Experimental Assessment of Runoff Generation Processes on Hillslope Scale in a Semiarid Region in Northern Tanzania

Abstract

In runoff modelling often saturated conductivities or infiltration rates for saturated surfaces are used to calculate the soil water balance and subsequently surface runoff. These infiltration rates are often too high and thus no or very little runoff is generated. However, after long dry periods even with small precipitation events surface runoff is observed. In this study we focus on the infiltration rates at different tensions to study their effect on surface runoff generation on a typical soil catena of northern Tanzania. The area is characterized by very little information on surface runoff and soil characteristics. Therefore, we measured soil infiltration and texture at the surface as well as overland flow volumes for individual rainfall events between October 10, 2010 and December 6, 2010 using a simple experimental setting. We examined rough water balance quantities for single rainfall events. A simplistic hydrological modelling of the surface runoff and accumulation was performed and compared to measured surface runoff. The study shows that infiltration at 0 water tension clearly overestimates infiltration quantities. However, even when soil characteristics like crusting of sealing are not considered, our simple approach yield much better surface runoff volumes using infiltration rates at higher soil water tensions instead of saturated infiltration conditions. The interaction between rainfall and soil surface conditions is relevant for understanding the hydrology of semiarid savannas with gentle slopes. However, we show that our approach integrating simple field measurements and basic hydrological budget models yield much better results than conventional approaches.