Regionalizing short-term rainfall affected by topography in semiarid Tunisia

The characteristics of fine time-scale rainfall are important in many hydrological applications, such as infiltration, erosion and flooding. The spatial properties of such rainfall are, however, seldom known, especially for arid and semi-arid areas. A better knowledge of fine time-scale rainfall and also comparison with daily rainfall may yield possibilities for disaggregation. For this purpose, rainfall data of different time scales, from 1-min to daily, from 25 stations during four years (1995–1998), were spatially analysed by means of spatial correlation, empirical orthogonal function (EOF) and hierarchical clustering. The results show that the spatial correlation is typically non-isotropic and varying, depending on topography and local meteorological settings. Similarly, spatial patterns of EOF are closely related to main atmospheric synoptic situations as influenced by orography and spatial dependence regarding areas with predominant convective and frontal rainfall. The clustering displayed different homogeneous sub-groups over the Tunisian Dorsal Mountains that can be used to better manage the limited water resources that often depend on fine time-scale rainfall variability