Community-based water demand management: socio-technical strategies for improving water security in Australian Indigenous communities

Sustainable water management in remote Australian communities is a delicate balance between sufficient and acceptable supply options and appropriate and effective demand approaches. This paper focus on the evaluation of community-based water demand management strategies piloted in four remote Aboriginal and Torres Strait Island communities in Australia. Findings of the pilot demonstrate that from a systems perspective, community-based demand management centred around education and encouragement of residents to conserve and use water efficiently, provide greater opportunities for long-term sustainable water management outcomes that support building of social capital. To ensure truly transformative management outcomes, a toolbox of socio-technological strategies should be used including, where possible, smart metering of water consumption and use of water-efficient devices. A key element of this approach to demand management calls for a shift away from business-as-usual policy towards a flexible learning approach that involves genuine collaboration between water managers and Indigenous communities.


Australian remote communities and water management
There are approximately 1187 remote communities in Australia, with the majority in Queensland, Western Australia and the Northern Territory (ABS 2006). More than one fifth of Indigenous peoples of Australia live in remote, isolated, and regional communities and make up over half the population in those areas; this includes mainland Aboriginal communities and those island communities of the Torres Strait Island peoples (ABS 2016a). Indigenous communities are usually not well provided for, and suffer from negative impacts of health, disunity, and gaps in education and health services (Moran and Corpus 2014). This has led to a significant gap between Indigenous and non-Indigenous health and wellbeing that, despite significant government investment, remains far from being closed (Altman 2006, Hunt and Smith 2006, Grey-Gardiner 2008, Hunt 2013a, 2013b, Australian Government Department of Prime Minister and Cabinet 2020).
Water management has been largely left out of the Closing the Gap framework (Vanweydeveld 2022), with a preliminary report by the Water Services Association of Australia confirming earlier empirical research (Bailie and Wayte 2006, Jackson et al 2019a, 2019b, Lansbury and Crosby 2022 that people living in remote communities have issues with water supply delays and quality maintenance of drinkable water, and concerns about the health of their communities. These structural vulnerabilities are exacerbated by climate change for those living in isolated and remote communities in Australia and the Torres Strait, particularly in relation to water security (Hoverman andAyre 2012, Lansbury Hall andCrosby 2022). Sources of drinking water for remote Australian communities is largely surface (dams) or groundwater (bores), depending on location. Remote and isolated community access to drinking water is impacted by the complexities of increasingly unreliable rainfall, and variable groundwater recharge, more extreme weather events such as cyclones and floods which all affect supply and quality of water to communities. Indigenous social and emotional wellbeing are also linked to the access of safe, reliable, clean water supply as it impacts Indigenous peoples ability to practice cultural ceremonies and maintain responsibilities in Caring for Country (Brugess et al 2005, Hunter 2007, Hoverman and Ayre 2012, Hall 2019, Lansbury Hall and Crosby 2022). The inequities on exposure, sensitivity and adaptive capacity towards climate change impacts on Aboriginal communities (Standen et al 2022) reinforces the need to address water security issues.
To bridge the divide between health and wellbeing in Indigenous and Non-Indigenous Australian communities, addressing the issues of water security in a sustainable, systemic way that considers all aspects and actors within the system is necessary (Jackson et al 2019b). Significant challenges exist such as: balancing a sustainable water supply with increasing demand; meeting the costs of delivery and maintenance; managing uncertainty due to lacking baseline consumption data; meeting energy demand; managing reliability of local services; and variable skills for repairs and maintenance (Jackson et al 2019b).
Despite the constraints to water availability and access outlined above, many Australian Indigenous communities have a high per capita water consumption e.g. >700l per person, per day (Yuen 2005, Beard et al 2013. Under current trends of increasing population growth in remote Indigenous communities (Taylor et al 2021) this high water demand is expected to continue to rise. As a result, the economic and environmental consequences are becomingly increasingly unsustainable, with impacts already being experienced at the community (severe water restrictions), local government (poor supply resilience) and state and national government levels (massive subsidy costs). Implementing effective CWDM strategies and improving water use efficiency is therefore critical to addressing the long-term cost-effective and ensuring a sustainable, secure and resilient water and energy future for remote Indigenous communities.
Several prior studies have highlighted water supply and demand issues in remote Aboriginal and Torres Strait Island community contexts (Grey-Gardiner 2008, Grey-Gardner and Taylor Willis et al (2006), Ross et al (2014) and Jackson (2021) observed leaking pipes and poorly functioning system parts; often identified as symptoms of ageing infrastructure and poor maintenance. Further, (Pearce et al 2007) found that children playing with water in hot weather to cool down was a main contributor to high water use; flagging cultural usage and water habits as contributing factors (see also Yuen 2005).
In relation to high water Beal et al (2014Beal et al ( , 2018 highlighted that outdoor use of water was disproportionately large compared to household use in a remote Aboriginal community in far north Queensland. Many of these outdoor water use drivers related to health and well-being such as dust suppression and cleaning, cooling surfaces in the absence of air conditioning and cultural activities (hunting, fishing, ceremonies) . Beal et al (2016a), demonstrated though high-resolution smart metering data empirically-based modelling that a 35% reduction in water demand is achievable by applying targeted demand management strategies. These shifts in demand have the potential to save up to AUD$100 000 per annum for diesel-reliant island communities alone (Beal et al 2016a).
Despite a growing body of literature on Australian Indigenous water rights (Toussaint et al 2005, Jackson and Altman 2009, Tan and Jackson 2013, Jackson 2019, more is needed to understand remote and isolated community residential water use and drivers of consumption (Beal et al 2016a, Jackson 2019. Accompanying this water demand research, is a greater need to focus on attitude and behaviours of citizens on water use; particularly how community participation and ownership can be enabled alongside education on water issues-which we discuss later in this paper.

Improving water security through demand management
Water demand management broadly defines any actions or strategies that aim to promote and improve water conservation or efficiency (Brooks 2006). Several studies across the globe have compellingly demonstrated the effectiveness of demand management strategies in reducing potable water consumption through: technical approaches such as installation of water-efficient devices (e.g. Liu et al 2017, Hutton et al 2020, voluntary and behavioural approaches such as tariff structures (e.g. Sahin et al 2018, Rahim et al 2021, social-based marketing (e.g. Walton and Hume 2011, Tourigny andFilion 2020), and mandatory behavioural approaches such as water restrictions (e.g. Ward 2008, Browne et al 2021).
There are multiple benefits to CWDM other than conservation of local water sources. A major benefit is reducing the energy use associated with providing water to communities, which in turn has multiple benefits including reducing diesel use many remote communities are reliant on high cost, greenhouse gas emitting diesel powered electricity generation (Beal et al 2016b. In the Torres Strait Islands, for example, energy-intensive desalination plants are operated continually through dry season and much of the wet seasons to supplement supply. An estimate of costs to supply water in these regions are as much as up to seven times that of urban supply costs (Beal et al 2016a). As rainfall patterns become less predictable, desalination has shifted from being a supplementary supply to a critical supply source in these island communities (Richards andSchäfer 2003, Werner andSchäfer 2007). In addition to high costs, there are reliability issues, and communities relying on diesel generation may regularly experience power interruptions for more than 24 h (Longden et al 2022). This has flow on effects to the water supply and health and safety implications for communities.
There is increasing focus on community participation in water governance and decision-making processes however, service providers are proceeding at varied rates and in different ways as they navigate delivery programs and goals (Jackson et al 2019b). Top-down, hierarchical, and technocratic frameworks of decision making often clash with meaningful community engagement and deadlines (Commonwealth of Australia 2014, Queensland Productivity Commission 2017). Water management still preferences a technical, engineering approach, often at the expense of understanding the complex social nuances of water systems, including diverse cultural understandings and meanings people attribute to water, (Checkland 1993, Jiménez et al 2014, Howarth and Monasterolo 2017. In recognising the need to move towards systemic and community-based collaborative management of water and energy in remote Indigenous communities, research conducted by Jackson et al (2019b) identified a range of challenges; including constraints from within governance arrangements and processes that perpetuate or arise from narrow thinking and short-term solutions and poor coordination across agencies; Economic and financial barriers included the power imbalances between community members and government employees, ultimately limiting their agency to effectively engage in water management processes. Issues related to poor data and information management and limited skills and capacity were identified as barriers within communities, as well as cross-cultural and community engagement skills within agencies. Limited connection into employment pathways and training, was also identified as being important for long-term sustainability and motivation for communities. The final category of cultural values and norms highlighted that at the core the differing worldviews and relationships with water between Western and Indigenous management as well as the sheer diversity of languages and cultural differences across Indigenous Australia were all challenging.
With this paper, we share findings from the RICES research project that sought to address some of the complexities and challenges ensuring year-round water security while balancing the water needs of the community. In the next sections, we will outline the methods and key findings of the three stages of the RICES project, before explaining in more detail, the management strategies and general approach to CWDM that have emerged from Stage three.

Research approach
To achieve the aim, the project was designed in three stages with each stage aligning to the research objectives. A participatory action research approach included community and stakeholder involvement across the research stages and activities. Stage one gathered baseline qualitative and quantitative data to characterise water consumption activities, attitudes, and challenges as well as water-related energy use in remote and isolated Aboriginal and Torres Strait Islander communities. Stage two trialled co-developed CWDM strategies in four remote Aboriginal and Torres Strait Islander communities (two mainland and two island communities). Stage three evaluated the CWDM strategies and identified which CWDM strategies were suitable and pragmatic for the context and suggested a pathway to promote long-term, efficient use of water in remote communities. Community engagement was a major component across all stages of the CWDM trial and included individual and group activities, combined and separate community and stakeholder events and the use of a range of social media and face-to-face communications.
The project methods, including participant recruitment, survey methodology and implementation, data generation, storage, and management, was reviewed by the Griffith University Indigenous Research Unit and cleared by the Human Ethics office (GU Ref No: ENG/15/14/HREC). As part of this ethics approval, individual participants are not identified.

Overview of communities
The key characteristics for each of the communities are provided in table 1. Community 1 (C1) is situated in the Central Australian arid (desert) zone in the Northern Territory. Community 2 (C2) is a mainland tropical coastal town in Cape York, Queensland (Qld). Community 3 (C3) and Community 4 (C4) are a tropical island communities located in the outer (C3) and inner (C4) Torres Strait Island group in the Coral Sea, Far North Qld. Data on the number of participating households for each community are also presented in table 1. Overall, given the acknowledged challenges inherent in recruiting remote and isolated Aboriginal and Torres Strait Island households ( Jamieson et al 2012), and due to the small populations, the participating household sample size was statistically solid, representing between 17% and 40% of total Aboriginal and/or Torres Strait Island households in each community (table 1). In terms of family composition, age, gender balance, and household stock, the participating households were generally representative of each community

Household recruitment process
Based on discussions with local project industry partners (Indigenous and non-Indigenous) a number of local council members and community representatives from several Aboriginal and Torres Strait Islander communities, were approached about their willingness to be involved in the project. Four communities across the Northern Territory and Queensland agreed to participate in the project which represented a total of 330 people from 77 households across the four remote communities (table 1). Household participants were recruited through door knocking, public workshops and/or were approached by council and/or the RICES team and invited to participate. All recruitment was conducted through a process of free, prior, and informed consent according to the Griffith University research ethics approval. In some communities where required, an interpreter (e.g. community-based Indigenous Engagement Officer or Environmental Health Officer) was present for the recruitment and the participant interviews process.

Stage one: baseline data gathering and analysis
Residential-scale water consumption was monitored using state-of-the-art, high-resolution, digital water meters and logging equipment installed at all the participating households. Smart energy meters were also installed to measure total household energy and hot water-related energy for a sub-sample of houses in each community. Using the high-resolution datasets from the participating households, a sample of received data was extracted from the database for two, two-week periods selected to represent the wet and dry seasons, and disaggregated into all end-use events (e.g. shower, clothes washer, tap, leaks, outdoor, bath, toilet) using the flow trace software Autoflow (Nguyen et al 2015). Concomitantly with meter and logger installation, a water fixture/appliance stock (e.g. clothes washer, toilet, shower) survey was conducted at each participating home which facilitated the disaggregation of trace flows from each home and provided a valuable snapshot of water consumption behaviours within each home.
Semi-structured interviews were administered face-to-face and standard demographic information and water use-related questions were designed to evaluate the existing uses of water by each household, as well as the knowledge and attitudes of different water users, and the capacity they had to conserve water. Questions were designed for the following categories of information relating to water demand management: water consumption; water values and behaviours; and attitudes toward community water security. Further details of this can be found in supplementary file A and Aldirawi et al (2019).

Stage two: trialling community-based water demand management strategies
Informed by a discussion of individual household water use patterns generated from smart meters (see figure 1 as an example) participants identified a range of potential water demand management activities (both technical and behavioural). The technical strategies were deliberately limited to devices that could be accessed by the community from their local shop or closest town: a tap timer, a soaker hose and a 'trigger-gun' hose nozzle. (Note, all these were provided free of charge to the participants during the trials). The behavioural WDM actions were s were co-developed with project participants with input from council staff, water managers, state government and RICES industry partners (see table 2).
The key steps of the trial were: 1. Smart water meter monitoring and analysis 2. Co-designing CWDM strategies with participants 3. Seeking community participant commitment to trial at least two CWDM strategies 4. Prompts and encouragement 5. Monitoring changes to water use 6. Participant and council evaluation of the CWDM trial.
The CWDM mechanisms proposed were co-designed with participants and informed by community engagement and change theories including community-based social marketing (McKenzie-Mohr 2000) and social practice theory (Shove et al 2012). The trials were carried out in four communities between 2016 and 2019. The participating householders themselves selected at least two CWDM strategies that they felt were the most suitable for their household water use activities (e.g. if the baseline data indicated that garden watering was a major activity then often a tap timer was trialled).

Stage three: evaluating and assessing the most appropriate CWDM strategies
As with the baseline water consumption patterns and household water use behaviour data, the CWDM strategies were evaluated using both qualitative and quantitative methods. Smart meter data enabled monitor of any water demand changes from the participating households during the trial period. Semi-structured surveys were used to illicit participant observation and opinions of the CWDM strategies that were employed in each household. Council officers were also consulted on their preferred CWDM strategies perspective. It was important to seek a service provider perspective to be consistent with the premise of the CWDM general approach which encourages feasible and practical options that would be embraced by both the water users and the water providers within a community. Following an evaluation by participants and council officers, CWDM strategies were identified as appropriate for further piloting in mainland and island communities. In addition to the trialled CDWM strategies, a number of other approaches, generated from the qualitative data were also considered.

Limitations
Project limitations related primarily to sample size and potential social desirability bias (Russell and Fielding 2010) (i.e. respondent tending to bias their responses to make themselves appear in a more favourable light), in the qualitative data responses. Over time, the number of participating households declined due mainly to technical and maintenance issues with the metering technology and some participants moving house. For example, in the C1 community, the attrition of households was significant by the end of the project due to unexpected hardware faults in the water meters and sorry business (in this situation, no-one occupied three houses for at least 6-12 months after family members had deceased in the houses as per cultural practises). The consequence of a lower sample size is the reduced confidence that the results are representative of community variability or are statistically robust. However, due to the inherently small community size in remote Australia, the number of participating households ranged from 20% to 40% of total community households for much of the project for at least three of the four communities. We have indicated where we are not confident with the C1 results due to low sample size. Social desirability bias on water use behaviours was able to be minimised by validating the survey responses with measured water use data from the smart meters.

Water consumption from smart metering data
Over the four-year monitoring period, across the four communities, per household water use averaged 2017 litres per household per day (l/hh/d) and 467 litres per person per day (l/p/d). A breakdown of average water end uses for the monitoring period shows that outdoor water use, leaks and showers were the main activities on both a per household and per person basis (figure 2). Overall, outdoor water use comprised most of the consumption, averaging 60% of total use, with an average of 50% used in summer (wet) and 70% in the winter (dry) months. Comparing water use both spatially and temporally provides more understanding of the patterns of use at different times of the year (figure 3). In this respect, water use (mainly outdoor) varied over time between each community. Outdoor water use was higher during prolonged periods of dry weather, which is usually associated with the winter months in Northern Australia (figure 3). Indoor water use activities generally remained homogenous over time and location. Results show that outdoor water consumption should be the focus of water conservation efforts rather than indoor use.
In many Torres Strait Island communities there are severe water restrictions during the dry season (May-November) and this can result in the treated, piped water supply being physically turned off by the council for up to 16 h a day (i.e. controlled access to the mains water supply). This is not the case for the C2 community which has 24 h a day, 7 d a week (24/7) access to treated drinking water, although water restriction notifications are issued in the hot and dry winter months in Cape York. For this community, where was no alternative water supply (unlike C3 and C4) to offset mains water use for outdoor activities during the dry season, such as dust suppression, cleaning and cooling purposes (see Beal et al 2018). For the C1 community, the opposite is true where the hot, dry period is in summer (December-February) which typically sees high water use, especially for outdoor use (figure 3).

Health and well-being drivers of water use
The results from the qualitative data indicate that outdoor water use is inextricably linked to health and wellbeing (table 3). From analyses of the survey responses, participant discussions, water end-use data and council consultation it emerged that several key drivers were contributing to the observed high outdoor water use activities. Following baseline analysis, interviews were conducted with all participants to further explore and identify the drivers (i.e. reasons and motivations) behind the different outdoor water uses. The analysis of those results identified that drivers of high outdoor water use are closely linked to necessary day to day functioning e.g. (group celebrations and festivities, tombstone openings, sorry business, children's play,  watering gardens, trees and establishing ground cover). Results highlighted in table 3 indicate that outdoor cleaning including cleaning food and hunting equipment was the most prevalent use of outdoor water, followed by watering gardens, social and cultural gatherings, dust control and heat relief. However, it must be noted that there are likely multiple uses for the same water use label. For example, children's play and using the swimming pool was categorised under social and cultural gatherings, it is also likely that swimming was also used for heat relief. This may also explain the lower numbers in the heat relief category-where researchers observed and heard anecdotes of participants soaking the ground to cool earth and generate an evaporative cooling effect with the prevailing wind-especially important to provide a cool area during social gatherings. The links between water use and health are discussed in more detail in Beal et al (2018).

Trialling CWDM strategies (stage 2)
This section provides a general summary of the findings of Stage two in relation to the effectiveness of how each step could enable CWDM implementation. In this regard, the findings of the six steps of the trial (detailed in the Methods section) are described in table 4. Note that the qualitative and quantitative evaluation of the CDWM strategies are discussed in the following section 4.3 (stage 3).

Quantitative evaluation-water consumption reductions
Overall, there was a 33% reduction in average water demand following the CWDM trial (table 5).
Comparisons of water end-use 12 months after the CWDM trials for C2, C3 and C4, at the same time of the year (thus controlling for weather influences) indicated a reduction in water use across most end-uses, leaks, and outdoor use. The average water use readings for C1 indicated a slight increase in water demand following the CWDM trial. This slight increase (12%) in water use could be due to several factors including the hot and dry weather at the time of the trial, the absence of some participants during the trial period, and the consequently very low sample size toward the end of the project which reduced the reliability of the data. Enabled participants to self-select tailored strategies to be trialled-thereby supporting longer-term uptake and buy-in. 3. Commitment to trial at least two CWDM strategies A written statement with photo taken to support the commitment during the period of the trial.
Verbal or written commitment supported participants in remembering and engaging with ongoing uptake in trial period.

Prompts and encouragements
Included written letter and visual feedback from the project team; public council notices of community-wide water levels.
Enabled keeping the program front of mind for participants and ensuring continued action from householders.

Participant and council evaluation of the CWDM strategies
Survey questions to all participants, workshops for volunteering participants to discuss the process, overall results and next steps.
Enabled objective (e.g. non-researcher) viewpoint of the success of the program and its different elements as well as learning outcomes.

Qualitative evaluation-survey responses
The RICES team sought feedback about the appropriateness and usefulness of the CWDM strategies that were trialled. In response to the question 'How useful did you find the actions that you used during the trial' participants reported feedback of actual water consumption data and benchmarking of individual household water use with others in the community as favoured CWDM options from RICES participants (figure 4). Participants were also generally in favour of CWDM strategies including: (1) council-led community workshops, (2) schools water conservation education, and (3) social and traditional media community announcements about current (real time) community water use and ways to save water (especially during the dry season). Each of these was employed in at least one community.

Identification of suitable CWDM strategies
In general, CWDM activities that involve feedback on household water use, and education and encouragement around how, why, and when to save water were popular with participants and were relatively low cost. Providing feedback on water use and benchmarking with other households, can still be done without smart meter water data, however, this would require more training and coordination by councils and service providers-though still at a generally low cost.
There was general support for the idea of creating local water champions in the community; to provide regular communication and feedback, and to promote a proactive community approach to water conservation. It was however, seen as important that this responsibility not be devolved from council/service provider to community members, and that leadership be shown by responsible parties to improve communication and coordination. There is potential to explore expansion of the scope of existing water officer (technical) role to incorporate more community engagement and partnership.
Many participants agreed that sharing water stories with friends and family, and even as a school activity for example, sharing stories was an appropriate and effective way for the Elders and Traditional Owners to transfer some of their traditional stories of accessing and saving water. This story sharing approach would also highlight the existing strengths of traditional (existing) water literacy in the community, with a view to further embedding local Indigenous knowledge into water conservation strategies.

Discussion
The trialled CWDM strategies were discussed with the project participants and council officers to test likelihood of engagement by community members, usefulness in terms of change outcomes and acceptability within Aboriginal and Torres Strait Islander culture and alignment with cultural protocols. It is acknowledged however, that these strategies may not be culturally or otherwise appropriate or workable across the breadth, depth and diversity of Australian Indigenous communities and language groups. Socio-demographics, existing household water infrastructure, proximity to major cities, and other household characteristics will influence preferences for CWDM strategies. Additionally, not all the CWDM strategies in isolation would work with all households in a community, although used in combination there is increased likelihood of a long-term change in how water is valued and used in communities. Below is a discussion of a range of factors to consider that may improve uptake of CWDM strategies in remote communities.

Policy and planning considerations for future CWDM programs
Each community had locally-specific limitations and opportunities for achieving water efficient outcomes. In addition to geography and environment, barriers specific to more effective engagement in demand management in remote Indigenous communities have been identified under categories of governance arrangements, cultural norms and values, economic and finance factors, capacity, education and awareness and data management and availability (Jackson 2019). Successful, and long-term water demand management strategies require a suite of tools to be implemented over time that acknowledge these barriers and seek to overcome them. As identified in Jackson et al (2019a), and drawing from the RICES project findings, water management policy in remote communities needs to broaden beyond the current model to consider the following: • Knowledge of the existing enabling environment (e.g. available funds, resources, expertise, past programs, community will and buy-in) and how this can assist or impede new water demand management directions. This knowledge will identify the realistic goals and manage expectations for both community and external parties. This is essentially the fit for purpose, fit for place rule that is critical for setting pragmatic and community-based water demand management approaches.
• Co-designing water demand management programs with Aboriginal and Torres Strait Islander representation from the start-not just consultation, there needs to be involvement of Indigenous peoples from the inception of the program. • Water demand management programs must budget sufficient costs for relationship building and community engagement (including repeat travel and in-community events). • As part of the collaborative approach, government needs to draw on local Indigenous knowledge about weather patterns, water supplies, historical water literacy around water conservation, water quality and relationships with water to make a demand management program relevant, culturally appropriate, and beneficial to local community. • Local capacity building where community members have the opportunity to become trained and knowledgeable in water management processes, e.g. this may include future consideration of a traineeship enabling environmental health workers to carry out minor plumbing repairs in emergency situations in remote Aboriginal and Torres Strait Island communities that do not have ready access to a licensed plumber such as the current Western Australian model (Government of WA 2016). • There must be an open policy of safe to fail for the community-based approach to demand management.
Not just financially (again this needs to be budgeted into programs) but also technically (e.g. water efficient devices may not be immediately used) and socially (communities may not initially engage strongly with education workshops). Research shows that unrealistic expectations, inadequate budgeting and an insufficient enabling environment are key ingredients for poor outcomes from water management programs in Indigenous communities (Jackson et al 2019a). Allowing some room to fail would include (yet not be limited to) the following approaches: * having a flexible budget; * realistic expectations for outcomes; * identifying and learning from 'poor' outcomes; * setting realistic timelines for programs; and * conducting a pilot CWDM program prior to a main roll out.

Re-thinking water restrictions and disconnections
Water restrictions (either fully controlled disconnections or through public notices with no strict enforcement) is currently the main demand management approach in remote Indigenous communities. It is generally agreed that water restrictions, especially ones that involve disconnection of town water supply for prolonged periods during the day such as in some outer Torres Strait islands , are currently considered the only effective option to ensuring there is sufficient water supply during the dry season. Limiting access however, to a community's only treated drinking water source is not a preferred option based on participant feedback as is not consistent with the SDG 6.1 goal of 'equitable access to safe and affordable drinking water for all' . Discussions with participants revealed that many householders would fill baths and buckets with water prior to the mains water cut off to ensure drinking water was available throughout the day. However, not everyone had the opportunity to store water or access alternative drinking water sources during the restriction times, most notably the elderly community members. This has also been reported in other studies where inaccessible and/or intermittent of drinking water supplies has been a barrier achieving SDG6.1 (Zozmann et al 2022). Therefore, water demand management strategies beyond enforcement measures are likely to yield a more equitable, sustainable, and resilient water supply. These findings further emphasise the need to consider alternative options for managing the long-term security of remote Australian water supplies using community-based education and encouragement approaches (see supplementary file B).

Balancing water conservation and public health
The research presented here focusses mainly on community approaches to ensure safely managed drinking water supplies are used in an efficient and sustainable way. Equally important, however, is ensuring an adequate supply of safely managed drinking water for positive environmental and public health outcomes in communities. For example, while it is essential that excess outdoor water use is curtailed through leak management and appropriate outdoor water conserving behaviours, there still needs to be a reliable supply for water use activities associated with healthy living practices such as dust and temperature control (Torzillo et al 2008).
Similarly, results show that indoor water use is also used for 'healthy living practices' (Torzillo et al 2008) such as washing bodies, clothes, bedding and towels. In this respect, water conservation messages should be very clearly directed towards leak reporting and repairs of water-based health hardware rather than indoor water consumption reduction in general. Further, Beal et al (2019) reported data demonstrating how in many communities with alternative water supplies, rainwater is the preferred drinking water source-not the treated town water. The balance between managing demand to ensure sustainable water supplies while continuing strong health promotion messages around using water to encourage healthy living practices, is a challenging element of demand-side management in remote communities. Nevertheless, there are excellent opportunities to address this type of challenge through broader roll out of community-based water management approaches.

Energy and cost implications
Previously reported data has demonstrated the reduction in energy and operating costs from applying smart metering and community-based approaches (see Beal et al 2016aBeal et al , 2019. For example, reduced bore pumping and desalination energy demand from lower water use (from CWDM) were estimated at 24% and up to 65%, respectively . Associated savings of 22% of business-as-usual costs have been reported from decreased energy (and water) consumption (Beal et al 2016a).

Suggested approach for community-based WDM
Building on the key insights from the research, the suggested approach toward community-based water demand management for remote Aboriginal and Torres Strait Islander communities is shown in figure 5. There are important elements to the approach, noted here: • Community-based demand management tools.
• External funding (this relates to a reduced reliance on capital and operating funding from local and state government and external providers but it must be emphasised that on-going strategic support for remote community councils is essential and that any saved costs through successful CWDM programs could be redirected to other community services). • Essential, ongoing collaboration between council/service provider and community.
• Time, trust, and safety to fail.
Each of the elements above are shown in figure 5; they are a pathway that is deliberately simple and broad; recognising that each community will require tailored CWDM programs, co-designed with local community and stakeholders. The five key elements are considered critical, overarching principles for transitioning to a more community-based approach to not only water demand management but to overall security and resilience of water and water-related energy supplies. The suggested pathway acknowledges that creating sustained behaviour change is not a simple and short-term process in any community, particularly in remote settings that require strong cultural, historical, governance, geographical and environmental considerations.

Key elements of the approach
Essential, ongoing collaboration between council/service provider and community. As an Indigenous community transitions to a resilient and sustainable water system, there must be an equal and ongoing relationship between community and council/service provider to ensure optimal engagement with CWDM tools. For example: • Council ensuring up to date water use notices are available through preferred community channels.
• Community aware of and adhering to water alerts about low water supply or temporary restrictions to outdoor use. • Residents, non-residents and council use of tap timers and other water efficient devices.
• Running (council) and attending (community) local workshops and water conservation education activities.
Capital and operating costs to local and state government and external providers. As the CWDM tools that are adopted in communities' transition from engineering and technical focus to more community-based CWDM tools, there will be a concomitant reduction in the reliance on money, resources, energy demand and associated direct and indirect expenses of high, ongoing water consumption.
Time, trust, safe to fail. Over time as the emphasis on council/service provider based CWDM approaches reduce and there is room for community-based CWDM approaches to 'fail and improve' , there is likely to be an increase in trust and confidence within the community that they have a sustainable, resilient, and ultimately independent water supply.

Conclusions
This paper has presented and discussed the findings from a large collaborative project investigating community-based approaches to water management in remote and isolated communities in Australia and the Torres Straits. Socio-technical approaches were employed to trial and evaluate a number of water demand management strategies. Water reductions up to 40% of pre-CWDM trial consumption were achieved in individual communities-though long-term reductions will require sustained and consistent efforts from councils/service providers to provide positive messaging to support community action e.g. they need to include the why and how of water conservation in their on-going messaging to community. Importantly, outdoor water use activities were found to be associated with cultural and environmental health practises, prompting the concern that simply enforcing water restrictions, or cutting off mains water supplies for prolonged periods of time, could result in detrimental health and wellbeing. The following are key conclusions from the research: • To help address the range of high-water use drivers, behaviours and attitudes, both community and council/service provider-led water conservation actions is needed within a broader water demand management program. • In all four communities, individualised water use feedback, including comparisons with the water use of other households was a CWDM strategy from both the community and council/service provider perspective. • Successful and long-term water demand management strategies require a suite of tools to be implemented over time. This is especially true for CWDM in remote Indigenous communities. Each community also has different limitations and opportunities for achieving water efficient outcomes and these must be understood and respected. • In the early-mid stages of implementation, communities need a safe to fail approach to allow some longterm behaviour change patterns to occur and to promote greater trust between local community members, council/ service providers, and external parties. • There is a need to co-design any WDM program with Indigenous representation from the start and for collaboration to be truly effective. Governments must budget sufficient community engagement costs into any water demand management program. • Indoor and outdoor water conservation messaging needs to avoid discouraging the use of water for key healthy living practices essential for human health (washing bodies, washing clothes, washing bedding etc).
A suggested community-based water demand management approach was described that acknowledges that creating sustained and sustainable changes in management inclusive of communities is not a simple and short-term process in any community, particularly in remote settings that require strong cultural, historical, governance, geographical and environmental considerations.

Data availability statement
The data cannot be made publicly available upon publication because no suitable repository exists for hosting data in this field of study. The data that support the findings of this study are available upon reasonable request from the authors.