Stability-based Level-of-Service classification under snow-covered free-flow conditions using sliding-window variability and catastrophe theory

Snow-covered pavements pose substantial challenges to expressway operation in cold regions, where environmental factors rather than traffic demand often dominate safety risks. Conventional Level-of-Service (LOS) methods, primarily based on traffic volume and speed, are insufficient to reflect the impact of pavement snow conditions on traffic operational stability. To address this gap, this study develops an SCF-based stability classification framework using the snow coverage fraction (SCF) as a quantitative environmental indicator. A sliding-window variability analysis was applied to evaluate local speed dispersion of heavy and light vehicles under different SCF. Critical transition points in operational stability were identified, and three SCF thresholds (7.2%, 19.0%, and 29.3%) were determined to construct a four-level LOS classification system. Results show that speed dispersion increases significantly when SCF exceeds 20%, particularly for heavy vehicles, indicating elevated operational risk. At lower SCF levels, inconsistent driver responses lead to greater fluctuation in speed adjustment, highlighting the need for differentiated management strategies. The proposed classification provides quantitative support for snow removal prioritization and traffic monitoring. By incorporating pavement environmental conditions into LOS evaluation, the framework offers a practical and implementable tool for winter expressway management in cold regions, contributing to improved operational safety and more refined decision-making.