The accuracy of two modified versions of a simple real-time semi-distributed flood forecasting model (RS model) is investigated. These are adaptive rainfall-runoff models obtained from the RS model by modifying the rainfall loss component, which is based on a combination of the runoff coefficient and Φ-index approaches, throughout an ensemble of homogeneous zones into which the basin is sub-divided. Both models assume the total loss of all rainfall in an initial period determined by the two early points of the hydrograph rising limb. The first version (RS1 model) assumes a subsequent constant loss rate while the other (RS2 model) assumes a subsequent loss rate proportional to the rainfall rate. A comparison of the different models was performed through their application to events observed on two Italian basins. The results revealed a significant sensitivity of flow forecasts to the variations in loss representation, particularly to the selection of the first rainfall producing direct runoff. The RS and RS2 models provided reasonably appropriate flood forecasting for lead times up to 6h.
A semi-distributed adaptive model for flood forecasting with different representations of losses
CORRADINI, Corrado;
1992
Abstract
The accuracy of two modified versions of a simple real-time semi-distributed flood forecasting model (RS model) is investigated. These are adaptive rainfall-runoff models obtained from the RS model by modifying the rainfall loss component, which is based on a combination of the runoff coefficient and Φ-index approaches, throughout an ensemble of homogeneous zones into which the basin is sub-divided. Both models assume the total loss of all rainfall in an initial period determined by the two early points of the hydrograph rising limb. The first version (RS1 model) assumes a subsequent constant loss rate while the other (RS2 model) assumes a subsequent loss rate proportional to the rainfall rate. A comparison of the different models was performed through their application to events observed on two Italian basins. The results revealed a significant sensitivity of flow forecasts to the variations in loss representation, particularly to the selection of the first rainfall producing direct runoff. The RS and RS2 models provided reasonably appropriate flood forecasting for lead times up to 6h.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
