Estimating volume of water harvested by farm dams in Murray-darling basin

Abstract

With limited water resources available, farmers in many parts of Australia have resorted to building farm dams to supplement water for irrigation and stock. While the impact of an individual farm dam on a catchment is relatively small, the cumulative impact on runoff from a large number of farm dams on may be significant. To give an example of scale, the Murray-Darling Basin (MDB) has over 650,000 farm dams at present. This paper aims to quantify the effect of the farm dams on the water resources in the Murray-Darling Basin. Through the Water Act 2007, the Bureau of Meteorology (the BoM) has the responsibility to produce an annual National Water Account (NWA) which provides information on water stores and fluxes, water rights and water use across Australia. The Australian Water Resource Assessment Modelling System (AWRAMS) is being developed by CSIRO and the BoM through the Water Information Research and Development Alliance (WIRADA) initiative. AWRAMS is a new integrated continental hydrological simulation system that has two modeling components to represent processes between the atmosphere and the landscape (AWRA-L) and in gauged rivers (AWRA-R), including all major water storages and fluxes in and between these components (surface, subsurface and groundwater). One of the fluxes described in the NWA is the runoff to rivers/reservoirs accounting for interception and storage by farm dams. The Spatial Tool for Estimating Farm Dam Impacts (STEDI), developed by SKM, was used in the production of the NWAs for 2010 and 2011. STEDI carries out a water balance for each farm dam at each monthly time step using contributing catchment inflows, rainfall, evaporation and on-farm demand. The contributing catchment inflows were obtained by lumping the contributing catchment's runoff generated by the grid-based AWRA-L model. Running the STEDI model is time consuming in terms of preparing and inputting the data into the model and applying it over large areas as the number of farm dams in a model run must be less than 10,000 (requiring 105 regions in the case of the MDB). In order to streamline and speed up the production of the NWA, a farm dam model (FDM) similar to STEDI was coded in FORTRAN, run over the NWA Canberra region and then evaluated by comparing the results to those from the original STEDI model. Following successful evaluation, the developed FDM model was applied to the recent NWAs (NWA 2011, NWA 2012, NWA 2013 and NWA 2014) over the MDB. The current FDM approach used in the NWA has two limitations: all farm dams are assumed to be directly connected to the catchment outlet and the spatial variation in the input is ignored by lumping them. Also, currently the farm dam modelling is undertaken separate from the AWRAMS although there are plans to integrate the two in the future. This paper will present the development of a farm dam model and its application to the MDB for the production of the NWA reports. Future direction of farm dam modeling through model integration with the BoM's operational AWRAMS is also presented at the end of the paper.