species living together. Pages: 409-448 In: R.N. Chapman (ed.) Animal Ecology.
McGraw-Hill.
Vos, C.C. and P. Opdam. 1993. Landscape Ecology of a Stressed Environment.
Chapman and Hall.
Wallace, A.R. 1962. The geographical distribution of animals. Reprint, New York: 2
vol.; Hafner.
Whittaker, R.H. 1972. Communities and Ecosystems. The Macmillan Company.
Wiens, J.A. and B.T. Milne. 1989. Scaling of ‘landscapes’ in landscape ecology, or,
landscape ecology from a beetle’s perspective. Landscape Ecology 3: 87-96.
Willis, J.C. 1922. Age and area: A study in geographical distribution and origin in
species. Cambridge University Press.

110
Woinarski, J.C.Z. 1993. A cut and paste community: Birds of monsoon rainforests
in Kakadu National park, Northern Territory. EMU 93: 100-120.

111
 Chapter

8
MONITORING AND MEASURING
FOREST BIODIVERSITY IN THAILAN D

B. Boontawee1, C. Plengklai1 and A. Kao-sa-ardl

INTRODUCTION
During the past decade “biological diversity” or “biodiversity ” has become one of
the most popular topics for discussion both as scientific and political issues and at
national, regional and global levels. The main theme of the discussion is the
contribution of biological diversity to social and economic development; losses of
biological diversity and their causes; how to manage the remaining biological
diversity for sustainable utilization; the measurement for conservation (in situ
and ex situ) of biological diversity and techniques for measuring and monitoring
of biological diversity.
“Biodiversity” is defined as the variety and variability among living organisms
and the ecological systems in which they occur. Generally, biodiversity is divided
into three components, i.e. ecosystem or ecological diversity, species diversity and
genetic diversity. (Kapoor-Vijay 1992, Sandlund et al. 1992). It has been recognized
that tropical forests are the major source of global biodiversity, and they are also
great producers of biological resources for human welfare.

1 Royal Forest Department, Bankhen, Bangkok, Thailand.

113
 Biological resources have provided tremendous benefits to human beings and
McNeely et al. (1990) have classified their direct and indirect values. Despite this,
biological resources are still being destroyed at high rates, as mentioned previously
Therefore, appropriate measures are needed to conserve the existing biological
resources, so as to maintain and to improve their productivity.
This paper is intended to summarize the status and activities undertaken in
measuring and monitoring of biodiversity in Thailand, with particular reference to
forests and forest tree species.
Thailand is located in the Southeast Asian region, covering the latitudes between
6 and 20o N. The country’s area is 513,115 km with a forest area of 136,698 km or
2 2

about 27% of the country’s area (RFD 1992). The range of elevation is from sea
level up to 2,200 m. Climatic conditions vary from the lowland humid tropics to
alpine and/or subtropical types.
Thailand is one of the most bio-resource rich countries of the world. This is due
to its biogeographical location, which is at the junction of the three main floristic
regions, namely the Indo-Burmese, the Indo-Chinese and the Malesian regions
(Smitinand 1994). The Indo-Burmese floristic region is in the northern, northwestern,
and western parts of the country. The Indo-Chinese floristic region is found in the
northeast, whereas the Malesian floristic type is found in the southern peninsular
and in the eastern part of the country.
Due to population pressure, the forest area in Thailand has been depleted rapidly
Both identified and unidentified ecosystems, species and their genetic resources
are being eroded. There is an urgent need to explore, identify, protect, and manage
the available forest biodiversity properly for future sustainable utilization.

MONITORING OF FOREST BIODIVERSITY

Assessment of Forest Area
As mentioned above the present forested area in Thailand is 136,698 km2, which is
about 27% of the country’s area. Figures illustrated in Table 1 show that the rate of
deforestation during the past 5 years is relatively high, about 3,000 km 2 per year.
This reflects the fact that the status of forest biodiversity, in terms of habitat, forest
ecosystem, species, population and genetic diversities seems to be endangered.
The major causes of deforestation are due to (1) population pressure increasing
the demand on land and forest products and (2) the improvement of the country’s
physical infrastructure, such as road and dam construction etc.
Despite a shift in government policy on environmental conservation, the rate of
forest degradation and deforestation is still high. The core bio-resource and unique
forest ecosystem areas have been identified and declared as biological protected
areas in various forms and for various functions. These protected areas, including

114
MEASURING FOREST BIODIVERSITY
To manage forest biodiversity for sustainable utilization, it is essential to know and/
or to understand the structure and characteristics of forest ecosystems. The most
commonly used method of measuring forest biodiversity is to establish sample plots
replicated on a number of sites of the same, or similar forest ecosystems. Various
techniques of forest inventory have been applied for assessment, measuring and
monitoring of forest biodiversity in Thailand. The common techniques are the line-
plot and plotless (point centred quarters) methods for large scale inventory, the
block/plot with replication method for intensive study, and the establishment of
permanent plots for monitoring of population dynamics in long term ecological
research site studies (LTERS). Based on species-area curve studies, the plot size
for measuring of forest biodiversity is between 0.1 - 1.0 hectare, depending on the
forest types (Thammincha 1993); that is, the greater the stand density, the larger
sample plots used. Within the sampled plots, number of species, number of
individuals, trees, saplings, seedlings and undergrowth are counted to assess species
richness and tree density. For species diversity, the Shannon-Wiener’s species
diversity index, the dominance and rarity index etc. can be calculated. In such ways,
numbers of sample plots are set up and data on species diversity are also com-
piled. Currently, the measurement of biological diversity in Thailand has been
expanded to cover gene and species levels (Changtragoon and Chaisurisri 1994).
However, at the ecosystem level, measurement is still in the infancy stage.

TREE DENSITY AND SPECIES RICHNESS

Tree density and species richness have been largely studied in various forest
ecosystems. It is clearly shown that the highest tree density and species richness are
found in the tropical rain forest, especially the Malayan mixed dipterocarp sub-
ecosystem, in the southern peninsular of Thailand (Table 5). In this ecosystem, the
highest tree density and the highest number of species were recorded as 1,540 tree/
ha and 109 species/ha respectively (Kiratiprayoon 1986). Among the deciduous
forest ecosystems, the dry dipterocarp forest, in all cases, contains a greater number
of tree species and higher plant densities than the mixed deciduous forest, both
with and without teak.
Under favourable conditions, such as in the Dipterocarpus tuberculatus subtype,
which is one of the sub-ecosystems of the dry dipterocarp forest, tree density in this
forest can be as high as 789 trees/ha with 37 species/ha (Visaratana et al 1986).

119
ECOSYSTEM DIVERSITY
Measurement of forest biodiversity at the ecosystem level has been conducted to
develop the means for sustainable management of tropical forests in the country,
either for optimal production of goods and services, or for conservation of
biodiversity.
This study has been aimed at understanding the dynamics of interrelationships
between seasonal evergreen and dry deciduous forests in order to formulate means
to explain how species richness and patterns of commonness and rarity within each
forest type are related to overall patterns of forest composition, and environmental
factors, especially moisture.
With these questions in mind, a permanent plot of 50 ha. covering a seasonal
primary evergreen forest at Huai Kha Khaeng Wildlife Sanctuary has been set up,
and the methodologies used for ecosystem study are as follows.

Technical plan
Activity 1 : Permanent plot
A 500 x 1000 m permanent plot was established with a 20 m interval permanent
grid. In the plot trees have been tagged for their identity. Their diameters have
been measured. This activity was expected to be completed in 1995.
Enumeration of species composition was planned to include trees and climbing
plants with dbh larger than 1 cm. Permanent demographic records of these plants
will be kept and will be subjected to recensus after 3 years, and later every 5 years.

Activity 2 : Supplemental observatio n of artificial populations of seeds and seedlings

Patterns of regeneration within the studied plot will be observed by establishing
artificial populations of seeds and seedlings along the moisture gradient. In this
case, growth and survival will be recorded.

Activity 3 : Environmental monitoring and phenology observation

During this study, several environmental factors will be monitored.
An automated weather station has been set up to monitor rainfall, air
temperature, relative humidity, wind speed (including wind direction) and solar
radiation. Soil moisture will also be analyzed by appropriate techniques.
Phenology of approximately 80 tree species will be recorded biweekly in the
first year using a focal tree approach, i.e. observing the activity of 5 adult indivi-
duals. When the general patterns of phenology have been discerned, the key

123
species will be observed with larger sample sizes in subsequent years.
When this study is completed, it is anticipated that this information will provide
a guideline for proper management of these types of forest habitats on a sustainable
basis. However, it is still in an initial stage. This will be the starting point for other
sites.
Huai Kha Khaeng, as a World Heritage Site and considered as an important bio-
resource of the country, more studies are being planned such as effects of forest fire
on forest ecosystems, natural regeneration of forest tree species and their distribution,
and wildlife dynamics etc.

REFERENCES
Bawa, SK. 1976. Breeding of Tropical Hardwood: An Evaluation of Underlying
Bases, Current Status and Future Prospects. In Tropical Trees: Variation, Breeding
and Conservation, J. Burley, and B.T. Styles (eds.). Academic Press, London, pp.
43-59.
Boyle, T.J.B., C. Liengsiri, and C. Piewluang. 1991. Genetic Studies in a Tropical
Pine Pinus kesiya II. Genetic Variation among four populations in northern
Thailand, J. Tropic. For. Sci. 3(4): 308-317.
Bunyavejchewin, S. 1983. Analysis of the tropical dry deciduous forest of Thailand,
I. Characteristics of the dominance-type. Nat. His. Bull. Siam Sot. Vol 31 (2):
109-122.
Butcher, PA., J.C. Bell and G.F. Moran. 1992. Patterns of Genetic Diversity and Nature
of the Breeding System in Melaleuca alternifolia. Aust. J. Bot. 40: 367-375.
Changtragoon, S. and K. Chaisurisri. 1994. Private communication.
Changtragoon, S. and R. Finkeldey. 1994. Genetic Variation of Pinus merkusii in
Thailand. II. Diversity and Differentiation at Isozyme Gene Loci and
Characterisation of the Mating System. J. Trop. For. (submitted).
Changtragoon, S. and A.E. Szmidt. 1994. Application of isozyme gene marker for
clone identification of plus trees in Azadirachta indica var. siamensis grafted clone
bank (unpublished paper).
Dhanmanonda, P. and P. Sahunalu. 1992. Research on natural teak forests. In “Teak
Conference : 50 th Anniversary of the Huay Tak Teak Plantation, in Lampang”,
Royal Forest Department, Bangkok, Thailand. 15-55 pp. (In Thai)
Hamrick, J.L., M.J.W. Godt and S.L. Sherman-Broyles. 1992. Factors Influencing
Levels of Genetic Diversity in Woody Plant Species. New Forests 6: 95-124.

124
Joly, H.L., M. Zeh-Nlo, P. Danthu and C. Aygalent. 1992. Population Genetics of
African Acacia, Acacia Albida I. Genetic Diversity of Populations from West
Africa. Aust. J. Bot. 40: 59-73.
Kajornsrichon, S. 1988. Some ecological characteristics of natural pine stands at
Ban Wat Chan Royal Project, Amphoe Mae Cham, Changwat Chiang Mai M.Sc.
Thesis, Kasetsart University 171 p. (In Thai).
Kapoor-Vijay, P. 1992. Biological Diversity and Genetic Resources. The
Commonwealth Science Council, The Chameleon Press Ltd., London.
Kiratiprayoon, S. 1986. Comparative study on the structure of the rattan bearing
tropical rain forest. MSc. Thesis, Kasetsart University.
Kiratiprayoon, S., J. Luangjame, P. Damrongthai and M. Tarumatsawat. 1995. Species
diversity of the second growth at Ngao Demonstration Forest, Lampang
Province. Chapter 16, this volume.
Krebs, G.J. 1972. Ecology. Harper and Row Publ., NY.
Liengsiri, C., E.C. Yeh, and T.J.B. Boyle. 1995. Genetic Structure of A Tropical Forest
Tree, Pterocarpus macrocarpus Kurz, in Thailand, For. Ecol. Manag. 74: 13-22.
Loveless, M.D. and J.L. Hamrick. 1987. Distribution de la variation en especies de
arboles tropicales. Pages 165-175 In: D.A. Clark, R. Dirzo and N. Fetcher (eds.).
Ecologia y ecofisiologia de plantas en los bosques mesoamericanos. Revista de
Biologia Tropical 35: Suppl.
McNeely, J.A., K.R. Miller, W.V. Reid, R.A. Mittermeier and T.B. Werner. 1990.
Conserving the World’s Biological Diversity. IUCN, WRI, CI, WWF-US, World
Bank.
NBU. 1992. Thailand Country Study on Biodiversity. National Biodiversity Unit,
Ministry of Science, Technology and Environment, Bangkok, Thailand.
Nanakom, V. 1993. Forest biodiversity in Thailand. In “Man and Nature: Biodiversity
Crisis and Solution to Sustainability. “Office of the Prime Minister, Thailand (in
Thai) 55-67 pp.
Nilroung, S. 1986. Structural characteristics, rate of gap formation and turnover
rate in dry dipterocarp forest at Sakaerat. M.Sc. Thesis, Kasetsart University (in
Thai).
RFD. 1992. Forestry Statistics of Thailand. Royal Forest Department Bangkok,
Thailand (Thai and English).
Sahunalu, P., M. Chamreanpruk, B. Puriyakom, P. Dhanmanonda, W. Suwannapin
and B. Prachaiyo. 1979. Structure of three forest types in the Prom Basin,

125
 Chaiyaphum province. For. Res. Bull. No. 63 Fac. Forestry Kasetsart University
(in Thai).
Sandlund, O.T., K. Hindar and A.D.H. Brown. 1992. Conservation of Biodiversity
for Sustainable Development, Scandinavian Press, Oslo.
Smitinand, T. 1994. Private communication.
Soonhuae, P., C. Piewluang and T.J.B. Boyle. 1994. Population Genetics of Dalbergia
cochinchinensis Pierre and Implications for Genetic Conservation. ASEAN Forest
Tree Seed Centre Technical Publication No.18.
TFSMP. 1993. Thai Forestry Sector Master Plan Vol 5. Royal Forest Department,
Bangkok, Thailand (in Thai and in English)
Thammincha, S. 1933. Establishment of permanent plots for monitoring tropical
forests. Pages 245-248 In: Proceedings of the Stand Inventory Technologies : An
International Multiple Resource Conference. American Soc. for Phot. and Remote
Sensing.
Vannaprasert, M. 1985. Structural characteristics and gap size distribution of the
hill evergreen forest at Doi Pui, Chiang Mai. M. SC. Thesis, Kasetsart University
(in Thai).
Visaratana, T. 1983. Structural characteristics and canopy gap regeneration of the
dry evergreen forest at Sakaerat Environmental Research Station. M. SC. Thesis,
Kasetsart University (in Thai).
Visaratana, T, K. Pitprecha, S. Kiratiprayoon, T. Kampan and K. Higuchi. 1986.
Structural characteristics and species composition of dry dipterocarp forest
(Dipterocarpus tuberculatus Roxb. Community type) at Salak Phra Wildlife
Sanctuary. Technical Paper No. 10 Forest Ecology Section, RFD (in Thai).

126
 Chapter

9
MONITORING OF FOREST BIODIVERSITY :
POLICY AND RESEARCH ISSUES

M. N. Sallehl and N. Manokaran2

INTRODUCTION
Biodiversity refers to the entire range of variation among plants, animals and
microorganisms, across all levels of the biological hierarchy from genes to ecosystems
(Solbrig 1991). It is usually considered in terms of numbers of species, particularly
of vertebrates and higher plants but is more than that. At the species level, tropical
forests are exceptionally rich, containing half of all vertebrates and vascular plant
species so far documented and, allowing for the much larger numbers of inverte-
brate species, many yet to be discovered, possibly 90% of the world’s total species
(McNeely et al. 1991).
Environmental degradation, and with it habitat loss, has led to loss of biodiver-
sity worldwide. There is no accurate account of the loss of species and this loss
may occur even before species are discovered. An estimate indicates that since the
year 1600,724 known species have become extinct (McNeely et al. 1991). Another

1 Director-General, Forest Research Institute Malaysia (FRIM), Kepong, 52109 Kuala Lumpur.
2 Senior Research officer, Forest Research Institute Malaysia (FRIM), Kepong, 52109 Kuala Lumpur.

127
estimate places the potential loss of biodiversity at 15,000 to 50,000 species per year
from the 1990s onwards, due mainly to tropical deforestation (Reid and Miller 1988).

GLOBAL INITIATIVES ON BIODIVERSITY

The first steps to environmental degradation can be traced to the Industrial
Revolution. Pollution of the atmosphere and the waterways began. Forests were
systematically removed. Soil degradation began. The rapid economic growth
following the Second World War rapidly accelerated environmental degradation.
By the early second half of the twentieth century, concern for the global environ-
ment translated into the United Nations Conference on the Human Environment
held in June 1972 in Stockholm, Sweden. The first ever meeting of the international
community on environment addressed the relationship between environment and
development at the global level. The Stockholm Declaration of the Conference
addressed biodiversity concerns in three of the 26 principles enunciated. The need
to safeguard flora and fauna and especially representative samples of natural
ecosystems for the benefit of present and future generations through careful plan-
ning and management is stressed in one principle. The need to safeguard and
wisely manage the heritage of wildlife and its habitat is stressed in another princi-
ple. The need to prevent pollution of the seas that is harmful to marine life is stressed
in a third principle.
A decade after Stockholm, environmental degradation had continued unabated.
Global warming, ozone layer depletion, marine pollution, acid rain, deforestation,
loss of biodiversity, and soil degradation had become serious environmental issues
by the early 1980s.
The United Nations General Assembly established the World Commission on
Environment and Development in 1983 with the mandate to re-examine the criti-
cal environment and development issues and to formulate realistic proposals for
dealing with them. The 1987 Report of the Commission, entitled “Our Common
Future”, focused its attention on several issues, one being the loss of species and
genetic resources.The Commission deemed it necessary to place this issue on
political agendas as a major economic and resource issue. The Commission felt
that reservoirs of biological diversity needed to be developed economically and
envisaged the necessity of protecting large areas for future needs, with assistance
for conservation coming from international agencies.
The 1987 Report led to preparations for holding the United Nations Conference
on Environment and Development (UNCED) in Rio de Janeiro in June 1992. A

128
series of preparatory meetings, or PREPCOM meetings, were held at govern-
mental level, beginning in 1990, and culminating in UNCED. A voluminous
500 page, 40 chapter document, Agenda 21, an action programme for environment
and development issues, was a major output at Rio. A chapter, Conservation of
Biological Diversity, addresses various issues, including that of monitoring. The
Conference in Rio was exactly 20 years after Stockholm.
Around the time of the Stockholm Conference, and shortly thereafter, a number
of international instruments had been developed for the protection of biological
diversity. The most important are:
Convention on Wetlands of International Importance Especially as Waterfowl
Habitat (RAMSAR) 1971
Convention Concerning the Protection of World Cultural and Natural Heritage
(PARIS) 1972
Convention on International Trade in Endangered Species of Wild Fauna and
Flora (CITES) (WASHINGTON) 1973
Convention on the Conservation of Migratory Species of Wild Animals (CMS)
(BONN) 1979
These international instruments are all, however, inherently limited in
application, however liberally their texts are interpreted. RAMSAR is confined to
only wetland habitats. CITES is essentially about trade in species. CMS deals only
with migratory organisms, and the Convention on World Heritage natural sites is
limited in scope. Collectively, the coverage of all these conventions is only sectoral
in nature and none has been designed to protect biological diversity as a primary
objective. A common thread in all these conventions is the element of monitoring.
In 1987, the United Nations Environment Programme (UNEP), established as a
consequence of the Stockholm Conference, formally recognised the need for
concerted international action to protect biodiversity. It initiated a series of
intergovernmental meetings that developed a negotiated document, the Convention
on Biological Diversity, that was brought to UNCED for signature by heads of
governments. The ratification of the treaty by governments is in progress.
The chapter on Conservation of Biological Diversity in Agenda 21 places
emphasis on monitoring of biodiversity. However Agenda 21 has no legal standing.
The Convention on Biological Diversity is a legal instrument. In one article titled
“Identification and Monitoring”, Parties to the treaty are required to monitor
components of biological diversity, especially those requiring urgent conservation
measures and those which offer the greatest potential for sustainable use. Also,

129
effects of activities that are likely to have significant adverse impacts on the
conservation and sustainable use of biological diversity are required to be moni-
tored through sampling and other techniques.
At the governmental level, efforts are underway to put the Convention into
operation. The first Conference of Parties was held in December 1994 to initiate the
process of giving the Convention operational form.

THE MONITORING PROCESS

In the context of forest biodiversity, or for that matter, any form of biodiversity,
what does monitoring involve? It involves the gathering of data to enable the
detection of changes in the status, security and utilisation of biological diversity for
the purpose of improving the effectiveness of management of that diversity (UNEP
1993).
As a condition, therefore, monitoring requires the building up of an information
baseline which is practically an incremental process. Such an information baseline
on forest biodiversity would allow for more enlightened resource planning. Such
an information baseline of forest biodiversity would usually include habitats and
species, but rarely genetic diversity.
Continuous enlargement of the information baseline, filling key gaps in the
information coverage, is required for more effective planning of biodiversity
management. Continuous updating of the information baseline through repeated
collection of data will be part of the monitoring process.

FOREST MANAGEMENT
Before one considers the issue of monitoring of forest biodiversity, one really needs
to consider the issue of what is forest management in practice at the moment. Salleh
(1991, 1992) has discussed this issue at length. Are forests being managed for
biodiversity at all? How are tropical forests, where much of the world’s terrestrial
biodiversity resides, being managed and for what? The question of policy and
research issues in the context of monitoring of forest biodiversity would need to be
addressed with reference to answers to the above questions.
Forests fulfil many functions. They are the source of timber and nontimber
goods, and provide numerous services. Nontimber goods include rattan, bamboo,
fuelwood and extractives (dyes, gums, incense, latexes, oils, resins, and including
those leading to development of pharmaceuticals), and food such as fish, game,

130
fruits, nuts, honey and spices. Services provided by forests include soil and
watershed protection, soil generation, energy supply, conservation of biological
diversity, regulation of climate, recycling of nutrients, carbon sequestration, oxy-
gen release and tourism and recreation. Interactions between a certain number of
species and genetic diversity uphold the cyclical relations within the forest
ecosystem, and thereby maintain ecological services.
Forests in almost all countries, whether temperate or tropical, have been utilised
primarily for the production of timber as it has been considered the only option for
economic growth. The sustainability of forestry practices is now being questioned,
especially pertaining to the tropics. The study of the International Tropical Timber
Organisation (ITTO) and the International Institute for Environment and
Development (IIED) (Poore et al. 1989), for example, reported that less than one
tenth of a percent of tropical forests are managed on a sustainable basis.
Malaysia, where forest management has been practised for nearly a century
has, like most other nations, focused solely on the production of timber in forests
outside of those totally protected, even though the forests produce a whole spectrum
of other products. Management and harvesting plans are prepared only for logging
of timber and, at most, other nontimber products such as rattan are only briefly
mentioned. Management to ensure sustained supply of the multitude of other
economically important products such as rattan, bamboo, fruits, wildlife, medici-
nal plants, resins and other products from any particular locality is difficult, if not
impossible.
In the management for timber supply, the basic approach is to assess the stand-
ing stock of the forest, determine the growth rate of the forest from research plots,
determine the economic yield that should be produced and thus determine the
rotation or cycle to be used. The forest is then divided into blocks equivalent to the
number of years in the cycle. This is a simple straightforward exercise. Timber har-
vesting is controlled by blocks, and after harvesting, silvicultural operations are
undertaken to assist the growth of seedlings or saplings, depending upon the
silvicultural system being used. Seldom, if ever, are considerations given to other
resources, except for minimising damage to rivers and streams, and to soil erosion
when preventive or conservation measures are put in place. Timber harvesting is
also not permissible in very steep areas or environmentally sensitive areas. No
attention is given to managing biodiversity, nontimber products or other
environmental aspects in the timber production forest.
Clearly there has to be a policy decision taken at the highest level to shift em-
phasis of management from only timber production to management of the forest

131
for other goods and services as well. Could timber be produced in conjunction
with soil and watershed protection? What technological changes in timber
harvesting are required to ensure this? Would the conservation of biological diver-
sity be compatible with sustainable production of timber? Could tropical forests
be managed for multiple uses all at the same time ? If not, should zoning of forests
be considered for optimum utilisation, as a further refinement of the system in
place in Malaysia where the Permanent Forest Estate is classified into Protective,
Productive and Amenity forests?
A major problem in arriving at a reasonable valuation of multiple functions of
tropical forests is that of lack of data, including monitoring systems. That is why
only timber resources have been considered and nontimber goods and forest services
have been ignored in national accounting systems. As tropical forests are almost
all in Third World countries, international funding and collaboration is required to
develop the data collecting and monitoring systems. Without such data there is
every likelihood that tropical forests will continue to be viewed solely from the
perspective of timber in the national accounting systems. Management for timber
alone for shortterm gains would lead to disruption of the ecological processes and
the complex species interactions that guarantee the sustainability of the natural
forest ecosystem.

FOREST BIODIVERSITY BASELINE DATA

This then brings us back to the question of data gathering in relation to monitoring
of forest biodiversity. What kind of data on biodiversity is available and what kind
of data requires to be collected so that a monitoring process could be implemented,
to detect changes in the status, security and utilisation of biological diversity. A
discussion of this in relation to the Malaysian context may help to highlight relevant
issues.
The beginning of documentation of forest biodiversity could be dated to the
last couple of centuries, during the colonial period, when botanical collections were
carried out leading to publications such as The Flora of British India by Hooker (1894)
and Palmae by Beccari and Hooker (1894). During this period there were similar

132
efforts at documentation in the region, in countries now collectively known as
ASEAN (Association of Southeast Asian Nations).
In the then colonial territory of Malaya, work on flora led to publications such
as Flora of the Malay Peninsula by Ridley (192225), Commercial Timber Trees by
Foxworthy (1927), Wayside Trees of Malaya by Corner (1940), Foresters’ Manual of
Dipterocarps by Symington (1943), Pocket Check List of Timber Trees by Kochummen
(1979). Work on fauna, in many cases of a general nature, led to publications such
as Introduction to Malayan Birds by Madoc (1976), The Birds of Borneo by Smythies
(1981), The Butterflies of the Malay Peninsula by Corbet and Pendlebury (1978),
Moths of Borneo with Special Reference to Mt. Kinabalu by Holloway (1976), Poisonous
Snakes of the Malay Peninsula by Lim (1991), Introduction to Mammals of Singapore
and Malaya by Harrison (1966), and The Birds of the Malay Peninsula by Medway and
Wells (1976).
The tree flora of Peninsular Malaysia has been documented in four volumes,
edited by Whitmore (1972, 1973) and Ng (1978-1989). This voluminous work, carried
out by 15 authors over a period of about 20 years, covers all the families of flower-
ing plant species forming trees reaching 90 cm girth (28.6 cm diameter), and covers
gymnosperms as well. Herbarium collections at the Forest Research Institute
Malaysia and at Singapore were the main source of the studies, with reference to
collections in Europe on critical matters. A total of 99 families comprising over
3000 species were covered. Work has started, along similar lines, on the tree flora
of the Malaysian states of Sabah and Sarawak in the island of Borneo. Based on
both herbarium and extensive field work, the rattan flora of Malaysia has been
documented by Dransfield (1979, 1984, 1992). Taxonomic accounts of the bamboos
(Wong 1992) and termites (Tho 1992) of Peninsular Malaysia were completed
recently.
Much is now known about the habitats, food, behaviour and distribution of
large mammals such as the tiger, elephant, rhino, deer and ‘seladang’. This is the
result mainly of continuing studies associated with the wildlife conservation
programme of the Department of Wildlife and National Parks.
Data gathering has included documentation of the forest types of Peninsular
Malaysia [( see Symington (1943) and Wyatt Smith (1964)]. This broad classification,

133
shown in Table 1, applies to Sabah and Sarawak as well.

Climactic climax forest Edaphic forest

Lowland dipterocarp forest Heath (or Kerangas/Kerapah) forest
Hill dipterocarp forest Forest over limestone
Upper dipterocarp forest Forest over ultramfic outcrops
Montane oak forest Beach stand vegetation
Lower ericaceous forest Mangrove forest
Montane subalpine vegetation Brackish-water forest
Semi-evergreen seasonal forest Peat swamp forest
Fresh-water swamp forest
Seasonal swamp forest

Table 1: The Rain Forest Types of Malaysia [Adapted from Symington (1943) and Wyatt-
Smith (1964)].

MONITORING OF FOREST BIODIVERSITY IN MALAYSIA

UNEP (1993) describes monitoring of biodiversity in four categories at the natio-
nal level, these being monitoring of genetic diversity, species diversity, habitats and
protected areas. There is virtually no baseline data in Malaysia on genetic diversity
of flora and fauna of the forests, and this is generally a global phenemenon. There
has been some monitoring of species and of habitats or ecosystems.
Remote sensing monitoring through aerial photography has provided
information on change in the extent of forested land. Over the period 1970 to 1989,
forested land for the whole of Malaysia had been reduced by 22.7%, mainly due to
forest conversion to agriculture (Manokaran 1992). Almost all the forests cleared
were lowland forest, with its multitude of habitats. As a result of this forest clea-
rance, and therefore of loss of habitats, reduction in populations of wildlife such as
the tiger, elephant, rhinoceros and ‘seladang’ (wild cattle) occurred (Anonymous
1998a, 1998b).
Monitoring of forest resources, essentially timber resources, has been, and is, an
ongoing process. The Forest Resources Reconnaissance Survey (FRRS), initiated in
1962 and completed in 1969 and designed to assess the extent, distribution and

134
nature of the forest resource on all forested lands throughout Peninsular Malaysia,
provided initial baseline data and maps for planning purposes. Up until then, a
limited amount of qualitative and quantitative data did exist for some of the forest
reserves. The FRRS was based on aerial photointerpretation and sufficient
quantitative survey on the ground.
The FRRS led to the first national forest inventory of Peninsular Malaysia (1970-
1972), a project that consisted primarily of reviewing and complementing where
required, the results of the FRRS, with the purpose of preparing a forest development
plan (Anonymous 1973). The inventory involved field sampling of the main
commercial species and species groups. Mangrove forests were excluded from the
survey. The national forest inventory was repeated in 1981-82 when the most
important commercial species of rattan were included in the survey. A third
inventory was carried out in 1991-92. A similar inventory of the mixed dipterocarp
forests of the state of Sarawak was carried out between 1969 and 1972, and the
main commercial species and species groups were enumerated (Anonymous 1974).
To date the inventory has not been repeated.
Tree species diversity, and populations of these tree species have been moni-
tored in detail for long periods of time in 2ha plots in primary lowland and hill
dipterocarp forests in Peninsular Malaysia. Results of the monitoring of trees of 10
cm diameter and larger for periods of 13,36 and 38 years have been documented
(see Manokaran and Kochummen 1987, Manokaran 1988), and monitoring still
continues. A 50ha plot established in primary forest in Pasoh Forest Reserve, a
lowland dipterocarp forest, beginning in 1985, has provided baseline data on all
woody species of 1 cm diameter and larger (Manokaran et al. 1990, Kochummen et
al. 1990, Kochummen et al. 1992, Manokaran et al. 1991, Manokaran et al. 1992a and
Saw et al. 1991). Arecensus in 1990 (Manokaran et al. 1992b) and repeated monitoring
at five-year intervals thereafter is expected to provide information on tree species
diversity and population change over time. In the same forest, Wells (unpublished)
has monitored populations of birds for about two decades, and Ratnam and Lim
(unpublished) have monitored small mammal populations for several years now.
Of the total land area of Malaysia of 32.86 million hectares, 19.2 million hectares
or 58.4 percent are covered with natural forest (Table 2a). National parks and wild-
life sanctuaries cover an area of 2.14 million hectares. Except perhaps for certain
groups of wildlife, no proper inventory has been carried out of species or habitats
within these protected areas.
The National Forestry Policy 1978 and the National Forestry Act 1984 catered
for the establishment of the Permanent Forest Estate (PFE). The original area of the

135
PFE of 12.73 million hectares is being increased to 14.05 by the gazetting of state-
land forest (Table 2b). While the Productive Forest of the PFE is for timber Production
in perpetuity, the Protective Forest is for the protection of watersheds and the
environment. There has been no inventory of species or habitats within these
protected areas.

Region Permanent National Stateland Total Total

Forest Park & Wildlife Forest Natural land
Estate (PFE) Sanctuary Forest area

Peninsular 4.70+ 0.74* 0.78 5.97 13.16

Malaysia
Sabah 3.35 0.40* 0.93 4.54 7.37
Sarawak 6.00+ 1.00 1.70 8.69 12.33

Total 14.05 2.14 3.41 19.20 32.86

Table 2a: Natural Forest Land Use Pattern in Malaysia, 1992 (million hectares).
* 0.19 and 0.14 million hectares respectively in Peninsular Malaysia and Sabah are located within
the PFE
+ 0.06 and 0.01 million hectares respectively in Peninsular Malaysia and Sabah are plantation forests
within the PFE

Source: Ministry of Primary Industries, Malaysia

Region Protection Production Total forested

Forest Forest area under PFE
Peninsular 1.90 2.80 4.70+
Malaysia
Sabah 0.25 3.10 3.35
Sarawak 1.00 5.00 6.00*
Total 3.15 10.90 14.05
Table 2b: Permanent Forest Estate in Malaysia, 1992 (million hectares).
+ Includes 0.24 million hectares to be gazetted
* Includes 1.50 million hectares to be gazetted
Source: Ministry of Primary Industries, Malaysia

136
 MONITORING POLICY AND RESEARCH ISSUES
In the absence of any clear policy on biological diversity, baseline information
gathering would probably proceed in an uncoordinated manner. Many aspects of
biological diversity may be neglected altogether. In the forestry sector, for example,
only timber species may be considered to the exclusion of all other aspects of
biological diversity. This is generally the case at the moment. A policy on biological
diversity would help to focus on these other aspects as well as to strategise the way
in which biodiversity considerations are included in planning for development.
The policy would need to cover not only the question of gathering of baseline data
but also that of a monitoring programme and of an information management sys-
tem. In parallel with the development, signing and ratification of the Convention
on Biological Diversity several nations have been developing such a policy. In
Malaysia a National Policy on Biological Diversity has been developed.
Policy documents themselves would be ineffective unless they are backed by
an adequate legislative framework and the necessary political commitment
towards enforcement. This legislative framework should fully integrate other
aspects of land use, in particular agriculture, with forestry practices since these
other land use practices generally impinge on the wellbeing of forest biodiversity.
Expansion of areas for agriculture (especially for cash crops) and aquaculture are
almost always at the expense of forested lands, and this issue has to be properly
addressed in the legislative framework.
Monitoring is the repeated standardised collection of data on certain parameters
that could indicate the status and use of resources. Time-series data are nonexistent
except for a few “megafauna” species, such as mammals and birds, tropical forests
and the landuse estimates produced by FAO (UNEP 1993). In Peninsular Malaysia,
with baseline information on tree flora available, monitoring of timber resources
has been made possible. Aerial surveys have also made it possible to monitor the
extent of the major forest types over time.
A focus of research in this region should be the identification of fragile or sensitive
ecosystems or habitats such as steepland areas and wetlands, and subsequent
monitoring of these sites. Often these are endangered or threatened by human
activity. In the same vein, endangered or threatened species need to be identified
and their populations and habitats monitored. In all these a multidisciplinary
approach is required.
Baseline data of forest biodiversity, especially in tropical countries, is woefully
inadequate, thus hindering any attempt at monitoring. This is due to a large extent
on the lack of trained personnel in the field of taxonomy. For example, except for

137
reasonable expertise in the tree flora, local expertise on shrub, herbaceous and
epiphytic flora is totally inadequate. The problem is compounded by the fact that
institutional infrastructure is extremely weak or is lacking altogether. Manpower
and infrastructure development is therefore an area where international funding
and collaboration is required to develop data collecting and monitoring systems.
Conservation International (1992) discusses aspects of these issues in its Rapid
Assessment Program, a biological inventory programme designed to meet the
information needs necessary to catalyse conservation action and improve
biodiversity protection. UNEP (1993) devotes a whole section to these issues, and
besides addressing the questions of institutional capacities and human resources,
also discusses points such as national legislation, technological facilities, informa-
tion resources, and data management and monitoring capacity.
In tropical countries, species richness, compounded by poorly developed
manpower and infrastructure facilities, is likely to ensure that comprehensive
inventories of all species in the forest ecosystem will remain a Herculean task. Not
only would such inventories, if at all possible, be highly time consuming, they would
also be too costly For example, Janzen (1993) estimates that an intensive all taxa
biological inventory analysis would cost US $25 million over 5 years for each and
every site in the tropics. Monitoring of forest biodiversity would therefore have to
focus on certain key parameters.
UNEP (1993), in noting that rapid advances in remote sensing, environmental
measurement techniques and information technology have greatly facilitated the
monitoring of biodiversity, provides a provisional list of key parameters under four
categories for monitoring biodiversity at the country level (monitoring genetic
diversity, species monitoring, habitat monitoring, protected areas monitoring).
Taken together with the M ‘ inimum set of indicators for monitoring biodiversity at
the country level’ of Reid et al. (1993) and the ‘Parameters that an early warning
network must monitor at the country level’ of the Global Biodiversity Strategy of
WRI/IUCN/UNEP (1992), UNEP (1993) considers that a framework is in place for
determining priorities and goals for biodiversity planning, that generates the data
necessary to monitor how well the country is doing in achieving its strategic plan-
ning objectives, and that supplies the early warning information necessary for the
rapid response to new threats.
Genetic diversity of forest species in the tropics remains unexplored. Early
studies are just beginning and it is anticipated that baseline information gathered,
and any monitoring thereafter, will have implications on forest management
practices.

138
 Developing countries are gene-rich but the economic potential of forest
biodiversity in these countries remains largely untapped. Developed countries are
technology-rich, and efforts at baseline information gathering and monitoring
should include collaboration between the North and the South in developing more
economic use of components of this biodiversity. That biodiversity knows no
boundary would also mean that regional cooperation is required in biodiversity
prospecting.
A research issue of great importance to countries like Malaysia is whether
management of forests for sustained production of timber will conflict with
conservation of biological diversity, both at the gene and species levels. The few
studies of timber harvesting effects on biodiversity so far carried out in Peninsular
Malaysia have been limited to selected locations and concentrated particularly
on the trees or other woody plants in the floral composition, and on some verte-
brate groups, particularly birds, among the fauna. Vincent and Binkley (1992) argue
that dominant-use management, managing some forests for timber as a dominant
use, and others for nontimber values, is more efficient than multiple-use manage-
ment, based on the economic theory of “comparative advantage”. As Production
Forests form about 57 percent of the total natural forest area in Malaysia, it would
be in the best interests of forest management to monitor the effects of harvesting at
a range of intensities on the inventory of species and their spectrum of distribution
and activity. Ultimately guidelines could be developed on forest exploitation and
management to conserve biodiversity, by the most costeffective means, in concert
with sustainable production of timber and other benefits. Such guidelines should
make it mandatory for Environment Impact Assessments (EIA) to be carried out
before timber harvesting, and this to be complemented by environmental auditing
during and after the harvesting activities. EIA and environmental auditing should
be legislated to be part of timber harvesting operations.
There is also the issue of global warming and acid rain and their effects on
forest biodiversity. Although uncertainties remain, there is now much broader
scientific consensus that the problem of global warming is real, while increased
levels of acid rain have been detected in tropical countries like Malaysia which are
on the path towards industrialisation. These phenomena are expected to affect
forest biodiversity in unpredictable ways over a long time frame, and some kind of
monitoring process would eventually have to be set in motion.

CONCLUSIONS
Most of the world’s forest biodiversity is located in tropical forests which are almost

139
entirely found in developing countries. Baseline information on biodiversity in
these countries is woefully inadequate, other than perhaps on the extent of forested
areas, timber resources and “megafauna” such as mammals and birds. Weaknesses
in the field of human resources (especially of taxonomists), institutional capacities,
technological facilities, information resources and data management capacities are
the reasons for this state of affairs. This is compounded by the fact that in forest
management, timber is the only option considered to the exclusion of all other aspects
of forest biodiversity. In such a scenario, therefore, the monitoring of forest
biodiversity is a most difficult task.
Policy formulation or changes are required that would shift focus to other aspects
of forest biodiversity from solely timber. This would have to be supported by
international funding and collaboration that would alleviate weaknesses in the
baseline information gathering capacities of the countries concerned. The
international support should also help to develop a monitoring system, initially
making use of information available, and eventually based on the wider baseline
information gathered.
The documentation of each and every species for the purpose of monitoring
may be an unrealistic goal although such documentation could be an ultimate target.
Nevertheless parameters or indicators that have been provisionally identified could
form the database of the monitoring system.
Monitoring of forest biodiversity should not just be in protected areas but also
in areas designated as permanent timber production areas. Research should
therefore focus also on the effects of timber harvesting on other species diversity,
and a monitoring regime should lead to developing guidelines on timber harvesting
that would include conservation of biodiversity in these permanent timber
production areas.
In the final analysis, only policies on biodiversity backed by legislation and
political commitment, and supported by international funding and collaboration,
would lead to meaningful progress in baseline data gathering and monitoring. The
legislative framework should integrate other land use practices such as agricul-
ture with forestry, as these are very closely inter-twined. Collaboration between
the gene-rich countries of the South and the technology-rich countries of the North
should also include efforts at greater utilisation of components of biodiversity. In
all these a reasonably efficient monitoring system is essential for detecting changes
in the status, security and utilisation of forest biodiversity that would allow steps
to be taken to manage biodiversity effectively.

140
REFERENCES
Anonymous. 1973. Forestry and Forest Industries Development: Malaysia. A
National Forest Inventory of West Malaysia 1970-1972. FO:DP/MAL/72/009
Technical Report 5. United Nations Development Programme, Food and
Agriculture Organisation of the United Nations, Rome.
Anonymous. 1974. Forestry and Forest Industries Development: Malaysia. An
Inventory of Mixed Dipterocarp Forests of Sarawak 1969-1972. FO: DP/MAL/
72/009 Technical Report 2. United Nations Development Programme, Food
and Agriculture Organisation of the United Nations, Rome.
Anonymous. 1988a. Peninsular Malaysia: conservation of biological diversity and
forest ecosystems. A briefing document prepared by the IUCN Tropical Forest
Programme. International Union of Conservation of Nature and Natural
Resources, United Kingdom.
Anonymous. 1988b. Sabah and Sarawak: conservation of biological diversity and
forest ecosystems. A briefing document prepared by the IUCN Tropical Forest
Programme. International Union of Conservation of Nature and Natural
Resources, United Kingdom.
Beccari, O. and J.D. Hooker. 1894. Palmae. In: J.D. Hooker and L. Reeve (eds.).
Flora of British India, London.
Conservation International. 1992. Rapid assessment program. Global conservation
priority setting. Unpublished.
Corbet, A.S. and H.M. Pendlebury. 1978. The Butterflies of the Malay Peninsula.
Third edition revised by J.N. Eliot. Malayan Nature Society, Kuala Lumpur.
Comer, E.J.H. 1940. Wayside Trees of Malaya. Government Printer, Singapore.
Dransfield, J. 1979. A Manual of the Rattans of the Malay Peninsula. Malayan
Forest Records No. 29, Forestry Department, Kuala Lumpur.
Dransfield, J. 1984. The Rattans of Sabah. Sabah Forest Records No. 13. Forest
Department, Sabah.
Dransfield, J. 1992. The Rattans of Sarawak. Sarawak Forest Department and the
Royal Botanic Gardens Kew.
Foxworthy F.W. 1927. Commercial Timber Trees of the Malay Peninsula. Malayan
Forest Records No. 3, Forestry Department, Kuala Lumpur.
Harrison, J.L. 1966. Introduction to Mammals of Singapore and Malaya. Malayan
Nature Society, Singapore Branch.
Holloway, J.D. 1976. Moths of Borneo with special reference to Mt. Kinabalu.

141
 Malayan Nature Society, Kuala Lumpur.
Hooker, J.D. 1894. The Flora of British India. L. Reeve, London.
Janzen, D.H. 1993. Taxonomy: universal and essential infrastructure for
development and management of tropical wildland biodiversity. Proceedings
of the Norway/UNEP Expert Conference on Biodiversity. Trondheim, Norway.
Kochummen, K.M. 1979. Pocket Check List of Timber Trees. Malay. For. Rec. No.
17, Second revision.
Kochummen, K.M., J.V. Lafrankie and N. Manokaran. 1990. Floristic composition
of Pasoh Forest Reserve, a lowland rain forest in Peninsular Malaysia. Journal
of Tropical Forest Science 3(l): 1-13.
Kochummen, K.M., J.V. Lafrankie and N. Manokaran. 1992. Diversity of trees and
shrubs in Malaya at regional and local levels. Malayan Nature Journal 45(14):
545-554.
Lim, B.L. 1991. Poisonous Snakes of the Malay Peninsula. 3rd. ed. Malayan Nature
Society Special Publication, Kuala Lumpur.
Madoc, G.C. 1976. Introduction to Malayan Birds. 2nd impression. The Malayan
Nature Society, Kuala Lumpur.
Manokaran, N. 1988. Population Dynamics of Tropical Forest Trees. Ph. D. thesis,
University of Aberdeen.
Manokaran, N. and K.M. Kochummen. 1987. Recruitment, growth and mortality
of tree species in a lowland Dipterocarp forest in Peninsular Malaysia. J. Trop.
Ecol. 3: 315-330.
Manokaran, N., J.V. Lafrankie, K.M. Kochummen, E.S. Quah, J.E. Klahn, P.S. Ashton
and S.P. Hubbell. 1990. Methodology for the fifty hectare research plot at Pasoh
Forest Reserve. Research Pamphlet No. 104, Forest Research Institute Malaysia,
Kepong.
Manokaran, N., J.V. Lafrankie and Roslan Ismail. 1991. Structure and composition
of the Dipterocarpaceae in a lowland rain forest in Peninsular Malaysia. Pages
317-331 In: I. Soerianegara, S.S. Tjitrosomo, R.C. Umaly and I. Umboh (eds.),
Fourth RoundTable Conference on dipterocarpaceae. Biotrop Special Publication
No. 41, SEAMEO BIOTROP, Bogor.
Manokaran, N. 1992. An overview of biodiversity in Malaysia. Journal of Tropical
Forest Science 5(2): 271-290.
Manokaran, N., J.V. Lafrankie, K.M. Kochummen, E.S. Quah, J.E. Klahn, P. S. Ashton
and S. P. Hubbell. 1992. Stand table and distribution of species in the fifty

142
 hectare research plot at Pasoh Forest Reserve. FRIM Research Data No. 1.
Manokaran, N., Abd. Rahman Kassim, Azman Hassan, E.S. Quah and P.F. Chong.
1992. Shortterm population dynamics of dipterocarp trees in a lowland rain
forest in Peninsular Malaysia. Journal of Tropical Forest Science 5(l): 97-112.
McNeely, J.A., K.R. Miller, W.V. Reid, R.A. Mittermeier and T. B. Werner. 1991.
Conserving the World’s Biological Diversity. World Resources Institute, World
Conservation Union, World Bank, World Wildlife Fund, Conservation
International: Washington, D. C. and Gland, Switzerland.
Medway, Lord and D.R. Wells. 1976. Birds of the Malay Peninsula. Penerbit
Universiti Malaya.
Ng, F. S. P. (ed.). 1978. Tree Flora of Malaya. Vol. 3. Longman, Kuala Lumpur.
Ng, F. S.P. (ed.). 1989. Tree Flora of Malaya. Vol. 4. Longman, Kuala Lumpur.
Poore, D., P. Burgess, J. Palmer, S. Rietbergen and T. Synnott. 1989. No Timber
Without Trees. Earthscan Publications Ltd. London.
Reid, W.V. and K.R. Miller. 1988. Keeping Options Alive: the Scientific Basis for
Conserving Biodiversity. World Resources Institute, Washington, D. C.
Reid, W.V., J.A. McNeely, D.B. Tunstall, D.A. Bryant and M. Winograd. 1993.
Biodiversity Indicators for Policymakers. World Resources Institute, Washington.
Ridley, H.N. 192-225. Flora of the Malay Peninsula. L. Reeve, London.
Salleh Mohd. Nor. 1991. Towards sustainable forest management in 2020. Paper
presented at the Inaugural Lectures on Environment to commemorate the
Langkawi Declaration on Environment, Kuala Lumpur, 27 October 1991.
Salleh Mohd. Nor 1992. Sustainable forest management. Paper presented at the
Conference on Environment and Development in the Islamic World, Kuala
Lumpur, l0-14 August 1992.
Saw L.G., J.V. Lafrankie, K.M. Kochummen and S.K. Yap. 1991. Trees bearing fruits
in a Malaysian rain forest. Economic Botany 45(l): 120-136.
Smythies, B.E. 1981. The Birds of Borneo. 3rd edition. Sabah Society and the
Malayan Nature Society, Kuala Lumpur.
Solbrig, O. (ed.). 1991. From Genes to Ecosystems: a Research Agenda for
Biodiversity. IUBS, Paris.
Symington, C.F. 1943. Foresters’ Manual of Dipterocarps. Malayan Forest Records
No. 16. (New edn. 1974, Penerbit Universiti Malaya, Kuala Lumpur).
Tho, Y.P. 1992. Termites of Peninsular Malaysia. Malayan Forest Records No. 36,
Forest Research Institute Malaysia, Kepong.

143
United Nations Environment Programme. 1993. Guidelines for country studies on
biological diversity. UNEP/Bio. Div./ Guidelines/CS/Rev.2.
Vincent, J.R. and C.S. Binkley. 1992. Multipleuse forestry: a fallacy of composition.
Development Discussion Paper No. 421. Harvard Institute for International
Development, Cambridge, Massachusetts.
Whitmore, T.C. (ed.). 1972. Tree Flora of Malaya. Vol. 1. Longman, Kuala Lumpur.
Whitmore, T.C. (ed.). 1973. Tree Flora of Malaya. Vol. 2. Longman, Kuala Lumpur.
Wong, K.M. 1992. Systematics of Bamboo in Peninsular Malaysia. Ph. D. thesis,
University of Malaya.
World Commission on Environment and Development. 1987. Our Common Future.
Oxford University Press, Oxford.
WRI/IUCN/UNEP 1992. Global Biodiversity Strategy. Policymakers’ Guide.
WRI/IUCN/UNEP.
Wyatt-Smith, J. 1964. A preliminary vegetation map of Malaya with descriptions of
the vegetation types. J. Trop. Geogr. 18: 200-213.

144
 Chapter

10
MEASUREMENT OF GENETIC DIVERSITY
WITH SPECIAL REFERENC E
TO THE ADAPTIVE POTENTIAL
OF POPULATIONS

Hans-Rolf Gregoriusl

INTRODUCTION
In principle, the use of any measure of biodiversity is restricted to the real or modeled
biological phenomena for which it is developed. A central biological phenomenon
is the capacity to persist (maintain identity) by adjusting to changing conditions
(adaptation) a fundamental characteristic of life per se. Therefore, studies of the
relationship between adaptive capacity and biodiversity represent one of the pivotal
points in ecological and evolutionary research.
Elaboration of distinctive features of diversity measures that are of relevance
to problems of adaptation requires a clear concept of adaptation and its basic
mechanisms. In order to allow specification of the position in a broader framework
that any particular measure takes, the present demonstration will be based on
concepts of adaptation that apply to the level of the ecosystem just as well as to
that of the individual. (The definitions of terms compiled in Table 1 sketch the con-

1 Abteilung fur Forstgenetik und Forstpflanzenzuchtung, Universitat Göttingen, Busgenweg 2, D

37077 Göttingen.

145
cept applied in the present considerations; for a detailed system theoretical
formulation of the concept of adaptation - see Gregorius 1993). Among these levels
the smallest unit of biological organization that shows continuity in time and is
capable of adaptation is to be found in the population, and the mechanisms of
adaptation are part of its genetic system. In this sense the population is the smallest
unit of evolution and adaptation* . At lower levels of organization, such as the
individual or cell, (physiological) adaptation lacks continuity simply because of
limitation of life span.

Adaptedness - adjustedness of a character to an environment in the sense that the vital functions of its
carriers are not impaired in this environment.
Adaptation - (1) any process ultimately leading to a state of adaptedness (thus securing the persistence,
integrity, or identity of the underlying system); (2) an adapted character.
Regulatory adaptation - adaptation not involving a change in system state (such as specified by
genotype, genetic structure, species spectrum, etc.).
Structural adaptation - adaptation involving changes in system state.
Adaptability - of a system state; capacity of a system state to adapt (regulatorily or structurally) the
system’s response (output) to a specified environment (system input).
Adaptive potential - of a system state; set of all environmental conditions to which a system state is
adaptable.

Table 1: Terminology of adaptation

While adaptive processes in populations clearly serve their physical perpetua-

tion (genealogical continuity, see Gregorius 1994), the system identity maintained
by adaptive processes at the ecosystem level is less obvious. The most appealing
candidate for specification of an ecosystem’s identity might be its species spec-
trum (including producers, consumers, and decomposers). However, changes in
the species spectrum occur regularly and continuously, and it is not clear how such
changes can be conceived of as a constituent of an adaptive system intrinsic to the
ecosystem. Anyway, this concept would contain circular reasoning by using the
species spectrum as a mechanism of maintaining the species spectrum.
On the other hand, the characteristics of the species spectrum determine the
capacity of an ecosystem to sustain the nutrient cycle and to balance the nutrient

* since in asexually reproducing species no genetic recombination takes place between individuals,
each individual line of descent can be conceived of as representing a separate population.

146
 budget under the limitations set by the external (primary. mainly physical or abio-
tic) environment. (The book of DeAngelis, 1992, gives a review of current concepts
and modelling efforts of nutrient cycling; for the hierarchical aspects of balancing
nutrient budgets of ecosystems that are relevant in regulatory and structural
adaptation, see e.g. Bums et al. 1991 or Ulrich 1993). Hence, more fundamentally,
the identity of an ecosystem is definable by characteristics of its nutrient cycle, and
the biological entities mediating and balancing this cycle are specified by the spe-
cies spectrum. Among the most important members of this spectrum are the “key
species”, i.e., those species which are indispensable for sustenance of the nutrient
cycle under the respective external environmental conditions (the term “key species
spectrum” will be used to emphasize this fact).
From the point of view of the population or species, all higher levels of
organization, including interspecific or other interactions at the ecosystem level,
define environmental conditions to which the population must adapt in order to
persist. These conditions can be of a probiotic type in the sense that they lower the
adaptive pressure on the population, or they can be of an antibiotic type, in which
case they add to the adaptive pressure. Examples of probiosis known from
interactions among tree species in forest ecosystems result in reduction of rates of
pest attack, improved wind protection, stabilization of water supply, or improve-
ment of soil structure and decomposition. To a certain extent such probiotic effects
are of course offset by antibiotic actions, mostly in the form of competition (for the
recently revived interest in the roles of probiosis frequently referred to as
“facilitation” versus competition in ecological research, see e.g. Bertness et al. 1993).
An ecosystem’s stability is significantly determined by the strength of adap-
tive pressures on its populations due to the interactions inherent in the species
spectrum. Probiotic interactions effectively increase the adaptive potential of the
benefiting populations by providing a buffer against exogenous environmental
pressures and by this stabilize the ecosystem (see Figure 1 for an illustration).
Genuinely antibiotic interactions can be expected to have an adverse effect, since
they introduce additional stress. Nevertheless, regulation of population size by
competition, for example, may only superficially resemble antibiotic interaction,
since unregulated growth could endanger the persistence of the concerned species
or even its ecosystem. Moreover, different species may show both pro and antibiotic
interactions, so that, in total, the pro- must at least offset the antibiotic interactions
in order to have a positive effect on ecosystem stability.
This adaptability-oriented concept of ecosystem stability centers upon
populations as units of adaptation. It assigns to the interactions within the species

147
community the role of generators of endogenous (milieu) and modifiers of
exogenous environmental conditions to which the populations have to adapt. The
basic determinant of the species spectrum lies in the sustenance of the nutrient
cycle in the respective external environment. The species spectrum itself, however,
does not constitute an adaptive
. system, since a species is replaceable by others as
long as its function in the nutrient cycle remains unaffected. Ecosystems are resis-
tant to environmental disturbances or fluctuations only to the degree to which
their populations can adapt to them. Thus the adaptability of the population
stabilizes the key species spectrum.
The intrinsic mechanisms of ecosystem stability, and thus of adaptation to
varying environmental conditions, are therefore to be sought in the genetic sys-
tems and structures of the populations. Any environmental change that exhausts
the adaptive capacity of a key species of the species spectrum destabilizes the
ecosystem. This explains the specialization of the present paper on investigations
of the relationships between adaptive potentials and a particular component of
biodiversity, namely intraspecific genetic variation.

ADAPTIVE POTENTIALS OF POPULATIONS

As stated in Table 1. the specification of adaptive potentials requires definition of a
system and its state. The system now comprises the population, and its state
encompasses all features of its genetic system relating to physiological adaptation
(where, controlled by its genotype, the individual adjusts its vital functions to its
environment; regulatory adaptation) and to evolutionary adaptation (where the
population adjusts its genetic structure to its environment; structural adaptation).
Accordingly there are two types of adaptive potential, a physiological and an
evolutionary.

Physiological adaptive potential

Physiological adaptation apparently involves two types of environment, which an
individual must harmonize,
1. an epigenetic environment, which induces the individual’s genotype to express a
character which is adjusted to
2. an adaptational environment.

148
made up of those environmental conditions for which epigenetic and adaptational
effects harmonize to guarantee physiological adaptability or adaptedness.
An extension of physiological adaptive potentials of individual genotypes to
the population must be based on distributions of genotypes, i.e., on the genetic
structure of the population. In accordance with the above definitions, physialogical
adaptedness of a specified genetic structure can be claimed for an environment, if this
structure includes a number of individuals that is physiologically adapted to this
environment and that is sufficient to maintain a viable population size.

Evolutionary adaptive potential

The establishment of genetic states of physiological adaptability or adaptedness is
an evolutionary principle, and the pertaining process of harmonizing epigenetic
and adaptational environments is one of the main characteristics of evolutionary
adaptation. Another such characteristic consists in the capacity of providing
physiologically adapted genetic structures (in the above-defined sense) under
varying environmental conditions. Any of these processes is governed by the
population’s genetic system and is realized through changes in the genetic structure.
This identifies the genetic structure as the state of the system that determines the
evolutionary adaptive potential of a population. Hence, the evolutionary adaptive
potential of a specified genetic structure of a population can be defined by all
environmental conditions which, starting with the specified structure allow
evolution of a physiologically adapted genetic structure. In more special situations
where a population is considered at a given instant of time, the system state must
be complemented by the demographic components (age class distribution, sex ratio,
reproductively effective population size, etc.) in order to enable realistic
specifications of evolutionary adaptive potentials.

CHARACTERIZATION OF EVOLUTIONARY ADAPTIVE POTENTIALS

In principle, all non-transient genetic types in a population may contribute to its
evolutionary adaptive potential. This view gains support from the reasoning that
differences in genetic structure are associated with adaptive differences for at least
one environment, provided the differing genetic types are not generally
malfunctioning. However, different classes of environments may make basically
different adaptive demands on genetic structures.
When characterizing environments of populations, spatial distribution and
dynamics (and thus change in time) have to be included as constituent parts. The

150
adaptive significance of an environment encompasses the proportion of the
population on which it acts (its expanse), the rate at which it expands or diminishes,
and possibly the speed of modification of a continuous environmental variable such
as temperature. Expanse and speed are thus distributional and dynamic characteristics
of environments of populations to which certain characteristics of their genetic
structures must correspond in order to allow for evolutionary adaptation.
For example, the chance of a genetic type to contribute to the evolutionary
adaptive potential for fast or wide-spread environmental changes may increase
with the initial frequency of this type. Typical situations are realized in spatially
adaptively differentiated plant populations in which one of the local environments
expands as a consequence of some major environmental change. The genetic types
adapted to the (pre-existing) local environment in principle provide adaptive
potential for this change. Yet, if the expansion occurs at a high rate and if the initial
frequency of the adapted types is low, the selection process might be accompanied
by reductions in population size which are too drastic to allow restoration of a
viable population and therefore prevents adaptation. Adaptation to such quantitative
environmental dynamics is quite likely to depend on frequency distribution
characteristics of the genetic structure in combination with the speed of
environmental change.
Qualitative environmental dynamics, on the other hand, where previously largely
nonexisting environmental conditions spread, are adaptively more demanding in
that they may require the formation and multiplication of new gene complexes
(genotypes). Adaptability to such changes is very sensitive to their rates. since for
high rates the adaptive lag increases continuously and may ultimately completely
obstruct adaptation. For qualitative environmental dynamics, adaptability may
therefore completely depend on the presence of rare genetic variants such as can
be maintained by recurrent mutation or gene flow.
In order to distinguish between the adaptive consequences of the above two
classes of environments for the genetic structures of populations, the terms operating
and latent genetic potential were suggested (Stebbins and Hart1 1988, Bergmann et
al. 1990). The operating genetic potential consists of those genetic types which
contribute to the adaptedness of a population to its current environmental
conditions; these genetic types usually prevail. The adaptive reserve is to be found
in the rarer genetic variants which form the latent genetic potential for adaptive
demands of future environmental changes. Obviously the adaptive potentials for
quantitative and qualitative environmental dynamics are based on the operating
and latent genetic potential, respectively. The capacity to maintain latent potential

151
(which is adaptively inferior under the currently prevailing conditions) is a vital
part of a population’s capacity to preserve its adaptability and is therefore one of
the most important characteristics of genetic systems.
Numbers of prevalent genetic types and the existence of rare ones are also the
subject of the well established distinction between “major” and “minor”
polymorphisms (Lewontin 1985), which relates these to the above two types of
genetic potential. In particular, the rare types in minor polymorphisms can be
considered as candidates for the latent genetic potential. However, unambiguous
distinction between rare and prevalent types is only possible in the absence of types
with intermediate frequency, i.e., where frequency distributions are highly
concentrated. The observation of genetic types with intermediate frequencies could
indicate that adaptive processes are transforming the genetic structure into a new
state of adaptedness.
This line of thinking motivates consideration of characteristics of genetics frequency
profiles (in which genetic types are ordered by decreasing frequency; Gregorius 1992,
Gregorius and Bergmann 1994) as indicators of the existence of potentials for
adaptation to certain classes of environmental conditions. The significance of this
approach for interpretation of the majority of the common methods of measuring
genetic variation will be referred to later.

Adaptive potential and phenotypes

The above considerations reflect the common approach of correlating characteris-
tics of genetic frequency profiles with more or less well-specified classes of
environments in order to analyze adaptive potentials. Phenotypes as the direct
objects of adaptedness and adaptability are not explicitly regarded in this approach.
Yet phenotypic variation may reflect adaptive events when observable for traits,
the states of which attest adaptedness or lack of it. Traits, particular states of which
represent stress symptoms, belong to this category. In fact, while transient stress
symptoms may be the result of a physiological adaptation process, individually
persistent or even intensifying stress symptoms indicate adaptive failure with
respect to a current environmental pressure.
The latter situation calls for evolutionary adaptation, the potential for which
can only be recognized as genetic differences between groups that are distin-
guished by the intensity of the persistent stress symptoms (Gregorius 1994). The
prerequisite for existence of this potential is that the environmental factors
causing the reaction are also the ones to which adaptation is required, i.e. that
epigenetic and adaptational environments coincide. For stress traits this prerequisite

152
is realized by definition. As a consequence of this coincidence, the (evolutionary)
genetic potential available for adaptation to the stressful conditions includes those
genetic types that exhibit no or only minor persistent stress symptoms when
subjected to the stressor.
This approach to the description of evolutionary adaptive potential provides a
concrete example of the above general definition, in that it assigns to a fraction of a
population’s genetic structure those environments for which this fraction contains
the genetic potential for adaptation. This fraction also determines the initial
population size available for adaptation (the adaptively qualified population size) and
together with the genetic variability remaining within the adaptively qualified
population, the bounds for the population’s future adaptability are set. It is thus
important to obtain information on the degree to which adaptation to the stressing
conditions reduces the genetic potential for future adaptive processes. For the
experimental verification of an adaptive condition, estimations of fractions of
populations containing genetic potential for adaptation to current environmental
pressures as manifested in the phenotype are therefore highly desirable.
Practical problems in obtaining such estimates arise, of course, from poor
suitability of the available gene markers and the fact that stressing conditions do
not affect all members of the population equally Despite genetic disposition, stress
symptoms may not be expressed either because not all individuals are exposed to
the stressor or because the presence of other environmental conditions neutralize
the stress effects. Nevertheless, based on appropriate sampling
. methods there is a
practical solution to this problem even for studies of populations in situ , as will be
demonstrated later.

CHARACTERIZATION OF FREQUENCY PROFILES BY MEASURES OF

DIVERSITY
The common approach to the measurement of ecological or genetic diversity of a
collection of organisms aims at summarizing numbers and abundances of types
within the collection into a single value. In this sense, diversity represents itself as a
characteristic of a frequency profile q of defined types, where such profiles are
represented by vectors of relative type frequencies arranged in decreasing order
(i.e. q 1 q 2 = 1). The usage of relative in place of absolute frequencies is
implied by the fact that, in general, the diversity of a collection need not depend on
its size. Collection size is thus considered to represent an independent quality. Yet,
this does of course not exclude the possibility of strong (positive) correlations

153
 Apparently, among the characteristics not explicitly reflected by current diver-
sity and evenness measures are those distinguishing prevalent from rare types, as
is required for the recognition of operating and latent genetic potentials and the
associated states of adaptation. The following chapter will therefore be devoted to
distinguishability between rare and prevalent types (provided the latter are
defined) as an adaptively important characteristic of genetic frequency profiles.

DISTINGUISHABILITY BETWEEN PREVALENT AND RARE TYPES IN

FREQUENCY PROFILES
Genetic frequency profiles of populations are shaped by adaptational processes.
These include the possibility of temporary adaptive neutrality or quasi-neutrality
of certain genetic traits (the latter leading to more or less erratic profiles). As was
argued earlier in connection with operating and latent genetic potentials, states of
adaptedness are likely to be characterized by a number of prevalent genetic types
and a remainder of more or less rare types (for a more detailed discussion with
special reference to isoenzyme data collected in forest tree populations see
Gregorius and Bergmann 1994). In such genetic profiles, prevalent and rare types
(i.e. operating and latent potential) are separated by a distinct step in the frequency
profile, as is exemplified in the first three profiles in Figure 2. On the other hand,
certain processes of adaptation progressing from one state of adaptedness to
another have to pass through intermediate profile characteristics in which either
no distinct frequency steps exist (such as in the rightmost profile in Figure 2 ) or in
which several types are approximately evenly distributed.
Moreover, the above reasoning suggests that operating and latent genetic
potentials correspond to evolutionary adaptive potentials for quantitative and
qualitative environmental dynamics, respectively. Therefore, characteristics of’
genetic frequency profiles indicating differentiation between the two forms of
genetic potential and quantifying their sizes are of considerable interest in the
evaluation of adaptive potentials.
With the probable exception of certain mating systems, such as gametophytic
incompatibility, where large numbers of evenly distributed genetic types may
stably coexist irrespective of the adaptive pressures of the external environment,
the operating genetic potential for single traits can be expected to concentrate on
only relatively few types. The main reason is that even with frequency-dependence
the conditions for selective maintenance of large numbers of alleles are too specific
to be realizable in a continually varying environment (the first three fictional profiles

156
it sufficiently exceeds the other step sizes, then one might agree that the types located
to the left of the maximum step size in the profile are the prevalent types. The less
distinguished the maximum step size is from the rest, the more the ambiguity in
the assignment of prevalence increases. As an intrinsic part of the present concept,
complete ambiguity exists if all step sizes are equal and the profile is linearly
declining. In fact, the measure quantifies this ambiguity. Thus the large values
for the first three profiles in Figure 2 strongly suggest that all types to the left of
the maximum step size (which is also large) be considered as prevalent, so that
each of the first two profiles contains one prevalent type and the third profile contains
two such types. However, this approach to the estimation of numbers of prevalent
types will not be further expanded upon in the present paper.

ESTIMATION OF ADAPTIVE POTENTIALS

FOR SPECIAL ENVIRONMENTS
We now return to the above problem of estimating in natural populations genetic
potentials for adaptation to special environments which are recognizable by their
effects on stress traits. For problems of feasibility, such estimates must resort to
gene markers of at least partially unknown function, so that any method is limited
to the degree to which such gene markers can detect adaptive potentials. To sim-
plify the derivations, only two trait states, called “sensitive” and “tolerant” will be
considered. The applied notation and assumptions are listed in Table 2. The
assumption of stochastic independence between genotypes (D and G) and
environments (U) is generally difficult to defend in naturally regenerating plant
populations because of the possibility for the evolution of differential local
adaptations. However the assumption is likely to be realized at least approximate-
ly in many cases, since gene flow within habitats may be considerable and the
gene loci controlling the stress trait as well as those under observation may not
be involved in micro-spatial adaptive differentiation. In addition, application of
appropriate methods of structured sampling (Gregorius 1989) can aid fulfillment
of the condition of independence.
The aim is now to distinguish the group of individuals with tolerant disposition
from those with sensitive disposition on the basis of an observable genetic trait and
observable stress symptoms. In the present context, usage of the term "disposition”
presumes that the environment has no share in the modification of the concerned
characteristics, which leaves genetic effects as the only cause for individual
differences in disposition.

165
reacting and sensitively disposed individuals. Consequently, by equations (2)
and (3)

P(D=t).[P(G=g lD=t)-P(G=g ID=s) ] =

=P(G=g,R=t)-P(G=g,U#s,D=s)-P(D=t).P(G=g lD=s)
=P(G=g,R=t)-P(G=g,D=s).P(D#s)-P(D=t).P(G=g lD=s )
= P(R = t).P(G = g I R = t) - P(G = g I D = s).[P(U s). P(D = s)+ P(D = t)]
=P(R=t).P(G=g lR=t)-P(G=g lR=s).P(R= t )
=P(R=t).[P(G=g lR=t)-P(G=g IR=s) ] (4)

This equation establishes the fundamental relation between the distribution

of an observable genetic trait across phenotypic stress states and across (non-
observable) stress dispositions. Since the term “disposition” addresses a purely
genetically determined character, carriers of the same (multilocus) genotype
coding for tolerant or sensitive disposition cannot appear in both sets [D=t] and
[D=s]. Hence, P(D=t) equals the fraction of the population containing the genetic
potential for adaptation to the stressing conditions, which, in agreement with the
previously introduced terminology, will be called the adaptively qualified frac-
tion and the estimation of which is the subject of this section.
The practical problem, however, stems from the fact that the observable genetic
traits may only partially separate one class of stress disposition from the other in
that carriers of a particular genotype may occur in either the class of sensitive and
of tolerant disposition. It is therefore of central importance to determine the degree
to which any particular genetic trait contributes to the distinction between the classes
of stress disposition. The proportion of individuals with tolerant disposition that
can be distinguished from those with sensitive disposition by the observable genetic
trait is therefore all that we can identify from the target quantity P(D=t). This estimate
constitutes a lower bound.
The principle of such a distinction involves for each attribute (observable
genotype) the determination of the number of members of one set (tolerants) left
after subtraction of those members from the other set (sensitives) showing the same
attribute. The sum of these differences over all attributes specifies the extent to
which the first set differs from the second. To exclude the possibility that
characteristics other than those specified by the attributes of interest affect the
measurement, the size of the second set must be the same as that of the first set.

167
potential for adaptation in the following generations (i.e. the evolutionary adap-
tive potential) by the values for the allele frequencies. However, it is the system
combining these genes into genotypes that determines the capacity to realize the
inherent genetic adaptive potential.
Variability of the observable genetic trait - A problem of more general signifi-
cance arises with the usage of highly variable genetic traits as can result from the
inclusion of large numbers of polymorphic loci into a study. At the extreme, each
member of the population is distinguished from each other by its observable multi-
locus genotype, which enforces a genetic distance of 1 between the groups of
sensitive and tolerant individuals both at the level of disposition and phenotype.
Thus = P(D=t) = P(R=t) irrespective of the joint distribution of D and U which
is a contradiction. The reason for this is to be found in the complete association
between G and L . which results from the fact that no genotype is repeated across
the environments; the basic prerequisite of stochastic independence between G
and U is thus invalidated.
Therefore, in contrast with many other applications such as tracing descent,
highly variable genetic traits may not constitute the appropriate tool for the
estimation of adaptive potentials. It may therefore be preferable, as was done in the
above interpretation of the results of Konnert, to compute or for each of
several genetic traits of intermediate variability and take the maximum of these
values over traits as an estimation of the adaptively qualified fraction (or its part
among the phenotypically tolerant individuals).

REFERENCE
Bergmann, F., H.R. Gregorius and J.B. Larsen. 1990. Levels of genetic variation in
Europe at silver fir (Abies alba). Are they related to the species’ decline? Genetica
82: l-10.
Bertness G., Mark D.H.G. Shumway and W.H. Scott. 1993. Competition and
facilitation in marsh plants. Am. Nat. 142(4): 718-724.
Bums, TX, B.C. Patten and M. Higashi. 1991. Hierarchical evolution in ecological
networks: environs and selection. Pages: 221-239 In: M. Higashi, T.P. Burns (eds.).
Theoretical Studies of Ecosystems. Cambridge University Press.
DeAngelis, D.L. 1992. Dynamics of Nutrient Cycling and Food Webs. Chapman
and Hall DeBenedictis P.A. 1973. On the correlations between certain diversity
indices. Amer. Natur. 107: 295-302.

171
Gregorius, H.R. 1978. The concept of genetic diversity and its formal relationship
to heterozygosity and genetic distance. Math. Biosci. 41: 53-271.
Gregorius, H.R. 1987. The relationship between the concepts of genetic diversity
and differentiation. Theor. Appl. Genetics 74: 397401.
Gregorius, H.R. 1989. The attribution of phenotypic variation to genetic or
environmental variation in ecological studies. Pages: 3-15 In: F. Scholz, H.R.
Gregorius, D. Rudin (eds.). Genetic Effects of Air Pollutants in Forest Tree
Populations. Springer-Verlag.
Gregorius, H.-R. 1980. The probability of losing an allele when diploid genotypes
are sampled. Biometrics 36: 643-652.
Gregorius, H.-R. 1990. A diversity-independent measure of evenness. Amer. Nat.
136: 701-711.
Gregorius, H.-R. 1992. A commentary on current approaches to forest population
genetics. New Forests 6: 409420.
Gregorius, H.-R. 1993. Systemeigenschaften der Anpassung. Gottingen Research
Notes in Forest Genetics 16: l-25.
Gregorius, H.-R. 1994. Strategies for conserving genetic diversity in the face of global
change. Pages: 157-182 In: T.J.B. Boyle and C.E.B. Boyle (eds.). Biodiversity,
Temperate Ecosystems and Global Change. Springer-Verlag.
Gregorius, H.-R. and F. Bergmann. 1994. Analysis of isoenzyme genetic profiles
observed in forest tree populations. In: Ph. Baradat (ed.). Population Genetics
and Genetic Conservation of Forest Trees. SPB Academic Publishing, in press.
Hennink, S. and A.C. Zeven. 1991. The interpretation of Nei and-Shannon-Weaver
within population variation indices . Euphytica 51: 935-240.
Hurlbert, S.H. 1971. The nonconcept of species diversity: a critique and alternative
parameters. Ecology 52: 577-586.
Konnert, M. 1992. Genetische Untersuchungen in geschadigten
Weißtannenbestanden (Abies alba Mill.) Siidwestdeutschlands. Doctoral thesis
Gottingen.
Lewontin, R.C. 1985. Population genetics. Annual Review of Genetics 19: 81-109.
Nei, M. 1975. Molecular Population Genetics and Evolution. North-Holland,
Amsterdam, Oxford.
Peet, R.K. 1975. Relative diversity indices. Ecology 56: 496-498.
Pielou, E.C. 1975. Ecological Diversity. Wiley, New York.
Pielou, E.C. 1977. Mathematical Ecology. Wiley, New York.

172
 Chapter

11
MOLECULAR POPULATION GENETICS AN D
EVOLUTION: TWO MISSING ELEMENTS I N
STUDIES OF BIODIVERSITY

Alfred E. Szmidtl

INTRODUCTION
Biodiversity has become the subject of intensive debate among politicians and
scientists alike. Hardly a week goes by without a new report or conference about
biodiversity. As a result, a massive amount of information has accumulated covering
impressively diverse fields ranging from sociology to economics (e.g., Gershon
1992, Machlis 1992, Perrings et al. 1992). So pervasive is the use (and misuse) of this
term in the mass media and scientific media that by now it can be viewed as a new
buzzword, surpassing its recent predecessor: biotechnology (see Lovett (1994) for
opposite view on this subject).
Biodiversity refers to the variety and variability among living organisms and
the ecosystems in which they interact (Woodruff and Gall 1992). Therefore, it is
intrinsically associated with the genetic system. Genetics and evolution sometimes
enter the biodiversity debate but most of the current discussion is focused on
taxonomic and ecological aspects of biodiversity (Faith 1994, Platnick 1991,

l Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences,
S-901 83 Umeå Sweden.

177
Prendergast et al. 1993, Renner and Ricklefs 1994). When genetics is considered in
biodiversity programmes, it is often concerned with collection, long term
preservation and cataloguing of genetic variation detectable by an ad hoc chosen set
of genetic markers rather than with assessing the processes that created its present
pattern.
Rapid advances in molecular biology have furnished a wide array of new
methods to study genetics and evolution of plants. Many of these methods pro-
vide excellent means for acquiring genetic information relevant for biodiversity
conservation programmes. However, successful utilization of these methods
requires a good understanding of the type of genetic information they can provide.
In this contribution, I briefly describe currently available methods for the detec-
tion of genetic variation and suggest how they can be used in order to improve our
knowledge about the genetic components of tree biodiversity.

CURRENT APPROACHES TO STUDY BIODIVERSITY

Species inventories are regarded as crucial to solving the ‘biodiversity crisis’ ; see
Renner and Ricklefs, (1994) for discussion on this subject. Species richness and the
presence of rare species are the most frequently cited criteria for site selection for
conservationists (Prendergast et al. 1993). In the past, conservation decisions have
been directed towards saving prominent species, representative ecosystems and
threatened areas (Renner and Ricklefs 1994). However, species-rich areas frequent-
ly do not coincide for different taxa, and many rare species do not occur in the most
species-rich areas (Prendergast et al. 1993). Thus, as noted by Platnick (1991)
numbers of species alone are an inadequate guide to the relative importance of
individual areas. Furthermore, predictions of species distributions due to the cli-
mate change have neglected the evolutionary potentials that exist within species
(Eriksson et al. 1993). Thus, even taxonomists came to doubt whether detailed species
inventories are needed to identify areas for conservation (Renner and Ricklefs
1994). First, full counts of all organisms are impossible on any scale. Second, saving
species may not save all of their useful alleles and may not save the communities or
their evolutionary dynamic (Namkoong 1992). Third, attempts to save communi-
ties may not even save the species they presently contain (Namkoong 1992).

GENETIC ISSUES IN BIODIVERSITY CONSERVATION

It is sometimes suggested that the extinctions of species due to modification of

178
environment are taking place so rapidly that the question of loss of genetic varia-
bility within species is moot because all individuals may have disappeared before
drift has a chance to operate (Savolainen and Kärkkäinen 1992). The many ecological
causes of extinction often operate before lack of genetic variation becomes an
issue (Lande 1988). Although genetic research may not always be able to catch up
with the pace of species extinction, the decline of many species is slow enough to
warrant meaningful consideration of its genetic causes and effects. Moreover, the
loss of genetic variation is not the only potential outcome of environment
modification. Various genetic strategies and recommendations for biodiversity
conservation have been formulated (Eriksson et al. 1993, Gregorius 1991, Hedrick
and Miller 1992, Kresovich and McFerson 1992, Namkoong 1992). Most of these
initiatives stress the importance of documenting patterns of genetic variation, as
well as providing an understanding of the evolutionary determinants that influence
these patterns. .
Information on tree genetic variation, central to the design of appropriate
conservation strategies is still scarce. Intraspecific variation has not been thoroughly
studied in temperate regions, and there is only sparse information from tropical
areas (Bawa and Ashton 1991, Ehrlich and Daily 1993). Likewise, research is lacking
in the areas of taxonomy, phylogeny, reproductive biology and ecotypical
differentiation, especially of tropical trees (Williams 1991). It is ironic that the start
of interest in forest genetic resources was about the same time as that for crop
genetic resources, but never attracted adequate funding (Williams 1991).
Genetic variation is a result of changing evolutionary histories and in itself is of
value to the present and future individuals, populations, and species in which it
occurs (Namkoong 1992). It is a prerequisite for future evolution and biodiversity
conservation programmes should provide opportunities for it (Eriksson et al. 1993).
Therefore, programs conserving genes in forest trees should be based on
evolutionary concepts and the existing adaptations should be used when
populations are appointed as gene resource populations (Eriksson et al. 1993). It is
not clear however, how these adaptations can be assessed and evaluated. Obvious-
ly, we lack explicitly stated and realistic tactical means for implementation of the
proposed strategies.

COMPONENTS AND DETERMINANTS OF GENETIC VARIATION

The biological function of plants relies on an intimate interplay between three dis-
tinct genomes: nuclear, chloroplast and mitochondrial. All these genomes harbour

179
 genes which are vital to growth, photosynthesis, respiration and other biological
processes. Therefore, studies of genetic variation should consider all these three
components of the plant genetic system. Furthermore, each of these components
harbours different structural, RNA and regulatory genes as well as non-coding
sequences. The relative proportion of coding and non-coding sequences differs
among chloroplast, mitochondrial and nuclear DNA. A substantial part of the
nuclear genome contains non-coding sequences. In contrast, chloroplast genome
is dominated by coding sequences. Ideally, studies of genetic variation should
consider all these types of sequences.
Selection, mutation, drift, gene flow and mating system are among the most
important evolutionary determinants of genetic variation. The relative significance
of these determinants is likely to vary among genes, populations, species and
habitats. Large populations may be immune to drift but may be exposed to highly
variable selective pressures. Small populations may suffer loss of variation due to
an increased inbreeding and drift but experience only weak environmental pressures.
The amount and distribution of genetic variation is also closely associated with
the mode of its transmission and the rate of recombination among loci. In contrast
to biparentally inherited nuclear genes, cytoplasmic genes show predominantly
uniparental inheritance (Birky 1988, Clegg 1989). Chloroplast genes are generally
maternally inherited in angiosperms, and paternally inherited in gymnosperms
(Conde et al. 1979, Szmidt et al. 1987), while mitochondrial genes are maternally
inherited in most plants (Palmer 1992). As a consequence of these different modes
of inheritance, the extent of gene flow among populations may differ for biparen-
tally, maternally and paternally inherited genes (Birky et al. 1989). Therefore, the
extent of population differentiation is expected to vary among nuclear, maternally
and paternally inherited genes for the same set of populations (Birky et al. 1989;
Ennos 1994).
Because of the heterogeneous nature of evolutionary determinants of genetic
variation we need specific, explicitly formulated approaches to study their effects.
Most of our knowledge about the amount and distribution of genetic variation in
forest trees comes from surveys of genetic marker variation. The validity of this
knowledge depends on the type of genetic information provided by the markers
used to generate it. One important question concerning the informativeness of
genetic markers is: what is their genomic origin, transmission and function ? Another
important question is: which evolutionary determinant is most likely to affect the markers
employed in our studies? Depending on the nature of evolutionary determinants,
individual markers may or may not respond to them. For instance, in the absence

180
of selection, genetic variation will be affected by mutation, gene flow and drift.
These processes will affect all loci, regardless of their function. On the other hand,
selective forces are more likely to affect variation of the structural, regulatory and
RNA genes than that of non-coding portions of the genome. Thus, the patterns of
variation at selected loci may differ from that of neutral loci (Hattemer 1991).
The wide array of currently available molecular techniques has greatly
improved our ability to study genetic variation and to discern its evolutionary
significance. However, there still appears to be much confusion among biodiver-
sity students with regard to the genetic informativeness and feasibility of parti-
cular methods. In the next three sections, I briefly summarize these methods and
describe the type of genetic information they provide.

MOLECULAR METHODS FOR DETECTING GENETIC VARIATION

Enzyme markers
Enzyme markers represent electrophoretically detectable forms of enzymatic
proteins visualized by substrate-specific staining. The predominantly codominant
character of enzyme variation, low cost and technical simplicity of the analysis
(Weeden and Wendel 1989, Wendel and Weeden 1989) are among the main reasons
for the widespread use of this category of markers in studies of genetic variation.
In most cases, the genetic variation detectable by enzyme markers is associated
with biparentally inherited nuclear genes. The nuclear origin of these markers also
implies that their variation is influenced by recombination. Enzyme loci often show
a considerable degree of polymorphism that makes them particularly useful for
studies of gene flow, mating system and the effects of drift. A serious disadvan-
tage of enzyme markers is their low number and a highly restricted group of
structural genes they represent. Moreover, apparent technical simplicity of enzyme
markers is not as great as sometimes suggested due to the lack of standard
‘portable’ protocols that are applicable to different species.

Restriction fragment length polymorphism (RFLP) markers

Restriction analysis and fragment hybridization are the most common methods
for detecting genetic variation at the DNA level. When DNAs from two gene-
tically distinct individuals are analyzed by these methods, polymorphism some-
times appears due to differences in the number of DNA sites that are cleaved by a
restriction enzyme. The molecular basis of RFLPs is: loss or gain of a restriction site

181
 due to a point or length mutation, or inversion. Such events result in a length
difference in the DNA fragments detectable by restriction analysis or by fragment
hybridization. Detailed descriptions of RFLP markers can be found in the
published record e.g., Gillet (1991), Szmidt and Wang (1992 and references therein).
Depending on the genomic origin of the probe, it is possible to obtain RFLP
markers for both nuclear and cytoplasmic genomes, showing either biparental or
uniparental inheritance respectively Similar to enzyme markers, nuclear RFLPs
are typically codominant and display simple Mendelian inheritance. By using
probes with at least partially known sequences it is possible to develop RFLP
markers for both coding and non-coding sequences. A somewhat special category
of RFLP markers can be obtained by using probes homologous to short multiple
repeats (microsatellites). The genetic information yielded by this category of
markers is difficult to determine at present.
In contrast to protocols for the analysis of enzyme variation RFLP protocols are
easily standardized, and can be applied to various materials. Unfortunately, despite
many advantages of RFLP markers their use in population analysis is constrained
by several factors. First, RFLP analysis usually requires large amounts of DNA,
making non-destructive sampling difficult. Second, well documented probes are
still scarce, which seriously limits the number of available RFLP markers. The use
of random probes retrieved from libraries constructed with total DNA digests
often does not permit the unambiguous determination of whether the observed
RFLPs are of nuclear or cytoplasmic origin. Moreover, it is not possible to deter-
mine the functional significance of genetic variation detected by such probes. RFLP
variation is detected by restriction enzymes that have specific target sequence. As
not all mutations alter restriction enzyme sites many potentially important muta-
tions will go undetected. Therefore, a battery of restriction enzymes must be used
which seriously increases the time and cost of analysis. Finally, the technique is
complex and requires well equipped laboratories which further limits its use in
many countries.

PCR-based markers
The advent of the polymerase chain reaction (PCR; Saiki et al . 1988) has profoundly
improved both the speed and efficiency of detecting all types of sequence varia-
tion. The two most commonly used techniques employing PCR are DNA
amplification by two-primer extension (Mullis et al . 1986, Mullis and Faloona 1987,
Saiki et al. 1988 ) and random amplified polymorphic DNA (RAPD) that employs a
single primer of arbitrary sequence (Gillet 1991, Williams et al . 1990). In both

182
techniques oligonucleotide primers hybridize with complementary sequences,
located on template strands of single-stranded DNA. DNA polymerase begins
extending the primers at their open 3’ ends by adding nucleotides complementary
to the nucleotide sequence of the template. Primer extension continues in the 3’ 5’
direction until the end of the template is reached or until the termination of the
cycle. The cycle is then repeated by heat denaturation of the double stranded DNA
to separate it into single strands, followed by cooling to allow annealing and
subsequent extension of the primers (Gillet 1991, Williams et al. 1990). In this way,
any given DNA sequence can be amplified and studied. In the two-primer exten-
sion technique, the template nucleotide sequence to be amplified is known in
advance, so that the two primers can be chosen to be short sequences complemen-
tary to the 3’ ends of the template and its complement (Gillet 1991). The RAPD
method differs from the standard two-primer amplification in that sequence varia-
tion is detected using a single primer of arbitrary sequence instead of a pair of
template-specific primers, uses more cycles and lower annealing temperature
(Williams et al. 1990).
Variation of the amplification products can be analysed in several different
ways. Amplification products can be separated by size on agarose gels. This
approach is typically used to detect size variation of RAPD fragments, but is
insensitive to sequence differences among them. Sequence differences can be
detected either by digestion of individual fragments with restriction enzymes or
by mismatch analysis (see Cotton 1989, for detailed descriptions of the latter
method). The mismatch approaches to study DNA variation detect most muta-
tions. This is particularly important in studies of conserved coding sequences.
Restriction enzymes or mismatch detection are typically used to analyse sequence
variation among products generated by two-primer extension. The fragments are
then separated on acrylamide gels and visualized by silver staining (Bassam et al.
1992, Caetano-Anolles et al. 1991).
Depending on the primer, PCR-based markers can be either gene-specific or
random. RAPDs usually detect variation in non-coding regions of nuclear DNA
and show dominant Mendelian inheritance (e.g., Lu et al. in press, Roy et al. 1992).
In contrast, two-primer extension with gene-specific primers provides detailed
information about the genomic origin and coding function of the amplified sequence.
Markers produced by two-primer extension show codominant variation and bi-
or uniparental inheritance. Alignment of known gene sequences from various
organisms enables construction of ‘consensus’ primers that can be applied to other
species for which the sequence data is not available. For instance, in combination

183
with restriction fragment analysis, we used this approach for studies of chloro-
plast and mitochondrial gene variation in a wide range of plant species (Wang and
Szmidt, in preparation). Fast accumulation of sequence data warrants that soon
many additional nuclear and cytoplasmic genes can be studied in a similar way
Relatively low cost, simple standard extraction and amplification protocols
(McPherson et al. 1991) make PCR-based markers applicable to virtually any
species. In contrast to the RFLP analysis, PCR markers can be studied in a minute
amount of tissue allowing for non destructive sampling.

WHAT DO WE KNOW ABOUT GENETIC VARIATION

AND HOW REPRESENTATIVE IS THAT KNOWLEDGE?
Most available data on genetic variation in trees come from surveys of enzyme
variation (see Hamrick and Godt 1991, for recent compilation). This information is
slowly being enriched with data on RFLP and RAPD variation (see Szmidt 1991,
Szmidt and Wang 1992 for recent reviews). The question thus arises: to what extent
are these data sets representative of the genome as a whole? At best, enzyme surveys are
instructive as to the relative levels of variation in structural, biparentally inherited
nuclear genes. Current surveys of nuclear DNA variation typically employ either
random nuclear probes with unknown gene content or RAPD markers (e.g., Devey
et al. 1991, Roy et al. 1992, Lu et al. in press). Therefore, they are likely to be instruc-
tive as to the relative levels of variation in the non-coding portion of the nuclear
genome. Our knowledge about other nuclear genes in trees is restricted to varia-
tion in the copy number of RNA genes (Strauss and Tsai 1988, Gorman et al. 1992,
Govindaraju and Cullis 1992, Moran et al. 1992, Karvonen et al. 1994) and there is
very little information about variation in mitochondrial, chloroplast and regula-
tory genes.
As mentioned earlier, most genetic markers employed in studies of genetic
variation are of nuclear origin and do not provide information about variation of
cytoplasmic genes. A common observation in most studies of allozyme and nuclear
RFLP variation in trees is weak differentiation among populations (Hamrick and
Godt 1991). This appears to be true even for extremely distant populations (Szmidt
and Wang 1993). Available evidence for population differentiation with respect to
cytoplasmic genes shows a very different picture. Analysis of mitochondrial DNA
variation in Pinus contorta and P. banksiana revealed substantial differentiation
within and between these taxa (Dong and Wagner 1993). Lower, but still substan-
tial differentiation in the same set of species was found with respect to chloroplast

184
 DNA (Dong and Wagner 1994). Similar discordance with respect to the
apportionment of genetic variation between nuclear and chloroplast markers was
found in populations of P. densata (Wang et al. 1990, Wang and Szmidt 1990, Wang
and Szmidt 1994) and other Pinus species (Hong et al. 1993, Strauss et al. 1993). The
extent of these differences is likely to be a function of the relative amounts of
interpopulation pollen and seed flow (Ennos 1994). Research on gene flow has
lagged because gene flow rates have been assumed to be insignificant (Ellstrand
1992). Now that gene flow rates have been recognized to occur frequently at levels
that can influence the genetic fate of populations, we must re-evaluate the impor-
tance of gene flow in conservation biology (Ellstrand 1992). Simultaneous
measurements of population differentiation for markers with different modes of
inheritance can provide important information on this subject (Ennos 1994, Szmidt
and Wang 1992). Geographic variation may also be generated by historical patterns
of migration and inter-specific gene exchange. Examples of such variation have
been demonstrated for several boreal and tropical tree species (Sigurgeirsson and
Szmidt 1993a, Szmidt et al. 1993, Szmidt et al. 1988a, Szmidt and Wang 1993, Wang
and Szmidt 1994).
There have been very few reports indicating useful correlation between levels
and patterns of enzyme marker variation and adaptive morphological and
physiological traits (Savolainen and Karkkainen 1992). Thus, these data are not
informative with regard to adaptive patterns of genetic variation that are impor-
tant for recommendations for genetic conservation (Woodruff and Gall 1992). This
limits the usefulness of allozyme markers in monitoring the important genetic
changes induced by natural selection. Random DNA markers are not likely to fare
much better in this respect (Savolainen and Kärkkäinen 1992). Until recently,
population analysis of sequence differences underlying allelic variation has been
impractical because of the lack of sufficiently fast and efficient methods. This
situation has changed substantially with the introduction of PCR-reaction and new
methods for detecting nucleotide substitutions (see previous section). By devising
primers for specific structural, RNA and regulatory genes it is now possible to
select markers that are likely to show effects of selection upon allelic variation. We
are currently using this approach to study such variation in some boreal and tropical
trees (Szmidt and Wang, in preparation).
More genetic information is also necessary about the taxonomy and phylo-
genetic relationships among tree species. Such information is essential for the
determination of the demographic units that should be conserved (Lacy 1988).
Moreover, phylogenetic studies establish the basic data for identifying patterns of

185
historical biogeography, for testing hypotheses about the processes that produce
these patterns and to identify how unique histories determine contemporary
patterns; global patterns of biodiversity being a prime example (Renner and Ricklefs
1994). In the case of cytoplasmic DNAs, groups of associated restriction sites are
not separated by recombination and the ancestry of individual haplotypes may
remain recognizable even after many generations of sexual reproduction
(Whittemore and Schaal 1991). These features of cytoplasmic DNAs have made
them particularly useful for studies of phylogenetic relationships and gene flow
among tree populations (Wagner et al. 1987, Szmidt et al. 1988b, Strauss and Doerksen
1990, Strauss et al. 1990, Wagner et al. 1991, Sigurgeirsson and Szmidt 1993b, Wang
and Szmidt 1993). However, wide use of these markers is difficult because of the
lack of probes, low sensitivity of restriction enzymes and complex RFLP techno-
logy Recent development of primers specific for various mitochondrial and
chloroplast sequences (Paran and Michelmore 1993, Taberlet et al. 1991, Tripp et al.
1993), and mismatch detection of sequence variation offer much simpler and faster
means for construction of non-recombinant cytoplasmic markers.

WHAT TO DO WHEN THERE IS TOO MUCH TO DO?

Even with new molecular technology, it is impossible to evaluate the genetic varia-
tion and population structure of all species. This is especially true for species-rich
tropical forests. One way to circumvent this problem is to prioritize species for
attention. Broad genetic management strategies could be formulated for biological
groups by studying selected indicator species from a large group of species with
similar attributes (Woodruff and Gall, 1992). A multidimensional matrix for
classifying population genetic structure within life forms can be defined according
to distribution, density, mating system, and pollen vectors (Woodruff and Gall, 1992).
Representatives of each type can be studied to discover trends and generalisations
(Woodruff and Gall, 1992). Unfortunately, selectivity may make inventory prac-
tical but introduces biases resulting from the need to judge the intrinsic values of
species (Renner and Ricklefs, 1994). Such judgement requires that at least basic
biological attributes of individual species are known which is often not the case in
tropical regions. Detailed genetic studies of selected representatives should there-
fore be paralleled by ‘quick and dirty’ studies of additional taxa employing RAPDs
and a limited assortment of gene-specific markers. Information gained from such
heuristic assessments of genetic variation will help to interpolate genetic proper-
ties of additional species and to identify taxa deserving closer genetic scrutiny.

186
CONCLUDING REMARKS
Our present knowledge about genetic variation in forest trees is based on a highly
restricted number of biparentally-transmitted enzymatic proteins and non-coding
regions of the nuclear genome. It is doubtful that this data set gives even a remotely
accurate picture of the amount and patterns of genetic variation and significance of
individual evolutionary determinants affecting tree populations. The current arsenal
of molecular methods offers excellent opportunities for precise studies of indivi-
dual components of plant genetic systems. Unfortunately the choice of markers
for particular studies still appears to depend more on personal preference and
competence rather than on the capacity of individual markers to provide relevant
information about particular determinants. There is a prevailing perception that
genetic studies utilizing DNA markers are of limited use in population genetic
analysis because of financial, competence and supply restrictions. This is certainly
correct with respect to RFLP markers. For this reason, most studies of tree
populations still employ enzyme markers that are deemed easier and cheaper.
However, as demonstrated by our and other studies, PCR-based DNA markers are
much faster and cheaper than RFLPs. At the same time, they provide far more
precise information about the function and origin of the observed variation than
enzyme markers. Such information is essential for predicting the evolutionary
significance of this variation which should guide gene conservation decisions.

ACKNOWLEDGEMENTS
I wish to thank the organizers and sponsors of this Conference for the kind invita-
tion and financial support for my contribution. Additional financial support was
provided by the Swedish Council for Forestry and Agricultural Research (SJFR)
and the Kempe Foundation. Thanks are also due to Dr Xiao-Ru Wang and Meng-
Zhu Lu for stimulating discussions and suggestions.

REFERENCES
Bassam, B.J., G. Caetano-Anolles and P.M. Gresshoff. 1992. DNA amplification
fingerprinting of bacteria. Applied Microbiology and Biotechnology 38,70-76.
Bawa, K.S. and P.S. Ashton. 1991. Conservation of rare trees in tropical rain forests.
Pages: 62-71 In: D.A. Falk and K.E. Holsinger (eds.) Genetics and Conservation
of Rare Plants, Oxford University Press, New York.

187
Birky, C.W. 1988. Evolution and variation in plant chloroplast and mitochondrial
genomes. Pages: 23-53 In: L. Gottlieb and S. Jain (eds.) Plant Evolutionary Biology.
Chapman and Hall, London.
Birky, C.W., P. Fuerst and T. Maruyama. 1989. Organelle gene diversity under
migration, mutation, and drift: equilibrium expectations, approach to
equilibrium, effects of heteroplasmic cells, and comparison to nuclear genes.
Genetics 121, 613-627.
Caetano-Anolles, G., B.J. Bassam and P.M. Gresshoff. 1991. DNA amplification
fingerprinting: a strategy for genome analysis. Plant Molecular Biology Reporter
9, 294-307.
Clegg, M.T. 1989. Molecular diversity in plant populations. Pages: 98-115 In: A.H.D.
Brown, M.T. Clegg, A.L. Kahler and B.S. Weir (eds.) Plant Population Genetics,
Breeding and Genetic Resources. Sinauer, Sunderland, MA.
Conde, M.F., D.R. Pring and C.S. Levings III. 1979. Maternal inheritance of organelle
DNA’s in Zea mays-Zea perennis reciprocal crosses. Journal of Heredity 70,2-4.
Cotton, R.G.H. 1989. Detection of single base changes in nucleic acids. Biochemical
Journal 263, l-10.
Devey, M.E., K.D. Jermstad, C.G. Tauer and D.B. Neale. 1991. Inheritance of RFLP
loci in a loblolly pine 3-generation pedigree. Theoretical and Applied Genetics
83,238-242.
Dong, J.S. and D.B. Wagner. 1993. Taxonomic and population differentiation of
mitochondrial diversity in Pinus banksiana and Pinus contorta. Theoretical and
Applied Genetics 86, 573-578.
Dong, J. S. and D. B. Wagner. 1994. Paternally inherited chloroplast polymorphism
in Pinus Estimation of diversity and population subdivision, and tests of
disequilibrium with a maternally inherited mitochondrial polymorphism.
Genetics 136, 1187-1194.
Ehrlich, P.R. and G.C. Daily. 1993. Population extinction and saving biodiversity.
Ambio 22, 64-68.
Ellstrand, N.C. 1992. Gene flow by pollen implications for plant conservation
genetics. Oikos 63, 77-86.
Ennos, R. 1994. Estimating the relative rates of pollen and seed migration among
plant populations. Heredity 72, 250-259.
Eriksson, G., G. Namkoong and J.H. Roberds 1993. Dynamic gene conservation for
uncertain futures. Forest Ecology and Management 62, 15-37.

188
Faith, D.P. 1994. Genetic diversity and taxonomic priorities for conservation.
Biological Conservation 68,69-74.
Gershon, D. 1992. If biological diversity has a price, who sets it and who should
benefit. Nature 359,565.
Gillet, M.E. 1991. Genetic analysis of nuclear DNA restriction fragment patterns.
Genome 34,693-703.
Gorman, S.W., R.D. Teasdale and C.A. Cullis. 1992. Structure and organization of
the 5S rRNA genes (5S-DNA) in Pinus radiata (Pinaceae). Plant Systematics and
Evolution 183,223-234.
Govindaraju, D.R. and C.A. Cullis. 1992. Ribosomal DNA variation among
populations of a Pinus rigida Mill (pitch pine) ecosystem. 1. Distribution of copy
numbers. Heredity 69,133-140.
Gregorius, H-R. 1991. Gene conservation and the preservation of adaptability.
Birkhauser Verlag, Basel, 31-47.
Hamrick, J.L. and J.W. Godt. 1991. Allozyme diversity in plant species. Pages: 43-
63 In: A.H.D. Brown, M.T. Clegg, A.L. Kahler and B.S. Weir (eds.) Plant Population
Genetics, Breeding and Genetic Resources, Sinauer, Sunderland, Mass.
Hattemer, H.H. 1991. Measuring genetic variation. Pages: 2-19 In: G. MüllerStarck
and M. Ziehe (eds.) Genetic Variation in European Populations of Forest Trees.
Sauerlander’s Verlag, Frankfurt am Main.
Hedrick, P.W. and P.S. Miller. 1992. Conservation genetics - techniques and
fundamentals. Ecological Applications 2,30-46.
Hong, Y.P., V.D. Hipkins and S.H. Strauss. 1993. Chloroplast DNA diversity among
trees, populations and species in the California closed-cone pines (Pinus radia ta,
Pinus muricata and Pinus attenuata). Genetics 135,1187-1196.
Karvonen, P., A.E. Szmidt and 0. Savolainen. 1994. ITS2 length variation in the
ribosomal DNA of Picea abies and related species. Theoretical and Applied
Genetics 89,969-974.
Kresovich, S. and J.R. McFerson. 1992. Assessment and management of plant genetic
diversity - Considerations of intraspecific and interspecific variation. Field Crops
Research 29,185-204.
Lacy, R. 1988. A report on population genetics in conservation. Conservation
Biology 2,245-247.
Lande, R. 1988. Genetics and demography in biological conservation. Science 241,
1455-1460.

189
Lovett, J. 1994. Biodiversity - Not just a buzzword. Search 25,146.
Lu, M-Z., A.E. Szmidt and X-R. Wang (in press). Inheritance of RAPD fragments in
haploid and diploid tissues of Pinus sylvestris (L.). Heredity.
Machlis, G.E. 1992. The contribution of sociology to biodiversity research and
management. Biological Conservation 62,161-170.
McPherson, M.J., R.J. Oliver and S.J. Gurr. 1991. PCR A practical approach. Pages:
123-145 In: D. Rickwood and B.D. Hames (eds.) The Practical Approach Series,
IRL Press at Oxford University Press, Oxford, New York, Tokyo.
Moran, G.F., D. Smith, J.C. Bell and R. Appels. 1992. The SS-RNA genes in Pinus
radiata and the spacer region as a probe for relationships between Pinus species.
Plant Systematics and Evolution 183,209-221.
Mullis, K., F. Faloona, S. Scharf, R. Saiki, G. Horn, and H. Erlich. 1986. Specific
enzymatic amplification of DNA in vitro: the polymerase chain reaction. Pages:
263-273 In: Cold Spring Harbor Symposium on Quantitative Biology.
Mullis, K.B. and EA. Faloona. 1987. Specific synthesis of DNA in vitro via a
polymerase-catalyzed chain reaction. Methods in Enzymology 155,335-350.
Namkoong, G. 1992. Biodiversity - Issues in genetics, forestry and ethics. Forestry
Chronicle 68,438-443.
Palmer, J. D. 1992. Mitochondrial DNA in plant systematics: applications and
limitations. Pages: 36-49 In: P.S. Soltis, D.E. Soltis and J.J. Doyle (eds.) Molecular
Systematics of Plants, Chapman and Hall, New York London.
Paran, I. and R.W. Michelmore. 1993. Development of reliable PCR-based markers
linked to downy mildew resistance genes in lettuce. Theoretical and Applied
Genetics 85,985-993.
Perrings, C., C. Folke and K.G. Maler. 1992. The ecology and economics of
biodiversity loss - The research agenda. Ambio 21,20l-211.
Platnick, N.I. 1991. Patterns of biodiversity - Tropical vs temperate. Journal of
Natural History 25,1083-1088.
Prendergast, J.R., R.M. Quinn, J.H. Lawton, B.C. Eversham and D.W. Gibbons. 1993.
Rare species, the coincidence of diversity hotspots and conservation strategies.
Nature 365,335-337.
Renner, S.S. and R.E. Ricklefs. 1994. Systematics and biodiversity. Trends in Ecology
and Evolution 9,78.
Roy A., N. Frascaria, J. Mackay and J. Bousquet. 1992. Segregating random amplified
polymorphic DNAs (RAPDs) in Betula alleghaniensis. Theoretical and Applied
Genetics 85,173-180.

190
Saiki, R.K., D.H. Gelfand, S. Stoffel, S.J. Scharf, R. Higuchi, G.T. Horn, K.B. Mullis
and H.A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a
thermostable DNA polymerase. Science 239,487-491.
Savolainen, 0. and K. Kärkkäinen 1992. Effect of forest management on gene pools.
New Forests 6, 329-345.
Sigurgeirsson, A. and A.E. Szmidt. 1993a. Chloroplast DNA diversity, phylogenetics,
and hybridization in Picea. Pages: 369-382 In: J. Alden, J. L. Mastrantonio and S.
Odum (eds.) Forest Development in Cold Climates NATO ASI Series Plenum
Press, New York and London.
Sigurgeirsson, A. and A.E. Szmidt. 1993b. Phylogenetic and biogeographic
implications of chloroplast DNA variation in Picea. Nordic Journal of Botany
13,233-246.
Strauss, S.H. and C-H. Tsai 1988. Ribosomal gene number variability in Douglas-
fir. Journal of Heredity 79, 453-458,
Strauss, S. H. and A.H. Doerksen. 1990. Restriction fragment analysis of pine
phylogeny. Evolution 44, 1081-1096.
Strauss, S.H., A.H. Doerksen and J.R. Byrne. 1990. Evolutionary relationships of
Douglasfir and its relatives (genus Pseudotsuga) from DNA restriction fragment
analysis. Canadian Journal of Botany 68, 1502-1510.
Strauss, S.H., Y.P. Hong and V.D. Hipkins. 1993. High levels of population
differentiation for mitochondrial DNA haplotypes in Pinus radiata, muricata, and
attenuata. Theoretical and Applied Genetics 86, 605-611.
Szmidt, A.E., T. Alden and J-E. Hallgren. 1987. Paternal inheritance of chloroplast
DNA in Larix. Plant Molecular Biology 9, 59-64.
Szmidt, A. E., Y.A. El-Kassaby, A. Sigurgeirsson, T. Alden, D. Lindgren and J-E.
Hallgren. 1988a. Classifying seedlots of Picea sitchensis and P. glauca in zones of
introgression using restriction analysis of chloroplast DNA. Theoretical and
Applied Genetics 76,841-845.
Szmidt, A.E., A. Sigurgeirsson, X-R. Wang, J-E. Hallgren and D. Lindgren. 1988b.
Genetic relationships among Pinus species based on chloroplast DNA
polymorphism. Pages: 33-47 In: J-E. Hallgren (ed.) Molecular Genetics of Forest
Trees, Swedish University of Agricultural Sciences, Umea.
Szmidt, A. E. 1991. Phylogenetic and applied studies on chloroplast genome in
forest conifers. Pages: 185-196 In: S. Fineschi M.E. Malvolti, F. Cannata and H.H.
Hattemer (eds.) Biochemical Markers in the Population Genetics of Forest Trees.
SPB Academic Publishing bv, The Hague, The Netherlands.

191
Szmidt, A.E. and X-R. Wang. 1992. DNA markers in forest genetics. Pages: 79-94 In:
G. MüllerStarck and M. Ziehe (eds.) Genetic Variation in Forest Tree Populations
in Europe, Sauerlander’s Verlag, Frankfurt am Main.
Szmidt, A.E., S. Changtragoon and X-R. Wang. 1993. Contrasting patterns of genetic
diversity in two tropical pines: Pinus khasya (Royle) and P. merkusii (Jungh et de
Vriese). In: Proc. Fourth Congress of European Society for Evolutionary Biology,
Evolution 93, August 22-28, Montpellier, France, University of Montpellier II.
Szmidt, A.E. and X-R. Wang. 1993. Molecular systematics and genetic differentiation
of Pinus sylvestris (L.) and densiflora (Sieb. et Zucc.). Theoretical and Applied
Genetics 86,159-165.
Taberlet, P., L. Gielly, G. Pautou and J. Bouvet. 1991. Universal primers for
amplification of three noncoding regions of chloroplast DNA. Plant Molecular
Biology 17, 1105-1109.
Tripp, H., H. Reilander and A. Wild. 1993. A rapid method for amplification of
plastome DNA-fragments from Spinacia oleracea by PCR. Journal of Plant
Physiology 142, 115-116.
Wagner, D.B., G.R. Furnier, M.A. Saghai-Maroof, S.M. Williams, B.P. Dancik and
R.W. Allard. 1987. Chloroplast DNA polymorphisms in lodgepole and jack pines
and their hybrids. Proceedings of the National Academy of Sciences USA 84,
2097-2100.
Wagner, D.B., Z-X. Sun, D.R. Govindaraju and B.P. Dancik. 1991. Spatial patterns of
chloroplast DNA and cone morphology variation within populations of a Pinus
banksiana-Pinus conforfa sympatric region. The American Naturalist 138, 156-
170.
Wang, X-R., A.E. Szmidt, A. Lewandowski and Z.R. Wang. 1990. Evolutionary
analysis of Pinus densata (Masters), a putative Tertiary hybrid. 1. Allozyme
variation. Theoretical and Applied Genetics 80,635-640.
Wang, X-R. and A.E. Szmidt. 1990. Evolutionary analysis of Pinus densafa (Masters),
a putative Tertiary hybrid. 2. A study using species-specific chloroplast DNA
markers. Theoretical and Applied Genetics 80, 641-647.
Wang, X-R. and A.E. Szmidt. 1993. Chloroplast DNA-based phylogeny of Asian
Pinus species (Pinaceae). Plant Systematics and Evolution 188, 197-211.
Wang, X-R. and A.E. Szmidt. 1994. Hybridization and chloroplast DNA variation
in a Pinus species complex from Asia. Evolution 48, 1020-1031.
Weeden, N.F. and J.F. Wendel. 1989. Genetics of plant isozymes. Pages: 46-72 In:
D.E. Soltis and P.S. Soltis (eds.). Isozymes in Plant Biology. Chapman and Hall,

192
 London.
Wendel, J.F. and N.F. Weeden. 1989. Visualization and interpretation of plant
isozymes. Pages: 5-45 In: D.E. Soltis and P.S. Soltis (eds.) Isozymes in Plant
Biology. Chapman and Hall, London.
Whittemore, A.T. and B.A. Schaal. 1991. Interspecific gene flow in oaks. Proceedings
of the National Academy of Sciences USA88,2540-2544.
Williams, J.G.K., A.R. Kubelik, K. J. Livak, J.A. Rafalski and S.V. Tingey. 1990. DNA
polymorphisms amplified by arbitrary primers are useful as genetic markers.
Nucleic Acids Research l8, 6531-6535.
Williams, J.T. 1991. International aspects of biodiversity. Forestry Chronicle 68,
454-458.
Woodruff, D.S. and G.A.E. Gall. 1992. Genetics and conservation. Agriculture
Ecosystems and Environment 42,53-73.

193
 Chapter

12

GENETIC DIVERSITY OF NORWAY SPRUCE

(PICEA ABIES KARST.) POPULATION FROM
BIALOVIEZA PRIMEVAL FOREST (POLAND)
REVEALED BY ENZYME ELECTROPHORESIS
M. Krzakowal and A.F. Korczyk2

INTRODUCTION
Forest trees are long-living organisms subject to short- and long-term fluctuations
of the environment. Due to their typically outcrossing mating systems, they are
generally found to retain amounts of genetic variation which are roughly twice as
great as any other plant group (Bergmann 1991). Most of the genetic variation in
tree species is located within populations, and only a minor proportion is disti-
buted among populations (El-Kassaby 1991). The threat of destruction of natural
forest populations and attempts to conserve natural populations as well as a board
genetic base, both for the current and future genetic selection, are now a major field
of study (Krugman 1984).
The Norway spruce from Bialovieza Forest, the most undisturbed forest on the
lowland of Central Europe, is known for its late growth initiation and fast growing
characters. Late growth initiation is genetically determined, and consequently,

1 Department of Genetics, Adam Mickiewicz University, ul. Miedzychodzka 5,60-371 Poznan, Poland.
2 Forest Research Institute, 17-230 Bialovieza, Poland.

195
it results in the Norway spruce from Poland performing above average in many
provenances tests and being well-adapted for growth in Northern regions of Sweden,
Norway and North America.
Since the forest resources in Central Europe are highly threatened by air-
pollution or other manmade environmental changes, there is a particular need to
characterize the genetic variation within the natural populations in order to main-
tain genetic diversity.

MATERIAL AND METHODS

Genotypes of 131 trees aged between 35 - 250 years, and distributed over an area of
30 ha, were identified using seeds. Ten megametophytes per tree were examined
using seven enzyme systems. The enzyme systems are as follows: formate
dehydrogenase (FDH - EC 1.2.1.2), NADH - dependent dehydrogenase (NDH),
malate dehydrogenase (MDH - EC 1.1.1.37), glutamate dehydrogenase (GDH -
EC 1.4.1.2), shikimate dehydrogenase (SHDH - EC 1.1.1.25), leucine
aminopeptidase (LAP - EC 3.4.1.1), and glutamicoxalacetic transaminase (GOT -
EC 2.6.1.1.).
Electrophoresis was conducted in 12% starch gels in a lithium-boric buffer
system pH 8.1 (Scandalios 1969) for GOT and LAP, and in a triscitrate (pH 7)
buffer system (Muona et al. 1987) for other enzymes. The staining techniques were
taken from Cheliak and Pitel (l984). Genetic variation was examined for polymorphic
loci, using various measures of variation expected heterozygosity (He), observed
heterozygosity (Ho), Wright’s fixation index (F) and degree of polymorphism (Pg).
Principal Component Analysis, which is usually used for illustrating the variabi-
lity of quantitative characters, was adapted for discontinuous traits, and the results
are depicted by scatter diagrams, showing groups of genotypes for each enzyme
system. All trees were also compared according to all enzyme systems in a
dendrogram, constructed on the basis of Euclidean distances.

RESULTS AND DISCUSSION

Formate dehydrogenase (FDH). This enzyme system has been utilized for popu-
lation investigations of Norway spruce only quite recently (Bergmann and Reutz
1991). The enzyme is coded by one locus with two alleles in Norway spruce and
four variants in Douglas-fir (Lewandowski and Mejnartowicz 1992). The genotype
distribution (Figure 1) along two Principal Component axes (PC axes) shows the

196
ACKNOWLEDGEMENT
This work was supported by a grant from The World Bank Programme GEF 05/
21685 POL.

REFERENCES
Altuchow, J.P., K.B. Krutowskij, H.I. Gafarow, B.A. Duchariew and G.P. Morozow.
1986. Allozymnyj polimorfizm w prirodnoj populacji eli ewropejskoj (Picea
abies Karst.). Genetika XXII, 8:2135 - 2151. .
Bergmann, F. 1991. Causes and consequences of species - specific genetic variation
patterns in European forest tree species: examples with Norway spruce and
Silver fir. In: G.Müller-Starck and M. Zieche (eds.). Genetic variation in European
populations of forest trees. Sauerlander Verlag, Frankfurt am Main.
Bergmann, F. and W. Ruetz. 1991. Isozyme genetic variation and heterozygosity in
random tree samples and selected orchard clones from the same Norway spruce
populations. Forest Ecology and Management 46:39-47.
Bergmann, F. and F. Scholz. 1989. Selection effects of air pollution in Norway spruce
(Picea abies) populations. In: F. Scholz, H.R. Gregorius and D. Rudin (eds.).
Genetic effects of air pollutants in forest tree populations. Springer-Verlag, Berlin,
Heidelberg.
Cheliak, W.M. and J.A. Pitel. 1984. Techniques for starch gel electrophoresis on
enzymes from forest tree species. Can. Forest Serv. Inf. Rep. PI-X-42.
El-Kassaby, Y.A. 1991. Genetic variation within and among conifer populations:
Review and evolution of methods. In: S. Fineschi, M.E. Malvolti, F. Cannata and
H.H. Hattemer (eds.). Biochemical Markers in the Population Genetics of Forest
Trees. SPB Academic Publishing by Hague.
Ernst, S.G., D.E. Keathley and J.W. Hanover. 1987. Inheritance of isozymes in seed
and bud tissue of blue and Engelmann spruce. Genome 29:239-246.
Geburek, Th. and G. von Wuehlisch. 1989. Linkage analysis of isozyme gene loci in
Picea abies Karst. Heredity 62:185-191.
Harry, D.E. 1983. Identification of locus modifying the electrophoretic mobility of
malate dehydrogenase isozymes in Incense-cedar (Calocedrus decurrens) and its
implications for population studies. Biochemical Genetics 21:417-434.
Krugman, S.W. 1984. Policies, strategies and means for genetic conservation in
forestry. In: Plant genetic resources. A conservation imperative. C.W. Yeatman,

204
 D. Kafton and G. Wilkes (eds.). AAAS Selected Symposium 87:71-78.
Lagercrantz, U., N. Ryman and G. Stahl. 1988. Protein loci in diploid tissue of Norway
spruce (Picea abies K.): description and interpretation of electrophoretic
variability patterns. Hereditas 108:149-158.
Lewandowski, A. and W. Mejnatrowicz. 1992. Inheritance of formate dehydrogenase
in Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco). Genetica Polonica 33(4):295-
300.
Lundkwist, K. 1979. Allozyme frequency distribution in four Swedish populations
of Norway spruce (Picea abies K.). Hereditas 90:127-143.
Lundkwist, K. and D. Rudin. 1977. Genetic variation in eleven populations of Picea
abies as determined by isozyme analysis. Hereditas 8567-74.
Muona, O., R. Yazdani and G. Lundqvist. 1987. Analysis of linkage in Picea abies.
Hereditas 106:31-36.
Miiller-Starck, G., Ph. Baradat and F. Bergmann. 1992. Genetic variation within
European trees species. New Forests 6:23-47.
Papageorgiou, A.C., F. Bergmann, E. Gillet and H.H. Hattemer. 1993. Genetic analysis
of isoenzyme variation in Mediterranean Cypress (Cupressus sempemirens L.).
Silvae Genetica 42,2-3: 109-111.
Poulsen, H.D., V. Simonsen and H. Wellendorf. 1983. The inheritance of six isozymes
in Norway spruce Picea abies Karst. Forest Tree Improvement 16:13-33.
Scandalios, J. G. 1969. Genetic control of multiple forms of enzymes: a compilation
of recipies. Biochem. Genet. 4:297-320.
Yeh, EC. and Y.A. El-Kassaby. 1980. Enzyme variation in natural populations of
Sitka spruce (Picea sitchensis). 1. Genetic variation patterns among-trees from 10
IUFRO provenances. Can. J. For. Res. 10:415-422.

205
 Chapter

13
A PRACTICAL APPROACH
TO CONSERVATION OF GENETIC DIVERSIT Y
IN MALAYSIA: GENETIC RESOURCE ARE A

Lim Meng Tsail and Chin Tuck Yuan2

INTRODUCTION
The Permanent Forest Estate (PFE) in Peninsular Malaysia is classified into
Protection, Production, Amenity, Research and Education Forests. Protection forests
are forests which are not exploited for timber, but are maintained in their natural
state to protect the hilly areas and watersheds, and to conserve their genetic
resources. Production forests are forests that are managed for timber production
and they are logged under cutting cycles of between 30 years and 55 years.
Amenity forests are those that are set aside and used for recreation, eco-tourism
and public awareness in forestry while Research and Education forests are set aside
for the conduct of research, education and the conservation of biodiversity and
wildlife.

1 Faculty of Forestry, Universiti Pertanian Malaysia, 43400 UPM Serdang, Selangor, Malaysia.
2 Forestry Department HQ, Jalan Sultan Salahuddin, 50660 Kuala Lumpur, Malaysia.

207
 Prior to 1979, almost all production forests (Lowland Dipterocarp Forests) were
managed under the Malayan Uniform System (MUS) while currently, the system
of management is either the MUS or the Selective Management System (SMS)
which is more suitable for the Hill Forests. Under the MUS, the cutting limit was
fixed. All trees with dbh over 45 cm were removed in 55 year cutting cycles. Under
the SMS, the minimum cutting limit prescribed for dipterocarps is 50 cm dbh and
that for the non-dipterocarps 45 cm dbh, the residual stocking should have at least
32 sound commercial trees in the diameter class 30cm-45cm or its equivalent per
hectare, and the percentage of dipterocarp species in the residual stand, for trees
with 30 cm dbh and above, should not be less than in the original stand.
Under the SMS, an area earmarked for harvesting is first surveyed by a pre-
felling inventory (Pre-F) conducted at 10% sampling intensity. The results are then
used to determine appropriate cutting limits based on stocking, as well as to en-
sure that the conditions mentioned above are fulfilled. The trees are then marked
for felling based on the limits set by the department. The high cutting limit and the
retention of a minimum number of stems is to ensure that residual trees can reach
maturity (over 45 to 50 cm dbh) in 25 to 30 years.Where necessary, such as when
the logged-over forests are in a poor condition, the areas are treated to enhance
their regrowth.
As the SMS and the MUS are basically systems that selectively remove the
biggest trees of selected species, there is a strong possibility of genetic erosion
and degradation unless adequate adult trees and regeneration are assured. As
such, there is an urgent need to conserve the genetic diversity through genetic
management of representative populations of commercially exploited species. Also,
the conservation of species and genetic variation is not specifically stated or
addressed in SMS, although the conservation of genetic resources of the forest
should be an inherent and dynamic component of any forest management and
utilization plan (Boyle et al. 1993, Namkoong 1984, Salleh and Manokaran 1995,
Yeatman 1987,1992).
Effective conservation of genetic resources implies that the selected
representative populations can be maintained and regenerated from generation to
generation (Yeatman 1987). Populations of indigenous tree species can be effectively
maintained in-situ by natural regeneration or assisted by enrichment planting. Ex-
situ gene resource plantations may be considered or required if in-situ manage-
ment of a given species is not possible. The management of such genetic resources
both in-situ and ex-situ is a new field and much research (such as on the actual
genetic diversity and reproductive patterns) needs to be done in order to success-

208
fully maintain the selected populations.
A Genetic Resource Area (GRA) is one such strategy and is intended to con-
serve, manage, and utilize the gene pools of selected species that may be at genetic
risk . It maintains, in perpetuity, the genetic diversity and integrity of sample
populations of the selected species within the GRA (Yeatman 1992).

GRA
A GRA is a protocol for in-situ conservation of the genetic resources and diversity
of selected species within the managed forests of Peninsular Malaysia and elsewhere.
The objective of the management of the selected species is to maintain sufficient
numbers of that species in order that it can maintain its genetic diversity within
the population or subpopulations in the designated forest stands and ecosystems.
Specifically, the objectives are to maintain and perpetuate genetic diversity of
economically important species in forests that are harvested; to utilize such selected
genetic reservoirs for breeding and planting, research, education and possibly even
plantation purposes; to reduce and halt genetic erosion and minimize the risk of
genetic losses and maximize future options by maintaining a sound genetic base;
to integrate sound genetic management into forest operations; to establish an
operating scale of GRA and to demonstrate the feasibility, costs and benefits of
systematic conservation of forest genetic resources.
The selected, or target, species are species that are considered to be critically
important for the designated forest stands and ecosystems, and therefore worthy
of being selected for special consideration, for a number of reasons. They may be
commercially important species and in great demand, so that they are selectively
chosen for harvesting; they may be species that are potentially important but found
in such low numbers that their future may be threatened; they may be of high value
and rare species or slow growing ones that may be threatened by any removal; or
species that have low regeneration potential. Hence, there can be many species
which fit into some or all of the above criteria and thus are considered to be wor-
thy of being maintained in sufficient numbers to conserve their genetic diversity
and retain their capacity for regeneration.
In addition to the conservation of genetic diversity of the target species, GRA’s
can also serve as a potential source of certified seeds and plants. These sources of
known populations can be used for large-scale production and collection of seeds
and/or seedlings for reforestation purposes and as potential genetic resources for
further selection and improvement. By maintaining records of target species in

209
different areas, a reservoir of genetically diverse populations can be maintained
for research and further investigations into the genetic make-up of the species as
well as the breeding systems of the target species.

RESULTS
The GRA project reported here is the first such project in Malaysia. It is a joint
project between the Forest Department Peninsular Malaysia (FDPM), Johor State
Forestry Department (JSFD) and the ASEAN Forest Tree Seed Centre (AFTSC). The
project was initiated in November 1992 with the AFTSC providing the technical
expertise and training required to undertake the project. The JSFD allocated an area
of 5517 ha within the Ulu Sedili Forest Reserve as a project site. It comprises 30
compartments, 19 of which have been logged while 11 are not logged but sche-
duled to be logged in the future. FDPM provided the expertise and overall
management of the project, as well as data analyses with the assistance of two
advisers/collaborators.
The Ulu Sedili Forest Reserve Genetic Resource Area is situated about 40 km
north of Johor Bahru. The site is easily accessible by the Kota Tinggi-Mersing road
and a forest road. The Ulu Sedili Forest Reserve was gazetted as a Reserve in Nov.
8, 1951, and lies around latitude 1o55’N and longitude 103o45’E. The GRA has an
undulating topography with altitudes ranging between 50-750 m above sea level.
The average elevation is about 300m above sea level. The mean annual rainfall
recorded at the nearest rainfall station averages 2,480 mm.
The forest can be described as a Lowland Dipterocarp Forest. The upper or
emergent storey is characterized by a dominance of the trees from the family
Dipterocarpaceae, including species belonging to the genera Dipterocarpus, Shorea,
Dryobalanops, Neobalanocarpus, Hopea, and Anisoptera . Other common large trees in
the emergent and main canopy level include Dyera costulata, Gluta spp., Intsia
palembanica, Koompassia malaccensis, Melanorrhoea spp., Palaquium spp., Sindora spp.
and Heritiera spp. Other important (timber) species include trees from the families
of Burseraceae, Guttiferae, Myristicaceae, Myrtaceae, Sapotaceae, Annonaceae,
Euphorbiaceae, Flacourtiaceae and Rubiaceae.
The high diversity of the forest and the high commercial value of the various
timber species in the proposed GRA made the selection of target species some-
what difficult. Log production data from the compartments that were logged
previously, and recommendations of the JSFD officers in the field, were used. The
eight target species were identified initially on the basis of priority rating (Table 1).

210
 In terms of absolute numbers, there are about 108 trees per ha over 15 cm dbh
with Dipterocarps accounting for about 23 % of this number (equivalent to 24
trees). In contrast, there is an average of 2.5 trees per ha over 90 cm dbh, with
dipterocarps accounting for over 90% of the number (equivalent to 2.3 trees). The
mean numbers of trees of each target species on a per ha basis above 15 cm dbh
ranged from 0.1 (D. costulata) to 5.15 (D. aromatica) (Table 2). Several of the target
species have trees in the 90 cm and above dbh class they were S. singkawang, S.curtisii,
D. aromatica, N. heimii and D. costulata . There were a number of non-target species
with trees over 90 cm dbh. These include S. leprosula, S. palembanica and S. acuminata
as well as Scaphium sp. (family: Sterculiaceae).
The numbers of the target species were tabulated according to dbh classes and
the cumulative numbers above the respective size classes calculated (Table 2).
The total number of the target species within the compartment varied considera-
bly; D. aromatica occurred in high numbers, with over 1200 trees above 30 cm in
dbh, while D. costulata was found in very low numbers, with less than 12 trees
above 30 cm dbh in the whole compartment (of 357 ha) Assuming that 40-50 % of
a (sub-)population should be preserved for genetic diversity (Namkoong 1984), it
was decided that at least 50% of the trees over 30cm dbh should be left (uncut).
Thus the size class with the cumulative number of trees equal to half (or a little
lower than) the cumulative number at 30 cm was taken as the cutting limit for the
species. The basis is that trees above 30 cm dbh are trees that can breed/ reproduce.
In addition, absolute stocking was considered, such that no cutting was prescribed
for the low incidence species, A. Zaevis and D. costulata. Based on these assump-
tions, the following were considered and recommended.

a) No cutting of Anisoptera sp. and D. costulata.

b) Cutting limit of >60 cm dbh for all other Dipterocarps except Neobalanocarpus
heimii, whose cutting limit is >90 cm dbh.
c) Cutting limit of >55 cm dbh for all other non-dipterocarps except K. malaccensis
whose cutting limit is >60cm dbh.

The trees in the compartment are in the process of being marked for felling
and, following timber harvesting, another inventory (PostF) will be conducted to
determine the status of the residual stand and to determine the types of silvicul-
tural treatments needed.

212
DISCUSSION
The changes to the present system of management is relatively minor. The normal
SMS procedures of conducting a Pre-F, tree marking and then tendering will be
continued. GRA considerations are only applied at the stage between Pre-F and
logging by altering some of the cutting limits of species in contrast to cutting limits
based on commercial groups. Hence there is greater selection at the species level
rather than at the commercial group level.
An obvious advantage of the GRA approach is that it blends in with the current
management practices and it would result in minimum change to the harvesting
operations. Only the cutting limits are modified to ensure the maintenance of gene-
tic diversity following the harvesting operations. It was recognized that the
harvesting would have to be economical (that is with an out turn of at least 28 cubic
m per ha), otherwise there may be problems finding contractors to perform the
harvesting operation. The volume of timber available for harvest with the above
conditions was as estimated and projected to be still sufficiently economical for
the area to be tendered out normally.
As with any new system that is being introduced, it is not perfect. We realize
that there are limitations which should be remedied by further research and also
by follow-up inventories of the areas after logging. The choice of the 8 target
species selected in this report may seem limited. It is possible that additional, or
even different species may be targeted in other compartments of the GRA. How-
ever, if the numbers are increased, the conditions imposed on the loggers would
also increase and this may create problems in monitoring by the forestry field staff.
The retention of 50% of trees of the target species over 30 cm dbh is also recognized
as an important or critical assumption but we think we have to start at a point that
is acceptable to the forest managers and the loggers. A corollary to the retention of
50% is the question of what should constitute a sufficient or appropriate minimum
number of trees to be left behind. This is yet to be determined. The value of 50% of
the total number of trees above 30 cm dbh could be valid for trees that are found in
fairly large numbers but may not be meaningful when it comes to numbers less
than 50 or even 100 in a compartment of over 300 ha. While this minimum number
still needs to be investigated, the numbers of stems of target species that will be
left in the first GRA compartment after logging at the abovementioned cutting
conditions, will generally be over 100, except for S. curtisii.
Even with 100 trees per compartment, the average density is only about 0.3

214
trees per ha and may seem rather low. However, we must remember that this is a
figure for trees above 30 cm in dbh and if the numbers of smaller trees and sap-
lings are included the total number would increase quite substantially When the
numbers of saplings and trees above 15 cm but below 30 cm dbh are included, the
numbers do show a significant increase. When poles and saplings below 15 cm dbh
are also considered the numbers will probably increase even more, thus giving a
retention figure that is considerably greater than 50%.
During the logging process there are likely to be additional losses due to road
construction, felling and skidding. These losses will be estimated through a Post-
Felling inventory, but such losses are expected to be minimized through better
supervision. The additional information from the post-F will also be used for fur-
ther decisions on the silvicultural requirements of the compartment.
Since management of genetic diversity in Peninsular Malaysia is relatively
new, studies on population genetics; flowering and seed phenology, regeneration
ecology, seed storage and germination; ecosystem dynamics of (current and
potential) target species and inventory survey methods applicable for GRA
management need to be conducted.
Complementing the above in-situ study, a preliminary genetic screening of
three target species, namely, S. laevis, D. aromatica and N. heimii. was conducted
jointly with Forest Research Institute Malaysia (FRIM) to determine family esti-
mates of genetic diversity and mating system parameters. Preliminary results
indicate that the three species possess sufficient diversity and vigour (Wickneswari
Ratnam, pers comm.) (Table 3). Genetic screening will continue for the rest of the
target species. .
As a continuation of the study, Post-F inventories will be conducted in the
logged over compartments in order to investigate the status of the forest stands so
that comparison can be made with the Pre-F data and to determine if any treat-
ment is required. As there are 30 compartments in the GRA, there is much
information that needs to be gathered. The compartments that are to be logged will
need to have their target species and cutting limits determined first. Subsequently,
the overall success or failure of the GRA strategy needs to be evaluated. In addi-
tion to these, it is hoped that a network of Genetic Resource Areas in Peninsular
Malaysia can be developed in conjunction with operational forest management
and seed/seedling production; and subsequently a Register of Genetic Resources
for Peninsular Malaysia for reference and use in planning reforestation.

215
Species N Ni P A He F tm Fs Fe
S. laevis 70.5 11 45.5 1.5 0.182 0.327 1.498 0.999 0.199
D aroma tica 45.5 10 67.8 2.5 0.270 0.303 0.926 0.0 0.041
N . heimii 33.5 6 83.3 2.3 0.390 0.209 0.950 0.112 0.026

Table 3: Estimates of genetic diversity and mating system parameters of three target species.
N = mean number of seedlings per tree assayed
Ni = number of isoenzyme loci analysed
P = percentage of polymorphic loci
A = mean number of alleles per locus
He = mean expected heterozygosity
F = mean fixation index per locus
tm = multilocus outcrossing rate
Fs = mean fixation index among progenies
Fe = (1-t)/(1+t), the equilibrium coefficient

CONCLUSION
With the current world-wide concern for the depletion of the tropical rainforest,
this pilot project aims to develop a better understanding of, and so ensure
perpetuation of, the genetic diversity of target species. Knowledge gained from
the GRA Pilot Project should enhance sustained yield management of the forest.
Genetic diversity of the rainforest cannot be neglected and steps need to be taken
to maintain and use these renewable resources.
As with any system of forest management, SMS has its fair share of shortcomings
especially with regard to the response of the residual stands. Cutting limits are
based on the stem sizes mentioned above. However, with the GRA approach, the
cutting limits are also based on size but with due consideration for the numbers of
the selected individual species. Although this has resulted in a generally higher
cutting limit for most of the target species; the other non-target species were not
affected at all. Further refinement of the basis of determining the cutting limit should
be developed. Genetic screening is one way this can be achieved.
It is proposed that this GRA project will continue for at least 10 years, and it is
also hoped the GRA guidelines will be incorporated into the SMS well before then.
The stands will be monitored regularly, probably five (5) years after logging. The
status and demographics of each target species will be reviewed and silvicultural
measures prescribed to ensure their maintenance and enhancement. Broader

216
applications of these principles and methods will benefit Malaysian forestry by
ensuring socio-economic returns from the forests, sustaining irreplaceable genetic
diversity and forest ecosystems, and maintaining environmental protection
associated with the forest.

ACKNOWLEDGEMENTS
We would like to thank the organizers for support, the AFTSC for the initiative
and grant to launch this project, the JSFD for their co-operation and assistance
in implementing the project.

REFERENCES
Boyle, T.J.B., C. Liengsiri and A.G. Young. 1993. Principles for effective in situ
conservation of tropical forest tree species. Pages 128-136 In: Proceedings
International Symposium on Genetic Conservation and Production of Tropical
Forest Tree Seeds. ASEAN Forest Tree Seed Centre, Muak Lek, Thailand.
Namkoong, G. 1984. Strategies for gene conservation in forest tree breeding. Chapter
8. In: C.W. Yeatman, D. Kafton and G.Wilkes (eds.) Plant Genetic Resources - A
conservation imperative. Westview Press Inc. Boulder, Colorado.
Salleh, M.N. and N. Manokaran. 1995. Monitoring of forest biodiversity: Policy and
research issues. Chapter 9, this volume.
Yeatman, C.W. 1987. Conservation of genetic resources within managed natural
and manmade forests. Malaysian Forester 50 (l):l-18.
Yeatman, C.W. 1992. Objectives and guidelines for genetic resources areas in ASEAN.
Poster presentation at the International Symposium on Population Genetics and
Gene Conservation of Forest Trees in August 1992 in Carcans-Mausbuisson,
France.

217
 Chapter

14

GENETIC VARIABILITY OF TWO

RATTAN SPECIES
FROM ISOZYME MARKERS

M.C. Bon1, B. Ali H. Basri1 and, H.I. Joly2

INTRODUCTION
Rattans are spiny climbing plants belonging to the major group “Lepidocaryoid”
of the Palm family. They are represented altogether by 600 species in 13 genera and
distributed from West Africa to Fiji and from Southern China to Queensland. The
greatest diversity of genera and species is in the western part of Malesia, as repor-
ted by Dransfield and Manokaran (1993).
Widely used in the village economies of Asia for centuries, rattans also enjoy an
expanding and successful international market for furniture, basketry, binding,
cordage and a wealth of other purposes. Due to the depletion of the forests and
their natural resources, and as a result of the current rate of exploitation, some
commercial species are endangered. The necessity for an international conserva-
tion programme and for strategies for resource development has been recognized
by producing countries in South-East Asia. In Sabah (Malaysia), national agencies

1 CIRAD-Forêt/Yayasan Sabah, P.0. Box 60793,91017 Tawau, Malaysia.

2 CIRAD-Forêt, 45 Bis, avenue de la Belle Gabrielle, 94736 Nogent sur Mame, France.

219
like Innoprise Corporation Sdn Bhd are participating in this conservation effort
by managing protected areas and setting up field gene banks of native and exotic
rattan species. A total of almost 14’ha of conservation stands has so far been
established in Luasong, as has been presented by Garcia et al. (1993). Ex situ
conservation efforts have started with two of the commercially most important
large diameter cane species, Calamus manan and CaZamus subinermis.
Calamus manan occurs naturally in Peninsular Malaysia and Sumatra, and
Calamus subinermis in Sabah (North Borneo) and the Palawan Islands (the
Philippines) (Dransfield and Manokaran 1993). Both species are dioecious and
produce a large diameter cane. As a solitary species which can only be harvested
once, as it does not produce new shoots following cutting, Calamus manan is
considered more endangered than the clustering species Calamus subinermis, for
which commercial exploitation started more recently
To establish conservation programmes, involving both ex situ and in situ
conservation, as well as to develop breeding strategies for rattan species, it is of
primary importance to establish the variability of each species and the spatial
distribution of its genetic diversity. Along with the establishment of field trials,
CIIWD-For& and Innoprise Corporation are developing tools to evaluate the gene-
tic diversity of rattan species. This paper presents the first results concerning the
evaluation of the genetic diversity of C. manan and C. subinermis, using isozyme
markers.

MATERIALS AND METHODS

Expeditions to collect rattan seeds from already severely depleted natural forests
must compete with both legal and illegal cane harvesters. Added to this problem,
only a small proportion of the dioecious rattan plants bear fruits in any year.
Difficulties are exacerbated by a lack of precise information, within the region, on
the period of fruit maturation and also by uncritical use of vernacular names.
Harvested rattan seeds rapidly lose their germination ability and therefore have
to be sown immediately on return from the collecting expedition. The seedlings
raised in the nursery are finally transferred to the conservation stands in Luasong.
The material already established in the stands is described in Table 1. All
populations were analyzed for the survey of genetic diversity; one plant was sam-
pled per progeny. Leaf tissue was used for isozyme analysis in preference to seed
because it is abundantly available from any individual irrespective of sex and
flowering period. Young leaves from individual rattan plants aged 2 to 5 years were
used.

220
Interpretation of allozymic banding patterns was carried out based on general
information about enzyme structure (Pasteur et al. 1987; Wendel and Weeden 1989)
and the study of phenotype segregation in progenies according to Gillet and
Hattemer (1989), (Bon, in preparation).
Four parameters were used to assess genetic diversity: percentage of
polymorphic loci, average number of alleles per locus, average observed, and
average expected heterozygosity. A locus was considered to be polymorphic if
the frequency of the most common allele did not exceed 0.95. Putative monomor-
phic loci were not included in the computation of genetic measures. The distribution
of genetic diversity within and among populations was evaluated through
Wright’s F statistics (FIS ,FI IT , FSR)(Wright 1965). F-statistics were calculated for
individual alleles in each locus according to Wright (1965). F, can be estimated by
the formula:
l-FIT = (1-FIS)(1-FSR)

These parameters were estimated using the BIOSYS-1 software package

(Swofford and Selander 1981).

RESULTS
Genetic variation within populations
The value of genetic parameters are given for Calamus manan and Calamus subinermis
in Table 2. The average number of alleles per locus (A) for all populations combined
was 2.5 for the two species. The mean proportion of polymorphic loci (P) within the
species was 85% and 77%, for Calamus manan and Calamus subinermis, respectively.
The mean observed heterozygosity (Ho) across loci was 0.37 and 0.42, for C. manan
and C. subinermis, respectively and the mean expected heterozygosity (He) was
0.47 for both species. At the species level, the fixation index, F, was 0.21 and 0.11, for
C. manan and C. subinernzis, respectively. C. manan seems to exhibit a deficit in
heterozygotes while C. subinermis seems to be in Hardy-Weinberg equilibrium.

Genetic differentiation among popda tions

In each species single populations seem to be in Hardy-Weinberg equilibrium (FIS=
0.07 and FIS = 0.05, for C. manan and C. subinermis, respectively). The values of FST
(Table 3) show that there is some genetic differentiation among populations of C.
manan but that populations of C. subinermis from the west coast of Borneo do not
show genetic differentiation.

222
ever, more populations with larger sample sizes must be studied before a complete
picture of the genetic diversity of C. manan and C. subinermis can be derived but,
from a practical point of view, as the differentiation among populations is large
for Calamus manan, one can recommend that a sufficiently large number of
populations be sampled for conservation.

ACKNOWLEDGMENTS
The authors are indebted to David Alloysius from Luasong Forestry Centre under
Innoprise Corporation Sdn Bhd for sample collection, to Marc Chauviere from
CIRAD-Forêt and Awang Mohdar from Innoprise Corporation for cooperation.
The helpful assistance of the Forest Research Institute of Malaysia (FRIM) and
Forest Research Centre of Sandakan (FRC) has also been appreciated. This work
has been founded under an EEC research programme of the DGXI.

REFERENCES
Bon, M.C., F. Riccardi and M. Najar. 1992. Discrimination èlectrophorètique de clones
de Populus x Interamericana. Annales AFOCEL 1991: 25-38.
Dransfield, J. and N. Manokaran. 1993. Plant resources of South-East Asia. No.6
Rattans. Pudoc Scientific Publishers, Wageningen, 137p.
Garcia, C., D. Alloysius, Y. Yusuf, M. Chauviere and M.C. Bon (eds.). 1993. Genetic
conservation of rattans. C.R. ASEAN Seminar - Management and Conservation
of Biodiversity.; Nov. 29- Dec. 1; Kuala Lumpur, Malaysia, A.I.F.M. Kuala
Lumpur, Malaysia, 9p.
Gillet, E. and H.H. Hattemer. 1989. Genetic analysis of isoenzyme phenotypes using
single tree progenies. Heredity 63: 135-141.
Khasa, P.D. 1993. Genetique du Racosperma auriculiforme et du R. mangium. PhD
Thesis of Faculte Forestiére et de Geomatique University of Laval, Quebec.
Liengsiri, C., C. Piewluang and T.J.B. Boyle. 1990. Starch gel electrophoresis of
tropical trees: a manual. ASEAN-Canada Forest Tree Seed Centre. Muak Lek,
Saraburi Thailand.
Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a
small number of individuals. Genetics 89: 583-590.
Pasteur, N., G. Pasteur, F. Bonhomme, J. Catalan and J. Britton-Davidian. 1987.
Manuel technique de génétique par electrophorese des proteines. Technique et

224
 Documentation (Lavoisier). 217p.
Swofford, D.L. and R.B. Selander. 1981. BIOSYS-1: A FORTRAN program for the
comprehensive analysis of electrophoretic data in population genetics and
systematics. J.Hered. 72: 281-283.
Wahlund, S. 1928. Zusammensetzung vom Populationen und
Korrelationserscheinungen von Stanpunkl der Vererbunglehre aus betrachtet.
Hereditas 11: 65-106.
Wendel, J.F. and N.F. Weeden. 1989. Visualization and interpretation of plant
isozymes. Pages: 5-45 In: D.E. Soltis and P.S. Soltis (eds.). Isozymes in Plant
biology. Dioscorides Press, Portland.
Wright, S. 1965. The interpretation of population structure by F-statistic with special
regard to systems of mating. Evolution 19: 395-420.

225
 Chapter

15
GENE DIVERSITY STUD Y
BASED ON ISOZYME ANALYSIS
IN TEAK (TECTONA GRANDIS L.F.)
PROVENANCE S
A.W.S. Kertadikara1,2 and D. Prat1,3

INTRODUCTION
Teak, Tectona grandis L.f., (Verbenaceae) is one of the most economically important
tropical timber species. The tree is native to the tropical deciduous forests, of three
disjunct regions: (1) Central and East India, (2) Myanmar, Laos and north west
Thailand, and (3) Indonesia (Central and East Java). Altona (1922) supposed that
its presence in Indonesia might be due to human introductions.
International provenance trials in the 1970s, supported by FAO, DANIDA and
CIRAD-For& (formerly CTFT), have revealed large genetic differentiation
among populations for quantitative traits (Kaosa-Ard, 1986). The present study
was then focused on gene diversity within and among eight teak stands, both from

1 Laboratoire INRA-ENGREF de Sciences Forestières, Unité de Génétique des Populations d’Arbres

Forestiers, ENGREF, 14 rue Girardet, F-54042 Nancy Cedex, France.
2 present address: Jurusan Manajemen Hutan, Fakultas Kehutanan, IPB, POB 168, Bogor 16001,
Indonesia.
3 present address: Station d’Amélioration des Arbres Forestiers, INRA, F-45160 Ardon, France.

227
Isozyme techniques
Enzymes were extracted from young leaf parenchyma of 2 month-old plants.
Samples were ground in an extraction buffer (sodium tetraborate buffer 50 M, pH
8.3). Enzymes were separated by electrophoresis at 5 ±1 C, under an electric field
of 12 V.cm -1 for 4.5 to 5 h, in polyacrylamide vertical gel (stacking gel: 3%; running
gel: 9%). The electrode and gel buffer was Tris (90 mM), borate (10 mM), EDTA
disodium salt (0.05mM) at pH 8.38 (Kertadikara, 1992).
Out of 69 tested enzyme systems, the following 14 systems were recorded:
Alanine aminopeptidase (AAP, E.C. 3.4.11.1), Aspartate amino transferase
(AAT, E.C. 2.6.1.1), Acidic phosphatase (ACP, E.C. 3.1.3.2), Alcohol dehydrogenase
(ADH, E.C. l.l.l.l), Diaphorase (DIA, E.C. l.6.4.3), Endopeptidase (ENDO, E.C.
3.4.-.-), Carboxyl esterase (EST, E.C. 3.1.1.-), Fluorescent ß-D-glucosidase (B-GLU,
E.C. 3.2.1.21), Glycerate-2 dehydrogenase (G2DH, E.C. 1.1.1.29), Leucine
aminopeptidase (LAP, E.C. 3.4.11.1), Lactate dehydrogenase (LDH, E.C. 1.1.1.27),
Nicotinamine adenine dinucleotide dehydrogenase (NADHDH, E.C. 1.6.99.3),
Peroxidase (PER, E.C 1.11.1.7) and Superoxide dismutase (SOD, E.C. 1.15.1.1).
Genetic interpretations of isozyme patterns were assessed from the observed
segregations in ten half-sib progenies of Indian provenance I, (Kertadikara and
Prat, in preparation). Genotypes of all others individuals were then deduced
accordingly.
Within-population genetic variation was estimated by considering allele
frequency, average number of alleles per locus, proportion of polymorphic loci (at
criterion 95 %), observed and expected heterozygosity and fixation index. Mating
system parameters (multilocus outcrossing rate) were assessed according to
Ritland and Jain (1981) and Ritland (1983) for provenance I3. Each open-pollinated
family consisted of at least 20 seedlings.
The genetic structure among individual populations was estimated by using
the level of differentiation among demes 6 (Gregorius 1985, Gregorius and
Roberds 1986). Relationships among populations were analyzed according to Nei’s
(1978) and Gregorius’ (1984) genetic distances, and represented by a dendrogram
based on the unweighted pair-group method with arithmetic averaging (UPGMA)
described by Sneath and Sokal (l973).

RESULTS
Eighteen polymorphic isozyme loci (Aap, Aat-b, Aat-c, Acp-b, Adh, Dia-a, Da-b, Endo,

229
Est, ß-Glu, G,dh-a, Lap, Ldh, Nadh-dh, Per, Sod-a, Sod-b and Sod-c) encoding fourteen
enzyme systems were scored and considered for the analysis of genetic variation.
Two monomorphic loci (Aat- a and Sod-d )were also noticed.

Genetic diversity within provenances

Over 20 loci were recorded (including two monomorphic loci), with 82 alleles being
scored. On the average, over all unweighted populations, the number of alleles per
locus was 3.0, the percentage of polymorphic loci was 88 %, the observed
heterozygosity was 0.289, while the expected value after Hardy-Weinberg
equilibrium (gene diversity) was 0.362 (Table 2). The values were, however,
dependent on the sample size of population (9 to 263) which resulted from varia-
tions in germination rates.
The Indian provenance I1 showed the highest values for both observed and
expected heterozygosity. Average fixation indices were positive in each prove-
nance, confirming a trend of heterozygote deficiency. The x2 goodness of fit test
indicated a significant deviation towards an excess of homozygosity in most
provenances at Aat-b, Acp-b, Dia-a, Dia-b, and Sod- b loci. Negative values of fixation
index (mostly not significant) were, however, found in some loci in all prove-
nances.
Ten polymorphic loci were used in order to determine the outcrossing rate in
the Indian population I3. The multilocus outcrossing rate was 0.98.

Genetic diversity among provenances

Allelic frequencies were significantly different among provenances over all loci
(x2= 58.0***). Some major alleles, particularly DIA-b2 and DIA-b3, EST3 and EST5,
AAT-b2 and AAT-b5, G2DH-a1 and G2DH-a3 and LDH2 and LDH4, allowed the
differentiation of all Indian provenances in one hand and of African and Indo-
nesian provenances in the other hand. The Thai population was quite similar to
the African and Indonesian ones, except for Est and G2DH- a loci, for which it
exhibited intermediate allelic frequencies.
Using the methods of Gregorius and Roberds (1986) differentiation among
provenances ( ) was high for loci Aat-b, Ldh, Est and G2DH-a; and overall, the
differentiation among provenances was about 19 %. The Indian provenances were
the most distinct provenances, and African provenances the least distinct (Figure 1).

230
Dispersal of teak in south-east Asia
The hypothesis of teak introduction in to Java from India as a result of human
activities about 1500 to 700 years ago, as suggested by Altona (1922) cannot be
supported by our results. Such introduction could take place only from the Thai
gene pool. Could introduction have occurred naturally ?
In mainland Asia teak is typically a monsoon forest species, as is also the case
in the Indonesian archipelago, where teak is exclusively found in dry and sea-
sonal climates. According to a hypothetical map of Pleistocene climatic condi-
tions (Heaney, 1991) the seasonal forests covered a larger area than at present,
including the centre of the Sunda shelf, extending in an arc from southern Thailand
to eastern Java (Morley and Flenley, 1987). Glaciation caused a sea level decline,
permitting a connection between the Asian continent and the Indonesian
archipelago. It can be hypothesised that the pattern of teak dispersal was firstly
continuous from its centre of origin, somewhere in Myanmar (Meniaud, 1930), to
Java. The present ecological and geographical conditions led to the current dis-
junct area of teak. The palaeogeographic and palaeoclimatic data thus support the
hypothetical natural migration of teak to eastern Java from Myanmar or Thailand
during the Pleistocene. However, palynological data could not confirm this
hypothesis since teak pollen has never been found in sediments. Human interven-
tion was not absolutely required to explain the present distribution of teak in
Indonesia.

REFERENCES
Altona, T. 1922. Teak and Hindoos. Origin of teak in Bodjonegoro (Java). Tectona
15: 457-507.
Delaunay, J. 1977. Résultats d’essais de provenances de teck, Tectona grandis, six ans
aprés leur mise en place en Cóte d’Ivoire. Pages: 273-284 In: Third world
consultation of forest tree breeders, FAO/IUFRO, Canberra. Vol. I.
Egenti, L.C. 1978. The Danish/FAO international provenance trials of Tectona grandis
in Nigeria. Indian Forest. 104: 227-237.
Gregorius, H.R. 1984. A unique genetic distance. Biom. J. 26: 13-18.
Gregorius, H.R. 1985. Measurements of genetic differentiation in plant populations.
Pages: 276-285 In: Gregorius, H.R., (ed.). Population genetics in forestry. Lecture
notes in biomathematics 60. Heidelberg Berlin: Springer-Verlag.
Gregorius, H.R. and J.H. Roberds, 1986. Measurement of genetical differentiation

234
 among subpopulations. Theor. Appl. Genet. 71: 826-834.
Heaney, L.W. 1991. A synopsis of climatic and vegetational change in southeast
Asia. Climatic Change 19: 53-61.
Hedegart, T. 1973. Pollination of teak (Tectona grandis). 2. Silvae Genet. 22: 124128.
Kaosa-Ard, A. 1986. Teak in ASEAN. A Survey report. Muak Lek: ASEAN/
CANADA Forest Tree Seed Centre.
Kertadikara, .A.W.S. 1992. Variabilite genetique de quelques provenances de teck
(Tectona grandis L.f.) et leur aptitude à la multiplication vegetative. Nancy: Thése
Universite Nancy I.
Kertadikara, A.W.S. and D. Prat. 1995. (in press) Isozyme variation among teak
(Tectona grandis L.f.) provenances. Theor. Appl. Genet.
Madoffe, S.S. and J.A. Maghembe. 1988. Performance of teak (Tectona grandis L.f.)
provenances seventeen years after planting at Longuza, Tanzania. Silvae Genet.,
37: 175-178.
Meniaud, J. 1930. Le teck et sa propagation en Afrique tropicale. C.R. Assoc. Colonies
Sci., Comite Nat. Bois Coloniaux.
Morley, R.J. and J.R. Flenley. 1987. Late cainozoic vegetational and environmental
changes in the Malay archipelago. Pages: 50-59 In: Whitmore, T.C., (ed.)
Biogeographical evolution in the Malay archipelago. Oxford Monogr.
Biogeography (Vol. 4)
Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a
small number of individuals. Genetics 89: 583-590.
Ritland, K. 1983. Estimation of mating systems. Pages: 289-302 In: Tanksley, S.D.;
Orton, T.J., (eds.) Isozymes in plant genetics and breeding. Part A. Elsevier Sci.
Publsh, Amsterdam.
Ritland, K. and S.K. Jain. 1981. A model for the estimation of outcrossing and
genes frequencies using n independent loci. Heredity 47: 35-52.
Sneath, P.H.A. and R.R. Sokal. 1973. Numerical taxonomy. W.H. Freeman, San
Francisco.

235
 Chapter

16
SPECIES DIVERSITY OF SECOND GROWTH
AT NGAO DEMONSTRATION FOREST,
LAMPANG PROVINCE

Somboon Kiratiprayoonl, Jesada Luangjamel,

Pralong Damrongthail and Metinee Tarumatsawasl

INTRODUCTION
The total area of Thai forest areas is decreasing at an alarming rate. Statistics com-
plied over the past three decades (1961-1991) suggest that 0.46 million ha are
deforested annually (RFD 1992). These rapid and profound changes endanger both
the species diversity and genetic resources of Thailand’s forested ecosystems. If the
natural balance of the remaining forest ecosystems is to be maintained and shor-
tages of wood avoided there must be a concerted and timely move towards
conservation and protection of the Thai natural resources. However, before
sustainable management strategies to protect forest biodiversity in Thailand can be
developed the structure and function of the many diverse ecosystems must be bet-
ter understood (Boontawee et al., chapter 8, this volume).
As a point of departure we have chosen three questions relevant to the mea-
suring and monitoring of forest biodiversity in Thailand: 1) Are there different
patterns of diversity between different forest types occurring in Thailand? 2) Are

1 Royal Forest Department, 61 Phaholyothin Road, Bangkok 10900, Thailand.

237
there different patterns of diversity within the forest types? 3) What are the
implications of variable levels of diversity between and within forest types on
sustainable management practices?
The present study is an effort to provide a practical foundation for addressing
these questions. To do this we have established permanent lha plots in each of
three forest types common to northern Thailand: Dry Dipterocarp forest, Mixed
Deciduous forest and Seasonal Dry Evergreen forest.

LOCATION
The study was conducted at the Ngao Demonstration Forest in Ngao District,
Lampang Province. The site area is approximately 1,750 km2 located at 18o 20’ - 19o
05’ N and 99o 45’ - 100° 05’E. For the past ten years the precipitation has averaged
1100 mm/year and the temperature 25. ’ 65oC. The Ngao Demonstration Forest
contains three kinds of forest communities i.e. Dry Dipterocarp forest, Mixed
Deciduous forest and Evergreen forest. All three forest types were studied. This
paper presents the results from the Dry Dipterocarp and Mixed Deciduous forests.

METHODS AND MATERIALS

Site and plot preparation
Study sites were selected based upon three criteria 1) homogeneity of forest types,
2) reasonable access, and 3) protection from future human disturbance. In each
forest type a 100 m x 100 m permanent plot was set up and divided into 100 10 m x
10 m subplots. Smaller temporary subplots of 4 m x 4 m and 1 m x 1 m were
systematically established at a corner of each 10 m x 10 m subplot.

Data collection
From the permanent 100 m x 100 m plot, the following data was recorded:
1. In the 10 m x 10 m subplots all plants with DBH > 4.5 cm were tagged and the
species, DBH, and height were recorded. All plants in this group are designated
as being in the tree group.
2. In the temporary 4m x 4m subplots, all individuals taller than 1.3 m but with
DBH < 4.5 cm were identified and measured for DBH and height. This is
designated t he sapling group.

238
3. In the 1 m x 1 m temporary subplots all individuals less than 1.3 m tall were
identified and counted. This is designated the seedling group.

Data analysis
The data from the plots and subplots was analyzed for 3 measures of ecological
importance: the importance value index (IVI), the Shannon-Wiener index for
diversity, and the similarity index. The IV1 is a composite index based on mea-
sures of relative frequency, relative density, and relative dominance (Mueller-
Dombois and Ellenberg 1974). The Shannon-Wiener index, borrowed from
information theory, relates the proportional weight of the number of individuals
per species to the total sample belonging to all species (Shannon and Weaver 1949).
The similarity index is an arithmetic comparison of those values common to two
groups with the total value of both groups (Bray and Curtis 1957)

RESULTS
Species composition
In the Dry Dipterocarp forest the dominant species were Shorea siamensis Miq. in
the tree group, Cratoxylum formosum (Jack) Dyer in the sapling group and ferns in
the seedling group. The IVI for each of the dominant plants was 95.99, 35.60 and
54.82, respectively (Table 1).
In the Mixed Deciduous forest, the dominant species in the tree, sapling, and
seedling groups were Tectona grandis Linn, Pterocarpus macrocarpus Kurz. and
Kalanchoe verticillata Eli., respectively. The IVI of these dominant plants was 53.74,
66.20 and 32.60, respectively (Table 2).

Species diversity
For the Dry Dipterocarp forest community, the species diversity indices of the tree,
sapling and seedling groups were 3.51, 4.49, 3.63, respectively (Table 3). In the
Mixed Deciduous forest community, the highest species diversity index occurred
in the seedling group (3.87), followed by the sapling group (3.75), and the tree group
(3.72) (Table 4). Both forest communities had the highest number of species in the
seedling group, followed by the tree and the sapling groups. Some plant groups
had lower species index values but a higher number of species due to higher den-
sities of individuals from a single species unevenly distributed across the subplots.

239
 the sapling group of the Dry Dipterocarp forest with the highest Shannon-Wiener
species index value (4.49). In the tree group of the Mixed Deciduous forest the
distribution and magnitude of the IVIs of the dominant species was comparable
to the IVIs of dominant tree species in tropical rainforest communities
(Kiratiprayoon 1986).
In order to develop sound forest management practices based on sustainabi-
lity, it does not suffice to know only the IVIs of each species, in each plant group, in
each forest type, nor only values of species diversity for each plant group, in each
forest type. The interrelationship of diversity and importance between different
size groups must be appreciated. The tree group of the future will arise from the
saplings of today; the future saplings will come from current seedlings and future
seedlings, in turn, will be a result of the reproductive success of the mature tree
group.
Measures of similarity between size groups within the Dry Dipterocarp forest
and Mixed Deciduous forest studied provide two clear and contrasting examples.
In the Mixed Deciduous forest the similarity index between the tree and sapling
groups is 54.10 (Table 5). This suggests that a large proportion of the trees that are
important or dominant in the tree group, based on measures of relative frequency,
density, and dominance, also play an important role in the sapling group. If the
Mixed Deciduous forest was deforested, thereby losing the large trees (typically
the first to go), a similar group of species exhibiting similar patterns of interspe-
cific dominance would be available to replace the removed overstorey.
By contrast, the Dry Dipterocarp forest has a much lower similarity index value
between the tree and sapling groups (35.08) (Table 6). If the Dry Dipterocarp forest
was deforested it is less clear which species would become dominant. The major
structural component of the Dry Dipterocarp forest, Shorea siamensis, will be limited
in the near future (IVI=18.7 in the sapling group vs. 96.0 in the tree group). The
implications may be profound, not only for the 21 species that exist only in the tree
group (and would presumably disappear upon removal of the tree group through
deforestation) but also for those species existing only in the seedling and sapling
groups that may depend upon the structural homogeneity of the Shorea siamensis
dominated forest for their continued existence. The ecological uncertainty created
by a deforested and, therefore, radically altered environment may make these spe-
cies more susceptible to local-scale extinction.
Returning to the Mixed Deciduous forest example, we see that the high simi-
larity between the tree and sapling groups provides this forest type a relative mea-
sure of flexibility or resilience in the face of encroaching deforestation If, however,

243
ACKNOWLEDGMENTS
The authors gratefully acknowledge the assistance of Mr. Patrick John Baker, an
American scientist working at Huai Kha Khaeng Wildlife Sanctuary, for his edi-
ting. Also, the authors would like to thank the Demonstration Forest and Forest
Management Development Section and its staffs, especially Mr. Surapong
Chaveepak who helped the authors in the field work for data collection.

REFERENCES
Boontawee, B., C. Plengklai and A. Kao-sa-ard. 1995. Monitoring and Measuring
Forest Biodiversity in Thailand (Chapter 8, this volume).
Bray, J.R. and J.T. Curtis. 1957. An ordination of the upland forest communities of
southern Wisconsin. Ecol. Monographs 27: 325-349.
Kiratiprayoon, S. 1986. Comparative study on the stucture of the rattan bearing
tropical rain forest. M. SC. Thesis, Kasetsart University.
Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and Methods of Vegetation
Ecology. John Wiley & Sons, New York.
RFD. 1992. Forestry Statistics of Thailand. Royal Forest Department. Bangkok,
Thailand.
Shannon, C.E. and W. Weaver. 1949. The Mathematical Theory of Communication.
Univ. Illinois Press, Urbana.

245
 Chapter

17
BIODIVERSITY ASSESSMENT
OF FOREST TREE SPECIES
IN BIA NATIONAL PARK GHANA

A. A. Oteng-Yeboahl

INTRODUCTION
The Bia National Park (BNP) is located at the north-western tip of the Western Region
of Ghana, between longitudes 3o 02’and 3o 08’W and latitudes 6o 32’and 6o 37
North. Since 1975, no exploitation of any kind has taken place there. In 1985, Bia
National Park was designated a Biosphere Reserve and an UNESCO World Heri-
tage site.
The position of the park is intermediate between the Moist Evergreen (ME)
and the Moist semi-deciduous (MS) Ghanaian vegetation classification of Hall
and Swaine (1976) corresponding to the vegetation classification schemes of
Taylor (1952) and Mooney (1961). By its position, therefore, the Bia National Park
is expected to show all the characteristics of the two rain forest types. An unpu-
blished report by Hall, Swaine and Lock (1976) on the park indeed confirms this.
The situation of Bia National Park makes it ideal for base-line studies of forest

1 Dept. of Botany, University of Ghana, Legon, Ghana.

247
ecosystems in Ghana. One of these studies is the estimate of the relative abundance
of tree species in order to provide an age distribution structure for predictive yield
capacities purely for forest research purposes. A study of this kind is reported here.

MATERIALS AND METHODS

Ten permanent quadrats, each measuring 25m x 25m, were randomly demar-
cated along the north-south transect of the Park. The average distance among the
plots was about half a kilometre, with the distance between the nearest and the
farthest almost 3 kilometres. In each of the ten quadrats, woody plants with girth
over 20cm at breast height (1 metre from the ground level) were identified, counted
and their girth measured. Species identification followed Hutchinson and Dalziel
(195472), Irvine (1961), Hall and Swaine (1981), and Hawthorne (1990).
Calculations of means ( X), diameter (2r), radius (r), and area of girth were
made from the tree girth measurements. The cover dominance measure was
calculated from the total basal area of the trees. For convenience, the girth sizes
were grouped into five arbitrary classes ranging as follows: a) 20 - 50cm b) 51 -
80cm c) 81- 110cm d) 111 - 140cm e) over 141 cm.
The Ghanaian timber classification, in which trees are classified into Classes I, II
and III respectively for economically most desired, economically less desired, and
lesser known species was used. Sub classes and for classes I and II refer to
growth rate at above 50cm dbh with a- the fastest and the slowest.

RESULTS
Species Occurrence and Diversity
A total of 364 trees were encountered in the 10 quadrats. These belonged to 63
species in 24 plant families (Table 1). On the average therefore, there were 36.4
trees in 19.2 species belonging to 12 families per quadrat. The list of species, their
families, the total number of individual trees, their frequencies in the ten quadrats
and girth size ranges have been given in Table 2. Species with 5 or more frequencies
in the 10 quadrats were considered common. There were thirteen such species, which
included Anthonotha fragrans Baphia nitida, Bussea occiden talis, Cola gigan tea,
Corynan the pachyceras, Dialium dinklagei, Diospyros canalicula ta, Guibourtia ehie,
Nesogordonia papaverifera, Panda oleosa, Pycnanthus angolensis, S terculia oblong and
Strombosia glaucescens.

248
 saplings and young trees at different stages of development, which need to be
identified and nurtured for the future trade in timber.
It is interesting to note from the results above that four of the very common
species in the ten plots are indeed classified as important timber species. These are
Guibourtia ehie, Nesogordonia papaverifera, Pycnan thus angolensis and Strombosia
glaucescens. It is also worth noting that none of the most desired timber species (in
classes la, lb and 2a) particularly Entandrophragma cylindricum, E. utile, E. candollei
and Piptadeniastrum africanum were common; occurring only once in the total
observations. These species may be considered rare in the Bia National Park, and
every effort must be made to ensure that their stock does not die out. If possible,
enrichment planting of such species must be encouraged.
The abundance of species other than the currently desired ones of value for
timber, demand further studies for their genetic diversity and their importance
in terms of local uses by the people living near the park. There is a need, in future
studies, to assess the variability among the individual plants in each of these species,
in order to determine genetic variability for future silvicultural treatments and
breeding. This is especially necessary if these plants assume positions of value in
the future, or are over-exploited and therefore become rare.
The history of the timber industry in Ghana is quite chequered, previously
involving the utilisation of only a few timber trees, through an expansion to in-
volve additional species, now referred to as non-traditional timber trees is taking
place. This history has seen changes in the emphasis of timber management and
classification (Baidoe 1976) to now include several other species that were previous-
ly unclassified. A typical example is Guibourtia ehie, which was previously
unclassified, but which has now been classified with a class 2b rating because of
its promotion as export timber (Hall and Swaine 1981). Of the 364 trees examined
for girth density and basal area, only 95 of these (25%) are currently known to be
exploitable and therefore classified by the Forestry Department. The remaining
270 trees (74%) currently unknown for timber, may become classified in the future
when the stock of the desired ones get exhausted, or when there is a diversifi-
cation drive for all kinds of wood, as is currently happening in the country.
The Ghana Forestry Department refers to some 14 tree species as traditional
timber species (Forestry Dept. 1988). All of these tree species are currently classi-
fied in Class I, namely: Entandrophragma angolense, E. cylindricum, E. utile, Khaya
an thotheca, K. gradifoliola, K. ivorensis, Militia (Chlorophora) excelsa, Nanclea diderrichii,
and Tieghemella heckelii in Class la; Lovoa tridiilioides, Pericopsis elata, Terminalia
ivorensis and Triplochiton scleroxylon in Class lb; and Heretiera utilis in Class lc. The

254
observations at Bia National Park have shown the presence of five of these species,
En tandrophragma cylindricum, E. utile, Khaya ivorensis, Militia (Chlorophora) excelsa,
and Triplochi ton scleroxylon.
The status of the Bia National Park, both as a national park for the conservation
of fauna and floral species, and a World Heritage site and a Biosphere reserve of
UNESCO, will ensure the conservation of these tree species with potential to serve
as sources of genetic resources. The park is situated within a predominantly rural
farming community. As more lands around the park become converted into farms,
particularly for cocoa plantations, several non-timber forest products which the
people need for some of their other economic activities are also depleted. The park
appears to be the only refuge where such products will remain available. Already
pressure has begun to mount on the park to supply these products. A need for the
understanding of the status of the park by the local people is necessary at this point
in time. Several efforts, including education about sustainability and in situ and ex
situ conservation methods/ techniques of non-timber forest products etc. should
be undertaken and backed by appropriate incentives. These would ensure that the
contributions of the Bia National Park in conservation and sustainable develop-
ment of forest ecosystems are maintained for posterity.

ACKNOWLEDGEMENT
The African Academy of Sciences (AAS) Nairobi, Kenya provided funds for this
project, and the author is very grateful.
Mr. D.K. Abbiw of Botany Department, Legon helped in the identification of
the trees and Mr. J.Y. Amponsah helped in the field investigations.

REFERENCES
Baidoe, J.F. 1976. Yield regulation in the high forests of Ghana. Ghana Forest Journ.
2: 22-27.
Forestry Department. 1988. National Inventory Results. Government of Ghana,
Accra.
Ghartey, K.K.F. 1990. The evolution of forestry management in tropical high forest
in Ghana. A paper presented at conference. Sur la conservation et l’utilisation
rationnelle de la Foret dense d'Afrique Central et de l’ouest. ADB/IUCN/World
Bank, Abidjan.

255
Hall, J.B. and M.D. Swaine. 1976. Classification and ecology of closed canopy forest
in Ghana, J. Ecol. 64: 913-915.
Hall, J.B. and M.D. Swaine. 1981. Distribution and ecology of vascular plants in a
tropical rain forest: Forest Vegetation of Ghana. Dr. W. Junk Publishers, The
Hague.
Hall, J.B., M.D. Swaine and J.M. Lock. 1976. Report on visit to Bia National Park,
Dept. of Game and Wildlife, Accra. Mimeo.
Hawthorne, W. 1990. Field Guide to the Forest Trees of Ghana. Chatman: Natural
Resources Institute for Overseas Development Administration, London, Ghana
Forestry Series 1.
Hutchinson, J. and J.M. Dalziel. 1954-72. Flora of West Tropical Africa vol. I, II & III,
Crown Agents, London.
Irvine, F. R. 1961. Woody Plants of Ghana, OUP, London.
Mooney, J.W.C. 1961. Classification of the vegetation of the high forest zone of Ghana.
Pages: 85-86 In: Tropical Soils and Vegetation. Proceedings of the Abidjan
Symposium II, UNESCO Paris.
Taylor, W. 1952. The vegetation zones of the Gold Coast. Bull. Gold Coast Forestry
Dept. 4: l-12.

256
 Chapter

18
MEASURING AND INVENTORYIN G
ARTHROPOD DIVERSITY IN TEMPERATE AN D
TROPICAL FORESTS

N.E. Stork1

INTRODUCTION
The diversity of arthropods in forests is enormous. This is true of temperate as well
as of tropical forests (Hammond and Owen 1995, Hammond 1990, Stork 1991, Stork
and Brendell 1990). Hammond and Owen (1995), for example, recorded 959 spe-
cies of beetle in the 1000 hectare Richmond Park during their intensive five year
study of this woodland in southern Britain (Figure la). This represents a quarter of
all British species of beetles. By comparison, more than 6,000 species of beetle were
collected in a similar sized area of tropical forest in North Sulawesi in one year
(Hammond 1990, 1994 and unpublished, Stork and Brendell 1990) (Figure lb). If
we consider that beetles represent about 20% of the arthropod fauna in a forest
(Stork 1988, 1993), then the total arthropod fauna in these examples of a temperate
and a tropical forest may be at least 5,000 and 30,000 species respectively
Not surprisingly therefore, a full inventory of the arthropods in forests (let
alone fungi and microorganisms) has never been completed for any site,

1 Biodiversity Division, Entomology Department, The Natural History Museum, Cromwell Road,
London SW7 5BD, UK.

257
temperate or tropical, on Earth. For example, in spite of the efforts of researchers
from Oxford University in the 1950’s through to the 1980’s, the full list of arthro-
pod species from Wytham Wood - one of the best known temperate forest sites in
the world - has never been completed. There are several reasons for this. First,
arthropods are so taxonomically diverse and species rich that there are never
enough taxonomists to cover all groups. Second, temporal and spatial variation
in the abundance and distribution of species means that even intensive sampling
of a forest over several years will inevitably fail to collect all species that use, or
potentially use that forest. For this reason, accumulation curves for insects from
forests never reach an asymptote but continue to creep upwards (see Figure 2 in
Stork 1993) . At this near asymptote many of the new species being collected will
be transients (‘tourists’ sensu Moran and Southwood 1982), having little to do with
the area concerned. Over a 15 year period Owen (1991) collected 1782 species of
animals and 422 species of plants in her 0.5 hectare garden and many of these spe-
cies are tourists. The third and final reason is that no collecting method can collect
all species present in an area and, however intensively an area may be sampled and
whatever range of sampling methods are used, some species will be missed. The
introduction of novel methods of collecting are testimony to this. For example,
canopy fogging with knockdown insecticides has revealed an unimagined rich-
ness of insects in trees (Stork 1991). Noyes (1984) and Askew (1985) both described
the immense diversity of Chalcidoidea collected by canopy fogging from trees in
Brunei and Richmond Park, respectively (see Table 1). Probably more than 90% of
the species in the Brunei samples were new species, collected for the first time.
Sampling the fauna of leaf litter and soil using traditional Berlese-Tullgren funnel
methods is extremely difficult in tropical forests and the readily portable Winkler
bag has revolutionised the collecting of many arthropod groups for this habitat.
The first five species of the very small chrysomelid, Clavicornaltica, were described
from leaf litter samples in 1974 from Sri Lanka (Scherer, 1974). P.M. Hammond
(unpublished) collected a further 70 species from the Mulu National Park in Sarawak
in 1978 using the Winkler bag method. Subsequently, many other species - mostly
waiting to be described - have been collected from leaf litter using this method, in
pit fall traps, and even from trees by canopy fogging in other parts of South-east
Asia.
These comments are not intended to deter researchers planning intensive
sampling of forests for insects but rather to set the scale of the task involved. With
such immense diversity, is it possible to carry out research programmes aimed at
inventorying the arthropod fauna of forests or aimed at measuring the changes in

259
Obvious examples are termites and earthworms which are essential for the pro-
duction and maintenance of soil, whereas some species of thrips (Thysanoptera),
beetles and bees are essential for the pollination of plants. Other species are less
obviously important but are an integral part of the many complex foodwebs that
are to be found in forests (e.g. Memmott et al. 1994). Measurement of biodiversity
in forests, whether for studies of the diversity or health of the system, must in-
clude assessments of insects and other invertebrates.
As discussed below, for rapid o r longer term assessments of biodiversity,
inevitably choices have to be made as to which groups to use for inventorying.
Usually, birds and mammals and large plants are first choices with few insects
falling into this category. Butterflies, dragonflies and tiger beetles are perhaps, the
most obvious exceptions. However, as several authors have shown (Prendergast
et al . 1993, Yen 1987, Burbridge et al . 1992) the distributions of different insect groups
or other taxa do not necessarily correlate with each other, and the conservation
needs for different groups of insects may not be fairly represented by either these
insect taxa or other groups of animals and plants. For example, Prendergast et al.
(1993) showed that the richest areas of the UK for butterflies, dragonflies, liver-
worts, aquatic plants and breeding birds did not necessarily overlap, nor did their
centres of endemism. In addition, measurement of the diversity of groups such as
plants, birds and mammals may provide a poor indication of the conservation value
of some temperate woodlands, particularly managed woodlands (Harding and
Rose 1986). In these cases, the richness of saproxylic and other insects (particularly
Coleoptera and Diptera) are more useful indicators of the age and health of these
woodlands (e.g. Hammond and Harding 1991, Terrell-Nield 1990)

INDICATOR GROUPS
The term indicators has a variety of interpretations in different contexts. Some use
the term in a general way to mean a variety of measures of the condition of a system.
Reid et al . (1993), for example, provide a list of 22 biodiversity indicators (e.g.
species richness, species threatened with extinction, species used by local residents)
that can be used to measure the conservation status of an ecosystem. What they do
not attempt to do is to determine for which taxa these data should be compiled.
Clearly, we know so little about the Earth’s fauna and flora that it would be
impossible to measure these biodiversity indicators for all taxa and therefore cer-
tain taxonomic or morphological groups need to be selected as indicator groups. In
the terms of Kitching (1993) we need a predictor set of organisms such that changes

261
in their biological status will reflect similar changes in a wider group of organisms.
The term indicator group has been loosely used by many and has therefore come to
represent a very broad term. Hawksworth (in press) compiled a list of alternatives
for this term which are more specific in their application. For example, he suggests
that biomonitors are concerned with the monitoring changes in individual
populations over time. Here, I consider the term bioindicatorto be used for those
taxa which through measurements of their presence/absence, abundance,
distribution, species richness or other measure provide an indication of the health
or state of a broader group of insects or other community. Many species of
Chironimidae, for example, are extremely sensitive to changes in pH or pollution
levels of freshwater and are therefore used as indicators of water health, or rather
health of the freshwater animal community (Pinder and Morley in press). Similar-
ly, carabid beetles and macrolepidoptera are commonly used as indicators in
terrestrial ecosystems including temperate and tropical forests (e.g. Brown 1991,
Luff and Woiwod in press). Others have made cases for their own specialist groups,
such as Scarabaeidae (Halffter and Favila 1993) and Cicindelidae (Pearson and
Cassola 1992), to be considered as candidate indicator groups. Brown (1991) analysed
some of the qualities that might be required of an insect indicator group and de-
vised a scoring system. This system indicated that amongst terrestrial insects, ants
(Formicidae), some butterfly groups (Heliconiini, Ithomiinae), some beetle groups
(Carabidae, Cicindelidae and others) and Isoptera, scored highest and therefore
were important indicator groups. Brown’s system is somewhat idiosyncratic and
perhaps more appropriate for the neotropics than elsewhere. Clearly, which groups
are selected as indicators will depend on what they are supposed to be indicating,
the ecosystem concerned, geographical location and other circumstances. However,
as Prendergast et al. (1993) have shown, we have to be sure that our indicator groups
do truly provide a broad picture of the h ‘ ealth’of a wider range of groups. In this
sense the term p ‘ redictor set’is particularly appropriate (Kitching 1993).
What is lacking is a truly objective analysis of which groups of insects to select
as your indicators. This set would differ depending on the situation and the kind of
assessment being made. Brown’s (1991) set of values would not necessarily be
suitable for a temperate or freshwater community. One approach to this problem is
to select an appropriate s‘ hopping basket’of taxa (Stork 1994). For example, rather
than pick just one or two taxa which have their own peculiar biologies, it might be
more appropriate to select a larger group of say 10-20 taxa with a much broader
range of biologies and habitat preferences. This shopping basket of groups might
thus be a better predictor set than just one or two groups. Stork (1994) identified a

262
number of characteristics that might be considered when selecting a shopping bas-
ket of groups including the following: taxonomic breadth, ecological diversity
(e.g. ranges of body size, feeding guilds, reproductive biology). Where such
information is available, it would be useful to consider the sensitivity of different
taxa to natural and human influenced environmental factors such as pH, humi-
dity/water availability, temperature, light, CO2 and NO X concentration, and heavy
metals.
In addition to these biological factors there are important practical factors that
need consideration when selecting a predictor set or shopping basket of groups.
For example, these groups need to be ones that can be sampled cheaply. rapidly
and reliably Groups selected should not be taxonomically intractable and also there
must be expertise widely available or adequate literature sources such that taxa
can be readily identified.

RAPID ASSESSMENTS
In recent years there has been an increasing demand for inventories of temperate
and tropical forests. The UNCED meeting of heads of government in Rio de Janeiro
in 1992 and the resulting agreements, such as Agenda 21, the Biodiversity
Convention, and the Climate Change Convention, raised our awareness of the
plight of biodiversity. As a result of these agreements, and because of the perceived
urgency of the situation, many agencies such as the World Bank require rapid
assessments of biodiversity both to assist conservation priorities and for the
evaluation of the impact of large-scale development projects. Trueman and Cranston
(1994) highlight the different perceptions of such rapid assessments. They suggest
that the Biodiversity Rapid Resource Appraisal Study of the World Bank
Environmental Facility (GEF) is concerned with the production of an assessment in
a fixed and short time frame in order to provide practical conservation advice. Al-
most of necessity such assessments are desk-top studies using existing data from
herbaria and museum collections with perhaps some computer simulations or
computer mapping schemes to provide predictions. The United Nations
Environment Programme Country studies and many national Biodiversity Action
Plans are generally similar in their approach to rapid assessments of biodiversity.
The UNEP Guidelines for Country Studies recommends a large range of
measurements of biodiversity (see UNEP 1993) but only relatively few relate to
species level data. Implicit in these guidelines is the use of existing data rather than
the collection of new data from field surveys. Many of the measures they recommend

263
are socio-economic measures of biodiversity. This is not surprising as the sustaina-
ble utilisation of biodiversity is a high goal for many developing nations.
Furthermore, documents such as Country Studies and National Biodiversity Action
Plans, are highly political documents.
Another use of the term rapid assessment is more or less synonymous with fast
field survey (e.g. Parker et al. 1993). As Trueman and Cranston (1994) point out,
these field-based rapid assessments typically examine only trees, large mammals
and birds (sometimes reptiles and amphibians) and therefore consider the diver-
sity of a very small proportion of the total species in a forest or other habitat being
examined. Perhaps more than 95% of the fauna and flora (arthropods, other
invertebrates, fungi and other microorganisms) in the areas examined are not
considered. Usually these assessments are single visits of a month or less and
therefore, even if invertebrate groups or fungi were to be examined, the temporal
nature of these organisms could not be addressed. Most insects are seasonal even
in apparently aseasonal climates (see Paarmann 1976). However, it is usually possi-
ble to predict the most appropriate season for sampling the greatest diversity of
insects at a particular site.
How then can rapid assessments be carried out for insects? First, it is important
that some groups are recognised as front line indicator groups. As yet no coordina-
ted effort has been made to select such indicators, predictors or sampling packages,
although there is now much evidence from recent studies to make such a selection.
Second, standardised methods of sampling need to be agreed. For groups such as
amphibians and birds, such standards have been recognised (e.g. Heyer et al. 1993).
Third, methods of statistical analysis need to be agreed. The latter is probably the
area most lacking in agreement and understanding.

MORPHOSPECIES
For those biologists studying ecological patterns in species-rich communities, such
as insects in temperate or tropical forests, sorting samples and dealing with the
many species in them are major problems. For some studies, knowing the precise
name of a species may not be essential as long as some aspects of the feeding biolo-
gy or habitat can be determined. However, identifying species does provide some
measure of confidence in the level of sorting and provides access to the biological
literature associated with the named species. However, in most parts of the world
there are neither sufficient taxonomists nor adequate collections of authentically
named species to be able to identify all species found in samples. This problem is

264
further compounded by the fact that in many samples most of the species are new
to science and are therefore unnamed. In such cases researchers sort their samples
on morphological characters as near as possible to what they believe are species
so called ‘morphospecies’ (also known as ‘recognisable taxonomic units’ (RTUs) or
‘operational taxonomic units’ (OTUs)). It is currently acceptable that once having
defined the level of sorting (i.e. to morphospecies/RTUs/OTUs) in a paper, authors
then use the term species throughout, as if their morphospecies are directly
equivalent to real species. Thus as standards of sorting vary enormously, the term
morphospecies has been somewhat abused in the literature and needs further
definition. Editors of ecological journals often can be very demanding of authors
when it comes to the statistical procedures used in a paper. This is because standard
statistical tests and procedures are now widely recognised. Statistical standards
have been raised by the widespread use of commercial computer statistical packa-
ges. However, similar taxonomic standards do not exist and therefore levels of
taxonomic competence in sorting samples to species, varies greatly The current
lack of sound training in taxonomy in many universities and personal experience
of the level of species sorting of some postgraduates and more senior scientists
completing community ecology studies indicates that this is an area which needs
some serious attention.
Oliver and Beattie (1992) examine how well a technician compared to
professional taxonomists in their sorting to morphospecies of samples of Austra-
lian mosses, polychaetes, ants and spiders after just a few hours of formal training.
They found a very high level of agreement in the sorting (see Table 2) with 88% and
83% ‘1 to 1’ level agreement at the species level for spiders and ant species,
respectively However, a sample of one technician is not adequate to assess the
taxonomic competence of technicians after limited training. Also it should be
recognised that in spite of good correlations between the species sorting and the
total number of species, the evenness of the sample, guild composition and so on
may be quite different. In many insect samples there is often a long tail of single-
tons but it can be very difficult to assess whether these are really different species
from each other or whether some of these may represent very variable species.
Sexual differences between species in some groups (e.g. Pselaphidae) can make it
very difficult to match males and females.
Here I propose a simple system of grading the level of accuracy of sorting to
morphospecies and verifying this sorting (Figure 2). This system will not go as far
as providing a ‘certificate of taxonomic competence’, equivalent to standard
statistical procedures or well recognised standards of ‘good laboratory practice’,

265
Initial sorting to species:

i) Morphospecies level 1 - specimens sorted to species without being removed

from site (e.g. trees or large mammals) or without vouchers being removed and
prepared from the sample and prepared in standard manner (e.g. mounted and
labelled or pinned as for beetles).
ii) Morphospecies level 2 - Initial sorting to species in the sample being
supplemented by preparation of a voucher collection of a representative set of
specimens. Voucher collections of single or few specimens per species are usually
inadequate for assessment of the distinctiveness of the species.
iii) Morphospecies level 3 - Preparation of all specimens and sorting of all to species.
Thus all specimens form the voucher collection.

Verification of species sorting:

iv) Morphospecies level 4 - Sorting checked by a non-taxonomist.

v) Morphospecies level 5 - Sorting checked against a comprehensive and accurately
named museum/herbarium collection.
vi) Morphospecies level 6 - Species sorting checked by an appropriate specialist
taxonomist or taxonomists.
vii)Morphospecies level 7 - Species formally described or new records published
for existing species from the samples by taxonomists.

Figure 2: A system of levels for morphospecies sorting.

ACKNOWLEDGEMENTS
I thank the organising committee of the Symposium on Measuring and monitoring
biodiversity in tropical and temperate forests for inviting me to present a paper in
Chiang Mai, Thailand and for their kind hospitality. I thank also Dr Tim Boyle for
his support and encouragement and Peter Hammond for comments on a draft of
this paper.

267
REFERENCES
Askew, R.R. 1985. A London fog. Chalcid Forum 4: 17.
Brown, KS. 1991. Conservation of neotropical environments: insects as indicators.
Pages: 349-404 In: N.M. Collins and J.A. Thomas (eds.) The Conservation of
Insects and their Habitats. 15th Symposium of the Royal Entomological Society
of London, Academic Press, London.
Burbridge, A.H., K. Leichester, S. McDavitt and J.D. Majer. 1992. Ants as indicators
of disturbance at Yanchep National Park, Western Australia. Journal of the Royal
Society of Western Australia, 75: 89-95.
Halffter, G. and M.E. Favila. 1993. The Scarabaeinae (Insecta: Coleoptera) an animal
group for analysing, inventorying and monitoring biodiversity in tropical
rainforest and modified landscapes. Biology International, 27: 15-21.
Hammond, P.M. 1990. Insect abundance and diversity in the Dumoga-Bone National
Park, N. Sulawesi, with special reference to the beetle fauna of lowland rain
forest in the Toraut region. Pages: 197-254 In: W.J. Knight and J.D. Holloway
(eds.) Insects and the rain forests of South East Asia (Wallacea). Royal
Entomological Society of London.
Hammond, P.M. 1994. Practical approaches to the estimation of the extent of
biodiversity in speciose groups. Philosophical Transactions of the Royal Society
of London, Series B, 345: 119-136.
Hammond, P.M. and P.T. Harding. 1991. Saproxylic invertebrate assemblages in
British woodlands: their conservation significance and its evaluation. Pages:
29-37 In: H.J. Read (ed.) Pollard and veteran tree management. Richmond
Publishing, Slough.
Hammond, P.M. and J.A. Owen. 1995. The beetles of Richmond Park SSSI - A case
study. English Nature, Peterborough.
Harding, P.T. and F. Rose. 1986. Pasture-woodlands in lowland Britain. A review of
their importance for wildlife conservation. 89pp. Institute of Terrestrial Ecology,
Huntingdon.
Hawksworth, D.L. (in press). Tropical Areas Bioindicators System (TABS). UNEP
and FAO.
Heyer, W.R., M.A. Donnelly, L.A.C. Hayek and M.S. Foster. 1993. Measuring and
monitoring biological diversity. Standard methods for amphibians. Smithsonian
Institution Press, Washington D.C.
Kitching, R.L. 1993. Towards rapid biodiversity assessment - lessons following

268
 studies of arthropods of rainforest canopies. Pages: 26-30 In: Rapid Biodiversity
Assessment: Proceedings of the biodiversity assessment workshop Macquarie
University 1993. Research Unit for Biodiversity and Bioresources, School of
Biological Sciences, Macquarie University, Sydney.
Luff, M.L. and I.P. Woiwod (in press). Insects as indicators of land-use change: a
European perspective, focusing on moths and ground beetles. In: R. Harrington
and N.E. Stork (eds.). Insects in a changing environment 17th Symposium of
the Royal Entomological Society, Academic Press, London.
Memmott, J. H., C.J. Godfray and I.D. Gauld. 1994. The structure of a tropical host-
parasitoid community. Journal of Tropical Ecology 63: 521-540.
Moran, V.C. and T.R.E. Southwood. 1982. The guild composition of arthropod
communities in trees. Journal of Animal Ecology 51: 289-306.
Noyes, J.S. 1984. In a fog. Chalcid Forum 3: 4-5.
Oliver, I. and A.J. Beattie. 1992. A possible method for the rapid assessment of
biodiversity. Conservation Biology 7: 562-568.
Owen, J. 1991. The ecology of a garden: the first fifteen years. Cambridge University
Press.
Paarmann, W. 1976. Jahreszeitliche Aktivitat und fortpflanzungsrhythmik von
laufkäfern (Col., Carab.) im Kivugebiet (Ost-Zaire, Zentralafrika). Zoologische
Jahrbücher, Abteihing fur Systematik 103: 311-354.
Parker, T.A., A.H. Gentry, R.B. Foster, L.H. Emmons and J.V. Remsen. 1993. The
lowland dry forests of Santa Cruz, Bolivia: a global conservation priority.
Conservation International, Washington D.C.
Pearson, D.L. and F. Cassola. 1992. World-wide species richness patterns of Tiger
beetles (Coleoptera: Cicindelidae): indicator taxon for biodiversity and
conservation studies. Conservation Biology 6: 376-391.
Pinder, L.C.V. and Morley, D.J. (in press). Chironomidae as indicators of water quality
- with a comparison of the chironomid faunas of a series of contrasting Cumbrian
tarns. In: R. Harrington and N.E. Stork (eds.) Insects in a changing environment
Academic Press, London.
Prendergast, J.R., R.M. Quinn, J.H. Lawton, B.C. Eversham and D.W. Gibbons. 1993.
Rare species, the coincidence of diversity hotspots and conservation strategies.
Nature 365: 335-337.
Reid, W.V., J.A. McNeeley D.B. Tunstall, D.A. Bryant and M. Winograd. 1993.
Biodiversity indicators for policy-makers. World Resources Institute, Washington

269
 Chapter

19
MEASURING BIODIVERSITY DIVERSIT Y
OF MICROFUNGI IN THE WET TROPICS OF
NORTH QUEENSLAN D

Kevin D. Hyde1

INTRODUCTION
This chapter will focus on the microfungi of rainforests in the wet tropics of north
Queensland, Australia, choosing the fungi developing on palms and those in streams
to illustrate how little is known. with respect to other rainforest substrates in
north Queensland, the reality is we know very little about the microfungi present.
There have been very few studies and it appears that no taxonomic mycologists
have lived and worked in this region. The lack of information on the microfungi of
north Queensland is typical for the tropics as a whole. Australia is regarded as a
developed country and the situation is unsatisfactory. Most tropical rainforests
occur in developing countries where even less is known.
Between May 1989 and October 1992, I was able to study some of the micro-
fungi in the rainforests around Cairns. I focused my research on fungi developing
on wood submerged in streams and on fungi developing on palms. The data that I
am presenting is mostly the conclusion from my own research and I am aware of

1 Department of Ecology and Biodiversity, University of Hong Kong, Pokfulam Road, Hong Kong.

271
few other studies on th microfungi in north Queensland other than those in plant
pathology.
Anon (1992) estimated that 85% of Australian plants (estimated 44,000 species)
were either named or known, but that only 10% of fungi (estimated 200,000 species)
were named. In another estimate, Pascoe (1990) indicated that there could be as
many as 250,000 fungal species in Australia and that only 5% were known. There
may be many more species than even these estimates suggest. Pascoe (1990) based
his figures using a ratio of vascular plants (25,000) to fungal species (250,000) of
1:10. As will be illustrated with the studies of microfungi on palms, a ratio of 1:10
is very low.
A survey of the taxonomic mycologists in Australia in 1991 (Grgurinovic and
Hyde 1993) concluded that there were only 32 mycologists whose research
incorporated at least some taxonomic mycology - only one of these was working
on rainforest fungi. Can this small number of mycologists deal adequately with
the Australian mycoflora? Assuming a mycologist can describe 600 species during
a working lifetime (40 years), then 19,200 species (32 x 600) should be described in
the next 40 years. If we use Pascoe’s (1990) conservative figure of 250,000 estima-
ted fungal species, and assume optimistically that all Australian mycologists are
active (and the habitats are preserved), then the last Flora of Australia will be writ-
ten in the year 2515. This illustrates the dire plight of mycology in Australia.

WHY DO POLITICIANS NOT FUND MYCOLOGY?

Politicians and scientists alike are accountable for the decline of funding in myco-
logy and some of the reasons and misapprehensions as to why fungi are over-
looked and thus poorly funded are given:

1. Agriculturally and other economically important species of fungi are studied,

while most others are disregarded.
2. Most people prefer to destroy rather than collect fungi.
3. There is a misapprehension that most fungi are poisonous.
4. The larger tropical fungi are seasonally present for short periods (usually when
its raining!).
5. The public are unaware of the presence of fungi, most of which they can never
see.
6. The public are uninformed of the important role of most fungi.

272
7. Fungi rot people’s houses and grow on their bodies.
8. Most fungi are sessile.
9. Some fungi are offensive (stinkhorns).
10. Many fungi smell and some attract flies.
11. The ecological importance of fungi is often overlooked.
12. Mycologists do not promote mycology.

WHY SHOULD WE STUDY FUNGI?

With such a dreadful rating with the public, politicians and non-mycological
scientists alike, what prospect is there for mycology? Perhaps the greatest pro-
mise is in biodiversity studies with some of the potential benefits in the form of
novel compounds. Grgurinovic and Walker (1993) have given several reasons as to
why mycological herbaria should be maintained. I have expanded this below, to
give some reasons as to why we should study fungi:
1. Documentation of fungal biodiversity provides a basis for monitoring success
of conservation and management practices.
2. Control of plant and animal diseases caused by fungi.
3. Biological control of weeds (e.g. Mimosa pigra in the.northem territory).
4. Biological control of arthropods (e.g. Cordyceps, Mefarrhizum).
5. Medicinal properties (e.g. Cordyceps tea, antibiotics, immuno-suppressants).
6. Quarantine.
7. Role of fungi in nutrient cycling and thus management of ecosystems (e.g.
mycorrhizae, endophytes).
8. Attraction (e.g. Phallus sp.).
9. Biotechnology.
10. Mycotoxins (e.g. potent carcinogens - aflatoxins).
11. Pure science.
Numerous biotechnological products are produced industrially as a result of
fermentation utilising fungi (Table 1). However, there are millions of undiscovered
fungi world-wide that will have the genetic potential to produce a multitude of
novel compounds. The examples below illustrate some of the recent achievements
in this field of research.

273
Fungus Production

Ashbya gossypii Riboflavin (vitamin)

Aspergillus niger Citric acid, Gluconic acid,
a-amylase, protease
A. oryzae a-Amylase
A. terreus Itaconic acid
Mucor pusillus Rennin
Penicillium ch ysogenum Penicillin
Penicillium griseofulvin Griseofulvin (antibiotic)
Rhizopus nigricans Fumaric acid
Saccharomyces cerevisiae Industrial alcohol
Trichoderma reesei Industrial alcohol

Tabie 1: Some Biotechnological Uses of Fungi.

Cyclosporine
Cyclosporine, like penicillin one of the best known pharmaceutical successes, was
discovered as a metabolite of Tolypocladium inflatum Gams. This fungus was isola-
ted from soil in Hardanger, Norway in 1970 and was found to synthesise metabo-
lites that were later called cyclosporins. It is this highly modified strain that is now
used in the large scale production of cyclosporin A by fermentation (Bore1 and Kis
1991). Cyclosporin emerged as the prototype of a new generation of immuno-
suppressive drugs and is today important as the first-line treatment in organ
transplantation.

Other novel compounds from fungi

Extensive screening of fungi for the production of novel compounds is currently
under way and has resulted in the discovery of numerous potential pharmaceuti-
cal products (Nisbet and Fox 1991). Three examples are given here. Phomactin A,
isolated from a marine Phoma sp.,is a novel platelet activating factor (PAF)
antagonist. PAF may be implicated in many inflammatory, respiratory and
cardiovascular diseases (Sugano et al . 1991). Emeriamine, isolated from Emericella
quadrilineata is, a new inhibitor of long chain fatty acid oxidation and may have
potential utility as a therapeutic agent for treating diabetes (Kanamaru and Okazaki

274
1989). Two new depsidones, Auranticins A and B, were obtained from a mangrove
isolate of the fungus Preussia aurantica and were found to exhibit antibiotic activity
against Bacillus subtili and Staphylococcus aureus (Poch and Gloer 1991).

PALM MICROFUNGI
Fungi have been collected from eight of the sixteen native palm genera (38 species)
represented in Queensland. Most palm fungi (48 species) have been recorded from
Archon tophoenix alexandrae (F. Muell.) Wendl. and Drude. Many of these collections
on Archontophoenix are the result of a short visit to north Queensland by Matsushima
in May 1988 (Matsushima 1989), who lists 31 hyphomycetes, and of recent studies
by Frohlich and Hyde (1994) and Hyde (1993c,d,e, 1994a,b) who record 6
ascomycetes. Fungi on other palms are virtually unknown. Only 4 fungal species
have been recorded from Cocos nucifera L. (all plant pathogens) and 12 species from
Calamus spp. (Hennings 1903, Hyde 1992d, 1993d, 1994b, Hyde and Fröhlich 1994,
Hyde and Alcorn 1993, Matsushima 1989, Simmonds 1966). The low ratio of fungi
recorded on palms (1.5:1) in north Queensland reflects what little is known of the
mycoflora of all rainforest habitats.
Microfungi on palms include pathogens, endophytes and saprophytes. The
symptoms of pathogenic fungi are varied, but many form leaf spots and these are
often species or genus specific. Frohlich (1992), Fröhlich and Hyde (1994, 1995a, b, c)
and Hyde and Frohlich (1994) identified 28 taxa, including 2 new genera and 11
new species associated with palm leaf spots in north Queensland. The implica-
tions of fungal pathogens to quarantine and the ornamental palm industry
(estimated annual turnover of sixty five million dollars - Forsberg 1987) are enor-
mous and this is a glaring example of important research that has not been carried
out.
Endophytes develop within the tissue of plants and rarely produce external
symptoms. Their nutrition is derived from the plant and in return they may pro-
vide some form of protection (e.g. from grazing by insects). There is very little
available information on the endophytes of tropical plants (Rodrigues and Petrini
1995). Frohlich (personal communication) has recently isolated about 150 cultural
morphotypes from three single Licual a sp. palms in north Queensland, while I
have isolated similar numbers from Calamu ssp.
Saprophytic fungi are able to develop on a wide range of dead palm material
and are less likely to be host specific. However, one would expect some fungi to be
selective to specific palm species or genera. This was found to be true of mangrove

275
trees (Hyde 1990). The number of palm saprophytes, appears to be extremely high.
I have found several hundred taxa on palms in South East Asia (Hyde, unpubl.). In
Australia Archontophoenix alexandriae is the host of 48 saprophytes, but this is the
only palm examined in any detail.
It is possible to speculate on the numbers of palm fungi in Queensland. There
are 38 native palm species and we can estimate that there are about 3 plant patho-
gens, 100 endophytes and 10 saprophytes that can develop on each palm species. If
we are conservative and assume that 10% of microfungi are host specific, then in
north Queensland we can anticipate finding 430 fungal species on palms. If we are
less conservative and calculated that 25% are host specific, then 1073 microfungi
should occur on palms in north Queensland. These figures have marked signifi-
cance on the total numbers of fungi in Australia and world-wide. Pascoe (1990)
estimated that there were likely to be 250,000 fungal species in Australia, using a
host species to fungal species ratio of 1:10. Hawksworth (1991) calculated a world
total of 1.5 million fungi using a host species to fungal species ratio of 1:6. With
palms the host species to fungi ratio is (presumably) much higher.

FRESHWATER FUNGI
The microfungi (ascomycetes) occurring on wood in freshwater streams were
reviewed by Shearer (1993) who listed about 300 species. Of these only seven were
tropical, which reflects the lack of information on ascomycetes in tropical streams
globally. Only three of these seven species were Australian collections, an
inconceivably small number when compared to the number of habitats and the
role of these fungi in nutrient cycling. This lack of knowledge also applied to
other fungal types in freshwater habitats in the Australian tropics. With this
realisation, studies of the fungi colonising wood submerged in freshwater in far
north Queensland were initiated.
Forty fungi have now been collected on wood submerged in streams in north
Queensland, and 25 are formally described or listed (Hyde 1992a,b,c, 1993a,b,
1995a,b,c, Hyde and Seifert 1992). They are a distinct ecological assemblage and
belong to a diverse range of families. Taxonomic placement is not easy as few
mycologists have such a wide expertise. Only 6 of the 40 species are also found in
temperate streams, most species appearing to be restricted to the tropics. As little
work of this nature has been carried out in the tropics many of the fungi collected
are new to science; 7 new genera of ascomycetes, including 17 new species, one
new genus of synnematous hyphomycetes and one new genus of Coelomycetes

276
have been described. These fungi are apparently unrelated to host substrate and
are an extra group to add to biodiversity calculations.

WHY MEASURE FUNGAL DIVERSITY?

I have often heard the argument, why do we need to know what fungi are present
in an ecosystem? Why not just measure their activity or why not use isozyme stu-
dies to indicate fungal presence? I hope that I have succeeded in convincing you as
to why we should know what fungi are present in various habitats in the section
“Why should we study fungi ?“. As arguments for maintaining insect biodiversity
(Kim 1993) use impoverished biodiversity, loss of basic resources, and loss of
potential food resources, similarly, these reasons are applicable to fungi. Because
few fungal species are presently utilised in biotechnological processes, or in the
production of novel compounds, there is a huge potential for their use. Remarkably
little is known of microbial diversity, numerous species remain to be described,
and the genetic diversity within those that are known is scarcely studied (Bull
1991). We need to collect, identify, name and maintain these fungi before they
disappear. Furthermore, because of their integral role in ecosystems, e.g., nutrient
cycling, plant growth, food source, sensitivity to air pollution and perturbation,
fungi lend themselves to measuring and monitoring biodiversity (Rossman 1994).

THE NEED FOR HERBARIA AND CULTURE COLLECTIONS

An inventory of the microfungi present in any rainforest ecosystem is desirable,
but it is also important to conserve the fungi as dried material in herbaria and living
material in culture collections. There is a need for more national herbaria and cul-
ture collections, particularly within developing countries in the tropics from the
following perspectives:
1. Industrial perspecfive. Fungi and bacteria are an important source of novel
compounds used in medicine and food production. In South East Asia and China
there is a pool of genetic material in the form of fungi and bacteria that could be
important in developing any of these benefits and the material is presently not
available to industry.
2. Conservation and biodiversity perspective. With the rapid destruction of the world’s
rainforests and natural environments thousands of species become extinct every
year and this includes microorganisms. These microorganisms are pools of
genetic material that can be extremely useful to mankind in medicine, industry

277
 and numerous other avenues. We are therefore obliged to conserve as much of
this material as possible for future generations.
These arguments apply to many countries in the tropics. However, culture
collections can never hold more than a fraction of the fungal resource (17% of
accepted fungal species are presently maintained in culture collections) and the
only way to fully conserve diversity is by conservation of the ecosystems (Bull
1991, Olembo 1991).

In its simplest form, species diversity is the number of different species of orga-
nisms found in a particular habitat or ecosystem (Dighton 1994). One of the problems
in measuring the overall biodiversity of microfungi in rainforests is the large num-
ber of habitats one would need to sample and the diversity of organisms that would
be encountered. In a rainforest situation there are numerous microhabitats (Figure
1) and each of these would need to be sampled. This would require input from
numerous mycologists, since a soil fungi specialist is unlikely to be an authority
on palm fungi. Furthermore, diverse groups of fungi will be encountered, which
would again require the input of several mycologists, as no singie mycologist is
likely to be an expert in all fungal groups. Several techniques would be employed,
so that all groups of fungi present are isolated and catalogued. Finally, many of the
fungi collected in any survey would be new to science and their identification
may only be possible to genus or family level.
How then do we tackle the problem of measuring the immense biodiversity of
microfungi in rainforests? Do we choose indicator organisms or target groups, in a
similar way to those of entomologists. Could we train students or “biodiversity
technicians” for short periods, to sort fungi into recognisable taxonomic units
(Oliver and Beattie 1993), or do we approach the problem in a different way? A
need for world initiatives in biodiversity databases has been stressed by
Hawksworth and Mound (1991). Standard techniques for measuring diversity are
required before meaningful comparisons can be made between geographically
distinct regions and different habitats. Mycologists need to discuss and agree on
these methods and a workshop to discuss approaches may be an appropriate forum
(Rossman 1994). Such a workshop would consider techniques, from visual sam-
pling using plots or transects, to laboratory isolation and culturing. In the remain-
der of this chapter I will briefly discuss some techniques propose one method
for measuring biodiversity of microfungi in rainforest. It is my hope that it will
promote the much needed discussion amongst mycologists.

278
INVENTORY AND MONITORING
In Chapter 3, Burley and Gauld propose immediate assessment needs, new re-
search required and the need for rapid monitoring methods for all organisms. Many
of their recommendations also apply to microorganisms.
One way forward is to target genera or families of microorganisms or specific
habitats, as a measure of biodiversity. However, there may be problems in this
approach. If the ascomycete genus Xylaria was chosen to represent diversity in
tropical rainforests, then the results might be misleading. In a rainforest near
Lockerbie in north Queensland in March 1991, there was a large diversity of Xylaria
species, while in Pasoh rainforest in Malaysia in November 1991, Xylaria species
were scarce. If we extrapolated from these results we would conclude that the
diversity of fungi in Lockerbie was much greater. Although this may be true, it
may be equally true that the numerous Xylaria species at Lockerbie was the result
of a recent destructive cyclone. Similarly, choosing habitats, e.g. bamboo, to mea-
sure biodiversity of microorganisms may also generate deceptive results. The
number of native bamboo species in Australia is small when compared to Borneo
and a measure of fungal diversity on bamboo would be very biased towards large
numbers in the latter region. The timing of the survey would also be significant, as
fungal diversity would differ between wet to dry seasons.
An integrated approach may provide a viable solution to measuring biodiver-
sity in rainforests. Unfortunately it will be time consuming and require the exper-
tise of several mycologists. A permanent protected plot of 100 m 2 (selected to
incorporate high plant and habitat diversity) should be established under the
auspices of a local scientist. The plant species within the plot should be identified
and labelled where possible. A mycological inventory can then be carried out over
a period of several years, with input from appropriate specialists. The larger
basidiomycetes (e.g. polypores), ascomycetes (e.g. Xylaria), and some biological
groups of fungi (e.g. entomophagous fungi, freshwater fungi) can be collected
and identified (spatially and temporally) over the whole plot, as their numbers
would be manageable.
Microfungi present the biggest challenge as they are the most abundant mycota
(microfungi: larger fungi, about 30:12) and an inventory would be unmanage-
able in large plots. For these fungi it may be necessary to select smaller plots (10 m 2 )
or individual host trees and sample within these elements. Methods for measuring
fungal diversity in soils have been reviewed by Gray (1990) and by Frankland et al.

22 Figures calculated from described basidiomycetes vs other fungi (Hawksworth et al. 1983).

280
(1990), with a critique of some of the methods by Dighton (1994). Bills and Polishook
(1994) have measured abundance and diversity of microfungi in leaf litter of a
lowland rain forest in Costa Rica. Four litter samples were collected from four
arbitrarily selected sites, by standing at a single point and raking the surrounding
litter into containers. They isolated between 78-134 fungal species from each litter
sample, with few abundant species and a high proportion of rare species. Several
methods for measuring ecological diversity have also been reviewed by Magurran
(1988). It may be possible to adopt these, or similar techniques, as standards for
measuring diversity of soil and leaf litter microfungi.
There are flaws in these methods. The microfungi present on leaf litter may be
determined by the host component of the litter. If the litter contains palm remains,
then the fungal community will differ as the genera of fungi developing on living
and dead palm material is quite different from that associated with other hosts.
Members of the Graminaceae (e.g. bamboo), Zinziberaceae (e.g. Alpinia) also have
a somewhat unique mycota. Appropriate specialists may need to cover these
substrates.
At present there are no standard techniques for measuring fungal diversity and
measuring diversity of microfungi is one or our major challenges. Pilot studies to
establish convenient techniques and protocols are essential, so that mycologists
can eventually agree on international measuring and monitoring strategies.

CONCLUSION
In this paper I have presented information on the fungi known from two habitats
in north Queensland. I must emphasize that for each taxon there is usually a single
record and this is the only information that is available. We do not have distribu-
tion maps of the occurrence of the fungi as is published (or known) for most vascu-
lar plant species in north Queensland (Clarkson - personal communication). We
know nothing of the mycobiota of most habitats, and most fungi remain unknown.
With this lack of knowledge one would hope that enlightened government bodies
such as Australian Biological Resources Study (ABRS) might give priority to fund-
ing mycological research in Australia. However, in 1993 and 1994 only 3 mycologi-
cal projects ($213,071) were funded by ABRS as compared to numerous vascular
plant projects ($ 17,702,467). The lack of knowledge of microfungi in Australian
rainforests is reflected internationally, with most developing countries being in a
worse state.
Difficulties in measuring biodiversity of microfungi in rainforests and the need

281
to develop standard techniques is discussed. A method for measuring biodiver-
sity in rainforests is proposed, with the hope of promoting discussion amongst
mycologists.

ACKNOWLEDGEMENTS
Thanks are extended to C. Grgurinovic, D. Guest, I.J. Hodgkiss and J. Clarkson for
their comments on the manuscript.

REFERENCES
Anon. 1992. Preserving biodiversity. ECOS 74, 5.
Bills, G.F. and J.D. Polishook. 1994. Abundance and diversity of microfungi in leaf
litter of a lowland rainforest in Costa Rica. Mycologia 86, 187-198.
Borel, J.F. and Z.L. Kis. 1991. The discovery and development of cyclosporine
(Sandimmune). Transplantation Proceedings 23, 1867-1874.
Bull, A.T. 1991. Biotechnology and Biodiversity. Pages: 203-220 In: The Biodiversity
of Microorganisms and Invertebrates: Its role in Sustainable Agriculture. C.A.B.
International, Oxon.
Dighton, J. 1994. Analysis of micromycete communities in soil: a critique of methods.
Mycological Research 98, 796-798.
Frankland, J., J. Dighton and L. Boddy. 1990. Methods for studying fungi in soil and
forest litter. Pages: 343-404, In: R.G. Grigorova and J.R. Norris (eds.) Micobiology,
Vol 22. Academic Press, London, U.K..
Forsberg, L.I. 1987. Diseases of ornamental palms. Queensland Agricultural Journal
Sept.-Oct., 279-286.
Fröhlich, J. and K.D. Hyde. 1994. New Oxydothis species associated with palm leaf
spots in north Queensland, Australia. Mycological Research 98,213-218.
Fröhlich, J. and K.D. Hyde. 1995a. Guignardia candeloflamma sp. nov. causing leaf
spots of Pinanga spp. Mycological Research 99,110-112.
Fröhlich, J. and K.D. Hyde. 1995b. Astrosphaeriella fronsicola sp. nov. associated with
leaf spots of Oraniopsis. Mycological Research 99, (accepted).
Fröhlich, J. and K.D. Hyde. 1995c. Maculatipalma fionsicola gen. et sp. nov. causing
leaf spots on palm species in north Queensland with descriptions of related
genera: Apioplagiostoma and Plagiostoma. Mycological Research 99, (accepted).

282
Gray, T.R.G. 1990. Methods for studying the microbial ecology of soil,. Pages: 310-
342, In: R.G. Grigorova and J.R. Norris (eds.) Methods in Micobiology, Vol 22
Academic Press, London, U.K.
Grgurinovic, C. and K.D. Hyde. 1993. The status of taxonomic mycology in Australia
in 1991. Australian Plant Pathology 22, 42-47.
Grgurinovic, C. and J. Walker. 1993. Herbaria and their place in science: a mycological
and plant pathological perspective. Australian Plant Pathology 22, 14-18.
Hawksworth, D.L. 1991. Presidential address 1990: The fungal dimension of
biodiversity: magnitude, significance, conservation. Mycological Research 95,
641-655.
Hawksworth, D.L., B.C. Sutton and G.C. Ainsworth. 1983. Ainsworth and Bisby’s
Dictionary of the Fungi. C.A.B. International, Oxon, U.K.
Hennings, P. 1903. Fungi australiensis. Hedwigia 42 Beih.: 73-88.
Hyde, K.D. 1990. A comparison of the intertidal mycota of five mangrove tree species.
Asian Marine Biology 7,93-107.
Hyde, K.D. 1992a. Tropical Australian freshwater fungi. I. Some ascomycetes.
Australian Systematic Botany 5, 109-116.
Hyde, K.D. 1992b. Tropical Australian freshwater fungi. II. Annulatascus velatispora
gen. et sp. nov, A. bipolaris sp. nov. and Ophioceras dolichostomum (Ascomycetes).
Australian Systematic Botany 5, 117-124.
Hyde, K.D. 1992c. Tropical Australian freshwater fungi. IV. Halosarpheia aquatica
sp. nov., Garethjonesia lacunosispora gen. et sp. nov. and Ophioceras dolichostomum
(Ascomycetes). Australian Systematic Botany 5, 407-414.
Hyde, K.D. 1992d. Fungi from palms. I. The genus Linocarpon a, revision. Sydowia
44, 32-54.
Hyde, K.D. 1993a. Tropical Australian freshwater fungi. V. Bombardia sp., Jahnula
australiensis sp. nov., Savo yella aquatica sp. nov. and S. lignicola sp. nov. Australian
Systematic Botany 6, 161-167.
Hyde, K.D. 1993b. Tropical Australian Freshwater fungi. VI. Tiarosporella paludosa
and Clohesyomyce aquaticu gen. s et sp. nov. (Coelomycetes). Australian Systematic
Botany 6,169-173.
Hyde, K.D. 1993c. Fungi from palms. IV Palmicola archonfophoenicis gen. et sp. nav.
Sydowia 45, 15-20.
Hyde, K.D. 1993d. Fungi from palms. VII. The genus Oxydothis from rachides of
palms in north Queensland, including five new species. Sydowia 45, 226-240.

283
Hyde, K.D. 1993e. Fungi from palms. IX. Manokwaria notabilis gen. et sp. nov.
(Ascomycetes) from Irian Jaya and Australia. Sydowia 45, 246-251.
Hyde, K.D. 1994a. Fungi from palms. X. Lockerbia palmicola, a new cleistothecial
genus in the Sordariales. Sydowia 46, 23-28.
Hyde, K.D. 1994b. Fungi from palms. XIV The genus Oxydothis, a revision. Sydowia
46.
Hyde, K.D. 1995a. Tropical Australian freshwater fungi. VII. New genera and species
of ascomycetes. Nova Hedwigia 60, (accepted).
Hyde, K.D. 1995b. Tropical Australian freshwater fungi. VIII. Bertia convolutispora
sp. nov. Nova Hedwigia 60, (accepted).
Hyde, K.D. 1995c. Tropical Australian freshwater fungi. VIII. Vaginatispora aquatica
gen. et. sp. nov. and reflections on ascomata shape as taxonomic criteria. Nova
Hedwigia 60, (accepted).
Hyde, K. D. and J. L. Alcorn. 1993. Some disease-associated microorganisms on
plants of Cape York Peninsula and Torres Strait islands. Australasian Plant
Pathology 22, 73-83.
Hyde, K.D. and J. Frohlich. 1994. Mycosphaerella palmicola associated with leaf spots
of Cocos nucifera in Australia, Irian Jaya and Papua New Guinea. Mycological
Research 99, (accepted).
Hyde, K.D. and K.A. Seifert. 1992. Tropical Australian Freshwater fungi III.
Candelosynnema ranunculosporum, a new genus and species of synnematous
. hyphomycetes. Australian Systematic Botany 5, 401-405.
Kanamaru, T. and H. Okazaki. 1989. Emeriamine: a new inhibitor of long chain
fatty acid oxidation and its antidiabetic activity. Pages: 135-44 In: A.L. Demain,
G.A. Somkuti, J.C. Hunter-Cevera and H.W. Rossmoore (eds.) Novel Microbial
Products for Medicine and Agriculture. Elsevier.
Kim, K.C. 1993. Biodiversity, conservation and inventory: why insects matter.
Biodiversity and Conservation 2, 191-214.
Magurran, A.E. 1988. Ecological diversity and its measurement. Croom Helm,
London.
Matsushima T. 1989. Matsushima Mycological Memoirs No. 6. Publ. by the Author,
Kobe, Japan.
Nisbet, L.J. and F.M. Fox. 1991. The importance of Microbial Biodiversity to
biotechnology. Pages:229-244, In: D.L. Hawksworth (ed.) The Biodiversity of
Microorganisms and Invertebrates: Its role in Sustainable Agriculture. C.A.B.

284
 International, Oxon.
Olembo, R. 1991. Importance of Microorganisms and Invertebrates as Components
of Biodiversity. Pages: 7-16, In: D.L. Hawksworth (ed.) The Biodiversity of
Microorganisms and Invertebrates: Its role in Sustainable Agriculture. C.A.B.
International, Oxon.
Oliver, I. and A.J. Beattie. 1993. A possible method for the rapid assessment of
biodiversity. Conservation Biology 7: 562-568.
Pascoe, I.G. 1990. History of systematic mycology in Australia. Pages: 259-264 In:
P.S. Short (ed.) History of systematic botany in Australia Australian. Systematic
Botany Society, Inc.
Poch, G.P. and J.B. Gloer. 1991. Auranticins A and B: Two new depsidones from a
mangrove isolate of the fungus Preussia aurantiaca. Journal of Natural Products
54,213-217.
Rossman, A. Y. 1994. Fungi as part of the national biological survey. Inoculum 45, 4-
6.
Rodrigues, K. and 0. Petrini. 1995. Biodiversity of endophytic fungi in the tropics.
Abstract In: Fifth International Mycological Congress, Vancouver, Canada, 15-
22 August 1994.
Shearer, C.A. 1993. The freshwater ascomycetes. Nova Hedwigia 56, 1-33.
Simrnonds, J.H. 1966. Host index of plant diseases in Queensland. Queensland
Department of Primary Industries.
Sugano, M., A. Sato, Y. Iijima, T. Oshima, K. Furuya, H. Kuwano, T. Hata and H.
Hanzawa. 1991. Phomactin A: A novel PAF antagonist from a marine fungus
Phoma sp. Journal of the American Chemical Society 113, 5463-5464.

285
 Chapter

20
RESERVE SIZE AND IMPLICATIONS
FOR THE CONSERVATION OF BIODIVERSITY
IN THE ANDAMAN ISLANDS

Priya Davidar 1, Soubadr a Devy l, T. R. K . Yoganand l and T. Ganesh 1

INTRODUCTION
The Andaman and Nicobar islands in the Bay of Bengal are one of the most
biodiversity-rich regions of India, apart from the North-East and the Western Ghats.
There has been a spate of interest in the conservation of the biodiversity of these
islands. As this group, commonly known as the Bay Islands, consisting of more
than 300 islands, stretches from the North Andamans in the North to the Great
Nicobar Island in the south, spanning diverse latitudinal and climatic zones,
conservation and maintenance of the biodiversity of this region requires informa-
tion on species distribution patterns, the status of species, and the adequacy of
the current protected area network (Burley 1988, Jenkins 1988). The task of
conducting an inventory of all the species in an area is daunting, if not impossible,
especially in areas of high diversity such as tropical rain forests. However, pat-
terns of abundance and distribution of select taxa can provide insights for the
protection and management of the biodiversity of an area.

1 Salim Ali School of Ecology and Environmental Sciences, Pondicherry University, Kalapet,
Pondicherry 605 014, India.

287
 The Andaman and Nicobar islands have diverse forest types from giant
evergreen forests to deciduous forests and mangroves (Champion and Seth 1968).
These islands have come under pressure due to heavy commercial timber felling
operations and increasing human impact (Saldanha 1989). To protect the fauna
and flora, a total of 6 National Parks and 94 Sanctuaries have been created from
1977 to 1987. Ninety-three of these are whole islands, usually small in size. Recent-
ly, two more have been added but no information could he obtained on these.
We undertook a survey of butterflies and forest birds in the Andaman group
of islands from 1992-1994, to study patterns of species distributions and status.
Reliable records exist for birds and butterflies of the Andamans and Nicobars
(Evans 1932, Khatri 1989, Ripley and Beehler 1989) but not for most of the other
taxa. Birds and butterflies can be sighted easily, identified and surveyed rapidly
over large areas.

STUDY SITE
The Andaman and Nicobar chain of islands lie off South-East Asia and extend from
South-Western Myanmar to North-Western Sumatra lying between 6o 45’ N and
13o 41’ N latitude (Figure 1) (Srinivasan 1986). They are postulated to be part of the
Arakan Yomas mountain range of Myanmar, which lies submerged. They are true
oceanic islands as they were never connected to the continent during the Pleisto-
cene glaciation and maximum over-water colonization possibly occurred before
the Andaman sea expanded (Ripley and Beehler 1989). The Andaman group consists
of 4 large islands, North, Middle, Baratang and South Andaman Islands forming a
super island of over 5000 km2 in area, surrounded by archipelagoes and isolated
islands: The large islands have extensive human settlements and primary forest
has been reduced by degradation and deforestation. The Nicobar island group is a
chain of smaller scattered islands, the largest of which is Great Nicobar Island. The
Little Andamans lie about 67 km south of the Andaman chain and are separated
from the Nicobars by the 10 degree channel.
The climate is tropical and oceanic with rainfall from both the SW and NE
monsoons. The average annual rainfall is 3000 mm (State Statistical Bureau 1989).
The dry season extends from January to May with another short break in
September-October.
About 38 islands in the Andamans and Nicobars are inhabited and the human
population, which was 279,111 in the 1991 census (Census of India 1991), has more
than doubled since 1971 (115,133). Most of the population growth is due to im-

288
migration from mainland India, increasing the encroachments into forested land.
The 6 tribal groups form 12% of the population. There are 5 tribal reserves which
also serve as protected forest areas, and which cover a total area of 1069 km 2 . There
is a proposal to create two Biosphere Reserves, one in the North Andamans and
another on Great Nicobar Island (Pande et al. 1991).
Commercial forestry operations in the islands started around 1853 (Saldanha
1989). Over time the volume of timber extracted has increased from 49,000 m 3 in
1949 to 145,000 m 3 in 1986 (Saldanha 1989). The number of species exploited has
increased from 20 to 40 over this period. Almost 12 % of the land has been clearfelled
(Whitaker 1985). Since 1989, clearfelling has been stopped and over time wood
extraction for wood based industries will be phased out and expansion of current
industries banned. The level of extraction has also been reduced from 150,000 m 3 to
100,000 m3 per year (Sinha 1991).
These islands support a diverse forest vegetation (Champion and Seth 1968)
and taxa such as birds, mammals and plants have been well documented (Ripley
and Beehler 1989, Abdulali 1965, 1981, Rao 1986). However, little comprehensive
information exists on other taxa. The flora and fauna show affinities with
Myanmar, the Malay peninsula and the Indian subcontinent. Over 1416 species of
flowering plants and 120 species of pteridophytes have been recorded (Rao 1986).
About 187 species of flowering plants are endemic. There are almost twice as many
flowering plant species in the Andamans group as compared with the Nicobars.
Both groups share only about 28% of the angiosperm flora (Table 1). The Andaman
flora shows general affinities with S. E. Asia and that of the Nicobars with Malesia
(Rao 1986).
The Andaman butterflies were first recorded by Evans (1932). Ferrar (1951) listed
268 species from 9 families. Certain families were revised by Khatri (1989) who has
listed 270 species from 6 families. Of these, 150 were recorded in the Andaman
Islands, 84 in the Nicobars and 36 are common to both island groups (Table 1).
Ripley and Beehler (1989) did an analysis of the breeding birds of the Andaman
and Nicobar islands. They identified 104 species of breeding birds of which 92 spe-
cies are found in the Andaman group and 65 in the Nicobar group. There are 13
endemic species of birds and 86 endemic races. The Andamans and Nicobars share
53 species. The Nicobar avifauna appears to be a subset of the Andaman birds and
both are most closely allied with that of south-western Myanmar and the Malay
peninsula. Certain taxa are poorly represented whereas others are common. For
instance hawks, herons pigeons and kingfishers are well represented whereas the

290
passerines are an impoverished group (Ripley and Beehler 1989). The data on other
taxa are unreliable, however, a total of 10 species of amphibians and 83 species of
reptiles have been documented (Table 1).

NUMBER OF SPECIES
A N TOTAL E

PLANTS 1079 770 1416 187 Rao 1986

BUTTERFLIES 150 84 270 ? Khatri 1989

AMPHIBIANS 10* 2* Rao 1989

REPTILES 83* 23* Rao 1989

BIRDS 92 65 104 13 Ripley & Beehler 1989

MAMMALS# 55* 33* Rao 1989

Table 1: Number of Species of some Taxa with Endemics in the Andaman and Nicobar
Islands.

* = includes subspecies # = includes marine mammals

A = Andamans N = Nicobars E = Endemics

METHODS
In order to study the distributions of birds and butterflies, sites were selected in
different habitats in the main large islands and on islands of different sizes lying off
the main islands, to represent a north-south gradient. A total of 45 islands were
surveyed (Table 2). These include the large islands such as North Andaman Island,
Baratang, South Andaman Island, Rutland and the Little Andamans. Islands of
different sizes in the Labyrinth archipelago, Ritchie’s archipelago and other
islands lying off the Andaman ‘super-island’ were included in the study. The sur-

291
veys were conducted in the dry seasons from February to May, 1992, February to
May, 1993 and February, 1994 when birds and butterflies are more active and more
visible. The South Andamans, Labyrinth archipelago and the Little Andamans were
surveyed from February to May, 1992 for birds and butterflies. Baratang, the Ritchie’s
archipelago and seven islands off the North Andamans were surveyed for birds
from February to May, 1993, and the North Andaman island and 11 associated
islands were surveyed for birds and butterflies in February, 1994.
Birds were surveyed in all 45 islands, and butterflies in 12 islands in the North
Andamans, 12 islands in the South Andamans and in Little Andaman Island (Table
2). The different habitats encountered in each island were identified with the help
of Champion and Seth (1968) and classified as:

1. Evergreen forest; multistoreyed climax forest formations that occur mostly on

low alluvial land or on moist loamy hillsides with representative trees such as
Dipterocarpus spp., Canarium manii, Artocarpus spp. and Pometia pinnata.
2. Semi-evergreen forest; containing both evergreen and deciduous trees. Some
tree species are Dipterocarpus alatus, Pterygota alata, Albizia chinensis, Bombax
insigne, Artocarpus lakoocha and Pterocymbium tinctorium.
3. Deciduous forest; forests of lower stature growing on lower hills and in drier
areas. Common species are Pterocarpus dalbergioides, Terminalia bialata, Dalbergia
spp., Pterocymbium tinctorium, Albizia spp., and Tetrameles nudiflora.
4. Littoral forest dominated by Manilkara littoralis.
5. Disturbed areas and edges composed of secondary moist deciduous forests
resulting from selective felling with trees such Canarium euphyllum, Pterocymbium
tinctorium and Salmalia insignis, and the transition zone between clearings and
forest, invaded by weeds such as Chromoleana odorata and Lantana camara.

Forest birds in the Andaman group of islands were selected from the list of
breeding birds compiled by Ripley and Beehler (1989). A total of 47 species of
birds were selected out of the 104 listed. The herons, rails, ducks, brahminy kite
and white-bellied sea eagle were not included. In addition, kingfishers, swifts
and swallows, nocturnal species and birds of open fields were eliminated from
the study. The Narcondam hornbill was not included because it was not possible
to visit Narcondam Island.

292
Island Area S-E DC L DT BR BT
(km2)
NORTH ANDAMANS
NORTHANDAMAN @ 1000 + + + + +
LANDFALL 13 - + - + +
SOUND 12.7 - + + +
PAGET 4 - + - + -
NORTH REEF 3.4 - + - + -
EAST 3 - + + + +
POINT 0.8 - + - + -
REEF 0.6 - + + -
DELGARNO 0.5 - + + +
EXCELSIOR 0.4 - + + -
ROSS 0.3 + + + +
POCOCK 0.25 - + + +
AVES 0.25 - + + +
TURTLE 0.13 - + + +
CURLEW 0.07 - - + + +
TEMPLE 0.06 - + + + +
EGG 0.06 - + + + +

MIDDLE ANDAMANS AND BARATANG

BARATANG 230 + + + -
HAVELOCK 92 + + + -
JOHNLAWRENCE 35 + + + -
PEEL 23 + + + -
LONG 14 + + + -
WILSON 14 + + + -
NORTHPASSAGE 13 + + + -
NEIL 12.6 + + + -
NICHOLSON 1.8 + + + -
INGLIS 1.4 - + + -
GUITAR 1 - + + -
HUGH ROSE 0.6 - + + -
MIDDLE BUTTON 0.4 + + + -
NORTH BUTTON 0.25 - - + + -

293
Island Area E S-E DC L DT BR BT
(km2)

S
SOUTHANDAMAN 1348 + + + + + + +
RUTLAND 116 + + + + + + +
TARMUGLI 11.5 + + + + - + +
ALEXANDRIA 3.6 + + - + - + +
REDSKLN 3.3 + + - + - + +
NORTH CINQUE 1.6 - - + + - + +
MALAY 0.7 + - - + + + +
TWINS 0.44 - - + + - + -
ROSS 0.28 - - - - + + +
SNOB 0.22 - + - + - + +
JOLLY BUOY 0.12 - - - + + + +
CHESTER 0.09 + - - + - + +
GRUB 0.03 - - + - - + +

LTTTLEANDAMAN 675 + + + + + + +

Table 2: Islands Surveyed and Habitat Characteristics.

E = Evergreen, S-E = Semi-Evergreen, DC = Deciduous, L = Littoral, DT = Disturbed,
BR = Birds, BT = Butterflies.

The available habitat types in a site or island were classified into the above 5
categories. Transects of one km length were selected in different habitat types on
large islands or across habitats on smaller islands. The number of such transects
varied with the size of the habitat surveyed. The transects were walked in the
mornings between 7 am to 10 am, and all birds seen and heard were recorded and
identified using Ali and Ripley (1987) and King et al. (1975). Their distance along
the transect line and approximate perpendicular distance to the transect line were
noted. From this information species lists for each site and for each island were
prepared, together with the relative abundances of all species. The bird species
recorded on all the above mentioned sites and islands were ranked in ascending
order according to the number of islands on which each was recorded. These ranks
were transformed into an index of between l-5. Birds recorded from l-10 islands

294
were assigned a rank of l, ll-20 islands, a rank of 2 and up to 41-45, a rank of 5.
The data on relative abundances were obtained from a total of 5028 birds on
70 km length of transects.
Individual butterflies were recorded 5 m on either side of variable length tran-
sects (Pollard 1977). The length of the transect depended on the size of each habi-
tat. On smaller islands the transects cut across all habitats. Some butterflies were
collected for later identification.
Sampling was repeated for several days depending on the size of the habitat or
island and was discontinued when no new species were encountered.
Butterflies were identified to species, genus or family with the help of Wynter-
Blyth (1957), Khatri (1989), and voucher specimens kept at the Zoological Survey
of India office in Port Blair. Despite this, many specimens could not be identified.

RESULTS
There appears to be a north-south gradient in forest types from the North to the
South Andamans (Table 2). The vegetation of the North Andamans is predominant-
ly deciduous except for patches of evergreen forest on the main North Andaman
island and a small patch of semi-evergreen forest on Ross. The other islands had
deciduous vegetation. The proportion of evergreen and semi-evergreen forests
increases on smaller islands off the Middle Andamans and Baratang and are found
on even very small islands off the South Andamans. This probably reflects a north-
south rainfall gradient, with drier conditions in the north (Ellis 1989). Very small,
disturbed or very isolated islands had a secondary type of vegetation.
Of the 47 species of birds surveyed, all were recorded on South Andaman Island
and 43 on North Andaman Island. On islands < 5 km2 a total of 39 species were
recorded on islands off South Andaman Island and 32 on islands off North
Andaman Island (Table 3). The species missing from small islands off the North
Andamans were Terpsiphone paradisi, Dendrocitta bayleyi, Columba palumboides,
Copsychus malabaricus, Cuculus micropteris, Euystomus orientalis and Dryocopus
javensis. These species have restricted distributions on large islands or occur at
low densities on islands with evergreen forests (Davidar et al. unpublished).
Of the 65 species of butterflies recorded in the survey, 44 were found on South
Andaman island and 40 on North Andaman Island. On small islands (< 5 km2),
there were 43 species in islands off the South Andamans and 22 on islands off the
North Andamans. North Andaman Island recorded 4 species of evergreen specia-
lists, whereas there were 8 on South Andaman island. The small islands off the

295
 NORTH ANDAMAN SOUTH ANDAMAN
Main Island Islands Main Island Islands
<5KM2 <5KM2
BIRDS 43 32 47 39
BUTTERFLIES 40 22 44 43
EVERGREEN 4 0 8 7
SPECIALIST
BUTTERFLIES

Table 3: Comparison of Bird and Butterfly Species Richness in the North and South
Andamans.

North Andamans had no evergreen specialist butterflies whereas off the South
Andamans there were 7 species of evergreen specialists (Table 3).
From the data on relative abundances, the status of endemic forest species were
determined and rare species identified. Of the endemic forest species found in the
Andamans group, Dendrocitta bayleyi and Columba palumboides were found only
on the larger islands. Macropygia rufipennis occurred at low densities and the other
4 species were common (Table 4; Davidar et al. unpublished). Species that are very
rare are Coracina striata, Oriolus xanthornus, Chalcites xanthorhynchus and Terpsiphone
paradisi. These have restricted distributions on the large islands (Davidar et al.
unpublished).
Of the existing National Parks and Sanctuaries, 58 reserves are less than 1 km 2 ,
and 13 are less than 0.01 km 2. Only 4 reserves are greater than 30 km 2 in area
(Figure 2). A mean of 9 species of birds was found on islands < 1 km 2 , 24 on islands
l- 5 km 2 , 33 on islands 5-30 km2 and almost all species were recorded on islands
> 30 km2 . When butterflies are compared, islands < 1 km2 have a mean of 6 species,
l-5 km2 a mean of 16 species and 35 species on islands > 30 km2 (Figure 2). Many
species of birds and butterflies were not recorded on smaller islands. For birds
there is a gradual drop off in the total number of species recorded with island size,
whereas for butterflies, the pattern is not clear except for islands < 0.01 km 2 in area.
Many species of butterflies found on smaller islands were not recorded on the main
North Andaman and South Andaman island, and certain species were restricted to
particular islands (Table 5; Soubadra Devy et al. unpublished).

296
 SPECIES STATUS REMARKS

EMICS
Dendrocitta bayleyi Rare Restricted to large islands
Columba palumboides Uncommon Restricted to large islands
Macropygia rufipennis Uncommon Occurs at low densities
Centropus andamanensis Common
Dicrurus andantanensis Common
Spilornis elgini Common
Sturnus eythropygius Common

RARE FOREST BIRDS

Coracina striata
Oriolus xanthornus
Chalcites xanthorhynkhus
Terpsiphone paradisi

Table 4: Status of Endemic Forest Birds and Rare Species in the Andaman Islands.

NUMBER OF SPECIES
BIRDS BUTTERFLIES
(n*=47) (n*=65)
NORTH ANDAMAN ISLAND 43 40
SOUTH ANDAMAN ISLAND 47 44
LITTLE ANDAMANS 41 21
ISLANDS > 30 km2 47 57
ISLANDS > 5< 30km 2 44 31
43 40
ISLANDS < 1 km2 36 39
ISLANDS < 0.1 km2 20 21

Table 5: Total Number of Forest Birds and Butterfly Species Found in Islands of Particular
Sizes.
n = total number of species recorded

297
MacArthur and Wilson (1967), where the number of species on an island is related
to the size of the island and its degree of isolation from the mainland. The mean
number of forest bird species drops with island size. However, the distributional
pattern did not appear to be random as many species were not recorded at all on
small islands and all forest bird species were recorded only on islands larger than 30 km2.

The majority of reserves in the Andaman and Nicobar islands are whole islands
less than 1 km in area. The consequences of having many small reserves for forest
2

species is that not all the vegetation formations are included. Small islands usually
have drier forests which do not support habitat specialists. Large islands which
have sizable extent of evergreen forests are under severe human pressure. This re-
sults in an inadequate protection of the biodiversity of this region.
The current rate of conversion of evergreen forests to a secondary deciduous
type by selective felling and regeneration of commercially useful species will lead
to the extinction of many habitat specialist butterflies, and will probably influence
bird diversity as well, by reducing the quantity of fruits and other resources availa-
ble. Large islands and forests on large islands should be notified as reserves and
protected on a priority basis. There should also be adequate infrastructure to en-
force protection. The legal status of many of these reserves is unclear as the legal
procedures have not been completed in most cases. In many cases, there are no
staff, funds and equipment to manage these sanctuaries and National Parks
(Pande et al. 1991).
Islands are fragile ecosystems with high extinction rates because the species
are confined to small areas and suffer disproportionately from habitat-destruction
and other factors. Many of the species occurring in the Andaman and Nicobar
islands, particularly those belonging to obscure taxa are yet to be documented. It
is important that they are protected before it is too late.

ACKNOWLEDGMENTS
This project was funded by a grant from the Ministry of Environment, France. We
are grateful to the Forest Department of the Andaman and Nicobar islands for
permission to carry out this survey, and providing advice, logistical support and
unstinting help at all times. The interest and enthusiasm of many of the Forest staff
will not be forgotten. We thank Dr. J. M. Thiollay for contributing to the survey. Mr.
Ramana Athreya identified many butterfly specimens. We acknowledge the help of
Dr. J. P Puyravaud in correcting this manuscript.

299
REFERENCES
Abdulali, H. 1965. The birds of the Andaman and Nicobar islands. J. Bombay Nat.
Hist. Soc. 61: 483-571.
Abdulali, H. 1981. Additional notes on Andaman birds. J. Bombay Nat. Hist. Soc.
78: 46-49.
Ali, S. and S.D. Ripley. 1987. Handbook of the Birds of India and Pakistan, Compact
2nd Edn. Oxford University Press, New Delhi.
Burley, F.W. 1988. Monitoring biological diversity for setting priorities in
conservation. Pages 227-230 In: Wilson, E.O. (ed.). Biodiversity National
Academy Press, Washington D.C.
Census of India. 1991. Series I. India. Provisional Population Totals; Ministry of
Home Affairs, New Delhi.
Champion, H.G. and S.K. Seth. 1968. A Revised Survey of the Forest types of India.
Government of India Press, Delhi.
Ellis, J.L. 1989. Project Document of North Andaman Biosphere Reserve in
Andamans. Botanical Survey of India, Port Blair.
Evans, W.H. 1932. The Identification of Indian butterflies. Bombay Nat. Hist. Soc.,
Bombay. l-454.
Ferrar, M.L. 1951. On the butterflies of the Andaman and Nicobar islands. J. Bombay.
Nat. Hist. Soc. 47: 470-491.
Jenkins, R.E. 1988. Information management for the conservation of biodiversity.
Pages 231-239 In: Wilson, E. O. (ed.). Biodiversity. National Academy Press,
Washington D. C.
Khatri, T.C. 1989. A revised list of butterflies (Rhopalocera: Lepidoptera) from Bay
islands. J. Andaman Sci. Assoc 5: 57-61.
King, B., M. Woodcock and E.C. Dickinson. 1975. Collins Field Guide to the Birds of
South-East Asia. The Stephen Greene Press, Lexington, MA.
McArthur, R.H. and E.O. Wilson. 1967. The Theory of Island Biogeography. Princeton
University Press, NJ, USA.
Pande, P., A. Kothari and S. Singh (eds.). 1991. Directory of National Parks and
Sanctuaries in Andaman and Nicobar islands. New Delhi: IIPA.
Pollard, E. 1977. A method of assessing changes in the abundance of butterflies
Biol. Conserv. 12: 115-124.
Rao, M.K.V. 1986. A preliminary report on the angiosperms of Andaman-Nicobar
islands. J. Econ. Bot. 8: 07-184.

300
Ripley, S.D. and B. M. Beehler. 1989. Omithogeographic affinities of the Andaman
and Nicobar islands. J. Biogeogr. 16: 323-332.
Saldanha, C. J. 1989. Andaman, Nicobar and Lakshadweep. New Delhi, India: Oxford
and IBH.
Sinha, B.P. 1991. Forestry in Andaman and Nicobar islands, Sunday Observer.
Srinivasan, M.S. 1986. Geology of Andaman and Nicobar islands. J. Andam. Sci
Assn. 2: l-12.
State Statistical Bureau, Andaman and Nicobar Administration. 1989. Handbook
on Andamans district. Govt. Press, Port Blair.
Whitaker, R. 1985. Endangered Andamans. Environmental Services Group, WWF-
India, and MAB India, Department of Environment, New Delhi.
Wynter-Blyth, W.M. 1957. Butterflies of the Indian Region. Bombay Natural History
Society. Bombay.

301
 Chapter

21
METHODS FOR MEASUREMENT
OF SPECIES DIVERSITY

Jiragorn Gajaseni1 and Kansri Boonpragob2

There are numerous definitions of biodiversity. Most treat diversity at genetic,

species or ecosystem level. Current measures select different levels of the bio-
system for emphasis; the species, population, ecosystem, or landscape levels.
Species diversity measures can be divided into 3 main categories (Magurran
1988):
1. species richness indices,
2. species abundance models, and
3. indices based on the proportional abundance of species.

SPECIES RICHNESS INDICES

These indices are essentially a measure of the number of species in a defined
sampling unit. If the study areas can be successfully delimited in space and time,

1 Department of Biology, Faculty of Science, Chulalongkom University, Bangkok 10330, Thailand.

2 Department of Biology, Faculty of Science, Ramkhamhaeng University, Bangkok 10240, Thailand.

303
and the constituent species enumerated and identified, species richness provides
an extremely useful measure of species diversity. If, however, a sample rather
than a complete catalogue of species in the community is obtained, it becomes
necessary to distinguish between numerical species richness, which is defined as
the number of species per specified number of individuals or biomass, and species
density, which is the number of species per specified collection area. It is not always
possible to ensure that all sample sizes are equal and the number of species inva-
riably increases with sample size and sampling effort. Hurlbert (1971) modified the
technique called “rarefaction” for estimating the unbiased number of species
expected in each sample if all samples were of a standard size.
Species richness measures have great intuitive appeal so long as care is taken
with sample size. Species richness provides an instantly comprehensible expression
of diversity. However the great range of diversity indices, and models which go
beyond species richness, is evidence of the importance that many ecologists place
on information about the relative abundance of species.

SPECIES ABUNDANCE MODELS

There is no community in which all species would be equally common. Instead,
it is typically the situation that a few species are very abundant, some have me-
dium abundance, while most are represented by only a few individuals. These
common community patterns lead to the development of species abundance
models. The species abundance models are usually classified into 4 models: log
normal distribution, the geometric series, the logarithmic series, and MacArther’s
broken stick model (Magurran 1988). The diversity of a community may therefore
be described by referring to the model which provides the closest fit to the ob-
served pattern of species abundance.
These models describe the distribution of species abundance. Species abun-
dance models range from those which represent situations where there is high
evenness to those which characterize cases where the abundances of species are
very unequal. While the species abundance models provide the fullest descrip-
tion of diversity data, they are dependent on some fairly tedious model fitting
and for rapid calculation require the use of computers. In addition, problems may
arise if all the communities studied do not fit one model and it is desired to com-
pare them by means of a diversity index.

304
INDICES BASED ON THE PROPORTIONAL ABUNDANCES
OF SPECIES
Indices based on the proportional abundances of species provide an alternative
approach to the measurement of diversity. This type of diversity measure has en-
joyed a great deal of popularity in recent years. The most widely used indices are
Shannon’s index of diversity and Simpson’s index. Shannon’s index of diversity
is a useful method for comparing the diversity of different habitats, especially when
a number of replicates have been taken (Gajaseni and Gajaseni, unpublished data).
When the randomness of a sample cannot be guaranteed as, for instance, during
light trapping where different species of insects are differentially attracted to light,
or if the community is completely censused with every individual accounted for,
Brillouin’s index is the appropriate form of the diversity index (Pielou 1969, Pielou
1975).

DISCUSSION
The loss of biodiversity is one of the most profound global crises. Even though we
may still disagree on the definition of biodiversity or how to measure biodiversity
(Hurlbert 1971), there is unanimous agreement that biodiversity is being reduced
at an accelerating rate (Wilson 1988). The vast majority of the past and current efforts
to preserve biodiversity have focused upon species. Species inventory, mainly by
listing names, has been the most common measure. This old paradigm on the
methods for measurement of species diversity might satisfy one of the most
fundamental questions in biology " How many kinds of living thing are there?” The
answer to this question is still a matter of guess work. We have so far succeeded in
naming and describing only a very fraction of the total number of species present.
There are many satisfactory methods and statistical analyses to measure both
richness and equitability (evenness) of species diversity. However, one of the most
famous questions on species diversity still remains. The question was asked by a
very famous ecologist, the late G. Evelyn Hutchinson, in 1959, who asked: “Why
are there so many kinds of animals ?" Disregarding political values and ethical issues,
the quality of inventory data and their practicality will determine whether there
is sufficient quantity of information to serve any useful purpose (Renner and
Ricklefs 1994).
Any rationalized strategy for biodiversity conservation must be based on
information. To set aside conservation areas that will protect the fullest range of
species requires more complete knowledge of the distribution and abundance of

305
organisms than is currently available (Lubchenco et al. 1991). Uncertainties about
the number of species that exist, the rate at which the number is being eroded, and
the proportion of species threatened with extinction stress the need for alternative
approaches to the maintenance of biodiversity. While we are fundamentally igno-
rant as to whether or not any conservation area is a self-sustaining ecosystem, we
cannot even come close to attaining our goal of preserving biodiversity, if we
continue to focus our efforts primarily on species. Conserving species is not the
total solution. Efforts to preserve biodiversity must focus increasingly at the
ecosystem level because of the immense number of species, the majority of which
are currently unknown (Franklin 1993).
Conservation approaches at the levels of ecosystems and landscapes are the
only way to preserve biodiversity (Franklin 1993). Ecosystems have long been
described by traditional “biotic ecosystem models” by considering each species to
operate in fixed trophic levels within constraints imposed by the physical
environment (Caswell 1988). Therefore, populations of species at the various tro-
phic levels become the focal point of most studies. The “collect, classify and store
everything found in a given area” approach is the major method. This has limited
utility because biodiversity has been shaped by millions of years of interactions
between speciation and adaptation. Only recently, “functional ecosystem models”
have emerged (Caswell 1988). Under these models, ecosystems are perceived as
being composed of different functional elements through which energy and mate-
rial move through. Organisms are placed into one of two general categories: energy
capture and nutrient retention, and rate regulation. Instead of focusing on each
species, we must conserve the processes that are defined by species’ interactions
within self-sustaining ecosystems (Georgiadis and Balmford 1992). Both approa-
ches have been fruitful in the past, but the functional model of determining
environmental factors that control the rate of energy flow and the movement of
materials within ecosystems now appears to be a more promising approach for the
future. Therefore, taking an ecosystem approach we have to monitor and study
species diversity in order to adequately assess the roles of all types of interaction
between species to maintain “healthy functional ecosystems”.

REFERENCES
Caswell, H. 1988. Theory and models in ecology: A different perspective. Ecological
Modeling 43: 33-44.
Franklin, J.R. 1993. Preserving biodiversity: Species, ecosystems, or landscapes.

306
 Ecological Applications 3(2): 202-205.
Georgiadis, N. and A. Blamford. 1992. The calculus of conserving biological diversity.
Trends in Ecology and Evolution 7(10): 321-322.
Hurlbert, S.H. 1971. The non-concept of species diversity: a critique and alternative
parameters. Ecology 52, 577-586.
Lubchenco, J., et al. 1991. The sustainable biosphere initiative: An ecological research
agenda. Ecology 72: 371-412.
Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University
Press, New Jersey.
Pielou, E.C. 1969. An introduction to mathematical ecology. John Wiley and Sons,
New York.
Pielou, E.C. 1975. Ecological diversity. John Wiley and Sons, New York.
Renner, S.A. and R.E. Ricklefs, 1994. Systematics and biodiversity. Trends in Ecology
and Evolution 9: 78.
Wilson, E.O. 1988. Biodiversity. National Academy Press, Washington, D.C.

307
 Chapter

22
TOOLS TO DIAGNOSE FOREST INTEGRITY;
AN APPRAISAL METHOD SUBSTANTIATED BY
SILVI-STAR ASSESSMENT OF DIVERSITY
AND FOREST STRUCTURE
H. Koopl, H.D. Rijksenl and J. Wind2

INTRODUCTION
In order to evaluate the consequences of the rapid decline of tropical rain forests,
and assess the prospects for restoration of these forests, a functional monitoring
system is required. For proper land allocation for conservation of biological diver-
sity, it is imperative to know the structure of major classes of forest quality, from
almost unimpaired forests to early stages of regeneration. Insight into the ecolo-
gical conditions of different patches composing such classes is also important for
harvesting and conservation management strategies.
Advanced remote sensing techniques can provide substantial information on
the extent of serious forest damage. However, information concerning changes of
forest integrity due to harvesting is hard to obtain. Timber exploitation does in-
crease the scale and frequency of gap dynamics, but in the case of small scale selec-
tive cutting, only a temporary and limited effect on canopy closure may be caused.

1 Research for Forestry and Naturure Research, P.O. Box 23,6700 AA, Wageningen, The Netherlands.
2 DHV Consultants BV.

309
From the air, such a qualitative decline in forest integrity is hard to evaluate. It is
nonetheless important to evaluate the impact of different timber extraction me-
thods and the effects of forest management on the potential for restoration of
biological diversity and functional integrity. Also, such an evaluation technique
can be applied to assess the effectiveness of park protection and buffer zone
programmes.
According to the Concise Oxford Dictionary, integrity is defined as wholeness,
entirety and soundness. A forest is sound when it is able to maintain its structure
at a landscape scale in the face of all regular and incidental natural disturbance
factors, such as storms, flooding or drought. The main criterion for integrity is the
occurrence of all species of organisms and age classes in a particular proportion
of social organisation, as would occur in a natural situation, that is without
human interference (Halle et al. 1978). The loss of particular species is the symp-
tom of poor forest condition (Oldeman and Van der Meer 1988, Karr and Budley
1981).
With increased human induced forest dynamics, biodiversity in terms of numbers
of species may decline, remain the same or even increase (e.g. Swaine and Hall
1983, Abdulhadi et al. 1987). New species may invade the forest and serious shifts
in evenness occur. Pioneer species, that are rare in the almost undisturbed mosaic,
spread in an highly dynamic landscape, while previously common species become
rare or extinct. Species confined to the complex-structured mosaic of mature forest
patches, with a limited percentage of young recently disturbed patches (e.g. pri-
mates and birds, or plant and animal species confined to high air humidity) are
particularly susceptible to extinction.
It is almost impossible to make an inventory of species diversity and their rela-
tive population composition, in order to evaluate ecological integrity for planning
and management (Hommel 1990). It would not only be too laborious, but also
specialized taxonomic knowledge for species identification is inevitably deficient.
Other approaches are needed to rapidly assess forest integrity without detailed
inventories of all major groups of forest species. A diagnostic method for forest
integrity may well be developed analogous to a physician’s diagnosis of the loss
of health in a patient, based on external symptoms indicating the quality of in-
ternal functioning.
Here we have raised the question, whether indicators of change in scale and
frequency of gap dynamics, rather than direct measurements of biodiversity, can
be used to diagnose forest integrity. We hypothesized that rapid simple
measurement of the appropriate set of indicators could result in a feasible,

310
standardized appraisal technique.
The conditions for the development of a Rapid Ecological Assessment (REA)
are:
1) The measurements shall provide relevant data. That is, the data must reflect the
degree of decline of forest integrity or recovery in an objective and repeatable
way.
2) The method must be reliable. That is, it must be constant in its diagnosis.
3) The indicators used shall not be rare: an indicator function implies that parti-
cular life-forms or species must occur commonly in particular patches of the
specific classes of quality.
4) The method must be easily applicable even by people without detailed taxo-
nomic knowledge. This implies that the indicators must be readily recognisa-
ble while the time and effort to gather data on their occurrence shall be opti-
mised in terms of transect number and length.
The REA will be substantiated by SILVI-STAR assessment of diversity and forest
structure in pristine and more and less disturbed sites.

METHODS AND MATERIALS

Procedure of method design
Within the borders of a regional climate, external stress factors, such as wind,
flooding and drought, have determined frequency and extent of gap dynamics and
thus provided the conditions for evolution and present biodiversity (Koop 1981).
However, human activities in and around forests raise the levels of forest dyna-
mics. Therefore, an absolute standard of virgin, unimpaired forest can hardly be
found anywhere in the world. Nevertheless, examples of overuse and decline of
forest structures are common and widespread.
During recent decades, the scale and frequency of human induced forest
dynamics have increased drastically. Former non-mechanised small-scale selective
cutting has changed into large scale mechanised harvesting, especially for indus-
trial use. Due to high population pressure and economic needs, the frequency of
human induced dynamics increased as well. Less time is left for regeneration to
mature, late successional stages, characterized by big sized trees. Early success-
ional stages become predominant.
The natural forest is subject to a particular dynamic range of life-processes

311
from germination to decay. Trees and other organisms may die due to senescence
or calamities and predation. On a landscape scale, climatic factors may cause an
almost regular pattern of local calamities that lead to a shifting mosaic patchwork
of communities of young stages next to older ones. The older or so-called late
successional patches may contain higher numbers of species and therefore contri-
bute most to biodiversity (Budowski 1965, Jacobs 1988). Some species are strictly
dependent on those mature patches while others depend on the complex struc-
ture of the mosaic as a whole.
Human disturbance commonly increases the proportion of younger stages of
forest succession. Although affected by the extent of disturbance in space and time,
the system usually reverts to pioneer stages. The decrease in area of old growth
patches and the loss of connectivity between them causes a marked decline in
biodiversity and hence in forest integrity, which can only be restored in the very
long term.
Our method is based on measuring the proportional representation of different
successional stages in the forest mosaic patchwork compared to the natural mature
mosaic of sites on areas of similar geomorphologic conditions. To characterise the
different successional stages, a literature survey and interviews with experts in the
field of tropical rain forest biology resulted in a list of possible diagnostic features
in addition to relevant parameters discovered during the study itself.
After the identification of three major classes of forest disturbance (undisturbed,
selective cutting, abandoned shifting cultivation) in the field, a preliminary list of
diagnostic features was selected. These features have been tested for their suitabi-
lity to be included in the diagnosis method. Formulae have been developed to rate
forest integrity in a range from 100 to 0, where 100 represents scores of indicators
the same as in the reference plot of undisturbed forest.
For substantiation of the diagnostic method, the detailed forest structure and
complete species composition of both disturbed and undisturbed forest plots are
described according to the SILVI-STAR method (Koop 1989). In the same plots,
different disturbance units were subjected to a rating of forest integrity according
to the new diagnostic method. A disturbance unit is defined and delineated as a
more or less homogeneous patch of forest represented by a transect section that
has been prone to the same kind of disturbance. Disturbance units represented by
a transect length of less than 50 m were ignored, because mean values of presence
become too uncertain if only a restricted number of assessed plot& considered.
To prove the relation of indicators with forest structure, SILVI-STAR side views
of representative parts of the disturbance units were plotted. The abundance of

312
indicators was plotted on the side views and percentages of indicator presence
were calculated for each developmental phase.
For substantiation of the calculation of end scores of the diagnostic method, the
mean indicator scores of the 400 m 2 transect intervals were subjected to detrended
correspondence analyses (DCA) (Hill 1979). The results are compared with the
diagnostic method scores. For one site, in addition to the indicator scores, the
complete species composition of trees having a dbh greater than 5 cm, as well as
recognizable samplings and seedlings in 10 x 10 m 2 intervals along the transects,
was subjected to canonical correspondence analyses (CANOCO) (Ter Braak 1986).
Species were recorded as present or absent. Environmental variables, basal area
(mean value over 300 m 2 ), selective cutting, shifting cultivation, development phase
(metric classification of forest patches according to tree height), the number of
stumps and the crown area index (Koop 1989) were tested for significant
contributions to the variance of species composition. Those variables that
significantly attributed to the explanation of variance were subjected to a de-
trended correspondence analyses.

Study area
The study area lies just below the confluence of the Alas and Ketambe rivers in
Northern Sumatra (3o41’ N,97o39' E) at an altitude of about 350 m above sea level.
The Alas river flows through the Semangko Rift zone, a graben running the full
length of Sumatra. Here the graben is only a few kilometres wide and contains a
number of accumulation terraces (van Beek 1982). The lower slopes of the valley
consist mainly of crystalline metamorphic schists, altering in places with hard
limestones of late Palaeozoic origin (van Beek 1982). The area was selected because
of the known management history of several disturbed sites, in the direct vicinity
of a piece of undisturbed forest of a comparable site type.
From 1971 until 1990 an area from the confluence of the Alas and Ketambe rivers
up to 5 km downstream, especially on the easily accessible accumulation terraces
and lower slopes just inside the Gunung Leuser National Park, has been illegally
cleared. Surrounding these cleared fields, the forest has been selectively cut for
timber, mainly in 1985-1986. However, on several occasions, in 1974, 1984, and 1991,
the settlers were forced to leave their illegally occupied land and since then these
fields have been left unexploited. Forest on similar site types that supposedly has
not been disturbed to any significant extent can be found directly below the
confluence near the Ketambe research station in a piece of forest of about 200 ha
(Rijksen 1978).

313
 For an undisturbed reference plot of the terrace site type, a 10 ha topo-unit
belonging to “terrace 4” (van Schaik and Mirmanto 1985), 200 m south of the junc-
tion of the Ketambe and Alas rivers, was selected - from here on referred to as
“Ketambe 1”. Different disturbance regimes were studied along a 750 m transect
situated on the same terrace as the reference plot, 5 km downsteam from the
confluence. For an undisturbed reference plot of the lower slope, a 10 ha topo-unit
about 1200 m south-south east of the confluence was selected - from here on
referred to as “Ketambe 2”. For different disturbance regimes, three 70 to 130 m
long transects were selected on the lower slopes of the valley, just opposite the
reference plot and respectively, 2 and 3 km downstream from the confluence.

RESULTS
Description of the diagnostic method; indicator groups
Three groups of indicators have been developed, consisting of forest structure
indicators, light indicators and moisture indicators. The indicators that have been
used in the method are listed below:

(a) Forest structure indicators. Indicators that define reduced integrity include:
Lower basal area
Basal area can be estimated by a simple and quick distance independent
method called the Bitterlich method (Bitterlich 1948, Kramer and Akca 1982).
All trees which are viewed to be thicker than the thumb held at arm’s length are
counted. Depending on the thickness of the thumb and the distance between
thumb and eye, a conversion factor can be determined. Multiplying the
conversion factor with the number of trees thicker than the thumb gives an
estimate of the basal area. In the case of relative comparisons, only the number
of counted trees is used.
Presence of big trees (Smiet 1989).
Trees with a diameter greater than 50 and 100 cm are counted.
Maximum tree height (Budowski 1965).
Tree height is estimated in ten-meter classes.
A distinct layered structure (Budowski 1965, Jacobs 1988).
Young and old secondary forests have a single or double layer, while late
successional stages have a more multilayered structure. Therefore three forest
structure classes are recorded: one (1), two (2) or multilayered (3).

314
 Characteristic diameter distributions (Koop 1989).
Diameter distribution patterns in secondary forest tend to have a characteristic
distribution curve, which differs from the reverse J-shaped curve of undisturbed
forest. Therefore a choice has to be made between a secondary (1) or a reverse J-
curve (2).

Light indicators. Indicators for reduced integrity in this group include easily
recognizable empirically selected species groups or families of species that
indicate high light intensities.
Indicative groups of pioneer tree species
The Euphorbiaceae: Cecropia, Ochroma and Trema in South America (Budowski
1965) and Macaranga, Mallotus and Trema spp. in South East Asia are typical
pioneers. The number of stems of these tree species, that are well known by
local inhabitants, must be counted.
Light demanding species or groups of species
Species such as grasses less than 1 m (Graminaceae) (Budowski 1965), big species
of ginger, taller than 2 m (Zingiber spp.) (this study), big ferns taller than 1 m (in
this study Nephrolepis biserrata, Diplazium esculenturn, Pteridium aquilinum,
Christella papilio, Dicranopteris curranii, Cyathea borneensis) (Budowski 1965), and
herbaceous lianas e.g. Convulvulaceae in secondary forest (Budowski 1965) were
recorded as present or absent.
Light demanding exotic invader species
Species such as Lantana camara, Piper aduncum, Chromolena odorata are noted
separately because they generally indicate severe disturbance over longer
time periods. They tend to dominate the vegetation thus hindering natural
regeneration of other tree species. Only the presence of these species has to be
noted.

(c) Atmospheric moisture indicators. This indicator group consists of easily

recognizable species groups or families of species that are common and
widespread, that indicate high air humidity.
- Epiphytic ferns that grow lower than 5 m above ground level on small trees
and lianas (e.g. Antrophium spp.)
- Epiphytic filmy ferns on small trees and lianas (HymenophylIaceae, e.g.
Trichomanes spp., Hymenophyllum spp.)
- Epiphytic mosses (Richards 1984 and Pots 1982) of the feather type - in this

315
 study e.g. Himanthocladium plumula (Nees), Homaliodendron flabellatum (J.E.
SMN.) Fleisch., Pinnatella mucronata (Bosch and Lat.) Fleisch.; the hanging
type e.g. Neckeropsis gracilenta (Lat.) Fleisch.; and the ramicolous type, e.g.
Calyptothecium recurvulum (C. Müll.), Floribundaria pseudofloribunda Fleisch.
Floribundaria floribunda (Dozy and Molk.) Fleisch.
- Epiphyllous mosses on leaves -in this study e.g. Hepaticae (Richards 1984).
- The upper limit of the moss carpet on the boles of trees referred to as the
moss line by Richards (1984).
- Presence of “bole climbers”, being herbaceous species that stick with their
leaves to the tree bole (Oldeman 1978, Budowski 1965).
Only the presence of these growth forms has to be noted.
Light and atmospheric moisture indicators were assessed in adjacent 10 m x 10 m
blocks in a line transect according to their presence or absence. Only for coverage
of the herbaceous layer was the decimal scale of Londo (1984) used, and for pio-
neer species stem numbers were counted. Forest structure indicators were assessed
every 50 metres along the transect base line.
Besides the three groups of indicators mentioned above, direct indicators of
disturbance by former land use were assessed. They are not used for the calcula-
tion of end scores but indicate the nature of the disturbance.
- number of stumps (Smiet 1989) (N)
- presence of charcoal, burnt stumps or logs (+ or -)
- number of timber tree species of a Dbh more than 25 cm (N); most of the
time being primary forest tree species e.g. Dipterocarpaceae
- number of commercially valuable rattan species (N)
- number of planted exotic trees (N)
- presence of paths (Smiet 1989) (+ or -)
- presence of sawah dikes (+ or -)
Indicators that could not be used for the rapid appraisal method include indica-
tors related to direct measurement of area of different phases of regeneration, such
as innovation and aggrading gap phases (Oldeman and van der Meer 1988),
because mapping proved to be too time consuming. Other indicators could not
objectively be determined in the Sumatran rain forest, for example:
- crown shape -uniform, open and bright green in secondary forest, com-
pared with dark green and highly varying crown shape in primary forest

316
 (Budowski 1965)
- big leaves in secondary forest, compared with small leaves in primary forest
(Budowski 1965)
- abundance of big fruits and seeds in primary forest (Jacobs 1988).
For the undisturbed reference plot and for each disturbance unit, the mean values
of forest structure indicators and percentage presence of atmospheric moisture
indicators and light indicators were calculated. Final rating per disturbance unit is
achieved by scoring the indicators relative to the standard of the undisturbed
reference as follows:

S=D/R x 100
where:
S= indicator Score relative to reference standard
D= value of indicator in Disturbance unit
R= value of indicator in Reference

Thus the forest structure scores of estimated basal area, maximum tree height,
numbers of trees greater than 50 and 100 cm dbh, and the score for canopy layer-
ing and diameter distribution were calculated in this way. Atmospheric moisture
indicators were rated similarly, except that the direct indicator values were
substituted by the percentage frequency of an atmospheric moisture indicator in
the disturbance unit. Thus the scores of percentage presence of mosses on leaves,
bole climbers, epiphytic filmy ferns and epiphytic ferns less than 5 m above ground
level, and ramicolous, feather and hanging mosses were calculated.
Because light indicators increase with decreasing forest integrity, a reverse value
has to be calculated to get similar scores from 100 to 0 with increasingly disturbed
forest. Therefore, instead of the herbaceous cover (Cov) itself, assessed according
the 1 to 9 scale of Londo (1984), the inverse (10 - Cov) was used. This inverse value
decreases with decreasing forest integrity and thus behaves like indicators of
moisture and forest structure.
For other light indicators assessed as frequency of presence, the score is calcu-
lated differently. Their frequency in the undisturbed reference plot (R%) is regarded
as the standard. The difference between R% and the percentage in a disturbed forest
(D%) is then rate against the range that is left between R% and the maximum of
100% presence of light indicators, i.e. 100 - R%. Because this ratio (S’) still increases
with decreasing forest integrity, the inverse (100 - S’) is taken. Thus the new Score

317
This group of light indicators is confined to the highest light intensities and is only
rarely found in innovation gap phases. There it is represented by only a few
individuals. In the 10 ha phase map of Ketambe 1, only 5, and in Ketambe 2, only 3
such spots were encountered, but these spots were not traversed by the 1 km
diagnostic transects.
Atmospheric moisture indicators in the Ketambe 1 plot are absent in the
innovation gap phase and are most frequent in the tree phase. Differences in scores
between aggrading gap and aggrading tree are small, except for the light indicating
pioneer trees and herbaceous lianas.

Innovation Aggrading Aggrading Tree

Gap phase Gap phase tree phase phase
H<2m 2m<H<l0m l0m<H<20m H>20m

HERB_ COV 6.5 1.9 1.8 1.9

PIONEERT 100 % 71 % 33 % 13 %
LIGHTEXO 0% 0% 0% 0%
GRAMINEA 0% 0% 0% 0%
BGINGERS 0% 0% 6% 0%
BIGFERNS 50 % 14 % 13 % 0%
HERBLIAN 100 % 86 % 47 % 11 %

WOODLIAN 2.0 5.9 6.1 8.5

MOSSLINE 1.0 m 3.1 m 4.6 m 6.2 m
EPIFFERN 0% 71 % 66 % 81 %
FILMFERN 0% 36 % 47 % 49 %
RAMICMOS 0% 100 % 100 % 100 %
LEAF_MOS 0% 64 % 66 % 84 %
BOLECLIM 0% 7% 7% 22 %

Table 4: Percentages of light and atmospheric moisture indicator presence and mean
values of indicator values in the developmental phases in the 1000 m transect for the terrace
reference plot, Ketambe 1.

Atmospheric moisture indicators have lower frequencies in reference plot

Ketambe 2, while the average mossline height is much lower than in Ketambe l (2.2
against 5.2 m).

322
 The results of ordination of samples and indicators for sites on the lower slopes
(Figures 4 and 5) show similar results to Table 2. The first axis explains 55 % of the
variance (eigenvalue=0.538) and the second axis 7.5 % of the variance
(eigenvalue=0.078) and is correlated with the elevation of the plots. Plots on the

Innovation Aggrading Aggrading Tree

Gap phase Gap phase tree phase phase
H<2m 2m<H<l0m l0m<H<20m H>20m

HERB-COV 1.0 0.5 1.6 l.3

PIONEERT 0% 25 % 27 % 2 %
LIGHTEXO 0% 0% 0% 0 %
GRAMINEA 0% 0% 0% 0 %
BGKNGERS 0% 0% 0% 0 %
BIGFERNS 0% 25 % 0% 2 %
HERBLIAN 0% 0% 0% 0 %

WOODLIAN 2 3.5 4.5 5.8

MOSSLINE 2m 2.25 m 2.5 m 2.1 m
EPIFFEJXN 100 % 25 % 75 % 26 %
FILMFERN 0% 25 % 27 % 17 %
RAMICMOS 100 % 100 % 100 % 93 %
LEAF-MOS 0% 0% 0% 8%
BOLECLIM 0% 25 % 9% 11 %

Table 5: Percentages of light and atmospheric moisture indicator presence and mean
values of indicator values in the developmental phases in the 1000 m transect in the lower
slope reference plot, Ketambe 2. The number of plots for calculating the percentages of
Innovation and Aggrading plots are 4 and 11, and therefore give inaccurate outcomes.

ridge have fewer atmospheric moisture indicators than plots in the valley.
The correspondence analyses of the environmental variables, basal area (mean
value over 300 m 2 ), selective cutting, shifting cultivation and developmental phase
(metric classification of forest patches according tree height), the number of stumps
and the crown area index (Koop 1989) proved that only basal area, selective cut-
ting, shifting cultivation and developmental phase contributed significantly to the
variance of species composition of trees, saplings and seedlings, shrubs and high

325
DISCUSSION
The final rating of forest integrity suffers from the fact that variables of a totally
different nature are lumped together in a single figure. Nevertheless the method
presented in this paper is substantiated by scientific processing of a more exten-
sive data set on species composition and forest structure. The diagnostic method,
applied in the Gunung Leuser area, gives reliable scores in plots from undisturbed
forest to woodlands in which few trees are left. The sensitivity of the method in
open areas, however, is limited. Recent cutting, mowing or burning reduces the
presence of pioneer trees and big ginger species and therefore the absence of light
indicators can give artificially higher scores.
Not all indicators react similarly in the whole range of forest disturbance.
Atmospheric moisture indicators, including woody lianas, are sensitive to
disturbance caused by selective cutting, whereas in young and old abandoned
shifting cultivation they are absent. More severe disturbance is better covered by
the indicators of herbaceous cover, light indicators and basal area.
Although the diagnostic method has now been tested along a 3500 m transect in
Gunung Leuser National Park and in the Wanariset forest in East Kalimantan,
experience with this method is still limited. Therefore it is necessary to test the
diagnostic method on a larger scale.
In different areas of the world atmospheric moisture, forest structure and light
indicators can be used to score forest integrity. The range of sensitivity of the three
indicator groups also depends on the region in which the diagnostic method is
used. The structure and light indicators are most generally applicable, even in
regions without moist climates. Atmospheric moisture indicators, however, have
a more restricted applicability. Even in plots on the lower slopes in the Ketambe
forest the atmospheric moisture indicators are scarce. Two factors can be identified
that may cause this scarcity. First, reference plot Ketambe 2 is situated on a ridge
where the forest is more exposed to winds than in the flat and sheltered position of
plots on the flat terrace. The most exposed portion of the ridge, with slopes of 25
to 35 degrees to the east, north, and west, shows total absence of atmospheric
moisture indicators. The lower slopes in the north-south and the west-east transect
have the highest mosslines and frequencies of atmospheric moisture indicators,
even where the transect transverses an aggrading gap phase. The crest and upper
slopes of the ridge again have lower average mosslines and lower presence of
atmospheric moisture indicators. This implies that in areas of marked topographi-
cal differences, valleys as well as ridges have to be sampled and comparisons of
disturbance units elsewhere should be based on similar conditions of relief.

329
 A second explanation for fewer atmospheric moisture indicators is that the
level of illumination may have become a limiting factor. Richards (1984) mentions
air humidity as the critical factor for survival of feather, ramicolous and hanging
mosses but stresses that light is often the most important limiting factor. Although
low intensities of direct sun radiation favour high air humidity, it can be too dark
despite sufficient air humidity. The lower presence of light indicators in the
Ketambe 2 reference plot can be explained by the higher canopy closure (Table 2),
with only 3.3% of the area in innovation and aggrading gap phases. This explanation
is supported by a relatively higher score of atmospheric moisture indicators in the
aggrading gap and aggrading tree phases (Table 5), where light intensities are
probably higher because an emergent canopy of Parashorea is absent. For the same
reason, the tree phase plots adjacent to an aggrading tree or an aggrading gap
phase most often have common atmospheric moisture indicators. Oblique radia-
tion can penetrate under tree crowns beyond the vertical projections of the tree
crowns that delimit a canopy gap and the disturbance of falling trees often extends
underneath crown projections of taller trees. With a limited opening of the canopy
and an insignificant loss of air humidity, light is not the limiting factor anymore
and atmospheric moisture indicators may increase. This explains why, in one
selective cutting disturbance unit, an atmospheric moisture indicator (bole-climber)
score was higher than in the undisturbed reference plot.
The indicator list of the diagnostic method so far is restricted to vegetation
characteristics. Animal species that are highly indicative of primary forest can,
however, be included in the list. Species that are easily recognized by their fre-
quent and characteristic calls e.g. hombills (Bucerotidae) and barbets (Capitonidae)
are most suitable. Other signs of the presence of big mammals can provide addi-
tional information. However, such data will be far from complete and the chance of
finding the evidence of a species presence in the limited area of the transect is too
low for these data to be used in the diagnostic method. The absence of evidence
does not prove the absence of the species. However, forest characterized by forest
structure, light and moisture conditions does not express diversity of animal life
which may be affected by hunting pressure. Separate data are needed to assess the
impact of hunting.
Increasing surface percentage of canopy gaps results in increasing homo-
geneity in forest structure in secondary forest (Budowski 1965, Jacobs 1988) and
discontinuity of old growth patches, providing a more direct diagnosis of forest
integrity (see Dale et al., Chapter 4, this volume). Mapping areas of innovation and
aggrading gap phases needs a different scale of assessment and is too time

330
consuming. The percentage of canopy gaps is, however, reflected by the fre-
quency percentage of light indicators.
To apply the diagnostic method, a reference is necessary. This reference does
not need to be a strictly virgin forest but a forest stand that has been unexploited
for as long as possible will suffice. Comprehensive data from undisturbed plots
can serve as a reference for the design of management methods for restoration
management of heavily disturbed natural forest.

CONCLUSIONS
This method for rapid assessment of forest integrity is suitable for monitoring the
effects of forest-people interaction. A set of simple and easy recognizable
characteristics can assess the integrity of tropical rain forest. The method can be
applied by local rangers and villagers such as para-taxonomists after practical
training of 3-5 days, if supervised and monitored by a scientist who should also
process the data.
Forest structure indicators and light indicators are not dependent on differen-
ces in macro and micro-climate. Atmospheric moisture indicators are restricted to
rainforest. Because of micro-climatological differences of ridges and valleys in hilly
areas that influence the presence of atmospheric moisture indicators, both should
be sampled, and similar relief conditions should prevail in undisturbed reference
plots and disturbance units. Light- and atmospheric moisture indicators have
different ranges of sensitivity. Atmospheric moisture indicators proved to be sensi-
tive to even slight human disturbance by selective cutting whereas light indicators
differentiate more serious disturbance where atmospheric moisture indicators are
already absent.
Forest integrity is related to biodiversity. However, forest characterized by
structure, light, and moisture conditions does not reflect diversity of fauna1 species.
Separate data are needed to assess the impact of disturbance on wildlife. For a
wider application of the diagnostic method there is a need to conserve virgin forest
reference plots for all site types within a particular region. In case no virgin forests
are left, forest areas that have been least disturbed should be left to serve as future
reference plots.

ACKNOWLEDGEMENTS
LDP Indonesia. The field workers Ibrahim Kusuma, Usman S., Samsu A.R. and

331
Bukam from the village of Balalutu Aceh Tengara, the Director and Staff of the
National Park of Gunung Leuser for their hospitality and Drs. A. Touw, De Wilde,
Van Balgooy from the Rijksherbarium Leiden for determination of moss and tree
species.

LITERATURE
Abdulhadi, R., E. Mirmanto and K.A. Kartawinata. 1987. A lowland Dipterocarp
forest in Sekundur, North Sumatra, Indonesia: five years after mechanized
logging. In: A.J.G.H. Kostermans (ed.) Proc. third round table conf. on
Dipterocarps, UNESCO.
Beek, C.G.G., van. 1982. Een geomorfologische bodemkundige studie van het
Gunung Leuser Nationaal Park, Noord Sumatra, Indonesie. Utrechtse
geografische studies 26, Dept. of Geography, University of Utrecht.
Bitterlich, W. 1948. Die Winkelzahlprobe. Algemeine Forst- und Holzzeitung: 4-5.
Budowski, G. 1965. Distribution of tropical rain forest species in the light of
successional process. Turrialba 15: 40-42.
Halle, F., R.A.A. Oldeman and R.B. Tomlinson. 1978. Tropical trees and forests. An
architectural analysis. Springer, Berlin, Heidelberg, New York.
Hill, M.O. 1979. Decorana: a Fortran computer program for detrended
correspondence analyses and reciprocal averaging. Ecology and Systematics.
Cornell University, New York.
Hommel, W.F.M. 1990. A phytosociological study of a forest area in the humid
tropics (Ujung Kulon, West Java, Indonesia), Vegetatio 89: 39-54.
Jacobs, M. 1988. The tropical rain forest: a first encounter. Springer-Verlag, Berlin,
Heidelberg, New York.
Karr and Budley (1981). Cited in: Karr, J.R. 1990. Biological integrity and the goal of
environmental legislation: Lessons for conservation biology. Conservation
Biology (3): 244-250.
Koop, H. 1981. Structuur en dynamiek van twee natuurlijke bossen; het Neunburger
en Hasbrucher Urwald. Reeks Landbouwkundige Verslagen 901. PUDOC
Wageningen.
Koop, H. 1989. Forest dynamics, SILVI-STAR: a comprehensive monitoring system.
Springer-Verlag, Berlin, Heidelberg, New York.
Kramer, H. and A. Akca. 1982. Leitfaden fur Dendrometrie und Bestandesinventur.
Sauerlander’s Verlag. Frankfurt am Main.

332
Londo, G. 1984. The decimal scale for releves of permanent quadrats. Pages: 46-370
In: Knapp, R. (ed.) Sampling methods and taxon analyses in vegetation science.
Junk, The Hague.
Oldeman, R.A.A. 1978. Architecture and energy exchange of dicotyledonous trees
in forests. Pages: 535-560 In: Tomlinson and Zimmerman (eds.) Tropical trees as
living systems. Cambridge University, Cambridge.
Oldeman, R.A.A. and P.J. van der Meer. 1988. Diagnosis of Northern Queensland
rain forests: the impact of selection silviculture. Wageningen Agriculture
University.
Pocs, T. 1982. Tropical forest bryophytes. Pages 59-104 In: A.J.E. Smith (ed.)
Bryophyte Ecology. Chapman and Hall, London.
Richards, P.W. 3984. The ecology of tropical rain forest bryophytes. Pages 1233-
1270 In: R.M. Schuster (ed.) New manual of Bryology Vol. 2. The Hattory
Botanical Laboratory, Nichinan.
Rijksen, H.D. 1978. A field study on Sumatran orang utans (Pongo pygmaeus Abelii
Lesson 1827), ecology, behaviour and conservation. Communications
Agricultural University Wageningen 78-2, Veenman Wageningen.
Schaik, C.P. van and E. Mirmanto. 1985. Spatial variation in the structure of litterfall
of a Sumatran rain forest. Biotropica 17:196-205.
Smiet, A.C. 1989. Human impact on mountain forest in river Konto area: Vegetation
and transect studies. Konto river project, Project communications No 11 Malang.
Swaine, M.D. and J.B. Hall. 1983. Early succession of cleared forest land in Ghana.
Journal of Ecology 71(2):601-628: 1983
Ter Braak, C.J.F. 1986. Canonical correspondence analyses: a new technique for
multivariate direct gradient analyses. Ecology 67(5): 1167-1179.

333
 Chapter

23
A GIS APPROACH
TO MAPPING SPATIAL VARIATION
IN TROPICAL RAIN FOREST BIODIVERSITY

Alan Grainger 1, Sam Rose 1, Yongyut Trisurat 2,

Warren Y. Brockelman 3

INTRODUCTION
The focus of conservation planning has switched over the past twenty years from
protecting single threatened species to conserving biodiversity as a whole, but the
techniques required to support this new goal are still lacking. Some aspects of
biodiversity have been mapped, e.g. distributions of species density and centres of
endemism, but biodiversity has not yet made the transition from wide-ranging
concept to measurable quantity and this, together with poor data, is hampering
conservation scientists and planners alike. New approaches are therefore needed
that can enable biodiversity to be mapped comprehensively over large areas and
provide the basis for improved conservation planning methods.

1 School of Geography, University of Leeds, Leeds LS2 9JT, UK.

2 School of Environment, Resources and Development, Asian Institute of Technology, Bangkok,
Thailand.
3 Centre for Conservation Biology, Faculty of Science, Mahidol University, Bangkok 10400, Thailand.

335
 This paper suggests how geographic information system (GIS) techniques can
be used to map the biodiversity of tropical rain forests. From a GIS perspective,
biodiversity has vertical and horizontal structures, just like an ecosystem. Its ver-
tical structure comprises the sum of the spatial distributions of many species of
plants and animals, together with those of the ecosystems of which they are a part.
Each of these distributions has a horizontal structure, corresponding to their spa-
tial variation. By studying this variation conservation planners can identify areas
of exceptional richness, or areas containing the minimum critical populations of
species of high conservation priority. The first part of the paper looks at biodiver-
sity as a multi-dimensional phenomenon, while the second explores the role of
fragmentation in explaining the spatial heterogeneity of biodiversity.

BIODIVERSITY AS A MULTI-DIMENSIONAL PHENOMENON

Geographic Information Systerms
A geographic information system is a piece of computer software capable of the
integrated handling and analysis of spatial data (Burrough 1986). It is a sophisti-
cated database that stores data on the geographical location of spatial entities and
one or more of their attributes, and can test for relationships between different enti-
ties and attributes. Information is stored either in vector format, where it refers to
points, lines and polygons that correspond to actual features, or in raster format,
where it refers to cells in a uniform grid into which a study area is divided. Most
GIS programs now allow easy interchange between the two formats.
Human beings are limited in their ability to interpret complex geographical
information, especially when, for the sake of economy, multiple features are
combined on the same printed map, e.g. the distributions of towns, rivers and lakes.
A GIS simplifies the picture by mapping the point distribution of towns, the line
distribution of rivers and the polygon distribution of lakes in an area as separate
digital data layers. So, while on a paper map these are overlaid in an inflexible
way, on a GIS they can be overlaid selectively and flexibly for easier interpreta-
tion. They can also be quantitatively related.
The GIS approach is more than just an efficient method for presenting complex
geographical information. It also offers a new way of looking at that information,
by dividing it into major sets of individual features and allocating a separate digi-
tal map layer to each. A single feature can have multiple attribute layers associa-
ted with it, e.g. the attributes associated with an area of forest include its biomass
density (mass of carbon per ha) and species density (number of species per ha)

336
 To quantify and map biodiversity as such, these components must be combined
in a coherent and integrated way Little progress has been made in this endeavour
so far. There have been attempts to create an aggregate national biodiversity
index (WCMC 1992) and although this is valuable for some purposes, it omits a lot
of detail and its spatial component is restricted to political boundaries that usually
have little ecological meaning.
From a GIS perspective, however, the solution to this problem seems obvious:
simply accept that biodiversity must be regarded as a multi-dimensional
phenomenon comprising the sum of its spatially referenced components. The spa-
tial distribution of biodiversity in an area can therefore be portrayed as a set of
three related primary information layers, comprising ecosystem diversity, species
diversity and genetic diversity, generated by combining a large number of raw data
layers corresponding to the distributions of individual taxa (Figure 2). Maps of the
aggregate distribution of biodiversity can then be produced using additional
information layers derived from the primary layers (see below).

Ecosystem Diversity
The ecosystem diversity layer is the foundation on which the others are laid. In
purely natural conditions undisturbed by human impact this shows the actual
distribution of different ecosystem types. The detail shown in this layer depends
upon the scale of the map. At global and regional levels it would show a small
number of major ecosystems, or biome-types, such as tropical rain forest. At larger
scales various ecosystem sub-types would be visible, e.g. in the humid tropics,
evergreen lowland rain forest, semi-evergreen lowland rain forest, heath forest etc.
However, production of detailed global maps is constrained by incompatibilities
between different ecosystem classification systems.
Because human impacts on ecosystems have been widespread, such maps
usually only show the potential distributions of the major climax ecosystem types.
The real distribution of ecosystem diversity is very different, since many natural
ecosystems have been cleared, and those which remain have been degraded in
various ways, resulting in a complex mosaic of cleared, degraded and undisturbed
areas. Maps of the actual distributions of tropical rain forests today are few and
far between, and most of those which are available are out of date. IUCN recently
combined the best available maps of tropical rain forest in Africa and Asia in its
atlases (Collins et al. 1990, Sayer et al. 1992), but acknowledged the limitations of
the original data in many cases. FAO’s latest survey of tropical forests in 1990 did
not include any maps at all (FAO 1993, Grainger 1995a).

338
Species Diversity
The species diversity layer is ideally derived by combining thousands of indivi-
dual data layers, each showing the distribution of a single species. At the moment,
owing to lack of reliable data, this layer is usually restricted to measuring varia-
tion in aggregate species richness. But there is potential to devise more sophisti-
cated measures and use GIS techniques to identify patterns that take account of the
actual species present and their many qualities.
Moreover, because of lack of data, species distributions are frequently repre-
sented on maps either by points on a map (Figure 3a), continuous distributions
over a large area (Figure 3b), or in terms of simple presence or absence of a species
from a grid square, which may be as large as 1” x 1” (Figure 3c). Even if these maps
were to reflect natural distributions they would be misleading, because many
areas have not been properly surveyed and the apparent concentration of some
species distributions may reflect collection bias (Nelson et al. 1990). But such maps
only really portray potential natural distributions, since it is impossible to inven-
tory everywhere and they rarely take account of human disturbance.
While recognizing existing constraints, it is important to show the possibilities
that would be opened up were better spatial data to be collected, so as to streng-
then the efforts of major herbaria which are just beginning to produce digital maps
of the distribution of taxa. Hopefully, in future, many individual species distribu-
tion maps will be produced and stored on GIS databases, so making possible the
kinds of analysis proposed here.
For simplicity, in this paper the second primary layer is restricted to species
diversity only. But there are strong arguments for including other taxonomic
levels, such as genera and families, so that the layer refers to the full range of
taxonomic diversity.

Genetic Diversity
The third data layer shows genetic variation within the distributions of individual
species, and hence builds on the second layer. Considerable genetic variation may
occur in the populations of tropical tree species, and the provenance of trees is of
prime importance when breeding improved varieties for different purposes (Burley
and Wood 1976). Some species show high variability between populations in
different areas, so distinct subpopulations are recognised, though variability with-
in each sub-population is quite low. Other species have low variability between
populations but high variability within populations. Such spatial genetic variabi-

340
The Overall Appraisal of Biodiversity
As biodiversity is such a complex phenomenon there is no single way of portray-
ing it. However, synthetic appraisals of biodiversity as a whole can be produced
by using various mathematical operations to combine some or all of the primary
layers to give a range of derived layers suited to particular purposes. One such
layer, for example, might show the location of centres of endemism; another might
separate an area covered by a single major ecosystem type into distinct ‘biodiver-
sity zones” ranked by their priority for conservation.

SPATIAL HETEROGENEITY IN BIODIVERSITY

To map variation in biodiversity as a whole requires searching for spatial patterns
in the primary and derived layers. This section looks at how spatial heterogeneity
can be characterised by fragmentation and clustering in the ecosystem diversity
and species diversity layers.

Fragmentation of Ecosystem Distributions

Natural Fragmentation.
Ecosystem distributions are fragmented under natural conditions. The extent to
which natural ecosystem diversity is shown on maps depends on the chosen map
scale (and ecosystem classification system). As the scale increases, so does the natural
fragmentation of areas with apparently homogeneous vegetation cover. Thus,
tropical rain forest becomes fragmented into its major sub-types, e.g. evergreen
lowland rain forest, semi-evergreen lowland rain forest etc., eventually fragment-
ing even further to match the distribution of landscape diversity.

Artificial Fragmentation
In practice, the natural distribution of ecosystems is heavily fragmented as a result
of being replaced or degraded by human intervention. Fragmentation is most visi-
ble when forest is cleared. In areas where deforestation has just begun, a large
contiguous block of forest will exhibit local fragmentation on its margins but its
core will remain intact. As deforestation reaches a more advanced stage, for exam-
ple in Peninsular Malaysia, preferential clearance in some areas leads to more
widespread fragmentation (Figure 4).

342
Gardner 1991). There has also been considerable work on the species compositions
of ecosystem fragments from the perspective of island biogeography theory
(MacArthur and Wilson 1967, Harris 1984). However, there is great potential to go
further by combining the two approaches, in order to study how highly frag-
mented ecosystem diversity distributions affect the variation in species diversity
over large areas.

A Taxonomy of Degraded Forest Types

A prerequisite for such work is a taxonomy that can introduce some degree of or-
der into the artificial fragmentation of ecosystem distributions. Despite the large
area of secondary forest in the tropics, forest degradation has received little atten-
tion from biogeographers. But this has begun to change, mainly because of the
recognition of the importance of degraded forest in climate change, and the need
to map spatial variation in biomass density.
Degradation has been defined as a “temporary or permanent reduction in the
density, structure, species composition or productivity of vegetation cover”. Using
a set of indicators based on this definition an initial taxonomy has been proposed
for Southeast Asia (Grainger 1995b). It includes six main types of degraded forest:
1. Extractive Forest, which has had trees or herbaceous plants removed from it.
Examples include logged forest and forest modified by other extractive uses.
2. Damaged Forest, which has been degraded by drought, insect pests, pollution
etc.
3. Regenerated Forest, which is regrowing after clearance. Examples include
shifting cultivation forest fallow, and some forms of managed forest.
4. Planted Vegetation, which results from intentional tree planting, e.g. for
silviculture, agroforestry, or tree crop plantations.
5. Interrupted Successional Vegetation, which has a low tree component and is
usually dominated by grassland, e.g. Imperata cylindrica grasslands.
6. Dispersed Forest, which comprises whole landscapes that have been heavily
modified by human impact and now contain mixtures of forest trees, tree crops
and grasslands in a dispersed, heterogeneous mosaic. This is an advanced case
of forest fragmentation.
This taxonomy can be used to map ecosystem diversity in a way that takes account
of both natural and artificial fragmentation, so providing a better foundation for
mapping the actual distribution of species diversity than has been possible hither-
to.

344
Ecosystem Fragmentation and Conservation Planning
The study of fragmentation has major applications in conservation planning. The
locations of national parks and other protected areas are generally chosen so as to
include large continuous areas of different habitats where the viability of animal
species populations is high, and avoid areas where deforestation and forest
degradation have caused habitat fragmentation, although inevitably some parts of
protected areas were disturbed prior to their establishment.
New techniques are needed to map biodiversity distributions rapidly over large
areas, so that future protected areas can be sited to protect areas with the highest
biodiversity, and existing protected areas can be managed more effectively. Lack of
artificial fragmentation of habitats will be one criterion for choosing a potential
protected area, but on the other hand both natural and artificial fragmentation need
to be taken into account when undertaking management zonation. The long-term
integrity and biodiversity of a protected area depends on choosing the most suita-
ble management regime for each part of it, e.g. strict protection, public access for
recreational purposes etc. Many protected areas were established without prior
surveys of their biodiversity, and this deficiency may often have been remedied
before they were divided into different management zones. Consequently, the
choice of high public access areas may owe more to the proximity of roads than to
the use of biological criteria.
GIS and remote sensing techniques are now increasingly used to study the
distribution of animal species based on habitat distribution, because traditional
ground survey methods, such as counting animals, trapping, collecting droppings,
investigating feeding sites and mapping habitats on the ground, are very time
consuming. Habitat maps for particular species are produced using satellite ima-
gery supplemented by ground truth data (Wulf et al. 1988), and then inferences are
drawn concerning the density of animals and the viability of species populations in
that area.
This approach has been used to study the distribution of gibbons in the Khao
Yai National Park in Thailand (Figure 5). Two species are involved, the pileated
gibbon (Hylobates pileatus) and the white-handed gibbon (H. lar). Areas containing
tropical evergreen forest below 1,000 m above sea level were designated as gibbon
habitats, since areas above 1,000 m are usually occupied by hill evergreen forest
which has fewer fruit trees, and mixed deciduous forest supports gibbons but at a
lower density. Gibbons are also generally absent from the perimeters of protected
areas because of poaching, habitat degradation etc.

345
Fragmentation of Species Diversity
Spatial variation in species diversity reflects the sum of complex variations in the
distributions of thousands of individual species. To measure this, in principle, would
require superimposing all of these distributions and identifying the resulting spa-
tial patterns. Given the large number of taxa in the humid tropics - there are
approximately 1,800 genera and 30,000-50,000 species in Amazonia alone - this would
be almost impossible to do manually. But it would be an ideal task for GIS tech-
niques - assuming, of course, that reliable spatial datasets become available.

Measuring lndividual Species Heterogeneity

Under natural conditions, some species will actually have the kind of continuous
distributions shown on maps, while others will be fragmented into a number of
clusters, owing to historical reasons or natural ecosystem diversity. Furthermore,
because of artificial ecosystem fragmentation, even naturally continuous species
distributions will be fragmented to some extent.
The analysis of fragmentation in biodiversity often focuses on local endemics,
i.e. those occurring in fragments or clusters of restricted size in an area. Indeed, the
distribution of endemics is often used as a surrogate for the distribution of
biodiversity. But it can be argued that endemism is merely a special case of
fragmentation, and that in mapping the distribution of overall species diversity
the distributions of endemics should be measured using the same method as that
used for species whose distributions are fragmented over a larger area. In other
words, endemics would be represented by single fragments (or patches) and
fragmented species by multiple fragments.
Determining the degree of species fragmentation empirically will require a far
more detailed knowledge of individual species characteristics than is generally
available at the moment in the humid tropics. The apparent degree of fragmenta-
tion will also depend on the scale and mapping method used. For example,
distributions mapped by denoting the presence or absence of a species in 10 km
grid cells may appear to be continuous although such maps do not say anything
about the density of individuals inside the grid cell - and maps with 1 km grid cells
could show a very different pattern.
This can be illustrated by comparing three species from Amazonia. Hirtella
physophora is fairly widespread throughout the region (Prance 1982) and even
though it has not been found everywhere, the likelihood of its presence between
confirmed observations is high, so its distribution can be assumed to be conti-
nuous (Figure 7a). On the other hand, Eperua duckeana is apparently a local endemic,

348
 This approach could provide a more comprehensive picture of the variation in
biodiversity than has been available until now, encompassing many more groups
of plants and viewing endemism as only part of a much greater heterogeneity in
the distribution of biodiversity. A more elaborate approach would attempt to
encompass species with fragmented distributions in areas affected by human impact.
Their distributions will be related to patterns of natural and artificial ecosystem
fragmentation; by means of relationships that will need to be modelled
mathematically and incorporated into the mapping process.

Data Limitations
Implementing this approach in practice will encounter a number of obstacles. One
of these is the severe lack of spatial data on individual species and genus
distributions, largely due to the paucity of spatial information accompanying
botanical collections and the huge number of species and genera in the tropics.
Another obstacle is collection bias. Some species may appear to be concentrated
in certain areas, but only because these were visited by collecting expeditions and
others were not, perhaps because the areas are close to cities or long-established
research sites (Nelson et al. 1990). It is essential to compensate for this bias. For
example, a good starting point would be to build spatial computer models of species
distributions over their potential natural ranges, using maps of basic environmental
data (Clinebell 1993, Phillips et al. 1994). Even better would be to supplement this
by superimposing individual species distributions with the ecosystem diversity
layer.

CONCLUSIONS
This paper has proposed that one way to overcome the many obstacles which im-
pede the mapping of biodiversity as a whole is to simply accept that it is a multi-
dimensional phenomenon comprising various primary information layers. If
necessary, these can be combined to give various derived synthetic information
layers as appropriate for particular purposes.
Until now, the only way to map biodiversity over large areas has been in terms
of the variation in species density, or the density of endemic species. This paper has
argued that this is a rather simplistic approach, and should be superseded by one
which maps the spatial heterogeneity of all species, applying the same fragmenta-
tion index to both endemic species and species with wider but fragmented
distributions.

351
REFERENCES
Brown, S., L.R. Iverson and A.E. Lugo. 1994. Land use and biomass changes in
Peninsular Malaysia during 1972-82: a GIS approach. Pages 117-143 In: V.H.
Dale (ed.) Effects of Land Use Change on Atmospheric Carbon Dioxide
Concentrations. Springer-Verlag, New York.
Burley J. and P.J. Wood. 1976. A Manual on Species and Provenance Research with
Particular Reference to the Tropics. Oxford Forestry Institute, Oxford.
Burrough, PA. 1986. Principles of Geographic Information Systems for Land
Resources Assessment. Clarendon Press, Oxford.
Clinebell, R. 1993. Regression Analyses of Lowland Neotropical Forest Species
Richness. MSc Thesis, Washington University, St Louis, Missouri.
Collins, N.M., J.A. Sayer and T.C. Whitmore. 1991. The Conservation Atlas of Tropical
Forests: Asia and the Pacific. Macmillan, London.
Cowan, R.S. 1975. A monograph of the genus Eperua (Leguminosae:
Caesalpinoideae). Smithsonian Contributions to Botany No. 28.
FAO. 1993. Forest Resource Assessment 1990 - Tropical Countries. FAO Forestry
Paper No.112. UN Food and Agriculture Organization, Rome.
Forman, R.T.T. and M. Godron. 1986. Landscape Ecology. John Wiley, Chichester.
Grainger, A. 1995a. Forest areas and deforestation rates in the humid tropics: a
reassessment. Geographical Journal (submitted).
Grainger, A. 1995b. The degradation of tropical rain forest in Southeast Asia:
taxonomy and appraisal. In: M. Eden et al. (eds.) Land Degradation. Mansell
Publishers, London (in press).
Harris, L.D. 1984. The Fragmented Forest: lsland Biogeographic Theory and the
Preservation of Biotic Diversity. University of Chicago Press, Chicago.
MacArthur, R.H. and E.O. Wilson. 1967. The Theory of Island Biogeography.
Princeton University Press, Princeton.
McNeely J.A., K.R. Miller, WV. Reid, R.A. Mittermeier and T.B. Werner. 1990.
Conserving the World’s Biological Diversity. World Bank, Washington DC.
Nelson, B.W., C.A.C. Ferreira, M.F. da Silva and Kawsaki. 1990. Endemism centres,
refugia and botanical collection density in Brazilian Amazonia. Nature 345: 714-
716.
NRC. 1991. Managing Global Genetic Resources: Forest Trees. US National Research
Council. National Academy Press, Washington DC.
Phillips, O., A. Gentry, C. Reynel, P. Wilkin and B. Galvez-Durand, 1994. Quantitative
ethnobotany and Amazonian conservation. Conservation Biology (in press).

352
Pickett, S.T.A. and P.S. White (eds.). 1985. The Ecology of Natural Disturbance and
Patch Dynamics. Academic Press, Orlando, Florida.
Prance, G.T. 1982. Forest refuges: evidence from woody angiosperms. In: G.T. Prance
(ed.) Biological Diversification in the Humid Tropics. Columbia University Press,
New York.
Prance, G.T. 1990. The floristic composition of the forests of Central America. In:
A.H. Gentry (ed.) Four Neotropical Forests. Yale University Press, New Haven.
Sayer, J.A., C.S. Harcourt and N.M. Collins. 1992. The Conservation Atlas of Tropical
Forests: Africa. Macmillan, London.
Solbrig, O.T. 1991. The origin and function of biodiversity. Environment 33(5): 16-
20,34-38.
Trisurat, Y. 1994. Mapping gibbon distributions using GIS. A case study of Khao Yai
National Park, Thailand. Asian Institute of Technology. Ms.
Turner, M.G. and R.G. Gardener (eds.). 1991. Quantitative Methods in Landscape
Ecology. Springer, New York.
WCMC. 1992. Development of a National Biodiversity Index. World Conservation
Monitoring Centre, Cambridge. Ms.
Wong, I. 1971. The Present Land Use of West Malaysia 1966. Ministry of Agriculture
and Lands, Kuala Lumpur, Malaysia.
Wulf, R.R.D., R.E. Goosens, J.R. MacKinnon and W.S. Cai. 1988. Remote sensing for
wildlife management: Giant Panda mapping from Landsat MSS images.
Geocarto International 3: 41-47.

353
 Chapter

24
ASSESSING PRIORITY AREAS
FOR BIODIVERSITY AND
PROTECTED AREAS NETWORK S

C.R. Margulesl and I.J. Kitching2

INTRODUCTION
This paper is a brief review of a large body of scientific work by many colleagues,
principally, and in alphabetical order, D.P. Faith, C.J. Humphries, C.R. Margules,
A.O. Nicholls, R.L. Pressey, A.G. Rebelo, R.I. Vane-Wright and PH. Williams. A
more thorough treatment of the entire field of priority areas analysis will be provided
in a forthcoming book, edited by C. J. Humphries, C.R. Margules, R.L. Pressey and
R.I. Vane-Wright, and to be published by Oxford University Press.

PRIORITY AREAS FOR BIODIVERSITY

Biodiversity priority areas are defined as a set of areas within a prescribed
geographical region that together encompass all, or as much as possible, of the
biodiversity of that region. They are so-called because they should receive priority

1 CSIRO Division of Wildlife and Ecology, P.0. Box 84, Lyneham, A.C.T. 2602, Australia.
2 Biogeography and Conservation Laboratory, Department of Entomology, The Natural History
Museum, Cromwell Road, London SW7 5BD, Great Britain.

355
attention for protection and/or appropriate management. Such areas will not in
themselves be sufficient to ensure the long term maintenance of biodiversity, but
they should form the core of conservation plans for biodiversity protection.
To achieve the goal of identifying priority areas, there must be methods for
measuring biodiversity, determining appropriate levels of representation and a
cost-effective means of allocating available resources to secure it. The objective
of this paper is to overview methods for deriving explicit statements about the
relative contribution of different areas, both individually and in combination.
These methods provide for maximum flexibility, for it is recognized that there are
competing land uses and that these can and do impose severe limitations upon
biodiversity protection. However, it is also accepted that there are areas that are
indispensable if full representation is to be achieved, and such areas can also be
recognized by application of these methods.
Attributes of organisms, the currency of biological diversity, are the characters
that confer variety among taxa and hence govern their functional interactions.
Protecting these attributes has been described as an exercise in maximising option
value (e.g. IUCN/UNEP/WWF 1980, Faith 1994). However, the number of poten-
tial attributes of all taxa is phenomenal and to all intents and purposes unknow-
able. Yet the goal of biodiversity conservation is to preserve this variety. In theory,
the pattern of phylogenetic relationships among taxa can be used as an estimate of
the distribution of their attributes. In other words, attribute diversity can be pre-
dicted from taxonomic classifications. Thus an ideal goal for biodiversity priority
areas might be that they should sample and maintain populations of all of the
terminal taxa (usually species) of a classification. However, this goal must be revised
in view of our imperfect knowledge of terminal taxa, their phylogenetic
interrelationships and geographical ranges; and inadequate management
prescriptions for ensuring their continued persistence in the long term.

BIODIVERSITY SURROGATES
The search for biodiversity priority areas ideally should be based upon a know-
ledge of the entirety of biodiversity. Although some moves have been made to
acquire such knowledge (e.g. the all-taxon biodiversity inventories of Janzen 1993),
the task of discovering, naming and determining the systematic relationships of all
species is daunting. Complete inventories remain an impracticable option for most
areas for the foreseeable future. However, changes in land use continue and so
some type of biodiversity surrogate is required. Three kinds of surrogates are

356
realistically available: a subset of species or other taxa; ecological assemblages; and
environmental variables.
Species are the usual units with which biodiversity is measured (Vane-Wright
1992). Higher taxa (genera, families) may also be used provided that a relationship
can be established between the distributional patterns of the higher taxa and those
of the species that they contain (Williams and Gaston 1994). Higher taxa are also
cheaper and quicker to survey and identify. The search in the past has been for so-
called “indicator taxa”, small groups of species or other taxa that would act as
general biodiversity predictors. Beccaloni and Gaston (in press) demonstrated that
butterfly species richness at forest sites in the neotropics could be accurately
predicted if the total number of species of the subfamily Ithomiinae was known.
Nevertheless, it is clear that there is no such thing as a single indicator taxon that
can accurately predict overall species richness for all groups at all sites. Kremen
(1992) proposed “target taxon” for a group that could be demonstrated to be a better
than average indicator of a wider range of biodiversity. We would also argue in
favour of “focal taxa” (Ryti 1992), which are those about which we have good
information and are taxonomically tractable in the short-term. Roger Kitching
favours “predictor sets”, a set of groups which, when summed, would approximate
overall biodiversity. However, there is little evidence that any subset of taxa can
fully or completely represent biodiversity as a whole (e.g. Prendergast et al. 1993,
Williams and Gaston 1994).
An assemblage is a more generalized and heterogeneous entity than a taxon,
and can be one of a number of ill-defined classifications, such as community,
association, habitat type, ecosystem, etc. Assemblages can be derived subjectively
by using a small number of dominant or prominent species or they can result from
a numerical pattern analysis of field records of the locations of species. Assembla-
ges represent various combinations of species and also include an element of their
interactions. As such, they contain more ecological complexity than species.
However, being more inclusive and less well-defined entities than species and other
taxa, assemblages are yet further removed from the attributes of taxa. Furthermore,
affording protection to an area as a means to representing an assemblage will
probably miss some species (Pressey and Logan 1994), and some ecological
complexity because it will be difficult to determine whether a given subset of an
assemblage is representative of the whole.
Environments, a generic term covering land classifications based upon physical
and climatic characteristics (with or without a biotic component), are even more
remote from attributes. There is strong theoretical support for the use of

357
environments as biodiversity surrogates (Margules and Austin 1994). A network
of priority areas designed to represent environments may encompass both
unknown species and unknown distributional components of known species.
Furthermore, environmental data are more widely available, with a greater
consistency and detail across broad geographical areas, than unbiased biological
data. However, environments suffer from similar drawbacks to assemblages
regarding what is adequate representation, while the relationships between
environmental variables and the distribution patterns of taxa are often vague and
difficult to quantify. It is always possible that some species require an unrecog-
nised combination of environmental variables (Pressey and Logan 1994).
Given the present limitations on knowledge and resources and the goal of
adequate representation of each surrogate, it is likely that some combination of
surrogates will be required. Whatever, the end result should be a data set of the
chosen biodiversity surrogates in the form of maps, or stored electronically in a
form suitable for mapping. It is these data, analysed using the methods summarized
below, that can be used to identify biodiversity priority areas. Because species are
surrogates frequently employed in biodiversity assessments, the following
discussion will be simplified by using “species” instead of “attributes” (or “taxa”,
“assemblages”, “ecosystems”, etc.).

IDENTIFYING PRIORITY AREAS FOR BIODIVERSITY CONSERVATION

The methods for identifying biodiversity priority areas should be both explicit and
efficient. Efficiency is necessary because the amount of land and water realistically
available for the conservation of biodiversity is limited and there is a real possibi-
lity that this maximum limit (whatever it is) will be reached before biodiversity is
adequately represented. The methods must be explicit in order that results are
repeatable through independent verification, and hence networks of priority areas
can be more readily justified and defended. The methods endorsed here are based
upon four principles: complementarity, irreplaceability, vulnerability and viability.
The selection of biodiversity priority areas must proceed from the specifically
stated goal of representing all species of the study set. It follows from this that each
new area added to an existing network should contribute species not yet repre-
sented in that network (providing that complete representation has not yet been
established). This common sense observation reflects the principle of comple-
mentarity. Priority areas should complement one another in the species they con-

358
 tain, not just contribute repeats of species found in areas already acquired. It
follows from this that the area thatcontributes most towards the goal of complete
representation may not necessarily be one of the originally richest areas. This is
why scoring and ranking procedures, and procedures that allocate high value to
species richness regardless of the identities of those species are often so highly
inefficient (Pressey and Nicholls 1989). They fail to take account of species tum-
over from area to area. Complementarity is vital because it results in efficient
representation of species and consequently leads to the efficient allocation and use
of often limited conservation resources.
Simply identifying a set of biodiversity priority areas, however, is not sufficient,
as all sites within that set are not equal. Some sites are irreplaceable in that they
contain species that are not represented in any other site within the set. If these sites
were lost, they would compromise the goal of complete representation. There are
two types of irreplaceable site. The first, termed globally irreplaceable, comprises
those sites containing strictly endemic species. The loss of such sites would result
in the extinction of those species in the wild. Goal irreplaceable sites are sites that
could be replaced but only by substituting them with two or more additional sites.
The loss of goal-irreplaceable sites would lead to a less efficient solution and would
thus compromise the goal. The identification of irreplaceable sites is key to the
design of efficient protected areas networks. Irreplaceable sites are not negotiable
if the goal of representing all species is to be achieved. Their future is a matter of
policy, not planning. They form the core around which such networks should be
designed.
All other sites within the set of priority areas are termed flexible. A flexible site
can be substituted by another without compromising the goal. Flexibility refers to
the various spatial arrangements of priority areas that are available to achieve the
goal. All flexible sites are negotiable. They provide opportunities for trade-offs with
competing uses.
Once a set of biodiversity priority areas has been identified, the question then
becomes one of establishing priorities for action within that set. Some species are
more vulnerable than others to the effects of the various threats, natural and man-
made, that impinge upon them. In some cases, the threats are quantifiable and this
information can be used to set priorities for action. In addition, because the resources
available for conservation planning are limited, not all areas of the world can be
dealt with equally or at the same time. Assessments of vulnerability can aid in the
initial selection of places where the identification of sets of priority areas is most
pressing.

359
APPLICATION OF WORLDMAP TO OBTAIN SET OF PRIORITY AREAS
A number of heuristic algorithms that incorporate complementarity to identify
irreplaceable and flexible sites have been developed (e.g. Kirkpatrick 1983, Margules
et al. 1988, Margules 1989, Pressey and Nicholls 1989, Rebelo and Siegfried 1992,
Nicholls and Margules 1993). WORLDMAP (Williams 1994) also uses heuristic
algorithms to implement three of the most important elements for measuring and
assessing biodiversity: taxon richness, character diversity and complementarity.
Data files are created by entering the names of the taxa and a clade code (Vane-
Wright et al. 1994) to identify their positions in the cladogram. Distributional records
are then mapped onto individually designed map grids. Three measures of
biodiversity are included: species richness, character richness and character
combination, while a measure of range-size rarity (“endemism”) is also available.
Further details of these measures can be found in Williams (1994) and Williams (in
press). Sets of priority areas can then be determined by using a heuristic ‘near-
minimum’set option. There is a further option included at this stage for selecting
the minimum number of populations of each taxon to be represented, ranging
from 1 to 10.
WORLDMAP proceeds by first selecting the globally irreplaceable areas, that
is those that contain taxa with more restricted distributions than the representa-
tion goal. Other areas are then added in sequence so as to maximize the
complementary endemism score at each step, selecting preferentially those areas
with narrowly distributed taxa so as to find a near-minimum set of areas. Some of
these areas will be goal-irreplaceable in that they contain unique elements of
biodiversity with regard to the near-minimum set. The remaining areas are flexible.
The term “near-minimum” is used because the procedure employed is heuristic. As
such, it only examines a subset of all possible results and cannot guarantee to find
a global minimum set or sets.
The near-minimum set is then checked for redundant areas before being
prioritized using one of two procedures. The first method re-orders the areas by
their overall taxon richness score, without regard for complementarity or tie-
breaking. The second, used in single replicate analyses only, re-sequences the areas
using one of the measures of biodiversity or by endemism, with complementarity.
The results are then mapped onto the screen, with irreplaceable areas highlighted.
Flexible and other types of area are distinguished using other colours.

ASSESSING THE EFFICIENCY OF AN EXISTING RESERVE NETWORK

The procedure described above identifies a set of priority areas that most efficient-

360
ly represents the chosen set of taxa. This set could be used as the basis for an idealized
network of protected areas. But there are numerous constraints on the establish-
ment of an ideal system. Among these are preexisting networks of protected areas,
upon which future expansion and development must, in general, build. Such
networks can be entered into WORLDMAP and used to “seed” an analysis. Such a
constrained analysis can assess the efficiency of the network in achieving the sta-
ted goal and also identify the additional areas required to achieve a given level of
representation of biodiversity.
When such analyses are performed, it is often found that existing protected
areas networks are inefficient (Pressey 1994, Pressey and Tully 1994). In the past,
parks and sanctuaries, areas currently interpreted as affording protection to
biodiversity, have been set aside for numerous reasons, most of which have little
or nothing to do with representing biodiversity. The very first national parks (e.g.
Yellowstone National Park in the USA and Royal National Park near Sydney,
Australia) were designated primarily as areas of outstanding natural beauty and
intended as places of recreation and enjoyment for the general populace. A major
function of national parks in Thailand continues to be tourism (Gray et al. 1991) and
this can lead to conflicts of interest between development and biodiversity
conservation. Reserves may also be designated to protect rare or spectacular spe-
cies considered to be of particular importance. With such a limited aim, such
protected areas may succeed very well. But the most common reason for a particular’
area of land to be offered for wildlife conservation is that it is considered to be of
little use for human habitation or commercial exploitation. Protected areas tend to
be placed where they will cause least interference with extractive land uses (Runte
1979, Kirkpatrick 1987, Pressey and Tully 1994). In other words, they are chosen for
what they are not, rather than what they are. Even when protected areas are
designated because they are considered to include relatively untouched examples
of particular assemblages or ecosystems, there is usually very little information
available on just what and how much biodiversity is contained within them. Data
are generally restricted to large vertebrates and selected plant groups. Furthermore,
rarely is it considered to what extent, if any, a new protected area adds additional
biodiversity to the existing reserve network - i.e. complementarity is not considered.

CONCLUSIONS
Overall, selection of protected areas has frequently been opportunistic and ad hoc.
In some instances, site selection has been undertaken in response to a perceived,

361
although seldom quantified, external threat. Only rarely has a protected area been
established specifically to represent biodiversity. This has had several consequen-
ces for the conservation of biodiversity as a whole. First, many elements of bio-
diversity (taxa, assemblages, ecosystems) that are most in need of conservation are
not yet protected. As a result, the limited resources available for conservation are
being used less efficiently than might otherwise be the case. Many newly gazetted
protected areas simply add more examples of biodiversity that are already
represented within protected areas. They may add little to the biodiversity
complement already present in the existing reserve network (Harper and
Hawksworth, 1994). There is a ceiling on the total area that will be available for
biodiversity conservation. If future additions to networks do not take account of
complimentarity, then this limit may well be reached before all, or even a signifi-
cant proportion of biodiversity has been accounted for in the reserve network. Finally,
there is now an unbalanced representation of biodiversity within existing reserves
(Leader-Williams et al. 1990, Pressey et al. 1993).
Thus, if the goal is to represent all elements of biodiversity adequately, then the
existing methods for selecting and designating protected areas must be reassessed.
However, we must be careful not to judge past actions by our new standards. We
should employ the gift of hindsight to ensure that future decisions are made on a
more informed basis. The methods and procedures outlined above, which will be
substantially elaborated upon by Humphries et al. (in prep.), are designed to remedy
this situation.

ACKNOWLEDGEMENTS
In preparing this paper, we are deeply indebted to all of our colleagues mentioned
in the Introduction, but especially Paul Williams, for access to various papers, both
in press and in preparation.

REFERENCES
Beccaloni, G.W. and K.J. Gaston. (in press). Predicting the species richness of
neotropical forest butterflies: Ithomiinae (Lepidoptera: Nymphalidae) as
indicators. Biological Conservation 71.
Faith, D.P. 1994. Phylogenetic pattern and the quantification of organismal
biodiversity. Phil. Trans. R. Soc. Lond. B 345: 45-58
Gray, D., C. Piprell and M. Graham. 1991. National Parks of Thailand.

362
 Communication Resources (Thailand) Ltd, Bangkok.
Harper, J.L. and D.L. Hawksworth. 1994. Biodiversity: measurement and estimation.
Preface. Phil. Trans. R. Soc. Lond. B 345: 5-12.
Humphries, C.J., C.R. Margules, R.L. Pressey and R.I. Vane-Wright (eds.). in prep.
Priority areas analysis. Systematic methods for conserving biodiversity. Oxford
University Press, Oxford.
IUCN; UNEP and WWF. 1980. World conservation strategy, living resource
conservation for sustainable development. IUCN, UNEP and WWF, Gland.
Janzen, D.H. 1993. Taxonomy: universal and essential infrastructure for development
and management of tropical wildland biodiversity. Pages: l00-113 In: O.T.
Sandlund and P.J. Shei (eds.) Proceedings of the Norway/UNEP Conference on
Biodiversity, Trondheim, Norway. NINA.
Kirkpatrick, J.B. 1987. Forest reservation in Tasmania. Search 18: 138-142.
Kirkpatrick, J.B. 1983. An iterative method for establishing priorities for the selection
of nature reserves: an example from Tasmania. Biological Conservation 25: 127-
134.
Kremen, C. 1992. Butterflies as ecological and biodiversity indicators. Wings 16: 14-
17.
Leader-Williams, N., J. Farrison and M.J.B. Green. 1990. Designing protected areas
to conserve natural resources. Science progress 74: 189-204.
Margules, C.R. 1989. Introduction to some Australian developments in conservation
evaluation. Biological Conservation 50: l-11.
Margules, C.R. and M.P. Austin. 1994. Biological models for monitoring species
decline: the construction and use of data bases. Phil. Trans. R. Soc. Lond. B 344:
69-75.
Margules, C.R., A.O. Nicholls and R.L. Pressey. 1988. Selecting networks of reserves
to maximise biological diversity. Biological Conservation 43: 63-76.
Nicholls, A.O. and C.R. Margules. 1993. An upgraded reserve selection algorithm.
Biological Conservation 64: 165-169.
Prendergast, J., R.M. Quinn, J.H. Lawton, B.C. Eversham and D.W. Gibbons. 1993.
Rare species, the coincidence of diversity hotspots and conservation strategies.
Nature 365: 335-337.
Pressey, R.L. 1994. Ad hoc reservations: forward or backward steps in developing
representative reserve systems? Conservation Biology 8: 662-668.
Pressey, R.L., C.J. Humphries, C.R. Margules, R.I. Vane-Wright and P.H. Williams.
1993. Beyond opportunism: key principles for systematic reserve selection.

363
 Trends in Ecology and Evolution 8: 124-128.
Pressey, R.L. and V.S. Logan. 1994. Level of geographical subdivision and its effects
on assessments of reserve coverage: a review of regional studies. Conservation
Biology 8: 1037-1046.
Pressey, R.L. and A.O. Nicholls. 1989. Efficiency in conservation evaluation: scoring
versus iterative approaches. Biological Conservation 50: 199-218.
Pressey, R.L. and S.L. Tully. 1994. The cost of ad hoc reservation: a case study in
western New South Wales. Aust. J. Ecol. 19: 375-384.
Rebelo, A.G. and W.R. Siegfried. 1992. Where should nature reserves be located in
the Cape Floristic Region, South Africa? Models for the spatial configuration of
a reserve network aimed at maximising the protection of floral diversity.
Conservation Biology 6: 243-257.
Runte, A. 1979. National parks: the American experience. University of Nebraska
Press, Lincoln.
Ryti, R. 1992. Effect of the focal taxon on the selection of nature reserves. Ecological
Applications 2: 404-410.
Vane-Wright, R.I. 1992. 3. Species concepts. Pages: 13-16 In: B. Groombridge (ed.)
Global biodiversity: status of the Earth’s living resources. Chapman and Hall,
London.
Vane-Wright, R.I., C.R. Smith and I.J. Kitching. 1994. Systematic assessment of taxic
diversity by summation. Pages: 309-326 In: P.L. Forey, C.J. Humphries and R.I.
Vane-Wright (eds.) Systematics and conservation evaluation. Systematics
Association Special Volume 50. Clarendon Press, Oxford.
Williams, P.H., 1994. WORLDMAP priority areas for biodiversity. Using version 3.
Privately distributed, London.
Williams, PH. (in press). Biodiversity value and taxonomic relatedness. In: M.E.
WHochberg, J. Clobert and R. Barbault (eds.) Aspects of the genesis and
maintenance of biological diversity. Oxford University Press, Oxford.
Williams, PH. and K.J. Gaston, 1994. Measuring more of biodiversity: can higher-
taxon richness predict wholesale species richness? Biological Conservation 67:
211-217.

364
 Chapter

25
SCANNED, ZAPPED, TIMED, AND DIGITIZED!
ADVANCED TECHNOLOGIE S
FOR MEASURING AND MONITORING
VEGETATION DIVERSITY
H. Gyde Lund1, David L. Evans2 and David S. Linden3

INTRODUCTION
The Convention on Biodiversity, opened for signature at the 1992 United Nations
Conference on Environmental Development, marked a commitment by all the
nations of the world to conserve biological diversity, to sustain biological resources,
and to share the benefits arising from their use. Human subsistence depends on
vegetation either directly or indirectly for food, shelter, fuel, and health. Hence,
vegetation biodiversity surveys are becoming increasingly important at all levels
of interest. Measuring, monitoring and maintaining vegetation diversity is necessary
to the survival of humankind.
Of the many ways to describe vegetation diversity, the most common are extent,
structure, composition, biomass/production, and condition. For this paper:

1 USDA Forest Service, Washington, D.C., USA.

2 USDA Forest Service, Starkville, MS, USA.
3 Colorado State University, Fort Collins, CO, USA.

365
 Extent refers to vegetation area, spatial arrangement, and horizontal diversity.
We often express measures of extent in terms of area, amount of edge (length),
fragmentation, etc. Fragmentation is the deforested area plus an edge effect of
1 km; all isolated forest areas surrounded by deforestation with an area of 100
sq. km and all roads with an edge effect of 0.5 km (Tucker and Skole 1992).
Structure deals with vertical diversity - i.e. whether a forest stand is multi-layered
or not. Units of measure usually include vegetation heights and profiles.
Composition refers to species richness. We usually describe composition by
species number per unit area.
Biomass is vegetation weight, volume, or mass per unit area. When measured
over time, it is a production gauge.
Condition is a measure of health and is an interpretation derived from a de-
sired or preconceived ideal situation. Measurement of condition requires the
use of monitoring and permanent plots.

CONCEPTS IN VEGETATION DATA COLLECTION

While we may want information on vegetation diversity nearly everywhere, we
cannot measure it because of time and costs. Basically, we want to maximize the
vegetal information we can extract from remote sensing and reduce the amount of
time we have to spend in the field. To do this we use remote sensing coupled with
sampling. Through sampling we measure or observe a small part of the popula-
tion with the assumption that our observations represent the entire population.
There are two general approaches to sampling - one is to purposefully select
where we will make our observations. This is subjective sampling. The other is to
use a more random or unbiased method for selecting sample locations, or statis-
tical sampling. Generally, researchers prefer statistical sampling to subjective
sampling. With statistical sampling, we can calculate the reliability of the estimates
generated, whereas we cannot do this with subjective sampling. On the other hand,
subjective sampling is quicker and less costly than statistical sampling and is now
gaining some acceptance in the scientific community. Many people feel it is better
to have some information, subjective as it may be, rather than none.
No matter what kind of sampling one uses, some form of stratification is desira-
ble in advance of sample selection. Stratification is the dividing of the population
into relatively homogeneous classes. Through stratification, one can concentrate

366
samples in higher interest areas - such as forest lands versus agricultural lands.
Thus, stratification is more efficient than not stratifying.
Common themes used for stratification include land cover, vegetation type, soils,
topography (elevation and landform), and land use. Land cover and vegetation
type are the most common themes used for assessing vegetation. Stratification by
land cover and vegetation type requires a quick look at all the land areas of in-
terest. We often use remote sensing from satellites and aircraft for this process.
Sample-based remote sensing from satellites, aircraft (including drones (McGeer
and Holland 1993)), and the field can help in collecting data on vegetal composi-
tion, structure, biomass, and condition. Wall-to-wall remote sensing is then used to
expand the sampled data to the inventory unit as a whole.
Remote sensing is available from various sources such as mineral and oil
companies, military and intelligence departments, survey departments, census
bureaux, utility and electric companies, highway departments, natural resource
agencies, donor organizations, space agencies, and universities. Each may employ
a variety of tools for collecting data.

TOOLS OF THE TRADE

For this paper, we separate remote sensing and associated geo-positioning
instruments into scanners, zappers, and timers. Scanners are passive remote sen-
sing systems that pick up energy that originates from a source away from the sensor.
Examples are photography, videography, thermal imagery, etc.Zappers are active
remote sensing systems that send out an energy pulse that reflects back from the
target area. Examples are synthetic aperture radar and laser altimeters or profilers.
Timers are instruments that receive time signals from a constellation of earth-
orbiting satellites. These receivers are global positioning system (GPS) units that
we use to determine our geographic location.

SCANNED!
To determine vegetal cover extent, we generally use remote sensing. The sensors
may be mounted on a flying platform (helicopters, fixed-winged aircraft, satellites)
or hand-held and used on the ground. With some sensors, especially those that
operate in several parts of the spectrum, we can get an indication of composition
and condition. Imagery acquired in different seasons aids in species identifica-

367
tion, especially in temperate areas. To determine structure and biomass, we often
need high-resolution, stereo remote sensing coverage. Generally, the higher the
resolution, the more costly is the imagery.
We include optical and electro-optical sensors in our discussion of scanners.
A. Optical systems are those that one thinks of as traditional cameras.
1. Aerial photography at scales from 1:12,000 to 1:24,000 is the most widely-
used form of optical remote sensing imagery for vegetation surveys in the
United States. Modern camera systems and films can provide high-resolution
imagery over a broad scale-range. Aerial photographic systems record
reflected energy in the visible and near infrared portions of the spectrum.
Factors that define aerial photography utility include: the coverage, mission
date and time, the scale, film emulsion, the camera format, the lens focal
length, and the atmospheric conditions during the mission.
2. Hand-held photography Ground photos, taken at sample plot are useful
for documenting conditions that are difficult to quantify on a field form.
Vegetation attributes that can be interpreted from ground photography
include species composition, structure, biomass, and health or condition.
Takao (1992) has developed a camera system that can take stereo photo-
graphs in a 360-degree circle around the plot centre. From stereo photo-
graphs, one can make detailed measurements of the vegetation. As with
other forms of remote sensing, periodic photographs taken at the same
location provide a good means of monitoring changes in vegetation (Tappan
et al. 1994).
B. Multi-spectral or electro-optical systems can be configured to get information
from the ultraviolet through the visible, near, middle and thermal infrared to
the microwave portion of the spectrum. The middle and thermal infrared
portions of the spectrum are important in identifying and assessing the condi-
tion of vegetation. Sensing in several parts of the spectrum classes can help in
vegetation life form and sometimes species identification (Lillesand and Keifer
1979).
There are two classes of electro-optical sensors: non-imaging and imaging. Non-
imaging sensors acquire individual measurements rather than an array of
measurements that form an image. Spectrometers mounted on aircraft or bucket
trucks can acquire reflectance measurements from scene elements (vegetation,
water bodies, etc.) to calibrate airborne and satellite imagery or develop reference
data.

368
 These electro-optical systems are not limited by the sensitivity of chemical
reactions that occur when reflected light strikes the film in an aerial camera to
create an image. Information from electro-optical systems may be recorded in
analog or digital format. Video systems capture data in analog form, but most
electro-optical systems convert the incoming energy directly to digital data.
Although generally of lower spatial resolution than aerial photography, electro-
optical sensor data have advantages for natural resource applications. Image
analysts can directly manipulate the digital imagery using computer-based sys-
tems to rectify, classify, enhance, and display the imagery (Robinson and Dewitt
1983 and Norwood and Lansing 1983).
1. Satellite based systems.Digital imagery from remote sensors carried aboard
earth-orbiting satellites provides extensive area coverage.
Meteorological or weather satellites provide information for specialized natural
resource applications. Geo-synchronous satellites provide synoptic low resolution
and hourly coverage. Imagery from the Advanced High Resolution Radiometer
(AVHRR) carried aboard the United States National Oceanic and Atmospheric
Administration (NOAA) series of satellites has been used in assessing forest fuel
condition in Western United States and for developing national forest cover maps
for both the U.S.A. and Mexico (Eggen-McIntosh and Zhu 1992, Evans et al . 1992,
Zhu 1994, and Zhu and Evans 1994). AVHRR imagery has a nominal resolution of
1.1 kilometre at nadir, and daily coverage. A “scene” covers an approximate area of
1750 x 6000 km.
Multi-date AVHRR data are valuable for basic forest/non-forest mapping,
landcover change detection, and trend documentation in vegetation conditions,
especially in the temperate zones. We can get AVHRR data daily and therefore may
be able to develop near-cloud-free composites based on several consecutive days
of imagery. Furthermore, we can use these products, compiled over a one year
interval, to identify phenological vegetation characteristics in development of
spectral classifications for monitoring programs.
Development of an AVHRR-based vegetation cover map uses digital process-
ing techniques such as those described by Loveland et al. (1991) and Zhu and
Evans (1992 and 1994). Briefly, these procedures include: 1) development of cloud-
free, multi-date composites of the AVHRR data, 2) estimation of percent land-
cover components within AVHRR pixels, 3) classification of the AVHRR data into
land-cover categories, and 4) verification of the products by use of high-resolution
satellite (Landsat and SPOT) and other ancillary data (aerial photography, radar).

369
 The United States Landsat and French SPOT (Systeme Probatoire d'Observation
de la Terre) satellites provide easily accessible imagery with global coverage. Cir-
cling the earth in near-polar sunsynchronous orbits, the sensors aboard these
satellites acquire imagery at a consistent solar time during each daylight pass. Repeat
vertical coverage is available from a single Landsat satellite on an approximate 16
day cycle. When multiple satellites in the same series are operating, the repeat
frequency of vertical coverage is proportionally increased.
The current Landsat satellites (4 and 5) carry the Multi-spectral Scanner (MSS)
and the Thematic Mapper (TM). Both instruments are mechanical scanners that
employ a rotating mirror to acquire data in the cross track direction. A full Landsat
scene covers a land area of 185 by 185 kilometers. The Thematic Mapper has a
resolution of 30 meters in six bands of reflected energy extending from the blue
portion of the spectrum to the middle infrared and an emissive thermal infrared
band with approximately 120 meters resolution. Thematic Mapper data have been
available since 1982. There have been many successful examples of using Landsat
TM for mapping vegetative cover, forest condition, and type. The mapping of old
growth vegetation, forest type and structure in the Pacific Northwest Region of the
United States is one example (Steffenson and Wilson 1993). Landsat TM has also
been used to monitor subtle changes over time in vegetation such as that reported
on the Mark Twain National Forest in the U.S. (Platt et al. 1993).
The Multi-spectral Scanner has an 80-meter resolution in four spectral bands
in the green, red, and near-infrared portions of the spectrum. Multi-spectral scanner
data have been available since 1972. The current Landsat 5 is the last satellite in the
series to carry a multi-spectral scanner instrument. Although of significantly lower
resolution than the Thematic Mapper, MSS data are available for a span of more
than 20 years starting in 1972, making the data especially suitable for evaluating
landscape change.
The French SPOT satellites carry two High Resolution Visible (HRV) instru-
ments. Unlike the instruments carried aboard the Landsat satellites, the HRVs are
solid state instruments that image the entire swath of the flight path simultaneous-
ly. Each of these sensors aboard SPOT 1, 2 (in orbit) and 3 can acquire imagery in
the green, red, and near-infrared portions of the spectrum. SPOT 4, scheduled for
launch in the middle of the decade, will add a mid-infrared band to the SPOT HRVs.
A full SPOT scene covers a ground area of 60 by 60 kilometers. SPOT multi-spectral
imagery has a resolution of 20 meters. The SPOT HRVs also can get panchromatic
imagery with ten meter resolution. The capability to point these sensors off-nadir,
parallel to the spacecraft ground track, permits the acquisition of additional ima-

370
 We also can use aerial video for detailed analysis of forest attributes (Evans
and Beltz 1992). One scheme could employ video as a sampling tool to assess
large areas for specific information associated with forest health. We can link this
type of survey to information collected at field monitoring plots. Ground plots
would be established at a uniform density across all forest lands. These field plots
would be geo-referenced with global positioning system (GPS) units. We would
then use the plot coordinates for video mission planning and execution to ensure
accurate overflights of the field locations.
Video imagery would provide inexpensive and fairly detailed information
about the canopy structure for field plots and all areas along transects between
the plots. Stereo pairs from video can be used to generate 3-D anaglyphs for cano-
py characteristics visualization. One also can measure tree and stand attributes
using digital video and photogrammetric techniques. Information derived from
these analysis techniques can help researchers evaluate current conditions and
changes in the forest canopy over time.
Video systems have lower resolutions than comparable photographic systems
and currently lack calibration necessary for precision photogrammetric applica-
tions. They are well suited for many natural resource applications requiring sam-
ple or small area coverage. They are also cost effective for locating features such as
isolated groups of insect-damaged trees within a larger survey area. System
operators can evaluate video data during acquisition and change mission para-
meters as necessary. Improvements in camera design and the availability of high-
er definition recording formats such as Super VHS and HI8 video have increased
the resolution and utility for natural resource application. Image analysts can
manually interpret video imagery using a high resolution monitor and a playback
unit with freeze-frame capability. For enhancement and geo-referencing, the ana-
log data in individual video frames can be captured as digital data using a video
“framegrabber.” The low cost of video systems make them a good candidate for
many monitoring applications (Myhre et al. 1991).
Airborne electrooptical remote sensing systems cover a broad range of
capabilities. Airborne systems support working requirements and serve as test
beds to test new sensor designs. Nixon et al. (1985) showed the utility of multi-
band videography to assess vegetal condition and species.
One must recognize that the data from many current airborne digital remote
sensing systems are difficult and expensive to register to ground coordinates. In
addition, specialized software and knowledge may be necessary to extract useful
information from these data. Nevertheless, airborne systems have an extremely

373
3. Digital Cameras -The use of digital cameras is very new in forestry applica-
tions. The camera records images on a hard disk integrated with the camera.
The image is transferred via a SCSI or parallel connection to the computer or
transferred directly to the computer’s hard disk by the same connections.
Preliminary results by Bobbe et al. (1994) show that digital camera systems
mounted in aerial platforms can provide good quality imagery under a variety
of conditions. Digitized photographs from the ground can be used in fractal
analysis to help evaluate the health and vegetation vigour (Mizoue and
Masutani 1993).

ZAPPED!
Active remote sensing systems are those where the energy recorded initiates
from the sensor. Zappers include radar and lasers. Radar can provide area cover-
age whereas lasers provide point data.
A. Radar. Radar is an active remote sensor using reflected radio signals. The all-
weather capability of synthetic aperture radar (SAR) systems to collect
information makes them ideal for use in tropical forest regions with frequent
cloud coverage. Radar imagery can be collected from satellite and aircraft
platforms. The European Space Agency’s ERS-1 remote sensing satellite and
the proposed Canadian RADARSAT have view areas of 50 x 50 to 500 x
500 km. RADARSAT will have a ground resolution of about 25 x 28 m and
will be useful for monitoring severe changes in forest cover for areas > 100 ha.
Airborne SAR can be used to complement Landsat TM and SPOT for
information on geomorphology and vegetation texture, particularly if stereo
data are acquired. It is also useful for detecting changes in vegetation at larger
scales (Ahem 1994). Short wavelength radar may penetrate upper vegetation
layers and may provide information about forest understory diversity and the
ground. Radar also may provide forest volume estimates (Wu 1990) and bio-
mass assessments (Dobson et al. 1992 and Hussin et al. 1992). However, these
predictive measurements may be dependent on terrain characteristics (van Zyl
1993). Other work has demonstrated the possibility that radar imagery could
have utility in species group separation (Leckie 1990).
Used in combination with aerial videography and other sensor data and
ground data, radar has the potential to provide information on forest biodiver-
sity. Aerial video and high-resolution radar can supply information at the stand
and plot level for detailed forest canopy characterizations. Ideally, radar sys-

375
 terns should be capable of collecting data in short wavelengths with the
anticipated detection of multi-storied characteristics within tropical forests. An
aerial video system (colour or multi-spectral) could be flown with the radar.
This detailed information will be invaluable for forest health or condition
monitoring.
B. Lasers. Lasers operate by sending out a short burst of light timed to determine
how long the light takes to travel to a target. This time is converted into dis-
tance (Carr 1993). We use hand-held lasers to measure distances to trees and
their heights. Laser profilers, mounted in aircraft, can measure vegetation heights
above the terrain (Ritchie and Weltz 1992). When coupled with airborne
videography and global positioning systems, vegetation structure and biomass
may be determined.

TIMED!
Although they were introduced less than ten years ago to the forestry community,
almost everyone is now familiar with the use of GPS receivers and satellites to
decide one’s position. With a GPS receiver and time signals sent from a constella-
tion of earth-orbiting satellites, we can determine our position to within 30 to 100
meters at any location on our planet. With two receivers and with one located on a
known location (base station), we can determine our location to within centime-
tres using differential calculations (Hurn 1989).
There have been several articles in the past two years dealing with GPS and
aerial videography integration. Evans (1992) demonstrated the usefulness of
GPS with aerial videography and recommended the use of a gyro-stabilized camera
mount to minimize the effect of aircraft attitude variations. Bobbe (1992) dis-
cusses the use of real-time differential GPS with airborne videography and dis-
cusses how the GPS data can be used to mosaic and geometrically correct digitized
video data using manual methods. Bobbe et al. (1993) discuss similar procedures
using a multi-spectral video camera. Graham (1993) discusses how Society of Motion
Picture and Television Engineers (SMPTE) time coding is used to synchronize the
GPS and video data. SMPTE time codes allow digital data to be stored on the se-
cond audio track of a video tape. One time code is stored for each frame. Each code
holds 80 bits of data: 32 bits for GPS time, 32 bits of “user” data, and 16 bits of
synchronized pattern.

376
DIGITIZED!
Digitization refers to the placement of geographic coordinates on observations for
use in a Geographic Information System (GIS). Getting data into a GIS is essential
for resource modeling. As indicated above, many forms of remote sensing
automatically store data in digital form.
With the evolution of GPS, the determination of coordinates provides a quick
way to digitize field information so it may be entered into a GIS. GPS units, cou-
pled with field data recorders and portable computers, tag locational information
to field observations.
Using GPS receivers, it is possible to quickly and accurately register remote
sensing imagery for entry into a GIS. Linden et al. (1993) have developed a tech-
nique for using GPS to automate the digital mosaicing process of airborne
videography.
We can use analytical stereo-plotters to make precise measurements (heights,
lengths, widths) of individual plants. We also can use the same instrument to digi-
tize information for entry into a GIS. In addition, there are many types of scanners
and line followers available to convert photographs and maps to digital form for
entering into a GIS (Gibso et al. 1983).
Once data are in a GIS, species occurrence and richness may be modelled from
remote sensing, topographic and climatic data (Podolsk et al. 1992 and Steffenson
and Wilson 1993). Through a GIS and appropriate models, sample data can be
extrapolated using stratification criteria for information on vegetation extent,
composition, structure, biomass, and condition portrayed for the entire inventory
unit.

CONCLUSIONS AND RECOMMENDATIONS

Based upon our experience and observations we have four conclusions and
recommendations:
1. New technology is available to measure and monitor key vegetation diversity
attributes. Large area1 extent of vegetation can be determined from satellite
imagery. Seasonal and multi-spectral imagery is useful for determining
overstorey composition and condition. Modelling may have to be used for
understorey composition. Structure and biomass usually require some height
estimates that may be obtained through stereo imagery, radar or laser profilers.
Rapid updates of conditions on small areas can be done using airborne

377
 videography or digital cameras. Global positioning units are useful in linking
remote sensing and field plots with a GIS. All remote sensing efforts, however,
need ground verification either for accuracy assessment or to provide
information that one cannot get directly from the imagery.
We need continued research to learn the extent to which these technologies
can be used in multistage or multiphase sampling schemes. Training and
technical assistance in implementing the technology can be provided by
equipment vendors and by agencies such as the USDA Forest Service when
linked to government requests and agreements.
2. To be most effective, collection of vegetation diversity information should be a
part of a regular or multi-resource inventory program. Collecting data for bio-
diversity and then later visiting the same area for a timber or range resource
survey is needlessly wasteful. Concepts for integrating inventories are presented
in Lund (1986).
3. What to measure is not so much the question as where to measure. A common
understanding of just what is forest land is in itself a problem - forest land may
be defined administratively, by land use or by land cover. Land cover is the
most easy and most consistent attribute to detect from remote sensing .However,
vegetation diversity, or lack of it, is of concern on all lands - urban, agricultural,
rangelands, wetlands, as well as forest lands. We recommend that all lands be
inventoried and monitored for the five vegetation components discussed in
this paper.
4. Measurement of vegetation diversity is not a problem how to present the
information to the analyst and decision maker is. Those that design and carry
out inventories should strive to collect and present data in an unbiased man-
ner. Interpretations should be left to those for whom the inventory system was
designed.
We hope that through this paper we have introduced some emerging
technologies that are available for measuring and monitoring vegetation
diversity. Readers are encouraged to consult the references provided in this
paper for more details about specific technology.

ACKNOWLEDGEMENTS
Our thanks to the conference organizers for allowing us to present this paper. We
also thank Stan Bain, Bill Clerke, and Henry Lachowski (all USDA Forest Service

378
employees) for their input. Lastly, we thank Jim Culbert and Rich Calnan (USDA
Forest Service) for their support.

REFERENCES
Ahern, F. 1994. The tropical forest initiative: analyzing the data. Remote Sensing of
Canada 22(l): 2-3.
Bobbe, T.J. 1992. Real-time differential GPS for aerial surveying and remote sen-
sing. GPS World (July/August): 18-22.
Bobbe, T.J.; M.L. Hoppus and D.T. Evans. 1994. The evaluation of a small format
digital camera for aerial surveys. In: J. Greer (ed.) Proceedings Fifth Biennial
USDA Forest Service Remote Sensing Applications Conference. 11-15 April
1994. Portland, OR. Bethesda, MD: American Society for Photogrammetry and
Remote Sensing (in press).
Bobbe, T.J.; D. Reed and J. Schramek. 1993. Georeferenced airborne video imagery,
natural resource applications on the Tongass. J. Forestry 91(8): 34-37.
Carr, B. 1993. How to use laser technology to measure tree heights and survey stand
boundaries. Pages: 249-258 In: H.G. Lund, E. Landis and T. Atterbury (eds.)
Stand Inventory Technologies 92. Proceedings of the Stand Inventory
Technologies -An International Multiple Resource Conference. 13-17 September
1992. Portland, Oregon. Bethesda, MD: American Society for Photogrammetry
and Remote Sensing.
Dobson, M.C., E.F. UIaby, T. LeToan, A. Beaudoin, E.S. Kasischke and N. Christensen.
1992. Dependence of radar backscatter on coniferous forest biomass. IEEE Trans.
Geosc. and Remote Sensing 30(2): 412-415.
EggenMcIntosh, S., Z. Zhu. 1992. Mapping Mexico’s forest cover: an international
perspective. Pages: 88-93 In: J.D. Greer (ed.) Remote Sensing and Natural
Resource Management: Proceedings of the Fourth Forest Service Remote Sensing
Applications Conference. 6-11 April 1992. Orlando. Bethesda, MD: American
Society for Photogrammetry and Remote Sensing.
Evans, D.L. 1992. Using GPS to evaluate aerial video. GPS World (July/August):
24-29.
Evans, D.L. and R. Beltz. 1992. Aerial video for support of forest inventory.
Proceedings of the 3rd Biennial Workshop on Color Aerial Photography and
Videography in the Plant Sciences, Orlando, FL. Bethesda, MD: American Society
for Photogrammetry and Remote Sensing: 192-198.

379
Evans, D.L., Z. Zhu, S. Eggen-McIntosh, P Garcia Mayoral, J.L. Omelas de Anda.
1992. Mapping Mexico’s forest lands with advanced very high resolution
radiometer. Res. Note SO-367. New Orleans, LA: U.S. Department of Agriculture,
Forest Service, Southern Forest Experiment Station. 4 p.
Gibson, J.R., R.A. O’Neil, R.A. Neville, S.M. Till and W.D. McColl. 1983. A stereo
electro-optical line scanner imager for automated mapping. Proceedings of the
6th International Symposium on Automated Photography, Vol.II, Ottawa: 165-
176.
Graham, L.E. 1993. Airborne video for near-real-time vegetation mapping. Journal
of Forestry 91: 28-31.
Hurn, J. 1989. GPS - a guide to the next utility. Sunnyvale, CA: Trimble Navigation
Ltd. 76 p.
Hussin, Y.A., R.M. Reich and R.M. Hoffer. 1992. Effect of polarization of radar
backscatter in relation to slash pine stand biomass using aircraft SIR-B data.
Pages: 532-541 In: Proceedings ASPRS/ASCM/RT 92 Technical Papers Volume
5 - Resource Technology 92. Bethesda, MD: American Society for
Photogrammetry and Remote Sensing.
Lachowski, H. 1990. Report to the Remote Sensing Steering Committee. Remote
Sensing Working Group Notes. Unpublished Report. Salt Lake City, UT:
Nationwide Forestry Applications Program, Integration of Remote Sensing. 20
p.
Leckie, D.G. 1990. Synergism of synthetic aperture radar and visible/infrared data
for forest type discrimination. Photogramm. Engr. and Remote Sensing 56(9):
1237-1246.
Lillesand, T.M. and R.W. Keifer. 1979. Remote sensing and image interpretation.
New York: John Wiley and Sons. 612 p.
Linden, D.S., D.P. Mueller and L.G. Lewis. 1993. Airborne video automated
processing. Final Report. Fort Collins, CO: U.S. Department of Agriculture; Forest
Service; Methods Application Group. Misc. pagination.
Looyen, W.J., F.O. Zeppenfeld, J.C. Venema, G. van der Burg, Dirk H. Hoekman, J. J.
van der Sanden, PG. Reichert and E. Westinga. 1994. FAME: a forest assessment
and monitoring environment scenario. Pages: 317-329 In: A. Singh (ed.)
Proceedings UNEP and IUFRO International Workshop in cooperation with FAO
on Developing Large Environmental Data Bases for Sustainable Development.
14-16 July 1993. Nairobi, Kenya. GRID Information Series No. 22. Sioux Falls,
SD: United Nations Environment Programme.

380
Loveland, T.R., J.W. Merchant, D.O. Ohlen and J.F. Brown. 1991. Development of a
landcover characteristics database of the conterminous U.S. Photogramm. Engr.
and Remote Sensing 57(11): 1453-1463.
Lund, H.G. 1986. A primer on integrating resource inventories. Gen. Tech. Report
WO-49. Washington, DC: U.S. Department of Agriculture; Forest Service. 64 p.
Mizoue, N. and T. Masutani. 1993. Application of fractal dimension to quantifying
form of tree crowns. In: Proceedings of the IUFRO Conference on Advancement
in forest inventory and forest management sciences. 20-25 September 1993. Seoul,
Korea. Seoul, Korea: Forestry Research Institute: 133-138.
McGeer, T. and G. Holland. 1993. IDNDR supports tiny robotic aircraft for tropical
cyclone reconnaissance. STOP Disaster 15(Sept.-Oct.): 4-5.
McLucas, J. 1994. Converting coveted spy technology. Geo Info Systems 4(2): 51-52.
Myhre, R.J., L. Graham and C. Sumpter. 1991. Airborne videography a potential
tool for resource managers. Pages: 191-194 In: Proceedings: Resource Technology
90. 12-15 November 1990; Washington, DC: Georgetown University. Bethesda,
MD: American Society for Photogrammetry and Remote Sensing.
Nixon, P.R., D.E. Escobar and R.M. Menges. 1985 Use of multiband video system
for quick assessment of vegetal conditions and discrimination of plant species.
Remote Sensing of Environment 17: 203-208.
Norwood, VT. and J.C. Lansing Jr. 1983. Electra-optical imaging sensors. Pages:
335-367 In: Robert N. Colwell (ed.). Manual of Remote Sensing, Vol. I. Bethesda,
MD: American Society for Photogrammetry and Remote Sensing: Bethesda, MD.
Platt, B. M. Schanta, H. Lachowski, P. Maus, V. Landrum and J. Johnson. 1993. Forest
plan monitoring: role of remote sensing and GIS on the Mark Twain National
Forest. Pages: 342-348 In: H.G. Lund, E. Landis, T. Atterbury (eds.) Stand
Inventory Technologies 92. Proceedings of the Stand Inventory Technologies -
An International Multiple Resource Conference. 13-17 September 1992. Portland,
Oregon. Bethesda, MD: American Society for Photogrammetry and Remote
Sensing.
Podolsky R., J. Freilich and R. Knehr. 1992. Predicting plant species richness from
remotely sensed data in a high desert ecosystem. Pages: 57-63 In: Proceedings
ASPRS/ASCM/RT 92 Technical Papers Volume 4 - Remote Sensing and Data
Acquisition. Bethesda, MD: American Society for Photogrammetry and Remote
Sensing.
Ritchie, J.C. and M.A. Weltz. 1992. Using airborne laser to measure vegetation
properties. Pages: 395-404 In: Proceedings ASPRS/ASCM/RT 92 Technical

381
 Papers Volume 4 - Remote Sensing and Data Acquisition. Bethesda, MD:
American Society for Photogrammetry and Remote Sensing.
Robinson, B.F. and D.P. Dewitt. 1983. Electra-optical non-imaging sensors. Pages:
293-333 In: R.N. Colwell (ed.) Manual of Remote Sensing, Vol. I. Bethesda, MD:
American Society for Photogrammetry and Remote Sensing.
Steffenson, J.R. and A.E. Wilson. 1993. Existing vegetation mapping and vegetation
zone and plant association group models. Pages: 63-64 In: H.G. Lund, E. Landis,
T. Atterbury (eds.) Stand Inventory Technologies 92. Proceedings of the Stand
Inventory Technologies - An International Multiple Resource Conference. 13-17
September 1992. Portland, Oregon. Bethesda, MD: American Society for
Photogrammetry and Remote Sensing.
Takao, G. 1992. Photogrammetric measurement of stand disposition and each tree’s
dbh with a rotatory camera. Pages: 388-397 In: Proceedings of the Symposium
on Integrated Forest Management Information System. 13-18 October 1991.
Tsukuba, Japan. Tokyo, Japan: Japan Society of Forest Planning Press.
Tappan, G., E. Wood, A. Hadj and A. Bodian. 1994. Monitoring natural resource
changes in Senegal: preliminary evidence from field studies. Draft Report. Sioux
Falls, SD: U.S. Geological Survey; EROS Data Center. 37 p.
Tucker, C.J. and D.L. Skole. 1992. Tropical forest habitat fragmentation in the legal
Amazon of Brazil from 1978 to 1988. Abstract. Pages: 27-29 In: Proceedings World
Forest Watch Conference, May 1992. Sao Jose dos Campos, Brazil. Ispra, Italy:
Commission of the European Communities, Joint Research Centre: 65.
van Zyl, J.J. 1993. The effect of topography on radar scattering from vegetated areas.
IEEE Trans. Geosc. and Remote Sensing 31(l): 153-160.
Wu, ST. 1990. Assessment of tropical forest stand characteristics with multipolarized
SAR data acquired over a mountainous region in Costa Rica. IEEE Trans. Geosc.
and Remote Sensing 28(4): 752-755.
Zhu, Z. 1994. Forest density mapping in the lower 48 States: a regression procedure.
Res. Pap. SO280. New Orleans, LA: U.S. Department of Agriculture, Forest
Service, Southern Forest Experiment Station. 11 p.
Zhu, Z. and D.L. Evans. 1992. Mapping Midsouth forest distributions: AVHRR
satellite data and GIS help meet RPA mandate. J. For. 90(12): 27-30.
Zhu, Z. and D.L. Evans. 1994. U.S. forest types and predicted percent forest cover
from AVHRR data. Photogramm. Engr. and Remote Sensing 60(5): 525-531.

382
 Index

A
a-amylase 274
Abies alba 170
Adaptive potential M-154, 165-171
Advanced High Resolution Radiometer 369, 371, 374
Aerial photography 368
Agaonidae 260
Agenda 21 129, 263
Agriculture 21, 54
Agriculture, shifting cultivation 21, 312~313, 318, 325-326, 329, 344
Albizia spp. 252, 292
Alleles 23, 196-203, 216, 222, 229
Allozymes see Isozymes
Alstonia boonei 250
Amphibians 9, 89, 291
Amphimas pterocarpa 250
Aningeria robusta 250
Anisoptera spp. see Diptexocarpaceae
Anthonotha spp. 248,250
Anthropogenic habitats (see also Disturbed
habitats) 20,50
Antiaris toxicaria 250, 252-253
Antidesma laciniatum 250
Antrophium spp. 315
Ants 24, 28-30, 41, 266
Aphelinidae 260
Archontophoenix spp. 275-276
Arthropods 20, 31, 257-267
Artocarpus spp. 292
Ascomycetes 276, 280
Asdepias spp. 240
Ashbya gossypii 274
A s p e r g i l l u s spp. 274
Asystasiella nessiana 240
Atta cephalotes 28
Australia 34, 105, 271-281, 361
AVHRR see Advanced High Resolution Radiometer

383
Azadirachta spp. 121

B
Bacillus subtilis 275
Baphia spp. 248, 250
Baphicacanthus cusia 240
Barbets 330
Basidiomycetes 280
Binary data 8, 11, 16, 82
Biodiversity, definition 1,6, 113,127, 177
Biodiversity, loss of, causes 2, 12, 114, 127
Biodiversity, prospecting 22, 139
Biotechnological products 273-275
Birds (see also Ducks, Herons, Hornbills, etc) 6, 9, 30, 32, 60, 71, 75, 89, 288, 290-299, 330
Boerhavia chinensis 240
Bombax spp. 240, 292
Brandypus variegatus 58-59
Brahminy kite 292
Brazil 53-57
Breeding systems 23, 181, 186, 215-216, 223, 229-230
Brundtland Commission see World Commission on the Environment and
Development
Brunei 259
Bussea occidelitalis 248, 250
Butterflies 27, 32, 261, 288, 290-299, 357

C
Caluromys philander 58
Calyptothecium recurvulum 316
Canada 12, 15
Cananga latifolia 240
Canarium spp. 240, 292
Canopy fogging 259
Carabid beetles 262
Cecropia spp. 315
Celtis spp. 250, 252-253
Centropus andamanensis 297
Cerambycids 20
Chalcidoidea 259-260
Chalcites xanthorhynchus 296
Chiamocleis shudikarensis 58-59
Chironimidae 262
Chlo rophora excelsa 250, 254-255
Christella papilio 315
Chromoleana odorata 292, 315
Chrysophyllum albidum 250

384
Cicindelidae 262
CITES see Convention on International Trade in Endangered
Species
Citric acid 274
Cladogram 360
Clavicornaltica spp. 259
Cleistopholis patens 250
Climate change 21-23, 178
Climate Change Convention 263
Cocos nucifera 275
Coelomycetes 276-277
Cola spp. 248, 250
Collector s curves see Species accumulation curves
Columba palmnhoides 295-297
Complementarity 8,11,12,358-361
Connectivity 48,312
Conservation biology 68
Conservation, ex situ 220
Conservation, in situ 208, 217, 220, 341
Conservation, planning 22, 93, 335, 345
Conservation, priorities 22, 33, 40, 178, 336, 342, 355-362
Conservation, value 7, 19, 22
Contagion 50-51, 54, 56-57
Convention on Biological Diversity 129-130, 137, 263, 365
Convention on International Trade in
Endangered Species 129
Convulvulaceae 315
Copsychus malabaricus 295
Coracina striata 296
Cordyceps spp. 273
Corridors 74, 100
Corynanthe pachycera s 248, 250
Costa Rica 20, 28-33, 281
Cotylelobium spp. see Dipterocarpaceae
Cratoxylon formosum 239-240
Cuculus micropteris 295
Cyathea borneensis 315, 320
Cyclosporine 274

D
Dacryodes klaineana 250
Dalbergia spp. 121, 240, 292
Daniellia ogea 250
Databases
Deforestation rates 114, 237
Degradation 344, 345

385
Dendrocitta bayleyi 295-297
Desplatzia subericarpa 250
Dialium dinklagei 248, 250
Dicranopteris curranii 315
Dicrurus andantanensis 297
Dioscorea hispidn 240
Diospyros spp. 248, 250
Diplazium esculentuer 315
Dipterocarpaceae 40, 119, 121, 210-217, 240, 242-243, 316, 330
Dipterocarpus spp. see Dipterocarpaceae
Discoglypremna caloneura 250
Distemonanathus benthamianus 250, 253
Disturbance 311-331, 340
Disturbed habitats ( s e e also Anthropogenic
habitats) 24
Diversity, alpha 22, 30, 90
Diversity, beta 22, 30, 90
Diversity, community 36, 98
Diversity, definition 5
Diversity, ecosystem 2, 3, 116, 123-l24, 337-338, 342, 384, 351
Diversity, gamma 22, 30
Diversity, genetic 2, 3, 36, 98, 104, 107~108, 120-121,134, 138, 145-171,
177-187, 209, 215, 222-223, 230, 337-338, 340-341
Diversity, habitat 12, 98
Diversity, landscape 49, 98, 105, 108, 337, 342
Diversity profile 70-68
Diversity, species 2, 3, 36, 98, 104, 107-108, 119, 122, 135, 237-244, 248,
259, 278, 303-306, 337-338, 340, 342, 344, 348, 350-351
Diversity, structural 32, 92
Diversity, taxonomic 91-93
DNA fingerprinting 35
DNA micro-satellites 35
Dominance 50, 119
Douglas fir 71
Dryocopus javensis 295
Ducks 292
Dyera costulata 210-213

E
Economic valuation 20, 34
Ecosystem stability 147-l48
Ecosystem, function 21, 97
Ecosystem, structure 21
Edge effects 59
EIA see Environmental Impact Assessments
Elasmidae 260

386
Emericella quadrilineata 274
Enarltia polycarpa 250
Encyrtidae 260
Enicospilus madrigalae 28
Endemism 22, 25, 32-33, 36, 290-291, 296-297, 335, 337, 342, 348,
350~351, 359-360
Endophytes 275
Entandrophragma spp. 250, 252-255
Environmental Impact Assessment 139
Eperua duckiana 348-350
Epiphytes 315, 317, 324, 327
Eucalypts 24
Eucharitidae 260
Euglossine bees 58, 60
Eulophidae 260
Eupatorium 240
Eupelmidae 260
Euphorbiaceae 315
Eurystomus orientalis 295
Eurytomidae 260
Evenness 49, 68, 155-157, 304, 310
Evolution 177-187, 311
Evolutionary forces 180-181
Exotics 21, 47, 316, 324, 327
Extinction processes 179
Extinction rates 127
Extinction, causes 12, 179

F
Faidherbia albida 39
Felis onca 58-59
Ficus sp 250
Firewood see Fuel wood
Floribunda pseudofloribuuda 316
Forest management 130-132, 207-217,243
Fractal dimension 50-51,56-57,82,98-99
Fragmentation 21, 47-48, 336, 342-345, 348, 350-351, 366
Frequency profiles 153-165
Fuelwood 21
Fumaric acid 274
Funtumia spp 250

G
Gap analysis 73-75, 82-83
Gap dynamics 311-331
Gap crossing ability 50, 59, 61

387
Genes see Alleles
Gene flow 23, 121, 165, 180-181, 185
Genetic distance 169-171, 233
Genetic drift see Evolutionary forces
Genetic resource area 209-217
Genetic variation (see also Diversity, genetic) 23,27,40,195-203,222,229-230
Genomes 179-180, 184
Genotype 148-149, 151, 167-168, 197-204, 229
Geographic information systems 37-39, 48, 52, 54, 61, 72-74, 335-351, 377-378
Germplasm 21
Ghana 23, 38, 247-255
Gibbons 345, 347
GIS see Geographic information systems
Global positioning systems see Timers
Gluconic acid 274
Glyphaea brevis 250
GPS see Timers
Graminaceae 280, 315, 319, 324, 327
Grazing 21
Great Britain see United Kingdom (incorporating Great Britain)
Griseofulvin 274
Guibourtia ehie 248, 250, 254
Guyana 349

H
Habitat mapping 76-82
Habitat fragmentation see Fragmentation
Harvesting see Logging
Heliconiini 262
Hepaticae 316
Heritiera utilis 252, 254
Herons 292
Heterozygosity 197-203, 216, 222-223, 229
Hierarchy theory 102-108
Himanthocladium plumula 316
HirteIla physophora 348-350
Holoptelea grandis 251, 253
Homaliodendro n flabellatum 316
Hopea spp. see Dipterocarpaceae
Hombills 292-330
Hymenodictyon excelsum 240
Hymenophyllaceae 315
Hymenoptera 20, 31, 33

I
Ichneumonids 20, 28, 33

388
Imperata cylindrica 344
India 24, 32, 227-229, 230-231, 233-234, 287-299
Indicator, biodiversity 31, 36, 138
Indicator, ecological 31
Indicator groups 31, 258-264, 278, 314-331, 357
Indices, biodiversity 9-10, 35, 310-331
Indices, diversity 6, 10, 68-70, 89, 91-92, 107-108, 154, 304-305
Indices, landscape 46-57, 61
Indices, Shannon (Wiener) 8, 49, 69-70, 89, 119, 122, 239, 243, 305
Indices, similarity 243-244
Indices, Simpson s 8, 49, 69-70, 89, 305
Indices, spatial 59
Indonesia 227-228, 230-231, 233-234, 257-259, 313-331
Industrial alcohol 274
Inheritance 180, 182, 183, 185
Integrity 310-331
Inventory 34, 119, 135-136, 261, 280, 287, 310, 340, 356
Isoenzyrnes see Isozymes
lsolona cooperi 251
Isozymes 35, 39, 120-121, 156, 169-170, 181, 184-185, 195-203, 216,
220, 229-234
Itaconic acid 274
Ithomiinae 262, 357
Ivory Coast 228

J
Jack-knifing 69, 71

K
Kalanchoe verticillata 239-240
Keystone species 31
Khaya spp. 34, 252, 254-255
Kingfishers 292
Koompassia malaccensis 210-213

L
Lagerstroemia cuspidata 240
Landsat 369-372, 374
Landscape characteristics 49-57, 100, 102
Landscape, definition of 48-46
Landscape ecology 48-49, 68, 98-102, 104-105, 343
Landscape function 100, 104
Landscape pattern see Pattern
Landscape structure 100-102
lmnea spp. 240, 251
Lantana camara 292, 315

389
Laos 227
Lasers 376
Lianas 315, 319, 322, 324, 327, 329
Licuala spp. 275
Logging 21, 35, 4041, 71-72, 131, 210-216, 290, 309, 311-313,
318, 325-326, 329-330, 343-344
Lophira alata 251
Lovoa tridiilioides 254

M
Macarahnga hurifolia 251
Macropygia rufipennis 296-297
Malaysia 4041, 131-140, 207-217, 219-224, 259, 266, 280, 342
Mandonia altissima 252
Mangrove forest 118, 288
Manilkara littoralis 292
Massularia acumita 251
Mating systems see Breeding systems
Measurement, biodiversity 7-8, 87-90, 19-l 24, 153
Measurement, difficulties of 28-30, 98, 259
Measurement, diversity 7-8, 88
Memecylon afzelii 251
Metarrhizum spp. 273
Mexico 108
Microfungi 271-282
Migration see Gene flow
Millettia spp. 240, 242
Mimosa pigra 273
Minimum (viable) population size 31, 68
Mitragyna 240
Modelling 48, 52, 54, 57, 61, 76, 351, 377
Molecular methods 35, 39, 177-187
Monitoring 22, 39, 87, 100, 105, 107, 114-118, 130, 132, 134-139, 280,
309
Monitoring, objectives 23
Monodora tenufolia 251
Morphospecies 260, 264-267
Moms mesozygia 251, 253
Mosses 118, 266, 316-317, 324, 327, 330
Mouriri oligantha 349-350
Mucor pusilus 274
Mutation see Evolutionary forces
Myanmar 227, 233-234, 288
Mymaridae 260
Myrianthus spp. 251

390
N
Nanclea diderrichii 254
Napoleonaea vogelii 251
Natural selection see Evolutionary forces
Neckeropsis gracilenta 316
Neobalanocarpus spp. see Dipterocarpaceae
Nephrolepis bisserrata 315
Nesogordonia papaverifera 248, 251, 252-254
New Guinea 33
Niche width 9
Nutrient cycle 146-147

0
Ochromza spp. 315
Old-growth forests 9, 61, 330
Operational taxonomic units see Morphospecies
Ormyridae 260
Oriolus xanthornus 296
Orussids 20

P
Panda oleosa 248,251
Parashorea spp. see Dipterocarpaceae
Pattern 47-48, 52, 68, 75-76, 81-83, 93, 100, 104, 237-238
Penicillin 274
Penicillium spp. 274
Pcricopsis ela ta 254
Phallus sp. 273
Phenotype 152-153, 166
Phoma spp. 274
Phylogenetic assessment 33, 35, 92-93, 186, 356
Phylogenies see Phylogenetic assessment
Physiological adaptation 148, 150, 185
Picea abies 195-203
Pinnatella mucronata 316
Pinus spp. 92, 121, 185
Pioneer species 310, 322, 324, 327
Piper aduncum 315
Piptadeniastrum africanum 251-253
Pithecia pitlrecia 58
Poaching 345
Poland 195-203
Pollution 21-23, 27, 170, 201, 203, 262, 277, 344
Polychaetes 266
Pometia pinnata 292
Population pressure 21, 114, 311

391
Population structure 121,184,229
Predictor set 262,357
Preussia auranfica 275
Principal Component Analysis 196-203
Protease 274
Protected areas see Reserve network design, Reserve size
Pselaphidae 265
Pteridium aquilinum 315
Pteridophytes 118, 239
Pterocarpus spp. 121, 239-240, 242, 292
Pterocymbium tinctorium 292
Pteromalidae 260
Pterygota spp. 251, 292
Pycnanthus angolensis 248, 251-254

Q
Quadrats 12, 248
Qualitative variance 10
Quantitative data, difficulties 8

R
Radar imagery 54, 375-376
Rails 292
Randomly amplified polymorphic DNA 35, 182-183, 186
RAPD s see Randomly amplified polymorphic DNA
Rapid assessment 260, 263-264, 311-331
Rarity 9, 22, 25,33, 90-91, 119, 155-164, 178, 360
Rattan 219-224, 275, 316
Recognizable taxonomic units see Morphospecies
Remote sensing 35, 37, 48-49, 52, 134, 309, 316, 345, 366-378
Rennin 274
Research needs 38, 140
Reserve network design 355-362
Reserve size 99, 287-299
Restriction fragment length polymorphisms 35, 181-182, 184-187
RFLP s see Restriction fragment length polymorphisms
Rhizopus nigricans 274
Riboflavin (vitamin) 274
Richness (see also Species richness) 49

S
Saccharomyces cerevisiae 274
Salmalia insignis 292
Saprophytes 275
Satellite imagery 52, 54, 74,
Scale, importance of 12, 29, 68, 82, 98-108

392
Scanners 367-375
Scarab beetles 58, 60, 262
Schleichera oleosa 240
Scottelia klaineana 251
Selective cutting see Logging
Semecarpus reticulata Lec. 240
Shifting cultivation see Agriculture, shifting cultivation
Shorea spp. see Dipterocarpaceae
Signiphoridae 260
Simulation models see Modelling
Slash and burn 54
Sloths 59-60
Spatial ordination 76-82
Species abundance models 303-304
Species counts 7, 22-24,31
Species lists 8-10, 12, 294
Species, number of 33, 118, 272, 276, 290-291, 348
Species richness 7, 10, 12, 15, 21, 24-25, 31-33, 36, 39-40, 49, 68, 71,73-
74, 89, 104, 119, 122, 178, 261-262, 303-304, 340, 359-
360, 366
Species-accumulation curves 13-15, 32,259
Species-area relationship 31, 68, 94, 104, 119
Spiders 266
Spilornis elgini 297
SPOT see System Probatoire d Observation de la Terre
Spotted owl 61,99
Sri Lanka 259
SSR s see DNA microsatellites
Staphylococcus au reus 275
Sterculia spp. 248, 251
Stockholm Declaration 128
Stronrbosia glaucescens 248, 251-254
Sturnus erythropygius 297
Succession 9, 29, 311-312
Sustainable development 20
Sustainable management 27, 123, 237-238
Swallows 292
Swifts 292
System Probatoire d/Observation de la Terre 369-370, 372, 374

T
Tabernaemontana crassa 251
Tanaostigmatidae 260
Tanzania 228, 231
Teak 24, 120-121, 227-234, 239-240, 242
Teminalia spp. 240, 251-253, 292

393
Temporal effects 25,27
Terpsiphone paradisi 295-297
Tetrameles nudiflora 292
Thailand 113-124, 227-228, 230-231, 233-234, 237-244, 345-347,
361
Thrips 261
Tieghemella heckelii 254
Timers 367, 376, 378
Tolypocladium inflatum 274
Torymidae 260
Transects 294-295, 313, 316
Tree breeding 23, 340
Trema spp. 315
Trichilia spp. 251
Trichodermna reesei 274
Trichogrammatidae 260
Trichomanes spp. 315
Triplochiton scleroxylon 251-255
Tropical forests, threats 21

U
Uganda 29, 34-35, 37
UNCED see United Nations Conference on the Environment
and Development
United Kingdom 20, 31-32, 40, 257-259, 261
United Nations 128-129
United Nations Conference on the
Environment and Development 128, 129, 263, 365
United States 22, 35, 71-73, 75, 91,94, 99-100, 361
Urban expansion 21

V
Videography 372-374,376
Vitex pinnata 240

W
Wasps, ichneumonids see Ichneumonids
Wasps, mesostenine 25
Wasps, ophionine 25
WCED see World Commission on the Environment and
Development
Weighting 6-7, 22
Western meadowlark 71
White-bellied sea eagle 292
World Commission on the Environment
and Development 128

394
Wrightia tomentosa 240

X
Xylaria 251, 280
Xylia spp. 240
Xylopia quintasii 251

Z
Zappers 367, 375-376
Zingiberaceae 40, 280, 315, 319, 324, 327

395