Effective climate change refugia for coral reefs.

Abstract

Coral reefs are hyperdiverse ecosystems that are highly threatened by ocean warming and acidification and are vital to the livelihoods of millions of people (Hoegh-Guldberg et al., 2007). Climate change refugia, habitats with favourable environmental conditions that species may retreat to during inhospitable climatic conditions, offer potential safe havens from anthropogenic climate change (Keppel et al., 2012). Such refugia are increasingly important for conservation planning in terrestrial ecosystems (Keppel et al., 2015). Cacciapaglia & van Woesik (2015) take an important step towards identifying refugia for coral reefs. However, their predictions are based on only one global stressor (ocean warming) and made at a coarse scale (>80 km2). As a result, their findings are overly positive and of limited predictive power regarding the identification of effective future refugia for coral reefs. Considering the serious impacts of ocean acidification (Hoegh-Guldberg et al., 2007), it cannot be ignored in predictive modelling of future coral reef refugia. These impacts will be most severe at higher latitudes, constraining the ability of corals to escape ocean warming through migration. As a result, it has been suggested that few or no refugia will be able to safeguard the long-term persistence of coral reefs (van Hooidonk et al., 2014). Furthermore, ocean acidification and warming will exacerbate local stressors related to water pollution, overexploitation and coral-limiting environments (Hoegh-Guldberg et al., 2007). While Cacciapaglia & van Woesik (2015) did consider local stressors to some extent by masking areas receiving high river outflow and areas experiencing cold temperatures, they did not consider global ocean acidification. Hence, their results are likely to be excessively positive regarding the number and size of future coral reef refugia. Scale is a key consideration for modelling species distributions under future climates (Keppel et al., 2012). Indeed, using excessively coarse scales for species distributions modelling will produce inaccurate projections (Franklin et al., 2013). For coral reefs, key environmental parameters affecting coral performance and responses to ongoing ocean acidification and warming, such as temperature and aragonite saturation, may vary among individual reefs within coral reef complexes and even within reefs (Manzello et al., 2012; Guadayol et al., 2014). Although Cacciapaglia & van Woesik (2015) used a finer resolution (~9.2 × 9.2 km) than a previous attempt (1 × 1 degree latitude/longitude, van Hooidonk et al., 2014), this resolution is not relevant to fine-scale microclimatic habitats that exist on coral reefs. Even parts of a single reef system could potentially provide important refugia. For example, the proximity to seagrass beds can provide considerable buffering from the effects of ocean acidification (Manzello et al., 2012). Therefore, to adequately predict the existence and location of coral reef refugia, relevant environmental data need to be downscaled to sufficiently fine scales. In marine environments, this is particularly complex, as environmental variables vary with horizontal distance and depth (Guadayol et al., 2014). Refugia are species specific (Stewart et al., 2010; Keppel et al., 2012), as species display varying responses to environmental conditions (Cacciapaglia & van Woesik, 2015). The selection of species for predictive modelling therefore will have considerable impact on the forecast refugia for coral reef communities. Laudably, Cacciapaglia & van Woesik (2015) used 12 coral species to determine their proposed refugia. However, all species selected were generalists, which generally have greater ranges, dispersal potentials, genetic diversity, adaptive potential and resilience to environmental changes than specialist species (see Cacciapaglia & van Woesik, 2015). These rare specialists are most at risk from anthropogenic climate change and will likely depend on local, fine-scale refugia for persistence (Purvis et al., 2000). Therefore, the responses of rare specialist species need to be taken into account, if effective refugia for coral reef communities are to be identified. The identification of refugia for coral reefs is likely to be important for safeguarding the persistence of these communities under anthropogenic climate change, and Cacciapaglia & van Woesik (2015) present an important first step towards this goal. However, their analyses are overly simplistic, and the inclusion of ocean acidification into the modelling would have likely produced vastly different results (see van Hooidonk et al., 2014). Ideally, modelling of future distributions of coral reef communities would include ocean acidification and warming and indicators of other key stressors at appropriate scales for a wide variety of species. However, our ability to do this is currently constrained by the unavailability of sufficient spatial data at fine scales for key environmental parameters, such as ocean temperature and aragonite saturation. It is therefore vitally important to gain an understanding of the variability of marine environmental parameters in three-dimensional spaces at appropriate scales, and to use realistic modelling approaches, if effective refugia for coral reef communities are to be identified.