2.6.2.6 Parameterisation

The Great Artesian Basin comprises an alternating sequence of permeable sandstone aquifers and lower permeability siltstone and mudstone aquitards in the Surat cumulative management area (CMA). Flow in the aquifers is dominated by subhorizontal flow (QWC, 2012), which is governed by the horizontal hydraulic conductivity (Kh). Vertical leakage from the aquifers through the low permeability aquitards, which occurs at a much slower rate (QWC, 2012) is governed by the vertical hydraulic conductivity (Kv) of the aquitards. Storage values represent the capacity of an aquifer to release groundwater, where specific storage (Ss) relates to confined aquifers and specific yield (Sy) relates to unconfined aquifers. Parameter values are estimated during model calibration and uncertainty analysis as summarised below and described in GHD (2012):

‘Calibration of the steady-state regional model and transient sub-model was carried out using the PEST suite of software (Doherty, 2010) and adopting the following overall framework:

1. A preliminary calibration of the steady-state regional model to provide initial heads and boundary conditions for the transient sub-model;
2. Calibration of the transient sub-model using detailed groundwater level data for the WCM in the vicinity of the existing Kogan North/Daandine CSG gas field; and
3. Re-calibration of the steady-state regional model but this time adopting calibrated hydraulic conductivity values for the WCM (model layers 9, 10 and 11) from the transient calibration within the sub-model area.’ (GHD, 2012, p. 72)

Preliminary calibration of the steady-state Office of Groundwater Impact Assessment (OGIA) model replicated long-term mean conditions pre 1995 (i.e. pre coal seam gas (CSG) extraction) through reference to mean groundwater levels from 1541 boreholes (GHD, 2012). This approach maximises the size of the data set by collating information from a number of different sources, completing missing information on the strata monitored by each bore and calibrating to average groundwater levels at each bore largely independent of the date of the reading (GHD, 2012). GHD (2012) states that ‘… while there are some notable exceptions, these plots also suggest a general lack of long-term trends in groundwater levels, particularly post 1960. The use of calibration targets based on average groundwater levels is therefore not considered likely to bias the model calibration significantly.’

Modelled hydraulic conductivity and storage parameters were refined during the calibration process by adjusting initial hydraulic conductivity values within the adopted permissible range to minimise overall residuals in each layer (GHD, 2012). Further refinement of these parameter values was then achieved using pilot points that were adjusted to reduce the residual errors further. The calibrated mean hydraulic conductivity values for the main aquifer layers are generally higher than for the aquitard layers. Further pilot points were used to adjust the conductance of the General Head Boundary (GHB) cells located at the southern and eastern boundaries of the regional model (GHD, 2012). Table 7 lists the minimum, average and maximum Kh and Kv calibrated values used in the predictive model runs and uncertainty analysis.

Storage values for both the Walloon Coal Measures and the Bandanna Formation model layers are not well calibrated, since only limited transient calibrations have been possible to date (GHD, 2012). A more detailed transient sub-model is used to calibrate specific yield (Sy) and specific storage (Ss) parameter values for the Walloon Coal Measures model layers. The transient sub-model is constructed using 250 m × 250 m cells around the existing Daandine Coal Seam Gas Project production field, which has been operational since 2005 and where detailed groundwater levels are available. Storage values are varied spatially in the productive coal layers (Walloon Coal Measures and Bandanna Formation) based on a relationship with depth and constrained by previously modelled data. ‘As observed groundwater levels were only available for the productive coal of the WCM (model layer 10) in the Kogan North/Daandine area, only the storage value for this layer can be considered to have been calibrated using this model’ (GHD, 2012). A constant storage value for other model layers is assumed to be 5.0 x 10⁵ (m^-1). Storage values for the aquifer outcrop areas, Condamine Alluvium, Main Range Volcanics and alluvium outside the Condamine area (i.e. unconfined storage) are based on specific yield estimates. Table 7 lists the adopted storage values for each OGIA model layer.