Impact of land use on wind erosion and dust emission in the Sahel: a regional modelling approach
Résumé
The semi-arid Sahelian region is particularly prone to wind erosion due to low and variable precipitation rates producing low vegetation cover with a high interannual variability. Surfaces traditionally devoted to pasture are progressively turned to cultivated fields increasing the proportion of bare surface unprotected from wind erosion. Wind erosion could significantly evolve in the future due to climate change and to the increase in crop areas due to the growing population. To estimate the wind erosion and dust emission in the Sahel, we have developed a regional modelling approach in the frame of the CAVIARS project (Climate, Agriculture and Vegetation: Impacts on Aeolian eRosion in the Sahel). The approach is based on existing models that were adapted, improved and coupled to represent the main processes involved in the wind erosion and their dependence to climate parameters and agricultural practices. Natural vegetation (herbaceous) and cultivated vegetation (millet) are respectively modelled with the STEP vegetation model (Mougin et al., 1995) and the SARRA-H agronomic model (Baron et al., 2005). Both models have been improved to reproduce the dynamics of dry vegetation and to produce multi-annual simulations (e.g. Pierre et al., 2016). The wind erosion model (Marticorena and Bergametti, 1995) has been tested and adapted to represent the erosion fluxes over typical Sahelian vegetated surfaces (Pierre et al., 2014a, b). Surface crusting and its dynamics is represented as a function of soil texture, precipitation and vegetation cover. A main limitation was to represent the very high surface winds associated to the convective activity that are not reproduced by the large scale meteorological wind fields (Largeron et al., 2015). A specific parameterization of the distribution of the surface wind speeds have been developed based on high resolution simulations (CASCADE exercise). The application of the models at the regional scale required dedicated data sets, established for this purpose. Daily precipitation fields have been produced based on local measurements and grazing pressure maps have been established based on national data and/or agro-poastoral census. The proportion of cultivated fields and pastured lands are derived from the Sahelian land use map of the AGRHYMET (Tappan et al., 2004). The spatial and temporal distribution of the simulated dust emissions computed from the period 2000-2014 will be presented. The parameterization of the wind speed distribution associated with the convective activity allows to better capture the seasonal pattern of local wind erosion (Bergametti et al., 2017). The relative weight of climatic parameters, vegetation and cultivation will be estimated by comparing regional simulations with and without vegetation and cultivation. References