Incorporating topography-dependent groundwater storage in AWRA-L improves groundwater flux estimation

Abstract

The landscape component of the Australian Water Resources Assessment system, AWRA-L, is a grid-distributed biophysical model designed to simulate water storage in and flows between vegetation, soil, surface water and groundwater for the Australian continent [van Dijk, 2010a]. This study addresses three known issues with the representation of groundwater dynamics in version 0.5 of AWRA-L: 1. The saturated area fraction per grid cell, which controls surface runoff, soil and groundwater evaporation, is computed as the ratio of groundwater storage over a reference groundwater storage, rather than a function of groundwater storage and topography. 2. The entire groundwater store is ultimately accessible for evapotranspiration, while in reality only the fraction of the groundwater store in reach of plant roots is accessible. 3. Groundwater discharge to rivers is a function of groundwater storage without taking into account the connection status of the river. Ephemeral streams are therefore not well represented in AWRAL. These issues are addressed by using topography information in the calculation of saturated area fraction and baseflow. Topographic variability within an AWRA-grid cell is represented by hypsometric curves derived from a 9" Digital Elevation Model. The minimum elevation within the cell is assumed to be equal to the drainage elevation. The DEM data is transformed into potential groundwater storage by multiplying cell ground elevation by effective porosity. The saturated area fraction corresponding to a groundwater storage level is then obtained from hypsometric curves. By subtracting an extinction depth from the hypsometric curve, the fraction of groundwater storage available for evapotranspiration can be computed. This provides a mechanism for groundwater storage to be lower than the storage corresponding to drainage elevation. Under the assumption that the river system is disconnected or losing in this situation, base flow to rivers is only computed if storage is greater than the drainage elevation. Results for a limited number of test locations, representative for a range of Australian conditions, show that the results of the modified model are in better agreement with the conceptual understanding of groundwater dynamics at these locations.