A SENSITIVITY STUDY OF A COUPLED SOIL-VEGETATION BOUNDARY-LAYER SCHEME FOR USE IN ATMOSPHERIC MODELING

Abstract

During the last two decades, several land surface schemes foruse in climate, regional and/or mesoscale, hydrological and ecologicalmodels have been designed. For all of them, of great importance isa correct parameterization of the energy and momentum transportfrom the surface. Mainly, the land-surface schemes have beendeveloped in order to parameterize land-surface variations both ondaily and annual time scales. In this paper we describe the Land AirParameterization Surface Scheme (LAPS) designed for use inatmospheric models and mainly for use in simulations of thermalcirculations. It has two modules for parameterizing the processes atthe boundary-layer interface: one for a bare surface and one for avegetated one. Incorrect parameterization of land-surface processes andprescription of the surface parameters in atmospheric modeling, canresult in artificial changes of horizontal gradient of the sensible heatflux. Thus, an error in horizontal temperature gradient within thelower atmosphere may be introduced. Of course, the reliability of themodel depends on the quality of boundary-layer scheme implementedand its sensitivity to the bare soil and vegetation parameters. In this study, the sensitivity of the LAPS scheme toparameterization of bare soil processes, ground roughness length,vegetation parameters (stomatal resistance, leaf area index and canopyheight) and albedo, has been examined. Tests based on timeintegrations using real data were performed. For sensitivity tests withthe non-plant module the data sets for June 3, 1981; June 4, 1982;June 11, 1982 and June 24, 1982 from the experimental site atRimski Šančevi (Yugoslavia) were used. The performance of thevegetation module was tested by employing the data set forSeptember 8, 1988 measured at the experimental site of DeSinderhoeve (The Netherlands). The computed latent heat flux andthe ground temperature outputs were compared with the observations.Finally, the sensitivity of the LAPS scheme to the prescribedparameters was examined by using a simple statistical analysis.