AI-Driven Predictive Modeling of Dust–Rainfall Interactions in Urban Arid Climates: A Comparative Study of Tehran and the Southwestern United States

Dust-induced alterations in rainfall pose a growing challenge for urban regions located in arid and semi-arid climates, where water scarcity, rapid urbanization, and intensified dust emissions intersect to create complex atmospheric dynamics. This study addresses the critical problem of understanding how severe dust events influence precipitation formation in major urban environments—focusing specifically on Tehran and the southwestern United States—and why accurate prediction of these interactions is essential for climate adaptation and water-resource management. Despite advances in aerosol research, the nonlinear and highly variable nature of dust–rainfall relationships has limited traditional statistical and physical modeling approaches, underscoring the need for new predictive frameworks.To overcome these limitations, this research employs an AI-driven modeling strategy integrating satellite-derived aerosol optical depth (AOD), ground-based PM2.5 and PM10. concentrations, reanalysis meteorological variables, and urban microclimatic indicators. Machine learning algorithms—including Random Forest, Gradient Boosting, and LSTM-based temporal networks—were used to detect hidden patterns in dust–cloud–rainfall interactions and to quantify the sensitivity of precipitation outcomes to varying dust loadings under distinct regional conditions. Comparative evaluation between Tehran and U.S. arid regions enables identification of both shared and region-specific atmospheric mechanisms.Key findings reveal that AI models successfully capture nonlinear dust–rainfall relationships, outperforming conventional regression methods. Tehran demonstrates stronger rainfall suppression during high-intensity dust events, likely driven by enclosed topography and high anthropogenic aerosol mixing, whereas the U.S. Southwest exhibits mixed outcomes, including rainfall enhancement under moist monsoon conditions.This study demonstrates that AI-driven predictive modeling provides a powerful framework for understanding and forecasting dust-impacted rainfall in urban arid climates. The results highlight significant implications for water-resource planning, urban resilience, and the development of next-generation climate-adaptive forecasting systems