How could managed aquifer recharge be feasible in the Coleambally Irrigation Area?

ABSTRACT Managed aquifer recharge (MAR) has been proposed as an innovative water storage method to manage variable water availability in Australian agricultural regions by combatting water shortage whilst addressing groundwater overuse and reducing large evaporative losses from surface storages. MAR innovations involve the recharge of water into aquifers and extraction when needed. However, their uptake has been limited as they are associated with large amounts of uncertainty. We present an argumentation approach to reason about the feasibility of MAR in the Coleambally Irrigation Area in the Murrumbidgee catchment. Three scenarios that focus on managed aquifer recharge and two competing surface water scenarios are developed for the case study area, as well as associated recommendations aimed at the key stakeholders who can initiate change. The approach and scenarios show potential, with future work involving the implementation of the recommendations and the application of the approach to another case study for further refinement.


Introduction
The variable Australian climate necessitates the capture and storage of water in favourable periods for times of low supply (Kokic et al. 2007), to meet the demands of people, agriculture and industry.Traditional storage via large dams and surface storages (Arshad, Qureshi, and Jakeman 2013) results in large evaporative loss (Arshad et al. 2012).In the agriculturally important Murray-Darling Basin (MDB) (Rawluk et al. 2013), evaporative losses of up to 40% of a storage's capacity have reduced water use efficiency and led to large economic losses (Craig et al. 2005).Groundwater withdrawal to supplement surface water is commonplace although needs to be managed sustainably to avoid overexploitation of the resource (Arshad, Guillaume, and Ross 2014;Dillon et al. 2019).
Managed aquifer recharge (MAR) involves banking water underground, often through infiltration into a shallow unconfined aquifer or injection into a deep confined or semi-confined aquifer, to recover the water when needed (Yuan et al. 2016).In the southwestern United States, particularly Arizona and California, infiltration-based MAR systems are used to bank surface water for irrigated agriculture, among other uses (Megdal, Dillon, and Seasholes 2014;Scanlon et al. 2016).Favourable hydrogeology and well-established regulatory frameworks in these areas allow MAR schemes to utilise available surface water (Megdal, Dillon, and Seasholes 2014;Scanlon et al. 2016).Infiltration-based MAR schemes have also been developed to support agricultural irrigation in Spain and Israel (Dillon et al. 2019).In Australia's farming regions, MAR has been proposed to assist in storing water to meet water demand, particularly during drought, while minimising evaporative losses (Rawluk et al. 2013;Fuentes and Vervoort 2020).However, MAR in farming regions of Australia is not without constraints, including the lack of knowledge about the hydrogeology of an area, or unfavourable hydrogeology (Fuentes and Vervoort 2020;Arshad, Guillaume, and Ross 2014), policy (Ward-Noonan 2021;Ward and Dillon 2012;Dillon et al. 2020) or institutions There are only a small number of MAR schemes currently operating in an agricultural setting in Australia (Dillon et al. 2009).
Past analysis of MAR potential in Australia has typically applied a disciplinary approach, focusing on technical (Fuentes and Vervoort 2020;Arshad et al. 2012), financial (Arshad, Qureshi, and Jakeman 2013), governance (Ward and Dillon 2009) or social aspects (Rawluk et al. 2013).Fewer studies have considered multiple aspects of MAR feasibility.Examples include Khan et al. (2008), Arshad, Guillaume, and Ross (2014) and Ticehurst and Curtis (2017).Disciplinary studies of hydrogeological or economic feasibility, for example, are not sufficient to develop enough confidence to catalyse investment in MAR.Evidence of this is the limited MAR investment in Australia outside of government-led water reuse schemes.A broader set of feasibility criteria (Ticehurst and Curtis 2017) is necessary to provide compelling descriptions of how an investment could proceed and the system changes engendered.In this context, the Cotton Research & Development Corporation (CRDC) funded a project investigating the potential to implement MAR in key irrigated cotton growing regions of Australia.The brief of the project was to assess whether MAR was worth pursuing, or whether it should be laid to rest.That is, the research team explicitly needed to disprove the hypothesis that MAR was not feasible.
An argumentation approach was adopted as the foundation of the research, informed by case study selection, structured feasibility analysis and iterative development of scenarios.Although scenarios have previously been used to describe and backcast alternative possible states (Herman et al. 2015;Maier et al. 2016;Mahmoud et al. 2009;Notten et al. 2003Notten et al. , 2005;;Schuck et al. 2018), our specific integration for the purpose of argumentation is unique.This paper 1) demonstrates a method to construct an argument for investment in an innovation (here, MAR), and 2) provides scenarios describing how MAR could be feasible in a case study: the Coleambally Irrigation Area (CIA) in the Murrumbidgee valley.
It should be noted that while the paper does make a methodological contribution, it is intended primarily as place-based research, i.e. providing new scientific insights about a particular place (CIA and to some extent more broadly the Murrumbidgee catchment, the Murray-Darling Basin and Australia).The value of place-based research has been recognised in understanding community context in water resources in particular (Gerlak et al. 2018) and in facilitating integration of local knowledge and co-construction of solutions (Balvanera et al. 2017).More specifically, Beven (2007) and Beven and Alcock (2012) made the case for changing the modelling process 'from one in which general model structures are used in particular catchment applications to one in which modelling becomes a learning process about places' (Beven and Alcock 2012).The scenarios are developed as a contribution to this learning process.
The authors do note that these scenarios do not constitute endorsement from either the Coleambally Irrigation Co-operative Limited (CICL) or the CRDC for a particular course of action.Further details of the analysis are available in the project report (Guillaume et al. 2020).

Approach
Investing in innovative projects is associated with great uncertainty as the requirements and consequences of the project are only partly known.This challenges traditional uncertainty quantification and modelling techniques (Maier et al. 2016).The result is that the identified uncertainty may present as insurmountable risk and impede decision-making.To overcome this obstacle in this case study, an argumentation perspective of decision support was used, which highlights available information rather than uncertainty.
In general, argumentation involves the use of premises to determine the truth or acceptability of a conclusion.In our work, we identify a conclusion to be demonstrated (or hypothesis to be disproved), namely whether or not MAR is feasible, and seek to identify 1) scenarios that justify how MAR could be feasible, and 2) evidence that shows the scenarios are plausible.The burden of proof is explicitly on collecting sufficient evidence to convince specific agents of change that MAR is feasible.The approach therefore focuses on minimal information needs, i.e. can enough data/information be gathered and presented that support the innovation to ensure that the innovation cannot be dismissed?
Our use of scenarios inherits from scenario literature on backcasting and more generally normative transformational scenarios (Börjeson et al. 2006).Backcasting starts with a desired future and works backwards to identify how that future might be reached from the present.Börjeson et al. (2006) contrasts transformational and preserving scenarios, where the latter are used when it already seems possible to reach the target.In the case of an uncertain innovation such as MAR, our starting premise is that MAR is both transformative and requires transformation, and within an argumentation approach, we hypothesise that a plausible and desirable scenario exists in which MAR is deployed.Using construction of narrative scenarios based on both desktop analysis and stakeholder feedback, we then seek to construct such a scenario.The form in which MAR is finally deployed may involve radical changes to existing institutional structures that the scenario approach allows us to only hint at rather than having to confront these changes in the short term, allowing steps to be taken to open new options and reduce existing path dependency, rather than triggering resistance from the incumbent power structures by laying out the details of that path.The advantage of leaving out unnecessary details and focusing on minimal information needs is consistent with literature on beneficial uses of ignorance (Gross and Mcgoey 2015).
The general idea of backcasting to test the feasibility of alternative futures has been used before (e.g.Schuck et al. (2018)), but its specific application here is unique.In particular, we integrate it with a systematic framework of criteria, emphasise incremental transition, compare to alternative options, consider the level of detail needed to support the argument and ultimately focus on initial next steps rather than long-term commitments in order to encourage immediate action.In principle, each of these ideas defines specific premises and argumentation schemes (Walton, Reed, and Macagno 2008) that support our conclusion in the face of possible counterarguments, though it has not yet been necessary to formalise our approach in those terms.
Competing approaches include cost-benefit analysis, criteria-based feasibility analyses, and analyses of pros and cons, or barriers and enablers.Our approach makes explicit that the burden of proof is on us to convince specific agents of changes, and therefore helps identify how much and what information is needed, rather than trying to be 'objective'.Results from all but the most certain structured analyses can easily be rejected by picking at individual uncertainties.Within a backcasting approach, we instead seek out multiple compelling stories that hold together as a whole and bring focus back to immediate actions that 1) are a natural first step within each story, 2) are easy to support, and 3) bring us closer to the long-term outcome.This notably builds on literature about decision-making under deep uncertainty (Marchau et al. 2019) regarding robust low-regret short-term actions.There are many ways that an effective argument for an uncertain innovation could be made -we simply present the approach that we adopted.
The approach is described in Figure 1, outlining the steps involved, and for each step, the guiding questions to be addressed and the actions taken to assess the viability of the innovation (MAR) in the case study area (the Coleambally Irrigation Area -CIA).The steps are grouped into three sections: framing, scenario development and communication (Figure 1).The framing phase scopes realistic project outcomes based on available resources, determines individuals, groups and/or organisations that need to be convinced of the viability (or not) of the innovation ('agents of change'), and defines the case study area and scale at which to assess feasibility of the innovation.In the scenario development phase, scenarios that do (and do not) employ the innovation are outlined, supported by a systematic feasibility criteria assessment providing documentation at required information levels.Feedback on the developed scenarios is sought from the agents of change and others (e.g.project steering committee) in the communication phase, before recommendations that address initial steps moving towards implementation of the innovation are provided.

Framing: catchment and case study area
The Murrumbidgee catchment was identified by the project steering committee as a region where MAR may be suitable and well received based on previous work in the area.Through stakeholder engagement in the catchment, the focus area was further defined, with the Coleambally Irrigation Area (CIA) suggested as a suitable case study area to assess the feasibility of MAR in an Australian agricultural setting.Shifting from the regional scale of the Murrumbidgee catchment to that of the CIA allowed the development of scenarios arguing for MAR that are relevant to The Murrumbidgee catchment is located in the south-east of Australia, within the MDB (Figure 2) and covers 87,348 km 2 .Rainfall in the catchment varies spatially, from over 1500 mm annually in the mountainous east to less than 400 mm annually in the west (Khan et al. 2008).Soil types in the catchment vary, with a trend of increasing clay content from east to west (Usowicz et al. 2017).The Murrumbidgee River, the major river system in the catchment, is highly regulated with several instream dams (Green et al. 2011).Agricultural production in the area is highly valued (Kandasamy et al. 2014), with irrigated agriculture representing a substantial portion of this (Khan et al. 2008).
The CIA is managed by the CICL and comprises 790 km 2 in the mid-Murrumbidgee, supplied by a main canal of the Murrumbidgee River.The CIA is the more intensively irrigated portion of the Coleambally Irrigation Area of Operations (CIAO) (Figure 2), also managed by CICL.Current water entitlements are derived from both surface water and groundwater sources (operative Limited 2019).Water availability in the Murrumbidgee catchment has generally been decreasing over the past decades as a result of drought, water reform and increased demand from higher intensity agriculture (e.g.horticulture) which is almost absent in the CIA (Schuster, Kennedy, and Holley 2020;Schenk et al. 2014).To allow this region to continue to produce substantial agricultural goods and support its population economically and socially, well-planned water management is crucial.

Scenario development: feasibility criteria
Table 1 outlines the feasibility criteria (Step 5 in Figure 1) and guiding questions adapted from Ticehurst and Curtis (2017), and the methods used to assess MAR against each criterion in the CIA.Ticehurst and Curtis (2017) assessed the amount of evidence for each criterion relative to a chosen scenario before assigning a scale to reach a feasibility score.In contrast, this study used the feasibility assessment to develop multiple practicable and fundamentally different scenarios that either employ MAR or are competing options to MAR.

Communication: stakeholder engagement
We engaged with 29 individuals during the feasibility assessment and scenario development, most of whom were irrigators, industry representatives (e.g.CICL staff) or water managers (Appendix 1).Other groups were invited (e.g.cultural leaders, urban water managers, community drought and financial representatives) but ultimately did not participate.Fifteen individuals attended one or more of three stakeholder workshops, where we collectively identified the case study area (workshop 1), discussed the feasibility of MAR against the seven criteria (workshop 2), and critiqued and refined the proposed scenarios (workshop 3).The workshops focused on active, open discussion facilitated by the research team.Each workshop built upon the last.Fourteen key informant phone interviews informed our assessment of the demand for water (criteria 1), social acceptability of MAR (criteria 6) and the importance of consistent water supply.An outline of the questions asked during the interviews can be found in Appendix 2.Where gaps were observed, the views of individuals with relevant expertise were sought.
Within the argumentation framework, the evidence sought only needed to disprove the hypothesis that MAR was not feasible, such that the validity of data collection in stakeholder engagement lay in ensuring scenarios made a convincing case to take the next steps rather than seeking a representative sample or achieving some level of statistical confidence.The use of scenarios reduced the information required as it is sufficient for one rather than all scenarios to be convincing to the reader, details beyond the immediate next steps are only needed if they need to be resolved before the next steps can be taken, and remaining uncertainties are delegated to future work -meaning that those who will be involved simply need to be confident they will be able to address those uncertainties when needed.

Scenarios for potential MAR feasibility in Coleambally & recommendations
In this section, we demonstrate the potential viability of MAR in the CIA using three scenarios and contrast these against two surface storage scenarios that meet the same objectives (Figure 3).Each scenario is further explored using the seven feasibility criteria (Tables 2-4).

Multi-year water storage: don't miss a drop
This scenario would use MAR to bank water available during wet periods to then recover during dry years at prices competitive with the water market.The water source for this scenario could be supplementary entitlements, which are announced periodically after heavy rainfall when supply is greater than other users' needs (e.g.environmental flows, domestic users, native title rights) and satisfy other entitlements (MDBA 2019; WaterNSW 2019).These entitlements can go unused due to circumstances at the time of Table 1.The seven feasibility criteria and respective key considerations/questions, and the methods used to inform the analysis adapted from Ticehurst and Curtis (2017).MDB = Murray Darling Basin, MAR = Managed aquifer recharge.

Criteria
Example questions and considerations Methods used to assess Effective demand for products • Is there demand for more water or greater water security?
• Who wants the water and when?
Gauged the local demand for change to existing water management, using stakeholder workshops and phone interviews.

Water availability
• Is water available to be banked underground (e.g. unused surface water shares, surface water traded in when prices are low)?
Reviewed water allocation, use, carryover, and entitlement data to identify potential sources of water to be banked via MAR Technical feasibility • Is there space in the aquifer systems to store surface water for drier times?
• How can the water be recharged, stored and extracted?
Considered aquifer and soil characteristics, a MAR suitability assessment previously undertaken in the area, literature review of MAR methods, and how a monitoring program could be implemented Financial viability • Can MAR be a viable and profitable alternative to surface storage schemes?
• What are the critical factors affecting financial viability?
Assessed the levelized cost of MAR using differing water sources, compared to surface storage cost Environmental risks • Are there any significant effects on water quality and quantity (positive or negative)?
• What are the consequential impacts of any change on farmland and ecosystems?
Investigated effect of banking 'unused' water on downstream environments, water quality risks and potential benefits to aquifer health associated with MAR Social acceptability • Is it a socially acceptable option to irrigators and other stakeholders?
• What are people's values, knowledge and beliefs about MAR?
• Do they perceive risks about its implementation in their region?
Elicited opinions on MAR during stakeholder workshops and phone interviews Governance arrangement • Are the legislative and policy settings appropriate to support a MAR system?
• If not, how would they need to be changed?
Reviewed legislation and policy related to MAR at the federal, state and MDB level announcement such as farm dams already being full due to preceding rain or the announcement occurring after it can influence planting decisions.While MAR would start small to allow learning by doing, the complete supplementary entitlement for the CIA is substantial (~20 GL).This volume could be taken from the Murrumbidgee River within 4 days via the main canal for the area (Shahidi, Smith, and Gillies 2012); supplementary events often last this long (WaterNSW 2020).Due to the extensive canal system in the CIA (operative Limited 2019), it would be possible to direct water to multiple MAR sites, increasing recharge capacity (Miotliński et al. 2014).However, recharging this volume of water would take time and temporary storage of water to be recharged is likely necessary.Temporary storage would also assist in sediment settling and reducing clogging at recharge sites, which would promote the maintenance of recharge efficiency (Dillon, Peter, and Muhammad, Arshad 2016).In addition to unused supplementary entitlements, water entitlements that cannot be carried over to the next water year could also be banked in this scenario (e.g.conveyance entitlements (WaterNSW n.d..)).There is a need for development of explicit MAR policy in New South Wales (NSW) (Ward and Dillon 2009) to reduce uncertainty around the recovery of water recharged via MAR (this is relevant to all proposed MAR scenarios).Table 2. Assessment of feasibility criteria for Managed aquifer recharge (MAR) scenario (Don't miss a drop) and alternate surface storage scenario (Surface storage) that both store 'unused' allocations (e.g.supplementary entitlements).More details are provided in the project report (Guillaume et al. 2020).
Scenario Don't miss a drop Multi-year water storage Effective demand for products Variable and low allocations mean that water is valuable and fully using entitlements is worthwhile Water availability During long events with wet antecedent conditions, supplementary allocations that would otherwise be forgone by members could be banked Water entitlements that cannot be carried over could also be banked Technical feasibility High flows capture is limited by main channel intake and recharge rate across multiple infiltration and injection sites Large dams, likely with high evaporation, would be needed to allow for interannual storage Financial viability Levelised cost of MAR water is competitive with market value of water in dry years Large corporations in the region have found the building of dams viable, with government subsidies Environmental risk Capture of full supplementary water allocations may affect environmental flows, and be further regulated Social acceptability Multi-stage pilot projects may resolve community concerns and minimise initial investment risk There has been public opposition to other large dams built in the region Governance arrangements Injection is subject to NSW Aquifer Intervention Policy; any project will require project assessment process or new water recovery rules Other large dams have been successfully completed and audited

Multi-year water storage: surface storage
The competing option is to direct investment to the construction of large dams to store supplementary allocations and other underutilised water entitlements.
However, there is the potential for large amounts of water loss from dams, as evaporative losses in the area have been estimated to be up to 40% of the storage volume annually (Craig et al. 2005).Although evaporative losses could be minimised using MAR, there are also associated losses, especially when a saline aquifer is targeted (Ward, Simmons, and Dillon 2007;Khan et al. 2008).Critically, dams require a large initial investment (Benjamin 2018) and, unlike MAR, are difficult to scale up over time.They are, however, an established technology with fewer legal barriers to construction than a MAR scheme.

Resilience: MAR for community sustainability
In this scenario, the source of water for MAR recharge is the banking of a portion of more reliable entitlements (e.g.general security, conveyance).The aim of  banking such entitlements is to even out the peaks and troughs in water supply noted by operative Limited (2019) and reduce the district interannual variation in planted area (Roth 2014).As the water available during 'good' water years would be reduced, it is conceivable that during these years the area planted would reduce but be compensated for in dry years when the banked water is extracted and used.From a community perspective, this scenario could stabilise expenditure for local businesses, including cotton gins, thus supporting local townships through dry spells.Altering how current water allocations are used, as well as increasing dependence on (recharged) groundwater resources would need to be a gradual transition.This is to achieve required changes to the policy and regulation surrounding water entitlements, the mode of operation by CICL and, critically, a shift in mindset at all levels, from the government to the community to individuals.The first step would be to gain a better understanding of the community's need for water in dry years, and the values that recovered MAR water would support.Governments at all levels want to 'drought proof' regional areas (Department of Agriculture Water and the Environment 2020; Alexandra 2018), so co-investment of MAR would be an avenue to peruse under this scenario.

Resilience: investigate increasing community sustainability with high security entitlements or policy change
Here two scenarios without MAR are explored as possible means of increasing community sustainability, one which changes how high security entitlements are used, and the other changes water allocation policy during drought.They both present problems which are discussed below.High security water entitlements can provide water security during dry years, albeit at high cost (Seidl, Wheeler, and A 2020).Such entitlements are currently free to be traded on the water market and therefore are not specifically allocated to support community sustainability.High security water is also limited in quantity and is often purchased to water high value perennial crops (Khan et al. 2008).Lastly, high security water allocations are controlled at a regional scale (Khan et al. 2008), rather than by the irrigation cooperative which is the level of control targeted in the previous MAR scenario.
In the broader debate over water management in the MDB, policy reform has been proposed to change how water is allocated during droughts (Alexandra 2018).It has been suggested that a portion of environmental allocations could be sold on the water market during severe drought, providing income for the Commonwealth and lessening water stress of famers and the community (Davey, Drum, and Webster 2019).As the provision of survival water for drought depends solely on surface water, it faces many of the same difficulties associated with high security entitlements as a means of increasing community sustainability, including limits to availability based on dam storage and high costs seen on the water market.

Conjunctive use: integrated groundwater and surface water delivery
Groundwater use in the CIA is widespread, with ~38 GL extracted in the 2019/20 water year (operative Limited 2020).MAR could be a mechanism to further integrate the management of groundwater and surface water, with the aim of achieving 'conjunctive' use (Ticehurst and Curtis 2017).A reduced reliance on limited surface waters, which are associated with high evaporation rates (Craig et al. 2005), could be supported during dry years by extracting groundwater closer to where water is required.To replenish the aquifer, this would be offset during wet periods with reduced groundwater use, operation of the MAR scheme, as well as with conveyance water redeployed from channels in dry years.
Conjunctive water use is currently an opportunistic farm-scale decision (Evans, Evans, and Holland 2016).The CICL could facilitate collective decision-making on which form of water is most efficient for supply to an area at any given time.Such a conjunctive use system requires detailed monitoring of both surface water and groundwater systems (Ticehurst and Curtis 2017;Evans, Evans, and Holland 2016) and a key requirement of this scenario would be near-real-time operational monitoring to improve planning and coordination of surface water delivery and groundwater pumping.Currently, monitoring and reporting of groundwater in CIA is done at a coarse temporal scale (operative Limited 2019).

Recommendations
Each of the developed MAR scenarios (Figure 3) leads to recommendations tailored to the two targeted agents of change: CICL and CRDC (Figure 4).The recommendations are linked but framed differently for CICL as the potential local driver of, or investor in MAR, and CRDC as a potential facilitator or supporter who can influence or leverage other activities in this space.It should be noted that the recommendations outlined here for the CRDC could be applicable to other research corporations/investors who were looking to support MAR as an innovation.The recommendations emphasise that a partnership-based approach is necessary to achieve the potential industry wide benefits, with Coleambally acting as a pilot that could be applied to irrigated agriculture elsewhere in Australia.The recommendations also have flow on effects beyond the agricultural industry, for example, into the communities that surround agriculturally established areas.All the recommendations are 'no loss', i.e. applying the recommendations would result in advances in knowledge, capability, and relationships regardless of whether MAR is implemented in the long term and provide sufficient flexibility that there are many ways that the recommendations can be pursued.For example, low-cost pilot projects gain understanding about the entire system and improved groundwater monitoring could assist in developing conjunctive use, with or without MAR.The developed knowledge, capability, and relationships could also be leveraged by other agricultural areas interested in MAR.The recommendations should be pursued in a staged manner to minimise the need for large initial investment and access to resources.
One of the recommendations is to establish a MAR pilot in the CIA, for which there is strong local support.The costs of a pilot would be kept low by retrofitting current bores or small-scale infiltration basins.A pilot would provide invaluable practical experience in the planning, construction and operation of a MAR scheme in an agricultural setting in Australia.The pilot would also improve local hydrogeology understanding and could be used to catalyse the design of policy frameworks for MAR in NSW (e.g.water rights for MAR, recovery of MAR water and use of recovered water).Collaboration with state government (e.g.NSW Department of Planning and Environment and WaterNSW) and the Murray -Darling Basin Authority (MDBA) would be required for policy development.Ultimately, a well planned and executed pilot would enable irrigation companies like CICL to evaluate the true feasibility and value of MAR, research and development organisations to better direct future investment in MAR if appropriate, and policymakers to have greater confidence when establishing regulation around MAR.

Discussion
This research developed and applied an argumentation approach to tackle the uncertainty that plagues the traditional assessment of innovations.The Recommendations for CICL  development of scenarios was central to demonstrate potential innovation and feasibility of MAR in the CIA and to establish 'no loss' recommendations for staged progress towards MAR.
The notable work of Khan et al. (2008) identified five sites in the Murrumbidgee region that were associated with areas of aquifer extraction without adequate replenishment, leading to reduced groundwater pressures, which they concluded were hydrologically suitable for MAR.The authors found that MAR implementation was economically feasible given the higher costs to develop new surface storages and associated evaporative losses.Our work complements previous approaches to MAR research of Khan et al. (2008) by providing a holistic assessment of MAR across the seven feasibility criteria and focusing on a scale of analysis that supports engagement with local stakeholders and specifically addresses the needs of the agents of change that would champion the next phase of MAR assessment.Although MAR is the focus innovation in this paper, our argumentation approach (first column in Figure 1) could be adapted to other innovations, with the necessary modifications to the guiding questions (second column in Figure 1) and methods (last column in Figure 1 and Table 1).
A substantial area of uncertainty identified in this case study involves the governance arrangements for MAR.The differentiation of water recharged via MAR from native groundwater supplies, and how entitlements to these would operate, is not currently addressed in NSW policy (Ward and Dillon 2009), though project-based approvals for pilot studies would be possible.The MDBA, as well as other Australian states, do address MAR in policy and legislation (MDBA 2019;Thomson 2008;SA EPA 2004;Newland 2015), and there are Australian guidelines (NRMMC, EPHC & NHMRC 2009), providing confidence that the necessary regulatory framework could be established if continued work shows that initial MAR pilots should be scaled up.Policy recognition of MAR in NSW would be of benefit beyond the CICL.In NSW, it appears past fears around and resistance to MAR (Dillon et al. 2020) are beginning to be reconsidered (NSW Department of Planning and Environment 2022).
Another area of uncertainty that pertains to all MAR scenarios in this location, from a technical feasibility standpoint, is salinity.In the early 1990s, the district experienced rising water tables and increasing soil salinity (Khan et al. 2008).This was remedied by a united effort from both the government and the community to improve land and water management (Freak et al. 2022).There are MAR approaches that could be taken to avoid the problems of the past including, targeting low salinity areas of the shallow aquifer via infiltration basins, or targeting low salinity deep aquifers via injection wells (Khan et al. 2008;Christen, Prasad, and Khan 2001).A key component of a MAR scheme in an area at risk of rising water tables and salinity is monitoring.CICL is experienced in implementing long-term monitoring programs, having hundreds of piezometers already operating in the Coleambally Irrigation Area of Operations (operative Limited 2021).
The scenario development phase proved effective for sharing knowledge between local stakeholders and the research team, in particular, enabling local experience to contribute to the arguments or assumptions underpinning the scenarios.At a high level, MAR was viewed by the stakeholder workshop participants as an opportunity to increase the region's water storage capability.Concerns raised in the third workshop were used to clarify and refine each scenario, notably about the feasibility of capturing large supplementary water events, the need for clearer distinctions between the MAR and alternate scenarios, acknowledgement that trade on the water market should not be hindered and need for improved definition of community benefits.The workshops highlighted which MAR scenarios were of greatest interest and, contrary to the research team's expectation, the integrated groundwater and surface water delivery scenario was thought to hold the most promise.The community resilience MAR scenario was considered more difficult, requiring considerable changes to policy and CICL mode of operation, and therefore would need to be a long-term objective.The process of co-design and staged workshops introduced the stakeholders to the innovation early, which developed familiarity and confidence with the innovation and trust in the bona fides of the research team.It created an inclusive space for considered input from the local stakeholders and meant that there was no surprise when scenarios were introduced.Therefore, when implementing innovative technology consider a collaborative approach to codesign so as to not 'leave behind' the individuals and groups the innovation would impact (e.g.Klerkx et al. (2012)).
The resources available to this project shaped the approach (and depth) taken in the feasibility analysis, and this was explicitly acknowledged in the first step of the argumentation approach.Low resources are not an excuse not to act as first steps can be taken even in the face of great uncertainty.As climate change progresses, the risks of inaction are likely to rise as opportunities for MAR in wet years are missed.The lack of identified sites that are most suitable for MAR development in the area could also be viewed as a limitation; however, unlike more traditional MAR feasibility analysis (e.g.Khan et al. (2008), this is treated here as a secondary concern to gaining confidence that MAR could be feasible somewhere in the CIA.Another factor that shaped the approach was the research team, comprised of university researchers working alongside one agent of change (i.e.CICL).University researchers bring a different perspective to a problem than a consultant (or group of consultants) would.Where a consultant would likely work in a linear way, ultimately disengaging when the project is finished (Baskerville and Wood-Harper 1996), the research team here have continued to engage parties interested in MAR in the region.Ongoing engagement has been suggested to build community trust and support the uptake of innovations in the water space (Sharma et al. 2012;Thorne et al. 2018).
True impact of this analysis approach in the CIA, and broader irrigated agriculture industry, will stem from the uptake and realisation of the case study recommendations by the CICL and CRDC (and/or other research investors) (Figure 4).This would demonstrate that the large amount of uncertainty associated with innovative water management strategies such as MAR can be overcome.We demonstrated the value of the argumentation approach to research in an established irrigation district where there is undeniable need and demand across stakeholders to change how water is used and managed into the future.Continuing application of this approach in other case study contexts will help refine guidance on its implementation and understanding of its strengths and weaknesses.

Conclusion
The assessment of an innovation is plagued by large amounts of uncertainty that raises the likelihood of its abandonment.The use of an argumentation approach, supplemented with scenario development, in this research outlined situations in which MAR could be argued as feasible to provide water for irrigation to the CIA.Alternative scenarios using more established technology (i.e.surface storage) were also explored to put the innovations in context.Using a holistic set of feasibility criteria -demand for water, water availability, technical feasibility, financial viability, environmental risk, social acceptability and governance arrangements -allowed for a systematic evaluation and comparison of the MAR and alternative scenarios.'No loss' recommendations were then outlined to the agents of change based on the developed scenarios.Overall, this approach was successful in focusing on what we do know about an otherwise uncertain innovation by requiring MAR to be proved not feasible.

Figure 1 .
Figure 1.The argumentation approach developed for this study, with associated guiding questions and how this approach was applied in Coleambally Irrigation Area of Operations (CIA).CICL = Coleambally Irrigation Co-operative Limited, CRDC = Cotton Research & Development Corporation, MAR = managed aquifer recharge.

Figure 2 .
Figure 2. The Coleambally Irrigation Area and surrounds.

Figure 3 .
Figure 3. Potential of MAR building on scenarios within an argumentation approach.Managed aquifer recharge (MAR) scenarios (blue boxes) and alternate scenarios (grey boxes) with arguments for MAR in green.Section numbers are shown in square brackets.

Figure 4 .
Figure 4. Recommendations for initial steps towards implementation of manage aquifer recharge (MAR) in Coleambally Irrigation Area of Operations.Recommendations are provided for the Coleambally Irrigation Co-operative Limited (CICL) and Cotton Research & Development Corporation (CRDC).MDBA = Murray Darling Basin Authority, R&D = research and development.Although the recommendations here are aimed at the two case study agents of change (CICL and CRDC), they could be applicable to other MAR innovators, both in a local or supporting role.

Table 3 .
(Guillaume et al. 2020)e (MAR) scenario(MAR community sustainability)and alternate surface storage scenarios that store current entitlements or require policy change.More details are provided in the project report(Guillaume et al. 2020).While cotton is an opportunistic crop, evening out peaks and troughs could allow greater investment in supply chains, infrastructure, justify cost of pickers, and take-up of forward contracts, allowing farmers to farm and the community to endure extended dry

Table 4 .
(Guillaume et al. 2020)e (MAR) scenario that stores water to facilitate conjunctive water use.More details are provided in the project report(Guillaume et al. 2020).Possible now with approval through NSW Aquifer Intervention Policy, but decisions on groundwater use restrictions would need to account for drawdown to guarantee access in drought periods as well as clarifying rights to surface water post infiltration