Parsimonious monthly rainfall-runoff models for humid basins with different input requirements

Abstract The availability of data is a major problem for the widespread application of rainfall-runoff models. In order to assist the practicing hydrologists in the selection of suitable models, four levels of monthly water balance models are developed requiring different sets of input data. Essentially four types of models are considered according to the inputs necessary, which are respectively: precipitation and potential evapotranspiration; precipitation, temperature and humidity; precipitation and temperature; precipitation alone. The outputs from the models are monthly river flow and other water balance components. The models are examined on 91 small and medium sized basins, of which 85 are situated in Belgium and 6 in southern China, and they are capable of reproducing both the magnitude and timing of monthly and seasonal runoff, as well as changes in soil moisture conditions. Results are satisfactory and there are barely any differences between the model types, although the amount of information used is very different. Calibration is by automatic optimization without any subjective element. Application in forecasting and Monte Carlo simulation are exemplified.