Land surface heterogeneity on surface energy and water fluxes

We studied the importance of land surface heterogeneity on climate models using the MOSIAC Land-Surface Model (LSM). Preliminary analysis of results indicated there were errors in surface heat fluxes for certain geographical regions with contrasting cover such as forests, grasses, and crops when using only one cover class per grid. For spatially varying areas, two to four classes per grid typically captured most of the variation in surface energy and water fluxes. A Minimum Percent Cutoff approach to select the number of classes per grid (or tiles) was found the most efficient in terms of computer time and accuracy. In a comparison between 1/8 degree versus 1-degree grid resolutions, the finer resolution land cover data were more important than finer resolution atmospheric forcing data (e.g. precipitation and radiation) on latent heat flux estimation.