Assessment of GHRSST and OISST Datasets in Identification of Marine Heat-Waves and Heat-Spikes

Abstract

This study assesses the performance of two satellite-derived products, viz. Group for High-Resolution Sea-Surface Temperature (GHRSST) and Optimum Interpolation Sea-Surface Temperature (OISST), in detecting the Marine Heat-Waves (MHWs) and Marine Heat-Spikes (MHSs) after a detailed validation with in situ instruments in the Bay of Bengal (BoB). A significantly good correlation and low Root Mean Square Error (RMSE) between GHRSST (OISST) and the Research Moored Array for African-Asian-Australian Monsoon Analysis and Prediction buoy at 15�N and 90�E (RM15) were found to be 0.97 (0.96) and 0.30 (0.35)�C, respectively. In comparison with three near-coast Wave-Rider-Buoys (WRBs), GHRSST (OISST) shows limited ability to reproduce daily temperature variability with RMSEs ranging from 0.61 �C to 1.07 �C (0.51 �C-1.08 �C), respectively. GHRSST captures signals below 1-2 month periodicity better than OISST, however, both products efficiently capture low-frequency variability (> 4 months). Finally, regarding the number of MHW counts and duration, the GHRSST (36 and 382 days) and OISST (33 and 294 days) overestimate these quantities with respect to RM15 (25 and 287 days). This results in RM15 having a higher duration per MHW event (11.48 days) than GHRSST (10.61 days) and OISST (8.91 days). GHRSST exhibited its highest counts (duration) of 7 (82 days) during 2020 (2010), whereas RM15 showed only 3 (55 days) counts (duration) during 2020 (2010). Moreover, the mean daily intensities from satellite-derived SSTs (0.86 �C) are comparable to RM15 (0.87 �C). Therefore, our study suggests that the satellite-derived SSTs are suitable to capture MHW characteristics in the open-ocean region of the BoB, but their use in the coastal regions should be judiciously done. � 2004-2012 IEEE.