The ultimate constraints on the distribution of a species are usually climatic (Davis & Shaw 2001), so increasingly severe and rapid anthropogenic climate change is likely to result in increased rates of species decline or loss (Thomas et al. 2004). There is now strong theoretical (Jetz et al. 2007, Huntley et al. 2008, La Sorte & Jetz 2010) and empirical (Jiguet et al. 2007, Green et al. 2008, Gregory et al. 2009) evidence that suggests climate change will generally result in a poleward or altitudinal shift in species’ distribution, at least in boreal and temperate zones. These range shifts will occur because of both reductions in populations at the retreating range-margin and increases in abundance at the advancing range-margin associated with climatic amelioration. The relative strength of these two processes will largely determine the ultimate consequences of climate change for a species.
Management intervention may either reduce the negative effects of climate change at the retreating range-margin or increase the capacity of species to respond to climate amelioration at the advancing margin (Green & Pearce-Higgins 2010). These two options involve the promotion of different, although not necessarily competing, strategies, resulting in much debate over the most appropriate methods of climate change adaptation (e.g. Opdam & Wascher 2004, Heller & Zavaleta 2009, Hodgson et al. 2009), which are likely to vary between species and taxa. We summarize here the likely suitability of different options for birds, building on discussions at the 2010 BOU conference on climate change and birds, and use these to identify research priorities to help inform climate change adaptation. We focus on five main adaptation options, taking a sixth, a reduction in the severity of other threats, as a given:
1 increasing the size of key sites;
2 managing the condition of key sites;
3 the creation of new key sites;
4 increasing functional connectivity between key sites;
5 translocation to establish new key sites.
To illustrate these in practice, a habitat specialist species with a small contiguous global range may be protected within a network of protected areas of suitable habitat with latitudinal extent p (Fig. 1a). Assuming that climate change will result in a latitudinal shift in species distribution of s, if s < p, the current protected area network may remain sufficient (Fig. 1a). Depending on the degree of overlap between the future range and p, management to increase the size of protected areas (option 1) may be desirable to ensure that s < p (Fig. 1b). It may also be necessary to create additional protected areas of habitat within the potential future range if none exists (option 3; Fig. 1c). The relative shift in priority from option 1 to options 3 and 4 therefore depends upon the magnitude of s, which will be related to the severity of future climate change and the vulnerability of the species. Management to increase the resistance of vulnerable populations to climate change by improving the condition of key sites (option 2) will reduce s, shifting the focus of climate change adaptation from facilitating range change to maintaining the existing range (Fig. 1d). Management to increase connectivity within a species’ range (option 4) may also increase the resilience of that population to climate change and therefore reduce the magnitude of s. If all else fails, it may be possible simply to translocate individuals to establish new populations within the newly climatically suitable range (option 5).
Priorities for research
This illustration can help identify the research required to inform particular adaptation strategies. Before doing so, however, three more generic issues need to be acknowledged. First, we have only a good understanding of the population dynamics and population-level impacts of climate change for a very small number of well-studied species. Expanding this understanding for a wider range of taxa and environments will always be beneficial. Secondly, conservation strategies are dependent upon effective estimates of species’ distributions. Thirdly, accurate estimates of abundance and trends are required to provide information about the likely effectiveness of the current protected area network for a particular species. Where this is not the case, significant effort needs to be invested to acquire this knowledge.
Improving the understanding of vulnerability
An accurate estimate of s has been the focus of much climate change research on birds (e.g. Jetz et al. 2007, Huntley et al. 2008), based largely upon correlations between large-scale patterns of occurrence and broad measures of climate. Although subject to significant criticism (e.g. Araújo & Rahbek 2006, Beale et al. 2008), there is increasing evidence that species’ populations are responding broadly in the ways that climate envelope models predict (Jiguet et al. 2007, Green et al. 2008, Gregory et al. 2009), and such models may be used to inform management decisions to account for climate change (Hole et al. in press). However, the accuracy of estimates of vulnerability, particularly for individual species, may be improved by the incorporation of additional ecological constraints and greater mechanistic understanding of impacts (La Sorte & Jetz 2010). Given their role in planning conservation adaptation, maximizing the accuracy of such models is important. Additionally, more work should be conducted at a range of spatial scales to identify the potential for fine-scale habitat and micro-climatic variation to permit populations to persist in areas that might otherwise be assessed as climatically unsuitable. This has so far only been examined in a few invertebrates (e.g. Davies et al. 2006). Failing to account for these processes may result in an over-estimation of s in areas with a high degree of habitat, landscape and climate heterogeneity (Hodgson et al. 2009). Existing models may also over-estimate s as they do not incorporate the potential of micro-evolution to enable species to adapt to climate change. Certainly, some species have the potential for a rapid evolutionary adaptation in response to a changing climate (Bertaux et al. 2004, Bradshaw & Holzapfel 2006), although identifying such species to improve estimates of future vulnerability (Visser 2008) is a significant challenge.
Improving understanding of how to facilitate resistance
There is considerable uncertainty about the potential of site-based management to increase the resistance of populations to climate change (option 2). This could be considered a temporary strategy, perhaps buying time to create new appropriate habitat elsewhere (Gilbert et al. 2010), but also has ‘no regrets’, as it retains a species’ range until no longer possible or practical, while maximizing dispersing individuals to colonize new locations. Process models may reduce this uncertainty through the incorporation of demographic mechanisms by which climate change impacts upon populations (Pearce-Higgins & Gill 2010). Where these include mechanisms underpinning variation in demographic parameters and knowledge of how management may impact on those mechanisms, they may be used to inform climate change adaptation (Ratcliffe et al. 2005, Gilbert et al. 2010, Pearce-Higgins et al. 2010). This evidence-based approach has been applied successfully to existing conservation problems and should therefore be adaptable to climate change adaptation (Green & Pearce-Higgins 2010).
However, although there have been documented conservation successes for some species, these have not always been cost-effective and have generally failed to result in an improvement in the quality of the wider landscape (Wilson et al. 2009, Lawton et al. 2010). Research to quantify the potential for, and limits to, site-based management to increase the resistance of populations to climate change and therefore estimate the degree to which option 2 may result in a reduction in s should be a priority. Although this may be achievable for a range of well-studied rare and declining species, for many others such detailed studies do not exist, and it may be necessary to attempt adaptation management in the absence of good ecological understanding and robust science. In such cases, interventions should be trialled and monitored, and then adapted on the basis of observed responses (Innes et al. 1999).
Improving understanding of how to accommodate range shifts
Existing models used to estimate future range shifts are largely pattern-based, and do not yet take advantage of work to understand the likely processes underpinning range change. There is good evidence that habitat fragmentation has impeded the climate-driven range expansion of butterflies within the UK (Warren et al. 2001, Willis et al. 2009a), and therefore by implication that reducing habitat fragmentation will help to accommodate the response of populations to climate change. No such direct evidence exists as yet for birds, although fragmentation may reduce the ability of individuals to disperse (Lampila et al. 2005). The occupancy of isolated habitat patches may be enhanced by the provision of corridors (Hannon & Schmiegelow 2002) and stepping stones (Van Dorp & Opdam 1987), although research suggests that the proportion of species likely to benefit from such management is relatively small (Hodgson et al. 2009). Alternatively, management of the intervening landscape between key sites may increase the ability of species to disperse through that landscape (Donald & Evans 2006). This principle is also largely untested, but there is some evidence that the persistence of bird species within fragmented tropical habitats is related to the ability of those species to utilize the intervening matrix (Sekercioglu et al. 2002).
The development of biologically realistic process models that incorporate colonization and dispersal patterns (e.g. Anderson et al. 2009, Willis et al. 2009b) may result in more robust future projections and, particularly, provide a framework to assess the effectiveness of different options to increase functional connectivity. This is required to guide estimates of the ability of option 4 to bridge the gap between s and p (Vos et al. 2008), essential given the current policy advocacy in this area.
Improving understanding of the potential for translocation
The use of reintroduction techniques to restore species’ populations to parts of their former range from which they have been lost is an increasingly employed technique in conservation (Green & Pearce-Higgins 2010) and has been successfully applied to birds (e.g. Evans et al. 1999, 2009). We might therefore expect that translocation could be used to establish populations in areas of suitable habitat but previously unsuitable climate in the absence of natural colonization. The potential for this approach has already been demonstrated for invertebrates (Willis et al. 2009a), and similar experiments could usefully be conducted on birds. Our ability to do this will require an understanding of the mechanisms underpinning range-limits, and how to ensure that translocations are to suitable areas where those limits are no longer exceeded.
A number of potential management adaptation tools are available for the conservation practitioner to use in a changing climate (Green & Pearce-Higgins 2010). Scientists should aim to produce usable frameworks to help policy makers and practitioners decide which are the most appropriate to use (e.g. Hole et al. in press). However, given the high degree of uncertainty associated with both future projections of species’ response to climate change and assessments of the likely effectiveness of climate change adaptation, there is an urgent need for research to reduce this uncertainty. It is vital that robust and long-term monitoring of trends in biodiversity is adequately funded, as it is largely through the provision of such data that the impacts of climate change may be identified and understood. Additionally, the development of demographic models can provide a mechanism to synthesize available information into a usable framework to inform policy decisions. Short-term but relatively intensive ecological studies may be required to obtain the data required to parameterize such models, better to appraise the potential for both increasing functional connectivity (e.g. Bellamy et al. 1998, Foppen et al. 1999) and site-based management to increase resistance to climate change (e.g. Gilbert et al. 2010, Kleijn et al. 2010, Pearce-Higgins et al. 2010). For other species, decisions may have to be made on the basis of general principles in the absence of detailed knowledge, and the consequences of any such decisions should be monitored to help improve knowledge. More widely, there is an urgent need for experimental studies to test solutions that models suggest will be effective, which will require adaptive climate change adaptation. These are all tried and tested methods which have previously been applied to other conservation and ecological issues (e.g. Norris 2004), and therefore it should be possible to make considerable progress in informing the development of evidence-based climate change adaptation policy, given sufficient resources.
This viewpoint is based upon discussions at the 2010 annual BOU conference on Birds and Climate Change and we are grateful to all those who contributed to that discussion.