Diagnosing the impacts of landscape characteristics on hydrologic signatures in the Krycklan catchment in Sweden using a flexible hydrological model

Abstract

This study investigated the applicability and performance of a hydrological model (namely FLEXG) for simulating streamflow in boreal catchment, exemplified by the representative Krycklan catchment. The FLEXG simulated daily streamflow for the whole catchment and its sub-catchments. The Kling-Gupta Efficiency (KGE) and Nash-Sutcliffe Efficiency (NSE) demonstrated favorable model performance during the calibration period (KGE = 0.88 and NSE = 0.75) and the subsequent validation (KGE = 0.81 and NSE = 0.63) period for the Krycklan catchment. Notably, a high degree of congruence was observed across eight sub-catchments, with KGE values consistently exceeding the threshold of 0.7 during both calibration and validation periods. These findings underscore the FLEXG model's aptitude for simulating streamflow within boreal catchments, as evidenced by its ability to capture seasonal patterns across the catchment. Additionally, regression analysis was conducted to examine the intricate relationships between model parameters and varied catchment characteristics. These catchment characteristics were identified as pivotal factors influencing streamflow modelling in the study region. Specifically, it was established that catchment size shows a negative correlation with the splitter D (R2 = 0.7), a parameter intimately associated with both fast and slow recession periods. The proportion of forest cover was negatively correlated with the slow response reservoir recession coefficient, Kf (R2 = 0.57), whereas the proportion of wetland cover displayed a positive correlation with Ks (R2 = 0.54). These correlations underscore the substantive impact of land use patterns on streamflow generation dynamics within the boreal catchment context.