Land-surface and boundary layer processes in a semi-arid heterogeneous landscape

The European Field Experiment in a Desertification-threatened Area (EFEDA) provides a comprehensive land-surface dataset for a semiarid Mediterranean environment. It is used here to study heat and moisture transport processes in the atmospheric boundary layer (ABL), to derive grid-scale surface fluxes for numerical weather prediction models, and to evaluate the performance of the High-Resolution Limited-Area Model (HIRLAM).Boundary layer budgets were derived for two sub-regions of the EFEDA area with different moisture characteristics. The budget analysis is based on the synergistic combination of the observational dataset and a simple coupled canopy-mixed layer model. The mixed layer (ML) warming is balanced by a combination of the heat flux divergence and the radiative divergence (directly measured by aircraft). The large radiative warming is consistent with the high aerosol load and low visibility observed in the area. The moisture budgets reflect the influence on non-ABL scales on the entrainment moisture flux, which changes sign over time as the ABL grows into the observed heterogeneous moisture structure of the residual layer (RL). Accurate high-resolution vertical humidity profiles are needed to properly estimate the moisture flux divergence, which varies in space and time. The coupled canopy-ML slab model proves to be a valuable tool in this complex environment, if it is regularly provided with updated RL gradients. The potential for moisture flux divergence and associated ML moistening is higher at Barrax, where irrigation enhances the surface evapotranspiration.The area-aggregated fluxes (in particular of moisture) depend strongly on the location of the area boundaries, whenever a significant fraction of irrigated land is present. This confirms the importance to adequately account for tiles of irrigated land in surface schemes and the corresponding physiographic databases of large scale models. A simple way to accommodate a minimum information on canopy water status is proposed in terms of the distinction of at least two seasonal classes of irrigated crops, one of spring and one of summer growing cycles.The HIRLAM performance evaluation reveals model shortcomings essentially in four areas. The moisture assimilation makes the model surface and ABL too moist. The ABL entrainment description cannot resolve the observed temporal and spatial variations of entrainment moisture fluxes. The landuse and soil classification with its associated physiographic database attribute too much green vegetation to the EFEDA grid cells, thus causing a wet bias in the surface energy balance. The aerosol parameterization in the radiation code does not account for the typically higher aerosol load of semiarid environments, which introduces a high bias in solar and net radiation. Practical steps for model improvement are proposed. They focus on the landuse classification and the aerosol parameters, both adapted to dry Mediterranean landscapes