Uncertainty analysis


In the uncertainty analysis the uncertainty in groundwater model parameters, expressed through their prior parameter distributions, is propagated to the predictions. These prior parameter distributions are constrained by estimates of baseflow and coal seam gas (CSG) water production rates, maximum hydraulic conductivity ratios between layer 3 (Bungawalbin Member) and layer 4 (Kangaroo Creek Sandstone) and, for predictions in layers with groundwater level observations, distance-weighted groundwater level observations. As a result, each prediction in layer 1 (alluvium, Lamington Volcanics and unconfined parts of the sedimentary bedrock) has an individual parameter posterior distribution.

As discussed in the sensitivity analysis, the observation dataset contains limited information to constrain the parameters relevant to the predictions. The exceptions are the drainage parameter controlling CSG water production rate, the hydraulic properties of layer 6 (coal seams of the Walloon Coal Measures), recharge multipliers in zone 1 and 2 and the riverbed conductance multiplier.

The head observations identify zones in the model that, regardless of the parameter values, will have a mismatch between observed and simulated groundwater level that either is always below or always above the predefined acceptance threshold. These zones are regions in the model where simulated groundwater levels are not controlled by parameter values given the current parameterisation, but by the specified boundary conditions or model structure issues.

The median predicted change in groundwater level due to the additional coal resource development across the modelled domain at the prediction locations is less than 0.01 m, with the 95th percentile not exceeding 1 m. The drawdown due to additional coal resource development (additional drawdown) is realised the earliest at model nodes in layer 6, within the planned production period. For model nodes in layer 1, the median year of maximum change is around 2060, after production ceases.

The qualitative uncertainty analysis discusses the perceived effect of the major model assumptions on the model predictions. The assumptions that have the most potential to profoundly alter the results and conclusions of the groundwater modelling are the implementation of the coal resource development pathway (CRDP) and the characterisation of the hydraulic properties of the groundwater system, especially if additional observations of these properties would warrant the conceptual model to be revisited.

Last updated:
11 July 2017
Thumbnail images of the Clarence-Moreton bioregion

Product Finalisation date

20 October 2016