Deep Learning–Enabled Climate Justice Analytics: Integrating Water Contamination, Flood Hazard, and Health Outcome Disparities in the U.S. Gulf Coast

Abstract

Communities along the U.S. Gulf Coast are increasingly exposed to compound climate hazards, including drinking water contamination and flooding, with disproportionate health impacts on socioeconomically vulnerable populations. While prior studies often assess these risks in isolation, limited attention has been given to their combined effects on public health within an environmental justice framework. This study presents a deep learning–enabled climate justice analytics framework that integrates multi-source datasets on water contamination, flood hazards, health outcomes, and demographic vulnerability across Gulf Coast census tracts.

Authoritative federal datasets were compiled, including the U.S. Environmental Protection Agency Safe Drinking Water Information System, Federal Emergency Management Agency flood hazard maps, National Oceanic and Atmospheric Administration climate exposure data, Centers for Disease Control and Prevention health indicators, and U.S. Census–based socioeconomic metrics. A multi-modal, multi-task deep learning architecture was developed to jointly model environmental hazards and predict health burden indicators, while interpretable machine learning techniques were applied to identify dominant drivers of disproportionate risk.

Results reveal pronounced spatial clustering of multi-hazard exposure and health disparities, with low-income and minority communities experiencing significantly higher cumulative burdens. Model interpretability analyses highlight the synergistic influence of flood frequency and drinking water violations on adverse health outcomes. Scenario-based simulations further demonstrate how changes in flood intensity or contamination mitigation strategies may alter inequity patterns.

By translating complex environmental data into decision-ready indicators, this framework supports climate resilience planning and aligns with federal environmental justice initiatives such as Justice40. The approach is generalizable to other regions and hazard combinations, offering a scalable tool for evidence-based climate justice policy and investment prioritization.