The impact of vegetation and soil parameters in simulations of surface energy and water balance in the semi-arid sahel: A case study using SEBEX and HAPEX-Sahel data

A series of numerical experiments has been designed to understand the physics at the land surface–atmosphere interface in the Sahel and to find the major parameters and parameterizations that are crucial to simulate arid climate processes. Observational data sets from the Sahelian Energy Balance Experiment (SEBEX) and the Hydrological Atmospheric Pilot Experiment (HAPEX-Sahel) were used to help interpret the results. The observational errors in these two data sets are examined first. In the control simulations, the standard type-8 (shrubs with ground-cover) data typically used by general circulation models for vegetation in the Sahel produce more than 50% underestimation of average latent heat flux and 100% overestimation of average sensible heat flux. The diurnal cycle of surface temperature may be overestimated during the day and underestimated at night. Sensitivity experiments were conducted to identify the most important parameters that affect Sahel simulation and their role at the surface/atmosphere interface. It was found that proper parameter estimation of leaf area index, stomatal resistance, and hydraulic conductivity allow enough evaporation to occur in land surface simulations. Sensible heat flux is thus reduced by greater partitioning of energy to latent heat flux. Proper parameter estimation of thermal diffusivity also impacts sensible heat flux and diurnal variation of surface temperature. In addition, the water balance may be misrepresented with standard parameter values by maintaining an unnecessarily high amount of soil moisture due to less evaporation. In general, the error existing between observations and simulation is reduced substantially by proper calibration of the vegetation/soil parameters. q 2005 Elsevier Ltd. All rights reserved.