Simulation of flood hydrographs based on lumped and semi-distributed models for two tropical catchments in Kenya

The design and operation of flood management systems require computation of flood hydrographs for both design floods and flood forecasting purposes, since observed data are usually inadequate for these tasks. This is particularly relevant for most developing countries, i.e. mainly for tropical catchments. One possible way of obtaining information about flood hydrographs is through the use of rainfall-runoff models. Two such models, namely the Bochum model and the Nash Cascade-Diskin Infiltration model, which are semi-distributed and lumped models, respectively, were used in the present study. These models were applied to two catchments in Kenya with drainage areas of 6.71 km2 and 26.03 km2. A set of 13 selected rainfall-runoff events was used to calibrate and validate the models. The physical parameters required by the models were derived from catchment characteristics using GIS and remote sensing data while the conceptual parameters were obtained by optimization. The flood hydrographs simulated using the parameters so derived indicated that it is possible to use the two models in this tropical environment.