Assessing the water balance in the Sahel: Impact of small scale rainfall variability on runoff. Part 2: Idealized modeling of runoff sensitivity

As in many other semi-arid regions in the world, the Sahelian hydrological environment is characterized by a mosaic of small endoreic catchments with dry soil surface conditions producing mostly Hortonian runoff. Using an SCS-type event based rainfall–runoff model, an idealized modeling experiment of a Sahelian environment is set up to study the sensitivity of runoff to small scale rainfall variability. A set of 548 observed rain events is used to force the hydrological model to study the sensitivity of runoff to the time and space variability of rainfall input. The rainfall time variability sensitivity analysis shows that preserving the event rain depth without representing the main variabilities of the hyetograph intensities can translate into a runoff error of 65% in the worst case. On a virtual mosaic of 1-km2 catchments covering 10,000 km2, the simulated runoff shows a high sensitivity to a decrease of the spatial resolution of event rain fields from 1 × 1 km2 to 100 × 100 km2. For the catchments characterized by low runoff coefficients, which are the most sensitive to rainfall variability, at the coarsest spatial resolution of 100 × 100 km2, the global runoff computed from the 548 events is underestimated by 50% with respect to the runoff simulated from the 1 × 1 km2 resolution rain fields. The threshold resolution of 20 km was identified as a characteristic spatial scale, over which the performance of the model rapidly decreases. Looking at the influence of the number of available rain gauges, the effect of spatial aggregation depends on the density of the rain gauge network with lower effect for sparser networks.