Estimation of surface fluxes in a small agricultural area using the three-dimensional atmospheric model Meso-NH and remote sensing data

Abstract

To provide an accurate water budget over a whole basin, hydrological models need to know the spatial variability of evapotranspiration at the watershed scale. The three-dimensional (3D) atmospheric models can provide such estimations at a regional scale, since they calculate the different energy and water fluxes by accounting for the landscape heterogeneity with a mesh grid varying from a few metres to several kilometres. We have used such a transfer model (Meso-NH) at a high spatial scale (50 m) to simulate the small agricultural region of the Alpilles (4 km × 5 km), where an experiment took place in 1997 and included intense ground measurements on different types of crops and airborne and satellite data collection. It was the first time that this model was used at such a fine resolution. The aim of this paper is to analyze the effects of the various crops on the spatial variability of the main energy fluxes, particularly evapotranspiration. We also wished to validate Meso-NH from this important available dataset. All input parameters were derived from remote sensing or airborne data: leaf area index (LAI) and albedo were computed from polarization and directionality of the earth's reflectances (POLDER) images. Roughness length was estimated combining both a land-use map obtained from Satellite pour l'Observation de la Terre (SPOT) images and the POLDER images. Maps of the main energy fluxes and temperatures were simulated for two periods in April and June and showed large spatial variations because of differences in soil moisture and in roughness of the crop types. Comparisons between the simulations and the measurements gave satisfactory results. Thermal images acquired by the infrared airborne camera were in good agreement with the surface temperatures estimated by the model. Significant differences were observed when we compared, on the same area, the value of averaged fluxes with the value of fluxes calculated with averaged surface parameters. This was due to the nonlinearity processes associated with averaging of environmental variables. The interest in using a mesoscale model applied at microscale is that coherent structures can be observed in the surface boundary layer, particularly on transects of the vertical wind speed. Such structures cannot be simulated at a larger scale or analyzed with simplified models. Remote sensing data acquired at a fine spatial resolution are a useful tool to provide accurate surface parameters to such a model. This allows quantification of the effect of each crop type on the spatial variation of temperature and evapotranspiration and thus improves our knowledge of the water budget of an agricultural landscape and the watershed functioning.