Restoration is one of the basic tools of a conservation practitioner. It has traditionally involved returning a system to its state prior to some disturbance ¹⁴ . However, climate change challenges the very nature of restoration based on such a definition ¹⁵ . It brings into question the utility of historical benchmarks as restoration targets, the species or seed sources to be used, and the time frame for planning.

Rethinking how restoration efforts are applied in light of climate change has led to shifts in thinking as well as in practice. Most fundamentally, practicing restoration in a changing climate requires embracing uncertainty and accepting that the goals of a project may need to change over time ¹⁶ . Instead of relying on historical benchmarks, restoration efforts will likely need to look into the future and anticipate change—perhaps relying on a dynamic reference process that accounts for variability in reference ecosystems ^{16, 17} . Looking forward will also include rethinking the mix of species to be planted and potentially focusing on ecosystem function rather than particular assemblages of species. Restoration efforts have begun to make use of the same niche modelling methods that have been used to assess potential species responses to climate change ¹⁸ , but there is clear recognition that better models are needed ¹⁹ .

Many recent climate-adaptation efforts have involved restoration. Of the projects funded by the Wildlife Conservation Society’s Climate Adaptation Fund, at least 70% have involved restoration ²⁰ . These projects have included riparian restoration to enhance connectivity, coastal restoration to prevent storm damage, forest restoration to reduce fire risk, and prairie restorations to enhance watershed function. In addition to the projects that clearly engage in restoration, several projects are aimed at converting one type of ecosystem to another—not necessarily restoring a historical condition but preparing an ecosystem to function differently in a future climate. An example of this type of project involved converting forested areas to grasslands to facilitate marsh establishment in the face of sea-level rise. Whether such actions are classified as restoration could be debated, but clearly the lessons learned from decades of restoration will be essential to these new types of adaptation projects.

Systematic conservation planning ²¹ has been used widely around the world to help prioritize conservation efforts, particularly the location of protected areas ²² . Most conservation planning has been based on “static” representations of biodiversity across a region, an approach that is clearly challenged by climate-driven changes in the distribution of species and communities ²³ . Consequently, substantial thought has been put into how best to incorporate climate change into the conservation-planning process.

The cornerstone of initial approaches to integrate climate change into conservation planning was the use of correlative niche models to predict future distributions of species and ensure that these were adequately represented by present-day conservation efforts. The more sophisticated niche-based planning efforts included the ability of species to track changing habitat conditions through space and time ^{24, 25} and consideration of uncertainties in predicted distributions ^{26, 27} . Climate, however, is only one of many factors determining the distributions of species, and the relationship is complex, uncertain, and in many cases evolving ²⁸ . The magnitude of these ecological uncertainties compounded by the uncertainties associated with climate predictions ²⁹ led to calls to integrate climate change into planning using approaches that were more robust to uncertainty in predicted climate impacts ^{30– 32} .

An approach to planning that has gained some traction for conservation in a dynamic climate involves conserving the underlying geophysical variation in a region, also referred to as “conserving nature’s stage” ³³ . The rationale for this approach is that, theoretically, there should be a strong relationship between species distributions and geophysical settings (for example, elevation and geology) ³⁴ such that conserving representative examples of geophysical settings will protect representative ecological communities under both current and future climates ³⁵ . A similar approach that focuses on current and known patterns in a region emphasizes conserving connectivity between climatically diverse areas ^{31, 36, 37} .

Increasing the connectivity of landscapes to allow species to move in response to climate change is the most-often cited climate change adaption strategy ^{38– 40} . Traditionally, connectivity planning has focused on connecting patches of habitat with what amount to linear strips or stepping-stones of more habitat. Although this approach may allow species to move through the landscape, it may do little to facilitate movement into what might become newly suitable habitat. Early attempts to model connectivity expressly for addressing climate change involved projecting shifts in species distributions through time and either identifying overlap of current and future distributions or mapping pathways that tracked those shifting distributions ^{41, 42} . A more mechanistic application of this approach involved mapping potential corridors through climate-driven shifts in suitable levels of snowpack for wolverines in the northwestern United States ⁴³ . Other studies have taken different approaches to identifying important areas for species movement in the face of climate change—approaches that are less reliant on projected future changes in climate and species distributions. Brost and Beier ⁴⁴ and Beier ⁴⁵ focused on the geophysical settings mentioned in the previous section, and charted routes across the landscape that either connected similar geophysical setting or linked a diversity of settings. Nuñez and colleagues ⁴⁶ mapped routes through landscapes that connected slightly warmer patches of intact land to slightly cooler ones with routes that followed gentle temperature gradients and avoided human-impacted landscapes.

Although efforts continue to develop more relevant climate-connectivity methods, the critical question of whether corridors are really needed—or whether there might be other strategies and approaches that would achieve the same result—has been repeatedly raised ³⁷ . Some argue that the focus on corridors is misguided and that, alternatively, protecting large intact ecosystems should be prioritized ^{47, 48} . Others have argued that the benefits of increasing the size and number of corridors are fewer than those resulting from simply increasing the amount of protected land, which, if regularly distributed, would increase connectivity ⁴⁹ .

Although a growing wealth of studies predict increased extinction risk for species because of climate change ^{50, 51} , many of these vary enormously in their estimations. It is also increasingly recognized that the predictions of extinction risk do not reflect the number of species that have become vulnerable (or extinct) to date, nor do they match the number identified as threatened because of climate change on the International Union for the Conservation of Nature (IUCN) Red List (only 10.5% of the 22,176 species) ⁵² . A fundamental challenge has involved integrating the projections of species niche models (the most-often used tool for assessing the impacts of climate change on species) into the processes of real-world extinction assessments.

Simple measures of population size, geographic range size, and other indicators of current status already used as IUCN Red List criteria are likely to be good predictors of climate change-associated extinction risk ⁵³ . Such IUCN Red List criteria have been used to predict the risk of extinction in the absence of conservation action and the time lag between assessment and extinction ^{54, 55} . This time lag amounts to a warning period in which adaptation efforts can be taken to prevent extinction. Although there is a warning period, it is finite and thus delays in developing and implementing conservation plans after a species is identified as being threatened could be costly. Roughly half of listed species are likely to go extinct within 20 years of being listed as critically endangered ⁵⁴ .

An additional challenge is that most species risk assessments treat climate change as a problem driven by relatively slow, predictable, and continuous change in environmental conditions and fail to account for other important components of climate change, such as increasing extreme weather and climate events (for example, cyclones, floods, and drought) ^{56, 57} . It is increasingly recognized that the increases in frequencies and intensities of extreme events are critical determinants of patterns of biological diversity and will affect it differently from impacts resulting from steady climate change ⁵⁸ . A good example of this is how climate change will impact bat species: extreme maximum temperature is now considered a critical factor in the vulnerability of bats to climate change ⁵⁹ , but many studies (for example, 60) fail to consider it in projections of species distributions under climate change.

Even though our understanding of which extremes are most important and how they are shifting is limited ⁶¹ , there are good examples of assessments that do account for extremes. Recently, Ameca y Juárez and colleagues ⁶² produced a comprehensive analysis of the impacts of cyclones and droughts on terrestrial mammals, one of the few large-scale studies to consider exposure to extreme events. They followed this exposure analysis with an assessment of terrestrial mammal sensitivity to extreme weather and climate events, identifying biological traits that make large terrestrial mammals more susceptible to climate-induced population declines.

Alongside species risk assessments, the assessment of the vulnerability of species to climate change—as well as the climate-related vulnerability of places and natural resources in general—has emerged as an important step in the adaptation process ⁶³ . As with extinction risk assessments, many different approaches to assessing vulnerability have been proposed (for example, 64), each evaluating some subset or combination of sensitivity, exposure, and adaptive capacity ⁶⁵ . These vulnerability assessments serve not only to determine which species are likely to be most vulnerable but also to identify the factors that make a species vulnerable and thus potential conservation actions—adaptation measures—that can be taken to reduce vulnerability.

One of the relatively new tools in the conservation toolbox is assisted colonization—broadly defined by the IUCN as the movement of an organism outside of its native range to avoid extinction of populations due to current or future threats ⁶⁶ . There are some species, particularly endemics with relatively specific habitat requirements and poor dispersal abilities, that will be unable to move to suitable climates. When these species are threatened with extinction—because of either climate change or some other factor—it may become necessary to move them to prevent their loss.

The question of whether assisted colonization should even be considered as an option spawned a lively debate ^{67– 72} . Those opposed to assisted colonization argue that the history of invasive species has taught us that the potential impacts on the ecosystems into which organisms would be moved could be too great ⁶⁸ . Those in favour of keeping the option of assisted colonization on the table argue that it would likely be necessary for preventing the extinction of certain species ⁶⁹ , that the potential impact of translocated species is likely overstated for several reasons ^{69, 71} (including that the traits of species that will need to be moved are not those traits generally associated with invasive species ⁷³ ), and that the amount of change that systems will experience over the coming decades will likely overshadow the impacts of translocated individuals of a rare and declining species ⁷⁴ .

As the debate worked its ways through the scientific literature, many researchers started to ask more productive questions—facing the reality that assisted colonization was already being used. There were, for example, several early efforts to develop frameworks for determining under what circumstances assisted colonization would be a viable conservation option ^{75, 76} . Other studies have highlighted the importance of the timing of assisted colonization efforts ⁷⁷ , developed advanced modelling approaches for identifying potential sites for translocations ⁷⁸ , and explored the situations in which invasions will be less likely and hence assisted colonization a less risky venture ⁷⁹ . Furthermore, calls for the development of policies to address assisted colonization ⁸⁰ have begun to be met ^{81, 82} . Overall, it appears that, at least in the scientific literature, assisted colonization is gaining acceptance as a tool in the conservation toolbox and one that may not differ so much from other movements of species for conservation reasons ^{83, 84} .