Coral reefs worldwide face unprecedented cumulative anthropogenic effects of interacting local human pressures, global climate change and distal social processes. Reefs are also bound by the natural biophysical environment within which they exist. In this context, a key challenge for effective management is understanding how anthropogenic and biophysical conditions interact to drive distinct coral reef configurations. Here, we use machine learning to conduct explanatory predictions on reef ecosystems defined by both fish and benthic communities. Drawing on the most spatially extensive dataset available across the Hawaiian archipelago-20 anthropogenic and biophysical predictors over 620 survey sites-we model the occurrence of four distinct reef regimes and provide a novel approach to quantify the relative influence of human and environmental variables in shaping reef ecosystems. Our findings highlight the nuances of what underpins different coral reef regimes, the overwhelming importance of biophysical predictors and how a reef's natural setting may either expand or narrow the opportunity space for management interventions. The methods developed through this study can help inform reef practitioners and hold promises for replication across a broad range of ecosystems.
Keywords: Hawai‘i; boosted regression trees; ecology; interactions; management; regime shift.