Estimating storm-related coastal risk in Mexico using Bayesian networks and the occurrence of natural ecosystems

This study develops an integrative and thorough assessment of storm-related coastal risk in Mexico. Through Bayesian networks (BN), we explore an approach that blends an ecological understanding of coastal dynamics and concepts of coastal risk based on hazards, exposure, and vulnerability. We used two approaches. First, we calculate a coastal risk index based on coastal hazards (storms, flooding potential and shoreline change rate), exposure (occurrence and size of human settlements), and vulnerability (extension of natural ecosystems which provide storm protection: mangroves and coastal dunes). Then, we used a Bayesian network approach to perform a causal-driven and spatially explicit probabilistic assessment of storm-induced risks. This study shows that the population living on the coast has grown drastically over the last decades, resulting in more significant exposure of people and assets to risk. We showed the importance of natural ecosystems in reducing coastal risk and that the highest coastal risk was associated with (i) the reduced cover (or absence) of mangroves and coastal dunes, (ii) the conservation status of vegetation, and (iii) higher population density. Overall, the study shows that utilizing a Bayesian network is an effective tool for exploring coastal risk scenarios.