No abstract

This study sought to identify climate-change thermal-stress refugia for reef corals in the Indian and Pacific Oceans. A species distribution modeling approach was used to identify refugia for 12 coral species that differed considerably in their local response to thermal stress. We hypothesized that the local response of coral species to thermal stress might be similarly reflected as a regional response to climate change. We assessed the contemporary geographic range of each species and determined their temperature and irradiance preferences using a k-fold algorithm to randomly select training and evaluation sites. That information was applied to downscaled outputs of global climate models to predict where each species is likely to exist by the year 2100. Our model was run with and without a 1°C capacity to adapt to the rising ocean temperature. The results show a positive exponential relationship between the current area of habitat that coral species occupy and the predicted area of habitat that they will occupy by 2100. There was considerable decoupling between scales of response, however, and with further ocean warming some 'winners' at local scales will likely become 'losers' at regional scales. We predicted that nine of the 12 species examined will lose 24-50% of their current habitat. Most reductions are predicted to occur between the latitudes 5-15°, in both hemispheres. Yet when we modeled a 1°C capacity to adapt, two ubiquitous species, Acropora hyacinthus and Acropora digitifera, were predicted to retain much of their current habitat. By contrast, the thermally tolerant Porites lobata is expected to increase its current distribution by 14%, particularly southward along the east and west coasts of Australia. Five areas were identified as Indian Ocean refugia, and seven areas were identified as Pacific Ocean refugia for reef corals under climate change. All 12 of these reef-coral refugia deserve high-conservation status.

Coral reefs have recently experienced an unprecedented decline as the world's oceans continue to warm. Yet global climate models reveal a heterogeneously warming ocean, which has initiated a search for refuges, where corals may survive in the near future. We hypothesized that some turbid nearshore environments may act as climate-change refuges, shading corals from the harmful interaction between high sea-surface temperatures and high irradiance. We took a hierarchical Bayesian approach to determine the expected distribution of 12 coral species in the Indian and Pacific Oceans, between the latitudes 37°N and 37°S, under representative concentration pathway 8.5 (W m(-2) ) by 2100. The turbid nearshore refuges identified in this study were located between latitudes 20-30°N and 15-25°S, where there was a strong coupling between turbidity and tidal fluctuations. Our model predicts that turbidity will mitigate high temperature bleaching for 9% of shallow reef habitat (to 30 m depth) - habitat that was previously considered inhospitable under ocean warming. Our model also predicted that turbidity will protect some coral species more than others from climate-change-associated thermal stress. We also identified locations where consistently high turbidity will likely reduce irradiance to <250 μmol m(-2)  s(-1) , and predict that 16% of reef-coral habitat ≤30 m will preclude coral growth and reef development. Thus, protecting the turbid nearshore refuges identified in this study, particularly in the northwestern Hawaiian Islands, the northern Philippines, the Ryukyu Islands (Japan), eastern Vietnam, western and eastern Australia, New Caledonia, the northern Red Sea, and the Arabian Gulf, should become part of a judicious global strategy for reef-coral persistence under climate change.

Aim: Coral reefs are experiencing both an increasing frequency and intensity of anomalously warm ocean temperatures because of climate change. Studies show that the majority of coral populations will likely decline as temperatures continue to increase, although some previous species-distribution models predict that ubiquitous species, such as the primary reef-building coral species Porites lobata, will increase their distribution under projected climate change. These predictive models, however, assume that all individuals of a population are able to tolerate the entire range of environmental conditions within the species' geographic range. The effects of genetic isolation and local adaptation are not considered in species-distribution models that assume genetically contiguous populations. We aim to determine the effects of genetic isolation and local adaptation in species-distribution modelling of the ubiquitous species P. lobata under three climate change scenarios by comparing contiguous and isolated subpopulations.Location: Indian and Pacific Oceans. Methods:We ran a novel species-distribution model for P. lobata, segregated as five geographically isolated regions across the Indian and Pacific Oceans, and examined the species response to three climate-change scenarios (i.e., A2, A1B and B1, most recently considered as Representative Concentration Pathways 8.5, 6.0, and 4.5 Wm À2 ) by the year 2100.Results: In contrast with previous homogeneous species-distribution models that predict a global expansion of P. lobata, (~5 AE 1%), we predict major losses of suitable habitat for P. lobata in four of the five regions examined, particularly in the central Pacific Ocean (>99% AE <0.1% for all climate scenarios). Indeed, when geographic and genetic isolation were considered, our predictions suggested that P.lobata would lose between 50-52 AE 4% of its habitat, depending on the climatechange scenario, mainly in the Pacific Ocean.Main conclusions: Genetic isolation will likely play a major role in the persistence of coral species under climate change, and small isolated populations may be more vulnerable to climate change than populations in large, highly connected regions.

Thermal-stress events are changing the composition of many coral reefs worldwide. Yet, determining the rates of coral recovery and their long-term responses to increasing sea-surface temperatures is challenging. To do so, we first estimated coral recovery rates following past disturbances on reefs in southern Japan and Western Australia. Recovery rates varied between regions, with the reefs in southern Japan showing more rapid recovery rates (intrinsic rate of increase, r = 0.38 year−1) than reefs in Western Australia (r = 0.17 year−1). Second, we input these recovery rates into a novel, nonlinear hybrid-stochastic-dynamical system to predict the responses of Indo-Pacific coral populations to complex inter-annual temperature cycles into the year 2100. The coral recovery rates were overlaid on background increases in global sea-surface temperatures, under three different climate-change scenarios. The models predicted rapid recovery at both localities with the infrequent and low-magnitude temperature anomalies expected under a conservative climate-change scenario, Representative Concentration Pathway (RCP) 4.5. With moderate increases in ocean temperatures (RCP 6.0) the coral populations showed a bimodal response, with model runs showing either recovery or collapse. Under a business-as-usual climate-change scenario (RCP 8.5), with frequent and intense temperature anomalies, coral recovery was unlikely.