Landscape water balance components were calculated for Melbourne region over 2009–10. The table below provides contextual information about the landscape water balance closure and the relative magnitude of various components.
This landscape water balance does not consider the effects of irrigation.
Fluxes into the soil column (‘the landscape’) for the 2009–10 year are entered in the ‘flux in’ column and all the fluxes leaving the landscape are entered under the ‘flux out’ column as negative numbers. The change in soil moisture is calculated by subtracting the soil moisture on 1 July 2009 from the soil moisture on 30 June 2010. The water balance error is calculated as a volume (ML) as well as a percentage of the total flux in.
Water balance component
|
Flux in
|
Flux out
|
Precipitation on landscape
|
10,335,901
|
|
Evapotranspiration from landscape
|
|
-7,566,080
|
Run-off to connected surface water
|
|
-1,005,902
|
Run-off harvesting into off-channel private water store
|
|
NA
|
Recharge/discharge
|
6,343
|
|
Total
|
10,346,386
|
-9,240,407
|
Net flux
|
1,105,979
|
Change in soil moisture
|
1,744,463
|
Water balance error (ML)
|
-113,774
|
Water balance error (% of total inflow)
|
-1.1
|
NA = not applicable
The water balance error for Melbourne is quite small and the negative sign indicates that the change in soil moisture is greater than the net flux. Run-off harvesting into off-channel private water store has not been included in the balance and the inclusion of this term would increase the water balance error.
Explanatory notes
Precipitation on landscape
Data source
Bureau of Meteorology, National Climate Centre (NCC): raster spatial data; version 3 daily rainfall grids; geographical information system (GIS) layers; Australian Hydrological Geospatial Fabric (AHGF).
Data provider
Buerau of Meteorology.
Method
Monthly precipitation data was produced by the Bureau of Meteorology. It was based on daily data from approximately 6,500 rain gauge stations and interpolated to a 0.05 degrees (5 km) national grid (Jones et al. 2007). The precipitation data was consistent with that input into WaterDyn to generate the Line item 3.1 Soil moisture estimates.
The volume of precipitation falling on the landscape was calculated using precipitation grids that fell within the reporting region boundary. Storages were removed from the landscape area within the reporting region using the Australian Hydrological Geospatial Fabric.
Uncertainty
Uncertainty is ungraded.
Approximations, assumptions, caveats/limitations
- The precipitation estimates were subject to approximations associated with interpolating observation point data to a national grid detailed in Jones et al. (2007).
- Grid cells that intersected the reporting region boundary (i.e. had some part of the cell outside the region boundary) were included in the calculation of total precipitation. This had a limited influence on the calculation of total precipitation because the area by which total precipitation was multiplied was determined from the catchment region boundary, not from the grid cells themselves.
- Precipitation on the surface water features that are not considered explicitly in the account (e.g. some wetlands and lakes) is captured in the landscape precipitation.
Evapotranspiration from landscape
Data source
Bureau of Meteorology, National Climate Centre (NCC): raster spatial data; version 3 daily rainfall grids; geographical information system (GIS) layers: Australian Hydrological Geospatial Fabric (AHGF).
Data provider
Bureau of Meteorology.
Method
Total evapotranspiration was derived using the Commonwealth Scientific and Industrial Research Organisation (CSIRO) WaterDyn model. The total evapotranspiration estimate is the sum of transpiration and soil evaporation outputs from WaterDyn. Monthly vegetation cover is determined from remotely-sensed vegetation greenness from the SeaWiFS satellite.
The WaterDyn model calculates evapotranspiration on a national 0.05 degree (~5 km x 5 km) grid. The total volume of evapotranspiration from the landscape in 2009–10 was calculated from grid cells that fell within the reporting region boundary. Storages were removed from the landscape area within the reporting region using the Australian Hydrological Geospatial Fabric (AHGF).
Uncertainty
Estimated uncertainty is ungraded.
Approximations, assumptions, caveats/limitations
- The landscape evaporation estimates are subject to the assumptions of the WaterDyn model detailed in Raupach et al. (2008).
- WaterDyn is balanced according to the supply of water to the landscape via precipitation and therefore does not account for application of water via irrigation and lateral flow. WaterDyn also assumes no interaction between the landscape and groundwater storages.
- The vegetation fraction cover used in WaterDyn is a monthly climatological average (i.e. a single grid to represent all Januaries) and does not capture land use change from year-to-year.
- Grid cells that intersected the reporting region boundary, i.e., had some part of the cell outside the region boundary, were included in the calculation of total evapotranspiration. This had a limited influence on the calculation of total evapotranspiration because the area by which total evapotranspiration was multiplied was determined from the catchment region boundary, not from the grid cells themselves.
- Evaporation from the surface water features that are not considered explicitly in the account (e.g. some wetlands and lakes) is captured in the landscape evaporation.
Run-off to connected surface water
Data source
Bureau of Meteorology, National Climate Centre (NCC): raster spatial data; National Climate Centre (NCC) version 3 daily rainfall grids; geographical information system (GIS) layers; Australian hydrological geospatial fabric (AHGF)
Data provider
Bureau of Meteorology.
Method
Run-off was calculated as the average of ‘discharge’ from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) WaterDyn water balance model and ‘streamflow’ from the CSIRO AWRA-L water balance model. These two estimates were averaged because previous research suggested an ensemble of these models provided a better estimate of run-off than that given by either model individually (Bacon et al. 2010; Viney 2010).
These CSIRO water balance models require gridded daily meteorological inputs. These inputs include precipitation, maximum temperature, minimum temperature and incoming solar radiation for days and locations. Where solar radiation was unavailable for a particular day or location, an infilling procedure provided a value from the climatological average.
The models were run across the spatial extent of the Australian continent using meteorological inputs from the Bureau and parameter estimates from CSIRO.
Volumetric run-off was determined by multiplying the depth of the averaged run-off by the area of the reporting region. Average run-off was determined as the unweighted arithmetic mean of run-off from all grid points that fell within the region.
The run-off determined by this routine corresponds to the run-off expected from an unimpaired catchment (Bacon et al. 2010; Viney 2010). The impairment on run-off created by interception by small dams has been accounted for by running the averaged run-off through a local catchment reservoir water balance model in areas where there is adequate mapping of local catchment reservoirs.
Uncertainty
Estimated uncertainty is ungraded.
Approximations, assumptions, caveats/limitations
- The run-off estimates were modelled only and were not verified by real-time analysis of streamflow records. The reported value was an estimate of water that was likely to have entered the connected water store from the landscape.
- The run-off estimates were subject to the assumptions of the WaterDyn model detailed in Raupach et al. (2008) and the AWRA-L model detailed in van Dijk (2010).
- Where the volume of water intercepted by the off-channel private water store has been calculated, the run-off estimates inherit the approximations, assumptions and caveats of the local catchment reservoir model (STEDI) and parameters used.
- Grid cells that intersected the reporting region boundary (i.e. had some part of the cell outside the region boundary) were included in the calculation of total run-off. This had limited influence on the calculation of total run-off because the area by which average run-off was multiplied was determined from the catchment region boundary, not from the grid cells themselves.
