Measuring the effects of extreme weather events on pavement performance

This study investigates the effects of extreme weather events, particularly prolonged droughts and excessive heat, on pavement deterioration across Texas. By analyzing the interplay between severe environmental stressors and transportation infrastructure resilience, the research highlights how climatic extremes accelerate pavement degradation, emphasizing the urgency of climate-adaptive infrastructure planning. Methodologically, this study combines traditional statistical modeling and machine learning approaches to assess both time-invariant and time-varying factors influencing pavement performance. A dynamic panel data model is employed to capture temporal dependencies, while machine learning techniques, including random forest, support vector machine, XGBoost, and a deep learning encoder-decoder architecture, are utilized to identify nonlinear relationships and predictive patterns. The research synthesizes two extensive datasets: (1) pavement condition data records from the Texas Department of Transportation’s (TxDOT) Pavement Management Information System (PMIS), comprising over 500,000 observations of distress, roughness, and structural integrity from 2000 to 2018; and (2) meteorological data from 6,037 National Oceanic and Atomospheric Administration (NOAA) weather stations, including daily temperature extremes (e.g., summer days exceeding 100℉ (37.8℃)) and precipitation metrics such as prolonged dry spells. Key findings reveal significant adverse effects of extreme heat and drought conditions on pavement health. These results underscore the critical need for integrating climate resilience into pavement design, maintenance, and policy frameworks. By quantifying the vulnerability of infrastructure to environmental stressors, this study provides actionable insights for mitigating climate-induced degradation and enhancing the longevity of transportation systems in regions prone to climatic extremes.