**1. Introduction**

Strategic river basin planning consists of a complex, socially ambitious set of knowledge production practices, involving monitoring and assessment, expert-led analysis, and participatory planning [1,2]. This set of practices involves the production and synthesis of knowledge in multiple domains, the interpretation of key messages by policy actors, and the deployment of such messages in planning processes. Because they are complex, resource-intensive, and generally publicly funded, and because actors use them to justify particular investments or development trajectories, river basin planning practices deserve critical reflection.

In this perspective paper, we reflect on the processes and outcomes of two participatory river basin planning initiatives, to which—as process designers, implementers, and observers—we have contributed. One subset of the co-authors implemented an exploratory planning study for the Ayeyarwady river basin, completed in 2018 as a step towards a river basin master plan [3]. Another subset initiated Nepal's first participatory water resources development strategy for the Kamala, a 2050 km2 river basin in Nepal, an ongoing initiative as of 2019 [4]. These capability-constrained, post-conflict, democratizing settings offer vital insights into the strengths and limitations of approaches to strategic planning. The paper argues that IWRM-based planning requires co-productive models of planning. This argument is based on the authors' reflection on methodological challenges we navigated when designing and implementing river basin planning projects in Myanmar and Nepal.

The concept of strategic river basin planning has evolved since its emergence in the late 20th C. Until the 1990s, it essentially meant long-term infrastructural development planning, with relatively simple social or environmental analysis. Post-WWII water supply infrastructure was planned according to an engineering-oriented paradigm to meet certain objectives (such as irrigated agriculture and hydropower production) [1]. However, events preceding and during the 1990s revealed the social and environmental limitations of such paradigms to river basin development. Thailand and Nepal, for instance, debated the social acceptability of, and alternatives to, large hydropower dams [5,6]. After experiencing major floods in 1993 and 1995, the Netherlands came to recognise the limits of engineering practice, and the value of integrating spatial planning and water management [7,8]. In 1995, Australia announced a cap on diversions from the Murray Darling Basin, to avoid the ecological collapse of Australia's largest river system. At this time, a less fragmented, more coordinated, and systemic paradigm for water resources planning emerged, as reflected in the 1992 Dublin Statement on Water and Sustainable Development [9]. The Dublin Statement is the foundation for an integrated approach to planning water resources, subsequently promoted by development actors during the 2000s under the name of Integrated Water Resources Management (IWRM).

IWRM aspires to improve three "Es": efficiency, equity, and environmental sustainability [10]. It explicitly promotes the integration of multiple stakeholders, disciplines, and spatio-temporal scales [11]. However, within two decades of its emergence, practices to implement IWRM drew criticism for overly optimistic assumptions about how changes to water planning could deliver the three Es in unequal societies such as South Africa [10], Tanzania [12], and Nepal [13], and transboundary regions such as the Mekong basin. Contemporary IWRM is an ambiguous and diverse set of practices [11,14]. It includes top-down, principle-driven variants, as well as local-level, bottom-up, "expedient" versions [15,16]). The basin planning processes we reflect on in this paper have been influenced by highly aspirational IWRM principles, such as formulating a stakeholder-agreed development plan for the Ayeyarwady basin in Myanmar [17], or the desire among water agency professionals in Nepal to identify optimal development strategies using decision support systems. However, to realize—even partially—such aspirations, a river basin development plan would require meaningful participation and collaboration at, and across, multiple levels of governance [14].

Proponents of IWRM in developing countries face two notable political challenges. The first such challenge is asymmetry of knowledge and power, manifested as the uneven distribution of capability and authority between local and national government. Not only are water and land resources unevenly distributed in river basins, the capability of planners, as well as of affected people, is also concentrated at particular levels and locations (and decisions at one location or level can lead to unwanted consequences elsewhere in the system). For example, until recently, the unitary system of Nepal concentrated planning resources and capability at the centre. Power asymmetry can limit the recognition of local interests, knowledge, and socio-technical water management. It can prioritise national-level water resource development preferences. For example, large hydropower and inter-basin water diversion projects for irrigation have dominated planning conversations in Myanmar and Nepal, respectively [18,19]. The challenge of power asymmetry is exacerbated in contexts of data scarcity and uncertainty about river basins as social–ecological systems, where the state of the basin is influenced by political dynamics and narratives, as much as biophysical processes.

A second political challenge consists of the organizational mode by which IWRM initiatives have been delivered to developing countries. The recurring mode has been the international technical assistance project. Project modalities may constrain local interest and institutionalization, particularly when assistance is narrowly channelled. This risk is heightened in settings such as Myanmar and Nepal, which have restructured their water-related agencies and sought to establish new inter-agency water bodies. Although inter-agency coordination is challenging in any context, it is particularly acute in contexts of radical state restructuring, such as Nepal. Compounding these particular challenges, bureaucratic competition [20] and deficits of trust (between non-state and state actors, between lowerand high-level state actors) may arise.

The above conditions and dynamics have influenced the methods and techniques applied to river basin planning—the focus of this paper. Modern basin planning, which pre-dates IWRM, has favoured a particular set of expertise and stakeholders. For example, the expertise of hydrologists, engineers, lawyers, and national government officials tends to outweigh that of citizens, local officials, livelihood specialists, gender and social inclusion experts, and political economists [21]. The expertise and stakeholders favoured may insufficiently represent the breadth of water- and development-related concerns of people in large or complex river basins [22]. Even the production of disciplinary knowledge—such as a basin-scale surface water model—requires within-disciplinary diversity (e.g., rigorous peer-review) to be credible. The consumption of expert knowledge by non-experts requires accountability and transparency among knowledge producers (e.g., about the implications of uncertainty). When expertise, stakeholder concerns, interests, and relevant socio-technical options are inadequately included, legitimacy is compromised [18,23].

How can IWRM-based river basin planners increase the legitimacy of the strategic planning processes they design and facilitate? We engage with this question by reflecting on how particular methodological commitments, evident in our case studies from Myanmar and Nepal, exert influence on knowledge production and stakeholder participation. By "methodological commitments", we mean recurring preferences that we, as planning practitioners or stakeholders, exhibit towards particular methodologies (understood as conceptual models, and associated study designs and techniques). Such commitments can be explicit or tacit. The commitments discussed in this paper are inferences we have made, based on a review of key primary texts (e.g., terms of reference) and participant-observation.

To support such an interpretation, we use two models of how expertise and knowledge inform policy action. The first model is a rational choice model of decision-making [24,25]. In this model, an authorized decision maker (e.g., a minister, or ministerial council) makes decisions which allocate finite public resources so as to maximize societal utility, based on preferences voiced by citizens [26]. This model of expertise and choice assigns high responsibility to credentialed experts, who advise on the consequences of taking different socio-technical options. The second model is a co-productive model of decision-making. This model attaches relatively greater weight to the knowledge of non-credentialed experts. Through collaborative processes involving diverse actors, it seeks to produce *agreement on shared goals*, and to produce *knowledge relevant to achieving those goals* [27–30]. In the second model, authorities make decisions after recognizing, participating in, and responding to recommendations from co-productive processes (Section 3.2).

The two models of decision-making diverge on what constitutes actionable knowledge and how it should be produced (Section 3.2). Our case studies reveal the tensions that arise when both models co-exist in river basin planning. We describe our attempts to negotiate such tensions, and the consequences for planning of such negotiations. Our motivation is thus to reflect critically on the consequences of making particular methodological commitments, including those to which we contributed, for the purpose of improving IWRM-based planning.

Section 2 expands on the development contexts of Myanmar and Nepal. Section 3 then summarizes the original designs of the two river basin planning initiatives in Myanmar and Nepal, showing how their specific features originate in designs by water resource experts affiliated with international development partners. We also describe how specific commitments led to methodological tensions, which challenged us to revise or augment our original designs. Based on the insights from our two cases, Section 4 discusses implications for realizing IWRM-based river basin planning in developing countries. Section 5 concludes with recommendations.

#### **2. Water Resources Development Contexts: Myanmar and Nepal**

#### *2.1. Myanmar*

Since 2011, Myanmar's partial and contested democratization [18] has led to a notable increase in technical assistance by international development partners, to multiple sectors. In turn, since 2017, such assistance has yielded an efflorescence of water and water-related studies (e.g., [3,31–33]). Development technical assistance involves the promotion, by partners of national expertise (e.g., Australia and other donors in water resources modelling), of cooperation among partners to focus investment (e.g., an Australia–The Netherlands memorandum of cooperation around water resources assistance in Myanmar), as well as competition among partners promoting IWRM-based planning (e.g., Australia, The Netherlands) and those promoting infrastructural development (e.g., China, Japan, Korea).

In 2015, the World Bank initiated a Decision Support System and Basin Master Plan (herein, "Basin Master Plan") project for the Ayeyarwady river basin, as part of a \$100 M credit-financed initiative known as the Ayeyarwady Integrated River Basin Management Project. The 2018–2020 Basin Master Plan project is a major investment in evidence-based planning for Ayeyarwady. It aims to deliver a stakeholder-agreed basin development strategy for the Ayeyarwady, under the auspices of the National Water Resources Committee (NWRC), an inter-agency advisory body formed in 2015. Reaching agreement among a diversity of actors and interests is demanding in any context, let alone in the Ayeyarwady basin of Myanmar, where ethnic armed and quasi-state organizations contend for power and recognition against the Union government.
