Bringing resilience-thinking into water governance : Two illustrative case studies from South Africa and Cambodia

Resilience is a multidimensional concept that is increasingly used to understand environmental change in hydrological systems. Yet, the current discussion about water governance and resilience remains relatively limited, with resilience typically seen as a normative outcome for governance (i.e., to be resilient against change). Using a theoretical multiplicity approach, we explore how the theories of social-ecological systems (SES), resilience and interactive (water) governance can provide new insights for water governance studies. We propose a resilience – governance framework that captures the partly overlapping but distinct characteristics from these three theories. The framework aims to develop a more nuanced way of using resilience-thinking for water governance, viewing resilience as a function of three capacities (absorptive, adaptive and transformative capacity) and noting the simultaneous existence of three interpretations for resilience (as a property, process and outcome) across different scales. The framework also considers issues of power and equity, which are often missing from resilience framings. We illustrate the framework with two case studies – the Tonle Sap Lake in Cambodia and a small sub-catchment of the Limpopo River Basin in South Africa – to provide two distinct examples of the possibilities of resilient governance. Finally, we consider what the framework suggests more broadly for ongoing discussions around resilience and water governance, including the possibilities for governance to also ‘bounce forward ’ – i.e., transform – to a new, improved state. We argue that resilience-thinking may be valuable in understanding governance characteristics and guiding governance processes, in addition to seeing resilience (just) as a normative end-goal. In this way, the article supports an epistemological shift away from focusing on institutional structure, towards capturing the dynamic processes within governing systems.


Introduction
Governing water resources in a rapidly changing world increasingly requires a reimagining of water systems as complex systems in a way that captures both social and hydrological dynamics.Increasing awareness of complexity in social-hydrological systems has emphasised the need for a more nuanced understanding of change within governance responses to preserve and maintain key functions while adapting or transformingthat is, for them to be resilient.This is no simple task, however, given that water resources management has historically focused largely on technical solutions to well-defined problems, with assumptions of predictability, efficiency and control (Pahl-Wostl, 2015;Rockström et al., 2014;Rodina, 2019;Zeitoun et al., 2016).While human dimensions of complex, multi-causal water problems more broadly have been increasingly recognised (Pahl-Wostl, 2015), attention to social and political dimensions of water resilience remain limited, despite governance being widely understood as a core component in addressing water resilience issues (Rodina, 2019;Rodina et al., 2017).
Water governance is, broadly, the political, social, economic and administrative elementsconsisting of institutions, actors and their interactionsthat regulate, influence, and shape water and other related natural resource use and management, with the aim of moving towards a desirable state (Pahl-Wostl, 2015).Such an aim provides a logical link with resilience, which can be defined as the capacity of social and ecological systems to cope, adapt or transform in the face of change while maintaining core functions and identity (e.g., Folke et al., 2010;Holling, 2001).At the same time, there is an increasing understanding that resilience is not just a system characteristic, but can also be considered as both a learning process and an outcome (Moser et al., 2019).Resilience can thus be used as a lens through which to explore water and environmental-related stressors and their impacts on society, particularly for those whose livelihoods depend upon local ecosystems (Dharmawan and Nissa, 2020;Marschke and Berkes, 2006;Ross and Berkes, 2014).Growing interest in resilience goes hand-in-hand with alternatives that are emerging for traditional command-and-control water governance arrangements, such as adaptive governance (Huitema et al., 2009).
The growing use of resilience-thinking, both implicitly and explicitly, in water studies also brings challenges and unanswered questions that this article addresses.Firstly, resilience typically forms an end-goal for water governance (for example, to be resilient against a changing world), particularly in terms of maintaining integrity and functioning of human-built infrastructure such as dams.Resilience theory's potential contribution for understanding the characteristics and processes within governance arrangements themselves remain much less studied, and dominant water governance discourses such as Integrated Water Resources Management (IWRM) do not always explicitly consider issues of complexity, cross-scale interactions and nonlinear dynamics, for which resilience theory may be beneficial.Secondly, resilience theory itself risks missing issues of politics, power and equity that form the core of how governance operates.Normative interpretations of resilience have attempted to close this gap, though they still struggle with the nuanced consequences of interventions on different social groups (Barrett and Constas, 2014;Béné et al., 2012;Cote and Nightingale, 2012;Dewulf et al., 2019;Sendzimir, 2006).A system may also be resilient in an undesirable state, and even then, the definition of what is 'desirable' depends upon perspective and is typically determined by those with the loudest voice (Béné et al., 2012;Cinner and Barnes, 2019).Such considerations have led to promising, critical queries of resilience, considering the questions 'resilience for whom, and to what?When, where and why?' (Dewulf et al., 2019;Lebel et al., 2006;Meerow et al., 2016).Thirdly, the use of resilience-thinking in governance remains primarily focused on a system's ability to 'bounce back', rather than the possibility of 'bouncing forward'.Thus, the more transformative aspects of resilience in water governance seem to be partly overlooked, or at least underdeveloped.This is a particularly important consideration in combination with the previous issues raised regarding power, justice and equityfor example, a resilient institution may be one that perpetuates inequalities (Cleaver and Whaley, 2018;Ensor et al., 2021).
While resilience helps us recognise the complex realities of environmental change in interconnected social-hydrological systems, its potential in developing more resilient forms of water governance is not fully realised.There remains, therefore, a significant gap in the scholarship of water governance and resiliencenamely, how do the two concepts complement or counter one another?And in more practical terms: how could resilience-thinking help in understanding complex social-hydrological systems experiencing uncertainty and change and how we govern such change?This is particularly beneficial when a multitude of otherwise disconnected research fields are working on the dynamics of complex systems (Grove, 2018).Finally, how can issues of politics, power and equity be brought to the forefront of such analyses?
This article aims to answer these questions by developing a theoretical resilience-governance framework for understanding social--hydrological systems and how they and their governing systems respond to multiple, cross-scale social and environmental pressures.To do this, we simultaneously consider theories related to resilience, socialecological systems (SES) and interactive (water) governance, with the help of theoretical multiplicity (Karpouzoglou et al., 2016).We then illustrate the framework with two case studies from South Africa and Cambodia.These cases are used to discuss how a power-sensitive approach to resilience and governance may help reframe the existing water governance arrangements, providing new ways forward.Finally, we consider what our framework and the findings from the two cases suggest more broadly for the ongoing discussion of both resilience and water governance.

Methodology
Our study includes three main methodological steps that are closely connected: (i) a structured analysis of key theories related to socialecological systems, resilience and interactive (water) governance; (ii) development of the resilience-governance framework; and (iii) an illustration of the framework with two case studies.
We apply qualitative methods for all three steps.For the first step, we use a theoretical multiplicity approach to explore how theories overlap and inform each other while retaining the distinctiveness of the theories (Karpouzoglou et al., 2016).Such an approach is a suitable choice for this study, given that dealing with complex sustainability issues typically requires the simultaneous application of multiple theories and epistemologies (Karpouzoglou et al., 2016;Termeer and Dewulf, 2012).In our analysis, we look at the theories of social-ecological systems (SES), resilience and interactive (water) governance, recognising first their distinct characteristics and then focusing on their theoretical linkages and, importantly, the areas in which one theory may fill a conceptual gap in another (Sections 3.1-3.3).Building on the first step, we then propose a 'resilience-governance framework' that helps to bring the key perspectives and categorisations of different theories together in a structured manner, using a synthesising table and visualisation as the key elements of this second step (Section 4).
In the third step, we apply a basic case study method (Yin, 2014) to illustrate the proposed framework.To better understand the potential added value of the framework, we use two case studiesthe Tonle Sap Lake in Cambodia and the Doringlaagte basin within the Limpopo River Basin in South Africathat we authors have already studied before (e.g., Fallon et al., 2019;Jones and Sok, 2015;Keskinen, 2006;Keskinen and Varis, 2012;Nuorteva et al., 2010).Two distinct cases facilitate the recognition of both the similarities and differences provided by the framework as applied to two distinct contexts.The first case, Cambodia's unique Tonle Sap Lake, is a dynamic hydrological system with a multiscale governance arrangement and extensive documentation of environmental change due to human activity.We analyse the Tonle Sap case with the help of key scientific literature, policy documents and other secondary data, as well as the authors' own extensive experience in the region.In South Africa, the case focuses on a small aquifer within the Hout catchment within the Limpopo River Basin (the 'Doringlaagte'), which is experiencing groundwater depletion due to irrigated agriculture and climate variability and has a mixed governance arrangement of ineffective hierarchical governance combined with informal self-governance.The analysis of the Doringlaagte case builds on the first author's prior experience and interview data collected in the region, complemented by relevant scientific literature and policy documents.

Theories of social-ecological systems, resilience and governance
In this section, we synthesise key aspects of the theories of socialecological systems (SES), resilience, and interactive (water) governance, considering the linkages and discrepancies between the theories as relevant for studying water systems.In the following sections, we then combine the relevant perspectives and categorisations from these theories with the help of theoretical multiplicity to establish a resilience-governance framework and illustrate its use in the contexts of our two case studies.

Social-ecological systems (SES) theory
The embeddedness of water-related systems with society highlights the usefulness of studying them as interconnected systems as advocated by the social-ecological systems (SES) frameworki.e., as interacting and interrelated elements that form a whole, as described by its boundaries, structure and purpose.SES theory was developed as an A. Fallon et al. analytical framework to study intertwined social and natural systems, and thereafter to study the resilience of such systems (Colding and Barthel, 2019), which in itself is a huge body of literature that considers human-environment relations as interconnected and interdependent (Folke, 2006;Holling, 1973).Fundamentally, SES theory assumes that humans make conscious choices (both individually and collectively) that may significantly alter natural resource management outcomes (McGinnis and Ostrom, 2014).
In this article, we use SES theory as a basis for studying water systems.Hydrological systems are often complex and characterised by seasonality, nonlinearity and changing infrastructure, while situated within institutionally challenging contexts and mutually shaped by human activity.SES-thinking captures this complexity of hydrological systems as broader ecological systems (for example, the life-supporting role of rivers), and their close, dynamic relationship with humans (see also Wesselink et al., 2016;Gain et al., 2021).
There are several aspects of SES-thinking relevant to understanding hydrological systems (Gain et al., 2021).First, an SES is the sum of many interacting components which act both individually and in reaction to one another (Sendzimir, 2006;Walker et al., 2004).They also consist of a hierarchy of nested adaptive cycles across temporal and spatial scales, sometimes described as 'panarchy' (Gunderson and Holling, 2001;Holling, 2001).Due to their multiple components and scales, SES are typically both complex and adaptive, which makes their governance particularly challenging.Preiser et al. (2018) capture these characteristics through the following six principles of complex adaptive systems: (1) the diverse nature and structure of relationships between components in a system; (2) systems adapt and change, and this should be anticipated, even if unpredictable; (3) systems are dynamic and deeply uncertain, and their behaviour is affected by feedback loops (which may lead to tipping points, regime shifts or feedback structures); (4) systems are radically 'open' and embedded in a larger system, with boundaries that do not separate but, rather, intimately connect with their environment; (5) system components have multiple functions that change with context; and (6) systems have complex causalityinteractions are not linear or isolated causal trajectories.
However, SES theory cannot alone provide a holistic analytical framework for water systems and their governance.Due to its ecological origins, SES theory has struggled with core differences between social and ecological systems and how they can be analysed, which has implications for how we use SES-thinking in social-hydrological systems.For example, many SES-informed studies have been critiqued for overly focusing on ecological principles to analyse social dynamics within an SES (Cote and Nightingale, 2012), undermining the importance of more socially grounded questions such as the motivations and social institutions that operate outside of the immediate human-environment connection, though progress has been made in this arena (e.g., Hertz and Schlüter's (2015) use of the SES framework as a boundary object between disciplines).Such critiques have thus carried over into the SESinformed socio-hydrology due to its tendency to downplay social dynamics of environmental change and difficulty in dealing with human values, power relations and agency (Wesselink et al., 2016).This highlights the challenges that come with more the integration of more constructivist interpretations of water and its diverse meanings to people.Insights from the social sciences, such as social anthropology and political ecology are, however, increasingly used by scholars to further integrate issues such as power and social diversity into SES-informed analyses (see Armitage et al., 2012;Cote and Nightingale, 2012;Fabinyi et al., 2014).
With such shortcomings in mind, we also draw upon the concept of hydrosociology, which similarly relates water and society, and is rooted in the social sciences (particularly political ecology) and thus focuses on power relations, historical contexts and narratives through rich case study analyses (Boelens et al., 2016;Linton and Budds, 2014;Wesselink et al., 2016).However, this discipline has its own critiquesnamely, its tendency to over-focus on the social and political aspects of water use and management and constructivist framings of environmental issues.We therefore draw from both of these SES-related approaches, using the term 'social-hydrological systems' when describing water systems through an SES lens (for more discussion on hydrosociology and sociohydrology, see e.g.Budds et al., 2014;Linton and Budds, 2014;Sivakumar, 2012;Sivapalan et al., 2012;Wesselink et al., 2016).

Resilience theory
The concept of resilience is useful for navigating complexity and change in human-environment relations (Lebel et al., 2006;Olsson et al., 2015), and more specifically, water systems (Rockström et al., 2014;Rodina, 2019;Rodina et al., 2017).Resilience is the paradoxical and elastic condition of staying the same while simultaneously changing in response to disturbanceseither 'bouncing back' to a previous state, or 'bouncing forward' to a new and improved state with new system structures, functions and identity (also called a resilience regime) (Grove, 2018).SES resilience is broadly defined as the capacity of a social-ecological system to absorb, cope with, and adjust to changing social or environmental conditions while retaining key functions, structures and identity, including people's capacity to learn and engage in shaping transformative change (Dewulf et al., 2019;Stockholm Environmental Institute and Stockholm Resilience Center, 2016).
Although there are many perspectives on resilience 1 , we focus on three core capacities that go beyond its typical understanding as a conservative concept: absorptive, adaptive and transformative capacity (e.g., Béné et al., 2017Béné et al., , 2012;;Béné and Doyen, 2018;Folke et al., 2010;Matyas and Pelling, 2015;Walker et al., 2004).This conceptualisation of resilience embraces a system's ability to not only bounce back and persist (absorptive capacity), but also to adapt and transform into new system states (Béné and Doyen, 2018;Matyas and Pelling, 2015).
The first dimension, absorptive capacity, refers to a system's ability to absorb change, stressors or shocks while retaining its key functions, and is the most understood component of resilience.The general outcome of absorptive capacity is stability and persistence of the system components involved (Holling, 2001(Holling, , 1973;;Walker et al., 2004).Second, adaptive capacity is the ability of a system to incrementally adjust to change through reorganisation and learningalso referred to as a 'transition' (Foxon et al., 2009;Matyas and Pelling, 2015;Pelling, 2011).In terms of governance, transition management is an intermediary form of adaptation seeking to fully realise rights under existing political and governance systems (Pelling, 2011).Finally, transformative capacity refers to a system's ability to fundamentally shift into a new regime, or a 'new normal'.In such a situation, the current system reaches a critical threshold or tipping point, beyond which it may become impossible to return to its previous state.Transformation may be sudden in response to a shock or occur over a longer time scale in response to slow-changing variables and feedbacks.Transformation is an important component of resilience in complex natural systems, which can rarely return to the exact previous system configuration.In terms of social systems, transformation is usually associated with technological innovation, institutional reform, and deep cultural and behavioural shifts which challenge the status quo (O'Brien, 2012;Pelling, 2011).
In the context of water systems, resilience is commonly viewed as an outcome or goal (e.g., to build resilience of water infrastructure to climate change).However, Moser et al. (2019) identify two further 1 Resilience's long history has been well-documented across disciplines (e.g.see Grove, 2018;Martin-Breen and Anderies, 2011;Sendzimir, 2006).Though rooted in engineering, ecology and psychology, more normative interpretations are increasingly used in the social sciences domain.Many researchers adopt more than one framing (e.g., Matyas and Pelling, 2015).Resilience is increasingly used as a transdisciplinary 'hybrid' concept, blending descriptive and normative uses, which Brand and Jax (2007) argue promotes communication and serves as a bridging or boundary object across disciplines and sectors.
A. Fallon et al. interpretations that we also consider: resilience as a trait (i.e., system property), and as a process.Viewing resilience as a trait stems from its ecological and engineering roots, referring to system properties that we can typically measure via indicators, such as the biodiversity of a lake.Resilience as a process focuses on (governance) actions or interventions, indicating that resilience is not only about what a system has but also what it does.This is central to our resilience-governance framework (Section 4), as it allows us to see where a governing system may not be adequately responsive to change, and to focus on the processual dynamics of a system, including the historical development of institutions.Finally, resilience can be interpreted as an emergent outcome from a diversity of cross-scale interactionsparticularly normative (i.e., valuebased) perspectives common in social sciences by researchers focused on temporary states of a system.This perspective also links to how agendas are shaped, trade-offs, and whose perspective on resilience is prioritised.
As with resilience theory more broadly (e.g., see Béné et al., 2012;Dewulf et al., 2019;Evans and Reid, 2013;Grove, 2018;Pelling, 2011), the use of resilience in (water) governance risks missing issues of politics, power and equity.This may be partly due to the fact that resilience theory does not easily translate into governance scholarship because of its engineering and ecological origins; social systems are inherently different to ecological systems, and we thus cannot use a 'copy-paste' mentality for them.While resilience scholars are developing social scientific framings of resilience that incorporate issues of power and politics (such as the concept of 'social resilience'), their practical implementation remains challenging (Béné et al., 2012).This is largely due to resilience theory being developed in isolation from critical social science understandings of the human dimensions of environmental change (Cote and Nightingale, 2012).However, some progress has been made recently in this realm, such as Dewulf et al.'s (2019) consideration of power and the political choices that influence how resilience definitions are negotiated and contested in water studies.

Interactive (water) governance theory
In recent natural resource governance theoretical discussions, there has been a focus on interactive governance as a response to an increasing awareness of social and ecological complexity; interactive governance therefore forms part of our theoretical basis for analysing complex social-hydrological systems.Interactive governance emphasises the dynamic interactions between a plurality of actors (state, market and civil society), and the associated outcomes in terms of a system's 'governability', or quality of governance (Kooiman and Bavinck, 2013).The interactive governance framework (IGF) has three core components: a system-to-be-governed, a governing system, and the governing interactions mediating the two (Jentoft et al., 2015;Kooiman et al., 2008Kooiman et al., , 2005;;Kooiman and Jentoft, 2009).
The IGF has many explicit theoretical overlaps with theories of both SES and resilience; for example, the IGF focuses on an interconnected governing system and system-to-be-governed, which broadly correlate with the social and ecological systems of SES (see Section 4), and highlights features of complexity, cross-scale interactions, dynamics and diversity in both systemsall of which are key features in SES and resilience-thinking.However, the IGF goes further than a descriptive tool for interrelated systems and determines how these features relate to the governability of a water or aquatic system (e.g., a small-scale fishery).Governability is defined as "the balanc[e] between the capacity of the governing system and the needs of the system-to-be-governed, with governance interactions playing an intermediary role" (Kooiman et al., 2008, p.5). Generally speaking, the more complex a system, the less governable it is.Importantly, the IGF also applies these criteria to the governing system itself (e.g., looking at stakeholder diversity or the complexity of institutional arrangements) and uses this information to judge the 'goodness of fit' between the system-to-be-governed and the governing system.
The IGF also applies useful categorisations related to the 'orders', 'modes' and 'elements' of governance.Among the three governance 'orders', first-order governance covers the day-to-day activities between people and organisations, and their efforts in identifying and solving problems; second-order governance includes the formal and informal institutional arrangements; and third-order governance (also known as meta governance) consists of the cultural values, principles and norms that shape how people think and act (Kooiman and Jentoft, 2009).The three governance 'modes' then describe the dominant institutional structure(s) of the system: hierarchical governance (state-led, top-down formal institutions); co-governance (hybrid of state, private sector and civil society); and self-governance (local-level ownership).Finally, the three governance 'elements' are the images, instruments and actions that take place throughout the governing system.For example, an image (such as the perception of water as an economic good) may influence instruments used (e.g., regulations on water pricing) to direct actions taken (e.g., less water used by certain actors due to cost).Governing interactions are also an important part of the IGF in identifying not only participatory processes, information sharing and learning, but also power relations and potential conflicts.When paired with information on the system-to-be-governed, this assessment can also determine the goodness of fit of the governing system.

Water governance
Due to our focus on water, we also consider the key theoretical aspects of water governance that complement interactive governance theory to form what we term 'interactive (water) governance'.Theories of water governance aim to describe in a systematic manner the diverse ways through which we as humans use, manage and develop water and related resources.This includes not only formal arrangements, such as laws and policies, but also informal ones, such as social norms and cultural values.There are multiple views on how this happens, with different publications emphasising, for example, power and politics (Zwarteveen et al., 2017), systems (e.g., Rogers and Hall, 2003;UN, 2002), functions and attributes (e.g., Jiménez et al., 2020), as well as actual governance outcomes (e.g., OECD, 2015).As such, the theories have close linkages to general theories of governance and to the theories related to environmental and natural resources governance.In the context of this article, we draw on the definition by Pahl-Wostl (2015), noting that water governance can be defined as the political, social, economic and administrative elementsconsisting of institutions, actors and their interactionsthat regulate, influence, and shape water resource use and management, with an aim to guide the resource towards a desirable state and away from an undesirable state.
The aim to move away from an undesirable state is particularly important given the increasing water quantity and quality problems seen globally, as well as pressures placed on water resources by trends such as climate change, pollution and urbanisation.Water governance must thus increasingly tackle challenges posed by complexity, uncertainty and global environmental change (Boltz et al., 2019;Galaz, 2007;Zeitoun et al., 2016).A key difficulty in governing complex systems undergoing change, however, is that governance cannot 'intervene' from outside the system, but rather is embedded within the system itself; governance arrangements work in a constant process of self-reflexivity and entanglement (Chandler, 2014;Chandler et al., 2020).They, like the system they govern, are complex and adaptive, with the social system being embedded within the ecological system.

Table 1
Selected perspectives and categorisations on social and environmental change among the three theories of resilience, social-ecological systems and interactive (water) governance (drawing on literature referred to in main text).Notes: a Here, we use the term 'hydro/ecological' to indicate the ecological system, with a focus on the hydrological cycle.
A. Fallon et al.
While resilience-informed water governance literature is emerging, it remains fragmented and draws predominantly on limited, engineering-oriented understandings of resilience (Rodina, 2019;Rodina et al., 2017).In addition, dominant global water discourses such as Integrated Water Resources Management (IWRM), do not fully recognise the importance of environmental characteristics such as system complexity, uncertainty and nonlinearity (Galaz, 2007).While resilience is often not explicitly mentioned in water governance theory beyond a normative end-goal, several governance approaches have emerged to challenge traditional command-and-control approaches, both implicitly and explicitly incorporating ideas from SES and resilience theories.One such approach is adaptive water governance and management, which focuses on learning to live with change and uncertainty through ongoing iterative management interventions for environmental change.It fundamentally views human and natural systems as intrinsically coupled, with resilience as a central, desirable attribute (Cleaver and Whaley, 2018).Experimentation, participation, diversity, cross-scale management, and polycentricityall often described as key features of resilienceare explored through such adaptive approaches (Biggs et al., 2015;Cosens and Williams, 2012;Huitema et al., 2009).This increased focus on adaptability and flexibility has also been applied to water law and legal frameworks more specifically (Cosens et al., 2021;Hill Clarvis et al., 2014).Meanwhile, concepts of polycentricity and bricolage emphasise the need for modular governance arrangements to increase resilience to water-related stressors (e.g., via redundancy), and to deal with uncertainty and how power shapes outcomes within such governance arrangements (Cosens and Williams, 2012;Morrison et al., 2019;Pahl-Wostl et al., 2012).

Establishing a resilience-governance framework
Building on the theories of social-ecological systems (SES), resilience and interactive (water) governance, we propose a resilience-governance framework that seeks to combine the relevant perspectives and categorisations from these three theories into one systematic framework.Importantly, this framework is not purely technical, and seeks to raise several value-based and context-specific questions through its usage.
Table 1 synthesises the key overlaps and differences in how the theories of resilience, SES and interactive (water) governance deal with issues around change and complexity.As shown, resilience and SES theory have many overlaps, particularly in their framings of change and a 'systems' approach; this is unsurprising, given the SES framework was primarily built to analyse resilience in complex systems (Folke, 2006).
Drawing on the previous section, we identify three key issues to be considered while conducting a resilience-governance analysis.First, resilience is not inherently 'good' or 'bad', and thus undesirable resilience in a system must also be recognisedfor example, an autocratic government regime resistant to change (Cinner and Barnes 2019).Second, resilience depends upon perspective; is a person more resilient because they can migrate away from a hazardous environment, or are they forced to do so due to low absorptive capacity?Whose perspective counts?(Marschke and Berkes 2006;Bornstein 2013;Keskinen et al., 2013;Ensor et al. 2021).Here, agency to act is central.Third, resilience of one social group or scale may come at the expense of another, and thus the power-sensitive question 'resilience for whom?' must be considered (Béné et al. 2012;Dewulf et al. 2019).Importantly, attention must be paid to whether a marginalised group is truly uplifted by a resilience-focused intervention, or merely burdened with the responsibility of adapting to externally produced threats (and potentially romanticised for indigenous resilience), without the interrogation of broader socio-political structures that made them vulnerable in the first place (Grove, 2018;Reid, 2019).
The interactive governance framework, alongside certain aspects of water governance, provides insights that may serve to 'fill' certain gaps in resilience theory and the SES framework, particularly around the power-related issues previously discussed.Governance studies often focus on social dynamics and institutionsparticularly collaborative and participatory approaches, and the role of informal institutionsand thus tend to have a more nuanced understanding of competing interests and the socio-economic trade-offs involved in natural resource management.Meanwhile, resilience and SES perspectives risk generalisations and the prioritisation of the system as a whole, rather than targeting specific social components.They do, however, complement governance studies by challenging (often dominant) command-and-Fig.1. Visualising the connection between a social-ecological system and the interactive governance framework's (IGF) governing system and system-to-begoverned.While intertwined, we describe social and ecological systems as two interlinked spheres.The system-to-be-governed is primarily the ecological system (which due to our emphasis could also be called the hydro/ecological system), but also includes a large proportion of the social system due to human interactions with nature.The governing system is positioned in the social system, though overlaps with the ecological system again due to the human activities taking place at the intersection of social and ecological systems (represented in light green in the figure).These systems span multiple spatial scales, and key interactions are represented by arrows.(Adapted from Kooiman et al., 2008;Kooiman and Jentoft, 2009).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)control approaches to dynamic environments, with a more elaborated understanding of change and its diverse characteristics.
The proposed resilience-governance framework is represented by two conceptual diagrams that accentuate the linkages between the three theories.The first diagram (Fig. 1) builds on the IGF as a way to describe the governance of complex SES, while the second diagram (Fig. 2) links this conceptualisation with the three dimensions of resilience.
Fig. 1 incorporates the key categorisations used in the IGF, including three governance modes (hierarchical/co-/self) and three orders of governance (day-to-day activities/institutions/values).As mentioned in Section 3.3 and shown in Table 1, the IGF implicitly considers some aspects of resilience-thinking, as both the governing system and the system-to-be-governed can have properties of diversity, complexity, dynamics, and scale.Further, the framework looks at the effectiveness of a governing system to respond to governance challenges (a core component of resiliencethe capacity to respond to change).
The governing interactions, as indicated by the arrows in Fig. 1, also represent aspects such as power relations between different actors and organisations across the governing systemwhether the system is hierarchical or built around principles of self-or co-governance (e.g., as seen in more polycentric arrangementssee Morrison et al., 2019).Drawing on social theory, power can be understood as not only the domination of one actor or group over another, but also the capacity to influence the 'rules of the game' to their advantage (e.g., see Cleaver, 2012;Lukes, 2005;Saravan, 2010;Swyngedouw, 2006).Within the IGF, power thus emerges through the formation and mobilisation of images and instruments and identifying for whom the resulting action is beneficial.Power relations also influence how participatory processes operate, whose knowledge is prioritised, and how conflicting needs are resolved.the focus rather being on the interplay between them.The conceptualisation helps us identify where in the governing system the three capacities of resilience emerge, both desirably and non-desirably (Shaw et al. 2014).
It is important to realise that resilience outcomes are determined by assessing how institutional processes shape interactions and the resulting actions taken within the governing system that contribute to (or undermine) resilience in the interconnected governing system and system-to-be-governed (roughly correlating with the interconnected social and ecological systems respectively), resulting in either persistence, transition or transformation.This builds on the classic 'ball-andcup' conceptualisation of resilience, with the 'ball' representing the system and the 'cups' the different regimes, or steady states for the system (the deeper the cup, the higher the resilience).After a disturbance or pressure, the system may be pushed over a critical threshold or tipping point into an alternative steady state (see Scheffer, 2009;Walker et al., 2004;Willems, 1970); this depiction does not, however, automatically differentiate desirable or undesirable resilience (e.g., see Cinner and Barnes 2019 for discussion on social-ecological traps).In this article, we use two interlinked spheres (instead of one 'ball') to acknowledge the interconnectivity of social and ecological systems while appreciating the different ways in which they respond to pressures and disturbances, and the potential mismatch that may accompany this.For example, issues of power and resilience trade-offs are considered through the governing interactions (as in Fig. 1), as well as the differentiation of how benefits and trade-offs are distributed across society ('resilience for whom?').
Together, the two figures establish the essence of our resilience-governance framework, providing a structured way to bring resilience-thinking into both the theory and practice of water governance.We suggest that the framework in a given governance context builds on two main parts: a visualisation, emphasising the key governance linkages across the three resilience dimensions, and a table synthesising the key elements of governance and resilience (see next section).

Applying the resilience-governance framework
We now illustrate how the resilience-governance framework may be used through two short case studies: first, the Tonle Sap Lake in Cambodia, followed by the Doringlaagte Basin of the Limpopo River Basin in South Africa.We then synthesise the two cases with the help of Table 2.

The Tonle Sap Lake, Cambodia
The Tonle Sap is a complex and biodiverse tropical lake in Southeast Asia, and part of the Lower Mekong River Basin.Often referred to as the 'beating heart' of Cambodia, the lake's unique and dynamicyet predictableflood pulse system drives a diverse ecosystem, which in turn supports an estimated 1.7 million people living on the lake and its floodplains and provides up to 80% of the country's protein consumption (Arias et al., 2019;Chen et al., 2021;Keskinen, 2006;Middleton and Un, 2017;Sithirith, 2011).Communities around the lake have lived in a symbiotic relationship with the lake's ecological system for centuries, developing a synchronised and dynamic 'livelihood pulse' that fluctuates with the seasonal flood pulse (Keskinen, 2006).
The lake is undergoing significant ecological change, including habitat loss, agricultural land conversion, ecological flow disruptions and fisheries decline (Arias et al., 2014;Uk et al., 2018).These have largely arisen from human actions such as built infrastructure (including upstream hydropower development), climate change, and point and non-point sources of pollution.Drivers of change in the Tonle Sap are cross-scalarfor example hydropower development occurring at larger regional scale in the Mekong upstream.These dams have already caused major impacts to the Tonle Sap's flood pulse, particularly affecting its aquatic productivity by, for example, blocking fish migration routes and sediment flow (Keskinen et al., 2013).Such pressures occur within a complex multi-level hierarchy of ineffective, overlapping and contested governance arrangements (Dore et al., 2012;Hirsch, 2016;Jones and Sok, 2015;Käkönen, 2020;Keskinen and Varis, 2012).This hierarchical governance structure has arisen largely due to Cambodia's tumultuous and often violent history (e.g., see Blake, 2019).The lake's social system, like its hydro/ecological system, is highly diverse and dynamic; many ethnic groups co-exist around the lake, livelihoods shift with the water level fluctuations; and outward migration from the lake to nearby cities for work is increasing.In terms of resilience, diversity within a system is typically deemed necessary for resilience; however, social diversity may only be beneficial where it corresponds with agency (Cinner and Barnes, 2019).This is seen in the Tonle Sap, where high disparities in wealth, opportunity and even conflict emerge as a part of the lake's social diversity.Meanwhile, the governing system's formal institutions have low diversity in terms of democratic participation, which corresponds to low adaptive capacity.
A key feature that emerges using our resilience-governance framework is the differentiation between desirable and undesirable resilience.For example, the hierarchical governance mode in the Tonle Sap, and Cambodia more broadly, is highly resilient in terms of absorptive capacity, due to its long-standing and increasingly authoritarian system.The governing system is thus inflexible, lacks diversity, and is characterised by institutions ill-fitted to the complex challenges facing the Tonle Sap (Keskinen and Varis, 2012).Consequentially, this has impacts across both spatial and temporal scales by undermining the absorptive capacity of local communities as well as the system's adaptive and transformative capacities as a whole.
While the lake's communities are traditionally adept at 'living with the floods' (Middleton and Un, 2017), their capacity to cope with unusual, unanticipated or combined socio-economic and environmental change is weak (Nuorteva et al., 2010).This is an important distinction in the communities' resilience to environmental change; there is a difference between being resilient to natural, predictable (even if highly dynamic) water-level fluctuations, and conversely being resilient to unexpected and significant shifts away from these expected fluctuations.This is currently being witnessed in the Tonle Sap, due to impacts from upstream hydropower development and other activities.
The question 'resilience for whom?' distinguishes further the winners and losers in the Tonle Sap's governance arrangement.As already discussed, resilience of those in power is generally strong, while Fig. 3. Three dimensions of resilience across the Tonle Sap's governing system and system-to-be-governed, with emphasis on fisheries.The core elements of absorptive, adaptive and transformative capacity are shown across the two interlinked systems, with arrows marking key causal linkages across the two.Actions and instruments are represented in black italics, while blue, red and purple are used to indicate resilience outcomes (persistence, transition and transformation, respectively).Key outcomes are emphasised in bold, and are represented non-normatively (i.e., without value-based judgements).Elements in grey text are deemphasised for clarity, with italics again indicating actions/instruments.In the bottom-centre, a simple conceptualisation based on Fig. 2 shows the corresponding resilience outcome of the system-to-be-governed (green) and the governing system (yellow).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)inequality is growing because of ongoing environmental changes, at the local scale between wealthier communities (e.g., farmers) and poor fishers living on the lake (Nuorteva et al., 2010;Salmivaara et al., 2016).Both short-term absorptive strategies and medium-term adaptive strategies to unusual environmental change (such as the selling of farming equipment), are typically more accessible to wealthier households; coping efforts by poor households meanwhile may result in poverty traps due to mounting loans and loss of land tenure.This erodes both their absorptive and adaptive capacity further, with migration becoming a main coping strategy (Nguyen et al., 2015).However, the matter of perspective also arises here: does migration make one more, or less, resilient?The answer lies again in agency; it depends on whether migration is a choice.
The key aspects of resilience within the governing system and system-to-be-governed in the Tonle Sap Lake are visualised in Fig. 3, with emphasis on fisheries and the linkages occurring across the system, both positively and negatively.
As Fig. 3 shows, there are numerous linkages and multi-causal pathways of change in the Tonle Sap.Institutional persistence (second-order governance) emerges as highly influential across the entire systemfor example, by impacting the adaptive capacity of the systemto-be-governed across scales (i.e., the lake and its inhabitants).Interestingly, institutional changessuch as the recent Deep Fisheries Reform whereby commercial fishing lots were suddenly cancelled by Cambodia's Prime Ministerhave had limited impact on the (positive) resilience of both the lake's fisheries and those dependent upon them (Jones and Sok, 2015).Rather, the system-to-be-governedi.e., the lake system and related nature-based livelihoodscontinues its trajectory towards significant change.While this on-going transformation is mainly caused by the impacts that hydropower development causes to the seasonal flood pulse and Tonle Sap ecosystem, it also links to overfishing and the inability of both the old and new fishing system in preventing unsustainable practices.
Persistent images held within the governmentfor example, a longstanding 'hydraulic society' and irrigation ideologiesstifle the possibility of institutional and technical innovation and learning which form key aspects of desirable resilience (as shown in Fig. 3).Over time, this may undermine the capacity of both the governing system and systemto-be-governed to adapt to change brought about by both internal and external pressures.Fig. 3 also shows the emergence of possible 'rigidity traps' in the systemfor example, illegal fishing practices degrade the lake's fisheries, leading to fishers investing in better gear to increase their catch, often taking out loans to do so.However, this further degrades the fisheries, thus leading to low returns and a reinforcement of debt.
The analysis suggests a mixture of resilience capacities and outcomes.The system-to-be-governed (i.e., the lake ecosystem and related communities) continues to transition, with the potential to transform into a new system state in the future.This is due to its weakening flood pulse and degrading fisheries, with a disparity of adaptive capacity between different social groups.Meanwhile, images in the third-order governance remain persistent, maintaining the status quo of a predominantly inflexible and inequitable hierarchical governance arrangement (with little diversity), regardless of surface-level policy changes made in the second-order governance.There is also little evidence of transformative capacity in the first-order governance, and more fundamental change here may require long-term shifts in the third-order governance (values, norms and principles), which typically operate over a longer timescale.Referring back to Fig. 2 ′ s 'cup-and-ball' conceptualisation, we may even conceive that the green 'system-to-begoverned' sphere is undergoing a hydro/ecological transformation into another steady state, while the yellow 'governing system' sphere sinks even deeper into the 'cup' of resilience; or, in the language of the IGF, there is an 'ill-fit' between the governing system and system-to-begoverned (bottom-centre of Fig. 3).

The Doringlaagte Basin, South Africa
Our second case study is a small basin of the Hout Catchment in the Limpopo River Basin in South Africa (locally known as the 'Doringlaagte').The basin covers 509 km 2 area and drains into the Limpopo River in the northeast.The area covers two connected aquifers episodically recharged and experiences a semi-arid environment.Farmers in the area depend entirely on groundwater for irrigation (for crops such as potatoes), and thus water management primarily occurs at the farmlevel.The basin is subject to both biophysical and human pressuresnamely, climate variability (which is increasing due to climate change) and excessive groundwater pumping for irrigation.
In the area, long-term groundwater recharge trends and variability are affected by climate variability, which may impact long-term sustainability, and thereby resilience, of the coupled human and natural systems (Ebrahim et al., 2019;Fallon et al., 2019).Due to inadequate monitoring, data is low, though recent integrated hydrogeological modelling suggested that current irrigation (from only 0.9% of the catchment area) appears to be close to the aquifer's sustainability limit (Ebrahim et al., 2019).Day-to-day activities of farmers (first-order governance) influence the system-to-be-governed's resilience, despite poor (perceived) institutional support.Furthermore, one staff member in the Department of Water and Sanitation (DWS) produced in-depth quarterly groundwater status reports and disseminated these to farmers, who felt they were useful in understanding groundwater dynamics regarding pumping activities and seasonal climate variability (Fallon et al., 2019).
The resilience of the catchment's groundwater resources to socialecological pressures is undermined by weak regulation and a degraded monitoring network to aid water management measures.Farms are subject to water licenses by DWS and local cooperation between farmers is commonplace, though at times strained.Groundwater management options have been discussed (e.g., irrigation planning, land-use planning, and seasonal climate forecasting), though informal and ad-hoc management practices currently dominate the governing system due to weak formal regulation by DWS (Ebrahim et al., 2019;Fallon et al., 2019).Attempts to establish a Water User Association (WUA) have been unsuccessful, though responsibility has shifted to the informally established Agricultural Union.The catchment's water problems can also be seen as a symptom of South Africa's broader 'wicked' water problems (Fallon et al., 2021), showing the impact of cross-scale pressures.Primarily, water and land access in post-apartheid South Africa remains highly inequitable, with a majority in the hands of white farmers (van Koppen and Schreiner, 2014).
Due to the catchment's small size, diversity between resource users is low, with most groundwater use carried out by commercial (white) farmers.However, there are also emergingtypically black, historically disadvantagedsmallholder farmers in the area, who are often excluded from Agricultural Union meetings, and lack the administrative literacy needed to understand the licensing system.It can therefore be conceived that wealthier commercial farmers will be able to buffer socialecological change more readily than their emerging counterparts.However, commercial farmers' resilience is still self-perceived to be inadequate.
The Doringlaagte is a nested system embedded within a larger sociopolitical context that affects its resilience.Some key expressions of resilience across the governing system and system-to-be-governed are shown below in Fig. 4. As shown, there is a mixture of absorptive, adaptive and transformative capacities across the governance orders, which has uncertain but potentially remarkable implications for the aquifer.For example, the visualisation suggests transformation in the second-order governance; a transformative shift in post-apartheid South Africa's water law and policy in 1998 (Herrfahrdt-Pähle et al., 2020) led, in theory, to more equitable water rights for previously disadvantaged groups, and a recognition of water as a public resource rather than a riparian right (Republic of South Africa, 1998).However, decades later water access remains inequitable, and implementation of the 1998 National Water Act has been slow (Schreiner, 2013).
Resilience-thinking can shed light on two important aspects of this governance shift.First, it suggests that rapid, transformative institutional changeif not accompanied by change in cultural values and daily actions taken on the groundmay have limited practical impact across the governing system.In South Africa, persistent daily practices, alongside an unwillingness (or capacity) to implement such drastic legislative changes, undermined the ability of the new legal framework to truly transform the country's water governance (Schreiner, 2013).In the Doringlaagte, many farmers still perceive water to be a riparian right (i.e., the view that only those owning land adjacent to rivers are entitled to use the water)2 , which again suggests a slower shift in principles, regardless of institutional change regarding water and land rights.This is particularly important for groundwater, because land tenure also remains highly inequitable, although some farmers interviewed are engaged in a mentorship scheme for emerging farmers to increase the number of black-owned farms, signalling a potential shift in these longheld beliefs.
The resilience trade-off across different parts of the governing system may partly be explained by a misplaced focus on institutional design over the social realities of post-apartheid South Africa (Meissner et al., 2016).Despite the government's emphasis on IWRM principles in the third-order governance (e.g., decentralisation), it did not have the necessary resources for such a profound governance shiftfor example, as seen by the failure to establish local Catchment Management Agencies (Kemerink et al., 2013;Movik, 2011;Schreiner, 2013).Schreiner (2013, p. 6) refers to this as the National Water Act's (Act 36 of 1998) Fig. 4. Three dimensions of resilience across the Doringlaagte's governing system and system-to-be-governed, with an emphasis on groundwater use.The core elements of absorptive, adaptive and transformative capacity are shown across the two interlinked systems, with arrows marking key causal linkages across the two.Actions and instruments are represented in black italics, while blue, red and purple are used to indicate resilience outcomes (persistence, transition and transformation, respectively).Key outcomes are emphasised in bold, and are represented non-normatively (i.e., without value-based judgements).Elements in grey text are de-emphasised for clarity, with italics again indicating actions/instruments.In the bottom-centre, a simple conceptualisation based on Fig. 2 shows the corresponding resilience outcome of the system-to-be-governed (green) and the governing system (yellow).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) "Volkswagen vs the Rolls Royce issue", in that it might have been better to write a 'Volkswagen' piece of legislation that was more suited to the country's limited resource capabilities, rather than the 'Rolls Royce' of IWRM legislation that has largely been unimplementable.A conceptualisation of this ill-fit between a rapidly transformed governing system (in terms of the formal institutions) and a slowly transitioning system-to-be-governed is shown in the bottom-centre of Fig. 4.

Synthesising the case studies
The resilience-governance framework shows various aspects of resilience across the two case studies.Table 2 brings these together to highlight key aspects of both.Geographical scale is an important aspect of both the Tonle Sap Lake and Doringlaagte, with both cases existing as sub-components of nested systems within the transboundary Mekong River Basin and Limpopo River Basin respectively.Causal relationships occur across spatial scales in both cases, both within and across the governing systems and systems-to-be-governed.For example, the Tonle Sap's system-to-be-governed is influenced by other national-level decisions within the broader Mekong governing system (for example, local ecological externalities caused by nested levels of uncoordinated decision-making to continue building dams by other Mekong countries).Meanwhile, regional environmental drivers in the Limpopo (such as increasing climate variability) influence how farmers at the local level respond and adapt.
Elements of resilience as a non-normative property emerge in the governing systems, such as Cambodia's institutional persistence and inflexibility, and low institutional diversity.Normatively, different perspectives highlight different resilience capacities, as well as outcomes, such as the differences between ethnic or fishing gears-based groups in the Tonle Sap, or between white commercial farmers and Black emerging farmers in the Doringlaagte.Meanwhile, trade-offs also highlight different resilience perspectives.For example, Cambodia's resilient hierarchical governing system comes at the expense of locallevel resilience to the multiple stressors impacting the Tonle Sap.Short-term coping strategies may also undermine longer-term resilience.This is seen through poor fishers in the Tonle Sap taking out loans for fishing equipment to increase their short-term catch, which also reduces the resilience of both their livelihoods (by reducing financial absorptive capacity) and the fishery itself.
Finally, resilience interactions are suggested across governance orders.Rapid policy changes within the second-order governance (the Fishery Reform in Cambodia and the National Water Act in South Africa) seem to have limited impact in both cases on first-order governance (day-to-day activities) regarding natural resources, potentially due to a concurrent slower rate of change in the third-order governance (see Section 6.1).

Discussion
In this section, we consider the theoretical and practical implications of our study and the proposed resilience-governance framework, followed by reflections on the framework's application to the cases, methodological considerations, and ways forward.

What does resilience-thinking bring to water governance?
We argue that current uses of resilience-thinking in water governance contexts often under-utilise its full potential due to an incomplete understanding of resilience theoryin particular, an over-focus on absorptive capacity and persistence.Governance studies often focus on either preserving a system in the face of stressors and change, or on enhancing adaptive capacity to adjust to externally driven change.We therefore suggest that there is scope for expanding the use of resilience in water governance studies in terms of its conceptualisation as a property, process and outcome.
First, we can identify properties of resilience (normative and nonnormative) not only in the system-to-be-governed, but also within the governing system itselffor example, the level of stakeholder diversity, polycentricity, and institutional flexibility.Non-normative understandings of resilience may be helpful in identifying persistence and rigidity traps within a governing system.Resilience-thinking can also encourage 'big-picture thinking', and a clearer way to understand nested systems and complex cross-scale interactions within and between the governing system and system-to-be-governed, which may be obscured in other governance framings.
As visualised in the 'resilience outcome' boxes at the bottom of Figs. 3 and 4, it can also be argued that despite their interlinkages, resilience outcomes of a system-to-be-governed and governing system are not necessarily coupled.The Tonle Sap case indicated that it is possible for a system-to-be-governed (the Tonle Sap's fisheries) to undergo significant changeand even transformationwhile the general governing system persists in its existing, partly reformed form.The Doringlaagte case, on the other hand, showed how the transformation of a formal institutional setting does not necessarily lead to hydro/ ecological transformation.This is a potentially valuable addition to how we conceptualise and understand social-hydrological systems and the differences between social and hydro/ecological systems in terms of how they respond to pressures.
Related to this, the results also suggest that resilience within the governing system itself is not necessarily homogenous; the three governance orders of the IGF seem to be malleable to different degrees, and thus respond differently to change.Day-to-day actions (first-order governance) may have surprisingly fast effects on the trajectory of a system.Meanwhile, structural institutional change (second-order governance) can be fast or slow, as seen in both Cambodia and South Africa's rapid legislative changes.While rapid changes in (possibly inequitable) laws and policies may be desirable, they do not always have the intended result in the day-to-day activities of resource users (firstorder governance).This is often due to implementation challenges (Schreiner 2013), but may also be due to a slower rate of change in stakeholders' images, values and principles (third-order governance)revealing a possible 'window of opportunity' for encouraging longerterm, or even transformative, change.
Our results indicate that more attention should therefore be paid to the transformative capacity of a governing system in water governance studies.This may help us imagine a system beyond preserving the status quo and pre-existing (potentially untenable) system functions.This shift in focus may also reduce the burden of adaptation often placed upon vulnerable groups and emphasise the need for more significant structural changes (Evans and Reid, 2013;Grove, 2018).The importance of transformative capacity is particularly evident within the third-order governance, where cultural values, principles and norms are typically, and understandably, difficult to changeand could therefore prevent change (for good and for bad) elsewhere in the system.

Towards a more critical reading of resilience-governance studies?
For resilience-thinking to complement water governance studies fully, it must be incorporated criticallythat is, with consideration of its desired and actual impacts across society.There is a risk of resilience and SES language depoliticising water issues through inadequate consideration of socio-political issues, whereby power relations determine the 'winners' and 'losers' of governance arrangements and may mask key relationships that stimulate or hinder change.A power-sensitive approach to water and resilience is therefore important (Dewulf et al., 2019), which may require a shift towards analyses that focus on cultural values and historical contexts (Cleaver and Whaley, 2018;Cote and Nightingale, 2012).We see that the resilience-governance framework also facilitates this shift due to its consideration of third-order governance and the historical development of infrastructure and water policies, as well as cross-scale trade-offs whereby the resilience of one scale of social group may come at the expense of another (Cinner and Barnes, 2019;Dewulf et al., 2019).
A more critical and power-sensitive use of resilience could ameliorate its current focus on maintaining the status quo or incremental change, and move governance towards transformation ('bouncing forward').Resilience, when used uncritically, risks normalising undesirable conditionsthe most vulnerable and marginalised in society must endure threats and shocks imposed upon them, rather than gain the freedom and agency to transform their situation (Evans and Reid, 2013).In the same way, resilience may also romanticise local people's adaptive capacity.Thus, we contend that the transformative element of resilience could receive more focus in water governance studies than it currently does.This is particularly important in contexts such as our case studies, where challenging the status quo may be necessary for achieving broader equity and justice goals.
Regardless of which of the three resilience capacities are being analysed (see Fig. 2), we maintain the most important aspect is to retain a power-sensitive perspective that focuses on whose benefits, perspectives and values are prioritised across all three capacities.This avoids overly naïve or black-and-white thinking around resilience in water governance systems, such as the assumption that transformation would automatically lead to positive outcomes for all.As seen in the South Africa case, for example, transformation in one system component (here, formal water institutions) does not automatically lead to more equitable outcomes for all water users, and may even lead to 'elite capture' whereby vast changes are again steered by the most powerful players for their own gain (Schreiner, 2013).
A core issue in integrating resilience and SES-thinking into water governance is that social systems are fundamentally different to ecological systems, and thus cannot be analysed in the same way; doing so may further depoliticise the issues faced (Béné et al., 2012).Therefore, while it is useful to understand resilience properties in a governing system, they may not necessarily duplicate their original ecological understandings.A more fundamental contribution of resilience and SESthinking to water governance studies may therefore be an epistemological shift away from focusing on institutional structure, towards capturing the processes and relations supporting said structure (Cote and Nightingale, 2012).However, this requires a more critical, socially grounded approach to resilience than is often found in current water governance studies.
Critical use of resilience-thinking may also help us understand where the burden of resilience lies within a system.We found that some normative properties, such as stakeholder diversity, are insufficient for building resilience, and may even contribute to conflict; attention must therefore be paid to stakeholder agency and participation in decisionmaking processes.Resilience-thinking may then help us to understand better the relational and processual dynamics of governance arrangements, rather than focusing on the existence of certain properties or structures.
Finally, diverse epistemic perspectives on resilience may make it a useful boundary object across disciplines to address some of the challenges around cross-disciplinary knowledge integration, as well as encouraging researchers to reflect upon their own interpretations while being aware of others' (Brand and Jax, 2007;Ensor et al., 2021).This is particularly important when addressing complex water problems, where multiple perspectives must be deliberated (Brand and Jax, 2007;Fallon et al., 2021;Moore, 2013).

Reflections on the framework and case studies
Our study provides also possibility for important methodological reflections on the use of resilience-thinking in water governance as well as the related framework and case studies.We found theoretical multiplicity (Karpouzoglou et al., 2016) useful for exploring theories with different conceptual originsfor example, resilience and SES in their engineering and ecological roots, and water governance's roots in social theory and political ecology.Choices had to be made, however, regarding where to focus (e.g., choosing the IGF as a basis for our governance frame, rather than others) and what kind of perspectives and characterisations to include in the resilience-governance framework.
Including an additional theoretical layer, namely resilience, into the already-multifaceted concept of governance also easily leads to an increasing diversity of elements and linkages, as visualised by the case studies and particularly the related Figs. 3 and 4. Yet, we feel a certain level of complexity is required to identify and understand the many linkagesthough, after this initial visualisation, it may also be possible to focus on the most relevant linkages.The case studies also indicate how gaining a 'snapshot' of a dynamic system is difficult due to constantly moving system components and processes through time and space.In our analysis, this made it challenging to distinguish between the three often-overlapping (and partially subjective) resilience dimensions included in the proposed resilience-governance framework.
We also found it challenging to situate the impacts of infrastructure within the visualisations, as it does not clearly align with either the social or ecological system, nor the governing system or system-to-begoverned. Yet, infrastructure may fundamentally alter both systems, as shown in the Tonle Sap case, where hydropower dams are a key driver for flow changes and related impacts on both social and ecological systems through e.g., decline in fisheries.Future work could therefore make the role of infrastructure as a driver of social and environmental change clearerfor example, through the use of a social-technicalecological systems framework (e.g. as used by Ahlborg et al., 2019;Chester et al., 2015).
Nonetheless, we are overall pleased with how the framework worked for two case study contexts, as it helped to emphasise the diversity of linkages across scales and systems and, in general, provided a structured way to consider resilience-thinking in those governance contexts.Yet, the framework's visualisation for the two case studies were still limited; they did not, for example, fully capture the diversity of actors involved and may inadvertently have put focus more on ecological rather than on social processes.Efforts could therefore be made to fully incorporate the two without losing comprehension.Furthermore, it was difficult to visualise spatial and political scales for the cases, although both were important aspects of the analysis.
Resilience is a slippery, and at times ambiguous, concept with many possible interpretations.Integrating resilience and water governance theory is therefore not straightforward, laden with assumptions and a risk of oversimplifying social and environmental challenges.Thus, we feel that future work should focus on more practically oriented uses.The role of resilience as a boundary object in multidisciplinary research around complex water problems could also be explored more, along with increased focus on opportunities for governance interventions.We do, however, appreciate the challenges associated with such pluralistic approaches, which is perhaps why many practitioners favour more simplistic frameworks and lenses.Nevertheless, complex water governance challenges require approaches that encompass not only environmental complexity, but also the many perspectives, needs, and values involved.

Conclusions
In this article, we considered how the theories of social-ecological systems (SES), resilience and interactive (water) governance could be used simultaneously for analysing complex systems undergoing significant social and environmental change.We found there is potential in integrating resilience-thinking more thoroughly into water governance, and used theoretical multiplicity to build a resilience-governance framework to do so.We found that the theories had many logical overlaps, though the numerous interpretations of resilience complicated this in practice.Resilience-thinking can, however, help us understand complex social-hydrological systems through incorporating issues of complexity, dynamics, nonlinearity and scale.Furthermore, resilience A. Fallon et al. properties emerged within governance arrangements themselves, including stakeholder diversity and nested institutions.
Importantly, we found that a crucial aspect of a resilience-governance analysis is to maintain a power-sensitive perspective focusing on whose benefits, perspectives and values are prioritised across the system.This may emerge through identifying resilience tradeoffs (such as the resilience of one social group at the expense of another), or the prioritisation of transformation in contexts such as those seen through the case studies, where challenging the status quo may be necessary for achieving broader equity and justice goals.
Insights were provided by both case studies -Cambodia's Tonle Sap Lake and South Africa's Doringlaagte.In both cases, there was an issue of formal institutions being incapable of properly preparing for and responding to environmental change.Both case studies indicated the potential of principles and norms in influencing day-to-day actions by water users, and therefore their resilience.Interestingly, both cases have undergone (seemingly) huge transformative institutional change (fisheries reforms in Cambodia and post-apartheid legislative changes in South Africa) that did not, however, fully translate into practice.This suggests two important lessons.First, more emphasis should be placed on understanding the historical origins and thus the role that third-order governance (i.e., deeper changes in norms and cultural values) plays in both enabling and hindering change and related responses.Second, considering the resilience of social and ecological systems as connected and co-producing entities may help better to see the possible mismatch between the resilience capacities of a system-to-be-governed and its governing system (as illustrated with the two spheres in Figs. 2, 3 and 4).
Current uses of resilience-thinking in water governance theory typically focus on resilience as a normative goal in dealing with social and environmental change.While this is as such understandable, we argue that resilience theory is more nuanced and developed than its current integration into water governance.A better consideration of the different capacities and interpretations of resilience will, we argue, provide a more useful basis to contribute to water governance, particularly in the context of dynamic social-hydrological systems.For resilience theory to be used more comprehensively in water governance, its own issues around social and political elements must be addressed simultaneously.We propose the resilience-governance framework as one way of doing so, whereby resilience is not only viewed as a goal, but also as a non-normative property of the governing system itself and an ongoing, interactive process between the governing system and systemto-be-governed.By doing so, the full potential of resilience-thinking may be better realised within a water governance context, and thereby inform the development of a more resilient and equitable governance system.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2
Fig. 2.A visualisation of the resilience-governance framework.The governing system and system-to-be-governed correlate with Fig.1's visualisation, also representing the interlinked social and ecological systems (with the social component represented in yellow and the ecological component represented in green).These overlay the three resilience capacities (absorptive, adaptive and transformative capacity), elements of which can be uncovered across the two interlinked systems.Note that these capacities are not fixed, and the systems may shift between states of stability, flexibility and change.Resilience outcomes for these interlinked systems are then represented by the bottom 'ball and cup' conceptualisation related to persistence, transition and transformation.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 2
Synthesising the linkages between resilience and governance in the two case studies of the Tonle Sap and the Doringlaagte.