Abstract
We investigated projected changes in sea surface temperature (SST) and the associated impacts on spawning habitat for skipjack tuna (Katsuwonus pelamis) in the Coral Triangle region (CT). A multimodel aggregate of SST CMIP5 models for the CT region, based on a comprehensive skill validation assessment, was used to identify the five best performing of 36 models tested for inclusion in a regional multimodel ensemble. Monthly 1 degrees SST multimodel aggregate projections for the CT region under RCP8.5 show that increases in SST, as high as 2.8 degrees C (mean value), will likely occur by the end of this century. Using these estimates of SST change, we applied three parameterizations of skipjack tuna spawning temperatures to assess the potential for change in spawning habitat within the CT region. The three spawning temperature parameterizations were as follows: (a) a square-wave function derived from catch data with boundaries at 26 and 30 degrees C; (b) a symmetric Gaussian function derived from the SEAPODYM models; and (c) an asymmetric Gaussian function that modifies the SEAPODYM curve in (b) to include the results of relevant physiological experiments. All three parameterizations show similar geographic patterns, with the amount of favourable spawning habitat decreasing throughout the central, equatorial CT region and increasing at higher latitudes. However, the three parameterizations show marked differences in the modelled magnitude of change, with an asymmetric Gaussian function (ASGF) showing a regionwide average of 66.1% decline in favourable spawning habitat between 2015 and 2099. These projected changes in tuna spawning habitats are likely to have important consequences on local and regional fisheries management in the CT region.