THE COMPONENTS OF A SVAT SCHEME AND THEIR EFFECTS ON A GCMS HYDROLOGICAL CYCLE

'Bucket'-type land surface models are being replaced in some general circulation model (GCM) climate studies by 'SVAT' (Surface Vegetation-Atmosphere Transfer) models, which feature an explicit treatment of vegetation control over the surface energy balance. The evaporation calculations of a typical SVAT model differ from those of a bucket model in at least four fundamental ways: (a) the SVAT model typically allows a greater variety of environmental stresses to limit evapotranspiration; (b) it generally includes a canopy interception reservoir; (c) the control of the land surface over evaporation in a SVAT model is influenced by the atmosphere; and (d) the land surface control in a SVAT model varies on short time scales. Global fields and grid square diurnal cycles illustrate the hydrological cycle produced in a 20-year simulation with an atmospheric GCM coupled to a SVAT model. A sensitivity analysis then examines the relative importance of the SVAT/bucket differences in terms of their effects on the simulated hydrological cycle. The interception reservoir exerts more control over global evaporation than does either the vapor pressure deficit stress or the temperature stress. The effect of the temperature stress is, in fact, insignificant. The time variability of land surface control over the surface energy balance in a SVAT model significantly increases moisture convergence over land.