Observations and predictions


Three types of data were included in the observation database to constrain the groundwater model parameters in the uncertainty analysis. These include historical groundwater levels, coal seam gas (CSG) water production forecasts and average streamflow in September and October as proxy for the surface water – groundwater flux. This section describes these observation datasets. Section provides a detailed discussion on how these observations are used to constrain the groundwater model parameters.

A total of 188 bores are available in the observation database. Only 30 bores have transient observations. Of the 188 bores, only 43 have surveyed spatial coordinates and ground level and are deemed reliable. While the entire dataset is used to guide initial model development, only the data associated with the 43 reliable bores is used in the uncertainty analysis.

The total CSG water production rates forecasted by Metgaso Limited (Parsons Brinckerhoff, 2013) provide a point of reference for the simulated CSG water production rates.

Likewise, the observed streamflow at the Casino surface water model node (CLM_008) in September and October will provide a point of reference to the simulated average surface water – groundwater flux, as streamflow in these months is most likely dominated by the groundwater contribution.

A total of 38 parameters were varied during the sensitivity and uncertainty analysis. Ten-thousand parameter combinations were generated for the entire parameter space for the model sequence using a maximin Latin Hypercube design. Within the operational constraints of the project it was possible to successfully evaluate a total of 3877 parameter combinations. Recharge, drainage conductance and riverbed conductance were found to be the three most sensitive parameters for head observations in layer 1 (alluvium, Lamington Volcanics and unconfined parts of the sedimentary bedrock). The impact on layer 3 (Bungawalbin Member) is dominated by the vertical hydraulic conductivity anisotropy of layer 3, followed by the horizontal hydraulic conductivity in layers 3 and 4 (Kangaroo Creek Sandstone). The drain conductance of the CSG wells and the storage coefficient and hydraulic conductivity of layer 6 (coal seams of the Walloon Coal Measures) controlled the amount of water produced from the depressurisation process. As a result, these parameters have a large impact on the drawdown prediction as this defines the stress on the system.

The model generated heads corresponding to head observations are mostly not sensitive to those parameters dominating the change in drawdown due to additional coal resource development (additional drawdown) given the current model implementation. The head observations, therefore, have limited scope to constrain the parameters that generate the most uncertainty in the additional drawdown predictions. The average surface water – groundwater flux is most sensitive to the riverbed conductance followed by the drainage conductance of the drainage boundary at the top of the model and the recharge multipliers of recharge in the alluvium and volcanics.

Last updated:
23 October 2018
Thumbnail images of the Clarence-Moreton bioregion

Product Finalisation date