- Email: [email protected]
Protected areas in climate space: What will the future bring?
Biological Conservation 144 (2011) 2119–2125
Contents lists available at ScienceDirect
Biological Conservation journal homepage: www.elsevier.com/locate/biocon
Protected areas in climate space: What will the future bring? John A. Wiens a,b,⇑, Nathaniel E. Seavy a, Dennis Jongsomjit a a b
PRBO Conservation Science, 3820 Cypress Dr. #11, Petaluma, CA 94954, USA School of Plant Biology, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia
a r t i c l e
i n f o
Article history: Received 13 February 2011 Received in revised form 28 April 2011 Accepted 2 May 2011 Available online 26 May 2011 Keywords: California Climate change Climate space Conservation Novel climates Protected areas
a b s t r a c t Protected areas for conservation are intended to contain the environmental conditions that enable species and ecosystems to persist. The locations of such areas are fixed, but the environment within them may change, especially with climate change. To illustrate how multiple climate factors may change in relation to protection status, we used Principal Components Analysis to construct a climate space for California based on eight climate variables assessed at an 800-m resolution. We used projections of future climate derived from a downscaled regional climate model in conjunction with the IPCC SRES A2 scenario to assess how the climate space might shift under future conditions and to identify the combinations of conditions that may no longer occur in the state (disappearing climates) or that will be new to the state (novel climates). Disappearing climates, which were generally toward cooler and/or wetter extremes of the climate space, represented only 0.5% of California’s land area but occurred disproportionately more often in conservation areas that were fully protected. Novel climates (5.8% of California) also occurred disproportionately in fully protected areas; in most cases these climates were characterized by hotter and drier combinations with more seasonal precipitation. The disproportionate occurrence of both novel and disappearing future climates in currently protected areas may create challenges to conservation of the status quo, but such areas may also be ‘‘hotspots of opportunity’’ for responding to the extremes of climate change. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction To live long and prosper, individuals and species require habitat that meets their needs for survival and reproduction – their ecological niche. A major focus of conservation is on protecting or managing places that contain such essential habitat while removing or mitigating other threats. This premise – build it and they will come (or stay) – underlies the efforts of many non-governmental and government entities to set aside protected areas for nature conservation. These efforts have been hugely successful. Globally, some 10–15% of the earth’s surface is under some form of protection (Chape et al., 2005; Soutullo, 2010), and some of the world’s most imperiled species can only be found in such protected areas. The locations of protected areas are geographically fixed. The environment within these areas, however, is not. As awareness of the reality and potential impacts of climate change has increased, it has become apparent that maintaining the conservation value of a protected area may not always be possible (Cole and Yung, 2010). As the climate changes, conditions within a protected area will also be altered, unleashing a cascade of changes in habitats,
⇑ Corresponding author. Address: 7233 NW Valley View Drive, Corvallis, OR 97330, USA. Tel.: +1 703 268 1869; fax: +1 707 765 1685. E-mail address: [email protected] (J.A. Wiens). 0006-3207/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2011.05.002
species, and biological communities. Protecting areas for conservation will increasingly resemble shooting at a moving target. It is important to anticipate how these cascading changes might affect the value of existing protected areas (Griffith et al., 2009). Species-distribution models (SDMs) or habitat suitability models (HSMs) have been used to project the spatial distribution of species or their habitats under future climatic conditions (e.g., Elith and Leathwick, 2009; Stralberg et al., 2009; Thuiller and Münkemüller, 2010) and therefore which species may remain in, leave, or enter protected areas (Kharouba and Kerr, 2010). The premise is that a species can occupy areas that fall within the ‘‘climate space’’ that specifies the combinations of acceptable climatic conditions (e.g., Holdridge, 1947; Whittaker, 1975; Gates, 1980) and that the species’ distributional limits occur where climatic factors fall outside that envelope. Because the climate space is defined independently of geographic locations, the challenge is to determine where combinations of factors in the climate space map onto the locations of existing protected areas in geographical space, both now and in the future. To do this, one would need to know the multivariate set of suitable climatic conditions for each species of interest. Here we take a step back to ask, independently of any single species, how existing protected areas are represented and distributed in the current climate space of a region and how this representation and distribution may change in the future. To illustrate the approach, we
2120
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
construct a climate space for the state of California and determine how that climate space is projected to change in the future. Although there are many ways that these data might be used to describe shifts in climate, we focus on two of the most dramatic by identifying novel climates (portions of climate space that do not presently occur in California but may appear with future climate change) and disappearing climates (portions of the current climate space that may disappear from the state in the future) (c.f. Williams et al., 2007; Ackerly et al., 2010). We then map the geographic locations of these novel and disappearing climates and evaluate the degree to which they may be represented in areas under different levels of existing protection in California. Because effective management of protected areas will require information about potential future conditions in addition to the current conservation value of such places, identifying future novel and disappearing suites of climate conditions may be particularly useful in evaluating how well the current portfolio of protected areas will encompass the full range of future conditions in the state. We limit our analysis to the state of California because this corresponds with the geographic scope of the state agencies and nongovernmental organizations responsible for identifying and planning the protection of conservation areas in the state (as visualized, for example, in State Wildlife Action Plans; Bunn et al., 2007). This focus also aligns with the scope of analysis we have used in previous investigations of shifts in bird distributions within the state (Stralberg et al., 2009; Wiens et al., 2009) and in other studies now underway. California is a tremendously diverse state in terms of topography and climate and it is a regional ‘‘hotspot’’ of biological diversity within North America (Flather et al., 2008; Stein, 2008), so it represents a suitable arena in which to develop and illustrate this approach. In its general features, the approach should be applicable to a variety of other geographies over a range of spatial scales.
2. Methods 2.1. Data 2.1.1. Current and future climate data We used two sources of information to describe the current and future climates of California. Current climate conditions were summarized using 30-year (1971–2000) monthly climate normals interpolated at an 800-m grid resolution by the PRISM Group (Daly et al., 1994, 2000). Future climate conditions were summarized using projections from a regional climate model (RCM), RegCM3 (Pal et al., 2007) at a 30-km resolution with emissions trajectories taken from the Intergovernmental Panel on Climate Change (IPCC) SRES A2 scenario and boundary conditions from the National Center for Atmospheric Research (NCAR) Community Climate System Model (CCSM3.0; Snyder and Sloan, 2005), an atmosphere–ocean global climate model (GCM) run from 1870 to 2099. The RCM was run from 2038 to 2069. For this time period, monthly temperature and precipitation outputs were averaged across years to obtain one set of monthly values. These 30-km resolution results were then statistically downscaled using the change factor method (Wilby et al., 2004) to an 800-m resolution to match the PRISM grid (for details see Stralberg et al., 2009; Wiens et al., 2009). Because our intent was to develop and illustrate the climatespace approach rather than to explore a comprehensive array of possible future climates (as is done in ensemble models; Murphy et al., 2004; Lopez et al., 2009), we used only a single GCM. In comparison with the Geophysical Fluid Dynamics Laboratory model (GDFL CM2.1; Delworth et al., 2006) that we (Stralberg et al., 2009; Wiens et al., 2009) and Ackerly et al. (2010) have used, CCSM3.0 projects a drier and slightly cooler future climate for much of California. The Canadian Centre for Climate Modelling
and Analysis climate model (CCCMA_CGCM3_1.1; see http:// www.ec.gc.ca/ccmac-cccma/default.asp?lang=En&n=1299529F-1, accessed 26 January 2011), also used by Ackerly et al. (2010), projects warmer and wetter conditions than does CCSM3.0 (see http:// meteora.ucsd.edu/cap/cccc_model_prelim.html#kdpdt, accessed 26 January 2011, for model comparisons). For the current and future climate data, we summarized monthly minimum, maximum, and average temperatures and precipitation using eight bioclimatic variables (BIOCLIM; Nix, 1986; Beaumont et al., 2005) that were previously identified as biologically relevant to California wildlife (Stralberg et al., 2009). These were: Annual mean temperature, calculated as the 12-month average of mean monthly temperature. Mean diurnal temperature range, calculated from the 12-month average of the difference between mean maximum and mean minimum temperature for each month. Isothermality, calculated as the ratio of mean diurnal temperature range to the annual temperature range (maximum temperature of the warmest month – minimum temperature of the coldest month). Isothermality compares monthly variability (mean diurnal temperature range) to annual variability (annual temperature range). When monthly variability is greater than annual variability, isothermality is >1; when monthly variability is less than annual variability, isothermality is <1. Temperature seasonality, calculated as the 12-month standard deviation of mean monthly temperature. Mean temperature of warmest quarter, calculated as the average temperature of the warmest 3-month period. Annual precipitation, the 12-month total of mean monthly precipitation. Precipitation seasonality, calculated as the 12-month coefficient of variation of mean monthly precipitation. Precipitation of driest quarter, calculated as the average precipitation for the driest 3-month period. 2.1.2. Definition of protected areas To quantify the geographic distribution of protected areas in California, we used the USGS Gap Analysis Program Protected Area Database for the United States (http://www.protectedlands.net/ images/PADUSTechnical_Report_Oct_2009.pdf, accessed 26 January 2011). This data base identifies four levels of GAP conservation status, which we combined into three levels for analysis: 1 = fully protected (protection of natural resources with natural disturbances either allowed or suppressed, chiefly parks, wilderness areas, wildlife refuges, and nature reserves; Gap status codes 1 and 2); 2 = extractive (protection of natural resources with extractive uses allowed, chiefly state or national forests, multiple-use parks, military reservations, or recreation areas; Gap status code 3); and 3 = unprotected (no mandate for protection of natural resources, chiefly private lands; Gap status code 4). The 764 Fully Protected areas cover 75,987 km2 in California (average size = 9562 ha; range = 0.2–1,068,205 ha); the 347 extractive use areas cover 122,820 km2 (average size = 31,290 ha; range = 0.2–640,867 ha) [comparable data for the unprotected areas, which cover 209,675 km2, cannot be derived because much of this land is in private ownership and parcel sizes are not available]. The geographic distributions of the three categories of protection used in our analyses are shown in Fig. 1. 2.2. Analyses and mapping Although climatic variation in one dimension (e.g., temperature) is relatively easy to conceptualize and display, in actuality climatic variation is multidimensional. Even with only eight climatic variables, the process for conceptualizing and displaying changes in
2121
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
tify pixels as (1) persisting climate: future pixels that were located inside the current climate polygon; (2) disappearing climate: current pixels that were located outside of the future climate polygon; and (3) novel climate: future pixels that were located outside of the current climate polygon. Although it is important to determine which combinations of climate conditions are represented or not represented in the climate space of existing protected areas, conservation planning is framed in terms of geographic space. It is therefore important to consider how novel and disappearing climate spaces translate into real locations on the ground. Accordingly, we mapped the geographic locations of the novel and disappearing climates and compared these locations with those of the protected areas of different protection status. 3. Results 3.1. Current and future California climate space
Fig. 1. Locations of areas of different protection status in California.
this multidimensional space is not straightforward. One approach to this problem has been simply to reduce climate to two of the most important variables and then display shifts in two-dimensional space (e.g., Seavy et al., 2008; Ackerly et al., 2010). This approach is intuitively appealing, but it fails to capture some of the subtle shifts in climatic conditions. An alternative approach is to use multivariate statistics to collapse a suite of climate variables into a smaller set of variables that incorporate as much of the variation as possible (e.g., Comrie and Glenn, 1998; Abatzoglou et al., 2009). For example, multivariate geographical clustering has been used to group agricultural landscapes by multiple environmental factors (Williams et al., 2008) or delineate the boundaries of ecoregions (Hargrove and Hoffman, 1999). To reduce the dimensionality of the eight climatic variables for the current climate, we used Principal Components Analysis (PCA; McGarigal et al., 2000) with climate values for every 800-m pixel in California as inputs to derive the PCA dimensions and factor loadings. The same PCA model was then used to define the future climate space (i.e., we assumed that the relationships among climate variables that determined the PCA dimensions will not change, even though the position of an individual pixel may shift from the current to the future climate space). We used the two axes that explained the majority of variation in the multivariate climate data to portray the climate space. Using these two PCA axes, we plotted each of the 800-m California pixels in the two-dimensional current and future climate spaces. As climate conditions in California change in the future, the climate space as a whole will shift to incorporate new portions of the PCA space that are not currently represented anywhere in the state – these will be the novel climates. Other combinations of climate conditions may disappear entirely – these are the disappearing climates. To assess these changes, we compared areas of overlap and nonoverlap between the current and future climate spaces for California. For the current and future climatic conditions, we used 2-dimensional kernel smoothing to generate a density surface of the climate-space pixels (in PCA space) from which we derived polygons encompassing the areas of the current and future climate space that contained 99.9% of the pixels. We then used these polygons to iden-
Three PCA axes had eigenvalues >1 (PC1 = 3.42, PC2 = 2.40, PC3 = 1.02). The first two axes explained 73% of the variation in the data. We used these two axes to define the multivariate climate space for California. Factor loadings of the eight variables that entered into the analysis (Table 1) suggest that PC1 represents a gradient from warmer, drier areas (negative values) to cooler, wetter areas (positive loadings). PC2 appears to be more strongly associated with seasonality, contrasting areas with a relatively stable temperature and strongly seasonal precipitation (such as coastal areas; positive values) with areas showing greater seasonal variation in temperatures (such as interior mountain and desert areas; negative values) (Fig. 2). The overall climate space is strongly asymmetrical, with cooler, wetter, and either more seasonal temperatures (lower right) or more seasonal precipitation (upper right) conditions predominating (Fig. 3A). 3.2. What will be gained or lost as the climate space shifts? According to the projections of the GCM and RCM we used, future shifts in the California climate space (Fig. 3B) will cause some combinations of climate factors that currently exist to disappear from the state, while other new combinations of climate factors will emerge. Because the bulk of the climate space remains in roughly the same position in the PCA space, these changes occur along the periphery of the climate space. The general trajectory of change is toward lower values of PC1 and somewhat lower values of PC2; therefore, the disappearing climates occur on the trailing edge of this climate-space shift (Fig. 3B). These disappearing climate conditions are generally toward the cooler and wetter (and, to a lesser degree, hotter and drier, and more seasonal) extremes of the climate space. Geographically, the disappearing climates occur in the northern California coast and areas of the Mono Basin, Death Valley, and the southeastern Great Basin (Fig. 4). Despite concerns about the impending loss Table 1 Factor loadings of climate variables on PCA axes 1 and 2. Climate variable
Abbreviation
Factor loadings PC1
Annual mean temperature Mean diurnal range Isothermality Temperature seasonality Mean temperature of warmest quarter Annual precipitation Precipitation seasonality Precipitation of driest quarter
ann temp temp range isotherm temp seas warmest-q temp ann precip precip seas driest-q precip
0.47 0.31 0.05 0.29 0.50 0.43 0.05 0.40
PC2 0.15 0.01 0.56 0.52 0.04 0.03 0.53 0.34
2122
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
Fig. 2. Geographic distributions of scores on the first and second PCA axes.
Fig. 3. Left: distribution of 800-m pixels in the current climate space for California defined by the first and second axes of a Principal Components Analysis of eight bioclimatic variables. Arrows indicate the direction and magnitude of loading of the eight variables on the two PCA axes (abbreviations in Table 1). Right: shifts in the climate space from current to future conditions projected by the climate-change scenario modeled, indicating the distributions of pixels with disappearing and with emerging, novel climate conditions.
of climates and habitats in high-elevation areas of the Sierra Nevada and elsewhere (Dirnböck et al., 2010), few of these climates are projected to disappear entirely (Fig. 4). Emerging, novel climates show quite a different pattern. It is likely that areas characterized as consistently hotter and drier (but with more seasonal precipitation) than now occur will emerge (Fig. 3B). Many of these new climate combinations will appear in the Sonoran and Colorado deserts and in portions of Death Valley and the Mojave Desert. But areas with cooler conditions and greater precipitation in the driest quarter of the year are also likely to
appear. Geographically, these areas will occur in the southern California coast, central California coast ranges, northern California coast, and Klamath Mountains (Fig. 4). 3.3. How are novel and disappearing climates captured by current protected areas? We calculated the area of novel and disappearing climates that occurred in each protected-area category and compared these values to those for California as a whole (Table 2). Areas with
2123
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
Fig. 4. Geographic locations of pixels with disappearing (A) and novel (B) climate conditions, as shown in Fig. 3.
disappearing climates represent only a small area (0.35%) of the state (Fig. 4). If the occurrence of disappearing climates in Fully Protected areas was directly related to the area of the state that is in this protection category (18.6%), we would expect disappearing climates to occur over some 263 km2. Instead, the projected occurrence of disappearing climates in Fully Protected areas is disproportionately great – roughly 2.5 times greater than expected (Table 2). On the other hand, areas subjected to extractive resource use (chiefly State and National Forests and Bureau of Land Management lands) are projected to contain over 3 times less disappearing climate than would be expected on the basis of the proportion of the state in this protection category (30.1%; Table 2). New, novel climates are projected to occur over some 5.3% of the state, with somewhat disproportionate gains in Fully Protected areas (1.22 times greater than expected; Table 2). Unprotected areas, which comprise over half of the lands in California, are projected to contain slightly less area of disappearing climate and of novel climate (1.18 times and 1.11 times less, respectively) than expected. Overall, the observed proportions of disappearing and of novel climate pixels in the three protection categories differ significantly from what would be expected if they were distributed independently of protection category (X 22 = 1330, P < 0.001 and X 22 = 502, P < 0.001, respectively). 4. Discussion Protected areas are fixed in space, but the environments they contain change over time, altering the conservation value of such
areas. Absent episodic large disturbances such as fires, hurricanes, or volcanic eruptions, these changes have in the past been relatively slow, driven by the processes of ecological succession and changing land use in the surroundings. Because these changes have also been localized, they have been amenable to management (e.g., prescribed burning, targeted grazing), allowing the conservation values of protected areas to be retained or restored. Now, however, the pace of environmental change is quickening and threatens to move ecological systems beyond the zone of recent historical environmental variations. These changes are increasingly driven by broad-scale forces such as climate change or economic globalization that lie beyond the reach of local-scale conservation and management. As the distributions of species shift in response to climate change, there will be losses and additions of species to local communities (Peterson et al., 2002; Loarie et al., 2008, 2009), in some situations resulting in ‘‘no-analog’’ assemblages that have no contemporary counterpart (Root and Schneider, 1993; Stralberg et al., 2009). Webs of species interactions will be disrupted and the species that remain will encounter new challenges from species invading from elsewhere. Things will be different. Because these projected changes in species occurrences and communities are driven by future changes in climate, we have taken a step back to ask how the occurrence and distribution of the climate conditions themselves may change over a region under a scenario of future global and regional climate change and what this may mean for the future of protected areas. We use the state of California as an illustration, but the approach should be applicable to other areas at multiple scales.
Table 2 The land area (km2) of California that is classified as novel or disappearing climate in three protection categories. The percentages of the total land area of the state in the protection categories and in disappearing and novel climates are given in parentheses. The area of disappearing or novel climates that would be expected for a protection category given its percentage of the land area of the state as a whole is given in brackets. Climate type
Disappearing Novel All land area
Protection category All California
Fully protected
Extractive use
Unprotected
1 412 (0.35%) 21 630 (5.3%) 408 483
669 [263] 4 920 [4 024] 75 987 (18.6%)
130 [425] 6 733 [6 504] 122 820 (30.1%)
613 [725] 9 978 [11 103] 209 675 (51.3%)
2124
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
In general, the overall range and composition of climate conditions found in California, as represented by the multivariate climate space, may not change much. Although the climate of individual locations will change, the bulk of the climate space represented in Fig. 3B will persist somewhere in the state under the future scenario we have modeled. As the climate space shifts, however, some extreme climate conditions on the periphery of the climate space are likely to disappear, while other new extreme conditions may emerge. Our analysis suggests that such changes in climate are likely to be disproportionately great in areas that are currently most fully protected – National Parks, Wilderness Areas, National Wildlife Refuges, State Parks, and the like. Nationally, many of these protected areas already occur in extreme environments, at high elevations or in places with low productivity, what Scott et al. (2001) have labeled ‘‘rocks and ice.’’ The same is true internationally – fully protected areas are often in places that are less exploitable, more remote, more rugged, and more extreme environmentally than the places where people live and work, predisposing them to be more vulnerable to climate changes. California is no different – think of Death Valley or the iconic National Parks along the crest of the Sierra Nevada, places that are already near the fringes of the current climate space. If species with small ranges disproportionately occupy places with ‘‘rare climates’’ (i.e., at the periphery of the climate space), as suggested by Ohlemüller et al. (2008), they also may be particularly vulnerable to climate change. It is reassuring that only a small proportion of the California climate space is projected to disappear. Because the forces that determine shifts in the climate space are beyond the reach of management, however, nothing can be done to forestall the disappearance of some climate conditions and the consequences of those changes. Simply because some combinations of climate conditions may no longer occur in the state, however, does not mean that the species that presently occupy those areas will also disappear. Because the shifts in climate conditions are dictated by how climate features are interrelated in the multivariate climate space, species whose bioclimatic niche envelopes are determined by only a subset of those features may continue to find suitable conditions in those areas. For other species, the disappearance of some combinations of climate features may affect only a portion of their distribution, so they may continue to occur elsewhere in the state where their requirements are met. In some situations, however, disappearing climates may be associated with the extinction of local populations, as has been projected for American pika (Ochotona princeps; Grayson, 2005; but see Millar and Westfall, 2010). The emergence of novel climates, on the other hand, will create settings for the development of assemblages of species that are also new to California. Some species that currently occur in these areas may remain, others may immigrate from elsewhere in the state, and yet others that do not presently occur in the state may colonize from surrounding regions as distributions shift in response to climate change. Such ‘‘no-analog’’ assemblages may emerge in many areas (Overpeck et al., 1992; Williams et al., 2007; Stralberg et al., 2009), but they may be especially likely where novel climates appear. In a sense, the locations of these emerging novel climates represent future ‘‘hotspots of opportunity.’’ How biotic communities assemble in these areas will depend on a host of factors – whether the conditions of the novel climates fall within the bioclimatic envelopes of species, the dispersal abilities of the species, intervening land uses that impede or promote movements, the presence of suitable habitat and resources, interactions with other species, and so on. In these places, management may need to be directed toward species not considered in current management plans. Alternatively, if the climate of a protected area disappears completely, management of species associated with these conditions might be deemphasized. It will be important to
conduct comprehensive monitoring to detect early signals of changes in distributions and abundances of species; this might most effectively be done as part of an Environmental Change Network (see http://www.ecn.ac.uk/, accessed 12 February 2011). Because Fully Protected areas are projected to experience both disproportionate losses of existing climates as well as disproportionate gains in novel climates, they may represent areas of particularly turbulent future climate conditions. Under these conditions, future management and conservation may need to be especially nimble. Shifting the management focus from individual species to broader assemblages or ecosystem properties, together with using adaptive management in an anticipatory rather than a reactive mode, may be necessary (Lawler et al., 2010). The climates that we project as disappearing from or emerging in California are not uniquely Californian, of course. These conditions may also occur, both now and in the future, in other areas of the Pacific Northwest (disappearing) and Desert Southwest (emerging). It is perhaps no accident that the locations of greatest change in the climate space occur about the fringes of the state (Fig. 4), largely as parts of ecoregions that extend into adjacent states (Miles and Goudey, 1997). Although the truncation of our analysis at the state boundaries is thus somewhat artificial, it is nonetheless pragmatic. The California Wildlife Action Plan (Bunn et al., 2007), produced in response to a Congressional mandate to develop plans for the conservation of wildlife and habitats, is intended to guide conservation and management of species at risk. The California Essential Connectivity Project (Spencer et al., 2010) projects potential land acquisitions and management to enhance linkages among protected areas under current conditions, and this report is also intended to influence the priorities of state agencies as well as environmental organizations. Our analysis is intended primarily to illustrate how a consideration of the current and potential future multivariate climate space for a region may provide insights about the possible effects of climate change on areas protected for wildlife management, wilderness, and conservation. To simplify the analysis, we have used only a single climate-change scenario, one RCM, and a selected set of bioclimatic variables from a broader array. Therefore, the results should not be taken as definitive or predictive, but illustrative of the ways in which climate conditions may shift in the future, how these shifts map onto geographical space, and how the extreme changes (disappearing and novel climates) may differentially affect areas of differing protection status. Clearly, however, changing climatic conditions pose new and challenging questions about the current network of protected areas. To date, strategies for adapting conservation plans to climate change have focused on enhancing resistance or resilience, while preparing for ecological transformation has received less attention (Heller and Zavaleta, 2009; Poiani et al., 2011). In other words, current strategies focus on ways to maintain what is here now rather than how to prepare for what will arrive in the future. The disappearance of current climatic conditions and the emergence of novel climatic conditions in California’s protected areas illustrate that management must address ecological transformation. After all, climate (together with substrate and topography) is a primary driver of the potential distributions of plants and animals, and thence of biodiversity (Whittaker, 1960).
Acknowledgments This paper benefitted from the advice of Grant Ballard, Chrissy Howell, Diana Stralberg, Sam Veloz, and two anonymous reviewers. Mark Snyder provided the data on future climates. Partial funding was provided by the USFWS California Landscape Conservation Cooperative. This is PRBO Publication No. 1810.
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
References Abatzoglou, J.T., Redmond, K.T., Edwards, L.M., 2009. Classification of regional climate variability in the State of California. Journal of Applied Meteorology and Climatology 48, 1527–1541. Ackerly, D.D., Loarie, S.R., Cornwell, W.K., Weiss, S.B., Hamilton, H., Branciforte, R., Kraft, N.J.B., 2010. The geography of climate change: implications for conservation biogeography. Diversity and Distributions 16, 476–487. Beaumont, L.J., Hughes, L., Poulsen, M., 2005. Predicting species distributions: use of climatic parameters in BIOCLIM and its impact on predictions of species’ current and future distributions. Ecological Modeling 186, 251–270. Bunn, D., Mummert, A., Hoshovsky, M., Gilardi, K., Shanks, S., 2007. California Wildlife: Conservation Challenges. University of California Davis Wildlife Health Center, California Department of Fish and Game, Sacramento, CA, USA. Chape, S., Harrison, J., Spalding, M., Lysenko, I., 2005. Measuring the extent and effectiveness of protected areas as an indicator for meeting global biodiversity targets. Philosophical Transactions of the Royal Society B 360, 443–455. Cole, D.N., Yung, L., (Eds.), 2010. Beyond Naturalness: Rethinking Park and Wilderness Stewardship in an Era of Rapid Change. Island Press, Washington, DC. Comrie, A.C., Glenn, E.C., 1998. Principal components-based regionalization of precipitation regimes across the southwest United States and northern Mexico, with an application to monsoon precipitation variability. Climate Research 10, 201–215. Daly, C., Neilson, R.P., Phillips, D.L., 1994. A statistical topographic model for mapping climatological precipitation over mountainous terrain. Journal of Applied Meteorology 33, 140–158. Daly, C., Gibson, W.P., Hannaway, D., Taylor, G., 2000. High-quality spatial climate data sets for the United States and beyond. Transactions of the American Society of Agricultural Engineers 43, 1957–1962. Delworth, T.L., Broccoli, A.J., Rosati, A., Stouffer, R.J., Balaji, V., Beesley, J.A., Cooke, W.F., Dixon, K.W., Dunne, J., Dunne, K.A., Durachta, J.W., Findell, K.L., Ginoux, P., Gnanadesikan, A., Gordon, C.T., Griffies, S.M., Gudgel, R., Harrison, M.J., Held, I.M., Hemler, R.S., Horowitz, L.W., Klein, S.A., Knutson, T.R., Kushner, P.J., Langenhorst, A.R., Lee, H.-C., Lin, S.-J., Lu, J., Malyshev, S.L., Milly, P.C.D., Ramaswamy, V., Russell, J., Schwarzkopf, M.D., Shevliakova, E., Sirutis, J.J., Spelman, M.J., Stern, W.F., Winton, M., Wittenberg, A.T., Wyman, B., Zeng, F., Zhangc, R., 2006. CM2 global coupled climate models. Part I: Formulation and simulation characteristics. Journal of Climate 19, 643–674. Dirnböck, T., Essel, F., Rabitsch, W., 2010. Disproportional risk for habitat loss of high-altitude endemic species under climate change. Global Change Biology 17, 990–996. Elith, J., Leathwick, J., 2009. Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics 40, 677–697. Flather, C.H., Knowles, M.S., McNees, J., 2008. Geographic Patterns of at-risk Species: a Technical Document Supporting the USDA Forest Service Interim Update of the 2000 RPA Assessment. General Technical Report RMRS-GTR-211. US Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO. Gates, D.M., 1980. Biophysical Ecology. Springer-Verlag, New York. Grayson, D.K., 2005. A brief history of Great Basin pikas. Journal of Biogeography 32, 2103–2111. Griffith, B., Scott, J.M., Adamcik, R., Ashe, D., Czech, B., Fischman, R., Gonzalez, P., Lawler, J., McGuire, A.D., Pidgorna, A., 2009. Climate change adaptation for the US National Wildlife Refuge System. Environmental Management 44, 1043–1052. Hargrove, W.W., Hoffman, F.N., 1999. Using multivariate clustering to characterize ecoregion borders. Computing in Science & Engineering 1, 18–25. Heller, N.E., Zavaleta, E.S., 2009. Biodiversity management in the face of climate change: a review of 22 years of recommendations. Biological Conservation 142, 14–32. Holdridge, L.R., 1947. Determination of world plant formations from simple climatic data. Science 105, 367–368. Kharouba, H.M., Kerr, J.T., 2010. Just passing through: global change and the conservation of biodiversity. Biological Conservation 143, 1094–1101. Lawler, J.J., Tear, T.H., Pyke, C., Shaw, M.R., Gonzalez, P., Kareiva, P., Hansen, L., Hannah, L., Klausmeyer, K., Aldous, A., Bienz, C., Pearsall, S., 2010. Resource management in a changing and uncertain climate. Frontiers in Ecology and the Environment 8, 35–43. Loarie, S.R., Carter, B.E., Hayhoe, K., McMahon, S., Moe, R., Knight, C.A., Ackerly, D.D., 2008. Climate change and the future of California’s endemic flora. PLoS ONE 3, e2502. doi:10.1371/journal.pone.0002502. Loarie, S.R., Duffy, P.B., Hamilton, H., Asner, G.P., Field, C.B., Ackerly, D.D., 2009. The velocity of climate change. Nature 462, 1052–1055. Lopez, A., Fung, F., New, M., Watts, G., Weston, A., Wilby, R.L., 2009. From climate model ensembles to climate change impacts and adaptation: a case study of
2125
water resource management in the southwest of England. Water Resources Research 45. doi:10.1029/2008WR007499. McGarigal, K., Cushman, S., Stafford, S., 2000. Multivariate Statistics for Wildlife and Ecology Research. Springer, New York. Miles, S.R., Goudey, C.B., 1997. Ecological Subregions of California. USDA Forest Service Pacific Southwest Region Report R5-EM-TP-005, San Francisco, CA. Millar, C.I., Westfall, R.D., 2010. Distribution and climatic relationships of the American pika (Ochotona princeps) in the Sierra Nevada and Western Great Basin, USA: periglacial landforms as refugia in warming climates. Arctic, Antarctic, and Alpine Research 42, 76–88. Murphy, J.M., Sexton, D.M.H., Barnett, D.N., Jones, G.S., Webb, M.J., Collins, M., Stainforth, D.A., 2004. Quantification of modelling uncertainties in a large ensemble of climate change simulations. Nature 430, 768–772. Nix, H.A., 1986. A biogeographic analysis of Australian elapid snakes. In: Longmore, R. (Ed.), Atlas of Elapid Snakes of Australia. Australian Flora and Fauna Series 7. Australian Government Publishing Service, Canberra, pp. 4–15. Ohlemüller, R., Anderson, B.J., Araújo, M.B., Butchart, S.H.M., Kudrna, O., Ridgely, R.S., Thomas, C.D., 2008. The coincidence of climatic and species rarity: high risk to small-range species from climate change. Biology Letters 4, 568–572. Overpeck, J.T., Webb, R.S., Webb III, T., 1992. Mapping eastern North American vegetation change of the past 18 ka: no-analogs and the future. Geology 20, 1071–1074. Pal, J.S., Giorgi, F., Bi, X., Elguindi, N., Solmon, F., Gao, X., Rauscher, S.A., Francisco, R., Zakey, A., Winter, J., Ashfaq, M., Syed, F.S., Bell, J.L., Diffenbaugh, N.S., Karmacharya, J., Konare, A., Martinez, D., da Rocha, R.P., Sloan, L.C., Steiner, A., 2007. Regional climate modeling for the developing world: the ICTP RegCM3 and RegCNET. Bulletin of the American Meteorological Society 88, 1395–1409. Peterson, A.T., Ortega-Huerta, M.A., Bartley, J., Sanchez-Cordero, V., Soberón, J., Buddemeier, R.H., Stockwell, D.R.B., 2002. Future projections for Mexican faunas under global climate change scenarios. Nature 41, 626–629. Poiani, K.A., Goldman, R.L., Hobson, J., Hoekstra, J.M., Nelson, K.S., 2011. Redesigning biodiversity conservation projects for climate change: examples from the field. Biodiversity and Conservation 20, 185–201. Root, T.L., Schneider, S.H., 1993. Can large-scale climatic models be linked with multiscale ecological studies? Conservation Biology 7, 256–270. Scott, J.M., Abbitt, R.J.F., Groves, C.R., 2001. What are we protecting? Conservation Biology in Practice 2, 18–19. Seavy, N.E., Dybala, K.E., Synder, M.A., 2008. Climate models and ornithology. Auk 125, 1–10. Snyder, M.A., Sloan, L.C., 2005. Transient future climate over the western United States using a regional climate model. Earth Interactions 9, 1–21. Soutullo, A., 2010. Extent of the global network of terrestrial protected areas. Conservation Biology 24, 362–365. Spencer, W.D., Beier, P., Penrod, K., Winters, K., Paulman, C., Rustigian-Romsos, H., Strittholt, J., Parisi, M., Pettler, A., 2010. California Essential Habitat Connectivity Project: A Strategy for Conserving a Connected California. Prepared for California Department of Transportation, California Department of Fish and Game, and Federal Highways Administration., (accessed 27.01.11). Stein, B.A., 2008. Biodiversity and the military mission. In: Benton, N., Ripley, J.D., Plowledge, F. (Eds.), Conserving Biodiversity on Military Lands: a Guide for Natural Resources Managers. NatureServe, Arlington, VA, pp. 2–33. Stralberg, D., Jongsomjit, D., Howell, C.A., Snyder, M.A., Alexander, J.D., Wiens, J.A., 2009. Re-shuffling of species with climate disruption: a no-analog future for California birds? PLoS ONE 4 (9), e6825. doi:10.1371/journal.pone.0006825. Thuiller, W., Münkemüller, T., 2010. Habitat suitability modeling. In: Moeller, A.P., Fielder, W., Berthold, P. (Eds.), Effects of Climate Change on Birds. Oxford University Press, Oxford, pp. 77–85. Whittaker, R.H., 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30, 279–338. Whittaker, R.H., 1975. Communities and Ecosystems. Macmillen, New York. Wiens, J.A., Stralberg, D., Jongsomjit, D., Howell, C.A., Snyder, M.A., 2009. Niches, models, and climate change: assessing the assumptions and uncertainties. In: Proceedings of the National Academy of Sciences USA 106, pp. 19729–19736. Wilby, R.L., Charles, S.P., Zorita, E., Timbal, B., Whetton, P., Mearns, L.O., 2004. Guidelines for use of Climate Scenarios Developed from Statistical Downscaling Methods. IPCC Task Group on Data and Scenario Support for Impact and Climate Analysis (TGICA). (accessed 27.01.11). Williams, J.W., Jackson, S.T., Kutzbacht, J.E., 2007. Projected distributions of novel and disappearing climates by 2100 AD. Proceedings of the National Academy of Sciences USA 104, 5738–5742. Williams, C.L., Hargrove, W.W., Leibman, M., James, D.E., 2008. Agroecoregionalization of Iowa using multivariate geographical clustering. Agriculture, Ecosystems and Environment 123, 161–174.
Contents lists available at ScienceDirect
Biological Conservation journal homepage: www.elsevier.com/locate/biocon
Protected areas in climate space: What will the future bring? John A. Wiens a,b,⇑, Nathaniel E. Seavy a, Dennis Jongsomjit a a b
PRBO Conservation Science, 3820 Cypress Dr. #11, Petaluma, CA 94954, USA School of Plant Biology, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia
a r t i c l e
i n f o
Article history: Received 13 February 2011 Received in revised form 28 April 2011 Accepted 2 May 2011 Available online 26 May 2011 Keywords: California Climate change Climate space Conservation Novel climates Protected areas
a b s t r a c t Protected areas for conservation are intended to contain the environmental conditions that enable species and ecosystems to persist. The locations of such areas are fixed, but the environment within them may change, especially with climate change. To illustrate how multiple climate factors may change in relation to protection status, we used Principal Components Analysis to construct a climate space for California based on eight climate variables assessed at an 800-m resolution. We used projections of future climate derived from a downscaled regional climate model in conjunction with the IPCC SRES A2 scenario to assess how the climate space might shift under future conditions and to identify the combinations of conditions that may no longer occur in the state (disappearing climates) or that will be new to the state (novel climates). Disappearing climates, which were generally toward cooler and/or wetter extremes of the climate space, represented only 0.5% of California’s land area but occurred disproportionately more often in conservation areas that were fully protected. Novel climates (5.8% of California) also occurred disproportionately in fully protected areas; in most cases these climates were characterized by hotter and drier combinations with more seasonal precipitation. The disproportionate occurrence of both novel and disappearing future climates in currently protected areas may create challenges to conservation of the status quo, but such areas may also be ‘‘hotspots of opportunity’’ for responding to the extremes of climate change. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction To live long and prosper, individuals and species require habitat that meets their needs for survival and reproduction – their ecological niche. A major focus of conservation is on protecting or managing places that contain such essential habitat while removing or mitigating other threats. This premise – build it and they will come (or stay) – underlies the efforts of many non-governmental and government entities to set aside protected areas for nature conservation. These efforts have been hugely successful. Globally, some 10–15% of the earth’s surface is under some form of protection (Chape et al., 2005; Soutullo, 2010), and some of the world’s most imperiled species can only be found in such protected areas. The locations of protected areas are geographically fixed. The environment within these areas, however, is not. As awareness of the reality and potential impacts of climate change has increased, it has become apparent that maintaining the conservation value of a protected area may not always be possible (Cole and Yung, 2010). As the climate changes, conditions within a protected area will also be altered, unleashing a cascade of changes in habitats,
⇑ Corresponding author. Address: 7233 NW Valley View Drive, Corvallis, OR 97330, USA. Tel.: +1 703 268 1869; fax: +1 707 765 1685. E-mail address: [email protected] (J.A. Wiens). 0006-3207/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2011.05.002
species, and biological communities. Protecting areas for conservation will increasingly resemble shooting at a moving target. It is important to anticipate how these cascading changes might affect the value of existing protected areas (Griffith et al., 2009). Species-distribution models (SDMs) or habitat suitability models (HSMs) have been used to project the spatial distribution of species or their habitats under future climatic conditions (e.g., Elith and Leathwick, 2009; Stralberg et al., 2009; Thuiller and Münkemüller, 2010) and therefore which species may remain in, leave, or enter protected areas (Kharouba and Kerr, 2010). The premise is that a species can occupy areas that fall within the ‘‘climate space’’ that specifies the combinations of acceptable climatic conditions (e.g., Holdridge, 1947; Whittaker, 1975; Gates, 1980) and that the species’ distributional limits occur where climatic factors fall outside that envelope. Because the climate space is defined independently of geographic locations, the challenge is to determine where combinations of factors in the climate space map onto the locations of existing protected areas in geographical space, both now and in the future. To do this, one would need to know the multivariate set of suitable climatic conditions for each species of interest. Here we take a step back to ask, independently of any single species, how existing protected areas are represented and distributed in the current climate space of a region and how this representation and distribution may change in the future. To illustrate the approach, we
2120
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
construct a climate space for the state of California and determine how that climate space is projected to change in the future. Although there are many ways that these data might be used to describe shifts in climate, we focus on two of the most dramatic by identifying novel climates (portions of climate space that do not presently occur in California but may appear with future climate change) and disappearing climates (portions of the current climate space that may disappear from the state in the future) (c.f. Williams et al., 2007; Ackerly et al., 2010). We then map the geographic locations of these novel and disappearing climates and evaluate the degree to which they may be represented in areas under different levels of existing protection in California. Because effective management of protected areas will require information about potential future conditions in addition to the current conservation value of such places, identifying future novel and disappearing suites of climate conditions may be particularly useful in evaluating how well the current portfolio of protected areas will encompass the full range of future conditions in the state. We limit our analysis to the state of California because this corresponds with the geographic scope of the state agencies and nongovernmental organizations responsible for identifying and planning the protection of conservation areas in the state (as visualized, for example, in State Wildlife Action Plans; Bunn et al., 2007). This focus also aligns with the scope of analysis we have used in previous investigations of shifts in bird distributions within the state (Stralberg et al., 2009; Wiens et al., 2009) and in other studies now underway. California is a tremendously diverse state in terms of topography and climate and it is a regional ‘‘hotspot’’ of biological diversity within North America (Flather et al., 2008; Stein, 2008), so it represents a suitable arena in which to develop and illustrate this approach. In its general features, the approach should be applicable to a variety of other geographies over a range of spatial scales.
2. Methods 2.1. Data 2.1.1. Current and future climate data We used two sources of information to describe the current and future climates of California. Current climate conditions were summarized using 30-year (1971–2000) monthly climate normals interpolated at an 800-m grid resolution by the PRISM Group (Daly et al., 1994, 2000). Future climate conditions were summarized using projections from a regional climate model (RCM), RegCM3 (Pal et al., 2007) at a 30-km resolution with emissions trajectories taken from the Intergovernmental Panel on Climate Change (IPCC) SRES A2 scenario and boundary conditions from the National Center for Atmospheric Research (NCAR) Community Climate System Model (CCSM3.0; Snyder and Sloan, 2005), an atmosphere–ocean global climate model (GCM) run from 1870 to 2099. The RCM was run from 2038 to 2069. For this time period, monthly temperature and precipitation outputs were averaged across years to obtain one set of monthly values. These 30-km resolution results were then statistically downscaled using the change factor method (Wilby et al., 2004) to an 800-m resolution to match the PRISM grid (for details see Stralberg et al., 2009; Wiens et al., 2009). Because our intent was to develop and illustrate the climatespace approach rather than to explore a comprehensive array of possible future climates (as is done in ensemble models; Murphy et al., 2004; Lopez et al., 2009), we used only a single GCM. In comparison with the Geophysical Fluid Dynamics Laboratory model (GDFL CM2.1; Delworth et al., 2006) that we (Stralberg et al., 2009; Wiens et al., 2009) and Ackerly et al. (2010) have used, CCSM3.0 projects a drier and slightly cooler future climate for much of California. The Canadian Centre for Climate Modelling
and Analysis climate model (CCCMA_CGCM3_1.1; see http:// www.ec.gc.ca/ccmac-cccma/default.asp?lang=En&n=1299529F-1, accessed 26 January 2011), also used by Ackerly et al. (2010), projects warmer and wetter conditions than does CCSM3.0 (see http:// meteora.ucsd.edu/cap/cccc_model_prelim.html#kdpdt, accessed 26 January 2011, for model comparisons). For the current and future climate data, we summarized monthly minimum, maximum, and average temperatures and precipitation using eight bioclimatic variables (BIOCLIM; Nix, 1986; Beaumont et al., 2005) that were previously identified as biologically relevant to California wildlife (Stralberg et al., 2009). These were: Annual mean temperature, calculated as the 12-month average of mean monthly temperature. Mean diurnal temperature range, calculated from the 12-month average of the difference between mean maximum and mean minimum temperature for each month. Isothermality, calculated as the ratio of mean diurnal temperature range to the annual temperature range (maximum temperature of the warmest month – minimum temperature of the coldest month). Isothermality compares monthly variability (mean diurnal temperature range) to annual variability (annual temperature range). When monthly variability is greater than annual variability, isothermality is >1; when monthly variability is less than annual variability, isothermality is <1. Temperature seasonality, calculated as the 12-month standard deviation of mean monthly temperature. Mean temperature of warmest quarter, calculated as the average temperature of the warmest 3-month period. Annual precipitation, the 12-month total of mean monthly precipitation. Precipitation seasonality, calculated as the 12-month coefficient of variation of mean monthly precipitation. Precipitation of driest quarter, calculated as the average precipitation for the driest 3-month period. 2.1.2. Definition of protected areas To quantify the geographic distribution of protected areas in California, we used the USGS Gap Analysis Program Protected Area Database for the United States (http://www.protectedlands.net/ images/PADUSTechnical_Report_Oct_2009.pdf, accessed 26 January 2011). This data base identifies four levels of GAP conservation status, which we combined into three levels for analysis: 1 = fully protected (protection of natural resources with natural disturbances either allowed or suppressed, chiefly parks, wilderness areas, wildlife refuges, and nature reserves; Gap status codes 1 and 2); 2 = extractive (protection of natural resources with extractive uses allowed, chiefly state or national forests, multiple-use parks, military reservations, or recreation areas; Gap status code 3); and 3 = unprotected (no mandate for protection of natural resources, chiefly private lands; Gap status code 4). The 764 Fully Protected areas cover 75,987 km2 in California (average size = 9562 ha; range = 0.2–1,068,205 ha); the 347 extractive use areas cover 122,820 km2 (average size = 31,290 ha; range = 0.2–640,867 ha) [comparable data for the unprotected areas, which cover 209,675 km2, cannot be derived because much of this land is in private ownership and parcel sizes are not available]. The geographic distributions of the three categories of protection used in our analyses are shown in Fig. 1. 2.2. Analyses and mapping Although climatic variation in one dimension (e.g., temperature) is relatively easy to conceptualize and display, in actuality climatic variation is multidimensional. Even with only eight climatic variables, the process for conceptualizing and displaying changes in
2121
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
tify pixels as (1) persisting climate: future pixels that were located inside the current climate polygon; (2) disappearing climate: current pixels that were located outside of the future climate polygon; and (3) novel climate: future pixels that were located outside of the current climate polygon. Although it is important to determine which combinations of climate conditions are represented or not represented in the climate space of existing protected areas, conservation planning is framed in terms of geographic space. It is therefore important to consider how novel and disappearing climate spaces translate into real locations on the ground. Accordingly, we mapped the geographic locations of the novel and disappearing climates and compared these locations with those of the protected areas of different protection status. 3. Results 3.1. Current and future California climate space
Fig. 1. Locations of areas of different protection status in California.
this multidimensional space is not straightforward. One approach to this problem has been simply to reduce climate to two of the most important variables and then display shifts in two-dimensional space (e.g., Seavy et al., 2008; Ackerly et al., 2010). This approach is intuitively appealing, but it fails to capture some of the subtle shifts in climatic conditions. An alternative approach is to use multivariate statistics to collapse a suite of climate variables into a smaller set of variables that incorporate as much of the variation as possible (e.g., Comrie and Glenn, 1998; Abatzoglou et al., 2009). For example, multivariate geographical clustering has been used to group agricultural landscapes by multiple environmental factors (Williams et al., 2008) or delineate the boundaries of ecoregions (Hargrove and Hoffman, 1999). To reduce the dimensionality of the eight climatic variables for the current climate, we used Principal Components Analysis (PCA; McGarigal et al., 2000) with climate values for every 800-m pixel in California as inputs to derive the PCA dimensions and factor loadings. The same PCA model was then used to define the future climate space (i.e., we assumed that the relationships among climate variables that determined the PCA dimensions will not change, even though the position of an individual pixel may shift from the current to the future climate space). We used the two axes that explained the majority of variation in the multivariate climate data to portray the climate space. Using these two PCA axes, we plotted each of the 800-m California pixels in the two-dimensional current and future climate spaces. As climate conditions in California change in the future, the climate space as a whole will shift to incorporate new portions of the PCA space that are not currently represented anywhere in the state – these will be the novel climates. Other combinations of climate conditions may disappear entirely – these are the disappearing climates. To assess these changes, we compared areas of overlap and nonoverlap between the current and future climate spaces for California. For the current and future climatic conditions, we used 2-dimensional kernel smoothing to generate a density surface of the climate-space pixels (in PCA space) from which we derived polygons encompassing the areas of the current and future climate space that contained 99.9% of the pixels. We then used these polygons to iden-
Three PCA axes had eigenvalues >1 (PC1 = 3.42, PC2 = 2.40, PC3 = 1.02). The first two axes explained 73% of the variation in the data. We used these two axes to define the multivariate climate space for California. Factor loadings of the eight variables that entered into the analysis (Table 1) suggest that PC1 represents a gradient from warmer, drier areas (negative values) to cooler, wetter areas (positive loadings). PC2 appears to be more strongly associated with seasonality, contrasting areas with a relatively stable temperature and strongly seasonal precipitation (such as coastal areas; positive values) with areas showing greater seasonal variation in temperatures (such as interior mountain and desert areas; negative values) (Fig. 2). The overall climate space is strongly asymmetrical, with cooler, wetter, and either more seasonal temperatures (lower right) or more seasonal precipitation (upper right) conditions predominating (Fig. 3A). 3.2. What will be gained or lost as the climate space shifts? According to the projections of the GCM and RCM we used, future shifts in the California climate space (Fig. 3B) will cause some combinations of climate factors that currently exist to disappear from the state, while other new combinations of climate factors will emerge. Because the bulk of the climate space remains in roughly the same position in the PCA space, these changes occur along the periphery of the climate space. The general trajectory of change is toward lower values of PC1 and somewhat lower values of PC2; therefore, the disappearing climates occur on the trailing edge of this climate-space shift (Fig. 3B). These disappearing climate conditions are generally toward the cooler and wetter (and, to a lesser degree, hotter and drier, and more seasonal) extremes of the climate space. Geographically, the disappearing climates occur in the northern California coast and areas of the Mono Basin, Death Valley, and the southeastern Great Basin (Fig. 4). Despite concerns about the impending loss Table 1 Factor loadings of climate variables on PCA axes 1 and 2. Climate variable
Abbreviation
Factor loadings PC1
Annual mean temperature Mean diurnal range Isothermality Temperature seasonality Mean temperature of warmest quarter Annual precipitation Precipitation seasonality Precipitation of driest quarter
ann temp temp range isotherm temp seas warmest-q temp ann precip precip seas driest-q precip
0.47 0.31 0.05 0.29 0.50 0.43 0.05 0.40
PC2 0.15 0.01 0.56 0.52 0.04 0.03 0.53 0.34
2122
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
Fig. 2. Geographic distributions of scores on the first and second PCA axes.
Fig. 3. Left: distribution of 800-m pixels in the current climate space for California defined by the first and second axes of a Principal Components Analysis of eight bioclimatic variables. Arrows indicate the direction and magnitude of loading of the eight variables on the two PCA axes (abbreviations in Table 1). Right: shifts in the climate space from current to future conditions projected by the climate-change scenario modeled, indicating the distributions of pixels with disappearing and with emerging, novel climate conditions.
of climates and habitats in high-elevation areas of the Sierra Nevada and elsewhere (Dirnböck et al., 2010), few of these climates are projected to disappear entirely (Fig. 4). Emerging, novel climates show quite a different pattern. It is likely that areas characterized as consistently hotter and drier (but with more seasonal precipitation) than now occur will emerge (Fig. 3B). Many of these new climate combinations will appear in the Sonoran and Colorado deserts and in portions of Death Valley and the Mojave Desert. But areas with cooler conditions and greater precipitation in the driest quarter of the year are also likely to
appear. Geographically, these areas will occur in the southern California coast, central California coast ranges, northern California coast, and Klamath Mountains (Fig. 4). 3.3. How are novel and disappearing climates captured by current protected areas? We calculated the area of novel and disappearing climates that occurred in each protected-area category and compared these values to those for California as a whole (Table 2). Areas with
2123
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
Fig. 4. Geographic locations of pixels with disappearing (A) and novel (B) climate conditions, as shown in Fig. 3.
disappearing climates represent only a small area (0.35%) of the state (Fig. 4). If the occurrence of disappearing climates in Fully Protected areas was directly related to the area of the state that is in this protection category (18.6%), we would expect disappearing climates to occur over some 263 km2. Instead, the projected occurrence of disappearing climates in Fully Protected areas is disproportionately great – roughly 2.5 times greater than expected (Table 2). On the other hand, areas subjected to extractive resource use (chiefly State and National Forests and Bureau of Land Management lands) are projected to contain over 3 times less disappearing climate than would be expected on the basis of the proportion of the state in this protection category (30.1%; Table 2). New, novel climates are projected to occur over some 5.3% of the state, with somewhat disproportionate gains in Fully Protected areas (1.22 times greater than expected; Table 2). Unprotected areas, which comprise over half of the lands in California, are projected to contain slightly less area of disappearing climate and of novel climate (1.18 times and 1.11 times less, respectively) than expected. Overall, the observed proportions of disappearing and of novel climate pixels in the three protection categories differ significantly from what would be expected if they were distributed independently of protection category (X 22 = 1330, P < 0.001 and X 22 = 502, P < 0.001, respectively). 4. Discussion Protected areas are fixed in space, but the environments they contain change over time, altering the conservation value of such
areas. Absent episodic large disturbances such as fires, hurricanes, or volcanic eruptions, these changes have in the past been relatively slow, driven by the processes of ecological succession and changing land use in the surroundings. Because these changes have also been localized, they have been amenable to management (e.g., prescribed burning, targeted grazing), allowing the conservation values of protected areas to be retained or restored. Now, however, the pace of environmental change is quickening and threatens to move ecological systems beyond the zone of recent historical environmental variations. These changes are increasingly driven by broad-scale forces such as climate change or economic globalization that lie beyond the reach of local-scale conservation and management. As the distributions of species shift in response to climate change, there will be losses and additions of species to local communities (Peterson et al., 2002; Loarie et al., 2008, 2009), in some situations resulting in ‘‘no-analog’’ assemblages that have no contemporary counterpart (Root and Schneider, 1993; Stralberg et al., 2009). Webs of species interactions will be disrupted and the species that remain will encounter new challenges from species invading from elsewhere. Things will be different. Because these projected changes in species occurrences and communities are driven by future changes in climate, we have taken a step back to ask how the occurrence and distribution of the climate conditions themselves may change over a region under a scenario of future global and regional climate change and what this may mean for the future of protected areas. We use the state of California as an illustration, but the approach should be applicable to other areas at multiple scales.
Table 2 The land area (km2) of California that is classified as novel or disappearing climate in three protection categories. The percentages of the total land area of the state in the protection categories and in disappearing and novel climates are given in parentheses. The area of disappearing or novel climates that would be expected for a protection category given its percentage of the land area of the state as a whole is given in brackets. Climate type
Disappearing Novel All land area
Protection category All California
Fully protected
Extractive use
Unprotected
1 412 (0.35%) 21 630 (5.3%) 408 483
669 [263] 4 920 [4 024] 75 987 (18.6%)
130 [425] 6 733 [6 504] 122 820 (30.1%)
613 [725] 9 978 [11 103] 209 675 (51.3%)
2124
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
In general, the overall range and composition of climate conditions found in California, as represented by the multivariate climate space, may not change much. Although the climate of individual locations will change, the bulk of the climate space represented in Fig. 3B will persist somewhere in the state under the future scenario we have modeled. As the climate space shifts, however, some extreme climate conditions on the periphery of the climate space are likely to disappear, while other new extreme conditions may emerge. Our analysis suggests that such changes in climate are likely to be disproportionately great in areas that are currently most fully protected – National Parks, Wilderness Areas, National Wildlife Refuges, State Parks, and the like. Nationally, many of these protected areas already occur in extreme environments, at high elevations or in places with low productivity, what Scott et al. (2001) have labeled ‘‘rocks and ice.’’ The same is true internationally – fully protected areas are often in places that are less exploitable, more remote, more rugged, and more extreme environmentally than the places where people live and work, predisposing them to be more vulnerable to climate changes. California is no different – think of Death Valley or the iconic National Parks along the crest of the Sierra Nevada, places that are already near the fringes of the current climate space. If species with small ranges disproportionately occupy places with ‘‘rare climates’’ (i.e., at the periphery of the climate space), as suggested by Ohlemüller et al. (2008), they also may be particularly vulnerable to climate change. It is reassuring that only a small proportion of the California climate space is projected to disappear. Because the forces that determine shifts in the climate space are beyond the reach of management, however, nothing can be done to forestall the disappearance of some climate conditions and the consequences of those changes. Simply because some combinations of climate conditions may no longer occur in the state, however, does not mean that the species that presently occupy those areas will also disappear. Because the shifts in climate conditions are dictated by how climate features are interrelated in the multivariate climate space, species whose bioclimatic niche envelopes are determined by only a subset of those features may continue to find suitable conditions in those areas. For other species, the disappearance of some combinations of climate features may affect only a portion of their distribution, so they may continue to occur elsewhere in the state where their requirements are met. In some situations, however, disappearing climates may be associated with the extinction of local populations, as has been projected for American pika (Ochotona princeps; Grayson, 2005; but see Millar and Westfall, 2010). The emergence of novel climates, on the other hand, will create settings for the development of assemblages of species that are also new to California. Some species that currently occur in these areas may remain, others may immigrate from elsewhere in the state, and yet others that do not presently occur in the state may colonize from surrounding regions as distributions shift in response to climate change. Such ‘‘no-analog’’ assemblages may emerge in many areas (Overpeck et al., 1992; Williams et al., 2007; Stralberg et al., 2009), but they may be especially likely where novel climates appear. In a sense, the locations of these emerging novel climates represent future ‘‘hotspots of opportunity.’’ How biotic communities assemble in these areas will depend on a host of factors – whether the conditions of the novel climates fall within the bioclimatic envelopes of species, the dispersal abilities of the species, intervening land uses that impede or promote movements, the presence of suitable habitat and resources, interactions with other species, and so on. In these places, management may need to be directed toward species not considered in current management plans. Alternatively, if the climate of a protected area disappears completely, management of species associated with these conditions might be deemphasized. It will be important to
conduct comprehensive monitoring to detect early signals of changes in distributions and abundances of species; this might most effectively be done as part of an Environmental Change Network (see http://www.ecn.ac.uk/, accessed 12 February 2011). Because Fully Protected areas are projected to experience both disproportionate losses of existing climates as well as disproportionate gains in novel climates, they may represent areas of particularly turbulent future climate conditions. Under these conditions, future management and conservation may need to be especially nimble. Shifting the management focus from individual species to broader assemblages or ecosystem properties, together with using adaptive management in an anticipatory rather than a reactive mode, may be necessary (Lawler et al., 2010). The climates that we project as disappearing from or emerging in California are not uniquely Californian, of course. These conditions may also occur, both now and in the future, in other areas of the Pacific Northwest (disappearing) and Desert Southwest (emerging). It is perhaps no accident that the locations of greatest change in the climate space occur about the fringes of the state (Fig. 4), largely as parts of ecoregions that extend into adjacent states (Miles and Goudey, 1997). Although the truncation of our analysis at the state boundaries is thus somewhat artificial, it is nonetheless pragmatic. The California Wildlife Action Plan (Bunn et al., 2007), produced in response to a Congressional mandate to develop plans for the conservation of wildlife and habitats, is intended to guide conservation and management of species at risk. The California Essential Connectivity Project (Spencer et al., 2010) projects potential land acquisitions and management to enhance linkages among protected areas under current conditions, and this report is also intended to influence the priorities of state agencies as well as environmental organizations. Our analysis is intended primarily to illustrate how a consideration of the current and potential future multivariate climate space for a region may provide insights about the possible effects of climate change on areas protected for wildlife management, wilderness, and conservation. To simplify the analysis, we have used only a single climate-change scenario, one RCM, and a selected set of bioclimatic variables from a broader array. Therefore, the results should not be taken as definitive or predictive, but illustrative of the ways in which climate conditions may shift in the future, how these shifts map onto geographical space, and how the extreme changes (disappearing and novel climates) may differentially affect areas of differing protection status. Clearly, however, changing climatic conditions pose new and challenging questions about the current network of protected areas. To date, strategies for adapting conservation plans to climate change have focused on enhancing resistance or resilience, while preparing for ecological transformation has received less attention (Heller and Zavaleta, 2009; Poiani et al., 2011). In other words, current strategies focus on ways to maintain what is here now rather than how to prepare for what will arrive in the future. The disappearance of current climatic conditions and the emergence of novel climatic conditions in California’s protected areas illustrate that management must address ecological transformation. After all, climate (together with substrate and topography) is a primary driver of the potential distributions of plants and animals, and thence of biodiversity (Whittaker, 1960).
Acknowledgments This paper benefitted from the advice of Grant Ballard, Chrissy Howell, Diana Stralberg, Sam Veloz, and two anonymous reviewers. Mark Snyder provided the data on future climates. Partial funding was provided by the USFWS California Landscape Conservation Cooperative. This is PRBO Publication No. 1810.
J.A. Wiens et al. / Biological Conservation 144 (2011) 2119–2125
References Abatzoglou, J.T., Redmond, K.T., Edwards, L.M., 2009. Classification of regional climate variability in the State of California. Journal of Applied Meteorology and Climatology 48, 1527–1541. Ackerly, D.D., Loarie, S.R., Cornwell, W.K., Weiss, S.B., Hamilton, H., Branciforte, R., Kraft, N.J.B., 2010. The geography of climate change: implications for conservation biogeography. Diversity and Distributions 16, 476–487. Beaumont, L.J., Hughes, L., Poulsen, M., 2005. Predicting species distributions: use of climatic parameters in BIOCLIM and its impact on predictions of species’ current and future distributions. Ecological Modeling 186, 251–270. Bunn, D., Mummert, A., Hoshovsky, M., Gilardi, K., Shanks, S., 2007. California Wildlife: Conservation Challenges. University of California Davis Wildlife Health Center, California Department of Fish and Game, Sacramento, CA, USA. Chape, S., Harrison, J., Spalding, M., Lysenko, I., 2005. Measuring the extent and effectiveness of protected areas as an indicator for meeting global biodiversity targets. Philosophical Transactions of the Royal Society B 360, 443–455. Cole, D.N., Yung, L., (Eds.), 2010. Beyond Naturalness: Rethinking Park and Wilderness Stewardship in an Era of Rapid Change. Island Press, Washington, DC. Comrie, A.C., Glenn, E.C., 1998. Principal components-based regionalization of precipitation regimes across the southwest United States and northern Mexico, with an application to monsoon precipitation variability. Climate Research 10, 201–215. Daly, C., Neilson, R.P., Phillips, D.L., 1994. A statistical topographic model for mapping climatological precipitation over mountainous terrain. Journal of Applied Meteorology 33, 140–158. Daly, C., Gibson, W.P., Hannaway, D., Taylor, G., 2000. High-quality spatial climate data sets for the United States and beyond. Transactions of the American Society of Agricultural Engineers 43, 1957–1962. Delworth, T.L., Broccoli, A.J., Rosati, A., Stouffer, R.J., Balaji, V., Beesley, J.A., Cooke, W.F., Dixon, K.W., Dunne, J., Dunne, K.A., Durachta, J.W., Findell, K.L., Ginoux, P., Gnanadesikan, A., Gordon, C.T., Griffies, S.M., Gudgel, R., Harrison, M.J., Held, I.M., Hemler, R.S., Horowitz, L.W., Klein, S.A., Knutson, T.R., Kushner, P.J., Langenhorst, A.R., Lee, H.-C., Lin, S.-J., Lu, J., Malyshev, S.L., Milly, P.C.D., Ramaswamy, V., Russell, J., Schwarzkopf, M.D., Shevliakova, E., Sirutis, J.J., Spelman, M.J., Stern, W.F., Winton, M., Wittenberg, A.T., Wyman, B., Zeng, F., Zhangc, R., 2006. CM2 global coupled climate models. Part I: Formulation and simulation characteristics. Journal of Climate 19, 643–674. Dirnböck, T., Essel, F., Rabitsch, W., 2010. Disproportional risk for habitat loss of high-altitude endemic species under climate change. Global Change Biology 17, 990–996. Elith, J., Leathwick, J., 2009. Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics 40, 677–697. Flather, C.H., Knowles, M.S., McNees, J., 2008. Geographic Patterns of at-risk Species: a Technical Document Supporting the USDA Forest Service Interim Update of the 2000 RPA Assessment. General Technical Report RMRS-GTR-211. US Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO. Gates, D.M., 1980. Biophysical Ecology. Springer-Verlag, New York. Grayson, D.K., 2005. A brief history of Great Basin pikas. Journal of Biogeography 32, 2103–2111. Griffith, B., Scott, J.M., Adamcik, R., Ashe, D., Czech, B., Fischman, R., Gonzalez, P., Lawler, J., McGuire, A.D., Pidgorna, A., 2009. Climate change adaptation for the US National Wildlife Refuge System. Environmental Management 44, 1043–1052. Hargrove, W.W., Hoffman, F.N., 1999. Using multivariate clustering to characterize ecoregion borders. Computing in Science & Engineering 1, 18–25. Heller, N.E., Zavaleta, E.S., 2009. Biodiversity management in the face of climate change: a review of 22 years of recommendations. Biological Conservation 142, 14–32. Holdridge, L.R., 1947. Determination of world plant formations from simple climatic data. Science 105, 367–368. Kharouba, H.M., Kerr, J.T., 2010. Just passing through: global change and the conservation of biodiversity. Biological Conservation 143, 1094–1101. Lawler, J.J., Tear, T.H., Pyke, C., Shaw, M.R., Gonzalez, P., Kareiva, P., Hansen, L., Hannah, L., Klausmeyer, K., Aldous, A., Bienz, C., Pearsall, S., 2010. Resource management in a changing and uncertain climate. Frontiers in Ecology and the Environment 8, 35–43. Loarie, S.R., Carter, B.E., Hayhoe, K., McMahon, S., Moe, R., Knight, C.A., Ackerly, D.D., 2008. Climate change and the future of California’s endemic flora. PLoS ONE 3, e2502. doi:10.1371/journal.pone.0002502. Loarie, S.R., Duffy, P.B., Hamilton, H., Asner, G.P., Field, C.B., Ackerly, D.D., 2009. The velocity of climate change. Nature 462, 1052–1055. Lopez, A., Fung, F., New, M., Watts, G., Weston, A., Wilby, R.L., 2009. From climate model ensembles to climate change impacts and adaptation: a case study of
2125
water resource management in the southwest of England. Water Resources Research 45. doi:10.1029/2008WR007499. McGarigal, K., Cushman, S., Stafford, S., 2000. Multivariate Statistics for Wildlife and Ecology Research. Springer, New York. Miles, S.R., Goudey, C.B., 1997. Ecological Subregions of California. USDA Forest Service Pacific Southwest Region Report R5-EM-TP-005, San Francisco, CA. Millar, C.I., Westfall, R.D., 2010. Distribution and climatic relationships of the American pika (Ochotona princeps) in the Sierra Nevada and Western Great Basin, USA: periglacial landforms as refugia in warming climates. Arctic, Antarctic, and Alpine Research 42, 76–88. Murphy, J.M., Sexton, D.M.H., Barnett, D.N., Jones, G.S., Webb, M.J., Collins, M., Stainforth, D.A., 2004. Quantification of modelling uncertainties in a large ensemble of climate change simulations. Nature 430, 768–772. Nix, H.A., 1986. A biogeographic analysis of Australian elapid snakes. In: Longmore, R. (Ed.), Atlas of Elapid Snakes of Australia. Australian Flora and Fauna Series 7. Australian Government Publishing Service, Canberra, pp. 4–15. Ohlemüller, R., Anderson, B.J., Araújo, M.B., Butchart, S.H.M., Kudrna, O., Ridgely, R.S., Thomas, C.D., 2008. The coincidence of climatic and species rarity: high risk to small-range species from climate change. Biology Letters 4, 568–572. Overpeck, J.T., Webb, R.S., Webb III, T., 1992. Mapping eastern North American vegetation change of the past 18 ka: no-analogs and the future. Geology 20, 1071–1074. Pal, J.S., Giorgi, F., Bi, X., Elguindi, N., Solmon, F., Gao, X., Rauscher, S.A., Francisco, R., Zakey, A., Winter, J., Ashfaq, M., Syed, F.S., Bell, J.L., Diffenbaugh, N.S., Karmacharya, J., Konare, A., Martinez, D., da Rocha, R.P., Sloan, L.C., Steiner, A., 2007. Regional climate modeling for the developing world: the ICTP RegCM3 and RegCNET. Bulletin of the American Meteorological Society 88, 1395–1409. Peterson, A.T., Ortega-Huerta, M.A., Bartley, J., Sanchez-Cordero, V., Soberón, J., Buddemeier, R.H., Stockwell, D.R.B., 2002. Future projections for Mexican faunas under global climate change scenarios. Nature 41, 626–629. Poiani, K.A., Goldman, R.L., Hobson, J., Hoekstra, J.M., Nelson, K.S., 2011. Redesigning biodiversity conservation projects for climate change: examples from the field. Biodiversity and Conservation 20, 185–201. Root, T.L., Schneider, S.H., 1993. Can large-scale climatic models be linked with multiscale ecological studies? Conservation Biology 7, 256–270. Scott, J.M., Abbitt, R.J.F., Groves, C.R., 2001. What are we protecting? Conservation Biology in Practice 2, 18–19. Seavy, N.E., Dybala, K.E., Synder, M.A., 2008. Climate models and ornithology. Auk 125, 1–10. Snyder, M.A., Sloan, L.C., 2005. Transient future climate over the western United States using a regional climate model. Earth Interactions 9, 1–21. Soutullo, A., 2010. Extent of the global network of terrestrial protected areas. Conservation Biology 24, 362–365. Spencer, W.D., Beier, P., Penrod, K., Winters, K., Paulman, C., Rustigian-Romsos, H., Strittholt, J., Parisi, M., Pettler, A., 2010. California Essential Habitat Connectivity Project: A Strategy for Conserving a Connected California. Prepared for California Department of Transportation, California Department of Fish and Game, and Federal Highways Administration.