| A high priority goal of the AMMA project is to better understand and model the influence of the spatio-temporal variability of surface processes on the atmospheric circulation patterns and the regional water cycle related to the West African Monsoon. This issue is being addressed under the auspices of the AMMA Land surface Model Intercomparison Project (ALMIP). The idea is to force state-of-the-art land surface models with the best quality and highest (space and time) resolution data available in order to better understand the key processes and their corresponding scales. In this paper, an overview of the ALMIP project objectives is presented, along with a description of input forcing data, some preliminary results, and a discussion of ongoing evaluation efforts. In terms of forcing, satellite-based data is used to reduce errors in data derived from atmospheric models which suffer from systematic biases: errors in the placement of the active monsoon precipitation zone are shown to be significantly reduced using such data. This is shown to have a significant impact on the LSM (land surface model) evapotranspiration, especially in the Sahel. The inter-model scatter in evapotranspiration and soil moisture storage change over the Sahel was found to be on the order of 20% of the multi-model mean values during the monsoon season, thus the models were in fairly good agreement and have far better agreement than such estimates from coupled land-atmosphere models. In terms of evaluation of simulations, indirect methods are required on large scales: the LSMs were able to produce spatial and temporal soil moisture patterns consistent with remotely sensed brightness temperature data over this region. At the grid scale, ALMIP surface sensible heat flux estimates had the same basic response (amplitude and phase) to the wet season as seen in aggregated fluxes from local scale observational sites. Finally, ALMIP is an ongoing project and perspectives for the next phase (at finer spatial scales) are presented. |
Export this paper
