A partial contributing area model for linking rainfall simulation data with hydrographs of a small arid watershed

Four years of runoff measurement (29 events) in Kangussano, Mali, were analyzed on the basis of the partial area contribution concept. The study region is semiarid and the use of runoff water to satisfy crop consumption is of high importance. A first-order basin which included two flow channels draining an area of 1.14 X 10(6) ha (1.14 km(2)) was used for this purpose. The objective of this study was to develop and test a conceptual model to predict runoff on natural catchments of about this size. The model assumes that a basin is composed of a large number of pixels (satellite picture elements). Each pixel covers an area of 900 m(2). The hydraulic properties of a representative pixel are determined by runoff simulation experiments. The model calculates the number of runoff generating pixels at a given time and rain depth. The areal runoff is a product of these two factors. Analysis of satellite images from LANDSAT and SPOT indicated similarities of soil cover complexes between the study area and an experimental area in Upper Volta. Results of runoff simulation from this area were adapted for the study area and used for the calculations. A good agreement between predicted and measured contributing area was obtained when the number of runoff generating pixels was 80 pixels/mm effective rain. Predicted and measured hydrographs were also in good agreement. It was therefore concluded that the model can satisfactorily predict actual hydrographs from data generated by rainfall-runoff simulators. Dividing the watershed to surface elements of pixel's size makes the model capable of linking remote sensing information with simulation data in order to predict areal runoff