Adelaide
30.3 Runoff harvesting into off-channel water store

Supporting Information

Runoff harvesting into off-channel water stores during the 2010–11 year was 18,643 ML.

Quantification Approach

Data Source

Bureau of Meteorology: National Climate Centre daily climate grids (rainfall, temperature and solar radiation); Australian Hydrological Geospatial Fabric (AHGF) waterbody feature class. Commonwealth Scientific and Industrial Research Organisation (CSIRO): Australian Water Resources Assessment – Landscape (AWRA-L) model parameters; monthly climatological average radiation grid data. South Australian Department for Water: Geographical Information System layers. Geoscience Australia: nine arc-second digital elevation model (DEM).

Provided by

Bureau of Meteorology; South Australian Department for Water.

Method

Rainfall runoff harvesting to off-channel water storages was estimated based on the AWRA-L (Van Dijk 2010) streamflow and Spatial Tool for Estimating Dam Impacts (STEDI) model (Sinclair Knight Merz 2010) farm dam water balance outputs.

Using climate grid data for the Adelaide region (including precipitation, temperature and solar radiation data), AWRA-L was used to estimate the runoff depth at each grid point within the region.

The Adelaide region has more than 10,000 off-channel water storages. This is the maximum number of off-channel water storages that the STEDI model can handle for a single region. Therefore, the Adelaide region was split into two smaller subregions. The region was divided using the boundaries of the AHGF contracted catchments between McLaren Vale and the Onkaparinga Valley. The northern region included the Barossa Valley, the Northern Adelaide Plains, the River Torrens Catchment, the Patawalonga Catchment and the Onkaparinga Catchment. The southern region included catchments throughout McLaren Vale and the Fleurieu Peninsula. 

Only off-channel water storages filled primarily by rainfall-runoff were considered. These were determined from waterbody mapping provided by the South Australian Department for Water, and excluded waterbodies that were within the Virginia Pipeline Scheme service area and waterbodies that were less than 20 metres away from a channel of second order or higher, or an active bore. A nine arc-second DEM was used to determine the catchment area of each off-channel water storage in the Adelaide region.

The average runoff depth across the Adelaide subregions was determined as the weighted mean of runoff occurring from the relevant grid points within the region boundary. Points were weighted based on the area they represented within the Adelaide landscape to remove edge effects (where the area represented is not wholly within the reporting region) and the effect of changing area with changing latitude.

The average runoff depth was converted to a volume by multiplying depth by the total area and was used as an input into the STEDI model. The model determines the available capacity of each off-channel water storage at each monthly time step. Runoff that does not exceed the available capacity is considered harvested, while runoff that spills from the storage is considered to have contributed to the surface water.

Assumptions, Limitations, Caveats and Approximations

  • The gridded climate input data were subject to approximations associated with interpolating observation point data to a national grid detailed in Jones et al. (2007).
  • The estimated volume of runoff harvested was subject to the assumptions associated with the STEDI model and the parameters used.
  • The spatial extent of waterbodies was subject to the assumptions and methods associated with the spatial data provided by the South Australian Department for Water.
  • The use of a nine arc-second DEM to determine catchment area may result in off-channel water storages being assigned a catchment much larger or smaller than the true catchment. In some cases an off-channel water storage may be assigned to the catchment of a stream line hundreds of metres away.

Uncertainty Information

The uncertainty estimate was not quantified.