Spatially explicit modeling of road icing susceptibility using GIS and a multi-criteria decision approach: A case study of Gümüşhane city (Türkiye)

This study develops and validates a GIS-based, spatio-temporal model to identify and prioritize road segments with high icing susceptibility in the Gümüşhane city center, Türkiye. Icing on urban roadways constitutes a significant threat to vehicular and pedestrian safety, particularly in regions characterized by complex topography and dense urban morphology. A Multi-Criteria Decision Analysis (MCDA) framework, employing the Analytic Hierarchy Process (AHP) with structured expert input, was used to synthesize ten primary causative factors. These include solar radiation, temperature, aspect, object height, precipitation, elevation, pavement material, slope, albedo, and road width — each representing a distinct topographic, climatic, or morphological influence on urban road icing. The AHP quantitatively established the relative influence of these factors, identifying solar radiation (28.5%), temperature (18.2%), and aspect (13.5%) as the most dominant drivers. The model generates monthly susceptibility maps that reveal "urban canyons"—narrow streets flanked by high-rise buildings—consistently exhibit the highest and most persistent susceptibility due to severely limited solar radiation. The accuracy of the model's spatial predictions was quantitatively validated through field observations, with 91.4% of recurring icing locations correctly identified within high or very high susceptibility zones. This research culminates not only in a robust scientific framework but also in a validated decision-support tool that has been successfully integrated into the Gümüşhane City Information System (GCIS), where it is now operationally used to optimize winter maintenance strategies. Ultimately, this work provides a replicable, data-driven approach for enhancing urban resilience and traffic safety in challenging topographical settings.