AbstractAimLarge‐scale variation in species richness results from multiple environmental controls. We proposed to identify the main factors that influence species richness of reef corals to provide insight into natural forces and their implications for future climate impacts.LocationIndian Ocean.MethodsWe applied quantile regression (QR) to predict coral species richness and to evaluate the influence of reef area, biogeographical connection – represented by reef area up‐current, geological provinces defined by plate tectonics, latitude, longitude and mainland‐island separation, the physico‐chemical factors of photosynthetically active radiation (PAR), salinity and primary productivity and seven properties of sea surface temperature (mean, minimum, maximum, standard deviation, range, kurtosis and skewness) on coral species richness. Predictions at three quantiles (0.50, 0.75 and 0.95) were compared against ordinary least squares regression (OLS) analysis.ResultsAll four models successfully predicted species richness, with the 95% QR model showing the best fit. According to this model, reef area, number of reefs up‐current and mean‐SST had positive effects while maximum and kurtosis‐SST had negative effects on richness. The distribution of observed richness values in relation to the regression line suggested this model revealed the main limiting factors. High predicted richness sites were located in the region from Western Australia to Central Indian Ocean Islands and southern India, the Mozambique Channel and East Africa and Southern Red Sea and Gulf of Aden. Sites with low predicted richness were Arabian/Persian Gulf and Gulf of Oman, South Africa and south‐west Madagascar, Gulf of Kutch, Bay of Bengal and the Mascarenes.ConclusionsTemperature properties played a prominent role in influencing species richness mainly as latent energy and stress rather than thermal stability. These thermal properties indicate the conditions needed to promote coral diversity and selecting climate refugia for conservation management.