Home > Research > Publications & Outputs > Flood frequency prediction for data limited cat...
View graph of relations

Flood frequency prediction for data limited catchments in the Czech Republic using a stochastic rainfall model and TOPMODEL.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published
<mark>Journal publication date</mark>08/1997
<mark>Journal</mark>Journal of Hydrology
Issue number1-4
Volume195
Number of pages23
Pages (from-to)256-278
Publication StatusPublished
<mark>Original language</mark>English

Abstract

A continuing problem in hydrology is the estimation of peak discharges for design purposes on catchments with only limited available data. A promising and elegant approach to this problem is the derived flood frequency curve pioneered by Eagleson (1972, WaterResour. Res., 8(4): 878–898). A number of studies using this approach have been published over the last 20 years but only a few have compared the predicted curves with observations. One exception used a simple stochastic rainfall model to drive a version of TOPMODEL (Beven, 1987, Earth Surf Processes Lardforms, 12: 69–82). The present study describes a new version of the stochastic rainfall simulator previously used with TOPMODEL and its application on three small catchments (1.87, 4.75 and 25.81 km2) in the Jizera Mountains in the Czech Republic. The rainfall model differentiates between high and low intensity events. The resulting rainfall statistics were checked by comparisons with measured data. The flood frequency curves predicted by the combined model were constrained by the regional estimates or a measured series for short return periods and used to predict longer return period flood magnitudes. Only one TOPMODEL parameter has to be adjusted—an effective average transmissivity. For the two smaller catchments also a rainfall parameter has to be adjusted depending on the size of the catchment. It is shown that the random sequence of rainstorms can have a significant effect on the predicted 100 year return period event, even for 1000 year simulations and without allowing for uncertainty in the rainfall model parameters.