(1-4).

Great technological progress in the past half century

Resource-Conserving Agriculture has not been reflected in major reductions in hunger and
Increases Yields in Developing poverty in developing countries.
However, many novel initiatives have emerged that are
Countries demonstrating that agriculture in poor countries can be
greatly improved. Here we evaluate how farmers in 286
projects in 57 countries have improved food crop productivity
J . N . P R E T T Y , * ,† A . D . N O B L E , ‡
since the early to mid 1990s, and at the same time increased
D. BOSSIO,§ J. DIXON,| R. E. HINE,†
F. W. T. PENNING DE VRIES,⊥ AND
both water use efficiency and carbon sequestration, and
J. I. L. MORISON† reduced pesticide use. These initiatives also offer the
prospects of resource conserving agriculture both reducing
Department of Biological Sciences and Centre for
adverse effects on the environment and contributing to
Environment and Society, University of Essex, Wivenhoe Park,
Colchester CO4 3SQ, U.K., International Water Management
important environmental goods and services (e.g., climate
Institute (IWMI), P.O. Box 1025, Kasetsart University, change mitigation).
Bangkok 10903, Thailand, International Water Management In the past 40 years, per capita world food production has
Institute (IWMI), P.O. Box 2075, Colombo, Sri Lanka, Impact grown by 17%, with average per capita food consumption in
Targeting and Assessment Program, CIMMYT, Apdo. Postal 2003 of 2780 kcal day-1 (5), where a majority of the chronically
6-641, 06600 Mexico, Mexico, and International Project Office hungry are small farmers who produce much of what they
for Monsoon Asia Integrated Regional Study, Institute for eat. Yet consumption in 33 poor countries is still less than
Atmospheric Sciences, Chinese Academy of Sciences, 2200 kcal day-1. Food demand will both grow and shift in the
P.O. Box 9804, Beijing, China coming decades, as (i) population growth increases absolute
demand for food; (ii) economic growth increases people’s
purchasing power; (iii) growing urbanization encourages
people to adopt new diets; and (iv) climate change threatens
Despite great recent progress, hunger and poverty remain both land and water resources.
widespread and agriculturally driven environmental Increased food supply is a necessary though not sufficient
damage is widely prevalent. The idea of agricultural condition for eliminating hunger and poverty. What is
sustainability centers on the need to develop technologies important is who produces the food, has access to the
and practices that do not have adverse effects on technology and knowledge to produce it, and has the
environmental goods and services, and that lead to purchasing power to acquire it. The great success of
improvements in food productivity. Here we show the industrialized agriculture in recent decades has masked
significant negative externalities, with environmental and
extent to which 286 recent interventions in 57 poor countries
health problems increasingly well-documented and costed,
covering 37 M ha (3% of the cultivated area in developing including in Ecuador, China, Germany, the Philippines, U.K.
countries) have increased productivity on 12.6 M farms and United States (6-11). There are also growing concerns
while improving the supply of critical environmental services. that such systems may not reduce food poverty. Poor farmers
The average crop yield increase was 79% (geometric need low-cost and readily available technologies and prac-
mean 64%). All crops showed water use efficiency gains, tices to increase local food production and to raise their
with the highest improvement in rainfed crops. Potential income. At the same time, land and water degradation is
carbon sequestered amounted to an average of 0.35 t C ha-1 increasingly posing a threat to food security and the
y-1. If a quarter of the total area under these farming livelihoods of rural people who often live on degradation-
systems adopted sustainability enhancing practices, we prone lands (12).
The idea of agricultural sustainability centers on food
estimate global sequestration could be 0.1 Gt C y-1. Of projects
production that makes the best use of nature’s goods and
with pesticide data, 77% resulted in a decline in pesticide services while not damaging these assets. Many different
use by 71% while yields grew by 42%. Although it is terms have come to be used to imply greater sustainability
uncertain whether these approaches can meet future in some agricultural systems over prevailing ones (both pre-
food needs, there are grounds for cautious optimism, industrial and industrialized) (13). Agricultural sustainability
particularly as poor farm households benefit more from in all cases, however, emphasizes the potential benefits that
their adoption. arise from making the best use of both good genotypes of
crops and animals and their ecological management. Agri-
cultural sustainability does not, therefore, mean ruling out
any technologies or practices on ideological grounds (e.g.,
Introduction genetically modified crop, organic practice)sprovided they
What is the best way to increase agricultural productivity in improve productivity for farmers, and do not harm the
developing countries that still, despite efforts over several environment (12-16).
decades, have some 800 million people short of food? The In this research, we concentrate on projects that have
question is controversial, with widely varying positions about made use of a variety of packages of resource-conserving
the types of inputs and technologies likely to be effective technologies and practices. These include the following: (1)
Integrated pest management, which uses ecosystem resilience
* Corresponding author e-mail: [email protected]; tel: +44- and diversity for pest, disease, and weed control, and seeks
1206-873323; fax: +44-1206-873416. only to use pesticides when other options are ineffective. (2)
† University of Essex.
‡ IWMI, Kasetsart University. Integrated nutrient management, which seeks both to balance
§ IWMI, Colombo, Sri Lanka. the need to fix nitrogen within farm systems with the need
| Impact Targeting and Assessment Program, CIMMYT. to import inorganic and organic sources of nutrients, and to
⊥ Chinese Academy of Sciences. reduce nutrient losses through erosion control. (3) Conser-
10.1021/es051670d CCC: $33.50  xxxx American Chemical Society VOL. xx, NO. xx, xxxx / ENVIRON. SCI. & TECHNOL. 9 A
Published on Web 00/00/0000 PAGE EST: 5.7
TABLE 1. Summary of FAO-World Bank Farming System Categories in Developing Regions and Number of Project Entries for This
Studya
agricultural no. project
number of land area cultivated area agricultural population per entries for
FAO farm system category subsystems (M ha) (M ha) population (M) cultivated hectare each category
1. smallholder irrigated 1 219 15 30 2.0 16
2. wetland rice 3 330 155 860 5.5 55
3. smallholder rainfed humid 11 2013 160 400 2.5 95
4. smallholder rainfed highland 10 842 150 520 3.5 40
5. smallholder rainfed dry/cold 19 3478 231 490 2.1 43
6. dualistic mixed 16 3116 414 190 0.5 20
7. coastal artisanal 4 70 11 60 5.5 2
8. urban-based and kitchen garden 6 na na 40 na 15
total 72 10068 1136 2590 2.28 286
a From Dixon and Gulliver (19); na ) not available.

TABLE 2. Summary of Adoption and Impact of Agricultural Sustainability Technologies and Practices on 286 Projects in 57
Countriesa
number of number of hectares under average % increase
FAO farm system category farmers adopting sustainable agriculture in crop yields
1. smallholder irrigated 177,287 357,940 129.8 ((21.5)
2. wetland rice 8,711,236 7,007,564 22.3 ((2.8)
3. smallholder rainfed humid 1,704,958 1,081,071 102.2 ((9.0)
4. smallholder rainfed highland 401,699 725,535 107.3 ((14.7)
5. smallholder rainfed dry/cold 604,804 737,896 99.2 ((12.5)
6. dualistic mixed 537,311 26,846,750 76.5 ((12.6)
7. coastal artisanal 220,000 160,000 62.0 ((20.0)
8. urban-based and kitchen garden 207,479 36,147 146.0 ((32.9)
all projects 12,564,774 36,952,903 79.2 ((4.5)
a Yield data from 360 crop project combinations; reported as % increase (thus a 100% increase is a doubling of yields). Standard errors are

given in brackets.

vation tillage, which reduces the amount of tillage, sometimes and rejections were based on a strict set of criteria (18). As
to zero, so that soil can be conserved and available moisture this was a purposive sample of “best practice” initiatives, the
used more efficiently. (4) Agroforestry, which incorporates findings are not representative of all farms in developing
multifunctional trees into agricultural systems, and collective countries.
management of nearby forest resources. (5) Aquaculture, We used a novel typology of farming systems developed
which incorporates fish, shrimps, and other aquatic resources by FAO for the World Bank to classify these projects (19) into
into farm systems, such as into irrigated rice fields and fish 8 broad categories based on the following social, economic,
ponds, and so leads to increases in protein production. (6) and biophysical criteria: (i) the available natural resource
Water harvesting in dryland areas, which can mean formerly base, including water, land, grazing areas, and forest; climate
abandoned and degraded lands can be cultivated, and and altitude; landscape, including slope; farm size, tenure,
additional crops can be grown on small patches of irrigated and organizations; and access to services including markets;
land owing to better rainwater retention. (7) Livestock and (ii) the dominant patterns of farm activities and
integration into farming systems, such as dairy cattle and household livelihoods, including field crops, livestock, trees,
poultry, including using zero-grazing. aquaculture, hunting and gathering, processing, and off-farm
Here we show the extent to which recent successful activities; and the main technologies used, which determine
interventions focusing on agricultural sustainability (some- the intensity of production and integration of crops, livestock
times called bright spots (17)) have increased total food crop and other activities.
productivity in developing regions. Our questions are as Table 1 contains a summary of the global land area and
follows: (i) To what extent can farmers increase per hectare population located in these eight major farm system cat-
and per farm food production by using low-cost and locally egories. On average, these sustain 2.28 people per cultivated
available technologies and inputs? (ii) What impacts do such hectare of land (range 0.5-5.5). A total of 72 farming
methods have on environmental goods and services (in subsystems have been identified across the developing
particular using the water use efficiency, carbon sequestra- regions, some of which comprised similar systems occurring
tion, and pesticide use as proxies to indicate changes in on different continents (e.g., wetland rice systems in East
adverse effects on the environment)? Asia/Pacific and in South Asia). A summary of all these
systems and their locations is contained in the Supporting
Methodology Information. In our study, system categories 2-5 are well-
We used both questionnaires and published reports by represented, with 40-95 projects in each. System categories
projects to assess adoption of sustainable agriculture and 1, 6, and 8 have 15-20 projects each, and category 7 has only
changes over time. As in earlier research (18), data were two.
triangulated from several sources, and cross-checked by Extent of Agricultural Sustainability and Impacts on
external reviewers and regional experts. This study involves Yields. Table 2 contains a summary of the location and extent
analysis of projects sampled once in time (n ) 218) and those of the 286 agricultural sustainability projects across the eight
sampled twice over a 4 year period to assess temporal changes categories of farming systems in 57 countries. In all, some
(n ) 68). Not all proposed cases were accepted for the dataset, 12.6 million (M) farmers on 37 M ha were engaged in

B 9 ENVIRON. SCI. & TECHNOL. / VOL. xx, NO. xx, xxxx

FIGURE 1. Histogram of change in crop yield after or with project,
compared to before or without project (n ) 360, mean ) 1.79, SD
0.91, median ) 1.50, geometric mean ) 1.64).
FIGURE 2. Box and whisker plot of change in crop yield after or
transitions toward agricultural sustainability in these 286
with project, compared to before or without project. Bold lines
projects. This is just over 3% of the total cultivated area shown within boxes indicate median value, box limits indicate interquartile
in Table 1. The largest number of farmers was in wetland range (i.e., 50% of values lie within the box), whiskers indicate
rice-based systems, mainly in Asia (category 2), and the largest highest and lowest, excluding outliers (O, 1.5-3 × box length
area was in dualistic mixed systems, mainly in southern Latin distance away from edge of box) or extremes (*, >3 × box length).
America (category 6). “Other” group consists of sugar cane (n ) 2), quinoa (1), oats (2).
We were able to show that agricultural sustainability is
spreading to more farmers and hectares. In the 68 randomly
re-sampled projects from the original study, there was a 56%
increase over the 4 years in the number of farmers (from 5.3
to 8.3 M), and 45% in the number of hectares (from 12.6 to
18.3 M). These resurveyed projects comprised 60% of the
farmers and 44% of the hectares in the original sample of 208
projects (18). In the earlier study, we reported that 89 projects
for which there was reliable yield data showed increases in
per hectare food production.
For the 360 reliable yield comparisons from 198 projects
that we now have, the mean relative increase was 79% across
the very wide variety of systems and crop types (see Table
B in the Supporting Information for full details of changes
in each farming system category). However, there was a wide
spread in results (Figure 1). While 25% of projects reported
relative yields >2.0, (i.e., 100% increase), half of all the projects
had yield increases of between 18% and 100%. The geometric
mean is a better indicator of the average for such data with
a positive skew, but this still shows a 64% increase in yield. FIGURE 3. Mean changes in crop yield after or with project,
However, the average hides large and statistically significant compared with before or without project. Vertical lines indicate (
differences among the main crops (Figures 2 and 3). In nearly SEM. “Other” group consists of sugar cane (n ) 2), quinoa (1), oats
all cases there was an increase in yield with the project. Only (2).
in rice were there 3 reports where yields decreased, and the
increase in rice was the lowest (mean ) 1.35), although it Though many technologies and practices were used in
constituted a third of all the crop data. Cotton showed a these projects, three types of technical improvement are likely
similarly small mean yield increase. to have played substantial roles in food production in-
The mean (2.84) and spread was largest in cassava and creases: (i) more efficient water use in both dryland and
sweet potato crops, although the sample is small. Soybean irrigated farming; (ii) improvements in organic matter
and groundnut showed mean increases of about 50%. Maize, accumulation in soils and carbon sequestration; and (iii) pest,
millet and sorghum, potatoes, and the other legumes group weed, and disease control emphasizing in-field biodiversity
(beans, pigeon peas, cowpea, chickpea) all showed mean and reduced pesticide (insecticide, herbicide, and fungicide)
yield increases of >100%, significantly higher than those for use.
cotton, rice, and groundnut (P < 0.05). For most of the main Impacts on Farm Water Use Efficiency. Widespread
field crops that are well represented in the survey, those with appreciation of the “global water crisis” recognizes that
low yields before intervention often showed larger relative scarcity of clean water is affecting food production and
improvements, either because of growth limiting environ- conservation of ecosystems. By 2025 it is predicted that most
ments, or perhaps reduced investment in developing these developing countries will face either physical or economic
crops, although potato showed large increases across the water scarcity (20). Water diverted from rivers increased 6-fold
range (Figure 4). between 1900 and 1995 (21), far outpacing population growth.

VOL. xx, NO. xx, xxxx / ENVIRON. SCI. & TECHNOL. 9 C

 nologies in a variety of crops and farm systems. Our results,
and others (24-25), demonstrate that the greatest op-
portunity for improvement in water productivity is in rainfed
agriculture. Better farm management, including supple-
mental irrigation and fertility management can significantly
reduce uncertainty, and thus avoid chronic low productivity
and crop failure that are characteristic of many rainfed
systems.
Impacts on Carbon Sequestration. The 1997 Kyoto
Protocol to the UN Framework Convention on Climate
Change established an international policy context for the
reduction of carbon emissions and increases in carbon sinks
to address the global challenge of anthropogenic interference
with the climate system. It is clear that both emission
reductions and sink growth will be necessary for mitigation
of current climate change trends (26-28). Carbon sequestra-
tion is defined as the capture and secure storage of carbon
that would otherwise be emitted to or remain in the
FIGURE 4. Relationship between relative changes in crop yield
atmosphere (29).
after (or with project) to yield before (or without project). Only field
crops with n > 9 shown. One of the actions farmers can take is to increase carbon
sinks in soil organic matter and above-ground biomass. We
Increasing demand for freshwater now threatens the integrity calculated the potential annual contributions being made in
of many aquatic ecosystems, and their associated environ- these 286 projects to carbon sink increases in soils and trees,
mental services (22). As agriculture accounts for 70% of using established carbon audit methods (30) (Table 4). As
current water withdrawals from rivers, improving the pro- the focus is on what sustainable methods can do to increase
ductivity of water use in agriculture is a growing challenge. quantities of soil and above-ground carbon, we did not take
The potential for increasing food production while account of existing stocks of carbon. Soil carbon sequestration
maintaining water-related ecosystem services rests on ca- is corrected for climate, as rates are higher in humid
pacity to increase water productivity (WP), i.e., by realizing compared with dry zones, and generally higher in temperate
more kg of food per unit of water. Sustainable agricultural than tropical areas (28-29).
practices may do this by (i) removing limitations on These projects were potentially sequestering 11.4 Mt C
productivity by enhancing soil fertility; (ii) reducing soil y-1 on 37 M ha. If scaled up, assuming that 25% of the areas
evaporation through conservation tillage; (iii) using more under the different farming system categories globally (Table
water-efficient varieties; (iv) reducing water losses to un- 1) adopted these same sustainability initiatives, this would
recoverable sinks; (v) boosting productivity by supplemental result in sequestration of 100 ((4) Mt C y-1 The average gain
irrigation in rainfed systems; and (vi) inducing microclimatic was 0.35 t C ha-1 y-1, and an average per household gain of
changes to reduce crop water requirements (23). We cal- 0.91 t C y-1. The per hectare gains vary from 0.15 t C ha-1
culated changes in WP for field crops in 144 projects from y-1 for smallholder irrigated systems (category 1) to 0.46 t C
the data set (Table 3) based on reported crop yields and ha-1 y-1 for category 3 systems. For most systems, per
average potential evapotranspiration (ETp), for each project households gains were in the range 0.05-0.5 t C y-1, with the
location during the relevant growing season. Actual evapo- much larger farms of southern Latin America using zero-
transpiration (ETa) was assumed to equal 80% of ETp, and tillage achieving the most at 14.9 t C y-1. Such gains in carbon
ETa to remain a constant at different levels of productivity. may offer new opportunities to households for income
WP gains were high in rainfed systems, and moderate in generation under emerging carbon trading schemes.
irrigated systems, and were in agreement with other studies Impacts on Pesticide Use. Integrated pest management
reporting ranges of WP (23). The very large increase for the (IPM) programs are beginning to show how pesticide use
vegetables and fruits is probably an overestimate as we did can be reduced and modified without yield penalties in a
not adjust ETp for new crops or lengthened cropping periods. variety of farm systems, such as in irrigated rice in Asia (31)
Variability was high due to the wide variety of practices and rainfed maize in Africa (32). In principle, there are four
represented in the dataset, but do indicate that gains in WP possible trajectories an agricultural system can take if IPM
are possible through adoption of sustainable farming tech- is introduced: (i) both pesticide use and yields increase (A);

TABLE 3. Summary of Changes in Water Productivity by Major Crop Type Arising from Adoption of Sustainable Agricultural
Technologies and Practices in 144 Projectsa
water productivity water productivity water productivity
before intervention after intervention gain
crop (kg food m-3 water ETa) (kg food m-3 water ETa) (kg food m-3 water ETa) % increase in WP
irrigated
rice (n )18) 1.03 ((0.22) 1.19 ((0.12) 0.16 ((0.04) 15.5%
cotton (n ) 8) 0.17 ((0.04) 0.22 ((0.05) 0.05 ((0.02) 29.4%
rainfed
cereals (n ) 80) 0.47 ((0.06) 0.80 ((0.09) 0.33 ((0.05) 70.2%
legumes (n)19) 0.43 ((0.07) 0.87 ((0.16) 0.44 ((0.11) 102.3%
roots and tubers (n)14) 2.79 ((0.73) 5.79 ((1.08) 3.00 ((0.65) 107.5%
urban and kitchen gardens
vegetables and fruits (n)5)
0.83 ((0.29) 2.96 ((0.97) 2.13 ((0.71) 256.6%
a Standard errors in brackets.

D 9 ENVIRON. SCI. & TECHNOL. / VOL. xx, NO. xx, xxxx

TABLE 4. Summary of Potential Carbon Sequestered in Soils and Above-Ground Biomass in the 286 Projectsa
carbon sequestered total carbon sequestered
per hectare carbon sequestered per household
FAO farm system category (t C ha-1 y-1) (Mt C y-1) (t C y-1)
1. smallholder irrigated 0.15 ((0.012) 0.011 0.06
2. wetland rice 0.34 ((0.035) 2.53 0.29
3. smallholder rainfed humid 0.46 ((0.034) 0.34 0.20
4. smallholder rainfed highland 0.36 ((0.022) 0.23 0.56
5. smallholder rainfed dry/cold 0.26 ((0.035) 0.20 0.32
6. dualistic mixed 0.32 ((0.023) 8.03 14.95
7. coastal artisanal 0.20 ((0.001) 0.032 0.15
8. urban-based and kitchen garden 0.24 ((0.061) 0.015 0.07
total 0.35 ((0.016) 11.38 0.91
a Standard errors in brackets.

Discussion
It is uncertain whether progress toward agricultural sus-
tainability, delivering benefits at the scale occurring in these
projects, will result in enough food to meet the future food
needs in developing countries after continued population
growth, urbanization, and the dietary transition to meat-
rich diets (39). Even the substantial increases reported here
may not be enough. However, more widespread adoption of
these resource conserving technologies, combined with other
innovations in crop and livestock genotypes, would con-
tribute to increased agricultural productivity (1, 16), par-
ticularly as evidence indicates that productivity can grow in
many farming systems as natural, social, and human capital
FIGURE 5. Changes in pesticide use and yields in 62 projects (A, assets also grow (40). Our findings also show that poor
n ) 10; C, n ) 5; D, n ) 47). households benefit substantially.
But improving agricultural sustainability alone will not
(ii) pesticide use increases but yields decline (B); (iii) both solve all food poverty problems. The challenge is to find ways
pesticide use and yields fall (C); or (iv) pesticide use declines, to improve all farmers’ access to productive technologies
but yields increase (D). and practices that are also resource conserving. The critical
The conventional wisdom is that pesticide use and yields priority is now international, national, and local policy and
are positively correlated, and so only trajectories moving institutional reforms (41) designed to benefit both food
into A and C are likely (33-34). A change into sector B would security and income growth at national and households levels,
be against economic rationale, as farmers’ profits would while improving the supply of critical technologies that
invariably fall and behavior change. A shift into sector D improve the supply of environmental goods and services.
would indicate that current pesticide use has negative yield
effects. This could be possible with excessive use of herbicides
Acknowledgments
or when pesticides cause outbreaks of secondary pests (35). We are grateful to all project staff and scientists who made
We analyzed the 62 IPM initiatives in 21 developing countries data available on projects, to earlier comments and sug-
in the dataset (Figure 5). The evidence on pesticide use is gestions from researchers involved in the IWMI Bright Spots
derived from data on both the number of sprays per hectare research program, to Noel Aloysius for input for some of the
and the amount of active ingredient per hectare. There is research, to David Tilman for comments on an earlier
only one case in sector B reported in recent literature (36), manuscript, and to two referees for their helpful comments.
and so this was not included. The research was funded by the U.K. Department for
Sector A contains 10 projects where pesticide use in- International Development. The views expressed in this paper
creased. These are mainly in zero-tillage and conservation are those of the authors and do not necessarily reflect the
agriculture systems, where reduced tillage creates benefits policies of their organizations.
for soil health and reduces off-site pollution and flooding
costs. These systems usually require increased use of Supporting Information Available
herbicides for weed control (37), though there are examples Table A1 containing full details of the classification system
of organic zero-tillage systems (38). The 5 cases in sector C developed by FAO (Dixon and Gulliver, ref 19) for farming
show a 4.2% ((5.0) decline in yields with a 93.3% ((6.7) fall systems. This separates farming systems into 8 types (ir-
in pesticide use. Most cases, however, are in category D where rigated; wetland rice based; smallholder rainfed humid;
pesticide use declined by 70.8% ((3.9) and yields increased smallholder rainfed highland; smallholder rainfed dry/cold;
by 41.6% ((10.5). While pesticide reduction is to be expected, dualistic; coastal artisanal fishing; urban-based) for six regions
as farmers substitute pesticides by information, the cause of of the world (Sub-Saharan Africa; Middle East and North
yield increases induced by IPM are complex. It is likely that Africa; Europe and Central Asia; South Asia; East Asia and
farmers who receive good quality field training will not only Pacific; Latin America and Caribbean). Table A2 summarizing
improve their pest management skills but also become more the location of the 286 projects in this study in these farming
efficient in other agronomic and ecological management systems types, and giving the impact of agricultural sus-
practices. They are also likely to invest cash saved from tainability in each farming system. Part C containing profiles
pesticides in other inputs such as higher quality seeds and of 47 of the 286 projects (11 in Latin America, 17 in Africa,
fertilizers. This analysis indicates considerable potential for and 19 in Asia) as examples of how the technologies were
avoiding environmental costs. adopted and their environmental and social outcomes. This

VOL. xx, NO. xx, xxxx / ENVIRON. SCI. & TECHNOL. 9 E

material is available free of charge via the Internet at http:// (24) Agarwal, A.; Narain, S. Dying Wisdom; Thomson Press: Farida-
pubs.acs.org. bad, India, 1997.
(25) Rockström, J.; Falkenmark, M. Semiarid crop production from
a hydrological perspective - gap between potential and actual
Literature Cited yields. Crit. Rev. Plant Sci. 2000, 19 (4), 319-346.
(1) Trewevas, A. Malthus foiled again and again. Nature 2002, 418, (26) IPCC. Climate Change 2001: Impacts, Adaptation and Vulner-
668-670. ability; Intergovernmental Panel on Climate Change: Geneva,
(2) Smil, V. Feeding the World; MIT Press: Cambridge MA, 2000. 2001.
(3) Tilman, D.; Cassman, K. G.; Matson, P. A.; Naylor, R.; Polasky, (27) Swingland, I., Ed. Carbon and Biodiversity; Earthscan: London,
S. Agricultural sustainability and intensive production practices. 2003.
Nature 2002, 418, 671-677. (28) Lal, R.; Griffin, M.; Apt, J.; Lave, L.; Morgan, M. G. Managing soil
(4) McNeely, J. A.; Scherr, S. J. Ecoagriculture; Island Press: carbon. Science 2004, 304, 393.
Washington, DC, 2003.
(29) Watson, R. T.; Noble, I. R.; Bolin, B.; Ravindranath, N. H.; Verardo,
(5) FAO. FAOSTAT Database; Rome, 2005.
D. J.; Dokken, D. J., Eds. IPCC Special Report on Land Use, Land-
(6) Crissman, C. C., Antle, J. M., Capalbo, S. M., Eds. Economic,
Use Change and Forestry; IPCC Secretariat: Geneva, 2000.
Environmental and Health Tradeoffs in Agriculture; CIP, Lima
(30) Pretty, J.; Ball, A. S.; Xiaoyun, L.; Ravindranath, N. H. The role
& Kluwer: Boston, MA, 1998.
of sustainable agriculture and renewable resource management
(7) Norse, D.; Ji, L.; Leshan, J.; Zheng, Z. Environmental Costs of
in reducing greenhouse gas emissions and increasing sinks in
Rice Production in China; Aileen Press: Bethesda, MD, 2001.
China and India. Philos. Trans. R. Soc., Ser. A 2002, 360, 1741-
(8) Waibel, H.; Fleischer, G.; Becker, H. The economic benefits of
1761.
pesticides: A case study from Germany. Agrarwirtschaft 1999,
48 (6), 219-230. (31) Eveleens, K. The History of IPM in Asia; FAO: Rome, 2004.
(9) Pingali, P. L.; Roger P. A. Impact of Pesticides on Farmers’ Health (32) Khan, Z. R.; Ampong-Nyarkko, K.; Chiliswa, P.; Hassanali, A.;
and the Rice Environment; Kluwer: Dordrecht, The Netherlands, Kimani, S.; Lwande, W.; Overholt, W. A.; Pickett, J. A.; Smart, L.
1995. E.; Wadhams, L. J.; Woodcock, M. Nature 1997, 388, 631-632.
(10) Pretty, J.; Brett, C.; Gee, D.; Hine, R.; Mason, C. F.; Morison, J. (33) Knutson, R. D.; Taylor, C. R.; Penson, J. B.; Smith, E. S. Economic
I. L.; Raven, H.; Rayment, M.; van der Bijl, G. An assessment of Impacts of Reduced Chemical Use; Knutson & Assoc.: College
the total external costs of UK agriculture. Agric. Syst. 2000, 65 Station, TX, 1990.
(2), 113-136. (34) Schmitz, P. M. Overview of cost-benefit assessment. In OECD
(11) Tegtmeier, E. M.; Duffy, M. D. External costs of agricultural Workshop on the Economics of Pesticide Risk Reduction in
production in the US. Int. J. Agric. Sust. 2004, 2, 1-20. Agriculture; OECD: Paris, 2001.
(12) Uphoff N. Agroecological Innovations; Earthscan: London, 2002. (35) Kenmore, P. E.; Carino, F. O.; Perez, C. A.; Dyck, V. A.; Gutierrez,
(13) National Research Council. Our Common Journey; National A. P. J. Plant Prot. Tropics 1984, 1 (1), 19-37.
Academy Press: Washington, DC, 2000. (36) Feder, G.; Murgai, R.; Quizon, J. B. Sending farmers back to
(14) Conway, G. R. The Doubly Green Revolution; Penguin: London, school: the impact of Farmer Field Schools in Indonesia. Rev.
1997. Agric. Econ. 2004, 26 (1), 45-62.
(15) Pretty, J. Agri-Culture: Reconnecting People, Land and Nature; (37) de Freitas, H. Transforming microcatchments in Santa Caterina,
Earthscan: London, 2002. Brazil. In Fertile Ground; Hinchcliffe, F., Thompson, J., Pretty,
(16) Nuffield Council on Bioethics. The Use of Genetically Modified J., Guijt, I., Shah, P., Eds.; IT Publications: London, 1999.
Crops in Developing Countries; London, 2004. (38) Petersen, P.; Tardin, J. M.; Marochi, F. Participatory development
(17) Scherr, S. J.; Yadav, S. Land Degradation in the Developing World; of no-tillage systems without herbicides for family farming.
IFPRI: Washington, DC, 1996. Environ. Dev. Sustainability 2000, 1, 235-252.
(18) Pretty, J.; Morison, J. I. L.; Hine, R. E. Reducing food poverty by (39) Delgado, C.; Rosegrant, M.; Steinfield, H.; Ehui, S.; Courbois, C.
increasing agricultural sustainability in developing countries. Livestock to 2020: The Next Food Revolution; IFPRI: Washington,
Agric. Ecosyst. Environ. 2003, 95 (1), 217-234. DC, 1999.
(19) Dixon, J.; Gulliver, A.; with Gibbon, D. Farming Systems and (40) Pretty, J. Social capital and the collective management of
Poverty; FAO: Rome, 2001. resources. Science 2003, 302, 1912-1915.
(20) International Water Management Institute. World Water Sce-
(41) Dasgupta, P. The economics of food. In Feeding the World
narios Analyses; Rijsberman, F., Ed.; Earthscan: London, 2000.
Population of More Than Eight Billion People; Waterlow, J. C.,
(21) Shiklomanov, L. A. World Water Resources: An Appraisal for the
Armstrong, D. G., Fowden, L., Riley, R., Eds.; Oxford University
21st Century; UNESCO: Paris, 1999.
Press: New York, 1998.
(22) Costanza, R.; d’Arge, R.; de Groot, R.; Farber, S.; Grasso, M.;
Hannon, B.; Limburg, K.; Naeem, S.; O’Neil, R. V.; Parvelo, J.;
Raskin, R. G.; Sutton, P.; van den Belt, M. The value of the world’s
ecosystem services and natural capital. Nature 1997, 387, 253- Received for review August 23, 2005. Revised manuscript
260. received November 17, 2005. Accepted November 18, 2005.
(23) Kijne, J. W., Barker, R., Molden, D., Eds. Water Productivity in
Agriculture: Limits and Opportunities for Improvement; CABI
Publishing: Wallingford, U.K., 2003. ES051670D

F 9 ENVIRON. SCI. & TECHNOL. / VOL. xx, NO. xx, xxxx PAGE EST: 5.7