Dynamic coastal vulnerability index: A machine learning approach to predict future impacts of climate change and human activity on coastal environments

Coastal regions are increasingly vulnerable to the compounded impacts of climate change and human activities, necessitating the development of more dynamic and predictive assessment frameworks. This study introduces a Dynamic Coastal Vulnerability Index (DCVI), employing machine learning techniques to forecast future changes in coastal vulnerability along the shoreline of Mazatlán, Mexico. By integrating physical, environmental, and socio-economic indicators, the model utilizes a Random Forest algorithm to assign weights objectively and capture complex, non-linear relationships among variables. Initial vulnerability classification was performed using ranking and weighted overlay methods, subsequently refined through machine learning-based optimization to enhance predictive accuracy. The spatial analysis indicates that, out of a total area of 3068.50 km², approximately 56.06 % of the region is classified as moderately vulnerable, 27.71 % as low vulnerability, 5.38 % as very high vulnerability, 3.59 % as high vulnerability, and 7.28 % as very low vulnerability. These results demonstrate a heterogeneous distribution of vulnerability across the study area, emphasizing the urgent need for targeted adaptation and resilience strategies. The findings highlight the advantages of dynamic, machine learning-enhanced methodologies over conventional static models for coastal vulnerability assessments. The DCVI framework developed in this study offers an advanced decision-support tool for policymakers, urban planners, and coastal managers aiming to mitigate the anticipated impacts of climate change and anthropogenic pressures on coastal environments. The primary intention of this research is to develop a dynamic, predictive model for coastal vulnerability assessment and to inform adaptive coastal management strategies under future environmental and socio-economic scenarios.