LETTERS PUBLISHED ONLINE: 1 JULY 2012 | DOI:10.1038/NCLIMATE1582 Projected response of an endangered marine turtle population to climate change Vincent S. Saba 1,2 * , Charles A. Stock 3 , James R. Spotila 4 , Frank V. Paladino 5 and Pilar Santidrián Tomillo 4,6 Assessing the potential impacts of climate change on individual species and populations is essential for the stewardship of ecosystems and biodiversity. Critically endangered leatherback turtles in the eastern Pacific Ocean are excellent candidates for such an assessment because their sensitivity to contemporary climate variability has been substantially studied 1–4 . If inciden- tal fisheries mortality is eliminated, this population still faces the challenge of recovery in a rapidly changing climate. Here we combined an Earth system model 5 , climate model projections assessed by the Intergovernmental Panel on Climate Change 6 and a population dynamics model to estimate a 7% per decade decline in the Costa Rica nesting population over the twenty- first century. Whereas changes in ocean conditions had a small effect on the population, the ∼2.5 ◦ C warming of the nesting beach was the primary driver of the decline through reduced hatching success and hatchling emergence rate. Hatchling sex ratio did not substantially change. Adjusting nesting phenol- ogy or changing nesting sites may not entirely prevent the decline, but could offset the decline rate. However, if future observations show a long-term decline in hatching success and emergence rate, anthropogenic climate mitigation of nests (for example, shading, irrigation) 7,8 may be able to preserve the nesting population. Climate change can affect both the marine and terrestrial habitat of marine turtles 9 . The population of eastern Pacific leatherback turtles (Dermochelys coriacea) nesting on the northwest coast of Costa Rica has been studied in terms of its sensitivity to contemporary climate variability in the nesting beach 1,2 and ocean 3,4 . Leatherbacks forage almost exclusively on gelatinous zooplankton and the preferred foraging hotspots are typically waters with high primary productivity of large phytoplankton (that is, upwelling and coastal zones) 4 . Foraging success and reproductive frequency of mature females are enhanced after periods (∼one year) of high primary productivity in the eastern equatorial Pacific 3 . In northwestern Costa Rica, hatching success, hatchling emergence rates and the proportion of male hatchlings all increase during cool and wet conditions 1,2 . Variability in both foraging success of mature females in the ocean and in the local climate in the nesting beach are primarily associated with the El Niño/Southern Oscillation 2,3 (ENSO). The La Niña phase of ENSO is associated with higher primary productivity of large phytoplankton/cooler sea surface temperature (SST) in the eastern equatorial Pacific 10 and enhanced precipitation/cooler air temperature in northwestern Costa Rica 2 . 1 Atmospheric and Oceanic Sciences Program, Princeton University, 300 Forrestal Road, Sayre Hall, Princeton, New Jersey 08544, USA, 2 National Oceanic and Atmospheric Administration, National Marine Fisheries Service, Northeast Fisheries Science Center, c/o Geophysical Fluid Dynamics Laboratory, 201 Forrestal Road, Princeton University Forrestal Campus, Princeton, New Jersey 08540, USA, 3 NOAA Geophysical Fluid Dynamics Laboratory, 201 Forrestal Road, Princeton University Forrestal Campus, Princeton, New Jersey 08540, USA, 4 Department of Biology, Drexel University, 123 PISB, 3245 Chestnut Street, Philadelphia, Pennsylvania 19104, USA, 5 Department of Biology, Indiana-Purdue University at Fort Wayne, 2101 E Coliseum Blvd, Fort Wayne, Indiana 46805, USA, 6 The Leatherback Trust, Goldring-Gund Marine Biology Station, Playa Grande, Costa Rica. *e-mail:Vincent.Saba@noaa.gov. Nesting seasons that follow La Niña events thus result in peaks in the number of nesting females 3 , higher than average hatching success and emergence rate 2 and a larger proportion of male hatchlings 1 , whereas the opposite holds true for El Niño events. Leatherback hatchling sex ratios in Costa Rica are typically female biased (88%; ref. 11) and La Niña events are essential for producing pulses of male recruitment (>50%; ref. 1) into the population. Historic egg poaching 12 and incidental fisheries mortality 13 have rendered the eastern Pacific leatherback population critically endangered 14 . Although some steps have been taken to reduce these losses (for example, beach protection), the sensitivity of eastern Pacific leatherbacks to contemporary climate variability makes it essential for recovery plans to consider the impacts of anthropogenic climate change. Climate projections assessed by the fourth assessment of the Intergovernmental Panel on Climate Change (IPCC AR4) show relatively strong agreement that much of Central America is very likely to (>90%) become warmer and is likely to (>66%) become drier over the twenty-first century 15 . Projections also suggest possible changes in the strength and frequency of El Niño and La Niña events, though there is less agreement on the sign and magnitude of these changes 16 . Here we describe the response of a population dynamics model for eastern Pacific leatherback turtles to changes in environmental forcing from an ensemble of climate model projections that were assessed in the IPCC AR4. We used a series of sensitivity simulations to identify the source of any climate-change-driven trends and then assessed the implications of adaptation through plastic responses. We developed a climate-forced population dynamics model (CLIMPOP) based on existing nesting beach and ocean models 1–4 (Methods and Supplementary Methods). We assumed that future ENSO-linked SST anomalies would remain robust indicators of productivity changes over the twenty-first century. This assumption was examined within the Geophysical Fluid Dynamics Laboratory’s Earth System Model (GFDL-ESM2.1; ref. 5), which included both physical climate and ocean biogeochemical dynamics. The SST anomaly remained highly correlated with large phytoplankton productivity throughout the 100-year projection to the year 2100 (r =−0.84, Fig. 1a). Reliance on SST as an indicator of large phytoplankton productivity allowed us to compare bias-corrected output 17 from 14 global climate models 6 that were assessed in the IPCC AR4 and select only those that captured observed covariance between SST anomalies in the eastern equatorial Pacific and air temperature/precipitation in northwestern Costa Rica to force 814 NATURE CLIMATE CHANGE | VOL 2 | NOVEMBER 2012 | www.nature.com/natureclimatechange © 2012 Macmillan Publishers Limited. All rights reserved.