Variability of precipitation for flood modelling in mountainous basins – Use
of weather radar data
The study addresses the effects of spatial and temporal variable convective precipitation on
discharge simulation in a mountainous catchment. Therefore, the process-oriented catchment
model TACD (see above) and different rainfall scenarios generated by different regionalization
methods were applied. In addition, weather radar data from an operational C-band radar were
adjusted using ground stations and the basin precipitation was calculated and use for catchment
modeling. The significance of the spatial and temporal variability of the precipitation input for
flood modeling in mountainous environments was clearly demonstrated. Radar data can help to
capture this variability of the precipitation, particularly for events with very unequal spatial rainfall
distribution. But they need to be adjusted using several reliable ground stations directly within or
nearby the investigated basin. For this a suitable methodology is proposed.

Key publications:

  • Uhlenbrook S., Tetzlaff D., 2004: Significance of spatial and temporal variability of convective
    precipitation for flood modeling in mountainous basins. In: Hydrological and meteorological coupling
    in mountain areas: Experiments and modeling (Hrsg. De Jong C., Collins D., Ranzi R.), 22 pages,
    Springer Verlag, in press.

  • Tetzlaff D., Uhlenbrook S., 2004: Appropriate data resolution and model complexity capturing
    distributed flood formation, J. of Hydrology, in prep.
The use of weather radar to examine the significance of the spatial and
temporal variability of precipitation for flood modelling in mountainous basins

by Stefan Uhlenbrook and Dörthe Tetzlaff

This study addresses the effects of spatial and temporally variable rainfall input data
on discharge simulation in a mountainous catchment. The process-oriented catchment
model TAC-D was applied successfully to the meso scale Brugga basin (40 km²) and the
sub-basin St. Wilhelmer Talbach (15.4 km²) located at Southern Black Forest Mountains,
southwest Germany. Then the calibrated model was run with two different precipitation
input data sets for three storm events:

  1. Basin precipitation was calculated using up to seven ground stations and a regionalization
    method that combines the inverse distance weighting method with an elevation
    gradient (80:20).

  2. Weather radar data from a C-band radar were adjusted using ground stations and the
    basin precipitation was calculated.

This resulted in quite different precipitation patterns, in particular for the storm cells with
limited spatial extent. Both precipitation data sets were transferred to 50 x 50 m² grid, which
served as spatial discretization of the TACD model. For the radar data an algorithm developed
by Jens Lange was used. The significance of the spatial and temporal variability of the
precipitation input for flood modelling in mountainous environments is clearly demonstrated.
Incorrect basin precipitation data can lead to large errors in flood modelling. Radar data can help
to capture this variability of the precipitation, particularly for events with very unequal rainfall
distribution. But they need to be adjusted using several ground stations. For this a suitable
methodology is proposed. Finally the use of radar data for hydrological modelling in mountainous
basins is discussed.
zum IHF