Repairing Australia’s inland river and groundwater systems: nine priority actions, benefits and the finance gap

Action 1. Riparian revegetation

Revegetation of degraded and cleared riparian zones is widely recognised as a critical action for restoring river and catchment health, with significant benefits for biodiversity and people owing to the many essential ecological functions and ecosystem services provided (Mohan et al. 2022). Riparian vegetation conservation and restoration is also a priority for effective climate-change adaptation, for example, by shading and cooling aquatic ecosystems, providing corridors for wildlife movement, and acting as a buffer to both terrestrial and aquatic systems from a range of extreme events (Capon et al. 2013). It also contributes significant habitat to rivers and streams, for example, contribution of large wood or ‘snags’, supports important cultural values, for example, remnant sacred living trees, and regulates carbon inputs which are essential for the effective functioning of river systems.

To cost riparian revegetation requirements at a national scale, we generated a continental riparian buffer-zone layer by using Geoscience Australia watercourse data (Crossman and Li 2015) with an Australian Albers Geocentric Datum of Australia 2020 projection. We assumed a 100-m-wide buffer zone for watercourses classified as major rivers, 50 m for the additional zone of minor rivers, and 200 m for the additional zone of lakes, based on suggested widths provided in Hansen et al. (2010). In each case, only that area additional to the previously calculated buffer area was included to avoid double counting.

Following Mappin et al. (2022), we then used the Vegetation Assets, States and Transitions (VAST) dataset (Thackway and Lesslie 2006; Lesslie et al. 2010), to determine the area within these buffer zones assigned to a range of condition and land-use categories. The VAST framework provides a continental classification, at 1-km² resolution, of land types in relation to their degree of modification from a pre-European state on the basis of vegetation cover and land use. VAST Category 2, for example, encompasses modified areas with native vegetation communities that have been subject to minimal use or production from relatively natural environments, such as timber harvesting or grazing, whereas VAST Category 3 pertains to areas that have been transformed by land uses, for example, through significant tree thinning for pasture production, but still maintain native vegetation communities with the capacity to regenerate (Lesslie et al. 2010). Areas assigned to VAST Category 5 are those where native vegetation has been cleared and replaced with cultivated vegetation. VAST 0 (residual bare), 1 (residual) and 2 (modified) are considered to be in good ecological condition and thus not requiring significant revegetation efforts. We assume that VAST 3 (transformed), 4 (largely replaced and degraded) and 5 (replaced – managed) can be restored, whereas areas classified as VAST 6 (replaced with human-made structures) are considered to be unrestorable.

The ABARES National Land Use Mapping (NLUM) raster dataset, at 250 m, identifies secondary thematic land uses across agricultural and non-agricultural lands for 2015–16. Costs per hectare (ha) were estimated based on a hypothetical large-scale (1000 ha) revegetation project on an abandoned post-agricultural site, by using a density of 2000 native plants per hectare, including 30 species of groundcover, shrub and trees (Andres et al. 2024). We assumed that cleared riparian areas (VAST Category 5) would require active revegetation to be repaired to a good, albeit modified, ecological condition (VAST Category 2). For cleared riparian areas in cropping, forestry, mining and conservation land (NLUM Categories 1.1, 1.2, 1.3, 2.2, 3.1, 3.3, 3.4, 3.5, 4.1, 4.3, 4.4, 4.5, 5.1, 5.8, 6.3, 6.5), we assumed a mean cost for active revegetation of A$5787 ha⁻¹ (values were converted to 2022 Australian dollar values) for tube stock, labour and travel (costs supplied by S. Andres from Andres et al. 2024). On grazing land (NLUM Categories 2.1, 3.2, 4.2, 5.2), we assumed such active restoration would cost twice this amount, an average of A$11,574 ha⁻¹ (2022 dollar value), including costs of tube stock, labour, travel, fencing and provision of off-site watering points for livestock. For degraded areas (VAST 3), we assumed, on the basis of recent studies (e.g. Zivec et al. 2023), that passive revegetation would be sufficient to achieve a VAST Category 2, estimating costs to be half of those for active restoration for each land-use category. Riparian lands in residential, industrial or infrastructure areas were excluded. On the basis of the method used, the mean cost per hectare of repairing riparian buffer zones across Australia was A$5909 for major rivers, A$6411 for minor rivers and A$5673 for lakes (2022 dollar values; Table 1).

While we recognise that streambank engineering works may also be required in some instances, costing these was beyond the scope of this assessment. However, we did consider ongoing operational costs associated with maintenance and monitoring, including replacing tube stock where initial plantings do not survive (Table 1). We assumed a cost of A$4.7 ha⁻¹ (2018 dollar value) to manage native vegetation plantings for weeds, pests, flooding, and fire on the basis of the annual marginal per-hectare cost spent by Australian Wildlife Conservancy (2016), Bush Heritage Australia (2016), the NSW and Queensland governments (as reported by Adams et al. 2011), and other State Governments, as reported by the Legislative Council of Tasmania (2012), for non-riparian native vegetation. We further assumed, as per Mappin et al. (2022), that revegetation actions would be understaken over a 30-year period, meaning that the area requiring management and monitoring will grow accordingly over this timeframe (refer to the Supplementary material for calculations).

We estimated the portion of costs that could be potentially financed through carbon markets, such as the Australian carbon credit scheme. We calculated carbon sequestration potential by using revised estimates of maximum potential biomass (MaxBio) across Australia (Roxburgh et al. 2019). MaxBio, a 250-m spatial raster dataset, represented the conservative maximum upper limit to above-ground biomass accumulation for any location attainable from native vegetation that has achieved a stable, mature state of growth. Areas designated within the protected-area estate, defined as ‘1.1 Nature Conservation’ in the Australian Land Use and Management Classification Version 8 (2015–16) dataset (Australian Bureau of Agricultural and Resource Economics and Sciences 2024), were removed from our revenue calculations as these areas are currently not eligible with respect to delivering Australia Carbon Credit Units under the Carbon Credits (Carbon Farming Initiative) Act 2011.

The MaxBio layer was converted to vectors, then values were multiplied by area of each polygon in the riparian buffer zones. Values were summed to derive dry-matter tonnes of carbon for major rivers, minor rivers and lakes, by VAST and NLUM classes. Values were halved under the assumption that 50% of the dry biomass is elemental carbon, the same value applied in the ‘Carbon Farming Initiative – Reforestation and Afforestation 2.0 Methodology Determination 2015’ (Federal Register of Legislation 2015). We assumed that the passive restoration areas had 25% less potential carbon sequestration than the value identified by MaxBio, owing to the assumption that 25% native vegetation already exists in those areas, as per Mappin et al. (2022). Values of elemental carbon were multiplied by 3.67 to convert from tonnes of carbon (Mg C) to tonnes of carbon dioxide equivalent (Mg CO₂e). We divided this by 25 to get the per-year abatement over the 25-year post-restoration period.

It is assumed that a strong ‘safeguard mechanism’ will create sufficient demand for carbon credits over time. We modelled the following two carbon-price scenarios annually from 2025 to 2054: Scenario (1), an estimate based on extrapolating the current spot price of carbon increasing by a fixed percentage, and Scenario (2), an estimate based on extrapolating the Clean Energy Regulator’s cost containment measures, which reflect current policy regarding the maximum compliance costs faced by facilities under the safeguard mechanism (Clean Energy Regulator 2024). Both scenarios are reasonably aligned with BloombergNEF’s cover of the period (BloombergNEF 2024). The scenarios are conservative compared with the interim values of emissions reductions used by the Australian Energy Market Operator from 2025 to 2050 (see appendix A.11 in Australian Energy Market Commission 2024). It is assumed that the Australian carbon credit units (ACCUs) available on the market are high-integrity, following the Federal Government’s commitment to implement the recommendations of the Independent Review of ACCUs (Chubb et al. 2022) and of the Climate Change Authority’s review of the ACCU scheme (Climate Change Authority 2023).

Indigenous organisations could play a key role in implementing this action, with significant social and environmental benefits to be gained by restoring Country and generating jobs through Indigenous businesses and ranger organisations. Services required may include construction of off-river livestock watering points, fencing, erosion control, plant propagation and planting.

Action 2. Incentivise landholders to retire riparian farmland

To further support riparian revegetation and riverbank health, we estimated the cost of incentivising landholders to voluntarily retire farmland from active production, with the exception of low-intensity grazing, from within the privately owned portion of the national riparian buffer. We calculated farm-cash income for broadacre industries within the riparian buffer zones identified in Action 1, paid out over 30 years (Australian Bureau of Agricultural and Resource Economics and Sciences 2022), to estimate the expected cost of forgoing cropping, forestry or grazing production on these lands (Table 1). Average annual farm-cash income estimates were derived from an Australia-wide, 1-km² resolution spatial layer for the 5-year period from 2016 to 2020 (assumed in 2022 Australian dollars; Australian Bureau of Agricultural and Resource Economics and Sciences 2022). We propose that such an action might initially be delivered via voluntary uptake but recognise that, should this be lower than required, may require increased incentives or mandatory measures. Voluntarily retired riparian farmland may, in many cases, be transferred to Indigenous custodianship for restoration.

Action 3. Restore overallocated river systems of the Murray–Darling Basin to sustainable levels of take

Multiple river systems across Australia, including six in the Murray–Darling Basin (MDB), remain at a high risk of water overallocation (National Water Commission 2014). With climate change already threatening water security in such areas (Prosser et al. 2021), returning these river systems to sustainable levels of take is a high priority to support the maintenance of critical ecological functions (e.g. safe drinking-water provision) and reduce risks to irrigated agriculture from salinity.

The best publicly available estimate for sustainable extraction levels in the MDB requires recovery of between 3,856,000 and 6,983,000 ML (Murray–Darling Basin Authority 2010). Given current water recovery targets presented in the Basin Plan of 3,200,000 ML (Hart 2016), a further 726,000 ML is still required to achieve the minimum target estimated under high uncertainty for sustainable environmental outcomes of 3,856,000 ML. We did not cost the volume of water that governments have committed to recovering to achieve the 3,200,000-ML target. On the basis of the average cost of water recovery by using infrastructure upgrades of A$5100 ML⁻¹ (2017 dollar values) and purchase of entitlements (A$2200 ML⁻¹ (2018 dollar values) (Wentworth Group of Concerned Scientists 2017), we assumed a cost of A$5100 ML⁻¹ (2017 dollar values) (Table 1). Data are not available to extend this assessment beyond the MDB.

We further identified that the Water Act 2007 provides limited opportunity for the delivery of cultural water (also termed Cultural Flows), aiming to re-establish events that specifically deliver cultural benefits. For instance, the Water Act 2007 (s50 4A. a and b), through recent amendments, requests the Murray–Darling Basin Authority (MDBA) to report on matters relevant to Indigenous people in relation to the management of Basin water resources and recognise and protect the interests of and support opportunities for Indigenous people. The federal Department of Climate Change, Environment, Energy and Water (DCCEEW) Secretary also has to report annually on engagement and the above-mentioned MDBA reports (Water Act 2027, S85F 1). Further, through policy and commitments, DCCEEW provides funding for cultural flow plans to be developed, an Aboriginal Water Entitlements Program and the ongoing participation of Indigenous people through a designated body Committee on Aboriginal Water Interest (CAWI). We propose there is an opportunity to ensure that any future recovery and delivery of water is genuinely designed and implemented for cultural benefit by Indigenous people, for the protection of water. There are too few examples of this taking place in contemporary water management frameworks.

Action 4. Restore lateral connectivity between rivers and their floodplains and wetlands through constraints management

Numerous physical and operational constraints currently impede environmental water delivery within river systems, reducing connectivity both longitudinally along rivers and laterally between rivers and their floodplains and wetlands. By limiting the flow of water, these constraints reduce the extent and quality of ecological responses to flows and inundation of low-lying floodplains and wetlands, resulting, for example, in smaller native fish populations and degraded health of river red gum forests and woodlands (New South Wales Department of Planning and Environment 2023).

Removing constraints to environmental flow delivery can involve a range of options, including the removal or modification of physical infrastructure such as bridges, roads and flood works, or purchases of voluntary easements on private land. State government business cases for constraint management in five priority areas of the MDB (Hume to Yarrawonga, Yarrawonga to Wakool Junction, the Goulburn, the Murrumbidgee and Menindee Lakes (Lower Darling) and the Lower Murray in South Australia) were assessed by Kahan et al. (2021). We have assumed the total cost of these projects, minus A$200 × 10⁶ already committed to constraints management by the Australian Government (Murdoch, 2020), as our estimate for this action (Table 1).

Wetlands, including billabongs and horseshoe lagoons, have significant cultural benefits for Indigenous people. Early accounts of Indigenous utilisation of wetland resources, and historic storylines, paint a rich narrative of connection between healthy communities and rivers (Moggridge and Thompson 2024). For instance, many species of small-bodied fish with little or no commercial or recreational benefit (in a contemporary sense) are of extreme significance to Indigenous communities (Humphries 2007). Nature repair actions need to extend beyond commercially and recreationally important species because a broad range of species traditionally had cultural significance. Thus, Indigenous groups must be adequately resourced to design and implement wetland recovery programs that seek to rehabilitate all aspects of wetland fauna and flora.

Action 5. Restore longitudinal connectivity, fish passage, by removing or modifying high-priority fish barriers

Freshwater fish and other fauna are threatened globally by habitat fragmentation and reduced flow connectivity (Harris et al. 2017). In Australia, fish passage and flow connectivity are limited by thousands of structures within channels and on floodplains, many of which are either legally non-compliant or unauthorised (Steinfeld and Kingsford 2013; Harris et al. 2017). Removal or remediation of these barriers can generate many benefits for freshwater fish and other biota, as well as generating co-benefits for recreational and commercial fisheries, tourism and First Nations (Makombe 2003). Furthermore, many of these structures were constructed either directly on, or required the destruction of, traditional fish traps, which were used by Indigenous people as a source of food and social cohesion. Significant numbers of fish traps were destroyed in the process of constructing dams and weirs.

Although national estimates of fish migration barriers do not exist, multiple regional studies inform an understanding of the extent of this problem. Baumgartner et al. (2014) estimated that there are more than 10,000 barriers in the New South Wales (NSW) portion of the MDB alone. We extrapolated from this estimate, recognising the high density of watercourses and modifications in this area, to assume at least 40,000 fish passage barriers exist nationally, with 5% of these likely to be a high priority for removal or modification on the basis of proportions identified in existing regional prioritisations (New South Wales Department of Primary Industries 2006; Lawson et al. 2010; Moore and McCann 2018). Costs vary significantly from removal of obsolete road crossings to construction of new bridges or culverts (Gordos et al. 2007), with costs for fishway installation on large barriers ranging from A$250,000 to A$1 × 10⁶ (2017 dollar values) per vertical metre (O’Connor et al. 2017). However, it should be noted that inflationary pressures over recent years are likely to have disproportionately increased costs since these estimates were published.

We assumed a mean cost of A$150,000 (2022 dollar values) for remediation of 95% of high-priority fish barriers and a mean cost of A$2 × 10⁶ (2022 dollar values) for the remining 5% of barriers, including 42 high-priority structures identified in the MDB in NSW (New South Wales Department of Primary Industries 2012). To estimate a national cost of this action, we then subtracted A$56.8 × 10⁶ already committed by the Australian Government for the Fish for the Future: Reconnecting the Northern Basin (New South Wales Department of Climate Change, Energy, the Environment and Water 2024). However, we note that progress against this initiative has been significantly hampered owing to significant cost increases in a post-covid world. Although the works program was designed and approved several years ago, not a single fishway has been completed under this program yet.

For high-level dams with a vertical height of >10 m, we estimated costs for advanced fishways. We estimated that 2790 vertical metres of dam require fishways, given that the total height of large dams in Australia is 17,438 m (Australian National Committee on Large Dams 2022) and an assumption that 16% of Australia’s total catchment area is obstructed by large dams on the basis of the NSW average (Harris et al. 2017). We then assumed a cost of A$1 × 10⁶ (2016 dollar values) per vertical metre for installation of advanced fishways (Australian Fisheries Management Forum 2016). Although existing fishways are present on ~3% of dams (Harris et al. 2017), we assumed that the majority of these require replacement or significant upgrades (New South Wales Department of Primary Industries 2012). We assumed operational costs for maintenance and monitoring to be 1% year⁻¹ of the total upfront capital cost (Table 1). In some instances, it may be worthwhile to review the functional relevance of these structures in a contemporary water management system. Many nations, internationally, are looking to remove large structures and these are leading to significant positive river health outcomes (Woodward et al. 2008).

Finally, we strongly recommend that in areas where traditional fish traps have been destroyed or built over, Indigenous groups are appropriately resourced to rehabilitate these sites. There is a significant amount of Indigenous ecological knowledge, which is required to effectively rehabilitate these sites in a culturally appropriate manner. This would require programs to re-establish local community connections, to educate regional communities on the importance of these structures and to re-establish historical narratives and storylines for future generations.

Action 6. Install cold-water pollution devices on priority large dams

Water released from dams can be colder than natural flows by up to 13°C in summer (Australian Fisheries Management Forum 2016), with this cold-water pollution (CWP) often extending significant distances downstream, up to 2,000,000 river kilometres annually, with deleterious consequences for the reproduction, development, growth, movement and survival of freshwater fauna (Lugg and Copeland 2014; Michie et al 2020). To address this problem, dams can be retrofitted with multi-level offtakes, or techniques to mix thermally stratified water bodies can be applied (Chaaya and Miller 2022). CWP mitigation can also offer further downstream benefits with respect to improved water quality and reduced transport of cyanobacteria (Chaaya and Miller 2022).

Of 93 dams studied in NSW in 2024, nine were associated with severe CWP, defined in relation to structures with deep intake of ≥10 m and large discharge (≥1000 ML day⁻¹) (Preece 2004). We therefore assumed that ~15% of Australia’s large dams will be high priorities for CWP devices. Costs for CWP solutions vary among dams but can range from <A$1 × 10⁶ to A$170 × 10⁶ (2000 dollar values; Sherman 2000 in Chaaya and Miller 2022). Projects proposed under the Northern Basin Toolkit program to address CWP comprised installation of a multi-level offtake at Pindari Dam for an estimated cost of ~A$14 × 10⁶ (2020 dollar values) and one at Glen Lyon Dam for A$3 × 10⁶ (2020 dollar values; Capon et al. 2020). On the basis of these figures, we assumed a mean cost per dam of A$8.5 × 10⁶ for initial capital works and labour. Operational costs were assumed to be 2% year⁻¹ of the total upfront capital cost (Table 1).

Action 7. Install fish diversion screening on all licensed irrigation pumps

Native fish populations in many parts of Australia are further threatened by irrigation pumps, which can remove hundreds to thousands of fish from watercourses daily, equating to millions of fish each year (Boys et al. 2021). Following extraction, these fish have limited, if any, opportunities to return to the main river channel (Baumgartner et al. 2009). Essentially, these fish become ‘lost’ from the main river system. Installation of ‘fish-friendly’ pump diversion screens can significantly (>90%) reduce fish injury and mortality, while simultaneously generating co-benefits for irrigators by reducing debris levels entrained by pumps and subsequent reductions in their operational costs (Boys et al. 2021; Rayner et al. 2023).

Rayner et al. (2023) estimated that there are ~4500 licensed irrigation pumps (>200 mm) in NSW requiring diversion screens. Average annual water sources from rivers, creeks or lakes for irrigation in NSW was 843,987 ML (Australian Bureau of Statistics 2021) between 2017 and 2020, which equates to ~187 ML per pump (assuming there are 4500 pumps). Australia’s average annual water soured from rivers, creeks or lakes in the same period is 2,141,578 ML (Australian Bureau of Statistics 2021). On the basis of the NSW ratio, we have assumed that ~11,418 pumps therefore require diversion screens nationally. The main benefit of a national screening program is to ensure that more fish, which are spawned in main channel environments, actually stay in the rivers.

On the basis of figures provided by Fish Screens Australia (see https://fishscreens.org.au/faqs/), we have assumed a cost of A$1000 (2022 dollar values) per megalitre of pump or channel capacity. Assuming this to be 187 ML per pump, a cost of A$187,000 per pump is estimated. Our national costs also exclude A$39.5 × 10⁶ (2022 dollar values) already committed under existing pump screening programs (Rayner et al. 2023). Operational costs for monitoring and maintenance are assumed to be 2% year⁻¹ of the total upfront capital cost (Table 1).

Action 8. Cap remaining open artesian bores and convert open bore-drains to pipe and trough systems in the Great Artesian Basin

Groundwater accounts for ~30% of all water use in Australia (Barnett et al. 2020; Walker et al. 2021). Groundwater is extremely significant to Indigenous people in a dry landscape such as Australia; there is extensive knowledge and values connecting to groundwater-dependant sites (Moggridge 2020). Exploitation of groundwater has resulted in declines in aquifer pressure as well as loss and degradation of groundwater-dependent ecosystems (GDEs; Fensham and Laffineur 2022). Capping free-flowing bores and converting inefficient, high-evaporation bore drains into pipe and trough systems is an effective way of repairing groundwater systems (Barnett et al. 2020) and GDEs (Fensham and Laffineur 2022), with co-benefits for farmers including improved grazing management (Pegler et al. 2002).

To cost this action, we assumed that effective groundwater management in the Great Artesian Basin, the largest and deepest aquifer in the world, requires all bores to be capped to ensure a reliable supply of water and to protect GDEs. We estimated that at least 331 uncapped bores remain in the GAB, given that recent bore rehabilitation projects have addressed 100 of the 431 open bores reported in 2019 (Great Artesian Basin Coordinating Committee 2019). Costs for bore capping range from A$14,131 to A$1.4 × 10⁶ per bore (2018 dollar values), with an assumed average cost of A$346,529 per bore (2018 dollar values) (Centre for International Economics and Resource and Policy Management 2003; Hassall and Associates Pty Ltd 2003). We also estimated that 4560 km of open bore-drain remain in the GAB, given 576 km of 5136 km of open bore drains reported in 2019 have already been converted to pipe and trough systems (Great Artesian Basin Coordinating Committee 2019). Costs of conversion are estimated at A$8485 km⁻¹ (2018 dollar values) (Centre for International Economics and Resource and Policy Management 2003; Hassall and Associates Pty Ltd 2003). Current Australian Government funding commitments of A$27.6 × 10⁶ were subtracted to generate total cost estimates (Table 1). Operational costs, for monitoring, maintenance and to address bore failure, were assumed to be 2% year⁻¹ of the total upfront capital cost.

Action 9. Restore groundwater extraction of the Murray–Darling Basin to sustainable levels of take

This action concerns returning groundwater extraction to sustainable levels in the MDB by strategic purchases of water licenses from willing sellers. In recognition of the need to protect groundwaters in the face of climate change and growing demand, the 2012 Basin Plan established sustainable levels of take for groundwater systems in the MDB. In the Upper Condamine alluvium, 3.25 GL still needs to be recovered, in addition to 32.5 GL recovered by June 2023, to achieve this target of 38.45 GL (Murray–Darling Basin Authority 2023). The cost of these acquisitions is assumed to be A$1906.90 ML⁻¹ (2018 dollar values) on the basis of those associated with the purchase of 35,697.40 ML of groundwater entitlements purchased through open tender in the Upper Condamine Alluvium for A$68,070,646 (2018 dollar values; Department of Climate Change, Energy, the Environment and Water 2023).