Effects of subgrid-scale heterogeneity of soil wetness and temperature on grid-scale evaporation and its parameterization

An analytical approach is presented to study the effects of spatial heterogeneity of land surface parameters in climate models. This approach does not use any parametric probability density functions for land surface parameters and attempts to account for the effects of subgrid heterogeneity by using grid-level parameters and correlation between parameters. The adequacy of the approach is illustrated by using the influence of spatial variability in soil wetness and temperature on evaporation. It is shown that both the stability of the lower atmosphere and the surface soil wetness conditions could introduce non-linearity in the estimation of grid-scale evaporation. For stable and near neutral conditions, the heterogeneity in surface temperature plays an important role in the estimation of grid-level latent heat flux. The heterogeneous effect of soil wetness is found to be important when the soil is under moisture constraint. With increasing soil wetness, heterogeneous effects of soil wetness diminish and become negligible as we approach saturation. It appears that effects of temperature and soil moisture heterogeneity effects are an order of magnitude larger compared with correlative effects of temperature and soil moisture for a range of soil wetness and temperature conditions. The subtle difficulties associated with non-linear and coupled influence of soil wetness and atmospheric stability on evaporation preclude the use of any simple aggregation procedure to represent subgrid-scale effects in climate models.