Content uploaded by Brian S Maitner
Author content
All content in this area was uploaded by Brian S Maitner on Jun 22, 2024
Content may be subject to copyright.
www.ecography.org
ECOGRAPHY
Ecography
943
––––––––––––––––––––––––––––––––––––––––
© 2020 e Authors. Ecography published by John Wiley & Sons Ltd on behalf of Nordic Society Oikos
is is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Subject Editor and
Editor-in-Chief: Miguel Araújo
Accepted 21 February 2020
43: 943–953, 2020
doi: 10.1111/ecog.05166
43 943–953
Limiting climate change to less than 2°C is the focus of international policy under
the climate convention (UNFCCC), and is essential to preventing extinctions, a
focus of the Convention on Biological Diversity (CBD). e post-2020 biodiversity
framework drafted by the CBD proposes conserving 30% of both land and oceans
by 2030. However, the combined impact on extinction risk of species from limiting
climate change and increasing the extent of protected and conserved areas has not been
assessed. Here we create conservation spatial plans to minimize extinction risk in the
tropics using data on 289 219 species and modeling two future greenhouse gas con-
centration pathways (RCP2.6 and 8.5) while varying the extent of terrestrial protected
land and conserved areas from <17% to 50%. We nd that limiting climate change to
2°C and conserving 30% of terrestrial area could more than halve aggregate extinction
risk compared with uncontrolled climate change and no increase in conserved area.
Keywords: area-based conservation, biodiversity, climate change, conservation
planning, extinction risk
30% land conservation and climate action reduces tropical
extinction risk by more than 50%
LeeHannah, Patrick R.Roehrdanz, Pablo A.Marquet, Brian J.Enquist, GuyMidgley,
WendyFoden, Jon C.Lovett, Richard T.Corlett, DerekCorcoran, Stuart H. M. Butchart,
BradBoyle, XiaoFeng, BrianMaitner, JavierFajardo, Brian J.McGill, CoryMerow,
NaiaMorueta-Holme, Erica A.Newman, Daniel S.Park, NielsRaes and Jens-ChristianSvenning
L. Hannah and P. R. Roehrdanz (https://orcid.org/0000-0003-4047-5011) ✉ (proehrdanz@conservation.org), e Moore Center for Science, Conservation
International, 2011 Crystal Dr., Arlington, VA 22202, USA. – P. A. Marquet, D. Corcoran and J. Fajardo (https://orcid.org/0000-0002-0990-9718),
Dept de Ecología, Facultad de Ciencias Biológicas, Ponticia Univ. Católica de Chile, Santiago, Chile. PAM, DC and JF also at: Instituto de Ecología y
Biodiversidad (IEB), Santiago, Chile. – PAM and B. J. Enquist, e Santa Fe Institute, USA, Santa Fe, NM, USA. – BJE, B. Boyle, X. Feng, B. Maitner
and E. A. Newman (https://orcid.org/0000-0001-6433-8594), Dept of Ecology and Evolutionary Biology, Univ. of Arizona, Tucson, AZ, USA. – G. Midgley,
Dept of Botany and Zoology, Stellenbosch Univ., Stellenbosch, South Africa. – W. Foden, Cape Research Centre, South African National Parks, Cape Town,
South Africa. – J. C. Lovett, School of Geography, e Univ. of Leeds, Leeds, UK, and: Royal Botanic Gardens, Kew, Richmond, Surrey, UK. – R. T. Corlett,
Centre for Integrative Conservation, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun, Yunnan, China. – S. H.
M. Butchart, BirdLife International, David Attenborough Building, Pembroke Street, Cambridge, UK, and: Dept of Zoology, Univ. of Cambridge,
Cambridge, UK. – B. J. McGill, School of Biology and Ecology, and Senator George J. Mitchell Center of Sustainability Solutions, Univ. of Maine, Orono,
ME, USA. – C. Merow, Dept of Ecology and Evolutionary Biology, Univ. of Connecticut, CT, USA. – N. Morueta-Holme (https://orcid.org/0000-0002-
0776-4092), Center for Macroecology, Evolution and Climate; GLOBE Institute; Univ. of Copenhagen, Copenhagen, Denmark. – D. S. Park (https://orcid.
org/0000-0003-2783-530X), Dept of Organismic and Evolutionary Biology, Harvard Univ., MA, USA. – N. Raes (https://orcid.org/0000-0002-4329-
4892), Naturalis Biodiversity Center, Leiden, the Netherlands. – J.-C. Svenning (https://orcid.org/0000-0002-3415-0862), Center for Biodiversity
Dynamics in a Changing World (BIOCHANGE), Dept of Biology, Aarhus Univ., Aarhus, Denmark.
Research
944
Introduction
Preventing human-driven extinctions of the species sharing the
planet with us is among the greatest environmental challenges
of our time (Ehrlich and Mooney 1983, Rockströmet al.
2009). e Convention on Biological Diversity (CBD) has
been agreed by 195 nations to require ‘…maintenance and
recovery of viable populations of species’ for the conserva-
tion of biodiversity (Convention on Biological Diversity
1992). In the rst draft of the post-2020 biodiversity frame-
work developed by the CBD, extinction risk reduction is a
goal for 2030 to achieve the longer-term 2050 vision of a
world of ‘living in harmony with nature’ (CBD Secretariat
2020). Global extinction rate is a benchmark for planetary
boundaries, dening both the sixth mass extinction event
and a possible boundary for the Anthropocene geological
epoch (Rockström et al. 2009, Dirzo et al. 2014,
Ceballosetal. 2015).
Conservation of natural areas and sustainable use of
nature are recognized as important instruments with which
extinctions and other loss of biodiversity may be avoided
(Convention on Biological Diversity 1992). A key target of
the draft post-2020 biodiversity framework is to conserve
sites of importance for biodiversity through protected and
conserved areas covering at least 30% of land and ocean
(CBD Secretariat 2020). On land, both conservation and
sustainable use are currently challenged by increasing large-
scale monoculture and industrial developments, (Austinetal.
2017) making some form of conservation or land-use man-
agement essential to realizing these aims in most regions.
Without such conservation, loss of species’ habitats will result
in increased extinction risk.
In addition to the local-to-global benet of reducing
extinctions and maintaining biodiversity, maintaining natural
systems provides a wealth of benets at all scales. A well-man-
aged system of conserved areas can provide vital ecosystem
services, such as water purication and retention, erosion
control, the reduction of ooding, maintaining river base
ows and opportunities for ecotourism as well as minimiz-
ing extinctions (Dudley and Stolton 2003, Mulongoy and
Gidda 2008). For example, protected areas containing forests
provide an important supply of drinking water to more than
a third of the world’s 100 largest cities (Dudley and Stolton
2003). e conservation and eective management of natu-
ral areas is currently recognized as an important nature-based
solution for climate change mitigation (Griscometal. 2017,
Dudleyetal. 2018, Marquetetal. 2019); these areas account
for approximately 20% of the CO2 sequestered by all terres-
trial ecosystems (Melilloetal. 2016).
Climate change is now altering ecological conditions
across the planet, both on land and in the oceans, and spe-
cies are shifting their distributions in response to these new
conditions (Bellardetal. 2012, Urban 2015, IPBES 2019).
Species’ distributions are determined by a combination of
climatic and other niche tolerances (including edaphic),
geographical barriers and competition, so shifting climatic
conditions result in changes in the areas occupied by species,
altering representation in conserved areas and changing or
reducing ecosystem benets (Fodenetal. 2007, Lenoiretal.
2008, Pecletal. 2017). Climate change is causing some spe-
cies to shift their distributions out of existing conserved areas
(Araújoetal. 2004, Heller and Zavaleta 2009, Johnstonetal.
2013, Urban 2015), while also enabling some species to colo-
nize new areas (Chenetal. 2011, Angelo and Daehler 2013).
Species’ climate niches are unique, meaning that species move
individualistically in response to climate change. Species
will be moving at dierent rates and in dierent directions,
resulting in altered combinations of species in any one
location over time.
e interaction between climate change and com-
plex landscapes makes knowing where to conserve species
more complicated. Conservation plans that minimize area
requirements based on species’ current locations and habi-
tat requirements will no longer be eective as those locations
and habitats change. Establishing protected and conserved
areas in the right places to facilitate species’ distributional
shifts through time can help avoid extinctions due to cli-
mate change (Williams et al. 2005, Hannah et al. 2007,
Phillipsetal. 2008, Alagadoretal. 2016, Bagchietal. 2018).
Until now it is unknown how much extinction risk reduction
is possible under dierent extents of conserved area consider-
ing the eects of both land use change and climate change
(though see Baillie and Zhang 2018).
To address these questions, we created conservation spa-
tial plans for 289 219 tropical plant and vertebrate species
that minimized both present and modeled future extinction
risk under two greenhouse gas concentration pathways –
RCP2.6 and 8.5 – across varying area of terrestrial conser-
vation, from current levels up to 50% conserved, including
30% conserved as suggested in the draft post-2020 frame-
work (CBD 2020). We used widely available vertebrate dis-
tribution and occurrence data as well as the most up-to-date
compilation of terrestrial plant data (Enquist et al. 2019).
Extinction risk results are expressed as the mean, maximum
and minimum of our 10-climate model ensemble. Extinction
risk can be assessed in multiple ways. e IUCN Red List
classies species into dierent categories of relative extinc-
tion risk, while multiple modeling approaches exist for esti-
mating probability of extinction including the species area
relationship (SAR). e Zonation conservation planning
software (Moilanenetal. 2014) incorporates SAR theory to
produce estimates of extinction risk for every species. We sum
these values across species to calculate an aggregate extinc-
tion risk (AER), ranging from 0 if no species are at risk to 1
if all species have gone extinct). We thensummarize this for
three major tropical biogeographic regions – the Neotropics,
Afrotropics, and Southeast Asian tropics for both climate sce-
narios and conservation from present to 50% of terrestrial
area. is allows us to assess, for the rst time, which natural
areas are required to minimize extinction risk of tropical spe-
cies in the face of climate change and how much extinction
risk we may avoid by conserving additional area.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
945
Material and methods
Overview
We modelled present and future geographic ranges of
94 337 plant species and 9722 vertebrate species across the
Neotropics, Afrotropics and Southeast Asian tropics. Our
methods are described in detail below, under the following:
1) climate models and RCPs – we used 10 GCM selected to
be consistent across the three regions and two RCP (2.6 and
8.5) to provide low and high brackets of future climate; 2)
spatial planning algorithm – we used Zonation conservation
planning software with settings customized to accommodate
range shifts under climate change; 3) extinction risk calcula-
tion – we used an AER index equal to the sum of individual
species’ extinction risks, based on SAR; 4) species distribu-
tion models – we used Maxent (Phillipsetal. 2006) to t
species distribution models at 30 arc-second (~1 km) resolu-
tion then project distributions into baseline climate all future
climate scenarios at 2.5 arc-minutes (~5 km) for analysis; 5)
environmental predictor variables – were taken from stan-
dard WorldClim variables plus a custom aridity index; 6) soils
data – global soilgrids variables; 7) land use and land cover
data – from global consensus landcover dataset; 8) plant data
– we used botanical occurrence data assembled and standard-
ized within the BIEN database; 9) vertebrate data – we used
publicly available occurrence records veried with IUCN
expert range polygons. Details are given below for each of
these methods.
Climate models and RCPs
We used 10 GCMs and two climate scenarios for the 2060–
2080 timeframe. e GCMs used to drive our biological
models include Access 1.0, BCC-CSM1.1, CNRM-cm5,
GFDL-cm3, MOHC-HADGEM2-es, NCAR-CCSM4,
LASG-FGOALS-g2, NCC-NORESM-m, MIROC-ESM
and MPI-ESM-lr. We used a low forcing scenario, RCP2.6,
to approximate climate change consistent with meeting the
Paris Agreement target of 2.0°C global mean temperature
change, and RCP8.5 to approximate business-as-usual, no
action on climate change.
Spatial planning algorithm
We used the Zonation Conservation Planning Software
(‘Zonation’), a tool that allows for simultaneous prioritization
of many thousands of conservation features (Moilanenetal.
2005, Moilanen 2007). We applied Zonation at a horizontal
spatial resolution of 2.5 arc-minutes (~5 km). Zonation can
be used for conservation prioritization under climate change.
is conguration of Zonation allows for simultaneous pri-
oritization of a species current range, its modelled future
range, and the connectivity between the two limited by a spe-
cies’ capacity to disperse. e protocol for running Zonation
under climate change projections is described in Kujalaetal.
(2013). Species current ranges were linked to projected future
ranges through an interaction layer (Kujalaetal. 2013). is
layer is transformed by a dispersal kernel with a parameter
to limit the interaction to the species total capacity to dis-
perse over the period of analysis. Total dispersal capacity was
assumed to be 100 km for vertebrates (roughly 1 km year−1)
and 10 km for vascular plants (roughly 0.1 km year−1) follow-
ing previously published studies (Warrenetal. 2013). Species
with too few occurrence records to produce a model were
included as point locations (Zonation term = ‘species of spe-
cial interest’ or SSI) where the range was dened as each grid
cell containing ≥1 occurrence record. Equal weighting was
used for all conservation features.
Extinction risk calculation
Zonation/climate seeks to maximize biodiversity represen-
tation – both present and future – and thereby minimize
extinction risk, at each area increment, based on a power-
law species area relationship (SAR) parameter that can be set
for each conservation feature. Here we used the default SAR
parameter (z = 0.25) for all areas in order to transform of the
proportion of species ranges conserved at each step of the
solution to the per species extinction risk where:
Extinctionrisk1 Proportionconserved 0.25=-
()
^
To summarize over all species at dierent time steps, we cal-
culated aggregate extinction risk (AER) at each step as the
sum of individual species’ remaining extinction risk across all
species normalized by the total number of species.
Species distribution models
Species distribution models were produced with Maxent
(Phillips et al. 2006) for species with >10 unique occur-
rence records (i.e. unique 1 km grid cells in the modelling
domain) (van Proosdijetal. 2016). Maxent settings followed
the recommendations of Merowetal. (2013, 2014) to pro-
duce relatively less complex models (e.g. limiting features to
linear, quadratic, and product functions) to minimize overt-
ting. Modelling domains were limited to a spatial buer of
within 500 km of any valid occurrence record. is likewise
limited the projected ranges to within 500 km of any veried
observation. Background sampling was a random sample of
10 000 points within the buered modelling domain. Five
model replicates were used in tting the model and an aver-
age of the ve replicates was used for the nal species model.
Parameters from the nal model were used to project spe-
cies suitability for both baseline and future climate scenarios.
Default values for regularization coecients were used and
30% of occurrence records were randomly reserved to assess
model performance.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
946
Environmental predictor variables
We chose the following bioclimatic variables from downscaled
20-year normals (Hijmansetal. 2005) baseline (1960–1990)
and future (2060–2080) climates based on pan-tropical cor-
relation analysis: mean annual temperature (BIO1), mean
diurnal temperature range (BIO2), seasonality of tempera-
ture (BIO4), minimum temperature of the coldest month
(BIO6), mean annual precipitation (BIO12), seasonality of
precipitation (BIO15). We also used an accumulated aridity
index that is the sum of the monthly aridity (annual precipi-
tation – PET) for the maximum run of consecutive months
where (PET > precipitation). Accumulated aridity index is
derived from global monthly extra-terrestrial solar radiation
data from Trabucco and Zomer (2019) and monthly maxi-
mum temperature, minimum temperature and precipitation
from Worldclim ver. 1.4 (Hijmansetal. 2005).
Soils data
Soils variables used in species distribution models were depth
to bedrock, pH, clay proportion, silt proportion and bulk
density. All soil-related variables were obtained from Soilgrids
ver. 1.0 (accessed February 2018) (Hengl et al. 2017).
Variables with multiple strata available are the mean of the
top 1 m (strata 1–4). Soils variables were included as it has
been shown that climate change analyses that do not incor-
porate soils variability can misrepresent edaphic specialists
(Corlett and Tomlinson 2020).
Land cover and land use data
Areas of existing built up land or intensive agriculture were
removed from the analysis and therefore those cells are not
part of the prioritization solution. Built up and agricultural
areas were dened as >50% of pixel coverage for ‘urban’ and
‘agriculture’ classes from the 1 km resolution global consen-
sus land cover dataset produced by Tuanmu and Jetz (2014)
that were aggregated to match the 2.5 arc-minute resolution
of analysis. Existing protected areas (IUCN and WCMC
2018) were solved rst using a hierarchical mask so that
the conservation priorities take the eects existing protec-
tion into account. e analysis domain was limited to the
Afrotropics, Neotropics and Indo-Malayan biogeographic
realms with the Indonesian plus Papua New Guinea portion
of the Australasian realm also included in the Asia domain
(Dinersteinetal. 2017).
Plant data
Vascular plant data was extracted from the BIEN ver. 4.1
database using the RBIEN package (Maitner et al. 2018).
Occurrence records with geographic information were
obtained for 275 372 species of which we modelled 94
337. e BIEN data mainly comprise herbarium collec-
tions, ecological plots and surveys (DeWalt et al. 1999,
Wiseretal. 2001, Enquist & Boyle 2012, Enquistetal. 2016,
Fegraus 2012, Peetetal. 2012, Forest Inventory and Analysis
National Program 2013, Anderson-Teixeiraetal. 2015). For
details of specimen data sources see Maitneret al. (2018).
A full listing of the herbaria data used are given in the
Acknowledgements section. e observations in the BIEN
database are the product of contributions by 1076 dierent
data contributors, including numerous individual herbaria,
and data indexers of herbarium or plot data. Of the her-
baria, 550+ are listed in Index Herbariorum. Additionally,
BIEN 4.1 includes data from RAINBIO, TEAM, e Royal
Botanical Garden of Sydney, Australia, and NeoTropTree.
Plot data within BIEN are from the CVS, NVS, SALVIAS,
VEGBANK, CTFS, FIA, MADIDI, and TEAM data net-
works and datasets (< http://bien.nceas.ucsb.edu/bien/
data-contributors/all/ >).
Vertebrate data
Point occurrences for tropical vertebrate species were compiled
from GBIF (GBIF download 2019; accessed through R pack-
age ‘rgbif’) and restricted range bird species records (Birdlife
International 2018) were combined for a total of 13 847 trop-
ical species, 9722 of which had sucient occurrence records
(>10) for modelling (van Proosdijetal. 2016). Occurrences
were ltered to include only those with specic georeferenc-
ing coordinates, observations more recent than 1950, and
human observations only (no fossil records or museum speci-
mens). Additionally, spatial outliers (more than 500 km from
IUCN range polygon or >98th percentile of latitude + longi-
tude) were removed prior to modelling. IUCN and Birdlife
International range map polygons for all available terrestrial
mammals, birds, reptiles were used to generate the species
list and as a means of occurrence record validation (Birdlife
International 2018, IUCN and UNEP WCMC 2018).
Results
Our results show that a large reduction in extinction risk can
be achieved by moving to a world in which 30% or 50% of
terrestrial area is conserved, corresponding to the proposed
action targets of the rst draft post-2020 CBD framework
(CBD 2020). Figure 1 illustrates the modeled reduction in
aggregate extinction risk with increasing area conserved, for
the Neotropics, Afrotropics and SE Asian tropics respectively.
e reduction in extinction risk of up to 82% (ensemble
mean) results from both reduced loss of habitat due to land
use change and better representation of changing habitats
for species moving in response to climate change. Outside of
fully natural landscapes, achieving these minimum extinction
results requires matching land uses to the habitat needs of
species at risk of extinction in a particular area. For instance,
shade coee may provide habitat for many bird species, but
not for understory tree species. Actual extinction risk may
be somewhat higher because multiple-use or sustainable use
landscapes may not provide suitable habitats for all species.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
947
e spatial distribution of natural areas needed to achieve
these reductions in extinction risk is illustrated for each of
the three major tropical regions in our study in Fig. 2A–C.
Many areas highlighted are recognized as global biodiver-
sity hotspots (Myersetal. 2000), indicating the continued
importance of conserving these regions of high endemism.
Many also correspond to areas recently identied as plant
rarity hotspots (Enquistetal. 2019). Mountain ranges and
other areas of high topographic diversity feature prominently,
likely due to existing high biodiversity, high endemism,
and comparatively high climate and microclimate diversity
(Rahbeketal. 2019a, b) coupled with the comparatively low
velocity of climate change in these areas (Loarieetal. 2009).
However, as the core area algorithm used (Moilanen 2007)
prioritizes representation of all species, and will try to link
niche-tracking to areas already conserved to the greatest
degree possible, spatial priorities capture many distinct eco-
systems and climate types including lowland habitats.
e areas our prioritization highlighted as important for
species on the move include many areas that are also high
conservation priority under current climate. is is because
conserving a species from present to future begins with con-
servation of the species’ current range, since this serves as
the starting point for any future dispersal. Restricted-range
species are concentrated in mountainous areas (Rahbeketal.
2019a, b), where species’ movements in response to climate
change will generally be upslope with warming temperatures
(Peters and Darling 1985, Halpin 1997) or into nearby areas
with suitable microclimates in complex terrain (Hannahetal.
2014, Rahbeketal. 2019b). Low velocity of climate change
in mountains (Loarieetal. 2009) and decreasing area with
elevation mean that species’ upslope movements will occur
over shorter distances in mountains (Serra-Diazetal. 2014),
further concentrating restricted-range species near current
montane centers of endemism (Enquistetal. 2019).
We observe a steep drop-o in modeled AER with
increasing conservation of (primarily montane) natural area
(Table 1). If conservation is limited to existing protected
areas, AER is projected to be high under both climate change
scenarios – ranging from 60 to 77% under both RCP 8.5 and
2.6. Moving to 30% land conservation combined with lower-
ing climate change (RCP2.6) results in reduction of modeled
AER by 52–68% (ensemble mean) across all regions when
compared with existing levels of conservation and higher
(RCP8.5) climate change. Unsurprisingly, the greatest extinc-
tion risk reduction could be achieved by conserving 50% of
terrestrial area with a low climate change scenario; this com-
bined action reduces modeled extinction risk by 72–82% in
the ensemble mean (Table 1). Conserving 30% or 50% of
land but with high climate change (RCP8.5), reduces mod-
eled extinction risk between 46–66% and 62–76% respec-
tively across the three regions under business-as-usual climate
change (RCP 8.5) in the ensemble mean. Reducing climate
change from business as usual (RCP 8.5) to the RCP 2.6
scenario which is more consistent with limiting mean global
temperature change to 2.0° could reduce extinction risk
5–6% more under 30% land conservation. In contrast, with
no increase in existing conservation area (<17% area), the
same climate policy is projected to result in roughly −2 to
12% additional decrease in ensemble mean extinction risk.
Figure1. Modeled extinction risk reduction with increasing land conservation. Ensemble mean Aggregate extinction risk (AER) versus %
terrestrial area conserved under RCP 2.6 (dashed lines) and RCP 8.5 (solid lines) for the Afrotropics (blue), Neotropics (red) and Asia
Tropics (green). Vertical lines show the aggregate extinction risk curve intersection with 17%, 30% and 50% terrestrial area conserved. AER
is the mean of individual species extinction risk at each increment of conserved natural land and is scaled from 0 to 1 (zero probability of
extinction to likely extinct).
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
948
Figure2. (A–C) Spatial prioritizations for land conservation to minimize extinction risk. Zonation spatial solutions to minimize extinction
risk in the Afrotropics (A), SE Asia tropics (B) and Neotropics (C). AER values in Fig. 1 are derived from these spatial solutions. Areas of
darkest green are highest priority areas to minimize extinction risk at 17% land conservation. Successively lighter shades of green represent
areas that minimize extinction risk at 30% and 50% land conservation. Existing conserved areas registered in the World Database of
Protected Areas are shown in yellow. Priorities are selected to maximize representation of all species in both baseline (1960–1990) and future
ranges (RCP8.5 2060–2080).
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
949
Discussion
Our analysis suggests that 30% land conservation combined
with climate change action could reduce extinction risk by
half or more across multiple conservation/climate combi-
nations in all three tropical regions. Models show that even
without lowered climate change (RCP8.5), 30% land con-
servation could reduce AER by at least half in the ensemble
mean, except in the Neotropics. e lower levels of modeled
reduction in the Neotropics are likely due to the high number
of rare species in the Andes, distributed over a relatively large
area that can not be completely captured with 30% land con-
servation. 30% land conservation could therefore pay major
benets to biodiversity conservation and is a strong target for
CBD post-2020 consideration.
e analysis also suggests that 50% land conservation
could lead to an even larger decrease in extinction risk, result-
ing in over 80% reduction in all regions when coupled with
lower climate change (RCP2.6). Roughly half of the world’s
ecoregions have already lost 50% or more of their original
natural habitat (Dinersteinetal. 2017) so this longer-term
vision of 50% land conservation would require restoration
in many areas. Careful targeting of this restoration can have
important long-term benets for biodiversity. Geographic
targeting at coarse scales can be accomplished by selecting
priority areas indicated in Fig. 2A–C. Finer-scale priorities
can be further elaborated by conducting local systematic
spatial planning for climate change using methods similar
to this study.
Natural areas can be conserved through a wide range of
mechanisms, from government-designated protected areas
to multiple-use land management such as community con-
servancies and other eective area-based conservation mea-
sures (OECM) (Dudleyetal. 2018, Dinersteinetal. 2019).
Which approach is the most appropriate will depend on the
local environmental and social context (Brownetal. 2003).
What is clear is that large-scale industrial land uses, particu-
larly monoculture and plantations, are becoming increasingly
prevalent, so achieving the results reported here will require
active policy or land use management intervention to main-
tain natural areas in high priority locations of highest value to
species’ present and future ranges.
e modeled reductions in extinction risk we report
depend on specic spatial congurations in our conserva-
tion solutions that in turn depend on GCM and species
model variants. e uncertainty associated with our GCM
Table 1. Aggregate extinction risk (AER) under varying land conservation areas and climate change. Ensemble mean, maximum and mini-
mum AER for high (RCP 8.5) and low (RCP 2.6) climate scenarios and three levels of land conservation (existing protected areas, 30% land
conservation and 50% land conservation) for Asia, Africa and the Neotropics. Values are summed extinction risk across all species, scaled
0 (zero extinction risk for all species) to 1 (100% extinction risk for all species). The ‘change’ columns represent the percent difference in
AER relative to the current protected areas under RCP 8.5. The extinction risk values in Table 1 are presented in continuous form in Fig. 1.
Ensemble mean
Mean ensemble
change Ensemble max
Max ensemble
change Ensemble min
Min ensemble
change
Neotropics
RCP8.5
Existing protected area* 0.686 – 0.765 – 0.631 –
30% land conservation#0.370 −46.1 0.468 −38.8 0.332 −47.4
50% land conservation 0.259 −62.2 0.368 −51.9 0.219 −65.3
RCP2.6
Existing protected area 0.693 −2.6 0.735 −3.4 0.675 −0.8
30% land conservation 0.332 −51.6 0.400 −58.2 0.301 −62.3
50% land conservation 0.192 −72.0 0.269 −74.9 0.159 −74.8
Afrotropics
RCP8.5
Existing protected area 0.628 – 0.657 – 0.601 –
30% land conservation 0.234 −62.7 0.267 −60.2 0.195 −67.6
50% land conservation 0.171 −72.8 0.217 −71.8 0.140 −76.7
RCP2.6
Existing protected area 0.639 1.8 0.653 −0.6 0.629 4.7
30% land conservation 0.197 −68.6 0.220 −66.5 0.178 −70.4
50% land conservation 0.114 −81.8 0.140 −78.7 0.097 −83.9
Asia Tropics
RCP8.5
Existing protected area 0.756 – 0.771 – 0.745 –
30% land conservation 0.261 −65.5 0.300 −61.1 0.227 −69.5
50% land conservation 0.182 −75.9 0.223 −71.1 0.152 −79.6
RCP2.6
Existing protected area 0.671 −11.2 0.762 −1.2 0.614 −17.6
30% land conservation 0.314 −58.5 0.360 −69.2 0.231 −69.0
50% land conservation 0.259 −65.7 0.313 −80.5 0.143 −80.8
* IUCN categories I–VI.
# IUCN categories I–VI + OECM.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
950
variants is relatively low. Across 10 general circulation models
(GCM), the average ratio of ensemble mean to maximum is
0.85 and of ensemble mean to minimum is 1.17 for the 30%
conserved area target. Species modelling uncertainty is higher
than the uncertainty associated with climate models (Araújo
and Guisan 2006). Modeling a large number of species over
a broad spatial domain at relatively ne resolution requires
substantial compute resources, which is then multiplied
when several GCMs and climate pathways are considered. It
is an acknowledged limitation of this study that we report the
results of only one species distribution model method, as it
has been found that choice of modeling method is often the
largest source of variation in species range projections (Diniz-
Filhoetal. 2009, Garciaetal. 2012). Exploring solution space
with multiple species models is important and is a priority
for our research group and we recommend that the results
presented here be vetted against alternative species modeling
techniques. Use of ensemble species modeling methods and
the uncertainties they represent is a key recommendation of
recently agreed upon standards for biodiversity assessments
(Araújoetal. 2019) and is therefore an opportunity to rene
the results presented here. Indeed, the approach described in
Kujalaet al. (2013) which was followed this study oers a
template to incorporate species model uncertainty as well as
climate model uncertainty in the prioritization.
Despite this limitation, there are multiple reasons for us
to believe that these results, using only one species distribu-
tion model method (Maxent) are likely robust at the scale
of our analysis. First, our conservation planning algorithm,
Zonation, seeks solutions that minimize extinction risk,
which means minimizing range loss in rare species. Rare spe-
cies are concentrated in tropical and Mediterranean moun-
tains (Enquist et al. 2019, Rahbek et al. 2019a, b). ese
areas will emerge as high priority under current and future
climate, regardless of species model. Second, we include spe-
cies that have too few records to be modeled. ese species
are included in the conservation prioritizations with their
current occurrences set equal to their future occurrences.
Since most species are rare (Enquistetal. 2019), nearly one-
third of the species driving our priorities are these rare species
with too few occurrences to model. ese species and the
spatial priorities to conserve them are independent of species
models. ird, species will be predominantly moving upslope
in tropical mountains, so while exact locations of conserva-
tion importance for individual species may be dependent on
species models, general tropical montane locations, becom-
ing even more concentrated as climate changes, dominate our
general global results. Since rare plant species are common
and >90% of our modelled species are plants, including non-
modelled species reduces the biases associated with exclusive
reliance on the more widespread species with enough records
to model. Finally, Rare species are concentrated where past
climate change could be tracked with limited movement
(Sandeletal. 2011), likely reecting frequent limited disper-
sal ability in these species (Fodenetal. 2013) so at the spatial
resolution of our study, their future ranges are likely to be
within the same selection unit as their present range, reduc-
ing or eliminating the uncertainty that arises from lack of
ability to simulate their future range.
e AER and spatial results reported here show that on
coarse scales and across broad domains, we know where to
conserve to be most eective in meeting international con-
servation goals, even as climate changes. Conserving the areas
highlighted in our results can pay large dividends in carbon
sequestration to reduce climate change, as well as provid-
ing other ecosystem services while reducing extinction risk
and maintaining biodiversity. However, the converse is not
necessarily true – maximizing carbon sequestration may not
automatically conserve these critical biodiversity and cli-
mate priorities (Di Marco et al. 2015). Carbon sequestra-
tion, both above-ground and below-ground, may be higher
in lowland ecosystems. A strategy to maximize per unit area
carbon sequestration by pushing agriculture and other devel-
opment into more marginal uplands will require clearing
more habitat to meet production needs, in the very habitats
that are most important to conserve (Di Marcoetal. 2015,
Rahbek et al. 2019a, b). In some cases, a strategy of con-
serving more upland area (to sequester the same amount of
carbon) could meet the same carbon goal, provide substantial
ecosystem services such as watershed protection, while having
much greater biodiversity benet and potentially, as tradeo,
allowing some lowland areas with high agricultural potential
to be developed.
Achieving these multiple biodiversity and ecosystem ser-
vice benets in the real world requires continuous, iterative
planning. No set of priorities can be completely or instanta-
neously realized in a world of multiple competing develop-
ment interests that play out over time. Rather, in the real
world, conservation and development planning moves incre-
mentally and often imperfectly. So while the spatial priorities
presented here can greatly reduce extinction risk, real world
planning to reduce extinction risk will need to accept com-
promises and reorder priorities as choices (many times less
than optimal) are made. More important than a perfect set
of priorities is an ongoing planning process that considers
climate change eects on biodiversity and ecosystem services.
Such a process can ensure that at every decision point about
which areas are to remain natural or to be developed, the
highest priority areas for biodiversity under climate change
are identied. While the highest priority area may not always
be conserved, systematic bias towards the highest priority
remaining sites will progressively drive solutions towards
reduction of extinction risk. For this reason, it is important
to establish systematic conservation planning for biodiver-
sity and climate change as a key process within government
oces in charge of conservation and development planning,
as they should continuously update priorities for conserva-
tion as solutions are implemented, as well as when knowledge
of changes in species distributions and other eects of climate
change increases.
Ongoing systematic conservation planning allows for
assimilation of new data and improved climate models. For
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
951
the conservation of biodiversity, uncertainties need not stand
in the way of action. By building systematic planning pro-
cesses that serve communities and nations for decades, pro-
gressive reduction in uncertainty will result in continually
improved conservation and development outcomes. e per-
fect should not be the enemy of the good; our land use plans
will never be perfect, but they can be good, they can incor-
porate consideration of conservation and species movements
due to climate change and they can continually improve.
Doing so can help ensure that all of the species that share this
planet with us will continue to thrive, and provide benets to
people, even as climate changes.
Data availability statement
Data are available from the Figshare Digital Repository: doi:
10.6084/m9.gshare.c.4868019 (Hannahetal. 2019).
Acknowledgements – is work is supported by the Global
Environment Facility grant GEF-5810 ‘Spatial Planning for
Area Conservation in Response to Climate Change (SPARC)’.
Additionally, this work was conducted as a part of the Botanical
Information and Ecology Network (BIEN) Working Group,
2008–2012. PIs BJE, RC, BB, SD and RKP were supported by
the National Center for Ecological Analysis and Synthesis, a center
funded by NSF (grant no. EF-0553768), the Univ. of California,
Santa Barbara, and the State of California. Additional support
was provided for JCD, the NCEAS Research Associate in the
Group. e BIEN Working Group was also supported by the
iPlant Collaborative (NSF no. DBI-0735191). We thank all the
data contributors, numerous herbaria who have contributed their
data to various data compiling organizations for the invaluable
data and support provided to BIEN. We especially thank the New
York Botanical Garden, Missouri Botanical Garden, Biodiversity
Center, Leiden, the Netherlands, the UNC Herbarium as well
as GBIF, REMIB, and SpeciesLink. e sta at iPlant and the
Texas Advanced Computing Center at the University of Texas
provided critical computational assistance. We thank the more
than 50 scientists have participated in our various BIEN working
group and sub-group meetings since 2008 including B. Blonder,
K. Engemann, E. Fegraus, J. Cavender-Bares, B. Dobrin,
K. Gendler, R. Jorgensen, G. Lopez-Gonzalez, L. Zhenyuan, S. McKay,
O. Phillips, J. Pickering, N. Swenson, C. Vriesendorp, K. Woods,
who participated in a working group meeting and D. Ackerly,
E. Garnier, R. Guralnick, W. Jetz, J. Macklin, N. Matasci,
S. Ramteke and A. Zanne participated in sub-group meetings. We
thank John J. Wiens and Michael Sanderson for prior comments
and encouragement. We also acknowledge the critical support
of iPlant from R. Jorgensen, S. Go, N. Matasci and R. Walls.
Further, the long-term vision, encouragement, and support of
F. Davis, S. Hampton, M. Jones, and the ever-helpful sta at NCEAS
were critical for the completion of this rst stage of the BIEN
working group. NMH was supported by the European Union’s
Horizon 2020 research and innovation program under the Marie
Sklodowska-Curie grant agreement No. 746334 and acknowledges
the Danish National Research Foundation for support to the Center
for Macroecology, Evolution and Climate (grant no. DNRF96).
CV was supported by a Marie Curie International Outgoing
Fellowship within the 7th European Community Framework
Program (DiversiTraits project, no. 221060) by the European
Research Council (ERC) Starting Grant Project (Grant ERC-
StG-2014-639706-CONSTRAINTS). JCS and BJE acknowledge
support from the Center for Informatics Research on Complexity
in Ecology (CIRCE), funded by the Aarhus University Research
Foundation under the AU Ideas program. XF and EAN were
supported by the University of Arizona Bridging Biodiversity and
Conservation Science program. CM acknowledges funding from
NSF Grant DBI-1913673. JCS considers this work a contribution
to his VILLUM Investigator project ‘Biodiversity Dynamics in a
Changing World’ funded by VILLUM FONDEN (grant 16549).
We acknowledge the herbaria that contributed data to this work:
HA, FCO, DUKE, MFU, UNEX, VDB, ASDM, AMD, BPI, BRI,
BRM, CLF, CNPO, L, LPB, AD, A, TAES, FEN, FHO, ANSM,
ASU, B, BCMEX, RAS, RB, TRH, AAH, ACOR, UI, AK, CAS,
ALCB, AKPM, EA, AAU, ALTA, ALU, AMES, AMNH, AMO,
CHAPA, GH, ANGU, ANSP, ARAN, AS, CICY, BAI, CIMI,
AUT, BA, BAA, BAB, CMMEX, BACP, BAF, BAJ, BAL, COCA,
CODAGEM, BARC, BAS, BBS, BC, BCN, BCRU, BEREA, BG,
BH, BIO, BISH, SEV, BLA, BM, BOCH, MJG, BOL, CVRD,
BOLV, BONN, DAV, BOUM, BR, DES, BREM, BRLU, BSB,
BUT, C, DS, CALI, CAN, CANB, CAY, EBUM, CBM, CEN,
CEPEC, CESJ, CHR, ENCB, CIIDIR, CINC, CLEMS, F, COA,
COAH, FCME, COFC, CP, COL, COLO, CONC, CORD, CPAP,
CPUN, CR, CRAI, FURB, CU, G, CRP, CS, CSU, CTES, CTESN,
CUZ, DAO, HB, DBN, DLF, DNA, DR, DUSS, E, HUA, EAC,
EIF, EIU, GES, GI, GLM, GMNHJ, K, GOET, GUA, EMMA,
HUAZ, ERA, ESA, FAA, FAU, FB, UVIC, FI, GZU, H, FLAS,
FLOR, HCIB, FR, FTG, FUEL, GB, HNT, GDA, HPL, GENT,
HUAA, HUJ, CGE, HAL, HAM, IAC, HAMAB, HAO, HAS, IB,
HASU, HBG, IBUG, HBR, HEID, IEB, HIP, IBGE, ICEL, ICN,
ILL, SF, HO, HRCB, HRP, HSS, HU, HUAL, HUEFS, HUEM,
HUFU, HUSA, HUT, IAA, HXBH, HYO, IAN, ILLS, HAC,
IPRN, IMSSM, FCQ, ABH, INEGI, INIF, BAFC, BBB, INPA,
IPA, NAS, INB, INM, MW, EAN, IZTA, ISKW, ISC, ISL, GAT,
JEPS, IBSC, UCSB, ISTC, ISU, IZAC, JACA, JBAG, JE, SD,
JUA, JYV, KIEL, ECON, KSC, TOYA, MPN, USF, TALL, RELC,
CATA, AQP, KMN, KMNH, KOELN, KOR, FRU, KPM, KSTC,
LAGU, TRTE, KSU, UESC, GRA, IBK, KTU, ACAD, MISSA,
KU, PSU, KYO, LA, LOMA, LW, SUU, UNITEC, TASH, NAC,
UBC, IEA, GMDRC, LD, M, LE, LEB, LIL, LINN, AV, HUCP,
QFA, LISE, MBML, NM, MT, FAUC, MACF, CATIE, LTB, LISI,
LISU, MEXU, LL, LOJA, LP, LPAG, MGC, LPD, LPS, IRVC,
MICH, JOTR, LSU, LBG, WOLL, LTR, MNHN, CDBI, LYJB,
MOL, DBG, AWH, NH, HSC, LMS, MELU, NZFRI, MA, UU,
MU, CSUSB, MAF, MAK, MB, KUN, MARY, MASS, MBK,
MBM, UCSC, UCS, JBGP, DSM, OBI, BESA, LSUM, FULD,
MCNS, ICESI, MEL, MEN, TUB, MERL, CGMS, MFA, FSU,
MG, HIB, MIL, DPU, TRT, BABY, ETH, YAMA, SCFS, SACT,
ER, JCT, JROH, SBBG, SAV, PDD, MIN, SJSU, MMMN,
PAMP, MNHM, OS, SDSU, BOTU, OXF, P, MOR, POM, MPU,
MPUC, MSB, MSC, CANU, SFV, RSA, CNS, WIN, MSUN,
CIB, MUR, MTMG, VIT, MUB, MVFA, SLPM, MVFQ, PGM,
MVJB, MVM, MY, PASA, N, UCMM, HGM, TAM, BOON,
UFS, MARS, CMM, NA, NU, UADY, UAMIZ, UC, NE, NHM,
NHMC, NHT, UFMA, NLH, UFRJ, UFRN, ULS, UMO, UNL,
UNM, US, NMB, NMNL, USP, NMR, NMSU, WIS, NSPM,
XAL, NSW, NT, ZMT, BRIT, MO, NCU, NY, TEX, U, UNCC,
NUM, O, CHSC, LINC, CHAS, ODU, CDA, OSA, OSC, OSH,
OULU, OWU, PACA, PAR, UPS, PE, PEL, SGO, PEUFR, PFC,
PH, PKDC, SI, PLAT, PMA, PORT, PR, QM, PRC, TRA, PRE,
PY, QCA, TROM, QCNE, QRS, UH, QUE, R, SAM, RBR, REG,
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
952
RFA, RIOC, RM, RNG, RYU, S, SALA, SANT, SAPS, SASK, SBT,
SEL, SIU, SJRP, SMDB, SMF, SNM, SOM, SP, SRFA, SPF, SPSF,
SQF, STL, STU, SVG, TAI, TAIF, TAMU, TAN, TEF, TENN,
TEPB, TFC, TI, TKPM, TNS, TO, TU, UAM, UB, UCR, UEC,
UFG, UFMT, UFP, UGDA, UJAT, ULM, UME, UNA, UNB,
UNR, UNSL, UPCB, UPEI, UPNA, USAS, USJ, USM, USNC,
USZ, UT, UTC, UTEP, UWO, V, VAL, VALD, VEN, VMSL,
VT, W, WAG, WAT, WII, WELT, WFU, WMNH, WS, WTU,
WU, Z, ZSS, ZT, CUVC, LZ, AAS, AFS, BHCB, CHAM, FM,
PERTH, SAN.
References
Alagador, D.etal. 2016. Climate change, species range shifts and
dispersal corridors: an evaluation of spatial conservation mod-
els. – Methods Ecol. Evol. 7: 853–866.
Anderson-Teixeira, K. J.et al. 2015. CTFS-ForestGEO: a world-
wide network monitoring forests in an era of global change.
– Global Change Biol. 21: 528–549.
Angelo, C. L. and Daehler, C. C. 2013. Upward expansion of re-
adapted grasses along a warming tropical elevation gradient.
– Ecography 36: 551–559.
Araújo, M. B. and Guisan, A. 2006. Five (or so) challenges for
species distribution modelling. – J. Biogeogr. 33: 1677–1688.
Araújo, M. B.etal. 2004. Would climate change drive species out
of reserves? An assessment of existing reserve-selection methods.
– Global Change Biol. 10: 1618–1626.
Araújo, M. B. et al. 2019. Standards for distribution models in
biodiversity assessments. – Sci. Adv. 5: eaat4858.
Austin, K. G.etal. 2017. Trends in size of tropical deforestation
events signal increasing dominance of industrial-scale drivers.
– Environ. Res. Lett. 12: 054009.
Bagchi, R.etal. 2018. Forecasting potential routes for movement
of endemic birds among important sites for biodiversity in the
Albertine Rift under projected climate change. – Ecography 41:
401–413.
Baillie, J. and Zhang, Y.-P. 2018. Space for nature. – Science 361: 1051.
Bellard, C.etal. 2012. Impacts of climate change on the future of
biodiversity. – Ecol. Lett. 15: 365–377.
Birdlife International 2018. Endemic bird area spatial data. – Birld-
life International, Cambridge.
Brown, J. H.etal. 2003. Management of the semi-natural matrix.
– In: Marquet, P. A. and Bradshaw, G. A. (eds), How landscapes
change. Springer, pp. 327–343.
CBD Secretariat 2020. Zero draft of the post-2020 Global Biodi-
versity Framework, CBD/WG2020/2/3. Convention on Bio-
logical Diversity, Montreal, Canada.
Ceballos, G.et al. 2015. Accelerated modern human-induced species
losses: entering the sixth mass extinction. – Sci. Adv. 1: e1400253.
Chen, I.-C.etal. 2011. Rapid range shifts of species associated with
high levels of climate warming. – Science 333: 1024–1026.
Corlett, R. T. and Tomlinson, K. W. 2020. Climate change and
edaphic specialists: irresistible force meets immovable object?
– Trends Ecol. Evol. in press.
DeWalt, S. J.etal. 1999 Ethnobotany of the Tacana: quantitative
inventories of two permanent plots of northwestern Bolivia.
– Econ. Bot. 53: 237–260.
Di Marco, M. et al. 2015. Synergies and tradeos in achieving
global biodiversity targets. – Conserv. Biol. 30: 189–195.
Dinerstein, E.etal. 2017. An ecoregion-based approach to protect-
ing half the terrestrial realm. – BioScience 67: 534–545.
Dinerstein, E.et al. 2019. A global deal for nature: guiding prin-
ciples, milestones, and targets. – Sci. Adv. 5: eaaw2869.
Diniz-Filho, J. A. F.etal. 2009. Partitioning and mapping uncer-
tainties in ensembles of forecasts of species turnover under cli-
mate change. – Ecography 32: 897–906.
Dirzo, R.etal. 2014. Defaunation in the Anthropocene. – Science
345: 401–406.
Dudley, N. and Stolton, S. (eds) 2003. Running pure: the impor-
tance of forest protected areas to drinking water. – WWF and
the World Bank, Gland, Switzerland and Washington DC.
Dudley, N. etal. 2018. e essential role of other eective area-
based conservation measures in achieving big bold conservation
targets. – Global Ecol. Conserv. 15: e00424.
Ehrlich, P. R. and Mooney, H. A. 1983. Extinction, substitution,
and ecosystem services. – BioScience 33: 248–254.
Enquist, B. and Boyle, B. 2012. SALVIAS – the SALVIAS vegeta-
tion inventory database. – Biodivers. Ecol. 4: 288–288.
Enquist, B. J. et al. 2016. Cyberinfrastructure for an integrated
botanical information network to investigate the ecological
impacts of global climate change on plant biodiversity. – PeerJ
Preprints < https://peerj.com/preprints/2615.pdf>.
Enquist, B. J.etal. 2019. e commonness of rarity: global and
future distribution of rarity across land plants. – Sci. Adv. 5:
eaaz0414.
Fegraus, E. 2012. Tropical Ecology Assessment and Monitoring
Network (TEAM Network). – Biodivers. Ecol. 4: 287–287.
Foden, W. et al. 2007. A changing climate is eroding the geo-
graphical range of the Namib Desert tree Aloe through popula-
tion declines and dispersal lags. – Divers. Distrib. 13: 645–653.
Foden, W. B. et al. 2013. Identifying the world’s most climate
change vulnerable species: a systematic trait-based assessment
of all birds, amphibians and corals. – PLoS One 8: e65427.
Forest Inventory and Analysis National Program 2013. – <www.
a.fs.fed.us/>.
Garcia, R. A. et al. 2012. Exploring consensus in 21st century
projections of climatically suitable areas for African vertebrates.
– Global Change Biol. 18: 1253–1269.
Griscom, B. W.etal. 2017. Natural climate solutions. – Proc. Natl
Acad. Sci. USA 114: 11645–11650.
Halpin, P. N. 1997. Global climate change and natural-area protec-
tion: management responses and research directions. – Ecol.
Appl. 7: 828–843.
Hannah, L.etal. 2007. Protected area needs in a changing climate.
– Front. Ecol. Environ. 5: 131–138.
Hannah, L.etal. 2014. ‘Fine-grain modeling of species’ response
to climate change: holdouts, stepping-stones, and microrefu-
gia.’ – Trends Ecol. Evol. 29: 390–397.
Hannah, L. et al. 2019. Data from: 30% land conservation and
climate action reduces tropical extinction risk by more than
50%. – Figshare Digital Repository, doi: 10.6084/m9.
gshare.c.4868019.
Heller, N. E. and Zavaleta, E. S. 2009. Biodiversity management
in the face of climate change: a review of 22 years of recom-
mendations. – Biol. Conserv. 142: 14–32.
Hengl, T.etal. 2017. SoilGrids250m: global gridded soil informa-
tion based on machine learning. – PLoS One 12: e0169748.
Hijmans, R. J.etal. 2005. Very high resolution interpolated climate
surfaces for global land areas. – Int. J. Climatol. 25: 1965–1978.
IPBES 2019. Summary for policymakers of the global assessment
report on biodiversity and ecosystem services of the Intergov-
ernmental Science-Policy Platform on Biodiversity and Ecosys-
tem Services.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
953
IUCN and UNEP-WCMC 2018. IUCN spatial data portal,
accessed June 2018. – UNEP-WCMC, Cambridge, UK.
IUCN and UNEP-WCMC 2018. e world database on protected
areas (WDPA) [On-line]. UNEP-WCMC, Cambridge, UK.
<www.protectedplanet.net>, downloaded April 2018.
Johnston, A.etal. 2013. Observed and predicted eects of climate
change on species abundance in protected areas. – Nat. Clim.
Change 3: 1055–1061.
Kujala, H.et al. 2013. Conservation planning with uncertain cli-
mate change projections. – PLoS One 8: e53315.
Lenoir, J.et al. 2008. A signicant upward shift in plant species
optimum elevation during the 20th century. – Science 320:
1768–1771.
Loarie, S. R.etal. 2009. e velocity of climate change. – Nature
462: 1052–1055.
Maitner, B. S.etal. 2018. e bien r package: a tool to access the
Botanical Information and Ecology Network (BIEN) database.
– Methods Ecol. Evol. 9: 373–379.
Marquet, P. A.etal. 2019. Protected area management and climate
change. – In: Lovejoy, T. and Hannah, L. (eds), Climate change
and biodiversity. Yale Univ. Press, pp. 283–296.
Melillo, J. M.et al. 2016. Protected areas’ role in climate-change
mitigation. – Ambio 45: 133–145.
Merow, C.etal. 2013. A practical guide to MaxEnt for modeling
species’ distributions: what it does, and why inputs and settings
matter. – Ecography 36: 1058–1069.
Merow, C. etal. 2014. What do we gain from simplicity versus
complexity in species distribution models? – Ecography 37:
1267–1281.
Moilanen, A. 2007. Landscape zonation, benet functions and
target-based planning: unifying reserve selection strategies.
– Biol. Conserv. 134: 571–579.
Moilanen, A.et al. 2005. Prioritizing multiple-use landscapes for
conservation: methods for large multi-species planning prob-
lems. – Proc. R. Soc. B 272: 1885–1891.
Moilanen, A.et al. 2014. Zonation spatial conservation planning
methods and software v.4. User manual
Mulongoy, K. J. and Gidda, S. B. 2008. e value of nature: eco-
logical, economic, cultural and social benets of protected areas.
– Secretariat of the Convention on Biological Diversity. Mon-
treal, Canada, 30 pages.
Myers, N. et al. 2000. Biodiversity hotspots for conservation
priorities. – Nature 403: 853–858.
Pecl, G. T.et al. 2017. Biodiversity redistribution under climate
change: Impacts on ecosystems and human well-being.
– Science 355: eaai9214.
Peet, R. K.etal. 2012. VegBank: a permanent, open-access archive
for vegetation plot data. – Biodivers. Ecol. 4: 233–241.
Peters, R. L. and Darling, J. D. S. 1985. e Greenhouse Eect
and Nature ReservesGlobal warming would diminish biological
diversity by causing extinctions among reserve species. – BioSci-
ence 35: 707–717.
Phillips, S. J.etal. 2006. Maximum entropy modeling of species
geographic distributions. – Ecol. Model. 190: 231–259.
Phillips, S. J.et al. 2008. Optimizing dispersal corridors for the cape
proteaceae using network ow. – Ecol. Appl. 18: 1200–1211.
Rahbek, C.etal. 2019a. Building mountain biodiversity: geological
and evolutionary processes. – Science 365: 1114–1119.
Rahbek, C.etal. 2019b. Humboldt’s enigma: what causes global
patterns of mountain biodiversity? – Science 365: 1108–1113.
Rockström, J.etal. 2009. Planetary boundaries: exploring the safe
operating space for humanity. – Ecol. Soc. 14.
Sandel, B. et al. 2011. e inuence of late quaternary climate-
change velocity on species endemism. – Science 334: 660–664.
Serra-Diaz, J. M.etal. 2014. Bioclimatic velocity: the pace of spe-
cies exposure to climate change. – Divers. Distrib. 20: 169–180.
Trabucco, A. and Zomer, R. 2019. Global aridity index and potential
evapotranspiration (ET0) climate database v2. gshare. Dataset.
– <https://doi.org/10.6084/m9.gshare.7504448.v3>.
Tuanmu, M.-N. and Jetz, W. 2014. A global 1-km consensus land-
cover product for biodiversity and ecosystem modelling.
– Global Ecol. Biogeogr. 23: 1031–1045.
Urban, M. C. 2015. Accelerating extinction risk from climate
change. – Science 348: 571–573.
van Proosdij, A. S. J.et al. 2016. Minimum required number of
specimen records to develop accurate species distribution mod-
els. – Ecography 39: 542–552.
Warren, R. et al. 2013. Quantifying the benet of early climate
change mitigation in avoiding biodiversity loss. – Nat. Clim.
Change 3: 678–682.
Williams, P. et al. 2005. Planning for climate change: identifying
minimum-dispersal corridors for the cape Proteaceae.
– Conserv. Biol. 19: 1063–1074.
Wiser, S. K.etal. 2001. Managing biodiversity information: devel-
opment of New Zealand’s National Vegetation Survey data-
bank. – N. Z. J. Ecol. 25: 1–17.
16000587, 2020, 7, Downloaded from https://nsojournals.onlinelibrary.wiley.com/doi/10.1111/ecog.05166, Wiley Online Library on [22/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License