Introduction

In 2020–2021, the Murray-Darling Basin (MDB) in south-eastern Australia accounted for 60% of all irrigated land in Australia (ABS 2022). This predominantly semiarid to arid hydrological basin is highly allocated and regulated, and groundwater is a key water resource for ecosystems, regional town water supply and irrigation.

Key challenges for the basin are associated with a changing climate and water management in highly allocated and sometimes overallocated environments. Observed and predicted changes in climate are predominantly towards reduced rainfall or changes in rainfall patterns, increased temperatures and evapotranspiration, and longer dry periods (Prosser et al. 2021; Whetton and Chiew 2021). Water management challenges involve uncertainty in firstly estimating limits to annual groundwater extraction (or sustainable diversion limits: SDLs) and seasonal adaption of actual water allocations depending on groundwater and climatic conditions.

There are several papers that outline risks and challenges related to surface water and groundwater in the Murray-Darling Basin or Southern Australia (Cook et al. 2022; Ross et al. 2022; Simmons et al. 2019; Van Dijk et al. 2006). Van Dijk et al. (2006) noted groundwater extraction as one of the six key risks to the shared water resources of the Murray-Darling Basin: that extraction of groundwater can reduce streamflow through the reduction of baseflow or inducing local recharge from streams. However, it provided only a short discussion of the impacts on groundwater resources and groundwater-dependent ecosystems (GDEs). In a recent review of these risks, Pittock et al. (2023) and Ross et al. (2022) defined risks to groundwater that included uncertainty around SDLs in a changing climate, lack of monitoring and knowledge on surface-water/groundwater connectivity and greater integration of surface-water and groundwater management. They observed a capability gap in multiscale groundwater management, at a whole of aquifer scale for water management planning and local scale for observing and responding to impacts from groundwater use. They noted that the current Murray-Darling Basin Plan does not fully consider groundwater resources or GDEs (Ross et al. 2022; WGCS 2012).

From 2015 to 2018 the Murray-Darling Basin Authority (MDBA) and National Centre for Groundwater Research and Training (NCGRT) Strategic Groundwater Research Partnership undertook a suite of research in the areas of surface-water/groundwater connectivity, improving estimates of groundwater recharge and identifying socioeconomic factors that will improve water management in the MDB (Simmons et al. 2019). They made recommendations for future research and assessment, which included: increasing data acquisition to improve estimates of fluxes between rivers and groundwater; stating uncertainty bounds in recharge estimates and groundwater modelling; developing groundwater management tools that consider the effects of extraction on GDEs and surface-water resources; and including socio-economic assessments in groundwater management modelling.

Cook et al. (2022) identified 18 current challenges to groundwater extraction, which were prioritised by surveys of representatives from the Australian groundwater community: government, research institutions and private industry. Despite increasing use of adaptive management and risk-based limits to extraction, they noted that one of the greatest perceived challenges for groundwater management in Australia was determining regional-scale groundwater extraction limits. Challenges related to the connections between groundwater and ecosystems included a difficulty in determining and implementing maximum drawdown criteria for groundwater for adaptive management (trigger levels) and determining the water needs of ecosystems. Managing groundwater impacts on surface water and predicting the impacts of climate change on groundwater availability were other highly ranked challenges.

Many of the risks, challenges and opportunities raised in these papers are sensitive to future shifts in climate and aridity. In particular, surface-water/groundwater interactions, groundwater recharge, and the long-term averages and transient thresholds for water requirements of GDEs are likely to respond to changes in climate. This report discusses recent and emerging advances in knowledge in these three areas. It provides an overview of the ongoing efforts towards understanding the effects of a changing climate on groundwater recharge, surface-water/groundwater interactions and responses of GDEs, and discusses opportunities to manage groundwater as part of a whole basin water system and increase the visibility of GW in the basin water planning.

Site description and historical groundwater management in the Murray-Darling Basin

The MDB covers an area of 1.06 million km2 in the south-east of Australia (Fig. 1), including sections of Queensland (Qld), New South Wales (NSW), Victoria (Vic) and South Australia (SA) and all of the Australian Capital Territory (ACT). The basin includes Köppen–Geiger climate types ranging from oceanic (Cfb) in the southern upland parts of the basin (11% of total area) and subhumid (Cfa) in the northern upland regions (26% of total area), to semiarid (Bsk and Bsh) and desert (arid) (Bwk and Bwh) in the western lowland parts of the basin (56 and 7% of total area respectively; Fig. 2).

Fig. 1
figure 1

Murray-Darling Basin system, including a simplified geological regions, b major alluvial groundwater systems and c location of the basin across the five states and territory in Australia. States: SA – South Australia, Qld – Queensland, NSW – New South Wales, ACT – Australian Capital Territory, Vic – Victoria

Full size image
Fig. 2
figure 2

Köppen–Geiger climate types for the Murray-Darling Basin

Full size image

The river network drains from the Great Dividing Range, which stretches along the eastern and south-eastern boundary of the MDB, through the hotter and drier parts of inland Australia before discharging to the ocean in the Mediterranean climate (Csb) in South Australia. The River Murray drains the southern part of the catchment, including the Loddon, Campaspe, Murrumbidgee, Goulburn-Broken and Lachlan rivers. The southern basin has winter-dominant rainfall, and higher and more reliable streamflow (Walker 2022). This, and the proximity to large cities in the south-east of Australia, has led to the earlier development of the southern basin and larger areas of irrigated agriculture.

The Darling River drains the northern basin, and tributaries include the Paroo, Warrego and Maranoa Rivers (draining desert and semiarid regions), and Condamine-Balone-Culgoa, Gwydir, Namoi and Macquarie rivers (draining semiarid and subtropical regions). The northern basin is located in hot semiarid and subtropical climates, with summer dominant rainfall and higher seasonal variation, resulting in lower and more variable streamflow and higher susceptibility to drought (Bureau of Meteorology 2020). During the Millennium Drought from 2001 to 2009, and a more recent drought in 2019, flow in the Darling River was reduced to almost zero (Mallen‐Cooper and Zampatti 2020; Vertessy et al. 2019).

Three broad groups of aquifers important for water resources are present in the MDB: fractured rock aquifers in the eastern upland and far western areas of the basin, alluvial deposits associated with the major tributary floodplains and porous rock aquifers in the northern and western parts of the basin (Fig. 1a; Hart et al. 2020). Deeper geology of the MDB is described in Kingham (1998), and includes the Great Artesian Basin (GAB).

Fractured rock aquifers occur throughout the Great Dividing Range in Queensland, New South Wales, the ACT and Victoria, and in the Mount Lofty Ranges of South Australia. The aquifers are generally local to intermediate in scale (defined by Tóth 1963), of moderate productivity and contribute to baseflow in upland rivers. Recharge is predominantly diffuse due to higher rainfall in these regions. Productivity is relatively low in the Lachlan Fold Belt (Foster and Gray 2000), moderate in the western Adelaide Fold Belt (Habermehl 2020) and generally higher in parts of the upper New England and eastern Adelaide Fold belts.

Major alluvial deposits consist of deposits of sands and gravels from the main river tributaries and exist in both upper paleovalleys, which are spatially defined by floodplain channels, and lower deep leads, which are more generally widespread reaches (Taylor et al. 2021). Key aquifer units are shown in Fig. 1b. The alluvial deposits are generally intermediate to regional groundwater systems with medium to high productivity and good quality water. These systems are often recharged by losing river systems (Lamontagne et al. 2014) in addition to diffuse recharge from rainfall (Crosbie et al. 2010b) and overbank flood recharge (Doble et al. 2014).

Porous rock aquifers are located within the arid and semiarid western Murray Basin (Fig. 1a), where stored groundwater is thought to be a result of paleorecharge from an era when the climate was wetter (Commander 2004). Water resources are generally saline in the central and western parts of the Murray Geological Basin, with fresher areas in the eastern region closer to recharge areas (Murray-Darling Basin Commission 1999), some of which is suitable for irrigation and consumption.

Dryland salinity is an issue in the fractured rock and upper alluvium aquifers in NSW and Victoria and the Murray Basin in South Australia and Victoria (Hart et al. 2020; Walker and Prosser 2021). Historical clearance of native vegetation and the development of irrigation districts has led to locally raised levels in the naturally saline groundwater. This has been expressed as salt scalds on the land surface or increased baseflow of saline groundwater to the floodplains and river (Barnett 1989). Saline baseflow has increased river salinity and caused a widespread decline in floodplain vegetation conditions (Doble et al. 2006; Holland et al. 2009; Overton et al. 2006). As a response, groundwater extraction networks to intercept saline groundwater prior to reaching the river valley (salt interception schemes), have been constructed in South Australia, NSW and Victoria (Middlemis et al. 2005; Munday et al. 2005).

Water management

Groundwater in Australia is managed by state or territory jurisdictions, under regulation and frameworks developed by the federal government. A more detailed overview of national groundwater management and legislation can be gained from Nelson et al. (2020), Cook et al. (2022), Rinaudo et al. (2020), Walker et al. (2021b). A brief overview is presented in the following.

The 1994 National Water Reform Framework Agreement instigated water allocation systems in Australia at a national scale and formed an agreement between the federal government and all states and territories (Nelson et al. 2020). This, and the following 2004 Intergovernmental Agreement on the National Water Initiative (NWI), founded the basis of a national water reform to balance allocation of water resources to balance economic, social and environmental benefits from water. The NWI committed states to develop comprehensive water plans, achieve sustainable water use in overallocated or stressed water systems, introduce registers of water rights and water accounting and expand trade in water rights (Council of Australian Governments 2004).

The Murray-Darling Basin Plan (2012), or Basin Plan, arose from an agreement by all basin governments to better manage all water resources within the MDB (Australian Government 2012). The Basin Plan specifies requirements for compliance with state-defined water diversion limits, setting aside water for ecosystems. Under the Basin Plan, long-term average sustainable diversion limits (SDLs) for each of the management areas are defined by state-developed groundwater resource plans. The MDBA has the authority to propose adjustments to the SDLs to reflect new or improved information relating to the groundwater of the groundwater SDL resource unit and minor changes were made most recently in 2018. The SDLs are considered an upper limit to extraction. Basin states can alter the annual groundwater entitlements as a proportion of the SDL, based on current local groundwater resource conditions, variations in recharge and observed effects on GDEs. The Murray-Darling Basin Plan is due to be reviewed in 2026.

The National Groundwater Strategic Framework (2016–2026) was developed by Australian federal, state and territory governments in 2017 into a 10-year vision to support a national approach to groundwater management (Council of Australian Governments 2016). It recommends 28 actions in 3 priority areas: sustainable extraction and optimal use, providing investment confidence and planning and managing now for the future.

The National Water Initiative, Murray-Darling Basin Plan and National Groundwater Strategic Framework are implemented at a regional or state scale, as water governance remains the responsibility of state and territory governments, even though the core vision and values are nationally or basin-wide consistent. They are built on a system of tradeable water entitlements, metering of water use, and water-source-specific rules based on regular planning cycles with community consultation. Specific provisions vary across water resource plan areas and many areas have a long history of water management (Guillaume et al. 2022), though typically with a focus on expanding on water use until the 1980s, with a shift towards sustainability since the 1990s.

Salinity management

Salinity has been a known issue in the MDB since the 1960s, and in the mid-1980s, increasing knowledge of the environmental and economic challenges that it posed led to three consecutive salinity strategies being developed by the basin states and the Commonwealth. The Salinity and Drainage Strategy 1988–2000 (S&DS) and Basin Salinity Management Strategy 2001–2015 (BSMS; Murray-Darling Basin Authority 2001) focused on addressing the impacts of dryland and irrigation salinity and initiated the development of salt interception schemes, a salinity register for accountability of salt management, and improvements to irrigation efficiency and drainage. The current Basin Salinity Management Strategy 2030 (BSMS2030; MDBA 2015) maintains a commitment to the existing basin salinity targets and the existing accountability framework. It fosters smarter, more cost-effective management of salt within the MDB, including salt interception operations and combined groundwater and river flow management. The implementation of these salinity management plans has resulted in improvements to river and dryland salinity. Salinity is recognised as being largely under control, but is still considered to be a risk to the basin which requires ongoing active management.

Biophysical challenges and opportunities for groundwater in the MDB

Trends in climate, groundwater use and their impacts on groundwater levels

Historical changes in groundwater level

Groundwater levels have been observed to be falling in the eight main alluvial groundwater systems in the MDB that represent almost 75% of groundwater use (Fu et al. 2022). Results from the statistical analysis of 910 bores in NSW by Fu et al. (2022) showed an overall decreasing trend in water-table levels across alluvial aquifers during the last 50 years (1971–2021). Only a small number of bores in the Gwydir, Condamine and Murrumbidgee alluviums showed rising groundwater levels. The declining level trends were thought to be attributable to changes in recharge due to a reduction in rainfall and potentially historical effects of groundwater extraction.

In their study, Fu et al. (2022) noted a decreasing trend in the overall number of groundwater level observations in these regions between 2016 and 2021, despite the increasing use of automated loggers, and that the number of bores used for observation had, in 2021, decreased to 10% of the number of bores observed in 2005.

Trends in groundwater use

Under the Basin Plan, groundwater extraction volumes for all SDL units across the MDB and compliance with SDLs must be reported. The reporting over the last decade shows that there has been an increasing trend in groundwater extraction from 2012–2013 to 2018–2019, but this variance in groundwater extraction was able to be explained by responses to rainfall variation, where rates of irrigation are increased to offset lower seasonal rainfall, rather than growth in new licences (Walker 2022). There was no significant increase in groundwater extraction for the non-Victorian portion of the MDB for the period 2003–2004 to 2020–2021. Although there has been an increasing trend in groundwater extraction within the Victorian portion of the MDB, groundwater thresholds were able to be maintained with reductions in seasonal allocations (Walker 2023). Taken together, these results suggest only a very minor potential increase in near future groundwater extraction volumes based on historical trends.

The overall lack of increase in groundwater extraction is due to a range of causes. For many other units, there is no demand for groundwater due to the high salinity, slow transmissivity or remoteness. In the NSW major alluvial systems, the extraction limit was reduced between the 2003–2004 and 2020–2021 irrigation seasons due to the new SDL under the Basin Plan. There has been a cap on entitlements for most other connected units. In regional groundwater systems, legacy impacts on groundwater levels may still occur due to decades-long time lags between stabilisation of groundwater extraction and stabilisation of groundwater levels.

It is possible that new commodities, market prices, technology changes or changes in government policy may cause groundwater extraction to increase through increased licencing or the activation of sleeper licences, where entitlements have been granted but not used or traded. An adaptive management framework should provide early warning of increasing risk to groundwater levels and allow an appropriate management response Walker et al. (2020b).

Potential changes in climate

Following the Millennium Drought of 2001–2009, many papers were published on climatic changes, for example Murphy and Timbal (2008), the effects on surface water and groundwater (van Dijk et al. 2013), and water and salt balances (Austin et al. 2010). The semiarid areas of south-eastern Australia have a high natural hydroclimatic variability, and are historically characterised by droughts and floods (Leblanc et al. 2012). Climate change has augmented both water scarcity and hydroclimatic variability, through increasing temperatures, decreased rainfall and more frequent droughts (Murphy and Timbal 2008), thereby reducing streamflow and groundwater recharge.

Crosbie et al. (2012) developed projections of future Köppen–Geiger climate classifications for the Australian continent based on five future warming scenarios, using 17 global climate models (GCMs). In the Murray-Darling Basin, the arid zone was predicted to increase by 89% under a future temperature of 2.4 °C (Fig. 3). The semiarid zone was predicted to increase by 18%, and the temperate zone decrease by 45% (Crosbie et al. 2012). Predicted changes from temperate to semiarid climates are most pronounced in the northern and south-western parts of the MDB. This is likely to have implications for water demand due to changes in crop water requirements, as well as groundwater recharge, surface runoff and vegetation coverage.

Fig. 3
figure 3

Existing and projected Köppen–Geiger climates for historical and future climate projections. Developed from Crosbie et al. (2012)

Full size image

More recently, Kirono et al. (2020) developed projections for four drought metrics (percent time spent in droughts, mean drought duration, mean drought frequency, and mean drought intensity) for the Australian continent using CMIP5 global climate model simulations. They found that the percent of time spent in drought is likely to increase in most parts of the MDB under a possible but extreme RCP8.5 emissions future (equivalent to 4.3° increase in temperature by 2100). The mean duration of droughts and intensity of droughts is likely to increase for most of the basin, particularly the northern and southern areas. There was found to be less of an increasing trend in the frequency of droughts, but an increase in the frequency of extreme droughts was possible. Future groundwater management in the MDB will need to not only consider management to annual averages or annual climate variations, but also to develop capacity and resilience to longer and more extreme drought events (Prosser et al. 2021; Walker et al. 2021b).

There are opportunities for better understanding the vulnerability of groundwater and GDEs to climate change to provide knowledge for prioritising actions to support climate resilience. The 2020 Basin Plan Evaluation considers the vulnerability of ecosystems to climate change as a function of groundwater through baseflow, vertical connectivity between groundwater, surface water and with GDEs, reduction in groundwater recharge through decreased overbank/unregulated flow events and groundwater quality, but makes no reference to the vulnerability of groundwater storages (MDBA 2020a). Recent work in the Murray-Darling Water and Environment Research Program has started to assess potential groundwater system resilience, stress and sustainability under historical and observed conditions (Rojas et al. 2021; 2023).

There are still many new opportunities for novel data acquisition to observe and model groundwater trends and responses to climate change and extraction. These include gravimetric and surface deformation analyses with high spatial and temporal coverage (Castellazzi and Schmid 2021), new geophysical methods for estimating groundwater recharge and changes in quality (Li et al. 2020, 2021) and expansion of soil-moisture sensing data and telemetry (Cooper et al. 2021; Soylu and Bras 2021; Tangdamrongsub et al. 2020). Integrated groundwater modelling can predict relative data worth and optimal location and frequency for the most valuable data observations.

Impacts on groundwater recharge from climate change

The reviews of the many studies into diffuse groundwater recharge over the past few decades have consistently shown that the three biggest predictors of recharge are climate, soils and vegetation (Crosbie et al. 2010a; Kim and Jackson 2012; Scanlon et al. 2006). While it is reasonable to assume that the physical properties of soils will change little over the time scale of decades, changes are expected in climate and vegetation coverage, leading to subsequent changes in diffuse recharge.

At a national scale, the Climate Change in Australia program (CSIRO and BoM 2015) has synthesised the understanding of the current climate, trends in historical climate and CMIP5 GCM projections of a future climate for use at the regional scale. For the southern part of the MDB the projections were for increased temperatures, decreased annual rainfall, particularly for winter/spring and an increase in rainfall intensity for heavy rainfall events. For the eastern and north-western regions of the MDB, the projections were for increased temperatures, uncertain changes to annual rainfall but decreased winter rainfall and an increase in rainfall intensity for heavy rainfall events.

The cycles of droughts and floods are well documented in the MDB (van Dijk et al. 2013) but there are no studies specifically attributing changes in recharge to climate change. The closest is in the south-east of SA (adjacent to the SW boundary of the MDB) where Fu et al. (2019) demonstrated that the observed reductions in recharge since 1970 were attributable to reduced winter rainfall, increased summer rainfall and increasing temperature and potential evapotranspiration. The same changes in climate are projected for the future in this region (and the southern MDB; CSIRO and BoM 2015).

There has been increasing amounts of research into the climate change impacts on groundwater worldwide over the past few decades, the processes involved and challenges to be solved have been the subject of several review papers (Amanambu et al. 2020; Atawneh et al. 2021; Smerdon 2017; Taylor et al. 2013; Walker et al. 2021b). Most of these studies use a similar method for making projections of diffuse recharge under a future climate: (1) select the representative concentration pathways (predicted future CO2 concentrations) to be investigated through a suite of global climate models (GCMs); (2) downscale the GCM results for use locally; and (3) run the downscaled future climate timeseries through a hydrological model. Each of these steps has many decision points and inherent assumptions, so different studies are rarely directly comparable.

There have been a number of studies investigating diffuse recharge under a future climate across the MDB. Littleboy et al. (2015) used the PERFECT model to investigate diffuse recharge and runoff under a future climate across NSW. This modelling used four CMIP3 GCMs and three downscaling variants (Evans et al. 2014) for a near future (2020–2039) and far future (2060–2079) relative to the (then) current climate (1990–2009). DELWP (2016) recommended that the change in future recharge across Victoria should be assumed to be the same as the change in future runoff as modelled using 42 CMIP5 GCMs using the SIMHYD model (Potter et al. 2016). This modelling provided near future (2031–2050) and far future (2056–2075) climates relative to current climate (1986–2005), the downscaling method used was seasonal scaling of climate parameters with daily scaling of rainfall. Crosbie et al. (2013) modelled future recharge at the continental scale. This study used 16 CMIP3 GCMS and three global warming scenarios with downscaling using seasonal scaling of climate parameters and daily scaling of rainfall to investigate a 2050 future climate relative to a 1990 baseline. The low global warming scenario of +1 °C is most representative of the near future and the high global warming scenario of +2.4 °C is most representative of the far future.

Each of the three studies acknowledged the uncertainties in the future diffuse recharge but, for this review, the mean result from NSW (Littleboy et al. 2015), the median result from Victoria (DELWP 2016) and the 50% probability of exceedance for the MDB (Crosbie et al. 2013) are compared as the ‘best’ estimate of the future diffuse recharge for each study. At the catchment scale it can be seen that the Victorian and NSW recharge predictions for the near future (Fig. 4a) both predominantly show a reduction in recharge, but for the far future (Fig. 4b), the Victorian projections are for further reductions in recharge, whereas the NSW projections (mostly) change to an increase in recharge. The national scale projections show a similar trend to those from NSW in that the near future (Fig. 4c) is projected to have a decrease in recharge, while most of the eastern part of NSW is predicted to have an increase in recharge for the far future, and for most of Victoria, slightly less of a decrease in recharge for the far future is indicated (Fig. 4d).

Fig. 4
figure 4

A comparison of the change in future recharge from the NSW (Littleboy et al. 2015) and Vic (DELWP 2016) state governments for the a near future and b far future, and a comparison to national scale modelling (Crosbie et al. 2013) for the c near future and d far future

Full size image

It was suggested from these studies that the choice of GCM is the largest determinant in the model chain that influences the magnitude of the future diffuse recharge projection (Crosbie et al. 2011). The choice of downscaling method and hydrological model are secondary to the choice of GCMs (Crosbie et al. 2011).

In contrast to diffuse recharge where there are many examples of climate change projections from around the world (and within the MDB), information on changes to localised recharge are far less common. The changes in recharge due to losing streams and overbank flooding are dependent upon changes in river flows rather than climate directly and so require the additional step of modelling river flows before climate change impacts on recharge can be assessed.

One example of predicting changes in localised recharge due to climate change covering a large proportion of the MDB was conducted for the NSW Groundwater Strategy (DPE 2022). This used a single dry future climate scenario to investigate potential risks to groundwater resources rather than a ‘best’ estimate of the future climate (Crosbie et al. 2022b). The sensitivity of the changes in recharge to changes in rainfall can be assessed as the slope of a scatterplot in a similar way that ‘elasticity’ is often used in rainfall-runoff modelling (Fig. 5). In this example, a change in diffuse recharge was predicted to be ~2.7 times the change in rainfall (elasticity = 2.7), but groundwater recharge was less sensitive to a change rainfall than total river flow, which had an elasticity of 4.6 (Fig. 5a). Localised recharge was more sensitive to changes in rainfall than diffuse recharge, with instream recharge to groundwater having an elasticity of 3.3, and flood recharge having an elasticity of 5.7 (Fig. 5b). This highlights the variability in the response of different components of the recharge to a change in climate and indicates that a greater understanding of the mechanisms of recharge (particularly in alluvial aquifers) is necessary for improving projections of recharge under future climates.

Fig. 5
figure 5

A comparison of the change in recharge versus the change in rainfall for different recharge components: a diffuse recharge and total stream flow, b localised recharge consisting of overbank flood recharge and in-stream recharge. This shows the relative sensitivity of different recharge components to changes in rainfall, relative to the total streamflow response. Data adapted from Crosbie et al. (2022b)

Full size image

Knowledge gaps in this space still include quantifying the effects on localised groundwater recharge from changes in river hydrographs, flood frequency and groundwater levels due to changes in climate. There are opportunities to improve understanding of how less frequent but larger floods will change local recharge in rivers, both within the river channel and through overbank flooding. The commonly used method of downscaling GCMs to use as multipliers for historical rainfall and meteorological data does not capture the full extent of potential changes to extreme climatic events—for example, the duration and frequency of drought or extent and return period of large floods. Improving predictions of future rainfall, runoff and recharge that include the impacts of extreme events is a key research growth area.

Surface-water/groundwater interactions and the impact of lowered groundwater levels on rivers

There are two different aspects to connectivity between groundwater and surface water in the MDB. The first describes the direction of flux between the river and groundwater and degree of saturated contact between the water sources. The second relates to the sensitivity of one source to management in the other source, or connectivity factor (Walker et al. 2020a), measured by a median time response of localised recharge to changed surface-water management. High connectivity suggests the need for both groundwater and surface-water sources to be managed as one. There is a need for metrics of connectivity to underpin concepts such as significant or high connectivity that may be used as thresholds for actions. NSW has defined ‘highly connected’ as ‘70% or more of the groundwater extraction volume is derived from streamflow within a single irrigation season’ (DPI Water 2015), but this is not consistent for other states.

Braaten and Gates (2003) developed a simple conceptualisation of the western flowing rivers in NSW, based on monitoring data and the understanding of processes from field studies (Evans and Kellett 1989; Macumber 1991). This consisted of four geomorphological zones:

  • Headwater catchments that are generally gaining streams where groundwater supplies baseflow from fractured rock aquifers

  • Narrow alluvial valleys where the surface water and groundwater are generally connected, the direction of groundwater flow can change seasonally but they are net losing streams

  • Wide alluvial plains in the arid zone where the groundwater level is well below the stream level leading to disconnected status

  • End of valley toward the confluence with the Murray/Barwon/Darling rivers where geological controls often force the groundwater level closer to the surface bringing connected conditions and net gaining of groundwater (and salt) to the surface-water system.

The basin states agreed that with some modification, this conceptualisation could be used across the MDB for addressing both short-term and long-term risks related to connectivity and stream depletion (Parsons-Brinkerhoff 2009; Parsons et al. 2008), although some care is required for generalisation outside of NSW (SKM 2012). To complete the framework for the MDB, a fifth section needs to be added; namely the western Murray/Darling/Barwon rivers themselves. In particular, in the western Murray, there is a general discharge of saline groundwater into the streams, requiring salt interception schemes and management of irrigation to improve river water quality.

Under the Basin Plan, there is a requirement for groundwater management plans that groundwater extraction should not compromise environmental water, including baseflow. While the early perspective was that these objectives should be relatively easy to achieve with environmental water releases, the recent drought in the northern MDB has shown this not to be the case (Stewardson et al. 2021). In the review of the Basin Plan, this led to an assessment that under climate change the ‘vertical’ connectivity between groundwater and surface water is vulnerable with implications for baseflow.

The headwater catchments have had a declining amount of baseflow and declining baseflow as a proportion of total flow (baseflow index) since before the millennium drought (CSIRO and SKM 2010). Since the Millennium Drought, it has been shown that many catchments have not returned to the rainfall-runoff regime that existed prior to this time (Peterson et al. 2021), and as these are generally poor aquifers and are generally not exploited for groundwater, there is some consensus that this is due to lower groundwater levels related to a drying climate (Fowler et al. 2022). Baseflow is a key vulnerability for the northern MDB, particularly in drought conditions (Walker 2022).

The narrow alluvial valleys are the most connected systems where changes in groundwater are most likely to affect the surface-water resources and vice versa (Richardson et al. 2008). Over recent decades, these areas have seen an increase in the prevalence of losing conditions (Crosbie et al. 2023a) and a decrease in the perenniality of the streams, particularly in the northern basin (Crosbie et al. 2022a). After the cap was imposed on surface-water extraction in the mid 90’s (Independent Audit Group 1996) the extraction of groundwater from the narrow alluvial valleys increased (Braaten and Gates 2003) until groundwater entitlements for connected systems were capped as part of the basin planning process (MDBA 2020c). The narrow connected nature of these systems mean that stream depletion due to groundwater extraction is seen relatively quickly compared to the wide alluvial plains (Glover and Balmer 1954).

The wide alluvial plains have deep water tables that are mostly disconnected from the surface water (Crosbie et al. 2023a; Lamontagne et al. 2014). The majority of groundwater extraction in the MDB occurs in these units (MDBA 2020b). The groundwater levels in these areas have been decreasing over recent decades (DPIE 2022; GMW 2022), attributed to increased groundwater extraction and a reduction in recharge (Fu et al. 2022). The poor connectivity of these systems (Walker 2022) means that falling water tables have disproportionately low impact on stream flow given the large extraction. Large losses can still occur from such reaches (Lamontagne et al. 2014), affecting delivery of consumptive and environmental water downstream.

The end of valleys are often subjected to geological constrictions that force groundwater close to the surface and can contribute to gaining river conditions. These are generally arid areas with saline groundwater so there is little extraction of groundwater, other than salt interception schemes along the Murray. With the groundwater close to the surface, almost terminal wetlands have developed with little surface-water flow contributing to the Murray system. Groundwater extraction from these systems can result in a reduction in evapotranspiration and decline in the condition of terrestrial GDEs at the end of valleys (Walker et al. 2020b). The decreased connectivity of pools along streams can be somewhat mitigated through environmental watering in regulated systems, which occurred in 2018 and 2019 for over 1,000 km of the Barwon-Darling system (Eco Logical Australia 2020; MDBA 2018). Replacing the source of water for the terrestrial GDEs at the end of valleys requires engineered flooding (Bond et al. 2014).

Integrated water management plans exist for a small proportion of the MDB, and accounts for a small proportion of extraction. The nature of these varies with linked surface- water/groundwater determinations on regulated highly connected NSW streams; cease-to-pump rules common to surface and groundwater in unregulated perennial streams in NSW and Victoria; protection of baseflow in the Adelaide Fold Belt and protection of spring flow in Queensland. Outside of these plans, the expectation is that the SDL, together with annual allocations and local management rules should avoid adverse impacts on stream flow. This assumption allows groundwater and surface-water plans to be mostly developed in isolation.

Historical increases in extraction have led to about 17% reduction in the median stream availability in the Namoi and about 6% in the Lachlan rivers (Walker 2022). Rassam et al. (2017) have shown how a reduction of this magnitude can affect low flows for the regulated Namoi. These results suggest that historical groundwater extraction has affected the baseflow required for some ecosystems across some regulated river valleys, while the impact may be more local in others. There are many knowledge gaps around how to manage rivers and groundwater to protect baseflows for ecosystems. Gaps include better definition and prioritisation of ecosystem baseflow requirements, managing well-connected groundwater systems with long response times and identifying reaches of major losses for environmental water releases.

Although groundwater extraction can negatively impact surface-water flows, there is also the potential to use groundwater more efficiently to minimise impacts on surface water. Opportunities include conjunctive use of surface-water and groundwater, surface-water storages and managed aquifer recharge to support adequate baseflow in vulnerable stream reaches or for GDEs (see section ‘Summary of challenges and opportunities’). Another opportunity is through better digital and data-based monitoring and assessment of groundwater levels. Supervised and unsupervised deep learning methods have been used to predict groundwater level time series in the Namoi region (Clark 2022; Clark et al. 2022). These methods were found to be useful for modelling current groundwater levels and while the prediction of future groundwater levels was not straightforward, the methods were successfully used for assessing the ‘what if’ scenarios with different groundwater extraction regimes.

Groundwater-dependent ecosystems: prioritisation, water requirements and thresholds

Over the past two decades, definition, detection and research related to GDEs has advanced, particularly in Australia (Barron et al. 2014; Boulton 2005; Eamus and Froend 2006; Gow et al. 2016; O’Grady et al. 2006) and more recently in other countries (Erostate et al. 2020; Gou et al. 2015; Howard and Merrifield 2010; Xu et al. 2022). As defined by Richardson et al. (2011), GDEs are ‘natural ecosystems that require access to groundwater to meet all or some of their water requirements on a permanent or intermittent basis, so as to maintain their communities of plants and animals, ecosystem processes and ecosystem services’. GDEs include springs, which demonstrate a surface expression of groundwater, cave and aquifer aquatic ecosystems and riparian ecosystems, where groundwater contributes to the hydrological environment and estuarine and marine ecosystems relying on submarine discharge of groundwater (Doody et al. 2017; Eamus et al. 2006; Golder Associates Pty Ltd 2021; Hatton and Evans 1998; Richardson et al. 2011).

Significant investment and effort have been directed to mapping GDEs nationally, improving spatial resolution (Castellazzi et al. 2019; Fildes et al. 2023; Rojas et al. 2023) to advance upon methods used in the “National Atlas of Groundwater Dependent Ecosystems” (Doody et al. (2017); herein ‘The GDE Atlas’). However, less investment has been directed to establishing the water requirements of GDEs, specifically using in-situ-data collection which are vital to underpin regional scaling methods. Unexpected and/or prolonged disconnection of groundwater from GDEs such as vegetation or adverse changes in groundwater quality (impacting stygofauna or river-bed dwelling fish for example), lead to deterioration of GDEs (Eamus and Froend 2006; Eamus et al. 2006; Sommer and Froend 2011). Quantified timing and duration of when GDEs access groundwater to supplement their water requirements and volumes of extraction for vegetation, remain relatively unknown (Cook et al. 2022; Crosbie et al. 2023b), aside from detailed local studies (Benyon et al. 2006; O’Grady et al. 2006; Thorburn and Walker 1994). Hence, a lack of robustly identified critical management thresholds for groundwater change continues to hamper the prioritisation and protection of GDEs. Without field validated quantitative estimates of GDE water requirements, challenges associated with their protection in the coming years under increasing climate change will continue to escalate.

Murray-Darling Basin groundwater-dependent ecosystems

In the context of the Murray-Darling Basin, GDEs and their importance receives little direct attention. The governing water management framework for the basin, the Murray-Darling Basin Plan (MDBA 2010) and associated Basin Wide Environmental Watering Strategy (MDBA 2019a), aim to sustainably manage water to protect ‘water-dependent ecosystems’/environmental assets and support industries (MDBA 2010). The definition of water-dependent ecosystems is much broader than that of GDEs, including ‘ecosystems that depend on periodic or sustained inundation, waterlogging or significant inputs of surface water or groundwater to continue functioning’ (NWC 2014).

Key objectives of the Basin Plan are to protect and restore water-dependent ecosystems as well as their ecosystem functions to ensure their resilience to climate change and other threats (MDBA 2010), incorporating water specifically allocated for the environment or ‘environmental flows’. However, environmental flows can only be delivered within the constraints of a highly managed hydrological system and flow on prohibitive costs of elevating water higher and further in the landscape than the river channel. This leads to extended periods of wetland and floodplain disconnection from the river channels, especially across elevated floodplain regions, resulting in long-term ecosystem degradation in the absence of at least average rainfall and reduction of biodiversity with continuing disconnection of wetlands and other peripheral water bodies (Cunningham et al. 2018; Doody et al. 2021; Moxham et al. 2018).

To date, at the basin scale, there has been greater emphasis on surface-water management to support water-dependent ecosystems ‘in areas that can be influenced with environmental water’ (MDBA 2019a), although limits on groundwater extraction have been set basin wide (Hart 2016; Walker et al. 2020b). Hence, far less monitoring and research has been undertaken to specifically identify GDEs and investigate risks to prevent future degradation, particularly to reduced surface–groundwater connectivity related to persistent drought (Ross et al. 2022) and in areas beyond the management zone of environmental flow delivery due to water delivery constraints. More generally, GDEs in the basin have been identified in relation to potential impacts of gas and coal mining developments (Baird and Burgin 2016; Doody et al. 2019; Jones et al. 2020) and at the state level where GDE-related information is required to meet water sharing and water resource plan legislative requirements (McDougall et al. 2017; Rohde et al. 2017).

Terrestrial groundwater-dependent ecosystems in the Murray-Darling Basin

Strong field-derived evidence demonstrates that the woody vegetation estate of the basin, composed of Eucalyptus camaldulensis (River Red Gum; 360,000 ha), E. largiflorens (Black Box; 409,000 ha), E. coolabah (Coolabah; 310,000 ha) and lesser-known species Acacia stenophylla (River Cooba), all rely intermittently on groundwater (Costelloe et al. 2008; Doody et al. 2009; Holland et al. 2006). While it may seem counterintuitive, it has been clearly shown that species such as E. largiflorens possess adaptations such as salt exclusion at the root interface, to enable extraction of highly saline groundwater (>40 dS m−1), providing a critical water resource, especially during prolonged drought periods (Doody et al. 2009; Holland et al. 2006). In some areas, E. camaldunesis along creeks and rivers predominantly use groundwater and baseflow-fed water sources to supplement their water resource requirements (Mensforth et al. 1994; Thorburn and Walker 1993). In a regional-scale assessment, Rojas et al. (2023) used modified versions of Shannon and Simpson diversity indices to add granularity to the GDE assessment and prioritize alluvial aquifers in the MDB, but generally, beyond the large polygon mapping provided by the national-scale GDE Atlas, a fine-scale map of GDE locations across the entire MDB is limited. Likewise, an understanding of temporal groundwater requirements is scarce. Given this paucity of information, an understanding of the groundwater requirements of woody vegetation across the basin is a significant knowledge gap to protecting these extensive ecosystems (Cook et al. 2022; Ross et al. 2022).

Based on potential GDE mapping across the basin from the GDE Atlas, it is likely that around 36% of the basin extent is occupied by terrestrial GDEs. Of these, 9% are mapped as high potential of being groundwater dependent and 2% are mapped as moderate (data not shown). Terrestrial GDEs of highest concern are those located outside the zone of influence of environmental water. As demonstrated by Doody et al. (2023) multidecadal disconnection of E. largiflorens from surface water, below average rainfall and location in salinized soils and highly saline groundwater, leads to a decline in the trajectory of tree biomass and ecosystem resilience, to eventually culminate in altered ecosystem states and vegetation community transitions (Fuller et al. 2019; Scheffer et al. 2001; Yin et al. 2022). As E. largiflorens are a widespread floodplain species that are more drought and salt tolerant than E. camaldulensis (Roberts and Marston 2011; Rogers and Ralph 2010), these communities are largely situated outside of the influence of environmental flows (Doody et al. 2021). They rely on groundwater to supplement their water requirements and are thus GDEs at high risk of degradation with increasing climate change and further reductions in water availability through drought and competing water resource requirements.

A first step toward managing GDEs in the MDB is to prioritise their fine spatial- and temporal-scale mapping, especially terrestrial GDEs. A risk assessment in relation to current water management practices and environmental constraints would highlight areas on declining trajectories in relation to recent and decadal flood inundation patterns and other climatic factors (Doody et al. 2023). This would indicate GDEs that are at high risk of transition to dryland or saline ecosystems. With management intervention these transitions could be avoided, or at the very least, high-value ecosystems could be prioritised for conservation using management interventions. Some advances in prioritisation of terrestrial groundwater-dependent vegetation communities in NSW from an ecological perspective (Dabovic et al. 2019), but an integrated and consultative approach is also required to identify ecosystems with high cultural and social values.

While determining water requirements of GDEs poses challenges, there are approaches and conceptualisations that can provide estimates of likely groundwater requirements (Crosbie et al. 2023b; O’Grady et al. 2011), and opportunities for finer scale mapping of vegetation canopy and condition responses to changes in groundwater using remote sensing products (Gao et al. 2021). However, the importance of collecting robust field data such as that presented in Benyon et al. (2006) and Doody et al. (2023) is vital to underpin any modelling and remote sensing approaches in the future. With the continuing threat of reduced water availability via climate change, predicting future threats to GDEs in the MDB and the monitoring of flow outcomes can assist with planning for limited and critical water resources (Gawne et al. 2020).

Challenges and opportunities for groundwater management

Managing groundwater in a changing climate

The current state of knowledge of the potential effects on groundwater from a changing climate indicates that adaptation will be needed to manage changes particularly in relation to maintaining and recovering groundwater levels or pressure, accounting for impacts on surface water, and incorporating groundwater as an adaptation option for water resources management.

Maintaining and recovering groundwater levels will be needed in priority areas to minimise the effects of climate shifts, including longer or more frequent droughts, on connected surface water and GDEs. Management options may include further reductions in water use, managing or supplementing aquifer recharge during wetter periods and local management of water levels including protection of GDEs and baseflow, and condition-based management including trigger levels. In regions where groundwater is highly connected, policy frameworks to further integrate groundwater and surface-water management would enable the management of groundwater and surface water as one resource and account for interactions between the two.

Groundwater is an adaptation option to current water resources management in the Murray-Darling Basin to support climate resilience, where climate resilience is defined as the ability to manage surface and groundwater resources within a safe operating space of water level and flow objectives, which contributes to the ability of ecosystems and communities to cope with extreme events or disturbances. These adaptation options include groundwater as an additional water source in areas where the resource is not yet fully developed, or where it is, developing the potential for treatment and use of brackish water sources (Barron et al. 2015). Aquifers could also be used as an additional storage option through water banking and managed recharge.

Adaptive and active groundwater management are a key part of groundwater management in the face of climate variability. Adaptive groundwater management has been defined as a flexible and responsive approach that uses targeted monitoring to improve management practices in an iterative way (Thomann et al. 2020). This includes reviewing of management plans as further information becomes available, using groundwater level thresholds to inform management and engage stakeholders in decision-making. Active groundwater management involves a more prescriptive and proactive approach to groundwater management, including the coordination of operations, managing groundwater storage and levels or using numerical models or digital twins, including socioeconomic functionality, to optimise recharge and extraction.

Adaptive management

Uncertainty pervades almost every area of hydrogeology in the Murray-Darling Basin, particularly subsurface hydrogeological properties and future information on climatic conditions and extraction patterns in response to markets, technology and climate change. While data and knowledge are increasing in many areas and domains, there is a need to manage groundwater in the context of this imperfect knowledge. Risk management is a structured approach aimed at achieving management objectives despite of uncertainties. The incorporation of risk management has been encouraged within water resource management plans in Australia to address uncertainty (Australian Government 2010). Adaptive management is one component of risk management which includes plans being continuously reviewed as new information becomes available.

The Basin Plan is underpinned by an adaptive management framework (MDBA 2017). This includes (1) setting clear objectives; (2) linking knowledge, management, evaluation and feedback over a period of time; (3) identifying and testing uncertainties; (4) using management as a tool to learn about the relevant system and change its management; (5) improving knowledge; and (6) having regard to the social, economic and technical aspects of management. In terms of groundwater management plans, management objectives are defined by resource condition indicators (metrics), such as groundwater levels, and acceptability thresholds (Anderson et al. 2014a, b). Even if plans are compliant with the SDL and local management rules, uncertainty may mean that the groundwater system evolves to being outside of desired range and adaptive management may be required to return the system to a state more aligned with objectives.

Adaptive management has been applied in the MDB in different ways. Water level response management is a process where the seasonal or annual groundwater allocation depends on the water level. Lower levels may mean a decrease in allocation, helping to maintain water levels within the desired range. The main examples of this are from Victoria (GMW 2022; Walker 2023). Almost all plans have a set of contingent thresholds, the breaching of which indicates that actions need to be taken, for example groundwater level thresholds in NSW (DPIE 2022; Walker 2023). For alluvial aquifers that are highly connected to perennial streams, cease-to-pump rules may be invoked if the streamflow falls below a defined level (GMW 2012; NOW 2015). For long-term processes such as climate change, a series of incremental adaptive changes may be required (MDBA 2019b). The time lags for groundwater responses to a change in management will affect the ability to use adaptive management (Thomann et al. 2020). Two recent studies have shown the need for ongoing review of groundwater resource management plans for their efficacy and resilience (Thomann et al. 2022; White et al. 2016) and for guidelines to be developed (Thomann et al. 2022).

In the context of long-term change, adaptive management can usefully be supported by identifying adaptive pathways describing sequences of options, indicators that need to be monitored, and circumstances in which action might need to be taken (Haasnoot et al. 2013; Marchau et al. 2019). This can help identify robust short-term actions, guide evaluation of the value of new information, and anticipate barriers to change. It can also help build a shared understanding of what decisions may be difficult to change in the future and possible irreversible impacts (Thomann et al. 2022).

Active management

Additional opportunities are apparent if one also considers coordination of operations, i.e., aim to actively manage the aquifer as a storage. This has primarily been discussed in the context of managed aquifer recharge (MAR) through the concept of “water banking” (Gonzalez et al. 2020; Page et al. 2022). MAR schemes supplement natural recharge, addressing imbalances between desired surface and groundwater storages that cannot be achieved through conjunctive use or to compensate for limitations of existing management. This can allow full utilisation of existing water allocations, support resilience in times of drought, or provide a tool to improve reliability, flexibility and efficiency of existing water delivery systems (Harvey et al. 2023). Page et al. (2022) concluded that the current Basin Plan supports water banking, but that existing water accounting systems would need to be adjusted and there is potential for demonstration sites to facilitate policy development while providing critical experience to support learning and wider adoption. Replacing fresh groundwater supplies with more climate-resilient sources such as desalinated brackish groundwater or recycled water, may become more economically viable with increasing effects of climate change and water scarcity (Walker et al. 2021b).

More generally, groundwater management has historically been considered fragmented in parts of Australia (Cook et al. 2022). Other opportunities for more coordinated active management include restoring, protecting and monitoring natural recharge mechanisms (Norman et al. 2022) and facilitating local coordination of conjunctive use of surface and groundwater (Ticehurst and Curtis 2019). Groundwater extraction has tended to increase in dry periods (MDBA 2020d) and has the potential to increase, given it is still below the SDL in many regions, although annual allocations are limited by seasonal conditions. There is therefore an opportunity to better understand groundwater use in regions historically served by surface-water irrigation networks and opportunities for active management of water sources to improve system outcomes.

With substantial investment in modernisation of irrigation areas and development of remote sensing, there is now more frequent and higher resolution information available on where water is in a landscape, including telemetered channel flows and water levels, timing of water delivery to farms, and soil moisture levels (Perera et al. 2016). This has the potential to help close the water balance both by enabling new analyses, e.g. quantifying seepage (Moavenshahidi et al. 2016) and by helping build living models or digital twins (Purcell et al. 2023) of systems that can then be used to actively track and plan pumping and recharge. Together with other information from diverse information systems (Sharples et al. 2020) including from impact assessments (WABSI 2019, 2021), these are sources of data that have historically not been available to groundwater models used for water plan development, and therefore open up new foundations on which to build supporting policy. There are already calls internationally to develop capacity for integrated storage management across the landscape, including groundwater and soil moisture (Burke et al. 2023).

Finer control over water flows and storage in the landscape raises the importance of agreeing on both local- and basin-scale objectives (Robinson et al. 2015), including for water levels and use. In addition to triggers related to impact, one might envisage targets for seasonal drawdown and recharge (Khan et al. 2008). By analogy with flood management for dams, it may be beneficial to manage the volume of headspace within the aquifer, informed by seasonal forecasts. In addition to water allocations, setting long-term minimum supply level objectives may also support resilience of a community, accompanied by prioritising certain groundwater uses in times of drought (Harvey et al. 2023). Building on examples of local managed aquifer recharge or water banking schemes (Vanderzalm et al. 2020), it seems likely that this type of active management will be run by local agencies, potentially with greater local community involvement. While this may require policy change in the long run, pilots are already possible through impact assessment processes.

The future vision for groundwater management in the MDB already goes beyond SDLs, with the “Statement of expectations for managing groundwater” (MDBA 2019b) emphasising the need to address localised risks, monitoring of impacts, and also collaboration outside the water resource planning context. Groundwater markets provide a mechanism for reallocation of groundwater (Wheeler et al. 2021). The MDB also has a maturing ecosystem of guidelines, expressing shared expectations of what good practice looks like, including those from specific jurisdictions, agencies and professional organisations, as well as developed through shared bodies such as the National Water Reform Committee and the Water Monitoring Standardisation Technical Committee (WaMStec). Compared to other regions, there are therefore significant mechanisms to transition to and adopt new governance arrangements, even if improvement is always possible (Chipperfield and Alexandra 2022).

Improving water literacy

It is generally recognised that scientific monitoring and modelling play a key role in groundwater management; however, decisions are only ever informed by science, not made by science. An effective groundwater management framework requires an effective and well-integrated knowledge system, within which new data and scientific understanding are effectively shared with decision-makers at multiple scales, local knowledge is valued, and there is a clear understanding of how new information can inform future decisions.

Adaptive groundwater management and active management of aquifers depends on effective continuous learning about the systems being managed across scales and sectors (Refsgaard et al. 2010; Thomann et al. 2020). To date, updating of groundwater system knowledge has generally been through periodic reviews which require significant effort to discover grey literature and analyse new data. A future vision for knowledge sharing involves frameworks and practices that allow for regular (annual or subannual) updating of groundwater system knowledge based on new data, knowledge repositories as public tools that are open to scrutiny and improvement and readily accessible, and the ability to incorporate understanding from a wide range of sources, including regulator, industry and private data and impact assessments (WABSI 2019, 2021). Digital systems architecture and human-centred design has a key role in integrating groundwater systems knowledge (Box and Lemon 2015; Vilas et al. 2020).

Building on existing systems (Sharples et al. 2020), sharing data across scales would be supported by frameworks for landholders to share data and observations of groundwater levels and groundwater system behaviour such as responses of surface water and GDEs to changes in groundwater. This has the potential to build on citizen science efforts (Nath and Kirschke 2023; Walker et al. 2021a), especially related to mining and energy developments (Jamieson et al. 2020) where there is a large amount of currently unavailable knowledge and data. Quality assurance, fair payment for information sharing, fair sharing of the costs and benefits of new information, and the development of protocols for minimum standards for data sharing would need to be part of this process.

In addition to sharing data, continuous learning requires tools to help share knowledge and incorporate new information into decision-making, including Indigenous cultural and local knowledge. As with other environmental sciences strongly influenced by context and place (Gerlak et al. 2018), groundwater faces the difficulty of progressing both place-based knowledge and fundamental advances. Model catalogues and repositories provide a starting point by supporting model management (Arnold et al. 2020). They can publicly document existing tools, including sufficient information to judge their scope of applicability, covering multiple versions over time, and managing access to both numerical and analytical flow and transport models. When a model has regular operational use, as in the case of active management described previously, it can be treated as a living model that is “continuously updated with new data, new geologic interpretations, and new model calibrations” (Refsgaard et al. 2010). More is needed, however, to be able to provide integrated communication pathways to share knowledge with groundwater users and interested community members. The effectiveness of such tools is ultimately also dependent on processes and models to help communities articulate and track what it is they value about their groundwater system.

Underpinning all of this is the need for shared vocabularies around groundwater system knowledge and management and shared understanding of the different values of groundwater by a diverse range of stakeholders. Although inclusion of Aboriginal cultural values into water planning has generally been slower than for landscape management, examples of engaging Aboriginal communities in water planning within New South Wales are described in Moggridge et al. (2019). Opportunities for stakeholders to meet together to hear varied perspectives on water values, challenges and potential solutions is a first step. Including social and cultural values of groundwater in water governance, management and modelling can also provide more nuanced understanding required for water management (Castilla-Rho et al. 2019; Heinrichs and Rojas 2022).

Summary of challenges and opportunities

The Murray-Darling Basin is a highly allocated and regulated basin, predominantly located in semiarid to arid parts of south-eastern Australia. There is already evidence of a changing climate in rainfall patterns and in a decline in groundwater levels in many parts of the MDB, and future climate predictions are for a drying climate, particularly in the southern part of the basin, with greater wet and dry climate extremes and an expansion of arid and semiarid climate zones.

Predictive modelling has indicated a likely reduction in groundwater recharge in the near future, particularly diffuse recharge from rainfall and local recharge from stream losses, accelerating the decline in groundwater levels. The long-term changes to local groundwater recharge from overbank flooding is less clear but is likely to be influenced by longer and more frequent droughts and longer periods of time between floods. Developing knowledge around the effects of extreme climate conditions such as longer, more intense or frequent droughts or floods on groundwater recharge, particularly local recharge, is a knowledge gap in this area.

In turn, lowered groundwater levels are likely to result in higher stream losses from surface water to groundwater, and increased proportions of river channels either with reduced groundwater baseflow or rivers changing from gaining to losing systems. Reduced baseflow to rivers and streams has implications for instream and riverine ecosystems, as baseflow from groundwater often provides water to support these systems through dry periods and droughts. To date there has been a management disconnect in the Murray-Darling Basin Plan between groundwater resources and other components of the water balance. The magnitude of effects from declining groundwater levels on surface-water/groundwater connections, increased river losses and reduced baseflow have not yet been quantified with sufficient precision, and all have implications for surface-water accounting and regulation.

The ecological response of terrestrial GDEs to declining levels and a changing climate is also complex and not relatively well understood. A lack of robustly quantified critical management thresholds for groundwater change continues to hamper the prioritisation and protection of GDEs. In the Murray-Darling Basin Plan, terrestrial GDEs receive very little attention given the high social and cultural value that they provide to the region.

The upcoming review of the Murray-Darling Basin Plan in 2026 gives an opportunity to reassess groundwater management using a more integrated approach, including groundwater, surface water and ecological systems. Many regions may be able to respond to uncertain climatic conditions through continual application of adaptive management. In other areas, climate shifts may push groundwater management beyond adaptive measures, requiring more significant changes to the management framework. Active management, including the integrated coordination of operations for groundwater and surface-water resources and infrastructure development for MAR and water banking is showing promise for improving the ability of ecosystems and communities to respond to climate extremes.

Managing water to support climate resilience of ecosystems and communities requires better definition of what constitutes resilience at a local or regional scale and agreement by stakeholders involved. Shared vocabularies around groundwater management, shared western and traditional knowledge, and a shared understanding of the different values of water by a diverse range of stakeholder groups are needed for this to occur. Facilitation of constructive dialogue working towards mutual agreement requires open and low-cost distribution of groundwater knowledge and data related to water and management, presented in such a way that makes it accessible and engaging to a broad range of stakeholders and decision makers.

Despite a changing climate, and potential increasing aridity and water scarcity, there are a number of key opportunities for development of knowledge, policy and management in the MDB, including:

  • Improving data acquisition to support system understanding and management through automated digital monitoring of groundwater, using novel data sources, and related ground truthing, to observe groundwater and ecosystem trends at a regional scale

  • Improving capability in predictive modelling of integrated groundwater, surface-water and GDE management through risk-based frameworks, living models or digital twins

  • Improving quantification of water requirements and tolerance thresholds of GDEs to changes in groundwater levels and the extent and timing of surface-water inundation, to develop understanding of groundwater and GDE vulnerability to climate change, in particular the impacts of extreme climate events such as droughts and floods

  • Developing agreed metrics to define thresholds for management of connected groundwater–surface-water systems and GDEs and prioritisation to conserve those of highest ecological, hydrological, cultural and social value through active groundwater management and environmental watering

  • Developing policy changes required to support active management of groundwater in combination with surface-water management and operations, including conjunctive use, managed aquifer recharge and water banking

  • Continual improvement in communication of groundwater information and knowledge sharing to decision makers at levels, including understanding community values of groundwater and GDEs.

The current focus on the Murray-Darling Basin in social and political discussions leading up to the 2026 review of the Basin Plan makes it timely to pursue and develop these opportunities to ensure effective management under future climate conditions.