Ecological Indicators 60 (2016) 192–201

Contents lists available at ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

On the suitability of input–output analysis for calculating

product-specific biodiversity footprints
Daniel Moran ∗ , Milda Petersone, Francesca Verones
Programme for Industrial Ecology, NTNU, Trondheim, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Recently it has been estimated that one third of biodiversity threats are driven by consumer demand from
Received 26 November 2014 outside the country in which the threat occurs. This occurs when the production of export goods exerts
Received in revised form 2 June 2015 pressure on vulnerable populations. While population biologists have in cases been able to establish links
Accepted 4 June 2015
between species threats and the causative industry(s), little has been done to trace this biodiversity foot-
Available online 17 July 2015
print from the directly implicated industry out to final consumers, a step that would open a wider variety
of policy responses. Here we investigate the suitability of multi-region input–output (MRIO) analysis for
Keywords:
tracing out links between particular species threats, directly implicated industries, and the countries and
Trade
Biodiversity consumer goods sectors ultimately driving these industries. Environmentally extended MRIO models are
MRIO understood to provide reliable results at a macroeconomic level but uncertainty increases as the models
Hybrid LCA are used to investigate individual sectors, companies, and products. In this study we examine several
Biodiversity footprint case studies (nickel mining in New Caledonia, coltan from the Democratic Republic of Congo, cut flowers
from Kenya, and forestry in Papua New Guinea) in order to understand how and when MRIO techniques
can be useful for studying biodiversity implicated supply chains. The study was conducted using the Eora
global input–output database that documents >5 billion global supply chains. Calculating the biodiver-
sity footprint at this level of detail, between specific threats, supply chains, and consumer goods, has not
been done before. These case studies provide interesting insights in their own right and also serve to
highlight the strengths and weaknesses of using input–output analysis techniques to calculate detailed
biodiversity footprints. We conclude that MRIO analysis, while no panacea, can be useful for outlining
supply chains and identifying which consumption sectors and trade and transformation steps can be
subjected to closer analysis in order to enable remedial action.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction measuring ecosystems’ health is difficult. There is no clear consen-

sus on any single best way to measure biodiversity health. Next, it is
The planet is currently undergoing its sixth great extinction often difficult to attribute species threats to specific human activ-
event (Chapin et al., 2000; Butchart et al., 2010; Dirzo et al., 2014; ities – a necessary prerequisite for organizing any socio-political
Tittensor et al., 2014). Species loss is occurring at a rate of two or response. Then, even in cases where a policy has been established,
more orders of magnitude greater than before the Anthropocene economic and social interests often collide with protection goals
(Pimm et al., 1995), and humans and their domesticated animals (Chapman et al., 2003; Luck et al., 2004) and illegal and unre-
currently account for >97% of terrestrial vertebrate biomass (Smil, ported activity thus continue to drive further species loss. Adding
2002). Yet despite full awareness and considerable research into the further to this complexity is the fact that in today’s globalized
problem of biodiversity loss there are few clear stratagems for ame- economy purchasers in households, business, and governments are
liorating the situation. In terms of tractable policies, saving species often far removed from the ecological impacts their consumption
seems to be proving a far more difficult goal than reducing GHG ultimately drives. Often consumers cannot directly see how their
emissions. This is because protecting biodiversity is both an eco- actions impact individual species.
logically and an economically complex challenge. To begin with, It is this last point – supply chain opacity and complexity
– that we investigate here. We use environmentally extended
input–output analysis to evaluate the supply chains of biodiversity-
∗ Corresponding author. Tel.: +47 915 94 183. implicated commodities. We present four case studies (nickel
E-mail address: [email protected] (D. Moran). mining in New Caledonia, coltan from the Democratic Republic of

http://dx.doi.org/10.1016/j.ecolind.2015.06.015
1470-160X/© 2015 Elsevier Ltd. All rights reserved.
 D. Moran et al. / Ecological Indicators 60 (2016) 192–201 193

Congo, cut flowers from Kenya, and forestry in Papua New Guinea) and Lenzen, 2009). For this study we used the Eora multi-region
and use these techniques to try and identify clear links between IO table (Lenzen et al., 2012a) which covers 187 countries with
a particular biodiversity threat, a causative industry, and the key a detail of 26–500 economics sectors per country for a total of
supply chains leading out from that industry to final consumers S = 15,909 sectors/goods. Countries in Eora have variable levels of
of implicated products. Identifying such threat pathways allows detail because Eora is composited from national IO tables and the
all the actors along a supply chain – traders, companies, govern- original native classifications are preserved. Eora was chosen from
ments, and households – to contribute to reducing the magnitude amongst the several global MRIO tables currently available (Tukker
or ecological intensity of a product’s supply chain. and Dietzenbacher, 2013) because of its superior country coverage.
Using input–output (IO) techniques to trace environmentally This is an important attribute because our study requires an MRIO
important flows is not new. The basic techniques have been devel- table that covers even smaller economies in biodiversity hotspots.
oped and refined since their introduction in the 1940s. IO is The Leontief calculus can be used to connect final consumers
regularly used to calculate carbon footprints, trace substances of with upstream biodiversity impacts. Leontief originally created
concern, and unravel the linkages between consumers and the his methods in order to calculate how much of a given primary
raw resources their consumption requires (Graedel and Allenby, resource, e.g. coal, was needed, across the entire economy, to satisfy
1995; Graedel et al., 2002; Reck et al., 2008; Peters et al., 2012; $1 of consumer demand for a particular product. By conceptualiz-
Reck and Graedel, 2012; Graedel et al., 2013). What is new is using ing pollution – or, in our case, biodiversity impacts – as a necessary
IO techniques to trace biodiversity-implicated commodities. In a input to production, the same methods can be employed to deter-
seminal study on biodiversity footprints Lenzen and colleagues mine how much biodiversity impact was exerted to produce $1 of a
(Lenzen et al., 2012b) used IO accounting and found that one third of particular good in that year. Input–output tables are retrospective
the species threats recorded on the IUCN’s Red List of Threatened accounts; they record the production recipe and trade volumes for
and Endangered Species were ultimately driven by consumption prior years. Improvements in technology or changes in trade pat-
demand outside the country in which the threat was exerted. terns can change the environmental impact of a sector in future
Lenzen et al.’s study worked in the aggregate, looking at over years.
3000 individual species threats and 15,000 industries across 187 One challenge in studying biodiversity is choosing an indicator
countries. In this study we use the same methods but look in depth with which to measure biodiversity pressure or loss. For their study
at few selected species, industries, and trade flows. The aim is to on total biodiversity impact, Lenzen et al. (2012b) quantified the
determine whether IO methods are suitable for investigating indi- biodiversity impact of a sector by using the total number of species
vidual species threats and related implicated product flows. Policy endangered due to primary production of that sector. Since in this
responses to biodiversity threats, especially when linked to traded study our focus is on determining the suitability of MRIO meth-
products, will require a high level of industry and product detail, ods for tracing individual supply chains, we may skip the difficult
a fact recognized by the European Commission (Lammerant et al., question of choosing how to measure biodiversity pressure. The
2014) and in the active interest in the biodiversity footprint. IO important thing is to determine which sector(s) are responsible for
methods could also be used to bolster green supply chain and green causing pressure and we do not need to measure the intensity with
certification programmes. With the selected case studies we seek which that pressure is exerted. Thus to construct the environmental
to ask whether IO methods are appropriate, and robust enough, for satellite account for the environmentally extended MRIO analysis,
studying individual species – industry – supply chain links. The case all sectors are given 0 biodiversity impact, except the selected sec-
studies were selected to provide a “best case” of use of MRIO tech- tor(s) which are given a value of 1. All numeric results, then, are
niques to biodiversity-implicated products through supply chains. expressed as dimensionless percentages of the total impact.
The cases were chosen for having clear links between species loss The Leontief method has been well explained (for overviews see
and a particular industry, and good trade data on that industry’s Wiedmann, 2009; Kitzes, 2013; Schaffartzik et al., 2014) but we
onward flows. briefly reiterate it here. Using the Eora MRIO table Ts×s document-
This paper proceeds as follows. First, we introduce IO and related ing the monetary transactions between S sectors, the biodiversity
methods in Sections 2 and 2.1. Then, in Section 2.2 we introduce the footprint F1×s in terms of a particular implicated commodity,
case studies and establish the links between biodiversity pressure resulting directly and indirectly from spending ys×1 of final con-
and one or more specific industries in each case. Next, in Section 3 −1 −1 −1
sumers is F = Qx̂ (I − Tx̂ ) y, where xs×1 denotes sectoral gross
we present numeric results from the IO and trade analysis. Finally, output, the ôperator denotes diagonalization, Is×s is an identity
in Section 4 we offer discussion about the results, including on matrix, and Q1×s is an environmental satellite account containing
reliability. Finally, Section 5 concludes. the value of other resources used as input in that sector. For each
case study a Q vector was constructed containing a single nonzero
unit element flagging the environmental input to the particu-
2. Methods
lar sector under consideration. In this environmentally extended
input–output analysis the units used in the satellite account are
2.1. Input–output analysis and structural path analysis
arbitrary; the result footprint will be expressed in the same unit as
used in the satellite account. In these studies footprint was mea-
Input–output tables provide a database of global trade flows,
sured not according to number of species or area impacted, but
as well as production and consumption recipes. Using an IO table −1
it is possible to identify supply chains such as: “A typical $10,000 merely as a share of a unit impact. The term Qx̂ contains the direct
automobile purchased in the US requires $1200 worth of Japanese biodiversity impact of each sector’s production, in terms of impact
−1 −1
steel parts, the production of which in turn require $600 worth per $ gross output. The term (I − Tx̂ ) (where the sub-term Tx̂ is
of Chinese rolled steel, the production of which in turn requires often abbreviated A, or technical coefficients matrix) is the classic
$200 worth of Australian iron ore.” The techniques of input–output Leontief inverse. All analysis was conducted in terms of producer’s
analysis were originally developed by Leontief (1986), and the prices.
Structural Path Analysis (SPA) (Defourny and Thorbecke, 1984; Suh By solving the Leontief inverse as a Taylor series expansion foot-
and Heijungs, 2007) used to extract individual supply chains from prints can be unravelled into individual paths. We abbreviate the
−1
aggregate results, has been developed and applied extensively since terms to q = Qx̂ and define a technical coefficients matrix A,
then (e.g. in (Lenzen, 2003, 2006; Peters and Hertwich, 2006; Wood expressing production recipes in each column as composition of
194 D. Moran et al. / Ecological Indicators 60 (2016) 192–201

−1
inputs needed to produce $1 of output, as A = Tx̂ . We find that Myers et al., 2000) with extraordinary floristic diversity and high
F = q(I − A)−1 which can be expanded into an infinite series of sums: local endemism (Morat et al., 2012; Wulff et al., 2013). It has
been listed as one of the most important metallophyte hotspots
   (Whiting et al., 2004) where plants have evolved and adapted to
Fi = qim ym + qil Alm ym + qil Akl Alm ym the metal rich ultramafic soils covering one third of the archipelago
m lm klm (Proctor, 2003). This unique diversity is threatened by economic
 activities, such as open-pit nickel mining and processing indus-
+ qij Ajk Alm ym + · · ·.
tries, causing severe environmental degradation (Pascal et al.,
iklm 2008; Jaffré et al., 2010f; Wulff et al., 2013). Human-induced fires
where each term qij Ajk . . . Alm ym represents a supply chain starting constitute a secondary threat. Since fires occur for a variety of
with an expenditure ym on a commodity produced by industry m, reasons without clear evidence linking threats related to human-
proceeding with an input Alm by another industry l that is required induced fires to production of a specific commodity we restrict
by industry m to produce a commodity worth ym . The supply chain ourselves to study the link between mining activity and biodiversity
continues via additional upstream suppliers of inputs k, and ends threats.
with a producing industry j that causes threats to species i. In par- The unique flora of New Caledonia contains 3371 species of vas-
ticular, cular plants, with more than 74% being endemic to the archipelago
(Morat et al., 2012). Moreover, recent studies (Wulff et al., 2013;
 Ibanez et al., 2014) suggest that many endemic plant species have
- m qim ym describes threats occurring directly at the place of final
demand (for example because of pollutants emitted by a house- highly restricted distribution ranges, covering only one or few local-
hold), ities with small population numbers. This makes them inherently
 vulnerable to extinction. Most (60%) of the plant species are local-
- lm qil Alm ym describes threats facilitated indirectly through one-
stage supply chains (for example because of emissions from a ized on ultramafic substrates (Proctor, 2003) and 90% of those are
factory producing consumer goods), endemic to New Caledonia (Morat, 1993). Endemic species include
 43 conifer species (7% of the world’s total conifer diversity) (Farjon
- klm qik Akl Alm ym describes threats facilitated indirectly through
two-stage supply chains (for example because of overfishing of and Page, 2003), which makes the archipelago’s conifer fauna one
marine species destined for processing and canning, and then of the richest in the world. All but one of these conifer species
finally sold), are included in the IUCN Red List, with status varying from “least
 concern” (LC) to “critically endangered” (CR). All of them occur on
- jklm qik Ajk Akl Alm ym describes threats facilitated indirectly
through three-stage supply chains (for example because of habi- ultramafic soils and the majority (68%) is restricted to this sub-
tat destruction by wheat grown for a flour mill making flour for a strate (Jaffré et al., 2010a). Among the conifer species is the critically
bakery selling bread to consumers, endangered (CR) Araucaria nemorosa with a small and fragmented
- and so on. distribution in the southern part of the archipelago occurring on
only 9.8 km2 (Kettle et al., 2007; Jaffré et al., 2010a). This species
has been classified as CR due to recently revealed negative genetic
As this is an infinite series, there are theoretically an infi-
effects on juvenile population and overall demographics due to
nite number of supply chains. However, the longest and smallest
habitat degradation mainly from mining activities and human
of these approach infinitesimally small magnitudes. For the Eora
induced fires (Kettle et al., 2007). None of Araucaria nemorosa’s
MRIO database, approximately 5 billion of these chains have a sta-
range is protected. Furthermore, a new hydro-metallurgic plant is
tistically meaningful magnitude. Structural Path Analysis is then
being built in the area (Jaffré et al., 2010a). The start of the mining
simply an enumeration, ranking, and optional filtering of these
activities in New Caledonia dates back to the second half of the 19th
paths.
century with wide-scale mining establishment in the last decades
of the 20th century (Ali and Grewal, 2006). The archipelago har-
2.2. Case studies bours between 20% and 30% of the global nickel deposits (Jaffré
et al., 2010a) and is now the world’s second leading producer of
Here we present four case studies using IO methods to trace one ferronickel (an alloy composed of nickel and iron) and one of the
flagged production sector to final consumers worldwide. First we top suppliers of nickel ore (USGS, 2013, 2014d). The archipelago is
establish, for each case study, the link between biodiversity threat home to some of the biggest nickel mines in the world from which
and a specific causative industry(s) in that country. Then, in the the Goro nickel-cobalt mine is the largest. It is planned that the
Section 3, Results, with a clear link between a threat and a respon- mine, will reach an annual capacity of 60,000 t nickel and 4650 t of
sible industry established the IO and SPA methods are employed in cobalt using high-pressure acid leaching technology, which allows
order to trace the outputs from that industry through global pro- extracting metals from low-grade ores, but leaving huge amounts
duction chains to end consumers. We also identify for each study of tailings (USGS, 2014a). The capacity, however, has not yet been
the top countries consuming those biodiversity-implicated prod- reached due to several incidents of acid spills from the mines, caus-
ucts and present some of the paths through which those implicated ing environmental damage and local outrage (USGS, 2012; Lefort
goods flow. Finally, in Section 4, Discussion, we discuss the results, and Burton, 2014, USGS, 2014a).
including areas which may contribute uncertainty to the results. Nevertheless, mining expansion can be expected in New Cale-
The selected case studies were chosen because of the availability donia since nickel and cobalt production accounts for almost 90%
of both good trade data to follow the resource through the supply of the total exports (USGS, 2014a). This is largely possible due
chain, as well as clear links between biodiversity loss and a spe- to increasing mining of low-grade deposits (Jaffré et al., 2010a),
cific industry. We study coltan mining in the Democratic Republic attractive tax rates, and weakly enforced environmental regula-
of Congo, nickel mining in New Caledonia, cut flower production in tions (Ali and Grewal, 2006). Together this gives reasons for concern
Kenya, and forestry in Papua New Guinea (Fig. 1). over the future of island’s unique biodiversity, particularly species
restricted to ultramafic soil. Mining activities potentially threatens
2.2.1. Nickel mining in New Caledonia the habitat and distribution of the endemic floral species through
New Caledonia, an island archipelago in the south-west Pacific habitat destruction, fragmentation, deposition of tailings and
(Fig. 1), is a global biodiversity hotspot (Mittermeier et al., 1996; pollution.
 D. Moran et al. / Ecological Indicators 60 (2016) 192–201 195

Fig. 1. Geographical locations of the four case studies in grey shades: DR Congo, Kenya, Papua New Guinea and in the blow-up New Caledonia.

2.2.2. Tantalum (coltan) mining in the Democratic Republic of of Lake Naivasha and its biodiversity. These impacts include habi-
Congo tat destruction around the lake and lake-wide papyrus decline
The price of tantalum – a rare earth metal used in a variety of (Everard and Harper, 2002; MacLean et al., 2003; Morrison and
electronics – spiked in the early 2000s. The resulting gold rush for Harper, 2009), chemical water pollution and eutrophication due
the DRCs easily-mined surface deposits was alleged to have been to runoff from settlements, local farming and industries (Kitaka
a major funding source for warlords during the country’s civil war. et al., 2002; Ndungu et al., 2013), overfishing (Hickley et al., 2004),
The UN called coltan “the engine of conflict in the DRC” (UN Security alien species introduction (Gherardi et al., 2011) and excessive
Council, 2001 §IV.215). In addition to the human cost of this con- water abstraction (Becht and Harper, 2002; Awange et al., 2013).
flict mineral, illegal mining camps within several national parks The water level declines in the lake due to water abstraction has
decimated populations of gorillas and other endangered, edible ani- been proven to be directly linked with cut-flower exports and
mals. According to an investigation (Hayes and Burge, 2003) miners constitutes a major threat for the future existence of the lake and
in one national park effectively eliminated the gorilla and elephant surrounding areas (Mekonnen et al., 2012; Awange et al., 2013).
populations there. One investigator reported that miners moving The horticulture boom has provided social benefits. An esti-
into the park began hunting elephants, gorillas, chimpanzees, buf- mated 90,000 direct jobs and 500,000 indirect jobs (KFC, 2014)
falos and antelopes but after one year had depleted these stocks are linked to cut-flower production, in addition to free housing,
and were eating only tortoises, birds and small animals (Redmond, schools and hospitals provided to employees and their families. The
2001). industry provides high return in terms of revenue per unit water
use.
However, the industrial farming expansion has been detrimen-
2.2.3. Flower horticulture at lake Naivasha, Kenya
tal for biodiversity in the area. Otiang’a-Owiti and Oswe (2007) used
Lake Naivasha, situated 80 km north-west of Nairobi in the Rift
the Living Planet Index methodology (Loh et al., 2005) to construct
Valley of Kenya is the second largest freshwater lake in the country
a composite (8 species) index of water bird health around the lake,
after Lake Victoria. It has been declared a wetland of international
from 1982 to 2002. Indexed to 1982 = 100%, their LPI index of bird
importance under the Ramsar Convention in 1995 and an Impor-
health had declined to 40% by 2002. A study of population changes
tant Bird Area in 2001, supporting over 350 different water bird
of the African Fish Eagle at the lake (Harper et al., 2002) also showed
species and also harbouring a large hippo population (Hippopota-
a decline from 1987 to 1999. Papyrus-vegetated area, an important
mus amphibious; classified by the IUCN as Vulnerable) among others
water bird habitat, has substantially decreased between 1955 and
(Ramsar, 2012; BirdLife International, 2014). In the last decades the
2000 (Everard and Harper, 2002).
area around the lake has grown to become the main site of Kenya’s
horticulture industry, from which the principle product is cut flow-
ers – Kenya’s second most profitable export after tea. Flower farms 2.2.4. Forestry in Papua New Guinea
are an important local income source and represent a major share Papua New Guinea (PNG), occupying the eastern part of the
of Kenya’s export earnings in general, yet flower farms are heavily island New Guinea and surrounding islands, supports one of the
responsible for the deteriorating health of Lake Naivasha (Otiang’a- world’s largest and biologically and culturally richest rain forests
Owiti and Oswe, 2007; Mekonnen et al., 2012; Awange et al., 2013). (Brooks et al., 2006; Gorenflo et al., 2012). Covering 28.3 million
Although the lake has always been used by humans, socio- hectares of land area (Shearman et al., 2009) it hosts most of
economic activities around the lake reached a new level in the PNG’s unique biodiversity, which is around 7% of the world’s total
1990s when the cut-flower industry expanded into ≈100 km2 of species (PNG 4th National Report on the CBD, 2010; Shearman and
irrigated land around the lake within a decade (Harper and Mavuti, Bryan, 2011). Furthermore, it is one of the last “wild areas” left
2004). Since then, Kenya’s floriculture industry, mostly located (Mittermeier et al., 1998; Sanderson et al., 2002) with many species
around Lake Naivasha (95% in 2005) (Mekonnen et al., 2012), has yet to be discovered. Despite that, PNG forests have been exploited
grown every year – from 11,000 t of flowers exported in 1988 to more rapidly than previously assumed (Shearman et al., 2009) and
125,000 t in 2013 (KFC, 2014). it is very likely that most of the accessible rainforests will disappear
It is estimated that human population around Lake Naivasha or be seriously degraded within the next two decades if current
has expanded to 250,000, which is around ten times more than in trends continue (Filer et al., 2009; Shearman et al., 2009, 2012).
the beginning of the boom of flower industry (Enniskillen, 2002). The PNG forest ecosystem covers around 70% of the land, most of
With expansion of social and economic activities around the lake, which is rainforest (Shearman et al., 2009). The landscape is rugged,
human impacts have increased threatening the wetland ecosystem with extensive areas reaching over 3000 m above sea level (PNG
196 D. Moran et al. / Ecological Indicators 60 (2016) 192–201

4th National Report on the CBD, 2010) creating very diverse habi- Table 1
– Top final consumers, with top implicated consumption goods, of embodied New
tats, many of which are inaccessible. There are still many gaps in
Caledonian nickel. Goods and services are domestically produced, using embodied
the knowledge about flora and fauna of Papua New Guinea (IUCN, New Caledonian nickel, unless an origin country is specified. Data are for 2011.
2008), therefore exact numbers of species are not known. Today
Top final consumers Top implicated final consumption goods (Percent
2316 of 26,318 species are assessed by IUCN (20% of them endemic),
(Percent of total impact) of total consumption in the consumer country)
and many more are assumed to be undiscovered (e.g. 150,000 insect
species are estimated to be unidentified in PNG) (Sekhran et al., Japan (18%) Non-residential construction (non-wooden) (10%);
Passenger motor cars (7%); Residential
1995; IUCN, 2008; PNG 4th National Report on the CBD, 2010). More
construction (non-wooden) (7%); Public
than 1060 new species, including 12 mammal species, were discov- construction of roads (4%)
ered and described on the island of New Guinea between 1998 and China (15%) Construction (42%); Other special industrial
2008 (WWF, 2011b). equipment (7%); Motor vehicles (3%);
Communication equipment (2%)
There are many activities driving deforestation and forest
USA (12%) Light truck and utility vehicle manufacturing (8%);
degradation in PNG, including subsistence agriculture, plantation Japanese Passenger motor cars (8%); Automobile
development, and mining (Shearman et al., 2009; Nelson et al., manufacturing (6%); Other non-residential
2014). Logging, however, a large share of which is export-driven, structures (4%)
has caused the largest part of forest change (Shearman et al., 2009). New Caledonia (9%) Construction (30%); Metal Products (14%);
Electrical and Machinery (10%); Administration
Shearman et al. compared land-cover maps between 1972 and 2002
(8%)
and concluded that a net 15% of primary rain forest was cleared Australia (8%) Residential building construction (16%);
and an additional 8.8% degraded to secondary forest (resulting in Non-residential building construction (9%);
soil depletion and increased vulnerability to fire) due to logging in Finished cars (8%); New Caledonian Mining and
Quarrying (8%)
PNG (Shearman et al., 2009). Lowland forest areas (65% of all PNG
France (5%) Construction work (22%); Motor vehicles, trailers
forests) are of particular concern, as they are more easily accessi- and semi-trailers (14%); Fabricated metal products,
ble than mountain forests and an estimated 30% has been cleared except machinery and equipment (13%);
or degraded in the past 30 year period, mainly due to logging Machinery and equipment n.e.c. (5%)
(Shearman and Bryan, 2011). South Korea (3%) Japanese Steel ships (15%); Public administration
and defense (9%); Motor vehicles and parts (7%);
Electronic machinery, equipment, and supplies
3. Results (7%)
Spain (3%) Residential buildings (13%); Civil engineering (6%);
Motor vehicles (6%); Non-residential buildings (5%)
Having established for each case study a link between biodiver- Italy (2%) Construction work (9%); Furniture; other
sity pressure and one or more specific industries in each country, manufactured goods n.e.c. (9%); Machinery and
we now report on the result of the input–output analysis of each of equipment n.e.c. (7%); New Caledonian Metal
Products (6%)
those produced commodities.
Germany (2%) Passenger cars and parts (8%); Installations and
other construction (5%); French Motor vehicles,
trailers and semi-trailers (4%); Spanish Motor
3.1. Nickel mining in New Caledonia
vehicles (3%)

The threat to biodiversity on New Caledonia is related to the

mining and building of a new metallurgical plant. Nickel can be
found in various products ranging from pots, pans and jewellery jewellery and silverware manufacturing industry → Final demand
to marine and aerospace applications. Its biggest use is as an allo- in USA in the jewellery and silverware manufacturing sector” is
ying element for stainless and heat-resisting steels (Nickel Institute, worth $400,000 of nickel and includes New Caledonian nickel
2014). Most often nickel and nickel products from New Caledonia embodied in jewellery consumed in USA.
are sent to Japan, Australia and France where they are used in steel
alloys. SPA results, and a table of top consumer countries with top
consumption categories (Table 1 and Fig. 2), show that most of the
steel is afterwards again exported to be used in various applications Final consumers of New Caledonian Nickel

in different countries. Japan, China and USA were the biggest final
consumers of New Caledonian nickel in 2000, having consumed 45% Others, $340.8
Japan, $255.7
of total production in 2011. The MRIO analysis shows that one final
use of New Caledonian nickel is Japanese passenger cars purchased
in the US, made using Japanese steels containing New Caledonian
alloys. Another top value supply route of nickel to final consumer,
being worth $7.9 million, shows nickel from New Caledonia used
in Japan steel industry for final demand in the Chinese construction Germany, $27.0

sector: “New Caledonia mining and quarrying industry → Japan pig China, $210.8
iron products → Japan crude steel (converters) products → Japan
Italy, $34.3
hot rolled steel products → Final demand in China in the construc-
tion sector”. Other examples from SPA of final consumers include Spain, $36.2

nickeled Japanese steel alloys used in car production in Australia

($4,800,000), wires and cables in railway construction in Japan South Korea, $38.1
($1,800,000), ship building in South Korea ($1,000,000) and storage France, $69.4
battery manufacture in the US ($100,000). Notably, nickel is widely USA, $166.7

used in precious metal alloys to increase the strength, as well as for New Caledonia, $132.6
Australia, $107.2
plating of inexpensive costume jewellery (Nickel Institute, 2009).
A representative path for this “New Caledonia mining and quar- Fig. 2. Final consumers of New Caledonia nickel (direct + indirect flows). Values in
rying industry → Japan other non-ferrous metals products → USA million $US, for 2011.
 D. Moran et al. / Ecological Indicators 60 (2016) 192–201 197

3.2. Tantalum (Coltan) mining in the Democratic Republic of Final consumers of Kenyan Flowers
Congo
Others, $245.2
UK, $413.9
Spain, $14.9
It is not difficult to establish a link between coltan mining and
biodiversity loss in DRC. Most conflict coltan from DRC ultimately
affected the population of gorilla, chimpanzee, African buffalo and China, $19.8
several antelope species both through habitat destruction and frag-
Italy, $25.4
mentation from mining and poaching by miners and other persons.
However, while the overall link between mining and biodiversity Japan, $32.9
loss is clear, due to poor information it is difficult to more precisely
prove a link between mining and any one particular species being France, $34.9

threatened.
USA, $41.9
We draw on a study by Moran et al. (2014) who constructed a
hybrid LCA-MRIO study for columbite–tantalum (coltan) ore mined
in the Democratic Republic of Congo (DRC). We include this case Germany, $56.5
Netherlands, $64.3 Kenya, $253.2
study not simply to repeat previous work but to highlight in the UK Kenya Netherlands Germany USA France Japan Italy China Spain Others
context of this paper the high level of detail possible to achieve
using Hybrid MRIO techniques. Moran et al. (2014) executed a Fig. 3. Final consumers of Kenyan cut flowers (direct + indirect flows). Values in
hybrid LCA-MRIO analysis by separating the DRC “Mining” sector million $US, for 2011.

in the Eora global MRIO into a “Coltan Mining” and “Other Min-
3.3. Flower horticulture at Lake Naivasha, Kenya
ing” sector. This disaggregation step was necessary, since a straight
IO analysis using the original IO table would just follow flows of
Roses are the main product of Kenya’s floriculture industry,
“Mining” products, not separated into coltan vs non-coltan ores,
leading the rose import in the EU with a 38% market share (KFC,
and such a broad product group is too mixed to provide useful
2014). However, also within Kenya itself, the consumption of
results on coltan flows. Their study used data from a range of other
flowers is high. The supply of fresh flowers for the western mar-
sources, including the UN, US Geological Survey, and an inves-
kets/consumers is dependent upon aviation (Everard and Harper,
tigative report by Nest (2011), to accomplish this disaggregation
2002) and 90% of the fresh flowers is transported to the Netherlands
and distinguish coltan flows from general DRC ore flows. With the
(Mekonnen et al., 2012), mostly ending up in Dutch flower auctions.
improved MRIO in hand the Moran et al. (2014) study used the
UK supermarkets are one major retail outlet for Kenyan flowers
same footprinting and SPA methods as discussed above to calcu-
(KFC, 2014). Other destinations include Japan, Russia and the USA.
late coltan footprints and identify the main supply chains conveying
Fig. 3 shows the top final consumers of Kenyan flowers according
embodied coltan.1
to the MRIO analyses. The UK, Netherlands, Germany France, and
Moran et al. (2014) traced conflict coltan to processors and
USA show up as top consumers. For this case study we do not eluci-
refiners in the US, Germany, and Kazakhstan, where it was then
date the particular sectorwise composition of flower purchases in
sold onward, primarily for use in electronics. Germany, USA, China,
each consumer country. Flowers are purchased across all sectors:
UK, and Japan were identified as top final destinations of prod-
by consumers at shops, by restaurants, by hospitals, by businesses,
ucts containing DRC-sourced coltan. Coltan is used in mobile
etc., and the particular sector in which consumption occurs is not
phone batteries and compact capacitors used in entertainment
important since in this case policy responses would most likely tar-
electronics. One representative path was “DRC → Rwanda → USA
get the product itself (flowers) rather than the sector in which those
processing → USA electronic capacitor, resistor, coil, transformer,
flowers are purchased. The horticulture contributed to a decline in
and other inductor manufacturing products → Final demand in
the volume and surface area of the lake and also caused the con-
South Korea in the radio, television, and communications equip-
centrations of pollutants to rise. This has potentially direct negative
ment sector” showing US$75,000 worth of coltan (0.22% of DRC
consequences for the species in and near the lake. This example
coltan mined in 2000); this path would capture flows of coltan
of a clear causal relationship between a single industry and an
into game console electronics and mobile phones purchased by
easily visible consumer product has resulted in much attention
Koreans. The study was also able to capture more obscure paths,
from national and international organizations (FWW and CC, 2008;
for example: “DRC → Rwanda → USA processing → USA electronic
Mekonnen and Hoekstra, 2010; WWF, 2011a; LNRA, 2014).
capacitor, resistor, coil, transformer, and other inductor manu-
Fig. 4
facturing products → Mexico manufacture of transport equipment
In this case study, flowers from Kenya went into the Netherlands
industry → Final demand in Mexico in the manufacture of trans-
Agriculture sector but from there, in the MRIO analysis the flow-
port equipment sector.” In this path $US 57,000 worth of coltan
ers become mixed with all other Agriculture products. In this case
that leaves DRC, is used in electronics within American vehi-
product-specific information reported by the industry association
cles, which are then sold to consumers in Mexico. This path
is a better data source. Another curious finding from the case study
includes both physically embodied coltan (e.g. used in the
is that Kenya itself is a major consumer of flowers. This finding
electronics, airbags, ignition, antilock brake system, or corrosion-
reflects the data reported by the MRIO database, but the database
resistant bolts) and also indirect flows, e.g. when coltan is used
could be incorrect.
for precision manufacturing tools which are used to build the
car.
3.4. Forestry in Papua New Guinea

Given the rapid growth and extensive area of forestry opera-

tions in PNG, and the high value of biodiversity in Papuan forests,
1
Parenthetically, we shall mention an important note regarding the word embod- it is not difficult to establish a connection between forestry activity
ied. In normal usage one would say a product contained embodied coltan if some
coltan were physically present in the final product. However when used in footprint
and biodiversity pressure in the country. The more area is defor-
analysis the term also refers to indirect usage, for example if a product is made in a ested or fragmented, the more species potentially lose (parts of)
factory using coltan-alloyed machine tools. their habitats. Thus they may get increasingly stressed and could
198 D. Moran et al. / Ecological Indicators 60 (2016) 192–201

Table 2
Final consumers of PNG Forestry – Top final consumers, with top implicated consumption goods, of Papuan forestry
South products. Goods and services are domestically produced, using embodied Papuan
New forestry goods, unless an origin country is specified. Data are for year 2011.
Korea, $5.2
UK, $9.7 Zealand,
Top final consumers Top implicated final consumption goods
$4.0
Italy, $10.7 (Percent of total impact) (Percent of total consumption in the consumer
Other, country)
USA, $11.3 $52.4
Germany, Papua New Guinea (64%) Education, Health and Other Services (17%);
$12.7 Public Administration (14%); Financial
Intermediation and Business Activities (13%);
China,
Wood and Paper (11%)
$40.0 Australia (8%) Papuan Education, Health and Other Services
Japan, (18%); Papuan Construction (11%); Papuan
Retail Trade (11%); Papuan Transport (7%)
$46.7 Papua New
Australia, Japan (7%) Residential construction (wooden) (16%);
$59.0 Guinea, Residential construction (non-wooden) (9%);
$449.2 Non-residential construction (non-wooden)
(7%); General eating and drinking places
(except coffee shops) (4%)
China (6%) Construction (48%); Educational services (4%);
Public administration and other sectors (4%);
Furniture and products of wood, bamboo, cane,
Fig. 4. Final consumers of PNG forestry products (direct + indirect flows). Values in
palm, straw, etc. (4%)
million $US, for 2011.
Germany (2%) Food products (16%); Papuan Agriculture
(13%); Papuan Food & Beverages (11%);
Installations and other construction (6%)
potentially go extinct. Since so many species are yet to be discov- USA (2%) Chinese Furniture and products of wood,
ered, the direct link from logging to extinction may be hard (species bamboo, cane, palm, straw, etc. (6%); General
may already be extinct now, without having been discovered first). Federal defence government services (5%);
General state and local government services
Raw logs are the main exported timber commodity and account
(5%); Light truck and utility vehicle
for 90% of timber exported by volume in PNG (Papua New Guinea manufacturing (3%)
Forest Authority, 2009). A total of 3 million m3 of roundwood Italy (2%) Papuan Food & Beverages (36%); Food products
was exported from PNG in 2012, making PNG the second-largest and beverages (16%); Hotel and restaurant
services (16%); Retail trade services, except of
exporter of tropical logs after Malaysia (FAO Statistics Division
motor vehicles and motorcycles; repair
(FAOSTAT) 2012). Even though timber exports have been increas- services of personal and household goods (6%)
ing for more than a decade, the timber industry, which according UK (1%) Papuan Food & Beverages (6%); Papuan Hotels
to Laurance et al. ‘suffers from endemic corruption’ (2011), gener- and Restaurants (3%); Papuan Transport (3%);
ates only a small fraction (5.6%, (United Nations Statistics Division, UK Social work activities (3%)
South Korea (1%) Transportation and warehousing (12%);
2012) of annual export earnings of PNG, and is dwarfed by gold,
Furniture (8%); Papuan Hotels and Restaurants
copper and platinum mining (together >50% of export earnings). (7%); Public administration and defence (7%)
Private logging companies with Malaysian ownership dominate the New Zealand (1%) Residential building construction (26%);
timber industry, further exporting raw wood for manufacturing in Non-residential building construction (16%);
Papuan Food & Beverages (4%); Civil
China, which then supplies consumers all over the world often look-
engineering (4%)
ing for aesthetics in tropical timber rather than necessity (Papua
New Guinea Forest Authority, 2009; Laurance et al., 2011, 2012;
Shearman et al., 2012). Little of the added value of further processed
used by Chinese industries at the same frequency as average sawn
timber ends up in PNG as almost no domestic wood processing is
wood this fact is not a problem since the Papuan wood footprint
taking place in PNG (Papua New Guinea Forest Authority, 2009).
will still end up in Chinese goods. However if Papuan wood enter-
We find that even though China is the major export destination
ing the Chinese Sawn Wood sector faces noticeably different fates
of PNG timber, most timber is processed and delivered onward to
than typical wood entering the sector, this difference will not be
final consumers elsewhere. In the Final Consumers Table 2 China
captured well by the MRIO analysis.
is only the 3rd largest foreign consumer of PNG timber, but it is
also a major re-exporter. Australia and Japan are the largest con-
sumers of PNG timber (after PNG itself), but at least one fifth ends 4. Discussion
up in Europe (Germany, Italy, UK, etc.) and the USA. Structural path
analysis results show that timber, much of which is hardwood, is Input–output and SPA techniques have been used for many
used in building sectors in Japan, China, Australia, New Zealand years across many applications to trace flows of environmental
and other countries most likely for house constructions, flooring, concern. In selecting these case studies we sought to find exam-
wooden panels etc. Japan is the main producer and consumer of ples in which a specific species threat is caused by one (or a
paper products made from PNG timber. As an example we find one small number) of particular industries, the products of which can
path showing $300,000 of PNG timber flowing from “Papua New then be traced through international supply chains. Originally,
Guinea wood and paper industry → Japan pulp products → Japan we had hoped to find case studies with species whose popula-
paper products → Final demand in Japan in the newspapers sector”. tion trends were being tracked by the Living Planet Index (LPI)
Another noteworthy consumption sector is USA Light Trucks: this (WWF, 2014), however we were unable to identify any species
value includes Papuan hardwoods flowing through various trade with LPI time series data that were under threat due to a small
and transformation paths and ending as veneers in decorative car number of specific industries. Often species are under threat from
interior panelling. Finally, similarly to Kenyan flowers we observe an array of causes and it is difficult to implicate a particular com-
that biodiversity-implicated timber enters the Chinese Sawn Wood modity as causing a particular threat. Establishing such a link
Products sector, but there became mixed with all other sawn wood is crucial in order to continue to a supply chain analysis and
products from all other sources. Assuming that Papuan wood is imagine a consumer-oriented policy response. The macroeconomic
 D. Moran et al. / Ecological Indicators 60 (2016) 192–201 199

evaluation conducted by Lenzen et al. (2012b) avoided this prob- a variety of reasons, also pose a threat, although we were unable
lem by assigning all threats, with varying degrees of probability, to to implicate any particular commodity as responsible for threats
all industries. This provides adequate aggregate results, but clear, from fire. Several species are classified as Critically Endangered
defensible, threat ↔ industry links are a necessary starting point because of mining activity, however better published time series
for most realistic policy responses. Even in these carefully selected data would help to more clearly establish a temporal connection
case studies it is not difficult to imagine the implicated industry, or between declining species health and rising mining activity. Finally,
individual companies within that industry, being able to plausibly the trade analysis does provide some useful results highlighting key
deny responsibility for any particular species threat. consumers of embodied New Caledonian nickel and some key flow
In addition to the difficulty of implicating specific companies paths. The data error in the MRIO database, reporting negligible
and industries there are also shortcomings in the MRIO data, even flows from New Caledonia to France, does call into question the
when using the most detailed available global trade database. The validity of the chosen MRIO for this trade analysis.
MRIO databases themselves may contain errors, though in general In the case of the flower industry around Lake Naivasha again
we may expect that MRIO data quality will continue to improve. IO it is not difficult to establish a causative link between the rise of
and SPA analyses are attended by several sources of uncertainty. horticulture and the decline of water bird and other species at the
These sources of error have been well discussed in the literature lake. The MRIO trade analysis, however, is less helpful since many
but we shall summarize them here again as they relate to our Kenyan flower exports arrive into a broad “Netherlands Agricul-
particular question about the accuracy of IO tools for assessing ture” sector where they become mixed with many other goods
product-specific biodiversity footprints. Biodiversity impacts could before being sent onwards to further destinations. While the results
be mischaracterized or mis-attributed to the appropriate causative clearly show that, in aggregate, flowers and many agricultural
sector (e.g. it could be difficult to discern whether an impact associ- goods are re-exported out of the Netherlands, improved data, in
ated with bamboo cultivation should be attributed to an Agriculture this case from an industry association, was needed in order to trace
or Forestry Products sector). In life-cycle analysis this type of error this flows more finely.
relates to so-called system boundaries: which activities and flows Papua New Guinea is known to be a biodiversity hotspot, yet
should be included, or excluded, from a product’s life-cycle inven- quantitative measures of biodiversity health there remains lack-
tory (Suh et al., 2004), and, related to this, are truncation errors in ing. The extensive, and growing, forestry industry is clearly a major
LCA assessments due to incomplete assessment of indirect flows threat driver and thus Papuan timber products are implicated in
(Lenzen, 2001). The economic data in the MRIO table itself, and biodiversity loss.
individual transaction data points within it, are subject to stochas- Finally, as in the other cases, in the coltan study there is a clear
tic error. However, recent studies (Moran and Wood, 2014) suggest link between a specific industry and biodiversity pressure. In the
that uncertainty arising from stochastic error is manageable. Aggre- trade analysis the coltan study unquestionably benefited from the
gation error affects all input–output analyses: the sectors in the superior, detailed, trade data added by the hybrid LCA-MRIO analy-
MRIO table are aggregated sectors and each sector may represent a sis. This indicates that hybrid LCA-MRIO techniques may be a good
bundle of goods or processes with diverse environmental profiles. approach for developing product-specific biodiversity footprints.
This can mean that the sector as a whole is a poor approximation Much of the trade data is provided by the base MRIO, but by adding
of its constituent flows. key elements of additional, more detailed data, even if based on
Aggregation error is a well-studied topic in IO analysis (Andrew models or assumptions, can help increase the accuracy and reliabil-
et al., 2009; Lenzen, 2011; Steen-Olsen et al., 2013) and generally ity of the final result. While product-level biodiversity footprints
is held to be a manageable problem (Alexeeva-Talebi et al., 2012; based on hybrid LCAs will likely remain only initial estimates for
Caron, 2012), mostly to be solved by using newer, higher resolu- some more years, these techniques in terms of research time hybrid
tion IO tables such as Eora. Nevertheless, this issue of sectoral detail LCA methods can provide good results with work on the order of
is particularly relevant for our study. Often developing economies single person-months.
have lower-resolution IO tables, for example reporting a single In general, the limiting factors in MRIO analysis for biodiversity
“Forestry” sector that could contain both sustainably-farmed soft- threats are the spatial detail, economic sectoral detail, and good
wood products and unsustainably harvested hardwood products. links between threats and implicated industries. Biodiversity infor-
These two goods obviously have different environmental profiles, mation is scarce in several ways: in many world regions there is
yet IO analysis treats them as a homogenous product when tracing still a lack of knowledge of species presence, even if they consti-
it through trade and transformation steps. The solution is to use tute biodiversity hotspots. Also, as this is a very complex topic,
higher resolution IO tables if available, or to find superior data with the actual, direct threats affecting individual species communities
which to split the sector into finer, disaggregated, products. This are often unknown, making it hard to find robust species-threat
latter technique is called hybrid MRIO or hybrid LCA analysis (Suh links. However, as ecology and biology, as well as economic stud-
and Huppes, 2000; Suh et al., 2003, 2004; Suh and Huppes, 2005; ies, advance, we are confident that the relevant data will become
Suh and Nakamura, 2007; Ewing et al., 2011; Weinzettel et al., available in future, allowing for an effective tracing of biodiver-
2014). Hybrid MRIO analysis provides generally superior results, sity related impacts of specific industries throughout the world
but has high data demands, on par with those of a product-specific economy.
life-cycle analysis (LCA) study. In this study we sought to exam-
ine whether macroeconomic MRIO analysis is detailed enough to
provide useful information to guide more detailed studies at the 5. Conclusions
product specific level. However, we found that three of the case
studies (Kenya, Papua New Guinea and New Caledonia) suffered These case studies provide interesting insights into the chal-
due to aggregation problems, in contrast to the case study of coltan lenges of connecting specific industries to particular biodiversity
(Congo), which used a hybrid LCA MRIO system for analysis. threats. They also highlight the possibilities, and difficulties, of
Let us now turn to the individual case studies and what lessons tracing these commodities through supply chains, in order to cal-
may be learned from each. culate detailed biodiversity footprints. Despite the challenges, IO
In the case of nickel mining in New Caledonia it is clear that techniques can be highly useful for outlining supply chains and
open-cast mining and associated land use change is driving threats identifying key consumption sectors, and this in turns can help
to endemic metallophyte species. However fires, lit to clear land for prioritize how and where further research is needed in order to
200 D. Moran et al. / Ecological Indicators 60 (2016) 192–201

provide robust, detailed data usable by governments or companies conservation actions to protect East African lakes from the negative impacts
to start forming economic responses beyond the point of imme- of alien species. Biol. Conserv. 144 (11), 2585–2596.
Gorenflo, L.J., Romaine, S., Mittermeier, R.A., Walker-Painemilla, K., 2012. Co-
diate impact. IO may be highly suitable for such initial screening occurrence of linguistic and biological diversity in biodiversity hotspots and
given its relatively low cost: with an MRIO database in hand, cal- high biodiversity wilderness areas. Proc. Natl. Acad. Sci. 109 (21), 8032–
culating a biodiversity footprint is relatively simple and does not 8037.
Graedel, T.E., Allenby, B.R., 1995. Industrial Ecology. AT&T-Prentice Hall, Englewood
involve the months of additional work that a more in-depth LCA Cliffs, NJ.
or similar study requires. Green labelling and certifications, trade Graedel, T.E., Bertrama, M., Fuse, K., Gordon, R.B., Lifset, R., Rechbergera, H., Spataria,
controls, corporate engagement, and sustainable purchasing have S., 2002. The contemporary European copper cycle: the characterization of tech-
nological copper cycles. Ecol. Econ. 42 (1–2), 9–26.
all been shown to be useful consumption-oriented tools for allevi-
Graedel, T.E., Harper, E.M., Nassar, N.T., Reck, B.K., 2013. On the materials basis of
ating environmental pressure by reducing demand for implicated modern society. Proc. Natl. Acad. Sci. 112 (20), 6295–6300.
products. IO analysis of product-specific biodiversity footprints can Harper, D., Mavuti, K., 2004. Lake Naivasha, Kenya: ecohydrology to guide the
management of a tropical protected area. Ecohydrol. Hydrobiol. 4 (3), 287–
provide a starting point for developing such measures.
305.
Harper, D.M., Harper, M.M., Virani, M.A., Smart, A., Childress, R.B., Adatia, R., Hender-
Acknowledgment son, I., Chege, B., 2002. Population fluctuations and their causes in the African
fish eagle, (Haliaeetus vocifer (Daudin)) at Lake Naivasha, Kenya. Hydrobiologia
488, 171–180.
This study was conducted in part in support of the TSUNAGARI Hayes, K., Burge, R., 2003. Coltan Mining in the Democratic Republic of Congo: how
project of the Belmont Forum. tantalum-using industries can commit to the reconstruction of the DRC. In: Flora
& Fauna International Conservation Reports. F. F. International. Flora & Fauna
International, Cambridge, UK.
References Hickley, P., Muchiri, M., Boar, R., Britton, R., Adams, C., Gichuru, N., Harper, D., 2004.
Habitat degradation and subsequent fishery collapse in Lakes Naivasha and
Alexeeva-Talebi, V., Böhringer, C., Löschel, A., Voigt, S., 2012. The value-added of Baringo, Kenya. Ecohydrol. Hydrobiol. 4 (4), 503–517.
sectoral disaggregation: implications on competitive consequences of climate Ibanez, T., Munzinger, J., Dagostini, G., Hequet, V., Rigault, F., Jaffré, T., Birnbaum,
change policies. Energy Econ. 34 (Suppl. 2), S127–S142. P., 2014. Structural and floristic diversity of mixed tropical rain forest in New
Ali, S.H., Grewal, A.S., 2006. The ecology and economy of indigenous resistance: Caledonia: New data from the New Caledonian Plant Inventory and Permanent
divergent perspectives on mining in New Caledonia. Contemp. Pacific 18 (2), Plot Network (NC-PIPPN). Appl. Veg. Sci. 17 (3), 386–397.
361–392. IUCN, 2008. The Pacific islands: an analysis of the status of species as listed on the
Andrew, R., Peters, G., Lennox, J., 2009. Approximation and regional aggregation in 2008 IUCN Red List of Threatened Species. H. Pippard. IUCN Regional Office for
multi-regional input–output analysis. Econ. Syst. Res. 21 (3), 311–335. Oceania.
Awange, J.L., Forootan, E., Kusche, J., Kiema, J.B.K., Omondi, P.A., Heck, B., Fleming, K., Jaffré, T., Munzinger, J., Lowry Ii, P.P., 2010a. Threats to the conifer species found on
Ohanya, S.O., Gonçalves, R.M., 2013. Understanding the decline of water storage New Caledonia’s ultramafic massifs and proposals for urgently needed measures
across the Ramser-Lake Naivasha using satellite-based methods. Adv. Water Res. to improve their protection. Biodiv. Conserv. 19 (5), 1485–1502.
60, 7–23. Jaffré, T., Munzinger, J., Lowry II., P., 2010f. Threats to the conifer species found on
Becht, R., Harper, D.M., 2002. Towards an understanding of human impact upon the New Caledonia’s ultramafic massifs and proposals for urgently needed measures
hydrology of Lake Naivasha, Kenya. Hydrobiologia 488, 1–11. to improve their protection. Biodiv Conserv. 19 (5), 1485–1502.
BirdLife International, 2014. Important Bird Areas factsheet: Lake Naivasha, Kettle, C.J., Hollingsworth, P.M., Jaffré, T., Moran, B., Ennos, R.A., 2007. Identifying
Retrieved from http://www.birdlife.org/datazone/sitefactsheet.php?id=6438 the early genetic consequences of habitat degradation in a highly threatened
(accessed 08.07.14). tropical conifer Araucaria nemorosa Laubenfels. Mol. Ecol. 16 (17), 3581–3591.
Brooks, T.M., Mittermeier, R.A., Fonseca, G.A.B.d., Gerlach, J., Hoffmann, M., Lam- KFC, 2014. Kenya Flower Council, Retrieved from http://www.kenyaflowercouncil.
oreux, J.F., Mittermeier, C.G., Pilgrim, J.D., Rodrigues, A.S.L., 2006. Global org/ (accessed 20.07.14).
biodiversity conservation priorities. Science 313 (5783), 58–61. Kitaka, N., Harper, D., Mavuti, K., Pacini, N., 2002. Chemical characteristics, with
Butchart, S.H.M., Walpole, M., Collen, B., van Strien, A., Scharlemann, J.P.W., Almond, particular reference to phosphorus, of the rivers draining into Lake Naivasha,
R.E.A., Baillie, J.E.M., Bomhard, B., Brown, C., Bruno, J., Carpenter, K.E., Carr, G.M., Kenya. Hydrobiologia 488 (1–3), 57–71.
Chanson, J., Chenery, A.M., Csirke, J., Davidson, N.C., Dentener, F., Foster, M., Galli, Kitzes, J., 2013. An introduction to environmentally-extended input–output analy-
A., Galloway, J.N., Genovesi, P., Gregory, R.D., Hockings, M., Kapos, V., Lamarque, sis. Resources 2 (4), 489–503.
J.-F., Leverington, F., Loh, J., McGeoch, M.A., McRae, L., Minasyan, A., Morcillo, Lammerant, J., Vertriest, L., Peters, R., Lawlor, N., Hernandez, G., Markowska, A.,
M.H., Oldfield, T.E.E., Pauly, D., Quader, S., Revenga, C., Sauer, J.R., Skolnik, B., Homeyer, I.V., Moran, D., 2012. Identification and mitigation of the negative
Spear, D., Stanwell-Smith, D., Stuart, S.N., Symes, A., Tierney, M., Tyrrell, T.D., impacts of EU demand for certain commodities on biodiversity in third countries
Vié, J.-C., Watson, R., 2010. Global biodiversity: indicators of recent declines. (ENV.B.2/ETU/2012/0045r).
Science 328 (5982), 1164–1168. Laurance, W.F., Kakul, T., Keenan, R.J., Sayer, J., Passingan, S., Clements, G.R., Villegas,
Caron, J., 2012. Estimating carbon leakage and the efficiency of border adjustments in F., Sodhi, N.S., 2011. Predatory corporations, failing governance, and the fate of
general equilibrium – does sectoral aggregation matter? Energy Econ. 34 (Suppl. forests in Papua New Guinea. Conserv. Lett. 4 (2), 95–100.
2), S111–S126. Laurance, W.F., Kakul, T., Tom, M., Wahya, R., Laurance, S.G., 2012. Defeating the
Chapin, F.S., Zavaleta, E.S., Ebiner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L., ‘resource curse’: Key priorities for conserving Papua New Guinea’s native forests.
Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C., Diaz, S., 2000. Con- Biol. Conserv. 151 (1), 35–40.
sequences of changing biodiversity. Nature 405, 234–242. Lefort, C., Burton, M., 2014. Protesters burn vehicles, buildings at New Caledonia
Chapman, S., Buttler, A., Francez, A.-J., Laggoun-Défarge, F., Vasander, H., Schloter, nickel mine. Reuters, Sydney.
M., Combe, J., Grosvernier, P., Harms, H., Epron, D., Gilbert, D., Mitchell, E., 2003. Lenzen, M., 2001. Errors in conventional and input–output-based life-cycle inven-
Exploitation of northern peatlands and biodiversity maintenance: a conflict tories. J. Ind. Ecol. 4 (4), 127–148.
between economy and ecology. Front. Ecol. Environ. 1 (10), 525–532. Lenzen, M., 2003. Environmentally important paths, linkages and key sectors in the
Defourny, J., Thorbecke, E., 1984. Structural path analysis and multiplier decompo- Australian economy. Struct. Change Econ. Dyn. 14 (1), 1–34.
sition within a social accounting matrix framework. Econ. J. 94, 111–136. Lenzen, M., 2006. Structural path analysis of ecosystem networks. Ecol. Modell. 200,
Dirzo, R., Young, H.S., Galetti, M., Ceballos, G., Isaac, N.J.B., Collen, B., 2014. Defauna- 334–342.
tion in the anthropocene. Science 345 (6195), 401–406. Lenzen, M., 2011. Aggregation versus disaggregation in input–output analysis of the
Enniskillen, A., 2002. The Lake Naivasha management plan – consensus-building to environment. Econ. Syst. Res. 23 (1), 73–89.
conserve an international gem. Hydrobiologia 488, 9–12. Lenzen, M., Kanemoto, K., Moran, D., Geschke, A., 2012a. Mapping the structure of
Everard, M., Harper, D., 2002. Towards the sustainability of the Lake Naivasha Ramsar the world economy. Environ. Sci. Technol. 46 (15), 8374–8381.
site and its catchment. Hydrobiologia 488 (1–3), 191–203. Lenzen, M., Moran, D., Kanemoto, K., Foran, B., Lobefaro, L., Geschke, A., 2012b. Inter-
Ewing, A., Thabrew, L., Perrone, D., Abkowitz, M., Hornberger, G., 2011. Insights on national trade drives biodiversity threats in developing nations. Nature 486,
the use of hybrid life cycle assessment for environmental footprinting. J. Ind. 109–112.
Ecol. 15 (6), 937–950. Leontief, W., 1986. Input–Output Economics. Oxford University Press, New York, NY,
FAO Statistics Division (FAOSTAT), 2012. ForeSTAT. Rome, Italy, Food and Agriculture USA.
Organization of the United Nations. LNRA., 2014. Lake Naivasha Riparian Association, Retrieved 21/07/2014.
Farjon, A., Page, C.N., 2003. Diversity in strategies for conifer conservation, the action Loh, J., Green, R.E., Ricketts, T., Lamoreux, J., Jenkins, M., Kapos, V., Randers, J., 2005.
plan and future developments. Acta Hortic. 615, 397–403. The Living Planet Index: using species population time series to track trends in
Filer, C., Keenan, R.J., Allen, B.J., McAlpine, J.R., 2009. Deforestation and forest degra- biodiversity. Philos. Trans. R. Soc. B: Biol. Sci. 360 (1454), 289–295.
dation in Papua New Guinea. Ann. Forest Sci. 66 (8), 813p811ndash;813p812. Luck, G.W., Ricketts, T.H., Daily, G.C., Imhoff, M., 2004. Alleviating spatial conflict
FWW and CC, 2008. Lake Naivasha withering under the assault of international between people and biodiversity. Proc. Natl. Acad. Sci. 101 (1), 182–186.
flower vendors. MacLean, I., Hassall, M., Boar, R., Nasirwa, O., 2003. Effects of habitat degradation on
Gherardi, F., Robert Britton, J., Mavuti, K.M., Pacini, N., Grey, J., Tricarico, E., Harper, avian guilds in East African papyrus Cyperus papyrus swamps. Bird Conserv. Int.
D.M., 2011. A review of allodiversity in Lake Naivasha, Kenya: developing 13 (4), 283–297.
 D. Moran et al. / Ecological Indicators 60 (2016) 192–201 201

Mekonnen, M.M., Hoekstra, A.Y., 2010. Mitigating the water footprint of export cut Papua New Guinea: [Nairobi, Kenya], Dept. of Environment and Conservation.
flowers from the Lake Naivasha Basin, Kenya. Delft the Netherlands. UNESCO- Conservation Resource Centre; Africa Centre for Resources and Environment.
IHE Institute for Water Education, pp. 54. Shearman, P., Bryan, J., 2011. A bioregional analysis of the distribution of rainforest
Mekonnen, M.M., Hoekstra, A.Y., Becht, R., 2012. Mitigating the Water Footprint of cover, deforestation and degradation in Papua New Guinea. Aust. Ecol. 36 (1),
Export Cut Flowers from the Lake Naivasha Basin, Kenya. Water Resour. Manage. 9–24.
26 (13), 3725–3742. Shearman, P., Bryan, J., Laurance, W.F., 2012. Are we approaching ‘peak timber’ in
Mittermeier, R.A., Myers, N., Tliomsen, J.B., Olivieri, S., 1998. Biodiversity hotspots the tropics? Biol. Conserv. 151 (1), 17–21.
and major tropical wilderness areas: Approaches to setting conservation prior- Shearman, P.L., Ash, J., MacKey, B., Bryan, J.E., Lokes, B., 2009. Forest conversion and
ities. Conserv. Biol. 12 (3), 516–520. degradation in Papua New Guinea 1972–2002. Biotropica 41 (3), 379–390.
Mittermeier, R.A., Werner, T.B., Lees, A., 1996. New Caledonia – a conservation Smil, V., 2002. The Earth’s Biosphere: Evolution, Dynamics, and Change. MIT Press,
imperative for an ancient land. Oryx 30 (02), 104–112. Cambridge, MA.
Moran, D., Wood, R., 2014. Convergence between The Eora, WIOD EXIOBASE, Steen-Olsen, K., Owen, A., Hertwich, E.G., Lenzen, M., 2013. Effects of aggregation
and OPENEU’s Consumption-Based Carbon Accounts. Econ. Syst. Res. 26 (3), in MRIO-based environmental accounting. In: 21st International Input–Output
245–261. Conference. Kitakyushu, Japan.
Moran, D.D., McBain, D., Kanemoto, K., Lenzen, M., Geschke, A., 2015. Global supply Suh, S., Heijungs, R., 2007. Power series expansion and structural analysis for life
chains of coltan. J. Ind. Ecol. 19 (3), 357–365. cycle assessment. Int. J. Life Cycle Assess. 12 (6), 381–390.
Morat, P., 1993. Our knowledge of the flora of New Caledonia: endemism and diver- Suh, S., Huppes, G., 2000. Gearing input–output model to LCA – Part I: general
sity in relation to vegetation types and substrates. Biodiv. Lett. 1 (3/4), 72–81. framework for hybrid approach. CML, Leiden University, Leiden, Netherlands.
Morat, P., Jaffré, T., Tronchet, F., Munzinger, J., Pillon, Y., Veillon, J.M., Chalopin, M., Suh, S., Huppes, G., 2005. Methods for Life Cycle Inventory of a product. J. Cleaner
Birnbaum, P., Rigault, F., Dagostini, G., Tinel, J., Lowry Ii, P.P., 2012. The taxonomic Prod. 13 (7), 687–697.
reference base Florical and characteristics of the native vascular flora of New Suh, S., Lenzen, M., Treloar, G.J., Hondo, H., Horvath, A., Huppes, G., Jolliet, O., Klann,
Caledonia. Adansonia 34 (2), 179–221. U., Krewitt, W., Moriguchi, Y., Munksgaard, J., Norris, G., 2003. System boundary
Morrison, E.H.J., Harper, D.M., 2009. Ecohydrological principles to underpin the selection in life-cycle inventories using hybrid approaches. Environ. Sci. Technol.
restoration of Cyperus papyrus at Lake Naivasha, Kenya. Ecohydrol. Hydrobiol. 38 (3), 657–664.
9 (1), 83–97. Suh, S., Lenzen, M., Treloar, G.J., Hondo, H., Horvath, A., Huppes, G., Jolliet, O., Klann,
Myers, N., Mittermeler, R.A., Mittermeler, C.G., Da Fonseca, G.A.B., Kent, J., 2000. U., Krewitt, W., Moriguchi, Y., Munksgaard, J., Norris, G., 2004. System boundary
Biodiversity hotspots for conservation priorities. Nature 403 (6772), 853–858. selection in life-cycle inventories. Environ. Sci. Technol. 38 (3), 657–664.
Ndungu, J., Augustijn, D.C.M., Hulscher, S.J.M.H., Kitaka, N., Mathooko, J., 2013. Suh, S., Nakamura, S., 2007. Five years in the area of input–output and hybrid LCA.
Spatio-temporal variations in the trophic status of Lake Naivasha, Kenya. Lakes Int. J. Life Cycle Assess. 12 (6), 351–352.
Reserv.: Res. Manage. 18 (4), 317–328. Tittensor, D.P., Walpole, M., Hill, S.L.L., Boyce, D.G., Britten, G.L., Burgess, N.D.,
Nelson, P.N., Gabriel, J., Filer, C., Banabas, M., Sayer, J.A., Curry, G.N., Koczberski, G., Butchart, S.H.M., Leadley, P.W., Regan, E.C., Alkemade, R., Baumung, R., Bellard, C.,
Venter, O., 2014. Oil palm and deforestation in Papua New Guinea. Conserv. Lett. Bouwman, L., Bowles-Newark, N.J., Chenery, A.M., Cheung, W.W.L., Christensen,
7 (3), 188–195. V., Cooper, H.D., Crowther, A.R., Dixon, M.J.R., Galli, A., Gaveau, V., Gregory, R.D.,
Nest, M., 2011. Coltan. Polity Press. Gutierrez, N.L., Hirsch, T.L., Höft, R., Januchowski-Hartley, S.R., Karmann, M.,
Nickel Institute, 2009. Nickel & Yewellery. Krug, C.B., Leverington, F.J., Loh, J., Lojenga, R.K., Malsch, K., Marques, A., Mor-
Nickel Institute, 2014. Nickel Metal – The Facts, Retrieved from http://www. gan, D.H.W., Mumby, P.J., Newbold, T., Noonan-Mooney, K., Pagad, S.N., Parks,
nickelinstitute.org/NickelUseInSociety/AboutNickel/NickelMetaltheFacts.aspx B.C., Pereira, H.M., Robertson, T., Rondinini, C., Santini, L., Scharlemann, J.P.W.,
(accessed 08.08.14). Schindler, S., Sumaila, U.R., Teh, L.S.L., van Kolck, J., Visconti, P., Ye, Y., 2014. A
Otiang’a-Owiti, G.E., Oswe, I.A., 2007. Human impact on lake ecosystems: The case mid-term analysis of progress toward international biodiversity targets. Science
of Lake Naivasha, Kenya. Afr. J. Aquat. Sci. 32 (1), 79–88. 346 (6206), 241–244.
Papua New Guinea Forest Authority, 2009. Papua New Guinea Forestry Outlook Tukker, A., Dietzenbacher, E., 2013. Global multiregional input–output frameworks:
Study Asia-Pacific Forestry Sector Outlook Study II. Working paper series. an introduction and outlook. Econ. Syst. Res. 25 (1), 1–19.
Bangkok. UN Security Council, 2001. Report of the Panel of Experts on the Illegal Exploitation
Pascal, M., De Forges, B.R., Le Guyader, H., Simberloff, D., 2008. Mining and other of Natural Resources and Other Forms of Wealth of the Democratic Republic of
threats to the New Caledonia biodiversity hotspot. Conserv. Biol. 22 (2), 498–499. the Congo.
Peters, G., Davis, S.J., Andrew, R.M., 2012. A synthesis of carbon in international trade. United Nations Statistics Division, 2012. UN COMTRADE (COMmodity TRADE Statis-
Biogeosci. Discuss. 9, 3949–4023. tics Database). United Nations Statistics Division, UNSD, New York, USA.
Peters, G.P., Hertwich, E.G., 2006. Structural analysis of international trade: environ- USGS, 2012. Minerals Yearbook 2011 New Caledonia. U. S. G. Survey.
mental impacts of Norway. Econ. Syst. Res. 18, 155–181. USGS, 2013. Minerals Yearbook 2011 Ferroalloys.
Pimm, S.L., Russell, G.J., Gittleman, J.L., Brooks, T.M., 1995. The future of biodiversity. USG, 2014a. Minerals Yearbook 2012 New Caledonia. U. S. G. Survey.
Science (Washington) 269 (5222), 347. USGS, 2014d. Nickel Mineral Commodity Summaries.
PNG 4th National Report on the CBD, 2010. Papua New Guinea’s Fourth National Weinzettel, J., Steen-Olsen, K., Hertwich, E.G., Borucke, M., Galli, A., 2014. Ecological
Report on the Convention on Biological Diversity. Port Moresby, Papua New footprint of nations: comparison of process analysis, and standard and hybrid
Guinea. multiregional input–output analysis. Ecol. Econ. 101 (0), 115–126.
Proctor, J., 2003. Vegetation and soil and plant chemistry on ultramafic rocks in the Whiting, S.N., Reeves, R.D., Richards, D., Johnson, M.S., Cooke, J.A., Malaisse, F., Paton,
tropical Far East. Perspect. Plant Ecol. Evol. Syst. 6 (1–2), 105–124. A., Smith, J.A.C., Angle, J.S., Chaney, R.L., Ginocchio, R., Jaffré, T., Johns, R., McIn-
Ramsar, 2012. The Annotated Ramsar List of Wetlands of International Importance: tyre, T., Purvis, O.W., Salt, D.E., Schat, H., Zhao, F.J., Baker, A.J.M., 2004. Research
Kenya, Retrieved from http://www.ramsar.org/cda/en/ramsar-documents-list- priorities for conservation of metallophyte biodiversity and their potential for
anno-kenya/main/ramsar/1-31-218%5E16536 4000 0 (accessed 08.08.14). restoration and site remediation. Restor. Ecol. 12 (1), 106–116.
Reck, B.K., Graedel, T.E., 2012. Challenges in metal recycling. Science 337 (6095). Wiedmann, T., 2009. Editorial: carbon footprint and input–output analysis – an
Reck, B.K., Müller, D.B., Rostkowski, K., Graedel, T.E., 2008. The anthropogenic introduction. Econ. Syst. Res. 21 (3), 175–186.
nickel cycle: Insights into use, trade, and recycling. Environ. Sci. Technol. 42, Wood, R., Lenzen, M., 2009. Structural path decomposition. Energy Econ. 31,
3394–3400. 335–341.
Redmond, I., 2001. Coltan boom, gorilla bust. Dian Fossey Gorilla Fund Europe & Wulff, A.S., Hollingsworth, P.M., Ahrends, A., Jaffré, T., Veillon, J.M., L’Huillier, L.,
Born Free Foundation. Fogliani, B., 2013. Conservation priorities in a biodiversity hotspot: analysis of
Sanderson, E.W., Jaiteh, M., Levy, M.A., Redford, K.H., Wannebo, A.V., Woolmer, G., narrow endemic plant species in New Caledonia. PLOS ONE 8 (9.).
2002. The human footprint and the last of the wild. BioScience 52 (10), 891–904. WWF, 2011a. Shared risk and opportunity in water resources: seeking a sustainable
Schaffartzik, A., Sachs, M., Wiedehofer, D., Eisenmenger, N., 2014. Environmentally future for Lake Naivasha. WWF Report.
Extended Input–Output Analysis. U. o. K. Institute for Social Ecology, Vienna. WWF, 2014. Living Planet Report 2014. WWF International, Geneva.
Sekhran, N., Miller, S., Papua New, E., Department of Conservation. Conservation WWF. W.M.P.O., 2011b. Final Frontier: Newly Discovered Species of New Guinea
Resource, R. Africa Centre for and Environmen, 1995. Papua New Guinea coun- 1998–2008. WWF Western Melanesia Programme Office.
try study on biological diversity/edited by N. Sekhran and S. Miller. Waigani,