Contaminant screening modelling

Abstract

The dataset was compiled by the Geological and Bioregional Assessment Program from multiple sources referenced within the dataset and/or metadata. A key part of the GBA project was quantifying the potential for a decline in the water quality of unconfined aquifers due to unintentional chemical release at the soil surface. To assess this hazard, a quantitative analysis of chemical migration pathways was undertaken, which involved the estimation of contaminant attenuation by dilution and dispersion in soil and groundwater. This provided a conservative screening approach to identify areas for further analysis. Attenuation calculations involved one-dimensional advection-dispersion (AD) simulations through the unsaturated zone, and three-dimensional AD solute transport within the surficial aquifers. Dilution factor relationships for the combined effect of attenuation in the unsaturated and saturated zone were used to construct spatial maps of the potential for impact on aquifer properties after accidental chemical spills. A higher dilution factor (therefore lower consequence of the surface contamination) was associated with deeper unsaturated zones characterised by heavier soils near the surface, and lower groundwater velocities due to lower hydraulic conductivity and/or hydraulic gradient in the saturated zone. The framework was applied across the Cooper Basin and Beetaloo Sub-basin and resulted in two types of maps. The first identified areas being more susceptible to contamination if soil remediation does not occur within a 10-year period. The second map shows the spatially variable combined dilution factors for a groundwater receptor, which may be used to develop site-specific management plans and mitigation measures.

Attribution

Geological and Bioregional Assessment Program

History

The methods to create this data set are described in the journal paper "Screening of causal pathway networks to identify potential impacts from gas developments". Data inputs and outputs to the chemical screening method are included in this dataset. Soil characteristics for each landscape class were obtained from clay, silt and sand percentiles for each grid cell from the Australian Soil Resource Information System (ASRIS) Atlas of Australian Soils (Johnston et al., 2003). Soil hydraulic parameters were derived from USDA class transfer functions from Carsel and Parrish (1988) and the van Genuchten (1980) and Mualem (1976) equations based on each landscape class soil characteristics (Mallants et al., 2021 in review). The Hydrus-1D modelling was used to estimate the change in chemical concentration from that released at the surface to the point where it reached the groundwater through advection and dispersion, which was then presented as dilution factors. Empirical equations were used to calculate DFUZ for the landscape grid, with varying soil properties, landscape type and depth to water table. Model results were used to derive two types of maps: one showing the susceptibility to groundwater contamination if soil remediation does not occur within 10 years after a leak occurred. The second map displays spatially variable overall dilution factors for a groundwater receptor. To derive the first map, a DFUZ of less than 1000 within 10 years of the chemical release (DF1000), was selected as a threshold to represent a level of impact on groundwater which would trigger a more detailed investigation, and potentially require soil remediation. The duration of 10 years was selected as it was considered reasonable that a surface spill would be detected and remediated within this time frame. This threshold correlated with 95% of hydraulic fracturing fluid chemicals identified in the SANTOS environmental management plan (Santos QNT Pty Ltd, 2019) that would have a risk quotient (RQ) < 1 (i.e are of no concern), and in most cases by several orders of magnitude less than 1. In identifying the likelihood of any impact on groundwater, depth to groundwater was the most important parameter. Depth to water table correlating with a DF1000, plus 10 m (for possible impact) and 5 m (for further investigation required) to account for uncertainty in the ability to accurately estimate the groundwater surface at a regional scale, was used to define the likelihood of impact being ‘not possible’, ‘possible’ or ‘further investigation required’ (Table 1). These impacts assume that the soil would not be remediated within 10 years following the spill. To derive the second map, groundwater dilution factors (DFGW) were estimated from a 3-D AD model representing advection and dispersion via groundwater flow in the aquifer, and applied spatially based on groundwater pore velocity (Mallants et al., 2021 in review). DFGW for a travel distance of 500 metres, from the centre of one grid cell to the boundary of the next, was estimated across the basins. The 3-D AD models were run until the peak concentration occurred 500 m from the contamination source. The combined dilution factor (DFT) was calculated as the product of DFUZ and DFGW, where DFUZ reflects the full contaminant plume arriving at the groundwater table after a very long travel time across the unsaturated zone (several hundred years).