Estimation of runoff volumes and peak rates from rural catchments in Kenya

Abstract

The estimation of runoff volumes and peak rates of runoff was carried out in two rural catchments in Kenya. The catchments viz., Sambret (7.02 km2) and Lagan (5.44 km2) lie in the headwaters of the lake victoria drainage basin. Both were selected on the basis of availability of hydrometeorological data required for calibration and testing of the models. In Sambret catchment, out of the 41 rainfall-runoff events, 25 were used for model calibration while 16 events were used for testing. The calibrated and verified models were further tested in the neighbouring Lagan catchment using 18 events. Results from the runoff volume models showed that the USSCS (United states Soil Conservation Service) curve number and the API (Antecedent precipitation Index) models gave satisfactory results in both Sambret and Lagan catchments. However, SCS curve number performed better than API model. Nash model predicted the peak runoff rates better than kinematic'wave and SCS triangular hydrograph models in both Sambret and Lagan. Kinematic wave gave acceptable results in Sambret while in Lagan it under predicted the peak runoff rates. The SCS triangular hydrograph over predicted the peak discharges in both catchments indicating that it has limited applicability in the headwaters of lake victoria drainage basin in Kenya.