Perceived drought exposure in small communities of the United States: Demographic, attitudinal, and metric-based...

Droughts in the U.S. are intensifying and, in many regions, becoming more frequent and longer under climate change, placing disproportionate pressure on small and under-resourced water systems. Yet, as with many environmental risks, public drought perception frequently diverges from technical assessments, weakening risk communication and complicating adaptation and policy efforts. However, the extent to which underlying factors shape this divergence remains underexplored. Using nationally representative survey data (N = 2551) linked to U.S. Drought Monitor (USDM) indicators, we examined how the public perceived drought exposure aligns with physical drought metrics and individual-level factors. We applied XGBoost as an interpretable machine learning model to identify the strongest environmental and cognitive predictors. Results showed that recent drought intensity, particularly cumulative intensity over the most recent two-year window, was the dominant predictor of perceived drought exposure. Age and religiosity emerged as the strongest sociodemographic predictors, with older and less religious individuals exhibiting higher perceptual accuracy. Notably, high self-reported familiarity with wastewater treatment and heightened concern over flooding were associated with misalignment between perceived and physical drought conditions, suggesting that technical knowledge and hazard sensitivity may not always enhance accuracy and may lead to cognitive biases. These findings suggest that drought awareness depends not only on physical exposure but also on how people process information based on personal beliefs and trust in institutions. Our integrated modeling approach provides a transferable framework for other national or regional contexts to enable cross-country comparisons of environmental risk perception and communication effectiveness. This work highlights the need for fundamentally rethinking the conventional top-down communication toward more contextual, trust-centered strategies for climate adaptation, particularly in vulnerable communities.