Integrated modelling of the water cycle in semi arid watersheds based on ground and satellite data: the SudMed project

Abstract

The SudMed project aims since 2002 at modelling the hydrological cycle in the Tensift semi arid watershed located in central Morocco. To reach these modelling objectives, emphasis is put on the use of high and low resolution remote sensing data, in the visible, near infrared, thermal, and microwave domains, to initialize, to force or to control the implementation of the process models. Fundamental studies have been conducted on Soil-Vegetation-Atmosphere Transfer modelling (SVAT), especially related to the various means of incorporating both ground and remote sensing observation into them. Satellite data have been used for monitoring the snow dynamic which is a major contribution to runoff issued from the mountains. Remote sensing image time series have also been used to map the land cover, based on NDVI time profiles analysis or temporal unmixing of low resolution pixels. Subsequently, remote sensing time series proved to be valuable for monitoring the development of vegetation and the crop water status, in order to estimate evapotranspiration, key information for irrigation management.

Evapotranspiration estimates in a traditional irrigated area in semi-arid Mediterranean. Comparison of four remote sensing-based models

“…The main parameters of SAMIR are basically related to crop types and to the soil. Thus, a land cover map of the study area was generated from the NDVI time series using the algorithm developed by Simonneaux et al, (2007). The main classes were trees, annual crops and bare soil.…”

“…This means that the error on LST will not have the same effect on METRIC-GEE and SW as compared to SPARSE. One advantage of SAMIR is to produce continuous daily ETa estimates which is valuable in assessing the basin water budget (Simonneaux et al, 2007;Le Page et al, 2012;Diarra et al, 2017). SPARSE is a model that can be used accurately to detect stress conditions if properly calibrated (Boulet et al, 2015).…”