A fully-distributed HBV model (WFLOW_HBV) was applied for hydrological simulation in the River Suir catchment in Ireland compared to the currently used semi-distributed URBS model. The results show that WFLOW_HBV has high potential to enhance flood forecasting in the study area.
Application of a fully-distributed HBV model in the River Suir catchment in Ireland
Flooding has become a common occurrence yet a devastating hazard in the River Suir catchment in Ireland. Clonmel, a town located in the downstream part, is the most affected area. In response, the Office of Public Works developed the Advanced Flood Forecasting System (Suir-AFFS) to cope with impending flooding. The semi-distributed Unified River Basin Simulator (14-split URBS) has been applied in Suir-AFFS for flood forecasting and has produced promising results. However, the model is subject to recalibration after dry seasons and therefore it is unsustainable in the long term.
Realizing the importance of flood forecast and preventive measures, the main goals of this initial research project were to explore the performance of the novel fully-distributed HBV-96 model (WFLOW_HBV) on hydrological simulation in the River Suir catchment and its potential to enhance flood forecasting in Suir-AFFS. It was based on the hypothesis that a fully-distributed model might yield more accurate results compared to a semi-distributed model since it can incorporate spatial heterogeneities of the catchment better.
The calibration and validation results of WFLOW_HBV show outstanding performance for the downstream sub-catchments (Nash-Sutcliffe efficiencies ranging from 0.87 to 0.95 during a test period), particularly Clonmel (NSE 0.95). However, the model results for some upstream and responsive sub-catchments are unsatisfactory and should be improved. WFLOW_HBV with runoff parameters subject to soil types generate more accurate discharge results than that with runoff parameters subject to sub-catchments. The conceivable reason is that the runoff parameters were calibrated for all sub-catchments simultaneously. Comparing to the 14-split URBS, WFLOW_HBV produced more accurate discharge results in both short and long terms. This is particularly the case for Clonmel.
WFLOW_HBV shows high potential to be applied in Suir-AFFS. Recalibration of the runoff parameter set regarding the sub-catchment order is expected to improve the simulated discharges. At an advanced state, data assimilation is also recommended when implementing the model for real-time forecasting.