| Abstract: | A dataset of hourly sea surface temperature (SST) from the period 1 January 1982 to 31 December 2012, and covering the global ocean at a resolution of 0.3 0.3, was created using a validated ocean mixed-layer model (MLSST). The model inputs were heat flux and surface wind speed obtained from the Coupled Forecast System Reanalysis dataset. Comparisons with in-situ data from the Tropical Atmosphere Ocean array and the National Data Buoy Center showed that the MLSST fitted very well with observations, with a mean bias of 0.07C, and a root-mean-square error (RMSE) and correlation coefficient of 0.37C and 0.98, respectively. Also, the MLSST fields successfully reproduced the diurnal cycle of SST in the in-situ data, with a mean bias of -0.005C and RMSE of 0.26C. The 31-year climatology revealed that the diurnal range was small across most regions, with higher values in the eastern and western equatorial Pacific, northern Indian Ocean, western Central America, northwestern Australia, and several coastal regions. Significant seasonal variation of diurnal SST existed in all basins. In the Atlantic and Pacific basins, this seasonal pattern was oriented north-south, following the variation in solar insolation, whereas in the Indian basin it was dominated by monsoonal variability. At the interannual scale, the results highlighted the relationship between diurnal and interannual variations of SST, and revealed that the diurnal warming in the central equatorial Pacific could be a potential climatic indicator for ENSO prediction. |
