The simulation of complex land cover in regional climate studies

The land surface exhibits small-scale heterogeneities at scales far smaller than even regional climate models are generally run. The degree of scale mismatch can range from relatively slight, as in the case of the HAPEX-Mobilhy study area, to extreme, as in the case of the BOREAS region. In the HAPEX-Mobilhy study area, the two main land cover types were coniferous forest and agricultural areas, with relatively minor sub-grid scale heterogeneity. In the case of the BOREAS area, the land cover consisted of a highly complex mixture of coniferous forest, deciduous forest, wetlands, lakes and burned areas, at scales often considerably below that of regional modelling grids. Land surface models attempting to address such a modelling problem must adopt some form of a mosaic approach to characterize the different land cover types present; however, the model physics must also be capable of handling the highly distinctive energy and moisture transfer processes within such sub-grid elements. Satellite data must be relied upon to provide key input variables, and the limitations of such data must be recognized. This presentation will describe improvements that have been made to CLASS, the Canadian Land Surface Scheme, to handle such modelling challenges, and work that is underway to begin modelling studies over the BOREAS area using CLASS coupled with the Canadian Regional Climate Model.