Sensitivity of marine heatwaves metrics to SST products, focusing on the Tropical Pacific

Abstract. Marine heatwaves (MHWs) are increasingly studied in climate sciences for their ecological impacts, for which accurate real-time bulletins and forecasts are essential. Yet, methodological choices in their detection affect metric estimates, underlining the need to better assess these sensitivities. This study provides a thorough assessment of the impact of Sea Surface Temperature (SST) product choice on MHW statistics, focusing on the tropical Pacific. MHW detection was performed on six daily gridded SST datasets: four widely used blended satellite observational products, one ocean reanalysis, and a multi-dataset ensemble mean computed from the four observational products. Sensitivity to SST products was evaluated for six MHW metrics (MHW days per year, number of events per year, duration, maximum intensity, cumulative intensity and onset rate) and for the Degree Heating Weeks (DHW), a widely used proxy for coral bleaching. Inter-product comparisons revealed a significant dispersion among MHW metric estimates, with the reanalysis GLORYS12v1 detecting fewer, longer and less intense MHWs while OISST detected more MHWs of shorter duration and higher intensity, likely related to their weak and strong high-frequency SST variability (periods shorter than 2 weeks) respectively. The sensitivity analysis showed that the onset rate was the most sensitive metric to SST product choice and the maximum intensity the most robust. Metrics uncertainties were quantified inside seven regions of the basin and were largest in the western Pacific Warm Pool. Co-occurrence analyses of MHWs revealed that, over the basin, 10 to 80 % of MHW days were detected simultaneously by all products, with the western Pacific Warm Pool showing the lowest agreement (10–40 %). Filtering MHWs by size also revealed that the detection of large-scale MHWs (> 5°x5°) was more consistent across products than smaller ones. Finally, over the studied period, inter-product differences tend to decrease with time. The DHW also revealed to be sensitive to SST products, with inter-product differences on DHW annual maximum reaching more than 1°C.weeks-1 and percentages of bleaching alert days (DHW ≥ 4°C.weeks-1) in common across products reaching 70 % at most across much of the basin. These findings contribute to a better understanding of how methodological choices affect the characterization of MHWs and DHW, and their associated uncertainties.