Articles | Volume 9, issue 4
Hydrol. Earth Syst. Sci., 9, 322–332, 2005
https://doi.org/10.5194/hess-9-322-2005

Special issue: Advances in flood forecasting

Hydrol. Earth Syst. Sci., 9, 322–332, 2005
https://doi.org/10.5194/hess-9-322-2005

  07 Oct 2005

07 Oct 2005

Precipitation forecasts and their uncertainty as input into hydrological models

M. Kobold and K. Sušelj M. Kobold and K. Sušelj
  • Environmental Agency of the Republic of Slovenia, Vojkova 1b, Ljubljana, Slovenia
  • Email for corresponding author: mira.kobold@gov.si

Abstract. Torrential streams and fast runoff are characteristic of most Slovenian rivers and extensive damage is caused almost every year by rainstorms affecting different regions of Slovenia. Rainfall-runoff models which are tools for runoff calculation can be used for flood forecasting. In Slovenia, the lag time between rainfall and runoff is only a few hours and on-line data are used only for now-casting. Predicted precipitation is necessary in flood forecasting some days ahead. The ECMWF (European Centre for Medium-Range Weather Forecasts) model gives general forecasts several days ahead while more detailed precipitation data with the ALADIN/SI model are available for two days ahead. Combining the weather forecasts with the information on catchment conditions and a hydrological forecasting model can give advance warning of potential flooding notwithstanding a certain degree of uncertainty in using precipitation forecasts based on meteorological models. Analysis of the sensitivity of the hydrological model to the rainfall error has shown that the deviation in runoff is much larger than the rainfall deviation. Therefore, verification of predicted precipitation for large precipitation events was performed with the ECMWF model. Measured precipitation data were interpolated on a regular grid and compared with the results from the ECMWF model. The deviation in predicted precipitation from interpolated measurements is shown with the model bias resulting from the inability of the model to predict the precipitation correctly and a bias for horizontal resolution of the model and natural variability of precipitation.