Watershed runoff prediction under semi arid conditions: application of semi analytical models

Abstract

In this study, a regionalized distributed parameter model, Agricultural Non-Point Source Pollution (AGNPS) and lumped parameter models, Soil Conservation Services - Curve Number (SCS-CN) and Rational Formula Methods, were utilized in predicting surface runoff (volume and peak flow rates) of specific rainstorm events in a semi-arid Watershed, Iiuni, Machakos, Kenya. The models were applied using input parameters derived from watershed characteristics (conditioning factors such as soil type, vegetative cover, slope, land use and causative factors like rainfall amount, distribution and duration). The watershed characteristics were obtained from a resource survey that was conducted and subsequent interpretations of topographic map, aerial photographs and Spot Imagery. Results of runoff volume prediction showed that SCS-CN and AGNPS models gave satisfactory results under the wet antecedent moisture condition (AMC III). The regression lines for the models were close to the 1:1 line with acceptable coefficients of determination (r). Likewise all statistics of comparison (goodness of fit criteria, R2;percentage error, %E; and t statistic) that were computed were acceptable with R2 being above 70% indicating good model predictions. When predicting peak runoff discharges, the AGNPS and Rational Formula models performed poorly and gave low values of R? ofless than 30%. This study provided some insight into semi arid watershed runoff dynamics through the determination of the distribution of runoff (e.g. direct runoff and baseflow) within the watershed. Analysis of watershed runoff data showed that a small fraction 10% (mean) of rainfall constituted direct runoff (rainfall excess) and direct runoff was a large fraction 66% (mean) of total runoff Consequently, most of the rainfall is retained within the watershed as expected due to the soil conservation measures in place. Therefore, it was possible to determine the proportion of storm runoff outflow (of large rainstorms) that could be conserved for dryland crop production within the watershed.