A comparison of two strategies on land surface heterogeneity used in a mesoscale β meteorological model

Results of case studies with a mesoscale β meteorological model applying two different strategies to treat subgrid-scale surface heterogeneity are compared with each other to evaluate the effects of these strategies on the predicted hydrologically relevant quantities. In the first strategy, the mosaic approach, different land-use types are considered as separate patches within a grid cell independently interacting with mean atmospheric field quantities of that grid cell. Feedback to the grid scale is accomplished by forming area-weighted quantities from the fluxes provided by the individual patches for the soil-biosphere-atmosphere interaction. In the second strategy, a higher resolution subgrid is established within each model grid cell and the soil-biosphere-atmosphere interaction is determined for each subgrid cell with its individual soil and biosphere conditions and near-surface meteorological forcing. Probability density functions are used to evaluate the statistical behavior of both the strategies. It is substantiated that the partitioning of the atmospheric radiative and moisture forcing at the surface as well as cloud and precipitation formation can significantly be affected by the type of strategy. Using the explicit subgrid strategy results in a shift in the partitioning of energy towards decreasing Bowen ratios as compared to the mosaic approach. For very heterogeneous surfaces with strongly varying soil types and plant species an area-weighted meteorological near-surface forcing as used in the mosaic approach may artificially reduce evapotranspiration. An explicit subgrid strategy or individual near-surface meteorological forcing within the mosaic approach seems to be more adequate under such surface conditions.