Mapping ecosystem and biophysical variable satellite for the study of hydric flux over the African mainland

Abstract

In the context of climate change, the aim of this study is to characterize the heterogeneity of the African continent in order to provide some elements to better understand and quantify surfaces process acting on hydric fluxes. This work is intented to update the double ECOCLIMAP-I database which is constituted by a land cover map and a dassets of land biophysical parameters. To this end, we use remotely sensed data acquired by the latest generation sensors MODIS and SPOT/VEGETATION between 2000 and 2007. During the first step, two methods of classifications has been developed for the mapping of different ecosystems. The first method, which is supervised, is obtained by combining information provided by the both global land cover map GLC2000 and ECOCLIMAP-I using an interactive analysis of MODIS leaf area index (LAI). It has been performed in the framework of the AMMA project to discriminate ecosystems over the western African Region. The second method is hybrid in that it combines k-NN clustering, hierarchical principles and the Fast Fourier Transform (FFT) on the basis of multi-annual NDVI data from SPOT/VEGETATION to identify ecosystems at the whole African continent. Then, methods for the estimation of land surface biophysical variables such as albedo, fractional vegetation cover and leaf area index has been developed and/or applied over the mainland. A statistical approach allows us to determine the contribution of bare soil albedo and vegetation albedo to the constitution of albedo as required in land surface models. After the application of the latter approach over the western african region, we demonstrate the robustness of the method by applying it over the entire mainland. The sensitivity of two land surface scenarios was studied by analysing two simulations with the same atmospheric forcing over the western African Region:one using the ECOCLIMAP-I classification and another using the new physiographic forcing specifically developed over the