Akademik Veri Yönetim Sistemi
8th International Congress of Thrace Scientific Research, Edirne, Türkiye, 20 - 21 Aralık 2025, ss.497-506, (Tam Metin Bildiri)
Reliable runoff prediction in ungauged basins (PUB) remains a major challenge in hydrology. This study investigates the contribution of landscape metrics in hydrological parameter regionalization using the conceptual-lumped GR6J rainfall-runoff model. Two catchments in the Western Regions of Türkiye (E524 and D354) were selected as a donor-receiver pair and were separately calibrated using daily precipitation and potential evapotranspiration data. The calibration of the GR6J Models achieved good performance with Nash-Sutcliffe Efficiency, (NSE) of 0.67 for both catchments. Calibrated hydrological models were both cross-validated, and were validated using regionalized parameters by landscape metrics to comprehend their effects on hydrological parameters. Landscape metrics were extracted using CORINE Land Cover data. The mean perimeter-area ratio (PARA_MN) and mean shape index (SHAPE_MN) of two catchments were processed in the FRAGSTATS software. While calibration process provided a reliable model predictions cross-validation resulted in poor performance with NSE < 0.40. The regionalized parameters yielded satisfactory model performances of NSE ranging between 0.55 and 0.57, which consistently outperformed direct parameter transfer through spatial cross validation. While findings are exploratory due to the limited number of catchments and the use of a linear regionalization framework, this study provides evidence that landscape metrics can contribute useful hydrological information for PUB practices.