Continuous spatially distributed simulation of surface and subsurface hydrological processes in a small semiarid catchment

Abstract

AbstractModelling the hydrologic processes of semiarid basins is specially challenging due to the specific characteristics of this type of environment. The discontinuity of the processes both in space and in time produces a highly complicated rainfall–runoff relationship. Runoff production and reinfiltration areas produce a discontinuous spatial pattern of runoff; channel flow is ephemeral and restricted to the wet season of the year or to heavy rainfalls, and many local saturated layers of soil separated by unsaturated areas can occur. The hydrological connection of the different parts of the basin produces a nonlinear and complex hydrologic behaviour because different response mechanisms of the basin are triggered as the hydrological connectivity increases. Similarly, as connectivity decreases, so does the information arriving at the monitored state variables used to calibrate the model and to understand the system. This paper describes the application of a distributed hydrologic model to describe the hydrologic behaviour of a small semiarid basin including surface and subsurface processes. The model is parameterized and calibrated to simulate a period of 71 days covering the transition of the basin from a dry to a wet state, during which most of the hydrologic processes of the basin are activated. The calibration strategy used different sources and types of information on the state of the basin and an efficient gradient based search algorithm to find the optimal values of the most sensitive parameters. The modelled hydrograph and the modelled internal description of the basin were analysed in the light of its known behaviour and the existing knowledge of semiarid hydrology reported in the literature. The results indicate that while the integrated behaviour of the basin represented by the output hydrograph can be well reproduced, the subsurface processes and the internal dynamics of the basin are less accurately simulated, mainly due to the lack of information on the geometry of the subsurface domain (soil depths and bedrock topography) and on the spatial distribution of the parameters. Copyright © 2007 John Wiley & Sons, Ltd.