Spatial aggregation of vegetation parameters in a coupled land surface-atmosphere model

The aggregation of heterogeneous land surface cover parameters to calculate effective area-average energy fluxes at microscale using a realistic coupled land surfaceatmosphere model are been investigated in this study. The Biosphere-Atmosphere Transfer Scheme (BATS) was merged with a two dimensional atmospheric boundary layer simulation model (ABLE) to develop a coupled model, BAT-ABLE. Observed land surface and atmospheric forcing data from the FIFE site in Kansas were used to initialize and run the time series of this model. The initial model states were obtained from seasonal runs of a stand-alone version of BATS using observed data. The horizontal model domain was set to 2 km, while the size of each patch was 1 km. The height of the first model grid level above the ground was set to 2 m to match the screen height at the FIFE site. Simple aggregation rules were used to aggregate BATS parameters to obtain area-average energy fluxes. Four cover types (short grass, long grass, mixed crop and irrigated crop) with five different patch combinations were tested using July 7, August 15, and October 11, 1987 observed data. The aggregation scheme worked well in almost all cases for these different days and times (am, noon, pm). However the aggregation scheme failed in the particular cases of artificially wet soil patches set in a landscape of dry soil. However, analysis of fluxes in this situation showed that this failure is not due to a difference in efficiency of atmospheric advection in this case, rather it is due to a real net change in the area-average turbulent fluxes returned to the atmosphere.