Résumé
Ecological services provided by coral reefs are threatened by complex dynamics of global and local drivers. Fostering an appropriate management at local scale may enhance the resilience of the coral reefs facing large-scale issues, more laboriously manageable. We propose to model a bespoke reef health index (RHI) using a comprehensive dataset of socio-ecological variables and the Boosted Regression Trees (BRT). The variation of the RHI was primarily explained by the cluster of physiographic variables (55.45%), to a lesser extent, by land cover (36.42%), and finally by sociological variables (8.12%). The dispersion and the mean of the slope, as well as the built-up area increasingly contributed to the satisfactory modeling of the RHI (14.07, 14.12, 15.37%, respectively). These results will hopefully be useful for the stakeholders tasking with integrated coastal sustainability.