Articles | Volume 6, issue 2
Nat. Hazards Earth Syst. Sci., 6, 229–236, 2006
https://doi.org/10.5194/nhess-6-229-2006

Special issue: Modelling, computer-assisted simulations and mapping of dangerous...

Nat. Hazards Earth Syst. Sci., 6, 229–236, 2006
https://doi.org/10.5194/nhess-6-229-2006

  19 Apr 2006

19 Apr 2006

The use of radar in hydrological modeling in the Czech Republic – case studies of flash floods

M. Šálek1, L. Brezková1, and P. Novák2 M. Šálek et al.
  • 1Czech Hydrometeorological Institute, Brno, Czech Republic
  • 2Czech Hydrometeorological Institute, Praha, Czech Republic

Abstract. Flash flood induced by severe convection is the hydrometeorological phenomenon that is very difficult to forecast. However, the implementation of radar measurements, especially radar-based Quantitative Precipitation Estimate (QPE) and/or radar-based quantitative Precipitation Nowcast (QPN) can improve this situation. If the radar is able to capture the development of severe convection and can produce reasonably accurate QPE in short time intervals (e.g. 10 min), then it can be used also with hydrological model.

A hydrological model named Hydrog was used for investigation of simulation and possible forecasts of two flash floods that took place in the Czech Republic in 2002 and 2003. The precipitation input consisted of mean-field-bias-adjusted or original radar 10-min estimates along with quantitative precipitation nowcasts up to 2 h based on COTREC method (extrapolation). Taking into account all the limited predictability of the severe convection development and the errors of the radar-based precipitation estimates, the aim of the simulations was to find out to what extend the hydrometeorological prediction system, specifically tuned for these events, was able to forecast a the flash floods. As assumed, the hydrometeorological simulations of the streamflow forecasts lagged behind the actual development but there is still some potential for successful warning, especially for areas where the flood hits lately.

Download
Altmetrics