Runoff modelling in northern Algeria using a distributed physically-based model integrating remote sensing data

This paper describes the possibility of using remote sensing data in order to obtain the information necessary to apply a distributed physically-based runoff model. Many of the classical hydrological models are based on transfer functions or lumping techniques due to limitations of the available ground data. Remote sensing techniques have rapidly developed these last years and offer a practicable way of limiting ground work to previously identified classes with the same physical properties. The established classification does not correspond to a usual pedological soil map but reflects the physical soil properties instead. These soil properties are first of all the infiltration patterns (or runoff generation) and the vulnerability towards erosion which can be determined by rainfall simulation. It is possible, furthermore, to obtain information about the antecedent soil moisture distribution by applying a principle component analysis to the satellite images, These detailed information is coherent for the whole coverage of the satellite image which allows the modelling of even large basins in a detailed way. No traditional field method is capable of furnishing a similar amount of detailed simultaneous data. This modelling technique was used in a project of the Algerian-German technical cooperation dealing with the introduction of new technologies to the Algerian Hydraulic Resources Administration. On an example in northern Algeria the application of this combined modelling technique is shown.