Sensitivity of the Biosphere–Atmosphere Transfer Scheme (BATS) to the Inclusion of Variable Soil Characteristics

Abstract

The soils data of Wilson and Henderson-Sellers have been incorporated into the land-surface parameterization scheme of the NCAR Community Climate Model after Dickinson. A stand-alone version of this land-surface scheme, termed the Biosphere-Atmosphere Transfer Scheme (BATS), has been tested in a series of sensitivity experiments designed to assess the sensitivity of the scheme to the inclusion of variable soil characteristics. The cases investigated were for conditions designed to represent a low-latitude, evergreen forest; a low-latitude sand desert; a high-latitude coniferous forest; high-latitude tundra; and prairie grasslands, each for a specified time of year. The tundra included spring snowmelt and the grassland incorporated snow accumulation. The sensitivity experiments included varying the soil texture from a coarse texture typical of sand through a medium texture typical of loam to a fine texture typical of clay. The sensitivity of the formulation to the specified total and upper soil column depth and the response to altering the parameterization of the soil albedo dependence upon soil wetness and snow-cover were also examined. The biosphere-atmosphere transfer scheme showed the greatest sensitivity to the soil texture variation, particularly to the associated variation in the hydraulic conductivity and diffusivity parameters. There was only a very small response to the change in the soil albedo dependence on wetness and, although the sensitivity to the snow-covered soil albedo via the response to roughness length/snow-masking depth was significant, the results were predictable. Changing the total depth of the active soil column produced a much smaller response than altering the depth of the upper soil layer, primarily because the degree of saturation of the upper layer plays an important role in the parameterized hydrology. Soil moisture responses can also be initiated by changes in vegetation characteristics such as the stomatal resistance through changed canopy interaction which modify the radiation and water budgets of the soil surface. Overall, this land-surface parameterization scheme shows considerable sensitivity to the choice of soil texture. This sensitivity seems to be at least comparable to that involving changes in vegetation characteristics and it may be more important because soil characteristics are very poorly known at a resolution appropriate for global climate models.