Predicting hourly flows at ungauged small rural catchments using a parsimonious hydrological model
MetadataShow full item record
Streamflow data is important for studies of water resources and flood management, but an inherent problem is that many catchments of interest are ungauged. The lack of data is particularly the case for small catchments, where flow data with high temporal resolution is needed. This paper presents an analysis of regionalizing parameters of the Distance Distribution Dynamics (DDD) rainfall-runoff model for predicting hourly flows at small-ungauged rural catchments. The performance of the model with hourly time resolution has been evaluated (calibrated and validated) for 41 small gauged catchments in Norway (areas from 1 km2–50 km2). The model parameters needing regionalization have been regionalized using three different methods: multiple regression, physical similarity (single-donor and pooling-group based methods), and a combination of the two methods. Seven independent catchments, which are not used in the evaluation, are used for validation of the regionalization methods. All the three methods (the multiple regression, pooling-group, and combined methods) perform satisfactorily (0.5 ≤ KGE < 0.75). The combined method (which combines multiple regression and pooling-group) performed slightly better than the other methods. Some model parameters, namely those describing recession characteristics, estimated by the regionalization methods, appear to be a better choice than those estimated locally from short period of hydro-meteorological data for some test catchments. The single-donor method did not perform satisfactorily. The satisfactory performance of the combined method shows that regionalization of DDD model parameters is possible by combining multiple regression and physical similarity methods.