Spatially varying impacts of sea surface temperature on coral bleaching: A geographically weighted regression approach.

Coral reefs are facing bleaching threats due to climate change. Previous analyses primarily quantify this risk by assessing the average global menace. However, various local factors amplify the effects of climate change in some regions, making the phenomenon spatially heterogeneous. Thus, this study examines the spatially varying effects of sea surface temperature (SST) on coral bleaching using Geographically Weighted Regression (GWR) to better understand regional variations in coral vulnerability. This machine learning algorithm incorporates geo-localization of observations to capture regionally varying relationships in the data, offering insights into the geographical patterns of the local factors influencing SST effects. The analyses use Coral Check's 7941 globally distributed observations of coral health, collected by professional scientists as well as trained and certified citizen scientists. These observations were assembled by marine experts and followed a standardized transect protocol. The research identifies areas most vulnerable to temperature-induced coral bleaching. First, stationary models revealed a statistically significant relationship between SST and coral bleaching, highlighting the critical global impact of temperature on coral reefs. Second, the GWR emphasizes that the most sensitive to temperature-induced bleaching are in Southern Africa and Southeast Asia. Third, this study predicts the implications of this impact using IPCC's representative concentration pathways of climate change, namely RCP 4.5 and RCP 8.5. The estimates reveal that by 2050 several seas around the equator will experience the highest levels of temperature-led coral bleaching. The findings underscore the need for a differentiated approach under the Paris Agreement on climate change to address coral reef bleaching and identify hotspot regions where targeted assistance is necessary.