Articles | Volume 6, issue 3
Nat. Hazards Earth Syst. Sci., 6, 407–417, 2006

Special issue: Advances in radar, multi-sensor and hydrological modelling...

Nat. Hazards Earth Syst. Sci., 6, 407–417, 2006

  24 May 2006

24 May 2006

The use of HBV model for flash flood forecasting

M. Kobold1 and M. Brilly2 M. Kobold and M. Brilly
  • 1Environmental Agency of the Republic of Slovenia, Vojkova 1b, 1000 Ljubljana, Slovenia
  • 2Faculty of Civil and Geodetic Engineering, University of Ljubljana, Jamova 2, 1000 Ljubljana, Slovenia

Abstract. The standard conceptual HBV model was originally developed with daily data and is normally operated on daily time step. But many floods in Slovenia are usually flash floods as result of intense frontal precipitation combined with orographic enhancement. Peak discharges are maintained only for hours or even minutes. To use the HBV model for flash flood forecasting, the version of HBV-96 has been applied on the catchment with complex topography with the time step of one hour. The recording raingauges giving hourly values of precipitation have been taken in calibration of the model. The uncertainty of simulated runoff is mainly the result of precipitation uncertainty associated with the mean areal precipitation and is higher for mountainous catchments. Therefore the influence of number of raingauges used to derive the areal precipitation by the method of Thiessen polygons was investigated.

The quantification of hydrological uncertainty has been performed by analysis of sensitivity of the HBV model to error in precipitation input. The results show that an error of 10% in the amount of precipitation causes an error of 17% in the peak of flood wave. The polynomial equations showing the relationship between the errors in rainfall amounts and peak discharges were derived for two water stations on the Savinja catchment. Simulated discharges of half-yearly runs demonstrate the applicability of the HBV model for flash flood forecasting using the mesoscale meteorological forecasts of ALADIN/SI model as input precipitation data.