Carmichael model review

GHD Pty Ltd (2013) reports on the numerical model for the Carmichael Mine. Some basic parameters are outlined in Table 5. The highest grid resolution, in the location of open-cut mine pits and underground mines, is 50 m. This coarsens toward the margins of the model to a maximum cell size of 1 km. The model contains 3,318,470 active cells. It contains 12 layers and has a transient modelling time of 196 years (Bleakley et al., 2014). The mining phase of the transient modelling run was from 2013 to 2071, with recovery phase post mining running out to 2209.

The Permian and Carboniferous rocks are represented in the model by five layers. Four model layers represent main resource seams (AB and D), interburden and overburden to the AB seam. The lowermost layer incorporates the relatively low permeability Joe Joe Group, as well as the lower sections of the Colinlea Sandstone, which includes the E and F coal seams (Bleakley et al., 2014).

Surface water bodies in the model domain are mapped as ephemeral in the Bureau of Meteorology’s Australian Hydrological Geospatial Fabric, and were essentially represented as drain cells in the model; river stand height and riverbed elevation are given the same value, meaning water may flow into the river as baseflow, but not from the river to an aquifer as leakage. Riverbed conductance values were assigned with the intention of making them high enough that flow from aquifers to surface water was controlled by the aquifer properties rather than by the riverbed conductance (GHD Pty Ltd, 2013). Underground mines and open-cut mine pits are represented by drain cells.

General-head boundaries (GHBs) are assigned to the outer edge of the active cells in the model. Outflow was simulated on the east boundary with GHB in the Cenozoic (Tertiary) units, and inflow was simulated on the north and west boundaries with GHB in the Triassic and Permian geologic units, and in the south with GHB in the Cenozoic units. No GHBs were assigned to the top model layer (Quaternary alluvium), as discharge from this unit is expected to be primarily through baseflow to surface water bodies (GHD Pty Ltd, 2013).

Initial long-term mean rainfall recharge rates were estimated using recharge–runoff modelling techniques. Recharge rates were refined during the calibration process by allowing them to vary automatically between 0.5 and 5.0 mm/year. This assisted with obtaining the best agreement between observed and modelled water levels. Recharge was enhanced near major streams to simulate stream interactions with shallow groundwater in alluvium.

Predicted near-surface impacts of the model include drawdown in the alluvium to the east of the mining area greater than 2 m, and drawdowns up to 1 m along the Carmichael River corridor (GHD Pty Ltd, 2013). This contrasts with drawdown at depth in the coal seams that are being targeted for mining, which is in excess of 300 m during the mining phase.

The model predicted a drawdown during the operational phase of up to 10 m along the western boundary of the model, where the Dunda beds, Clematis Group (Sandstone) and/or the Moolayember Formation are mapped as outcrop.

Of particular importance is the potential impact on springs that are situated close to the predicted cone of drawdown. The closest of these is the Doongmabulla Springs complex, which is thought to source water from the Clematis Group aquifer. In this instance the predicted drawdown was 0.5 m to 0.19 m (GHD Pty Ltd, 2013).

Subsequent to the initial model run, the model was re-run with a range of GHB scenarios at the western boundary to address conditions on the Commonwealth’s Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) Approval and was subject to an independent peer review.

Last updated:
6 December 2018
Thumbnail of the Galilee subregion

Product Finalisation date