Reid, L.B., Townley, L.R., and Smith, A.J. (2013), Impacts of CSG Production on Regional Groundwater Systems, IAH Congress, Perth, September.

The emergence of a coal seam gas (CSG) industry in Australia has been accompanied by concern about potential impacts on regional aquifers, including the Great Artesian Basin. The volumes of water withdrawn during production of CSG need to be understood in the context of water use from all sources, including both surface water and groundwater. Nevertheless, prediction of the potential impacts of CSG production has been difficult, for a number of reasons, including the limitations of readily available groundwater modelling software and a lack of understanding of the differences between approaches taken by reservoir engineers and hydrogeologists. The purpose of this paper is to explain (i) the differences between methods used by reservoir engineers and hydrogeologists, (ii) why traditional groundwater modelling, without modification, is not capable of predicting the quantity and quality of groundwater produced during gas production, and (iii) how impacts on the surrounding environment can be predicted with a greater level of confidence once modifications to traditional groundwater modelling methods are made. Multiphase reservoir models (e.g. based on ECLIPSE) predict desorption of gas, migration of gas towards production wells and co-production of water, but are not designed to predict regional impacts at the water table or where coal seams sub-crop alluvial aquifers. Regional scale groundwater flow models (e.g. based on MODFLOW and FEFLOW) represent changes in storage of water in a fundamentally different way, and cannot predict rates of production of water or depressurisation to match the predictions of reservoir models. A number of CSG projects have been approved in Australia in recent years. This paper will compare and contrast the approaches taken to predict potential impacts on regional aquifers, as described in publicly available Environmental Impact Statements. The Australian experience will be compared with what has been learned in the Powder River Basin in the USA, including the differences between predictions and observations, before and after the start of gas production. Modifications to traditional groundwater modelling software will be described, based on the use of TOUGH2 to predict multiphase behaviour and on internal modifications to MODFLOW.