Climate change threats to plant diversity in Europe

Abstract

Climate change has already triggered species distribution shifts in many parts of the world. Increasing impacts are expected for the future, yet few studies have aimed for a general understanding of the regional basis for species vulnerability. We projected late 21st century distributions for 1,350 European plants species under seven climate change scenarios. Application of the International Union for Conservation of Nature and Natural Resources Red List criteria to our projections shows that many European plant species could become severely threatened. More than half of the species we studied could be vulnerable or threatened by 2080. Expected species loss and turnover per pixel proved to be highly variable across scenarios (27-42% and 45-63% respectively, averaged over Europe) and across regions (2.5-86% and 17-86%, averaged over scenarios). Modeled species loss and turnover were found to depend strongly on the degree of change in just two climate variables describing temperature and moisture conditions. Despite the coarse scale of the analysis, species from mountains could be seen to be disproportionably sensitive to climate change (≈60% species loss). The boreal region was projected to lose few species, although gaining many others from immigration. The greatest changes are expected in the transition between the Mediterranean and Euro-Siberian regions. We found that risks of extinction for European plants may be large, even in moderate scenarios of climate change and despite inter-model variability.

• Intergovernmental Panel on Climate Change storylines
• species extinction
• species turnover
• niche-based model

Full text

Climate change threats to plant diversity in Europe
Wilfried Thuiller*
†‡§
, Sandra Lavorel*
¶
, Miguel B. Arau
´
jo*
‡储
, Martin T. Sykes**, and I. Colin Prentice
††
*Centre d’Ecologie Fonctionnelle et Evolutive, Centre National de la Recherche Scientifique-Unite´ Mixte de Recherche 5175, 1919 Route de Mende, 34293
Montpellier Cedex 5, France;
†
Climate Change Research Group, Kirstenbosch Research Center, National Botanical Institute, P兾Bag x7, Claremont 7735, Cape
Town, South Africa;
‡
Macroecology and Conservation Unit, University of E
´
vora, Estrada dos Leo˜ es, 7000-730 E
´
vora, Portugal;
¶
Laboratoire d’Ecologie Alpine,
Centre National de la Recherche Scientifique-Unite´ Mixte de Recherche 5553, Universite´ J. Fournier, B.P. 53X, 38041 Grenoble Cedex 9, France;
储
Biodiversity
Research Group, School of Geography and the Environment, Oxford University, Mansfield Road, Oxford OX1 3TB, United Kingdom; **Geobiosphere Science
Centre, Department of Physical Geography and Ecosystems Analysis, Lund University, So¨ lvegatan 12, 223 62 Lund, Sweden; and
††
QUEST, Department of
Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road, Bristol BS8 1RJ, United Kingdom
Edited by Harold A. Mooney, Stanford University, Stanford, CA, and approved April 26, 2005 (received for review December 31, 2004)
Climate change has already triggered species distribution shifts in
many parts of the world. Increasing impacts are expected for the
future, yet few studies have aimed for a general understanding of
the regional basis for species vulnerability. We projected late 21st
century distributions for 1,350 European plants species under
seven climate change scenarios. Application of the International
Union for Conservation of Nature and Natural Resources Red List
criteria to our projections shows that many European plant species
could become severely threatened. More than half of the species
we studied could be vulnerable or threatened by 2080. Expected
species loss and turnover per pixel proved to be highly variable
across scenarios (27–42% and 45–63% respectively, averaged over
Europe) and across regions (2.5– 86% and 17– 86%, averaged over
scenarios). Modeled species loss and turnover were found to
depend strongly on the degree of change in just two climate
variables describing temperature and moisture conditions. Despite
the coarse scale of the analysis, species from mountains could be
seen to be disproportionably sensitive to climate change (⬇60%
species loss). The boreal region was projected to lose few species,
although gaining many others from immigration. The greatest
changes are expected in the transition between the Mediterranean
and Euro-Siberian regions. We found that risks of extinction for
European plants may be large, even in moderate scenarios of
climate change and despite inter-model variability.
Intergovernmental Panel on Climate Change storylines 兩 species
extinction 兩 species turnover 兩 niche-based model
R
ecent rapid climate change is already af fecting a wide variety
of organisms (1, 2). Long-term data indicate that the anom-
alous climate of the past half-century is already affecting the
physiology, distribution, and phenology of some species in ways
that are consistent with theoretical predictions (3). Although
natural climate variation and nonclimatic factors such as land
transfor mation may well be responsible for some of these trends,
human-induced climate and atmospheric change are the most
parsimon ious explanation for many (3, 4).
Several studies have modeled future species distributions at
regional (5–8) and local scales (9, 10) and have extrapolated
alar ming extinction risks for the next century (11). However, few
studies have considered the consequences of multiple climate-
change scenarios (7, 8), which represent the outcome of different
assumptions about the future (12). Using four representative
scenarios and three dif ferent climate models (HadCM3,
CGCM2, and CSIRO2), and a range of niche-based modeling
techn iques implemented in
BIOMOD (13), we develop predictions
of the potential consequences for 1,350 plant species in Europe.
The ‘‘future climate’’ we contrast with today’s climate (averaged
f rom 1961 to 1990) is the projected mean for the period from
2051 to 2080.
The ‘‘bioclimatic envelope’’ describes the conditions under
which populations of a species persist in the presence of other
biot a as well as climatic constraints (6, 14). Future distributions
are projected on the assumption that current envelopes reflect
species’ environmental preferences, which will be ret ained under
climate change. This principle has strong support from studies
demonstrating the evolutionary conservatism of ec ological
n iches and the phylogenetic inertia of species across time scales
(15, 16) and comparative biogeographical studies (17, 18).
However, this approach also assumes inst antaneous species-
range change, it ignores physiological CO
2
responses, and it does
not capture details of population dynamics or biotic interactions
nor the lags in spatial range shifts associated with processes of
dispersal, establishment, and local extinction. To assess the
sensitivit y of projections to the most critical of these assump-
tions, we considered two c ontrasting assumptions about migra-
tion abilit y (7, 8, 11): either species are unable to disperse at all
on the time scale considered (no migration), or they have no
c onstraints to dispersal and establishment (universal migration).
The reality for most species is likely to fall between these
extremes, depending on their ability to migrate across frag-
mented landscapes (19). We calculated losses of climatically
suit able areas (‘‘species loss’’) assuming no migration and gains
(‘‘species gain’’) and turnover (‘‘species turnover’’) assuming
un iversal migration.
Methods
Data Sources. Species’ distribution data are available for 2,294
plants (20), comprising ⬇20% of the total European flora,
sampled between 1972 and 1996. Modeling was c onducted by
using available data for Europe on a 50 ⫻ 50 km grid. The
mapped area comprises western, northern and southern Europe,
but excludes most of the eastern European countries where
rec ording effort was both less uniform and less intensive (21).
Af ter removing species with ⬍20 records, we considered range
responses of 1,350 plant species of Europe. We assume this
sample can be taken as represent ative of the responses of
European plant species to climate change because it includes
most of the life forms and phytogeographic patterns found
among plant species in Europe.
Climate data were obtained from the Climatic Research Unit
(www.cr u.uea.ac.uk) and included mean annual, winter, and
summer precipitation, mean annual temperature and minimum
temperature of the coldest month (MTC), growing deg ree days
(⬎5°) and an index of moisture availability (22). These variables
were chosen because of their strong link with the physiology and
growth of plant species (23, 24). For instance, MTC discrimi-
nates species based on their ability to assimilate soil water and
nutrients, and continue cell div ision, differentiation and tissue
growth at low temperatures (lower limit), and chilling require-
ments for processes such as bud break and seed germination
(upper limit). The moisture index discriminates species through
This paper was submitted directly (Track II) to the PNAS office.
Freely available online through the PNAS open access option.
Abbreviation: IUCN, International Union for Conservation of Nature and Natural
Resources.
§
To whom correspondence should be addressed. E-mail: thuiller@sanbi.org.
© 2005 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409902102 PNAS
兩
June 7, 2005
兩
vol. 102
兩
no. 23
兩
8245–8250
ECOLOGY

processes related to phenology, rooting strateg y, leaf morphol-
ogy, and xylem vulnerability to cavit ation. However, because
there is surprisingly little experiment al work for any particular
species to guide the choice of bioclimatically limiting variables,
the variables are generic and represent a hypothetical minimum
basic set for n iche-based modeling. Climate dat a were averaged
for the 1961–1990 period. The data were supplied on a 10-foot
(1 f t ⫽ 0.3 m) grid covering Europe. They were aggregated by
averaging to 50 ⫻ 50 km Universal Transverse Mercator (UTM)
to match the species data grid. Niche-based models were cali-
brated on the 50 ⫻ 50 km UTM grid, and modeled species
distributions were projected back onto the 10⬘ grid for current
and future climate.
Future projections were derived by using climate model
outputs made available through the Intergovernmental Panel on
Climate Change (IPCC) Data Distribution Centre (ipcc-ddc.cru.
uea.ac.uk).
‡‡
The modeled climate anomalies were scaled based
on four scenarios proposed by the IPCC (12). The A1 scenario
describes a globalized world with rapid economic growth and
global population that peaks in mid-century and declines there-
af ter and assumes rapid introduction of new and more efficient
technologies. Concentrations of CO
2
increase from 380 ppm in
2000 to 800 ppm in 2080, and temperature rises by 3.6 K (12). The
A2 scenario describes a heterogeneous world with regionally
oriented economic development. Per capita ec onomic g rowth
and technological change are slower than in the other scenarios.
Global concentrations of CO
2
increase from 380 ppm in 2000 to
700 ppm in 2080, and temperature rises by 2.8 K. The B1 scenario
describes a convergent world with global population that peaks
in mid-century and declines thereafter, as in A1, but with a rapid
change toward a service and information economy and the
introduction of clean and resource-ef ficient technology. Con-
centrations of CO
2
increase from 380 ppm in 2000 to 520 ppm
in 2080, and temperature rises by 1.8 K. The B2 scenario
describes a world in which the emphasis is on local solutions to
socioec onomic and environment al sustainability. It is a world
with continuously increasing global population (at a rate lower
than A2), intermediate levels of economic development, and less
rapid and more diverse technological change than in the B1 and
A1 scenarios. Concentrations of CO
2
increase from 380 ppm in
2000 to 550 ppm in 2080, and temperature rises by 2.1 K (12).
We did not assess the impacts of land-use change, even though
this factor will potentially compound the effects of climate
change on species distributions (25). However, given the spatial
extent and resolution of our data and the magnitude of climate
change in most projections, the effect of land use would be most
likely overridden by climate (26, 27).
Niche-Based Models of Species Climatic Envelops. We used the
BIOMOD framework, which capitalizes on several widely used
n iche-based modeling techniques (generalized linear models,
generalized additive models, classification tree analysis, and
artificial neural net works) to provide alternative spatial projec-
tions (13). For each climate change scenario, models relating
species distributions to the seven bioclimatic variables were fitted
by using BIOMOD and projected into the future. Then, a consen-
sus principal component analysis was run to explore central
tendencies in projections and select the niche-based model
representing the greatest commonality among projections (8).
There is increasing evidence that model projections can be
extremely variable, and there remains a need to test the
ac curac y of models and to reduce uncertainties (8, 28, 29). One
recent analysis has however provided the first test of the
predictive accurac y of such models by using bird observed
species’ range shifts and climate change in two periods of the
recent past (30). This work provides validation of niche-based
models under climate change and demonstrated how uncer-
t aint y can be reduced by selecting the most consensual pro-
jections, as done in this study. We are therefore confident that
this strateg y prov ides a robust and defensible approach to
species range projections for the purposes of c onservation
plann ing and biodiversit y management.
To evaluate species extinction risks, we summed the number
of pixels lost, potentially gained (under universal mig ration), or
st able by each species for the different climate-change scenarios.
We assigned each species to an International Un ion for Con-
servation of Nature and Natural Resources (IUCN) threat
category (IUCN 2001). Those that were not listed were classified
as lower risk, depending on the projected reduction in range size
f rom present to 2080. Present and future range sizes (area of
oc cupancy) were estimated f rom the number of pixels where
species occurred. Loss in range size was calculated by subtracting
future potential range size from present potential range size. In
line with IUCN Red List criterion A3(c), the following thresh-
olds were then used to assign a species to a threat category
(IUCN 2001). Extinct is a species with a projected range loss of
100% in 50 or 80 years, critically endangered has a projected
range loss of ⬎80%, endangered has a projected range loss of
⬎50%, and vulnerable has a projected range loss of ⬎30%.
A lthough this Red Listing approach is simplistic and c onsiders
only the effects of climate change, it provides a synthetic
overview of species-specific threats due to climate change.
To evaluate the percentage of extinctions for a given area, we
summed the number of species lost (L) by pixel and related it to
current species richness by pixel. The procedure was the same to
assess the percentage of species gained (G) by pixel (under
assumptions that species could reach a new suitable climate
space). Percent age of species turnover by pixel, under the
assumption of universal migration, is then given by T ⫽ 100 ⫻
(L ⫹ G)兾(SR ⫹ G) where SR is the current species richness.
Results and Discussion
Many European species could be threatened by future climate
change (Fig. 1). Under the assumption of no migration, more
than half of the species we considered become vulnerable or
c ommitted to extinction by 2080. The impacts of climate change
are, naturally, less under the universal migration because of the
possibilit y for species to move across landscapes. Under the
no-migration assumption and the most severe climate change
scenario (A1-HadCM3), 22% of the species become critically
endangered (⬎80% range loss), and 2% extinct by 2080. These
numbers decrease for the other scenarios and climate models.
Under the universal migration assumption, the results are, as
ex pected, less severe. Under tA1-HadCM3, 67% of species
would be classified as low risk, whereas under B1-HadCM3, 76%
of the species would be at low risk.
Our results coincide with the direction of predictions made by
Thomas et al. (11), although the magnitude of the risks we
project is less [and note that we project distributions to 2080,
whereas Thomas et al. (11) only projected to 2050].
Niche-based modeling does not address the proximate causes
of species extinction. Nevertheless, any reduction in the potential
geographic range of a species is likely to lead to an increased risk
of local extinction (11). This conclusion is, in fact, the rationale
for building IUCN Red Lists (31). A decrease in range size
implies that smaller stochastic events af fect a larger proportion
of the species’ total population, especially in fragmented land-
scapes. If a species becomes restricted to a few sites, then local
cat astrophic events (such as droughts or disease outbreaks) or an
‡‡
Mitchell, T. D., Carter, T. R., Jones, P. D., Hulme, M. & New, M. (2004) A Comprehensive
Set of High-Resolution Grids of Monthly Climate for Europe and the Globe: The Observed
Record (1901–2000) and 16 Scenarios (2001–2100) (Tyndall Centre for Climate Change
Res., Norwich, U.K.), Working Paper 55.
8246
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409902102 Thuiller et al.

increase of land transformation by humans c ould easily cause the
extinction of that species (32).
Rates of species’ loss and turnover show great variation across
scenarios (Fig. 2). In A1-HadCM3, the mean European temper-
ature increases by up to 4.4 K, leading to a mean species loss of
42% and turnover of 63%. This scenario provides the widest
range of variability across Europe for both species loss (2.5–
86%) and turnover (22–90%). The percentage of species loss
c ould exceed 80% in some areas, such as northcentral Spain and
the Cevennes and Massif Central in France. B1-HadCM3 gives
the lowest expected mean percentage of species loss (27%),
reflecting the fact that this scenario has the lowest rate of
increase in CO
2
and temperature by 2080 (mean European
temperature increase of 2.7 K). Other scenarios show interme-
diate mean rates of species loss (⬇30%) and turnover (⬇50%).
The relationship between the modeled percent age of species
loss and the anomalies for the two most significantly correlated
bioclimatic variables, growing-degree days (representing ac cu-
mulated warmth) and a moisture availability index, was used to
unc over the potential causes of variations in predicted changes
in plant diversity across regions within and across scenarios (Fig.
3). The strong consistent linear relationship across scenarios
indicates that projected species loss from our models could be
estimated f rom these two predictors. The Spearman rank-
c orrelation values for the separate univariate relationships were
0.73 and 0.65, respectively. Multiple-linear regression by using
these two predictors explains 60% of the variance across sce-
narios. The temperature of the coldest month, although being an
import ant predictor of distributions for many species (6), did not
show a strong relationship with species loss overall and was
therefore not used in this analysis.
Regional deviations from the inferred relationship (positive
and negative residuals) can be interpreted as indications of
particularly high or low species vulnerability, because of ecolog-
Fig. 1. Proportion of species classified according to the IUCN Red List assessment under two extremes assumptions about species migration. EX, extinct; CR,
critically endangered; EN, endangered; VU, vulnerable; LR, lower risk.
Fig. 2. Estimated percentage of species loss and turnover. Upper extreme, upper quartile, median, lower quartile, and lower extreme are represented for each
box.
Thuiller et al. PNAS
兩
June 7, 2005
兩
vol. 102
兩
no. 23
兩
8247
ECOLOGY

ical and historical characteristics of the flora, and兾or specific
environment al conditions (Fig. 4). An excess of species loss (red
c olor) is shown for mount ain regions (mid-altitude Alps, mid-
altitude Pyrenees, central Spain, French Cevennes, Balkans,
Carpathians). Severe climatic conditions have occurred in moun-
t ains over evolutionary times, promoting highly specialized
species with strong adaptation to the limited opportunities for
growth and survival (33). The narrow habitat tolerances of the
mount ain flora, in conjunction with marginal habitats for many
species, are likely to promote higher rates of species loss for a
similar climate anomaly than in any other part of Europe (34).
By contrast, the southern Mediterranean and part of the Pan-
non ian regions have a negative residual for species loss (gray
c olor). Both regions are characterized by hot and dry summers
and are occupied by species that tolerate strong heat and
drought. Under the scenarios used here, these species are likely
to continue to be well adapted to future conditions.
We finally present mean percentages of species loss and
turnover by environmental zones (M. Metzger, unpublished
dat a) with the A1-HadCM3 scenario of maximum change to best
illustrate the spatial patterns (Fig. 5). The major spatial patterns
are similar over all scenarios. The northern Mediterranean
(52%), Lusitanian (60%) and Mediterranean mount ain (62%)
regions are the most sensitive regions; the Boreal (29%), north-
ern Alpine (25%), and Atlantic (31%) regions are consistently
less sensitive. Species turnover shows a somewhat different
pattern. The Boreal region could, in principle, gain many species
f rom further south, leading to a high species turnover (66%). The
Pannon ian region c ould also theoretically gain eastern Mediter-
ranean species and has a calculated turnover of 66%. Thus, these
regions stand to lose a substantial part of their plant species
diversit y, and (in time) to show a major change in floristic
c omposition. Projected species turnover peaks at the transition
bet ween the Mediterranean and continental regions (Fig. 5) with
extirpation of Euro-Siberian species and expansion for Medi-
terranean or Atlantic species. Southern Fennoscandia is also an
area of high potential turnover w ith the loss of boreal species and
gain of Euro-Siberian species.
These results cannot be taken as precise forecasts given the
uncert ainties in climate change scenarios, the coarse spatial
resolution of the analysis (35), and uncertainties in the mod-
eling techniques used (8, 29). The relatively coarse grid scale
of our study may hide potential refuges for species and
env ironmental heterogeneit y that c ould enhance species sur-
v ival, especially in mountain areas where our estimation of
risks of extinctions c ould be overestimated. On the other hand,
landscape f ragmentation c ould increase the vulnerability of
these refuges to fire or other disturbances, which in c ombina-
tion with the lack of propagule flow, c ould compromise the
survival of remnant populations. There are also major uncer-
t ainties due to lags associated w ith biotic processes. The
rec ogn ized time scales for assign ing species IUCN Red List
categories are not suited to evaluating the c onsequences of
slow-acting but persistent threats. We have substituted a time
scale of 80 years (instead of 20) for critically endangered,
endangered and vulnerable, respectively, over which to assess
Fig. 3. Relationships between the percentage of species loss and anomalies of moisture availability and growing-degree days.The colors correspond to different
climate change scenarios.
Fig. 4. Regional projections of the residuals from the multiple regression of
species loss against growing-degree days and moisture availability. Red colors
indicate an excess of species loss; gray colors indicate a deficit.
8248
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409902102 Thuiller et al.

declines. The extent of species losses may be overestimated,
because the plasticity of species and the survival of species in
favorable microhabitats is not considered. However, even if
the numbers are overestimated, patterns across regions may
st and (e.g., the rank ing of region in ter ms of vulnerabilit y to
loss). Species loss does not necessarily imply the immediate
loss of a species f rom a site, rather it may imply a potential lack
of reproductive success and recruitment that w ill tend to
extinction on a longer time scale (36). Mig ration rates are
likely to be species-specific, and resulting biotic interactions in
‘‘no-analogue’’ assemblages may alter species’ realized n iches.
L and use and associated habitat fragmentation are likely to
generally inhibit mig ration rates (19). Further, future species
distributions w ill likely be influenced by other environmental
factors than changing climate. The current atmospheric CO
2
c oncentration exceeds any experienced during the past 20 mil-
lion years (12). Plant physiological responses, including grow th
responses to increased atmospheric CO
2
and changes in water-
use efficienc y, are expected to ameliorate the response of some
plant functional types to climate change (37). On the other hand,
n itrogen deposition, the enhanced potential for invasion by
exotic species, or the promotion of more competitive native
species may change competitive interactions in plant commun i-
ties, yielding novel patterns of dominance and ecosystem
function (38).
Despite uncert ainties, our findings provide illustration of the
potential importance and the likely direction of climate change
ef fects. From a conservation perspective, a proportion of Eu-
ropean plant species c ould become vulnerable. The strong
positive relationship between projected species loss and changes
in bioclimatic variables implies that action to reduce greenhouse
gas emissions would also mitigate climate-change effects on
plant diversit y. However, even under the least severe scenario
c onsidered, the risks to biodiversity appear to be considerable.
Dif ferent regions are ex pected to respond differently to climate
change, with the greatest vulnerability in mountain regions and
the least in the southern Mediterranean and Pannonian regions.
Recent observations (39) and predictions (9) corroborate our
c onclusion regarding the climatic sensitivity of species in Euro-
pean mountain areas. We have also identified a broad transition
zone where the greatest species mixing is likely to occur. During
the Quaternary period, this region acted both as a crossing point
and a refuge zone for Boreo-alpine and Euro-Siberian species
(40). This transition zone will be a strategic region for plant-
species conservation in a changing climate.
M.B.A. thanks T. Lathi and R. Lampinen for providing a digital version
of Atlas Florae Europaeae. We thank M. Erhard (Potsdam-Institute for
Climate Impact Research, Potsdam, Germany) for repackaging and
aggregating original climate data into the Atlas Florae Europaeae g rid
and M. Metzger for his environmental classification of Europe. This
research was funded by the integrated projects of the European Com-
mission’s FP5 Advanced Terrestrial Ecosystem Analysis and Modeling
(EVK2-CT-2000-00075) and FP6 Assessing Large-Scale Env ironmental
Risks with Tested Methods (GOCE-CT-2003-506675).
1. Edwards, M. & Richardson, A. J. (2004) Nature 430, 881–884.
2. Parmesan, C. (1996) Nature 382, 765–766.
3. Hughes, L. (2000) Trends Ecol. Evol. 15, 56–61.
4. Parmesan, C. & Yohe, G. (2003) Nature 421, 37–42.
5. Erasmus, B. F. N., Van Jaarsweld, A. S., Chown, S. L., Kshatriya, M. & Wessels,
K. J. (2002) Glob. Change Biol. 8, 679–693.
6. Huntley, B., Berry, P. M., Cramer, W. & McDonald, A. P. (1995) J. Biogeogr.
22, 967–1001.
7. Peterson, A. T., Ortega-Huerta, M. A., Bartley, J., Sanchez-Cordero, V., Soberon,
J., Buddemeier, R. H. & Stockwell, D. R. B. (2002) Nature 416, 626–629.
8. Thuiller, W. (2004) Glob. Change Biol. 10, 2020–2027.
9. Gottfried, M., Pauli, H., Reiter, K. & Grabherr, G. (1995) Divers. Distrib. 5,
241–251.
10. Guisan, A. & Theurillat, J.-P. (2000) Integr. Assess. 1, 307–320.
11. Thomas, C. D., Cameron, A., Green, R. E., Bakkenes, M., Beaumont, L. J.,
Collingham, Y. C., Erasmus, B. F. N., Ferreira de Siquiera, M., Grainger, A.,
Hannah, L., et al. (2004) Nature 427, 145–148.
12. Intergovernmental Panel on Climate Change (2001) Climate Change 2001: Im-
pacts, Adaptation, and Vulnerability: Contribution of Working Group II to the Third
Asssessment Report of the IPCC (Cambridge Univ. Press, Cambridge, U.K.).
13. Thuiller, W. (2003) Glob. Change Biol. 9, 1353–1362.
14. Sykes, M. T., Prentice, I. C. & Cramer, W. (1996) J. Biogeogr. 23, 203–233.
15. Martinez-Meyer, E., Peterson, A. T. & Hargrove, E. C. (2004) Glob. Ecol.
Biogeogr. 13, 305–314.
16. Peterson, A. T., Soberon, J. & Sanchez-Cordero, V. (1999) Science 285,
1265–1267.
17. Beerling, D. J., Huntley, B. & Bailey, J. P. (1995) J. Veg. Sci. 6, 269–282.
18. Huntley, B., Bartlein, P. J. & Prentice, I. C. (1989) J. Biogeogr. 16,
551–560.
19. Higgins, S. I., Lavorel, S. & Revilla, E. (2003) Oikos 101, 354–366.
20. Jalas, J. & Suominen, J. (1972–1996) Altas Florae Europaeae (Comm. Mapping
Flora Europe Societas Biologica Fennica, Vanamo, Helsinki).
21. Williams, P. H., Humphries, C., Arau´jo, M. B., Lampinen, R., Hagemeijer, W.,
Gasc, J.-P. & Mitchell-Jones, T. (2000) Belg. J. Entomol. 2, 21–46.
Fig. 5. Spatial sensitivity of plant diversity in Europe ranked by biogeographic regions. Mean percentage of current species richness (Left) and species loss
(Center) and turnover (Right) by environmental zones under the A1-HadCM3 scenario.
Thuiller et al. PNAS
兩
June 7, 2005
兩
vol. 102
兩
no. 23
兩
8249
ECOLOGY

22. Gerten, D., Schaphof f, S., Haberlandt, U., Lucht, W. & Sitch, S. (2004) J.
Hydrol. (Amer sterdam) 286, 249–270.
23. Bartlein, P. J., Prentice, I. C. & Webb, T. (1986) J. Biogeogr. 13, 35–57.
24. Prentice, I. C., Cramer, W., Harrison, S. P., Leemans, R., Monserud, R. A. &
Solomon, A. M. (1992) J. Biogeogr. 19, 117–134.
25. Travis, J. M. J. (2003) Proc. R. Soc. London Ser. B 270, 467–473.
26. Pearson, R. G., Dawson, T. P. & Liu, C. (2004) Ecography 27, 285–298.
27. Thuiller, W., Arau´jo, M. B. & Lavorel, S. (2004) J. Biogeogr. 31, 353–361.
28. Arau´jo, M. B., Pearson, R. G., Thuiller, W. & Erhard, M. (2005) Glob. Change
Biol., in press.
29. Thuiller, W., Arau´jo, M. B., Pearson, R. G., Whitt aker, R. J., Brotons, L. &
Lavorel, S. (2004) Nature 430, 34.
30. Arau´jo, M. B., Whittaker, R. J., Ladle, L. & Erhard, M. (2005) Glob. Ecol.
Biogeogr., in press.
31. International Union for Conservation of Nature and Natural Resources (2001)
IUCN Red List Categor ies and Criteria (Int. Union Conserv. Nat. Nat. Resour.
Species Survival Commission, Cambridge, U.K.) Version 3.1.
32. Lawton, J. H. & May, R. M. (1995) Extinction Rates (Oxford Univ. Press,
Oxford).
33. Ko¨rner, C. (1999) Alpine Plant Life (Springer, Berlin).
34. Thuiller, W., L avorel, S. & Arau´jo, M. B. (2005) Glob. Ecol. Biogeog r.,in
press.
35. Weaver, A. J. & Zw iers, F. W. (2000) Nature 407, 571–572.
36. Chuine, I., Cambon, G. & Comtois, P. (2000) Glob. Change Biol . 6,
1–11.
37. Sitch, S., Smith, B., Prentice, I. C., Arneth, A., Bondeau, A., Cramer, W.,
Kaplan, J. O., Levis, S., Lucht, W., Sykes, M. T., et al. (2003) Glob. Change Biol.
9, 161–185.
38. Craine, J. M., Reich, P. B., Tilman, G. D., Ellsworth, D., Fargione, J., Knops,
J. & Naeem, S. (2003) Ecol. Letters 6, 623–630.
39. Walther, G.-R., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T. J. C.,
Fromentin, J.-M., Hoegh-Gudberg, O. & Bairlein, F. (2002) Nature 416,
389–395.
40. Taberlet, P. & Cheddadi, R. (2002) Science 297, 2009–2010.
8250
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0409902102 Thuiller et al.