A winter flash flood in Skenderaj, Kosovo: heavy rainfall analysis and early warning perspectives

Flash floods pose a significant risk to infrastructure in Kosovo, particularly in urban and riverine areas. This research focuses on an intense river flood event that took place on January 19, 2023, in the Skenderaj catchment. The study’s main goal was to establish a flash flood early warning system by combining sophisticated atmospheric modeling, hydrological evaluation, and rainfall hazard analysis. The ARW model with 2-km resolution effectively captured rainfall intensity and local flood occurrences, particularly around Skenderaj and Istog, whereas the 4-km NMM model better represented wider spatial precipitation patterns. Hydrological results demonstrated that precipitation strongly dictated river discharge and runoff dynamics, with the highest flows recorded in northern Albania. To validate and enhance forecast accuracy for flash flood warnings, datasets from the Global Flood Awareness System (GloFAS), the European Flood Awareness System (EFAS), and ERA5 reanalysis were incorporated. These resources prov ided essential information on antecedent conditions, such as soil moisture and snowmelt, which substantially influenced runoff and flood magnitude. The ECMWF Copernicus framework also contributed by supplying 24-hour river discharge forecasts for Kosovo’s basins, aiding in timely and spatially detailed flood alerts. The Novel Thunderstorm Alert System (NOTHAS) was updated to integrate crucial hydrological variables - including surface and convective runoff, snow water equivalent, soil moisture, and slope - thereby enhancing the precision of flood warnings. This improved system enabled effective classification of flood risk zones, thus identifying areas vulnerable to flash floods and landslides. The study highlights the crucial role of high-resolution weather modeling, hydrological insights, and integrated early warning systems in enhancing flash flood prediction and mitigation efforts.