A hydrological model modified for application to flood forecasting in medium and small-scale catchments

Abstract

Steep headwater catchments in areas that experience intense rainfall can generate rapid floods. The significant risk to life and property posed by rapid floods means that they are a subject of great interest to flood forecasting. However, the rapid rise and short duration of flood hydrographs presents a difficulty for standard precipitation-runoff models, which generally operate on a daily time-step. The Hydrologiska Byrans Vattenbalansavdelning (HBV) model is a conceptual distributed hydrological model, normally applied to simulation of snow-melt hydrographs that can operate with an hourly time-step. In this paper, we modified the HBV model by adding a hillslope overland flow component so that the model would have more physical meaning when applied to prediction of floods humid climate zones. Hourly models for small-, medium-, and large-sized floods were calibrated and validated for flood events from the Tangjing catchment, a medium and small catchment within a humid subtropical climate zone of China. The calibration period utilized 50 events from the period 1980–1999, and the validation period utilized 20 flood events from the period 2000–2004. The model performance was evaluated using the Nash–Sutcliffe efficiency index and indicators of compliance of the modeled result with the observed result in terms of flood peak magnitude, flood runoff volume, and duration of flood peak. These indicators demonstrated good agreement between the observed and simulated hydrographs for the calibration and validation periods, achieving a reliability that was adequate for the purpose of flood forecasting in accordance with the relevant Chinese standard. We also applied the HBV model in its original configuration to the flood events, but its performance was generally inferior to that of the modified model. The modified model outperformed the original model mainly with respect to simulation of flood peak magnitude and flood peak timing.