With minor modification an existing analytic model employing a simple statistical description of soil moisture dynamics is found to give accurate estimates of the mean water balance partioning; comparisons are made to a Monte Carlo numerical simulation of unsaturated moisture flow in a soil column. In the statistical-dynamicl model the complex temporal and spatial characteristics of the soil moisture profile are approximated by a single effective soil moisture: the equilibrium soil moisture (s0) for which the modeled long-term mean fluxes are in balance. The relation between this equilibrium soil moisture and the equivalent steady state moisture profile (the steady profile which transmits the long-term mean flow) is discussed. The test numerical integrations of the flow and conservation equations are designed to represent a deep one-dimensional soil column. Soil surface boundary conditions are defined by the interaction of the surface soil moisture with a stochastic event-based atmospheric forcing model, allowing for state dependent switching from flux boundary conditions (rainfall intensity and potential evaporation) to concentration conditions (ponded and dry surface), thus simulating infiltration-excess runoff production and soil-limited evaporation. The ability of the equilibrium model to estimate the mean water balance under a large range of soil textures (from clay to sand loam) and climate forcings (humid to arid) implies that the soil moisture profile may, for purposes of climate water balance, be represented in a simple fashion. ¿ American Geophysical Union 1994