Global Biodiversity
Conservation Priorities
T. M. Brooks,
1,2,3
*
R. A. Mittermeier,
1
G. A. B. da Fonseca,
1,4
J. Gerlach,
5,6
M. Hoffmann,
1
J. F. Lamoreux,
3
C. G. Mittermeier,
1
J. D. Pilgrim,
7
A. S. L. Rodrigues
5
The location of and threats to biodiversity are distributed unevenly, so prioritization is essential to
minimize biodiversity loss. To address this need, biodiversity conservation organizations have
proposed nine templates of global priorities over the past decade. Here, we review the concepts,
methods, results, impacts, and challenges of these prioritizations of conservation practice within
the theoretical irreplaceability/vulnerability framework of systematic conservation planning. Most
of the templates prioritize highly irreplaceable regions; some are reactive (prioritizing high
vulnerability), and others are proactive (prioritizing low vulnerability). We hope this synthesis
improves understanding of these prioritization approaches and that it results in more efficient
allocation of geographically flexible conservation funding.
H
uman actions are causing a biodiversity
crisis, with species extinction rates up
to 1000 times higher than background
(1). Moreover, the processes driving extinc-
tion are eroding the environmental services on
which humanity depends (2). People care most
about what is close to them, so most responses
to this crisis will be local or national (3). Thus,
approximately 90% of the $6 billion of annual
conservation funding originates in and is spent
within economically rich countries (4). How-
ever, this leaves globally flexible funding of
hundreds of millions of dollars annually from
multilateral agencies (such as the Global En-
vironment Facility), bilateral aid, and private
sources including environmentally focused cor-
porations, foundations, and individuals. These
resources are frequently the only ones available
where conservation is most needed, given that
biodiversity is unevenly distributed and the most
biodiverse places are often the most threatened
and poorest economically (5). Accordingly, geo-
graphically flexible resources exert dispropor-
tionate influence on conservation worldwide
and have a key role in the recently agreed-upon
intergovernmental 2010 target to reduce signif-
icantly the rate of biodiversity loss (6).
The development of strategies to best allo-
cate glob ally flexible conservation resources
has attracted considerable attention since the
pioneering work of Myers (7), resulting in
much progress as well as much controversy.
The wide variety of approaches has led to crit-
icism that there is duplication of effort and lack
of clarity (8). Although attempts have been
made to summarize conservation planning
strategies by scale (9), none has done so within
the framework of conservation planning (10).
We review the published concepts and methods
behind global biodiversity conservation priori-
tization, assess the remaining challenges, and
highlight how this synthesis can inform alloca-
tion of globally flexible resources.
Global Prioritization in Context
Nine major institutional templates of global bio-
diversity conservation prioritization have been
published over the past decade, each with
involvement from nongovernmental organiza-
tions (fig. S1). Conceptually, they all fit within
the framework of ‘ ‘irreplaceability’’ relative to
‘‘vulnerability’ ’ (Fig. 1), which is central to con-
servation planning theory (10). However, they
map onto different portions of the framework:
Most of the templates prioritize high irreplace-
ability, but some prioritize high
vulnerability and others priori tize
low vulnerability. These differences
are key to understanding how and
why the nine prioritizations differ,
yielding priority maps that cover
from less than one-tenth to more
than a third of Earth’s land surface
(Fig. 2).
Six of the nine templates of
global conservation priority incor-
porate irreplaceability—measures
of spatial conservation options
(10). The most common measure
of irreplaceability is plant (11–14)
or bird (15) endemism, often sup-
ported by terrestrial vertebrate en-
demism overall (11, 13, 14). The
logic for this is that greater the
number of endemic species in a
region, the more biodiversity is lost if that
region is lost (although, in a strict sense, any
location with even one endemic species is
irreplaceable). In addition to the number of
endemic species, other aspects of irreplaceabil-
ity have been proposed, including taxonomic
uniqueness, unusual phenomena, and global
rarity of major habitat types (16), but these re-
main difficult to quantify. Although species rich-
ness within a given area is popularly assumed
to be important in prioritization, none of the
approaches relies on species richness alone.
This is because species richness is driven by
common, widespread species; thus, strategies
focused on species richness tend to miss exactly
those biodiversity features most in need of
conservation (17, 18). Three approaches do not
incorporate irreplaceability (19–21).
The choice of irreplaceability measures is to
some degree subjective, in that data limitations
currently preclude the measurement of overall
biodiversity. Furthermore, these data constraints
mean that, with the exception of endemic bird
areas (15), the measures of irreplaceability used
in global conservation prioritization have been
derived from the opinions of specialists. Sub-
sequent tests of plant endemism estimates (22)
have shown this expert opinion to be quite ac-
curate. However, reliance on specialist opinion
means that results cannot be replicated, raising
questions concerning the transparency of the
approaches (8). It also prevents a formal mea-
surement of irreplaceability, which requires the
identities of individual biodiversity features, such
as species names, rather than just estimates of
their magnitude expressed as a number (8, 23).
Five of the templates of global conservation
priority incorporate vulnerability—measures of
temporal conservation options (10). A recent
classification of vulnerability (24) recognizes
four types of measures: (i) environmental and
spatial variables, (ii) land tenure, (iii) threatened
species, and (iv) expert opinion. Of these,
environmental and spatial variables have been
REVIEW
1
Conservation International, 1919 M Street, NW , Washington,
DC 20036, USA.
2
World Agroforestry Centre (ICRAF), Post
Office Box 35024, University of the Philippines, Los Ban
˜
os,
Laguna 4031, Philippines.
3
Department of Environmental
Sciences, University of Virginia, Charlottesville, VA 22904,
USA.
4
Departamento de Zoologia, Universidade Federal de
Minas Gerais, Belo Horizonte, MG 31270, Br azil.
5
Department
of Zoology, University of Cambridge, Downing Street,
Cambridge CB2 3EJ, UK.
6
Nature Protection Trust of
Seychelles, Post Office Box 207, Victoria, Mahe´, Seychelles.
7
BirdLife International in Indochina, 4/209 Doi Can Street, Ba
Dinh, Hanoi, Vietnam.
*To whom correspondence should be addressed. E-mail:
t.brooks@conservation.org
Vulnerability
B A
Irreplaceability
Proactive Reactive
CE
LW
FF
HBWA
EBA, CPD
MC, G200
BH
Proactive Reactive
Fig. 1. Global biodiversity conservation priority templates placed
within the conceptual framework of irreplaceability and vulner-
ability . T emplate names are spelled out in the Fig. 2 legend. (A)
Purely reactive (prioritizing low vulnerability) and purely pro-
active (prioritizing high vulnerability) approaches. (B)Approaches
that do not incorporate vulnerability as a criterion (all prioritize
high irreplaceability).
7 JULY 2006 VOL 313 SCIENCE www.sciencemag.org
58
used most frequently in global conservation
prioritization, measured as proportionate habitat
loss (11, 14, 20, 21). Species-area relationships
provide justification that habitat loss translates
into biodiversity loss (1). However, the use of
habitat loss as a measure of vulnerability has
several problems: It is difficult to assess with
the use of remote sensing for xeric and aquatic
systems, it does not incorporate threats such as
invasive species and hunting pressure, and it is
retrospective rather than predictive (24). The
frontier forests approach (19) uses absolute
forest cover as a measure.
In addition to habitat loss, land tenure—
measured as protected area coverage—has also
been incorporated into two approaches (16, 21).
Other possible surrogates not classified by
Wilson et al.(24) include human population
growth and density, which are widely thought to
be relevant (25–27) and were integral to two of
the systems (14, 20). None of the global con-
servation prioritization templates used threatened
species or expert opinion as measures of vul-
nerability. Political and institutional capacity
and governance (27) affect biodiversity indi-
rectly, but have not been incorporated to date.
This is true for climate change as well, which is
of concern given that its impact is likely to be
severe (28). Finally, although costs of conser-
vation generally increase as the threat increases,
no proposals for global biodiversity conserva-
tion priority have yet incorporated costs direct-
ly, despite the availabilit y of techniques to
do this at regional scales (29). Two of the tem-
plates of global conservation prioritization do
not incorporate vulnerability (12, 13), and the
remaining two incorporate it only peripherally
(15, 16).
The spatial units most commonly used in sys-
tematic conservation planning are equal-area
grids. However, data limitations have precluded
their use in the development of actual templates of
global biodiversity conservation priority to date.
Instead, all proposals, with the exception of mega-
diversity countries (13), are based on biogeo-
graphic units. Typically, these units are defined a
priori by specialist perception of the distribution of
biodiversity. For example, ‘‘ecoregions,’ ’ one of
the most commonly used such classifications, are
‘‘relatively large units of land containing a
characteristic set of natural communities that share
a large majority of their species, dynamics, and
environmental conditions’’ (16). Only in the
endemic bird areas approach are biogeographic
units defined a posteriori by the distributions of
the species concerned (15). Relative to equal-
area grids, biogeographic units bring advantages
of ecological relevance, whereas megadiversity
countries (13) bring political relevance.
Reliance on biogeographic spatial units raises
several complications. Various competing biore-
gional classifications are in use (30), a nd the
choice of system has considerable repercussions
for resulting conservation priorities. Furthermore,
when unequally sized units are used, priority may
be biased toward large areas as a consequence of
species-area relationships. Therefore, assessment
of global conservation priorities should factor out
area, either by taking residuals about a best-fit
line to a plot of species against area (18)orby
rescaling numbers of endemic species with the
use of a power function (23). Nevertheless, the
use of a priori bioregional units for global
conservation prioritization will be essential until
data of sufficient resolution become available
to enable the use of grids.
In Fig. 3, we map the overlay of the global
biodiversity conservation priority systems into
geographic space from the conceptual frame-
work of Fig. 1. Figure 3A illustrates the large
degree of overlap between templates that pri-
oritize highly vulnerable regions of high ir-
replaceability: tropical islands and mountains
(including montane Mesoamerica, the Andes,
the Brazilian Atlantic forest, Madagascar, mon-
tane Africa, the Western Ghats of India, Ma-
laysia, Indonesia, the Philippines, and Hawaii),
Mediterranean-type systems (including Califor-
nia, central Chile, coastal South Africa, south-
west Australia, and the Mediterranean itself),
and a few temperate forests (the Caucasus, the
central Asian mountains, the Himalaya, and
southwest China). Highly vulnerable regions of
lower irreplaceability (generally, the rest of
the northern temperate regions) are priori-
tized by fewer approaches. Figure 3B shows a
large amount of overlap between templates
for regions of low vulnerability but high ir-
replaceability, in particular the three major
tropical rainforests of Amazonia, the Congo,
and New Guinea. Regions of simultaneously
lower vulnerability and irreplaceability, such as
Fig. 2. Maps of the nine global biodiversity conservation priorit y templates:
CE,crisisecoregions(21); BH, biodiversity hot spots [(11), updated by (39)];
EBA, endemic bird areas (15); CPD, centers of plant diversity (12); MC,
megadiversity countries (13);G200,global200ecoregions[(16), updated by
(54)]; HBWA, high-biodiversity wilderness areas (14); FF, frontier forests (19);
LW,lastofthewild(20).
REVIEW
www.sciencemag.org SCIENCE VOL 313 7 JULY 2006
59
the boreal forests of Canada and Russia, and the
deserts of the western United States and central
Asia, are prioritized less often.
Two general observations are apparent. First,
most land (79%) is highlighted by at least one of
the prioritization systems. Second, despite this, a
noticeable pattern emerges from the overlay of
different approaches. There is significant overlap
among templates that prioritize irreplaceable re-
gions (11–16), among those that prioritize highly
vulnerable regions (11, 21), and among those that
prioritize regions of low vulnerability (14, 19, 20),
but not between approaches
across each of these three gen-
eral classes (table S1). This pro-
vides useful cross-verification of
priority regions (31).
These patterns of overlap
reflect two approaches to how
vulnerability is incorporated into
conservation in the broadest
sense: reactive (prioritizing areas
of high threat and high irreplace-
ability) and proactive (prioritiz-
ing areas of low threat but high
irreplaceability). The former are
considered the most urgent pri-
orities in conservation planning
theory (10) because unless im-
mediate conservation action is
taken within them, unique bio-
diversity will soon be lost. The
latter are often de facto prior-
ities, because the opportunities
for conservation in these are
considerable (32). Biodiversity
conservation clearly needs both
approaches, but the implemen-
tation of each may correspond
to different methods. On the
one hand, large-scale conserva-
tion initiatives may be possible
in wilderness areas, such as the
establishment of enormous pro-
tected areas (one example is the
3,800,000-ha Tumucumaque
National Park, created in the
Brazilian state of Amapa
´
in
2003). On the other hand, finely
tuned conservation will be es-
sential in regions of simulta-
neously high irreplaceability
and threat, where losing even
tiny patches of remnant habitat, such as the sites
identified by the Alliance for Zero Extinction
(33), would be tragic.
Impact of Global Prioritization
The appropriate measure of impact is the success
of prioritization in achieving its main goal: in-
fluencing globally flexible donors to invest in
regions where these funds can contribute most to
conservation. Precise data are unavailable for all
of the approaches (34), but hot spots alone have
mobilized at least $750 million of funding for
conservation in these regions (35). More specif-
ically, conservation funding mechanisms have
been established for several of the approaches,
such as the $100 million, 10-year Global
Conservation Fund focused on high-biodiversity
wilderness areas and hot spots, and the $125
million, 5-year Critical Ecosystem Partnership
Fund, aimed exclusively at hot spots. The Global
Environment Facility, the largest financial mech-
anism addressing biodiversity conservation, is
currently exploring a resource allocation
framework that builds on existing templates.
Both civil society and government organizations
often use the recognition given to regions as
global conservation priorities as justification
when applying for geographically flexible
funding. In addition, the global prioritization
systems must have had sizeable effects in the
cancellation, relocation, or mitigation of envi-
ronmentally harmful activities, even in the
absence of specific legislation. Unfortunately,
resources still fall an order of magnitude short of
required conservation funding (4). Nevertheless,
the dollar amounts are impressive, and represent
marked increases in conservation investment in
these regions.
Challenges Facing Global Prioritization
Limitations of data have thus far generally
restricted global conservation prioritization to
specialist estimates of irreplaceability, to habitat
loss as a measure of vulnerability, and to coarse
geographic units defined a priori. Over the past 5
years, spatial data sets have been compiled with the
potential to reduce these constraints, particularly for
mammals, birds, and amphibians (5). When these
maps are co m b i ned with assess-
ment of conservation status,
they enable the development of
threat metrics directly based on
threatened species (36). So far,
the main advances to global
prioritization enabled by these
new data are validation tests of
existing templates (31). Encour-
agingly, global gap analysis of
priorities for the representation
of terrestrial vertebrate species
in protected areas (36) and
initial regional assessment of
plants (37) yield results similar
to existing approaches (fig. S2).
Invertebrates represent the
bulk of eukaryotic diversity on
Earth with more than a million
known species and many more
yet to be described (5). The con-
servation status of only È3500
arthropods has been assessed
(5), so global conservation prior-
ity is far from being able to incor-
porate megadiverse invertebrate
taxa (8, 23). Although some re-
gional data shows little overlap
between priority areas for arthro-
pods and those for plant and
terrestrial vertebrate taxa (38),
preliminary global data for
groups such as tiger beetles and
termites suggest much higher
levels of congruence (39). Simi-
larly, pioneering techniques to
model overall irreplaceability by
combining point data for mega-
diverse taxa with environmental
data sets produce results com-
mensurate with existing conser-
vation priorities (40). These findings, although
encouraging, in no way preclude the need to use
primary invertebrate data in global conservation
prioritization as they become available.
Aquatic systems feature poorly in existing con-
servation templates. Only one conservation prior-
itization explicitly incorporates aquatic systems
(16). The most comprehensive study yet, albeit
restricted to tropical coral reef ecosystems, iden-
tified 10 priority regions based on endemism and
threat (41). Eight of these regions lie adjacent to
priority regions highlighted in Fig. 3, raising the
Fig. 3. Mapping the overlay of approac hes prioritizing reactive and proactive
conservation. (A) Reactive approac hes, corresponding to the right-hand side of Fig.
1A, which prioritize regions of high threat, and those that do not incorporate
vulnerability as a criterion (Fig. 1B); the latter are only mapped where they overlap
with the former. (B) Proactive approaches, corresponding to the left-hand side of
Fig. 1A, which prioritize regions of low threat, and those that do not incorporate
vulnerability as a criterion (Fig. 1B); again, the latter are only mapped where
they overlap with the former. Shading denotes the number of global biodiversity
conservation prioritization templates that prioritize the shaded region, in both
(A) and (B).
REVIEW
7 JULY 2006 VOL 313 SCIENCE www.sciencemag.org
60
possibility of correspondence between marine and
terrestrial priorities, despite the expectation that
surrogacy of conservation priorities will be low
between different environments (42). Efforts to
identify freshwater priorities lag further behind,
although initial studies reveal a highly uneven
distribution of freshwater fish endemism (39).
Most measurement of irreplaceability is species
based, raising the concern that phylogenetic di-
versity may slip through the net of global con-
servation priorities (8, 23, 43). However, analyses
for mammals (44) find that priority regions repre-
sent higher taxa and phylogenetic diversity better
than would be predicted by the degree to which
they represent species. Islands such as Madagas-
car and the Caribbean hold especially high con-
centrations of endemic genera and families (39).
A heterodox perspective argues that the terminal
tips of phylogenetic trees should be higher pri-
oritiesthandeeplineages(45). In any case, the
balance of work implies that even if phylogenetic
diversity is not explicitly targeted for conservation,
global prioritization based on species provides a
solid surrogate for evolutionary history.
That global conservation priority regions
capture phylogenetic history does not necessar-
ily mean that they represent evolutionary pro-
cess (8). For example, transition zones or
‘‘biogeographic crossroads,’’ frequently over-
looked by conservation prioritization, could be
of particular importance in driving speciation
(46). On the other hand, there is evidence that
areas of greatest importance in generating bio-
diversity are those of long-term climatic sta-
bility, especially where they occur in tropical
mountains (47), which are incorporated in most
approaches to global conservation prioritiza-
tion. The development of metrics for the main-
tenance of evolutionary process is in its infancy
and represents an emerging research front.
A final dimension that will prove important to
assess in the context of global conservation priori-
tization concerns ecosystem services (43). Al-
though the processes threatening biodiversity and
ecosystem services are likely similar, the relation-
ship between biodiversity per se and ecosystem
services remains unresolved (48). Thus, while it is
important to establish distinct goals for these con-
servation objectives (49), identification of syner-
gies between them is strategically vital. This
research avenue has barely been explored, and
questions of how global biodiversity conservation
priorities overlap with priority regions for carbon
sequestration, climate stabilization, maintenance of
water quality, minimization of outbreaks of pests
and diseases, and fisheries, for example, remain un-
answered. However, the correspondence between
conservation priorities and human populations
(25, 26) and poverty (4, 5) is an indication that the
conservation of areas of high biodiversity priority
will deliver high local ecosystem service benefits.
From Global to Local Priorities
The establishment of global conservation prior-
ities has been extremely influential in directing
resources toward broad regions. However, a
number of authors have pointed out that global
conservation prioritization has had little success
in informing actual conservation implementa-
tion (8, 23). Separate processes are necessary to
identify actual conservation targets and priorities
at much finer scales, because even within a region
as uniformly important as, for example, Mada-
gascar, biodiversity and threats are not evenly
distributed. Bottom-up processes of identification
of priorities are therefore essential to ensure the
implementation of area-based conservation (50).
Indeed, numerous efforts are underway to
identify targets for conservation implementation.
Many focus on the site scale, drawing on two
decades of work across nearly 170 countries in
the designation of important bird areas (51).
There is an obvious need to expand such work to
incorporate other taxa (52) and to prioritize the
most threatened and irreplaceable sites (33).
Such initiatives have recently gained strong
political support under the Convention on
Biological Diversity, through the development
of the Global Strategy for Plant Conservation
and the Programme of Work on Protected Areas.
Both mechanisms call for the identification,
recognition, and safeguarding of sites of bio-
diversity conservation importance. Meanwhile,
considerable attention is also targeted at the scale
of landscapes and seascapes to ensure not just
the representation of biodiversity but also of the
connectivity, spatial structure, and processes that
allow its persistence (53).
Global conservation planning is key for
strategic allocation of flexible resources. Despite
divergence in methods between the diffe rent
schemes, an overall picture is emerging in which
a few regions, particularly in the tropics and in
Mediterranean-type environments, are consist-
ently emphasized as priorities for biodiversity
conservation. It is crucial that the global donor
community channel sufficient resources to these
regions, at the very minimum. This focus will
continue to improve if the rigor and breadth of
biodiversity and threat data continue to be
consolidated, which is especially important given
the increased accountability demanded from
global donors. However, it is through the con-
servation of actual sites that biodiversity will
ultimately be preserved or lost, and thus drawing
the lessons of global conservation prioritization
down to a much finer scale is now the primary
concern for conservation planning.
References and Notes
1. S. L. Pimm, G. J. Russell, J. L. Gittleman, T. M. Brooks,
Science 269, 347 (1995).
2. Millennium Ecosystem Assessment, Ecosystems and Human
Well-Being (Island Press, Washington, DC, 2005).
3. M. L. Hunter Jr., A. Hutchinson, Conserv. Biol. 8, 1163
(1994).
4. A. James, K. J. Gaston, A. Balmford, Nature 401, 323
(1999).
5. J. E. M. Baillie et al., Global Species Assessment (IUCN,
Gland, Switzerland, 2004).
6. A. Balmford et al., Science 307, 212 (2005).
7. N. Myers, Environmentalist 8, 187 (1988).
8. G. M. Mace et al., Nature 405, 393 (2000).
9. K. H. Redford et al., Conserv. Biol. 17, 116 (2003).
10. C. R. Margules, R. L. Pressey, Nature 405, 243 (2000).
11. N. Myers et al., Nature 403, 853 (2000).
12. WWF, IUCN, Centres of Plant Diversity (WWF and IUCN,
Gland, Switzerland, 1994–1997).
13. R. A. Mittermeier, P. Robles Gil, C. G. Mittermeier,
Megadiversity (CEMEX, Mexico City, Mexico, 1997).
14. R. A. Mittermeier et al., Proc. Natl. Acad. Sci. U.S.A. 100,
10309 (2003).
15. A. J. Stattersfield, M. J. Crosby, A. J. Long, D. C. Wege,
Endemic Bird Areas of the World (BirdLife International,
Cambridge, UK, 1998).
16. D. M. Olson, E. Dinerstein, Conserv. Biol. 12,502
(1998).
17. C. D. L. Orme et al., Nature 436, 1016 (2005).
18. J. F. Lamoreux et al., Nature 440, 212 (2006).
19. D. Bryant, D. Nielsen, L. Tangley, Last Frontier Forests
(World Resources Institute, Washington, DC, 1997).
20. E. W. Sanderson et al., Bioscience 52, 891 (2002).
21. J. M. Hoekstra et al., Ecol. Lett. 8, 23 (2005).
22. G. A. Krupnick, W. J. Kress, Biodivers. Conserv. 12, 2237
(2003).
23. N. Brummitt, E. N. Lughadha, Conserv. Biol. 17, 1442
(2003).
24. K. Wilson et al., Environ. Manage. 35, 527 (2005).
25. R. P. Cincotta, J. Wisnewski, R. Engelman, Nature 404,
990 (2000).
26. A. Balmford et al., Science 291, 2616 (2001).
27. C. O’Connor, M. Marvier, P. Kareiva, Ecol. Lett. 6, 706
(2003).
28. C. Parmesan, G. Yohe, Nature 421, 37 (2003).
29. K. A. Wilson et al., Nature 440, 337 (2006).
30. P. Jepson, R. J. Whittaker, Conserv. Biol. 16, 42 (2002).
31. G. A. B. da Fonseca et al., Nature 405, 393 (2000).
32. M. Cardillo, G. M. Mace, J. L. Gittleman, A. Purvis, Proc.
Natl. Acad. Sci. U.S.A. 103, 4157 (2006).
33. T. H. Ricketts et al., Proc. Natl. Acad. Sci. U.S.A. 51,
18497 (2005).
34. B. S. Halpern et al., Conserv. Biol. 20, 56 (2006).
35. N. Myers, Bioscience 53, 916 (2003).
36. A. S. L. Rodrigues et al., Bioscience 54, 1092 (2004).
37. W. Ku
¨
per et al., Ann. Missouri Bot. Gard. 91,525
(2004).
38. A. P. Dobson, J. P. Rodriguez, W. M. Roberts, D. S. Wilcove,
Science 275, 550 (1997).
39. R. A. Mittermeier et al., Hotspots Revisited (CEMEX,
Mexico City, Mexico, 2004).
40. S. Ferrier et al., Bioscience 54, 1101 (2004).
41. C. M. Roberts et al., Science 295, 1280 (2002).
42. W. V. Reid, Trends Ecol. Evol. 13, 275 (1998).
43. P. Kareiva, M. Marvier, Am. Sci. 91, 344 (2003).
44. W. Sechrest et al., Proc. Natl. Acad. Sci. U.S.A. 99, 2067
(2002).
45. T. L. Erwin, Science 253, 750 (1991).
46. T. B. Smith, R. K. Wayne, D. J. Girman, M. W. Bruford,
Science 276, 1855 (1997).
47. J. Fjeldsa
˚
,J.C.Lovett,Biodivers. Conserv. 6, 325
(1997).
48. M. Loreau et al., Science 294, 804 (2001).
49. S. Sarkar, Bioscience 49, 405 (1999).
50. R. J. Whittaker et al., Divers. Distrib. 11, 3 (2005).
51. BirdLife International, State of the World’s Birds 2004
(BirdLife International, Cambridge, UK, 2004).
52. G. Eken et al., Bioscience 54, 1110 (2004).
53. R. M. Cowling, R. L. Pressey, M. Rouget, A. T. Lombard,
Biol. Conserv. 112, 191 (2003).
54. D. M. Olson, E. Dinerstein, Ann. Missouri Bot. Gard. 89,
199 (2002).
55. We thank G. Fabregas, D. Knox, T. Lacher, P. Langhammer,
N. Myers, and W. Turner for help with the manuscript,
and the Gordon and Betty Moore Foundation for
funding.
Supporting Online Material
www.sciencemag.org/cgi/content/full/313/5783/58/DC1
Figs. S1 and S2
Table S1
References and Notes
10.1126/science.1127609
REVIEW
www.sciencemag.org SCIENCE VOL 313 7 JULY 2006
61
![Fig. 2. Maps of the nine global biodiversity conservation priority templates: CE, crisis ecoregions (21); BH, biodiversity hot spots [(11), updated by (39)]; EBA, endemic bird areas (15); CPD, centers of plant diversity (12); MC, megadiversity countries (13); G200, global 200 ecoregions [(16), updated by (54)]; HBWA, high-biodiversity wilderness areas (14); FF, frontier forests (19); LW, last of the wild (20).](/figures/fig-2-maps-of-the-nine-global-biodiversity-conservation-3j7mxit8.webp)