10.3 Recharge from landscape
The volume presented in the water accounting statements (382,674 ML) represents the volume of groundwater recharge from the landscape in the Melbourne region during the 2012–13 year.
Recharge to groundwater is strongly linked to with rainfall in the region. Groundwater recharge to the fractured rock is not included in this assessment; recharge is calculated for sedimentary and basalt areas where lateral groundwater flow can be significant. Most of the recharge occurring in the fractured rocks is either discharged as baseflow to rivers or evapotranspired (ET) within a short distance of the site of recharge. In general, over a year fractured rock areas experience a small and insignificant change in groundwater storage. In addition, groundwater discharge to river as base flow and ET from groundwater are difficult to evaluate in this region.
The Bureau of Meteorology National Climate Centre (NCC) Version 3 daily rainfall grids, Version 3 annual rainfall grids, daily maximum temperature grids, daily minimum temperature grids, daily satellite observed solar radiation grids and daily vapour pressure deficit grids; CSIRO: Australian Soil Resources Information System (ASRIS) soil information; Bureau of Rural Sciences: Water 2010 land use mapping; Victorian Department of Environment and Primary Industries: bore locations, groundwater level data, and aquifer attribution.
Groundwater recharge was estimated using the WAVES model (Zhang and Dawes 1998; Dawes et al. 1998). WAVES is a one-dimensional soil–vegetation–atmosphere–transfer model that integrates water, carbon, and energy balances with a consistent level of process detail. The input datasets required for WAVES include climate, depth to water table, soil and vegetation data. The clipped sedimentary area was selected to estimate net recharge. The climate data used at selected points include rainfall, rainfall duration, maximum and minimum temperatures, vapour pressure deficit, and solar radiation. The relevant vegetation parameters required for modelling were selected from the WAVES user manual (Dawes et al. 1998). WAVES uses the soil hydraulic model of Broadbridge and White (1998) with saturated hydraulic conductivity, saturated moisture content, residual moisture content, inverse capillary length scale, and an empirical constant as input parameters calculated from soil properties accessed in the ASRIS database (Johnston et al. 2003).
The WAVES model has been used by CSIRO in its sustainable yields projects (Johnston et al. 2003) and the Bureau has built on this methodology. WAVES was run for all combinations of soil, vegetation and depth to water table at each climate point. A groundwater recharge value was estimated for each 1 km × 1 km pixel across the region using annual rainfall, dominant soil and vegetation, and depth to water table. This recharge value can be positive or negative, due to evapotranspiration. Recharge was determined by summing the pixels with a negative estimate (grey areas).
WAVES model recharge areas
Assumptions, limitations, caveats and approximations
- Assumptions made when developing the WAVES model (Dawes et al. 1998) are all applicable to the recharge estimations carried out for the Melbourne region.
- The Bureau of Rural Sciences's land use map of the Melbourne region was reclassified to three vegetation classes: annuals, perennials, and trees. The major vegetation classes modelled are C3 annual pasture, C3 perennial pasture, and eucalypt trees with a grassy understorey.
- Recharge was estimated to be within the clipped sedimentary area, considering the effects of shallow water table interpolated using kriging with external drift and the 9" Digital Elevation Model as an external driver following the methodology presented in Peterson et al. (2011).
- Diffuse recharge to groundwater from irrigation applied to the landscape is not included in the estimate.
- The uncertainty in the input parameters and the corresponding impacts on the modelled recharge values have not been studied.
- The uncertainty of the estimated recharge resulting from different recharge interpolation methods is not estimated.