A new parameterization for surface and groundwater interactions and its impact on water budgets with the variable infiltration capacity (VIC) land surface model

Abstract

This paper describes a new parameterization to represent surface and groundwater interaction dynamics for land surface models. With the new parameterization, effects of surface and groundwater interactions on soil moisture, evapotranspiration, runoff, and recharge can be dynamically taken into account. The new parameterization is implemented into the three‐layer variable infiltration capacity (VIC‐3L) model, which is a hydrologically based land surface scheme. The new version of VIC (called VIC‐ground) is applied to two watersheds in Pennsylvania over multiple years. Results show that VIC‐ground can properly simulate the movement of the daily groundwater table over multiple years at the study sites. Preliminary comparisons of VIC simulations with and without consideration of the dynamics of surface and groundwater interactions show an important impact of such interactions on the partitioning of water budget components. In particular, soil moisture of the lower layer from the VIC‐ground simulations is generally wetter than that from VIC‐3L. For the top thin soil layer and the upper layer of VIC‐3L, soil moisture is generally drier in VIC‐ground than that in VIC‐3L. Such characteristics of VIC‐ground result in lower surface runoff peaks and higher base flow, as well as generally less evapotranspiration compared to VIC‐3L at the two study sites. Results at both sites show that it takes 3–4 years to have the effects of the initializations of groundwater tables disappear when the groundwater table is initialized to be deeper than the observed level, while it takes much less time (e.g., about 1.5 years) if the groundwater table is initialized to be shallower than the observed level. In addition, the preliminary sensitivity studies at both sites show that there is a more significant persistent signature of the impact of the precipitation when its amount is halved (i.e., 0.5 ppt) than that when its amount is doubled (i.e., 2 ppt).