Are environmental water requirements being met in the Murray–Darling Basin, Australia?

Environmental water requirements (EWRs)

The EWRs used here were drawn from long-term water plans developed for catchments across the MDB. EWRs are standardised into river-flow categories from cease-to-flow through to large overbank flow events (Fig. 1). For each flow category at each site, each EWR provides a guide for achieving the ecological outcomes required to maintain or improve environmental condition and ecosystem functions. EWRs were developed using extensive local, traditional and scientific knowledge about each catchment’s environmental assets and ecosystem functions; they were collected in partnership with water managers, natural resource managers, environmental water holders, landholders and community members (NSW Department of Planning, Industry & Environment 2020a).

Fig. 1.

Five flow categories linked to scientifically known environmental benefits, for which we used a subset (small fresh, large fresh, bankfull and overbank) to assess achievement of environmental water requirements (EWRs) at different sites in the Murray–Darling Basin.

An EWR describes the characteristics of a flow event (magnitude, duration, timing, frequency, and maximum inter-event period) within a particular flow category (e.g. small fresh; Fig. 1) that are required for that event to achieve a specified ecological objective or set of objectives such as replenishing waterhole volumes, providing water at sufficient depth for fish passage, or a flow pulse occurring at the time of year optimal for fish spawning. There can be multiple EWRs within a flow category and many EWRs across multiple flow categories. Fig. 2 conceptually summarises the EWRs for the small fresh, large fresh and bankfull flow categories for the Darling River at the Louth gauge (425004) for the period 2013–2020.

Fig. 2.

The environmental water requirements (EWRs) for the small fresh 1 (SF1), large fresh 1 (LF1) and bankfull (BK1) flow categories for the Darling River at Louth (gauge 425004). EWRs are described for a specific flow rate and prescribe the timing, duration, frequency, and maximum inter-event period for flow pulses greater that than flow rate. EWRs for SF1 (flow rate of ≥1500 ML day⁻¹; duration of 10 days minimum; timing anytime, but ideally October to April; frequency annual; maximum inter-event period of 1 year), LF1 (flow rate ≥15,000 ML day⁻¹; duration of 15 days minimum; timing anytime, but ideally July to September; frequency of 75% of years; maximum inter-event period of 2 years) and BK1 (flow rate ≥30,000 ML day⁻¹; duration of 15 days minimum; timing anytime; frequency of 50% of years; maximum inter-event period of 4 years) (NSW Department of Planning, Industry & Environment 2020b).

Assessing EWR achievement

Thousands of EWRs have been developed across more than 150 sites on the rivers of the MDB by the Murray–Darling Basin Authority (MDBA) and state governments, representing outcomes for each of the major rivers, wetlands, and species. We evaluated 23 EWR sites on representative rivers that matched stream gauge sites used as hydrological indicator sites by the MDBA during the development of the Basin Plan (Murray–Darling Basin Authority 2010a; Swirepik et al. 2016). These hydrological indicator sites have at least one stream gauge in each river catchment and align with the hydrological river model sites in the Basin Plan (Murray–Darling Basin Authority 2012). We did not use sites where daily-timescale observed or modelled streamflow data were not available or where there were substantial observed data gaps or known model biases. Appendix A1 outlines the EWRs used and their achievement criteria. The qualitative hydrological metrics that define EWRs are broadly divided into flow (in-channel) and volume (in weir pools and storages) metrics. In this paper, we assessed only flow-based EWRs. These are consistently defined across all flow categories on the basis of the following parameters (Table 1): minimum flow threshold (which can also include a maximum flow threshold), start and end month, target frequency (which can include minimum and maximum target frequencies), number of events per year, the duration of events, the minimum spell length of flows above a threshold, which contribute to the duration of events, the maximum inter-event period, and the maximum within-event gap tolerance.

The method used to choose sites and EWR flow ranges closely reflected that employed by the MDBA to support the determination of the environmentally sustainable level of take (Murray–Darling Basin Authority 2012). We used MDBAuth/EWR_tool (ver. 1.0.6, see https://doi.org/10.5281/zenodo.7435847), Python code developed by the MDBA, to evaluate achievement of EWRs at these 23 sites. The EWR assessment tool tracks EWR achievement at streamflow gauges in real time and allows comparison of EWR achievement between hydrological model simulations. The tool can be run in the following two modes: (1) ‘real time testing’ through automatically downloading and evaluating observed stream flow data, and (2) ‘scenario testing’ through evaluating inputted timeseries data.

Real-time, or ‘observed’ streamflow data, were formatted as MDBA BIGMOD data and processed through the ‘scenario-testing’ option because this allowed for pre-processing of input streamflow data (such as gap filling) before EWR evaluation. The EWR tool was configured to recognise EWR achievement even when the minimum and maximum flow thresholds and the flow durations were up to 10% lower than the target value. These allowances were provided to account for uncertainty with the specification of EWR criteria and were consistent with allowances applied in the determination of the environmentally sustainable level of take (ESLT) (Murray–Darling Basin Authority 2011).

Each New South Wales (NSW) site assessed contained, on average, 23 EWRs, often including multiple distinct EWRs within each flow category. From all the EWRs available, we selected an EWR to represent the following flow categories (Fig. 1, 2): small fresh, large fresh, bankfull, and overbank flow. Cease-to-flow, very low flow and baseflow environmental water requirements were not assessed because, at the time of assessment, revisions were still being made to the EWR assessment tool for evaluating these flow categories. Where multiple distinct EWRs existed for a particular flow category at a site, we adopted a conservative approach, opting for EWRs with smaller flow rates or shorter durations. Where the flow rates and durations were identical, we assessed the EWR most likely to be regularly achieved by considering the required duration, recurrence frequency and timing window; we did not assess the maximum inter-event period as part of the EWR assessment. In cases where EWRs had both a preferred timing window and a satisfactory timing window, we assessed against the more lenient satisfactory timing window. Within the overbank and floodplain flow category, each site usually contained multiple EWRs which represented a range of overbank flows from small overbank flows (which are targeted to occur more frequently) to very large overbank flows, which can be achieved only during large natural flood events. We conservatively assessed only the smallest of the overbank and floodplain flows, omitting EWRs associated with moderate and major flooding from the evaluation.

EWR comparisons

To explore EWR achievement, we used both observed hydrological data as well as a comparison among hydrological model scenarios. To assess EWR achievement across the MDB against observed hydrological data, we evaluated the longest observed time period for each site from state-based stream gauge data portals (Table 2). The longest common time period across most sites with substantially continuous data was 43.5 years from 1 July 1979 to 31 December 2022. This period covered three severe droughts, including that in 1997–1998, the Millennium Drought (2001–2009; van Dijk et al. 2013) and that in 2017–2020, as well as periods of sustained high rainfall and flows, including the La Niña events of 2020–2023. We included an additional 6-month period from 1 July 2022 to 31 December 2022 rather than ending the assessment at the water year, because this additional period contained large flood events across the Basin that were outside the average flow conditions, and which contributed to enhancing EWR achievement.

Stream gauge data gaps for periods of less than 120 consecutive days were infilled using linear interpolation, an approach aligned with techniques used in MDBA assessments (Murray–Darling Basin Authority 2013; 2020a). Infilling was applied to 17 of the 23 datasets (Table 2), with stream gauges along the Darling River at Louth, Bourke Town and Menindee containing the highest percentage of missing days, ranging from 1.40 to 1.65%. Three sites, the Darling River at Louth, the combined flow at the Macquarie River downstream of Marebone Weir and at Marebone Break and the Edward River at Deniliquin, contained continuous gaps too long to reasonably fill with this approach and so we used a reduced time period without data gaps for these sites (Table 2). EWR achievement was then determined for the full 43.5 years of observed gauge data (1979–2022) as well as a shorter 10-year period that corresponds with the commencement of the Basin Plan (2012–2022).

EWR achievement was also compared across three hydrological model scenarios simulated over a 114-year period (1895–2009) by the MDBA to inform the Basin Plan (Murray–Darling Basin Authority 2012). EWR achievement was compared across the following three scenarios: (1) modelled without any water resource development and therefore near-natural conditions (without development model); (2) modelled without Basin Plan sustainable diversion limit (SDL) settings, before implementation of the Basin Plan (baseline model); and (3) modelled with Basin Plan implementation settings, 2750 GL including the SDL adjustment scenario of 605 GL less 70 GL reduction in the northern basin (Murray–Darling Basin Authority 2017) (Basin Plan model).

Assessing environmental water requirement (EWR) achievement

EWR achievement was assessed using the MDBAuth/EWR_tool (ver. 1.0.6, see https://doi.org/10.5281/zenodo.7435847). This tool provides a publicly available method to consistently assess EWR achievement across the MDB. By way of overview, the tool iterates over each day in the timeseries and identifies all events within each water year where the sequences of flow are equal to or greater than the specified EWR criteria (flow rate, timing window, duration, minimum spell length, within event gap tolerance). EWRs are generally assessed over water years unless an assessment criterion, such as the timing window, requires bridging of multiple water years.

Summary statistics for successful events over the duration of the timeseries are returned, including the total frequency of achievement and duration of inter-event periods. These summary statistics can then be compared against the targeted frequency of achievement to identify whether the EWRs were achieved over the length of the timeseries. Although EWR achievement can be assessed on individual flow events, or over shorter periods of time, we assessed EWR achievement in aggregate over the entire flow record as a means of identifying whether environmental water needs have been met over the long term, in line with the aims of the Basin Plan. We adopted a conservative approach, defining EWR achievement or success when the frequency of flow events for that EWR occurred at a rate equal to, or higher than, 90% of the required frequency.

For detailed information on the method for evaluating EWR achievement using the EWR tool, including EWR interpretation assumptions, see the EWR tool user manual available through the MDBA GitHub repository (see https://github.com/MDBAuth/EWR_tool).

EWR achievement needs to be assessed over an extended time period, because the natural background hydrological variability of rivers in the MDB (Puckridge et al. 1998) means that achievement is not predictable. For example, the large fresh EWR (LF1) for the Louth gauge (Fig. 2) states that the event needs to be achieved in 75% of years, but this does not mean this EWR is always achieved every 3 years in 4; in wetter periods, it may be achieved every year; however, in naturally dry periods, it may not be achieved for several years, even before water resource development (Leigh et al. 2010).

Hydrological context

The 43.5-year period of hydrological data analysed included several La Niña events (1988–1989; 1998–2001; 2007–2009; 2010–2012; 2020–2023; see the Bureau of Meteorology’s Long Range Weather and Climate at http://www.bom.gov.au/climate/) where rainfall, and therefore river discharge, was higher than average, often leading to widespread flooding (Chiew et al. 1998). Sequential years of La Niña conditions, as occurred in 2010–2012 and then again in 2020–2023, generated some of the largest floods in the MDB (Leigh et al. 2010) (Fig. 3). The time period used for analysis also covered several El Niño events (1982–1983; 1987–1988; 1991–1995, 1997–1998, 2002–2003; 2006–2007; 2009–2010; 2015–2016; see http://www.bom.gov.au/climate/), including one of the most severe droughts on record, the Millennium Drought (2001–2009) (van Dijk et al. 2013), and a shorter period of extreme drought (2017–2019) where inflows to the MDB were the lowest on record (Fig. 3) (Holgate et al. 2020).

Fig. 3.

Stream discharge (ML day⁻¹) at two gauges in the two major river catchments: (a) Darling River at Bourke (425003) and (b) Murray River downstream of Euston (414203) for the period 1 July 1979 – 31 December 2022. Discharges required to meet the EWR flow threshold for small freshes (green), large freshes (blue), bankfull (orange) and overbank (red) for each gauge are displayed. The time period of the Millennium Drought (2001–2010) is shaded grey.

Achieving environmental water requirements (EWRs) over the past 43.5 years

Across the MDB, there were only two sites where all assessed EWRs were achieved over the 43.5-year period, namely, the Moonie River at Gundablouie and the Campaspe River at Rochester, whereas 11 sites had 0% achievement (Table 3). Less than one-third (31%) of the 72 EWRs assessed at 23 sites were achieved (Fig. 4), with achievement variable across flow bands (Table 3, Fig. 4).

Fig. 4.

Environmental water requirements (EWRs) achieved and not achieved in relation to the four flow categories (see Fig. 1) analysed using historical flow data, for 72 EWRs across 23 sites of the Murray–Darling Basin’s major rivers, 1979–2022 (43.5 years), the percentage of all EWRs achieved and not achieved across all sites and a tabular breakdown of the number of sites, and associated percentage, where EWRs were achieved and not achieved for each of the four flow categories from the long-term water plans.

Over the assessed time period, EWRs were more likely to be achieved in the tributaries of the northern MDB (the Darling River tributaries), with sites further downstream and along the Murray River showing lower levels of achievement (Fig. 5). The Barwon River at Mungindi, Gwydir River at Yarraman Bridge and Macquarie River downstream of Marebone Weir achieved most of the EWRs assessed. However, along the main stem of the Darling River and the Murray River and its tributaries, fewer EWRs were achieved over the 43.5-year period, with 50% being achieved in the Darling River at Bourke, 25% at three sites on the Murray River and none achieved at the remaining four sites (Table 3, Fig. 5).

Fig. 5.

Percentage of environmental water requirements achieved at each stream gauge site (see Tables 1, 3) in the Murray–Darling Basin across four flow categories (see Fig. 1), on the basis of analyses of flow data, 1979–2022 (43.5 years); (a) sites ordered on the basis of percentage EWR achievement, showing southern basin sites (black bars) and northern basin sites (grey bars), see Table 3 for site codes; (b) gauges and percentage EWR achievement are displayed spatially across the MDB, with the circle colour depicting percentage of EWRs achieved.

Environmental water requirement (EWR) across flow categories

Across the MDB, there were similar levels of EWR achievement for small freshes, large freshes and bankfull flows, with slightly higher achievement for overbank (Fig. 6). In the northern Basin, 6 of 10 sites had not achieved the small fresh EWR. This included all sites on the Darling River (Fig. 6, Table 3). Only those sites on the upper tributaries, including on the Macquarie River, the Gwydir River at Yarraman Bridge, the Barwon River at Mungindi and the Moonie River at Gundablouie, met the small fresh EWR. A similar pattern occurred for large freshes, with 7 of 10 sites not meeting this EWR over the period, including all sites along the Darling River (Fig. 6, Table 3). Again, the large fresh EWR was met at similar sites in the northern tributaries, the Macquarie, Namoi and Moonie Rivers. The bankfull EWR was met at only 2 of 5 sites, Bourke on the Darling River, and the Moonie River (Fig. 6, Table 3). The overbank flow EWR was met at 7 of 10 sites across the northern basin, including the Namoi, Gwydir, Moonie and Condamine–Balonne (Narran River at Wilby Wilby) rivers, as well as the Barwon River at Mungindi and two sites along the Darling River (Fig. 6, Table 3).

Fig. 6.

Achievement of environmental water requirements (EWRs; achieved, green; not achieved, red; not assessed, white) in 1979–2022 at 23 sites in the Murray–Darling Basin analysed by flow category.

Achievement in the southern Basin was lower than in the northern MDB, over the 43.5-year period, with 8 of 10 sites assessed having not met the small fresh EWR, with achievement only downstream of Yarrawonga Weir and the Campaspe River on the Murray River. Likewise, 7 of 9 sites assessed had not met the large fresh EWR, which was achieved only on the Lachlan Valley and the Torrumbarry gauge on the Murray River. The bankfull EWR was met only on the Campaspe River, with all remaining sites failing (5 of 6); the same result was evident for overbank flows, with only the Campaspe River at Rochester meeting this EWR (Fig. 6, Table 3).

For Ramsar-listed wetlands of international importance across the Basin, only two stream gauges, namely upstream of the Narran Lakes Nature Reserve and the Gwydir Wetlands, achieved the overbank flow EWR (Table 4) over the 43.5-year period. All other assessed stream gauges above Ramsar-listed wetlands, including the Macquarie Marshes, did not meet the overbank flow EWR.

Long-term trends and detecting Basin Plan impacts on environmental water requirements

The total number of flow events meeting EWR flow thresholds for one site in the northern and southern MDB (Darling River at Bourke and Murray River downstream Euston respectively) were calculated for each decade over the 43.5-year period and compared. This analysis suggested a consistent decline in the total number of events in each of the four EWR flow categories in both the northern and southern Basin (Fig. 7). Only the small fresh EWR at the Murray River downstream of Euston showed a positive trend; the number of small freshes meeting this EWR flow threshold increased at this site over the past decade (Fig. 7a, c). When the duration of each flow event meeting EWR flow thresholds was compared across all decades, there was a consistent decline in duration in both the northern and southern Basin for all flow thresholds (Fig. 7b, d). This suggests that although the number of small freshes in the Murray River may have increased, their duration had not.

Fig. 7.

Decadal trends in total number of flow events for the four EWRs for (a) Murray River downstream Euston (414203) in the southern basin and (c) Darling River at Bourke (425003) in the northern basin, and decadal trends in the total number of days of each flow event above the four EWR flow thresholds for (b) Murray River downstream Euston (414203) in the southern basin and (d) Darling River at Bourke (425003) in the northern basin. Note, the final decade (2013–2022) is 9.5 years and not 10 years.

The largest decline in number and duration of thresholds of all four EWR categories occurred between the first decade (1983–1992) and the second decade (1992–2002) (Fig. 7), with some stabilisation of the declining trend in the last decade. However, other than the number of small freshes at the Murray River site, the other thresholds, at both the northern and southern sites, were 50% lower in the most recent decade than in the first decade 1983–1992.

The most recent decade of data was for 2012–2022 when the Basin Plan was partially implemented, during which water extractions were reduced and additional environmental water was made available as managed environmental flows (Swirepik et al. 2016). The percentage change between the 43.5-year flow period compared with the 10 years of Basin Plan flows (Table 5) suggested that there were some increases in EWR flow threshold achievement for small freshes, particularly in the southern Basin and where environmental flows can be delivered. Sites at the ends of river systems, such as the Barwon–Darling River, at sites downstream of Mungindi and the Murray River in South Australia had decreased EWRs achieved over the past 10 years (Table 5, Appendix A2).

To further compare EWR achievement and changes through time, we compared achievement on the basis of observed flow data from river gauges for the two time periods (43.5 years and 10 years under the Basin Plan), with achievement using the three modelled scenarios for the 114-year period (1895–2009). The three modelled scenarios included without development (i.e. near natural conditions), baseline (i.e. before Basin Plan) and Basin Plan implementation (i.e. 2750 GL, including the SDL adjustment scenario of 605 GL less 70 GL reduction in the northern basin). By design, this analysis evaluated only EWRs that were achieved in the ‘without development’ scenario, with EWRs that were not achieved being removed from the analysis. The ‘baseline’ scenario achieved just over one-third (35%), and the ‘Basin Plan’ scenario achieved 46%. In comparison, our observed data showed the EWR achievement for the 43.5- and 10-year period was 31 and 26% respectively (Table 6). The analysis of the modelled data supports the trend analysis above, suggesting that current Basin Plan settings should improve EWR achievement from 34% under baseline conditions to 46% under Basin Plan settings. However, this improvement has not yet been seen across all EWRs in the observed data over the past 10 years.

Environmental implications

EWRs were designed to represent the minimum water requirements needed to support important environmental assets and ecosystem functions. Where flow regimes fall short of meeting these requirements, ecosystem condition is likely to decline (NSW Department of Planning, Industry & Environment 2020a). Our analyses showed that at 65% of the stream gauge sites assessed over the 43.5 years, most EWRs evaluated had not been met, despite the evaluation period including both periods of drought and widespread flooding (Fig. 3).

The failure to meet the EWR thresholds across such an extended period is reflected in observed declines in species distribution and abundance and ecosystem functions across the MDB (Kingsford et al. 2017; Ryan et al. 2021). The degradation of flow-dependent ecosystems continues to be documented in condition monitoring and evaluation assessments, including the Sustainable Rivers Audits (Davies et al. 2010, 2012), the 2020 Murray–Daring Basin Plan evaluation (Murray–Darling Basin Authority 2020b), and the 2021 State of the Environment Report (Kingsford et al. 2015, 2017; Green and Moggridge 2021).

EWRs also include a parameter that specifies the maximum inter-event period, namely, the longest allowable time between flow events before a significant decline in the condition, survival, or viability of a specific population is likely to occur (NSW Department of Planning, Industry & Environment 2020c), with both the required frequency and maximum inter-event periods needing to be met to achieve environmental outcomes (NSW Department of Planning, Industry & Environment 2020a). Although our study did not assess inter-event frequencies, a brief evaluation showed instances where the maximum inter-event period was exceeded, meaning that if this parameter were assessed, overall EWR achievement would have been further reduced.

Factors contributing to the lack of environmental water requirement achievement

Determining the primary factors that affected EWR achievement over the different time periods assessed requires hind-cast simulations using suitable hydrological river models. This was not our focus; such work could be undertaken using consistent hydrological models across the MDB. However, we qualitatively assessed factors that, individually and collectively, may have influenced the outcomes.

Insufficient frequencies and volumes of water are likely to have reduced achievement of EWRs. Indeed, modelled EWR achievement for different water-resource development scenarios, reflecting varying levels of flow volumes, suggested that frequency and volume were key determinants of EWR achievement (Table 5). The ‘Basin Plan’ modelled scenario over the longer 114-year period achieved only 46% of the EWRs. An inadequate frequency of flow events that met EWR thresholds was a key factor in this low EWR achievement, suggesting that even under Basin Plan settings, full achievement of EWRs will be difficult.

Complexities with how achievement of EWRs is measured using the current assessment tool may have affected the results. Flow timing, duration and other non-volumetric parameters interact with frequency and volumetric parameters, affecting EWR achievement over particular time periods. Insufficient duration of events was sometimes a key factor, but this was not always the case. For example, EWRs at four sites (Campaspe River at Rochester, Loddon River at Appin South, Lachlan River at Booligal and Murrumbidgee River downstream of Maude Weir) required two small freshes within 1 year; however, during the 2022 flood, flows were sustained above the small fresh threshold for the entire year, thus affecting achievement of at least one of those EWRs during that period. Similarly, during that same flood period, some flows above EWR flow thresholds occurred outside the timing window and were not counted towards EWR achievement over that period. Finally, three of the EWRs (all small freshes) at three sites (Campaspe River at Rochester, Loddon River at Appin South and Gwydir River at Yarraman Bridge) had a maximum flow threshold whereby flows needed to be within a flow range to count towards EWR achievement. During the recent floods of 2020–2022, flows were higher than the maximum flow threshold and these EWRs were not counted as being achieved over that period. This suggests that during periods of high flows, when bankfull and overbank EWRs are being met, the concurrent achievement of small fresh EWRs with the maximum flow thresholds are not counted, which means that small fresh and large fresh achievement may be slightly higher than the assessment tool suggests.

In addition, the time period used to assess the EWR achievement may also have an impact. We found that for shorter assessment periods, the start of the time period and the resulting division of the timeseries into yearly blocks could affect the EWR achievement. The use of water years or calendar years can also influence the EWR achievement through the same process. Where there are no timing-window constraints identified in the EWR parameters, the EWR tool assumes that flows can cross over water years. The tool allocates the event to the year that contains most of the event duration. If the event is split evenly between water years, it is allocated to the earlier water year; however, if the event is long enough to be counted in both water years, it is still counted only in one. Further refinement of EWRs and the EWR tool are continuing. We consider that additional research is needed to further identify limitations on the use of EWRs and factors that affect EWR achievement in ways not envisioned when specifying EWRs. This research could help inform future reviews of EWRs as well as EWR assessment methods.

Although environmental flows are not a new concept (Arthington 2015) and are provided for under the Water Act 2007 (Cth) and Basin Plan, there are still challenges in the effective delivery of environmental flows. For example, management of overbank flows for environmental purposes potentially inundates floodplain corridors involving private property. Landholders affected by such flows may argue that their land suffers damage, and if the flows occur without prior landholder approval, may seek compensation. Structures built across rivers, such as roads, bridges and low-level crossings, also limit flow rates that can occur without potentially leading to structural damage. These physical channel-capacity constraints have been operationalised through limits on managed flow rates along certain river reaches, particularly in the Murrumbidgee, Goulburn and Murray rivers. During normal operations, river operators seek to constrain river flows to below these limits. At some sites downstream of constraints identified in the MDBA Constraints Relaxation Strategy (see Table 2), flow limits were smaller than the smallest overbank EWR assessed.

These channel-capacity constraints mean that during normal river operations including regulated river flow, held and planned environmental water cannot be used to meet overbank EWRs, with channel constraints limiting the effectiveness of river operators to provide overbank flows for environmental needs (Young et al. 2011; Kahan et al. 2021). In these cases, outcomes for floodplains and wetlands that rely on overbank EWR achievement are restricted mainly to periods of natural flooding or uncontrolled flow events such as dam spills. Finding approaches to overcome these challenges is crucial to ensuring the successful restoration of floodplain and wetland ecosystems and the species that depend on them.

Uncertainty in flow–ecology relationships and environmental water requirements

The relationships between environmental impacts and the failure to achieve EWRs and environmental outcomes remain uncertain. This reflects the limited empirical data on explicit relationships between flow and ecological response, which are often non-linear and complex (Lester et al. 2011; Colloff et al. 2018). Although many studies have linked altered flow and flooding regimes with environmental degradation (Bunn and Arthington 2002; Poff and Zimmerman 2010; Webb et al. 2010; Arthur et al. 2012; Davies et al. 2014; Colloff et al. 2018), the specific relationships between flow thresholds and ecological responses remain uncertain for many rivers.

To address some of this uncertainty, we used six conservative approaches. These included assessing key representative sites rather than the full domain of EWR sites; allowing 10% leeway for minimum and maximum flow thresholds and flow duration; using smaller flow EWRs for each flow category that were easier to achieve and limited the assessment of larger overbank events; applying the ‘satisfactory’ timing window instead of the ‘preferred’ timing window (the latter generally increased environmental outcomes); recording achievement if the frequency of events was equal to or greater than 90% required; and not limiting EWR achievement on the basis of lack of achievement of maximum inter-event periods.

EWR evaluation was also limited to in-channel stream gauges, which estimate stream flow on the basis of river height through ‘ratings curves’, which often do not adequately capture all important parameters of flow requirements, such as inundation extent during overbank flow events (Ren and Kingsford 2011). Additionally, any overbank flow that occurs externally to a stream gauge is not captured in this assessment, including any floodplain or wetland environmental watering that occurs by pumping water overbank. Consequently, accurately assessing floodplain flow requirements, essential for environmental outcomes, requires ancillary data and alternative methods. Better integration of in-channel gauge data with remote-sensing monitoring across floodplains could improve the specification and evaluation of overbank flow EWRs.

Adopting an adaptive management approach (Kingsford et al. 2011b; Roux and Foxcroft 2011) can also address uncertainties in flow–ecology relationships, as well as linking flows to ecosystem responses explicitly. This approach involves proposing, testing, evaluating and updating flow–ecology relationships over time to explicitly link objectives to outcomes. An agreed standard for defining EWRs and long-term follow-up evaluations of their ecological appropriateness and effectiveness are vital components of an effective environmental water management framework (Davies et al. 2014). Current arrangements require reviews within 5 years of implementation and offer opportunities to re-evaluate EWRs considering new information, including both ecological knowledge and climate-change impacts (NSW Department of Planning, Industry & Environment 2020a).

Is there enough environmental water to meet the environmental water requirement thresholds?

The Water Act 2007 (Cth) requires the Basin Plan to determine the environmentally sustainable level of take (ESLT) that is based on the ‘best available scientific knowledge and socio-economic analysis’ (Water Act 2007, S.21(4)(b), p. 46) and which, inter alia, does not compromise environmental outcomes, including key environmental assets and ecosystem functions (Water Act 2007, S.4(1)). The Guide to the Basin Plan (Murray–Darling Authority 2010b) indicated that it was necessary to recover between 3856 GL year⁻¹ (high risk of not achieving environmental outcomes) and 6983 GL year⁻¹ (low risk of not achieving environmental outcomes) to reinstate an ESLT. However, this volume was later revised down (Murray–Darling Authority 2011) and in 2012 the Commonwealth Government legislated a recovery volume of 2750 GL year⁻¹ plus 450 GL year⁻¹ for ‘enhanced environmental outcomes’. This recovery volume was further reduced by 70 GL year⁻¹ after a review conducted under S.6.06 of the Basin Plan (commonly known as the ‘Northern Basin Review’).

A CSIRO review of this water-recovery volume concluded that it was ‘not consistent with the currently stated environmental targets’ (Young et al. 2011, p. 29) and ‘the modelling indicates that the proposed SDLs [Sustainable Diversion Limits] would be highly unlikely to meet the specified ecological targets even in the absence of future climate change. Operational constraints are a key reason for this, but many achievable targets are also not met in the modelling’ (Young et al. 2011, p. 30). The SDLs legislated in the Basin Plan reflect inadequacies in the ESLT determinations and the lack of achievement of more than half (54%; Table 5, Basin Plan scenario) of EWRs we examined. As such, the Basin Plan water-recovery targets, even when fully implemented, are likely to increase some EWR achievement but may not be sufficient to achieve EWRs over the long term. All EWRs are unlikely to be achieved unless additional water is made available for environmental needs and there is a repeat of historical climate conditions.

Impacts of climate change on water availability

The achievement of EWRs into the future could be significantly influenced by alterations in rainfall and river inflows (Speer et al. 2021), characterised by deviations from historical climatic conditions, including increases in the severity and duration of drought conditions (Kirby et al. 2013). The MDB is already being affected by the consequences of climate change (Bureau of Meteorology 2020; Whetton and Chiew 2021; Quentin Grafton et al. 2022), including shifting rainfall patterns, water availability and fundamental alterations in hydrological relationships (Peterson et al. 2021), such as those between surface and groundwater systems (Crosbie et al. 2023). Moreover, the MDB is experiencing elevated mean surface temperatures and, consequently, increased evaporation rates. Lower runoff reduces the amount of water available for all uses, including for environmental purposes, reducing the effectiveness of water management practices. These climate-induced changes have already had a notable impact on river flows and water resources in the region, and their influence on EWR outcomes is expected to continue to grow (Chiew et al. 2009), affecting the supply of important ecosystem services (Colloff et al. 2016).

Ecosystems in the MDB have evolved over long-term historical climates, and the shifts posed by climate change will further reduce ecosystem resilience (Colloff and Baldwin 2010). Addressing these challenges will necessitate the adoption of adaptive water management strategies that consider the dynamic nature of climatic conditions and their effects on the MDB’s ecosystems. Furthermore, the impacts of climate change may require a triage approach to determine which ecosystems can be conserved using environmental watering, along with a re-evaluation of some EWRs to account for the effects of climate change (Schweizer et al. 2022). Future research and policy initiatives should focus on transparent monitoring of EWRs on the basis of best available science and implementing measures to safeguard the ecological integrity of the MDB under changing climate conditions.