protected areas and climate change

Protected Areas and Climate ChangeLee Hannah

Conservation International, Center for Applied Biodiversity Science,Arlington, Virginia, USA

The study of protected areas and climate change has now spanned two decades. Pioneer-ing work in the late 1980s recognized the potential implications of shifting species rangeboundaries for static protected areas. Many early recommendations for protected areadesign were general, emphasizing larger protected areas, buffer zones, and connectivitybetween reserves. There were limited practical tests of these suggestions. Developmentof modeling and conservation planning methods in the 1990s allowed more rigoroustesting of concepts of reserve and connectivity function in a changing climate. Thesestudies have shown decreasing species representation in existing reserves due to cli-mate change, and the ability of new protected areas to help slow loss of representationin mid-century scenarios. Connectivity on protected area periphery seems more effec-tive than corridors linking protected areas. However, corridors serving other purposes,such as large carnivore movement, may be useful for accommodating species rangeshifts as well. Assisted migration and ex situ management strategies to complementprotected areas are being explored. Finally, in scenarios of the latter half of the cen-tury, protected areas and connectivity become increasingly expensive and decreasinglyeffective, indicating the importance of reducing human-induced climate change.

Key words: protected areas; climate change; biodiversity; reserves; connectivity; as-sisted migration

Introduction

Understanding of the role of protected ar-eas as climate changes began in the late 1980swith a series of landmark journal articles andbook chapters by Rob Peters, Thomas E.Lovejoy, and others (Peters & Darling 1985;Peters & Myers 1991; Peters & Lovejoy 1992).These authors pointed out that paleoecolog-ical evidence shows that range shifts werea predominant species response to climatechange, suggesting that anthropogenic climatechange would be likely to result in changingrelationships between (dynamic) species rangelimits and (fixed) protected area boundaries.Peters illustrated this concept graphically in afigure that is reproduced below as Figure 1.This visual depiction has shaped much of the

Address for correspondence: Lee Hannah, Conservation International,Center for Applied Biodiversity Science, 2011 Crystal Drive, Arlington,VA 22202. [email protected]

subsequent debate about protected areas andclimate change because it incorporates the ma-jor critical factors—mobile species range limits,fixed reserve boundaries, and land-use change.

In the decade following Peters’s pioneeringwork, a small but growing number of authorsaddressed the impacts of climate change onprotected areas. Most of these contributionswere assessments of likely vegetation change orshifting species composition at global, national,and regional scales. Early modeling studies ofspecies range shifts using software tools, suchas BIOCLIM (Nix 1986) and DOMAIN (Car-penter et al. 1993) confirmed that changes inrange boundaries are likely to affect spatial re-lationships between species and reserves. Forinstance, the range boundaries of some Euro-pean plants have been projected to shift of tensor hundreds of kilometers in response to climatechange (Huntley et al. 1995).

The early conservation and climate cha-nge literature made recommendations about

Ann. N.Y. Acad. Sci. 1134: 201–212 (2008). C© 2008 New York Academy of Sciences.doi: 10.1196/annals.1439.009 201

202 Annals of the New York Academy of Sciences

Figure 1. Illustration of relationships betweenspecies range limits, reserve boundaries, and landuse change (from Peters & Lovejoy 1992). Thehatched area in panel A represents the range limitof a species. The fragmentation of the hatched ar-eas in panel B represents near-term habitat loss dueto land-use change. Panel C illustrates the combinedeffect of habitat fragmentation and climate change–induced range shift on representation of the species’range within the reserve. (Note: This figure illustratesthe loss of the reserve as the species’ range retreats,which is unlikely in response to changes in only onespecies; the reserve would remain important for manyspecies whose ranges don’t shift, as well as for othersthat may experience range shifts that bring them intothe reserve.)

protected areas, but few studies actively testedthese notions. A useful review of the field(Halpin 1997) found five broad categories ofmanagement recommendation, which the au-thor judged to be generic and in need oftesting. These five protected area prescrip-tions were: redundant reserves; larger reserves;buffer zones around protected areas; land-scape connectivity; and management of ex-isting threats and disturbance regimes. Theearly calls for increased reserve area have beenborne out in recent modeling studies. The term“buffer zone” has been partially displaced bydiscussions of connectivity and managementof the matrix in the recent conservation lit-erature, but the concept of multiple-use land-scapes that permit both human use and species’responses to climate change remains current.Limiting other threats, such as habitat loss, isa major theme in climate change conservationwritings today.

The late 1990s saw developments in con-servation planning, species distribution mod-eling, and climate model downscaling that al-lowed more rigorous testing of the early recom-mendations. Reserve selection algorithms ca-pable of testing the relative effectiveness of al-ternative protected area configurations becamecommonly available (Pressey & Cowling 2001).Models of species range shifts due to climatechange moved from early envelope models,such as BIOCLIM, to better-performing statis-tical tools, such as Generalized Additive Model-ing and MaxEntropy (Guisan & Zimmermann2000; Elith et al. 2006). Finer-grain climate pro-jections of 1–50 km were created for countrystudies, conservation assessments, and agricul-tural studies (Root & Schneider 1995). Thesefine-scale projections have been made widelyavailable for biological modeling (see for in-stance www.worldclim.org). The convergenceof improved modeling tools, reserve selectionalgorithms, and fine-scale climate projectionscombined to facilitate direct investigation ofthe effectiveness of protected areas and cor-ridors in a modeled environment. A series ofinvestigations in several parts of the world have

Hannah: Protected Areas and Climate Change 203

directly tested the relevance of protected areasize, connectivity, and spatial configuration toconserving species whose ranges were shiftingdue to climate change.

These investigations have shown that pro-tected areas can be an effective response to cli-mate change–driven range dynamics, that lim-iting human-induced climate change is a crit-ical counterpart to additional protected area,and that connectivity is essential for somespecies even in the early stages of anthropogenicclimate change. Findings to date suggest ele-ments important in the design of individualreserves, particularly for marine reserves inresponse to coral bleaching. Finally, new re-sponses, such as temporally and spatially flex-ible reserves and assisted migration are nowreceiving attention.

This review summarizes this decade ofprogress, examining findings relating to designof:

1) Protected area networks2) Connectivity and corridors3) Individual protected areas4) Mobile reserves5) Measures needed when protection alone

is not enough

Each section will summarize the peer-reviewed literature on the topic and mention se-lect agency literature contributions. Challengescurrently facing the field and in need of futureinvestigation are highlighted. The review con-cludes with comments on the degree to whichthese findings are being, and can be, integratedinto conservation practice.

Design of Protected Area Networks

The first conceptual element to be testedwith models is the prediction that climatechange might alter the species composition ofreserves. Multispecies tests targeted at examin-ing climate change effects on levels of species’representation in reserve networks were firstconducted in South Africa (Rutherford et al.

1999) and Europe (Araujo et al. 2004). Repre-sentation is the amount of area of each species’range included within protected areas, a mea-sure commonly used in conservation planningto assess adequacy of protection. These andother studies (e.g., Pearson & Dawson 2005)indicate that range shifts result in declines inspecies representation in some reserves, but, de-pending on assumptions about species’ disper-sal abilities, may result in increases in represen-tation in other reserves. The net effect of move-ments on representation in individual reservesand across whole networks depends strongly ondispersal abilities and intervening land uses thatmay impede dispersal (Pearson 2006).

There is reason to believe that the net effectof climate change–induced range shifts will bereduction in ranges, which has implications forconservation and risk of extinction. Warmingwill cause species to migrate upslope and to-ward the poles to track suitable climatic condi-tions (Walther et al. 2002). Mountains have lessarea at higher elevation due to their geometry,while in the southern hemisphere, continentshave less land area toward the poles due totheir shape. While northern hemisphere landarea increases poleward, species richness andendemism is lower in these areas due to Pleis-tocene glaciation.

Thomas and colleagues (2004) estimated theextinction risk associated with range shifts in sixgeographically diverse regions. They found thatextinction risk increased in all cases, even un-der full dispersal assumptions, which suggeststhat there is an overall net trend toward smallerspecies ranges with climate change. Such a netreduction in average range size would favor netloss of representation in reserves. Modeling re-sults examining this question directly have con-firmed net reduction in species representationwith climate change for specific regions (e.g.,Hannah et al. 2005).

The creation of new reserves is a logical stepto compensate for this loss of representation(Villers-Ruiz & Trejo-Vazquez 1998; Hannahet al. 2002; Hossell et al. 2003; Anderson &Martinez-Meyer 2004; Coulston & Riitters


2005). Additional protected areas have beensuggested to compensate for loss of habitatrepresentation in Canadian parks, where evengross vegetation type may be altered by climatechange (Scott et al. 2002). Creation of new andlarger reserves has been proposed as a responsein marine systems because of their large ex-tent and high degree of interconnectivity (Soto2001). Improved coverage of important plantareas has been suggested in Africa and for dryforest worldwide based on modeling (McCleanet al. 2006; Miles et al. 2006).

One method for systematically adding pro-tected areas for climate change is to link speciesdistribution models (SDMs) and reserve selec-tion algorithms. The SDM is used to simulatethe present and future ranges of a species, al-lowing the gain or loss of area protected tobe calculated. Reserves are added to compen-sate for lost representation, using the algorithm,until no further improvement in representa-tion can be achieved in the future (climatechange) scenario. The improvement in repre-sentation achievable with additional protectedareas can then be measured. The benchmarkagainst which changes are gauged is a targetlevel of representation (Pressey et al. 2007). Inpractice the amount of area to be protectedshould be tailored to each species, dependingon their population sizes and habitat require-ments. However, in these studies a simplifyingassumption is made that all species must meeta minimal representation (area of species rangeprotected) target.

An alternative response is to improve theclimatic representation of a reserve network(Pyke et al. 2005; Pyke & Fischer 2005). Thismethod adds area containing climates poorlyrepresented in the reserve system. Reservesare added incrementally until climatic repre-sentation goals are met. Climatic representa-tion is calculated using the climatic tolerancesof target species, which is important becausespecies response is the product of the amountof change in a particular climatic variable (e.g.,precipitation) and species sensitivity to changein that variable (Pyke & Fischer 2005). The end

result of the climatic representation methodmay therefore be dependent on the numberof species sampled, a trait it shares with speciesrepresentation approaches.

No systematic test has been conducted com-paring the climate-representation and species-representation approaches. Both approachesgreatly increase the chance of all species be-ing represented in a reserve network as climatechanges, while only the species-representationapproach can meet specific targets for the areaunder protection for each species. Since vi-able populations depend strongly on area con-served, the ability to set and meet area targetsfor priority species is a significant advantage.

The need for global coordination and ex-pansion of protected areas to deal with climatechange has been recognized (Hannah 2001).A key need is regionally or globally acceptedrepresentation targets for species. These wouldprovide objective benchmarks, allowing new ar-eas to be added to maintain targets as climatechanges. Not all species are represented in pro-tected areas in their current ranges, so an effortis needed to complete protected areas coverageglobally. This provides an opportunity to meetclimate-change and current protection needssimultaneously. A system of standards and time-lines for completion of global coverage is there-fore a key element in a global system of pro-tected areas. Such a system would best functionin a decentralized, voluntary fashion (Hannah2001).

A crucial question about adding protectedareas is whether sufficient viable range remainsin future climates for new protection to makea difference. This issue has recently been ad-dressed in a study which employed SDM andreserve selection algorithms in three regions(Hannah et al. 2007). This work built on themultispecies SDM efforts available in the CapeFloristic Region (Midgley et al. 2002), Mex-ico (Peterson et al. 2002), and Europe (Araujoet al. 2004) under climate changes projectedfor 2050. New protected areas were able tosubstantially restore species representation lostdue to climate change in 2050 scenarios in all


three regions. New protected areas brought thenumber of species meeting the representationtarget to 78% (Cape), 89% (Europe), and 94%(Mexico). These results strongly confirmed therelevance of protected areas as a conservationresponse to climate change.

Strong performance of new protected areasin these midcentury scenarios does not indicatethat protected areas are a panacea for climatechange. On the contrary, the fact that up to20% of species lost representation even withnew protected areas in the 2050 scenarios isan indication that more severe climate changelater in the century is highly likely to result insevere loss of representation. Variation betweenregions was high, indicating that performanceof new protected areas is dependent on topog-raphy, characteristics of specific taxa, and otherfactors. In this study, the Cape and Europe anal-yses used plant taxa. In Mexico, vertebrate taxawere used and the protected area needs werefar lower. Dispersal limits chosen by expertsfamiliar with the species were included in thesimulation, so the lower protected area need inMexico may be related to the relatively higherdispersal abilities of vertebrates. Reanalysis ofMexico using plant taxa might show new pro-tected area needs more similar to those foundfor the Cape and Europe.

Finally, early implementation of new pro-tected areas for climate change paid dividendsin all three regions. Area savings of 39–96%were realized when climate change and com-pleting representation of current ranges wereaddressed at once. This is because protectedarea networks are not perfect in representationof current ranges—new protected areas mustbe added to meet a uniform representation goalfor current ranges—and there are choices to bemade as to which areas to add. If these choicesare made in ways that favor areas needed tocompensate for climate change, a head starton climate-change solutions is gained, reduc-ing the amount of area needed to compensatefor climate change losses later on. Even if antic-ipation of climate change is imperfect, explicitattempts to address climate change in selec-

tion of new protected area for current rangeswill result in some area savings (Hannah et al.

2007).

Connectivity and Corridors

Early concepts of connectivity and corridorsassumed that species moving in response to cli-mate change would need pathways betweenone protected area that they were leaving andanother, destination protected area. This sim-ple conceptualization belied the complexity ofmultiple species, all moving simultaneously inresponse to climate change. The individualisticnature of species’ responses to climate changeis widely recognized in paleoecology, suggest-ing that future range changes will also findspecies moving each in their own characteris-tic way, depending on species-specific climatictolerances. Such individualistic movements arenot likely to align with one another closely or bereadily captured in linear source-to-destinationcorridors.

The first modeling test of connectivity for cli-mate change supported the individualistic view(Williams et al. 2005). This study used timestepmodels of plants (proteas) in the Cape regionof South Africa to find “chains” of connectiv-ity that would track suitable climate for mov-ing species. SDMs were produced for 315 pro-tea species each decade, from the present to2050, in planning unit grids with dimensions ofjust under 2 km. A reserve selection algorithmwas modified to search for chains of contigu-ous planning units that maintained suitable cli-mate in all timesteps (Fig. 2). The shortest chainwas one grid cell that maintained suitable cli-mate from present to 2050. For species whoseranges moved considerably, 5–6 grid chainsmight be required. An estimate of dispersal lim-itation for each species excluded distant plan-ning units from consideration. The algorithmselected chains within existing protected areasfirst, then selected new areas complementaryto existing parks and reserves until the repre-sentation target was met. This final network of


Figure 2. (A) Method for searching for “chains” of connectivity among planning units. When repeatedover an entire protected area network, this procedure provides a basis for adding protected areas that willhelp accommodate range shifts due to climate change. In practice, searching backwards in time reducesnonproductive searches and search time, since it avoids “dead-end” chains that do not persist in the future(for instance the top row of cells in the figure). (From Williams et al. 2005.) (B) Example of additional areasselected by such a method for the Cape region of South Africa. Light green indicates existing protected areas,orange indicates protected area additions for climate change based on chains analysis. Light grey is landtransformed to human use (e.g., agriculture), blue is ocean to the south and west, and non-fynbos vegetationto the North east. (Figure courtesy of Steven Phillips).

existing and new protected areas provided thebest possible protection for all species as climatechanges.

The most cost-effective connectivity linked acurrent protected area with unprotected habi-tat on its periphery. Few species actually movedfrom one protected area to another, and these

long-distance links would seldom be selectedbecause shorter, less area-intensive links existedfor most species. For smaller range shifts, cellsthat retained suitable climate could often befound within existing protected areas, or addedto the network at relatively smaller cost. Manyrange shifts, both large and small, could be


accommodated by chains that lay entirelywithin existing protected areas.

The effectiveness of new protected lands toprovide connectivity was eroding by 2050, asevidenced by the species with no suitable cli-mate or need for long chains with little or nooverlap with other chains. Further investigationis needed in this and other regions to determinethe capacity for connectivity to compensate forthe effects of climate change later in the century.It appears to be highly likely that connectivityand new protected areas will fail to maintainprotection for many species past 2050 in atleast some regions. This suggests that the next50 years is the key timeframe in which climatechange must be stopped to avoid large negativeconsequences for biodiversity.

The approach of Williams et al. (2005)has now been improved and modified usinglinear programming optimization techniques(Phillips, in press). Commercial linear program-ming softwares have been developed for in-dustrial optimization problems such as airlinerouting. They provide better optimization re-sults than the algorithms used in most con-servation planning software, but are less oftenused in conservation due to the cost of the soft-ware and the longer time required to find asolution. These commercial optimization soft-wares have now been applied to the climate-change reserve-selection problem. Early resultsof these analyses suggest that connectivity maybe achieved with even less area than in so-lutions developed with reserve-selection algo-rithms. The results do not, however, show muchevidence of improved capacity to maintain rep-resentation past 2050, as the erosion of effec-tiveness in later timesteps is driven by shrinkingsuitable climate space for species rather than byproblems of efficiency.

Design of IndividualProtected Areas

A strong literature is evolving on the designof protected areas for climate change, particu-

larly with reference to marine protected areas(MPAs). Recommendations emphasize designconcepts that promote either resilience or re-sistance to climate change (Salm et al. 2001;West & Salm 2003). Resistance is the abilityof the ecosystems within a protected area towithstand climate change without substantialbiological change. Resilience is the capacity ofecosystems (or species) to recover to a startingor reference condition after being damaged byclimate change.

The climate change impact most addressedin the protected area literature is coral bleach-ing (West & Salm 2003). This emphasis arisesbecause coral bleaching is an immediate threatfrom climate change to marine reserves (Salmet al. 2001). Bleaching has already affected mostof the world’s coral reefs and is expected togrow in severity as climate change accelerates(Hoegh-Guldberg 1999). MPAs may have animportant role in slowing this process and en-suring that healthy coral reefs persist in some re-gions, with important implications for tourismand local fisheries.

Resistance to coral bleaching can be con-ferred by cool upwellings and by sites in whichsedimentation and exposure to light are min-imized (West & Salm 2003). Design elementsthat favor incorporation of these features there-fore improve resistance within an MPA. Un-damaged reefs in these resistant areas may serveas important sources of recolonization for areasdamaged in bleaching events. However, undermore extreme climate change, the combinationof high water temperature and increasing acid-ity from the dissolution of atmospheric CO2 inseawater may eradicate coral reefs on regionalor global scales (Hoegh-Guldberg et al. 2007),outstripping the moderating effects of up-welling and making recolonization irrelevant.

Resilience is a biological property and ismore difficult than resistance to incorporateinto reserve design. Nonetheless, understand-ing of resilience is improving, and applica-tions, particularly in MPAs, are evolving (VanOppen & Gates 2006). Some species andsome populations recover more quickly after


bleaching. Some coral reef ecosystems aremore resilient than others (Bellwood et al.

2006). These resilient populations or ecosys-tems may be targeted in reserve design, for in-stance by placing boundaries to maximize areaof resilient populations. Design for resiliencemay bias species composition (toward quicker-recovering species), so may have to be balancedagainst species representation goals. Zoning toreduce tourism use post-bleaching may reducephysical contact and damage to reefs, whichmay enhance recovery (Salm et al. 2001).

Sea-level rise is a second major concern ad-dressed in the MPA design literature. Risingsea level over the coming centuries will al-ter the proportion of habitat types representedin MPAs. For example, some reserves in theChannel Islands of California will lose sea grasshabitat as sea level rises, since shelf areas thatcurrently harbor extensive sea grass beds willbecome submerged too deeply to support seagrass in the future. Individual sites can be se-lected or designed to incorporate areas thatwill lose less sea grass habitat under projectedchanges in bathymetry.

The terrestrial literature on reserve designfor climate change is less well developed, butseveral important design considerations havebeen identified. Halpin (1997) has pointed outthat changes in montane vegetation zones withwarming are dependent on the topography ofspecific mountain chains. This effect parallelsin principle the changes in marine habitat dueto sea-level rise. Reserve configuration can helpcompensate for these changes. For example, us-ing a watershed divide as a reserve boundarymay be a good choice for ease of demarca-tion, but for climate change a better choiceis to include high-elevation habitats on bothsides of the divide, since these will be pushedupslope and consequently reduced in area bywarming—setting a reserve limit along the di-vide would exclude half of these diminishinghabitats from protection.

Sea-level rise is a concern in terrestrial re-serve design as well as in marine settings.Coastal reserves will often be squeezed between

rising sea level and hard boundaries of inlandhuman land use (Shirley & Battaglia 2006). Re-serve design can help resolve this evolving con-flict. For example, in the Albemarle Sound ofNorth Carolina, acquisition of inland land isunderway to allow migration of coastal com-munities as salt-water intrusion erodes peat soils(Pearsall 2005).

Protected area responses to ecosystemthreshold changes is of growing concern. Vege-tation type changes may be compensated for byadding protected area to compensate for vege-tation types declining in protection. For exam-ple, in Canada, redesign of the entire protectedarea network has been recommended as bo-real forests become rarer and temperate forestsmore common. Increased protection in the farnorth may be necessary to ensure the conserva-tion of boreal forests that were formerly abun-dantly represented in protected areas furthersouth (Scott & Suffling 2000). Other thresholdchanges may prove even more challenging. Pinebark beetle outbreaks are decimating dominanttree species across large areas of western NorthAmerica and, due to climate change, are pen-etrating to latitudes and elevations previouslyimmune to outbreak. Conservation responsesto these events have been little examined, butprotection of resistant stands may be critical inlong-term conservation of these species.

Reserve management and design recom-mendations for major ecosystem types havebeen reviewed for several terrestrial and ma-rine biomes for the NGO community in a vol-ume by Hansen and Hoffman (2003). Amongthe suggestions of chapter authors of this vol-ume were reduction in current stressors, max-imizing habitat heterogeneity in reserves, andearly and progressive implementation of pro-tection and management actions.

Moveable Feast: Protection thatVaries in Space and Time

Protected areas whose boundaries or level ofprotection vary in time or space are a logical


response to shifting species ranges due to cli-mate change. If the target of protection (aspecies or ecosystem) is moving, why not haveprotection move with it? Altering protection intime is relatively common in current conserva-tion practice, while protection that varies acrossgeographic space has fewer analogs current.Both temporal and spatial variations may beadapted for improved conservation response toclimate change (Scott 2007).

Protected areas that change are not unusual,particularly in sport hunting and fisheries. Tem-porally variable seasonal protection is com-mon in these applications. Examples includeclosed seasons in hunting and seasonal fish-eries closures. The level of protection may varyover time, for example in catch limits or baglimits that vary year to year depending on aregulatory agencies’ assessment of populationsize.

Area closures are a management tool inmany multiple-use protected areas such as Na-tional Forests. In fisheries, area closures areemployed when seasonal closures are not suf-ficient to maintain a healthy population. Evenwithin National Parks and nature reserves, areaclosures and zoning-related restrictions, for in-stance to allow vegetation to recover fromoveruse, are commonly employed. When clo-sures rotate between adjacent managementunits, as they sometimes do in fisheries, forestreserves, or between zoning areas in parks, theeffect is of a mobile protected area. These ex-isting practices form a precedent for protectedareas that vary in time and space to respond toclimate change (Pressey et al. 2007). However, tobe effective in management of climate change,such tools would need to be applied in coor-dinated ways over large areas. This requiresadvanced planning, monitoring, and coordi-nation between parcels specifically for climatechange.

Mobile protected areas to deal with climatechange were first proposed for marine conser-vation (Soto 2001). Major areas of ocean cur-rent interaction, or “fronts,” are associated withhigh concentrations of large marine predators

such as tuna and swordfish. Fronts are not fixedin space and may move over the course of ayear or within a season in response to changesin patterns of circulation. Marine frontal zonesare likely to shift with climate change as sea sur-face temperatures and global circulation pat-terns change. Geographically fixed protectionfor fronts would require huge areas to capturethese present and future variations. Mobile pro-tected areas would provide for the conservationof these variable features (Soto 2001).

Conservation easements are a final class ofconservation tool that may be variable in time.Easements are bundles of use restrictions thatcan be reassessed, triggered, or removed as timepasses. A coordinated system of landholderagreements has been proposed as a mechanismfor providing mobile protection in response toclimate change (Hannah & Hansen 2005). Inthis application, easements might be used toprovide futures options to acquire land at a pre-specified price or to effect a specified land-usechange (such as reversion to forest) based onfuture conservation action needed to respondto climate change.

When Protection Alone Fails:Assisted Migration

Assisted migration may be necessary whenclimate change is so rapid that the speed ofa range shift exceeds a species’ dispersal ability(McLachlan et al. 2007). It may also be requiredwhere human land uses make establishment ofnew protected area or connectivity impossible.Assisting dispersal may be as simple as trans-porting seeds from a plant to a new location oras difficult as maintaining viable populationsof a top predator over several generations untilnewly suitable habitat is fully established.

There are numerous biological and policyissues that confront assisted migration, whichare only beginning to be explored in the lit-erature (McLachlan et al. 2007). In the shortterm, few species have suffered documentedclimate change–related declines which would


warrant assisted migration, so there is time toput technical and policy frameworks in placefor decisions about appropriate application ofassisted migration. One of the elements of acost-effective policy will be preference for as-sisted migration in settings in which adequateprotected areas are in place to ensure the sur-vival of the translocated populations.

Other measures, such as captive breeding,may be necessary complements to in situ con-servation in protected areas as well. For exam-ple, ex situ management may be a necessarycomplement to protection for conservation ofgenetic resources. Climate change, like habi-tat loss, may lead to genetic bottlenecks, in-breeding, and related genetic and population-level issues that may require monitoring andpossibly captive-breeding efforts to supplementwild populations. A genetic issue unique to cli-mate change is loss of local adaptation. Popu-lations at the trailing edge of range shifts maybe particularly important. These low-latitudelimits of species’ ranges have frequently beenrefugia during the predominantly cooler cli-mates of the past 2 million years, makingthem unique stores of genetic diversity (Hewitt2000). As climates continue to warm on to-day’s relatively warm (in Pleistocene terms) cli-mate, these low-latitude, trailing-edge popula-tions may fragment and face extinction (Hampe& Petit 2005). Some gene banking and ex situ

conservation of individuals from these popu-lations may be warranted as climate changecontinues.

Conclusion

Protected areas have been generally under-recognized as a policy response for prevent-ing extinctions due to climate change—nonew protected areas have yet been establishedspecifically to cope with climate change. Suf-ficient evidence now exists to indicate thatearly implementation of new protected areas islikely to substantially reduce the threat climatechange poses to biodiversity. Connectivity and

corridors are much more frequently proposedas responses to climate change. Connectivityon the periphery of existing protected areas hasbeen shown to be the most cost effective, sincespecies move individualistically, making oppor-tunities to establish multispecies corridors forclimate change rare. Corridors for other pur-poses, such as large predator dispersal, may bevaluable for climate change, however, and cor-ridors can and should be designed to serve mul-tiple purposes wherever possible.

Addition of protected areas to counter theeffects of climate change on the natural worldis complicated by the uneven coverage of cur-rent protected areas. Large gaps in represen-tation exist, even for current ranges of rela-tively well-known taxa, such as mammals andbirds (Rodrigues et al. 2004). This lack of ma-turity in global protection must be addressedin order for improvements for climate changeto take place. It presents the opportunity, how-ever, of closing both present and future (climatechange–induced) gaps at the same time. Sinceclimate change may be a motivating force toimprove the consistency of global protected ar-eas coverage, there is hope for meeting thesedual goals.

Connectivity within and adjacent to pro-tected areas is being found to be more cost ef-fective and feasible than long-distance connec-tivity or connections between protected areas.The effectiveness of new connectivity is heavilyinfluenced by the degree and extent of speciesrange shifts. Assisted migration, captive breed-ing, and other intensive management optionsneed to be implemented on a species-by-speciesbasis and have the potential to become ex-tremely expensive as communities disaggregateand reassemble as species follow individualisticrange changes under severe climate change. In-creasingly management-intensive options anddecreasing species representation in protectedareas can be expected if climate change contin-ues unabated. Conservation strategies for thelatter half of the century, and their cost and ef-fectiveness, therefore depend in large part onaction to bring climate change in check.


Conflicts of Interest

The author declares no conflicts of interest.

References

Anderson, R.P. & E. Martinez-Meyer. 2004. Model-ing species’ geographic distributions for preliminaryconservation assessments: an implementation withthe spiny pocket mice (Heteromys) of Ecuador. Biol.

Conserv. 116: 167–179.Araujo, M.B., M. Cabeza, et al. 2004. Would climate

change drive species out of reserves? An assessmentof existing reserve-selection methods. Global Change

Biol. 10: 1618–1626.Bellwood, D.R., A.S. Hoey, et al. 2006. Coral bleaching,

reef fish community phase shifts and the resilience ofcoral reefs. Global Change Biol. 12: 1587–1594.

Carpenter, G., A.N. Gillison, et al. 1993. DOMAIN—aflexible modeling procedure for mapping potentialdistributions of plants and animals. Biodiversity Con-

serv. 2: 667–680.Coulston, J.W. & K.H. Riitters. 2005. Preserving biodiver-

sity under current and future climates: a case study.Global Ecol. Biogeogr. 14: 31–38.

Elith, J., C.H. Graham, et al. 2006. Novel methods im-prove prediction of species’ distributions from occur-rence data. Ecography 29: 129–151.

Guisan, A. & N.E. Zimmermann. 2000. Predictive habi-tat distribution models in ecology. Ecol. Modell. 135:147–186.

Halpin, P.N. 1997. Global climate change and natural-area protection: management responses and re-search directions. Ecol. Appl. 7: 828–843.

Hampe, A. & R.J. Petit. 2005. Conserving biodiversityunder climate change: the rear edge matters. Ecol.

Lett. 8: 461–467.Hannah, L. 2001. The role of a global protected areas

system in conserving biodiversity in the face of cli-mate change. In Climate Change and Protected Areas. G.Visconti & M. Balaban. Kluwer. Dordrecht.

Hannah, L. & L.A. Hansen. 2005. Designing landscapesand seascapes for change. In Climate Change and Bio-

diversity. T.E. Lovejoy & L. Hannah. Yale UniversityPress. New Haven.

Hannah, L., G. Midgley, et al. 2007. Protected area needsin a changing climate. Front. Ecol. Environ. 5: 131–138.

Hannah, L., G.F. Midgley, et al. 2005. The view from theCape: Extinction risk, protected areas and climatechange. BioScience 55: 231–242.

Hannah, L., G.F. Midgley, et al. 2002. Conservation ofbiodiversity in a changing climate. Conserv. Biol. 16:11–15.

Hansen, L.B. & J. Hoffman. 2003. BUYING TIME:A User’s Manual for Building Resistance andResilience to Climate Change in Natural Systems.http://assets.panda.org/downloads/buyingtime_unfe.pdf. (accessed Feb. 11, 2008).

Hewitt, G. 2000. The genetic legacy of the Quaternaryice ages. Nature 405: 907–913.

Hoegh-Guldberg, O. 1999. Climate change, coral bleach-ing and the future of the world’s coral reefs. Mar.

Freshw. Res. 50: 839–866.Hoegh-Guldberg, O., P.J. Mumby, et al. 2007. Coral reefs

under rapid climate change and ocean acidification.Science 318: 1737–1742.

Hossell, J.E., N.E. Ellis, et al. 2003. Climate change and na-ture conservation: implications for policy and prac-tice in Britain and Ireland. J. Nat. Conserv. 11: 67–73.

Huntley, B., P.M. Berry, et al. 1995. Modelling presentand potential future ranges of some European higherplants using climate response surfaces. J. Biogeogr. 22:967–1001.

McClean, C.J., N. Doswald, et al. 2006. Potential impactsof climate change on Sub-Saharan African plant pri-ority area selection. Divers. Distrib. 12: 645–655.

McLachlan, J.S., J.J. Hellmann, et al. 2007. A frameworkfor debate of assisted migration in an era of climatechange. Conserv. Biol. 21: 297–302.

Midgley, G.F., L. Hannah, et al. 2002. Assessing the vul-nerability of species richness to anthropogenic cli-mate change in a biodiversity hotspot. Global Ecol.

Biogeogr. 11: 445–451.Miles, L., A.C. Newton, et al. 2006. A global overview

of the conservation status of tropical dry forests. J.

Biogeogr. 33: 491–505.Nix, H.A. 1986. BIOCLIM—a bioclimatic analysis and

prediction system. Pages 59–60. CSIRO Division ofLand and Water Resources Annual Report. CSIRO,Canberra, Australia.

Pearsall, S. 2005. Managing for future change on theAlbemarle Sound. In Climate Change and Biodiversity.T.E. Lovejoy & L. Hannah, Eds. Yale UniversityPress, New Haven.

Pearson, R.G. 2006. Climate change and the migra-tion capacity of species. Trends Ecol. Evol. 21: 111–113.

Pearson, R.G. & T.P. Dawson. 2005. Long-distance plantdispersal and habitat fragmentation: identifying con-servation targets for spatial landscape planning un-der climate change. Biol. Conserv. 123: 389–401.

Peters, R.L. & J.D.S. Darling. 1985. The GreenhouseEffect and Nature Reserves. BioScience 35: 707–717.

Peters, R.L. & T.E. Lovejoy. 1992. Global Warming and

Biological Diversity. Yale University Press. London.Peters, R.L. & J.P. Myers. 1991. Preserving biodiversity in

a changing climate. Issues Sci. Technol. 66–72.


Peterson, A.T., M.A. Ortega-Huerta, et al. 2002. Futureprojections for Mexican faunas under global climatechange scenarios. Nature 416: 626–629.

Pressey, R.L. & R.M. Cowling. 2001. Reserve selectionalgorithms and the real world. Conserv. Biol. 15: 275–277.

Pressey, R.L., M. Cabeza, et al. 2007. Conservation plan-ning in a changing world. Trends Ecol. Evol. 22: 583–592.

Pyke, C.R., S.J. Andelman, et al. 2005. Identifying priorityareas for bioclimatic representation under climatechange: a case study for Proteaceae in the CapeFloristic Region, South Africa. Biol. Conserv. 125:1–9.

Pyke, C.R. & D.T. Fischer. 2005. Selection of bioclimati-cally representative biological reserve systems underclimate change. Biol. Conserv. 121: 429–441.

Rodrigues, A.S.L., S.J. Andelman, et al. 2004. Effective-ness of the global protected area network in repre-senting species diversity. Nature 428: 640–643.

Root, T.L. & S.H. Schneider. 1995. Ecology and cli-mate: research strategies and implications. Science

269: 334–341.Rutherford, M.C., L.W. Powrie, et al. 1999. Climate

change in conservation areas of South Africa andits potential impact on floristic composition: a firstassessment. Divers. Distrib. 5: 253–262.

Salm, R.V., S.L. Coles, et al. 2001. Coral Bleaching andMarine Protected Areas. Pp. 1–107. The NatureConservatory. Honolulu, Hawaii.

Scott, D., J.R. Malcolm, et al. 2002. Climate change andmodelled biome representation in Canada’s national

park system: implications for system planning andpark mandates. Global Ecol. Biogeogr. 11: 475–484.

Scott, D. & R. Suffling. 2000. Climate Change and Canada’s

National Park System. Environment Canada. Toronto,Canada.

Scott, J.M. 2007. Presentation at the annual meeting ofthe Wildlife Society. Tucson, Arizona.

Shirley, L.J. & L.L. Battaglia. 2006. Assessing vegetationchange in coastal landscapes of the Northern Gulfof Mexico. Wetlands 26: 1057–1070.

Soto, C.G. 2001. The potential impacts of global climatechange on marine protected areas. Rev. Fish Biol. Fish-

eries 11: 181–195.Thomas, C.D., A. Cameron, et al. 2004. Extinction risk

from climate change. Nature 427: 145–148.Van Oppen, M.J.H. & R.D. Gates. 2006. Conservation

genetics and the resilience of reef-building corals.Mol. Ecol. 15: 3863–3883.

Villers-Ruiz, L. & I. Trejo-Vazquez. 1998. Climate changeon Mexican forests and natural protected areas.Global Environ. Change Human Policy Dimensions 8: 141–157.

Walther, G., E. Post, et al. 2002. Ecological responses torecent climate change. Nature 416: 389–395.

West, J.M. & R.V. Salm. 2003. Resistance and resilienceto coral bleaching: implications for coral reef con-servation and management. Conserv. Biol. 17: 956–967.

Williams, P., L. Hannah, et al. 2005. Planning for cli-mate change: identifying minimum-dispersal corri-dors for the Cape proteaceae. Conserv. Biol. 19: 1063–1074.

protected areas and climate change

Documents