The GAB covers parts of New South Wales, Northern Territory, Queensland and South Australia, each with their own water management regimes. Regular assessment of the groundwater resources of the GAB at a whole-of-basin scale is important for reviewing the effectiveness and interaction of the different management approaches. However, historically this type of assessment has been undertaken on an ad hoc basis, by different agencies, for different reporting units, using different methodologies and input datasets, meaning it is difficult to compare assessments.

The water balance presented here has been undertaken to test incorporating and communicating uncertainty estimates as part of a whole-of-GAB water balance. The water balance incorporates new work, where available, from the jurisdictions and from this project (specifically groundwater recharge estimates). This project has produced quantified uncertainty for a single component of the water balance - groundwater recharge, the largest input component of the water balance. The water balance, as presented in this report, is not intended to represent a comprehensive critical appraisal of the techniques used by previous workers to estimate each element of the water balance, nor was it intended to develop new techniques for the estimation of the water balance elements other than groundwater recharge.

The incorporation of a component of the uncertainty in the water balance of the GAB is new. The uncertainty in the water balance has always been there, in all past iterations, though it has not been quantified. This estimation of uncertainty is important in the communication of the water balance, as it highlights several key issues:

• using current information, a whole-of-basin water balance for the GAB is not sufficiently detailed to be of use to water managers

• local monitoring of groundwater levels and pressures remains the primary management tool for monitoring groundwater resources in the GAB.

The Project produced a point-in-time assessment of the water balance of the GAB, comparing inflows (including long-term average groundwater recharge) and outflows to the main regional aquifers, for the year 2019 (The year 2019 was the latest year for which data was available at the start of the Project).

Most components of the water balance cannot be directly measured and in some cases reported values are long-term averages (e.g. groundwater recharge) and estimated values for the reporting time period. As such, the water balance relies on indirect measurements, long term averages and approximations, which have a level of inherent uncertainty resulting from the underlying assumptions used in their estimation.

To calculate a whole-of GAB water balance, the project divided all the formations in the GAB into four different ‘aquifer groups’: Rolling Downs Aquifer Group; Cadna-owie Aquifer Group; Hutton – Injune Creek Aquifer Group; and Precipice Aquifer Group. This approach was developed by KCB (KCB 2016b,c,d) in Queensland and has been adopted for this Project, as it ensures all inflows and outflows from the GAB were included in the water balance.

The whole-of-GAB water balance has been calculated, based on water balances estimates calculated for each constituent sub-basin – the Eromanga, Carpentaria and Surat basins. This was done to provide a picture of variations in the water balance across different parts of the GAB.

The whole-of-GAB water balance, calculated using the 5th, 50th and 95th percentiles of modelled long term average groundwater recharge rates, estimates a range of storage change volumes of -859 GL, -29 GL and +1,212 GL respectively, in 2019. The large variation in estimated storage volumes, ranging from a negative to positive groundwater storage change, highlights the large uncertainty associated with the water balance.

Using 50th percentile modelled groundwater recharge rates, sub-basin water balances shows an increase in groundwater storage for the Eromanga Basin (-229 GL 5th percentile; 51 GL 50th percentile and 424 GL 95th percentile) and a decrease in groundwater storage of the Carpentaria Basin (-413 GL 5th percentile; -72 GL 50th percentile and 511 GL 95th percentile) and Surat Basin (-217 GL 5th percentile; -9 GL 50th percentile and 277GL 95th percentile).

Using 50th percentile modelled groundwater recharge rates for major aquifer groups across the basin, water balance estimates for the Cadna-owie Aquifer Group and Precipice Aquifer Group suggest a decrease in storage volumes, while water balance estimates for the Hutton - Injune Creek Aquifer Group and the Rolling Downs Aquifer Group suggest an increase in storage volumes.

While the whole-of-GAB, sub-basin and major aquifer water balances provide basin-wide perspectives of the groundwater resource components, they also highlight the high uncertainties associated with estimating groundwater recharge at a regional scale. The large range in groundwater storage values calculated for the water balance presented here, are too great to confidently provide a whole-of-GAB scale assessment of groundwater resources.

The techniques used to estimate some water balance components have improved, for example Queensland has developed a repeatable methodology for estimating unmetered groundwater extraction. This project shows that our ability to confidently model groundwater recharge to the GAB is still evolving, and will continue to improve as further investigations are undertaken. Limited hydrograph analysis showed that areas where formerly free-flowing artesian bores have been rehabilitated, through the Great Artesian Basin Sustainability Initiative and predecessor programs, are seeing stabilisation or even recovery of water levels. Monitoring bores and hydrograph analysis continue to be important elements of any water resource assessment.

Undertaking the water balance for the GAB and sub-basins has been useful for highlighting components of the water balance that should be considered carefully by groundwater resource managers and where necessary, targeted for future research eg. groundwater recharge, consistent GAB wide bore discharge estimates and evapotranspiration.