2.6.2.6 Parameterisation

Table 13 lists the parameter distributions used in the GW AEM for the Galilee subregion for sensitivity and uncertainty analysis. The analytic element modelling code TTim (Bakker, 2015) requires for aquifer layers to specify horizontal hydraulic conductivity (Kh) and specific storage (Ss) (for confined aquifers) or specific yield (Sy) (for unconfined aquifers). Aquitard layers require specification of vertical hydraulic conductivity (Kv) and specific storage (Ss).

In finite-difference groundwater modelling codes such as MODFLOW (Harbaugh et al., 2000), the vertical flow between model layers is controlled by the equivalent vertical hydraulic conductivity, computed based on the vertical hydraulic conductivity assigned to the model layers. In the analytic element groundwater code as implemented for the Galilee subregion, vertical flow between aquifers is controlled by the vertical hydraulic conductivity assigned to the interspersed aquitard. It is therefore not necessary to specify vertical hydraulic conductivity to aquifer layers or horizontal hydraulic conductivity to aquitard layers.

For most of the hydraulic properties of the hydrostratigraphic units, insufficient data are available to empirically establish a formal probability distribution. Some local information on aquifer properties is available from various technical reports. For example, Harrington et al. (2012) provides a preliminary review of the available data in the region, while Bleakley et al. (2014) is a more comprehensive summary of currently available hydraulic property information relevant to the Galilee Basin. Bleakley et al. (2014) contains hydraulic field test data, core test data and calibrated model parameters obtained from mining proponent reports, produced as part of the environmental impact statement assessment process. This product is, therefore, the main source to establish the ranges of hydraulic properties in the GW AEM.

Each parameter is assumed to be lognormally distributed, in line with international literature on distributions of hydraulic properties (Carrera et al., 2005). The specific yield of the alluvial hydrostratigraphic unit is the only parameter that is not transformed. The means of the distributions are based on those documented in Bleakley et al. (2014). In some cases, the specific storage values used are not exact matches to those provided in Bleakley et al. (2014) but are conservative estimates selected for modelling purposes (see source notes in Table 13).

The standard deviation is chosen such that the interquartile range, the range between the 25th and 75th percentile, approximately covers at least an order of magnitude. To illustrate the resulting range, Table 13 lists the mean of the distribution as well as the 5th, 50th and 95th percentile of the distribution (based on 10,000 samples). Although the ranges are based on the limited local information, they correspond well to ranges reported in international literature, such as in Batlle-Aguilar et al. (2016). No covariance between parameters is specified as no reliable information is available locally or in international literature. Not specifying covariance will result in conservative predictions as unlikely parameter combinations (e.g. high conductivity and low storage) are retained in the parameter distributions used in the uncertainty analysis.

Table 13 Parameter distributions used for the groundwater analytic element model (GW AEM) for the Galilee subregion

Data: Bioregional Assessment Programme (Dataset 1)

All parameters are considered to be normally distributed.