Comparison of in situ and satellite sea surface temperature data from South Australia and Tasmania: how reliable are satellite data as a proxy for coastal temperatures in temperate southern Australia?

Satellite sea surface temperature (SST) is widely used for biological modelling and ecological studies assuming it represents sub-surface in situ temperature (IST). We test this assumption at 32 coastal sites in southern Australia spanning a wide geographic range. Annual IST regimes are described and demonstrated to be highly correlated with SST. Mean annual deviations were generally small, varying spatially and seasonally (range 0-1ºC). No correlation between deviations and a range of site attributes was found, indicating the importance of site-specific factors. Seasonal deviations were not geographically consistent, being higher in South Australia during the summer (mean 1.4 ºC) than in Tasmania (mean 0.5 ºC). Generally small annual mean deviations between SST and IST justify using SST for broad scale ecological and climate change studies, but considerable deviations at some sites suggests using SST at smaller spatial and temporal scales is unlikely to be appropriate. In addition, SST data lack information on daily temperature fluctuation that may be biologically relevant. Excepting South Australia where spatially consistent summer deviations would allow a correction factor, this site-specific variation is hard to correct. In spite of this, studies that rely on SST should consider the implications of such variation on the level of certainty associated with temperature-based predictions.