Handling uncertainty through adaptiveness in planning approaches: comparing adaptive delta management and the water diplomacy framework

ABSTRACT

Planners and water managers seek to be adaptive to handle uncertainty through the use of planning approaches. In this paper, we study what type of adaptiveness is proposed and how this may be operationalized in planning approaches to adequately handle different uncertainties. We took a comparative case study approach to study two planning approaches: the water diplomacy framework (WDF) and adaptive delta management (ADM). We found that the approaches differ in their conceptualization of uncertainty and show that different types of adaptiveness are used in the approaches. While WDF builds on collaborative adaptive management as a set of ongoing adjustments and continuous learning to handle uncertainty, ADM deliberately attempts to anticipate future adaptations through a set of tools which allows for seizing opportunities and avoiding lock-in and lock-out mechanisms. We conclude that neither of the approaches is fully able to account for different uncertainties. Both approaches may benefit from specific insights in what uncertainty and adaptiveness entail for the development of water management plans.

KEYWORDS:

• Adaptiveness
• uncertainty
• planning approaches
• environmental planning
• water management

1. Introduction

Planners and water managers seek to be adaptive to handle change and uncertainty in spatial and environmental planning (Kwakkel, Walker, & Marchau, 2010). Handling uncertainty grows in importance due to the need to deal with climate change, characterized by the prevalence of long-term, stochastic uncertainty about its impacts (Lempert, Nakicenovic, Sarewitz, & Schlesinger, 2004). Planners and water managers can strive for adaptiveness in three ways. First, by adaptively managing resources through focused experiments and closely monitoring change (Caves, Bodner, Simms, Fisher, & Robertson, 2013; Walters, 1986). Second, by increasing the adaptive capacity of institutions and deliberate learning from past experiments (van den Brink, Meijerink, Termeer, & Gupta, 2014). A third possibility is to iteratively take decisions and create room for future adjustments to cope with as yet uncertain change when planning for interventions in the physical environment (Kato & Ahern, 2008).

How adaptiveness is used and specified for environmental planning practices and policy-making depends on assumptions about uncertainty. These assumptions influence how information is dealt with (Martens & van Weelden, 2014) and which type of adaptiveness is adequate to handle specific uncertainties (Zandvoort, van der Vlist, Klijn, & van den Brink, 2017). The adequacy of adaptiveness foremost depends on the nature of uncertainty with which planners and water managers are confronted (Brugnach, Dewulf, Pahl-Wostl, & Taillieu, 2008; Walker et al., 2003). Uncertainty can be of an ontic nature (in case of chaotic system behaviour), epistemic nature (in case of a lack of current knowledge) or ambiguous nature (in case of diverging frames or perspectives) (Brugnach et al., 2008; Kwakkel et al., 2010; Walker et al., 2003). In the context of planning problems, ambiguity relates but is not similar to disagreement (Brugnach et al., 2008). The difference relates to how ‘frame’ is defined: either as a cognitive representation or as an interactional co-construction (Dewulf, Craps, Bouwen, Taillieu, & Pahl-Wostl, 2005). We see ambiguity as originating from the existence of different cognitive representations pertaining to an issue, which we distinguish from disagreement as this pertains to the temporary absence of an unanimously accepted frame of understanding of an issue or its adequate handling (Zandvoort et al., 2017). This distinction helps to identify an adequate course of action to mitigate both ‘types’ of ambiguity: either seeking co-construction, including different frames and hedging against the influence of frames on planning or conflict resolution and frame alignment in a particular situation. Ontic uncertainty, due to the chaotic behaviour of a water system, is best tackled by adaptiveness in the system itself. Ambiguity is best dealt with (but not necessarily resolved) by adequate representation of different frames and values throughout an adaptive learning process and appropriate mechanisms to resolve disagreements (Brugnach et al., 2008; Zandvoort et al., 2017). To put it simply, increasing the height of a dike will not solve ambiguity among stakeholders, while building consensus will not diminish uncertainty about future water discharges.

When planners intend to handle uncertainty they may rely on planning approaches which make use of adaptiveness. We conceptualize planning approaches as combined sets of tools in a coherent framework which planners can use to structure planning processes and deal with multifaceted planning situations, akin to policy packages (Howlett, Fraser, Mukherjee, & Woo, 2015) and portfolios of tools (Aerts, Botzen, van der Veen, Krywkow, & Werners, 2008).

Multiple planning approaches are developed to handle uncertainty, especially in practices involved with planning in river deltas and other water systems (e.g. Berke & Lyles, 2013; Lempert et al., 2004). Here, we intentionally focus on planning approaches which originate from the messy day-to-day practice of planning. In such practices, complex water problems arise which are characterized by different uncertainties that need to be accounted for simultaneously. To get a coherent approach for multifaceted situations with complex problems, multiple uncertainties need to be addressed. We argue that when the resulting approaches do not offer insight in the consequences of different uncertainties, the congruent handling of uncertainty can become problematic. Moreover, it is yet unknown if approaches are adequate to handle different uncertainties when developing policy and what type of adaptiveness is expected to do so.

In this paper, we study what type of adaptiveness is proposed and how this is operationalized in planning approaches to handle different uncertainties. Our research question is: ‘How is adaptiveness operationalized in planning approaches to handle uncertainty in complex water problems?’ We start from the premise that analysis of planning approaches may offer insight in how adaptiveness can be operationalized to handle different uncertainties, specified as to the three natures described above. The structure of our paper is as follows. We first outline our research approach. Second, we describe the studied planning approaches in general, the uncertainties they address and their specification and use of adaptiveness. Third, we compare the approaches to see how adaptiveness is embedded in the approaches to handle uncertainty. Lastly, we discuss our findings and offer our conclusions.

2. Methods

To answer our research question we took a comparative case study approach (Engeli & Rothmayr Allison, 2014). A major advantage of comparative case studies is that they allow for a tailored analytic strategy to explore poorly understood phenomena by answering how and why questions (Meyer, 2001). We chose to compare not more than two approaches to allow for in-depth insight in the mechanisms involved in operationalizing adaptiveness to handle uncertainty. We sought for two similar approaches developed in different contexts because we assumed that the context may determine the use of concepts such as adaptiveness and uncertainty. To make this possible our case study design needed to adequately specify the analytical use of the comparison and the principles on which this is based (Hyett, Kenny, & Dickson-Swift, 2014).

There are multiple planning approaches, tools and instruments available for specific practices and contexts. Some examples are listed in Table 1. We distinguish planning approaches from tools and instruments because approaches can be seen as comprehensive packages of tools and instruments. We selected two planning approaches, adaptive delta management (ADM) (van Rhee, 2012) and the water diplomacy framework (WDF) (Islam & Susskind, 2012; Susskind & Islam, 2012) for several reasons (Table 2). First, they were both developed to deal with complex problems and uncertainty by providing a comprehensive planning approach with explicit use of adaptiveness. An additional reason to study these approaches is their origin in the messy practice of day-to-day planning. They originate from the audacious attempt of practitioners to get a grip on situations wherein environmental issues, fraught with uncertainty, interrelate with conflicts between values and stakes. Finally, both approaches are fully described and supported by a scholarly body of knowledge which forms their underlying rationality.

Table 1. Different planning approaches, tools and instruments.

Planning approaches	Tools	Instruments
Adaptive delta management	Adaptation pathways	Covenants
GeoDesign	Cost–benefit analysis	Rules
Integrated water resources management	Joint fact-finding	Standards
Strategic choice approach	Multi-criteria analysis	Subsidies
Water diplomacy	Real options analysis	Taxes
	Robust decision-making
	Scenario planning

Table 2. Case selection criteria and scores for five planning approaches.

Approach	Complex problems	Comprehensive description	Adaptiveness	Practitioner based	Scholarly embedded	Context
Adaptive delta management	✓	✓	✓	✓	✓	Netherlands
GeoDesign	✓	✓	✗	✗	✓	U.S.
Integrated water resources management	✓	∼	✗	∼	✓	United Nations
Strategic choice approach	✓	✓	✗	✗	✓	Great Britain
Water diplomacy	✓	✓	✓	∼	✓	U.S.

Our analytical focus was on the guidelines which offer the formal description of both approaches (Islam & Susskind, 2012; van Rhee, 2012). We deliberately chose to focus on the formal descriptions because we wanted to study their operationalization and not application. Study of the operationalization of approaches is necessary for choosing approaches and to study if successful applications are due to or despite the operationalization of an approach. We assumed that adaptiveness is used to handle uncertainty and were interested in two major elements: their relation in the formal operationalization and the contextual influence on this relationship (Figure 1). Our analysis included the following six steps (Table 3). We started with studying the formal descriptions to get a sense of the general themes related to uncertainty and adaptiveness in the approaches. Our next step was contextual immersion. We derived contextual data via citations in the guidelines and via conversations with the main authors of both approaches. To broaden our contextual insight, the water diplomacy workshop, held each year in Cambridge (MA), was visited. Also, several open interviews (Appendix 1) were conducted with the authors of the approaches to query their background and context, whereby uncertainty and operationalization of adaptiveness were discussed. We also held interviews with first users of ADM in the context of the Dutch Delta Programme, a multi-level governmental programme to develop water management strategies in which ADM was developed.

Figure 1. Conceptual framework.

Table 3. The used analytic strategy (Ayres, Kavanaugh, & Knafl, 2003).

Steps	Strategy	Analytic focus	Product
1	Analytic immersion in both approaches	Within each main guideline	Sense of the general themes related to uncertainty and adaptiveness
2	Contextual immersion in both approaches	Open conversations with developers of both approaches; contextual material (scholarly and non-scholarly)	Sense of the general history and broader (scholarly) context of both approaches
3	Immersion in each approach	Within each main guideline (Atlas.ti)	Identification of significant statements and themes (codes) related to uncertainty and adaptiveness
4	Comparison of statements	Across main guidelines (Atlas.ti)	Identification of commonalities and differences
5	Reconnection of established themes	Within formal descriptions and in check with contextual immersion	Corroborate identified themes to ascertain fidelity to the approaches’ context
6	Testing themes	Compare identified themes with general themes in the research question	Comparison of the material based on the found uncertainties; use of adaptiveness

Steps three and four in our research strategy were built on interpreting the main guidelines of both approaches in a sense-making process (Ayres, Kavanaugh, & Knafl, 2003; Schwartz-Shea & Yanow, 2012). We first immersed ourselves in the main guidelines. We used the programme Atlas.ti. to interprete the text and search for linkages between themes and variables. We looked at how uncertainty and adaptiveness are described in the approaches and identified statements and variables related to our research question. We undertook three formal coding sessions of the guidelines until the main themes (uncertainty, adaptiveness, their relation, context) were saturated with sub-variables and all occurrences in the formal descriptions were accounted for (Appendix 2). Next, we compared the underlying statements to identify commonalities and differences. In our fifth step, we reconnected to the context of the approaches to corroborate the identified themes and ascertain their fidelity. Our sixth and final step was to reflect on these findings by comparing both approaches in the context of the different natures of uncertainty and the specification and use of adaptiveness.

3. Results

We draw on the conceptual insights in the nature of uncertainty and distinct types of adaptiveness to elaborate how the approaches describe uncertainty and operationalize adaptiveness to deliver the outcomes they promise. First, we describe the WDF, then ADM to pinpoint their distinct characteristics. We illustrate WDF with the example of water management in the Sacramento-San Joaquin river delta and ADM with the example of replacing hydraulic structures in the Meuse river. Second, we compare the approaches and describe the found commonalities and differences.

3.1. Water diplomacy framework

WDF (Islam & Susskind, 2012; Susskind & Islam, 2012) combines insights from complexity sciences, consensus building approaches and negotiation theory. It is developed to overcome problems associated with integrated water resources management (IWRM) by building on continually evolving systems due to open boundaries and including the latest developments in negotiation theory, based on seeking mutual gains for multiple stakeholders. Water system boundaries are claimed to be set in IWRM. Assuming set boundaries leads to perceive allocation as a win–lose situation. WDF seeks a transition towards an open, networked management of water resources. In most cases such open boundaries lead to the impossibility of optimizing water allocation because of uncertainty and variability.

An open network-perspective reframes water systems as networks crossing science, policy and political fields. Water networks span across the boundaries of societal and natural processes, both influenced by and influencing the political domain. Water networks also cross different scales (e.g. spatial, temporal, jurisdictional, institutional) and within each scale they may cross different levels (Islam & Susskind, 2012, p. 49). Due to the open and interconnected perspective on networks across scales and levels, problems tend to be complex. WDF assumes that focussing on problems arising due to the complexity of water networks, dynamic and open systems allow for searching flexibility in values and stakes of actors. A network view enables actors to perceive water as a flexible resource which allows for win–win water resource allocation and mutual gains. Such mutual gains thinking is developed since the 1970s as a reflection on negotiating agreements (Fisher & Ury, 1981; Margerum, 2011).

The general rules WDF offers elaborate on the configuration and development of water systems, but the major emphasis is on how to steer collaborative processes to desired end-results. WDF equips planners to design a collaborative scheme among stakeholders to manage water networks and argues that system perspectives always depend on the values and views of stakeholders. This emphasis originates from the insight that there is ‘no agreement on the data that needed to be collected or how projections regarding future demand should be made [and] fundamental disagreement about … appropriate allocation formula’ (Islam & Susskind, 2012, p. 12). WDF thus starts from the premise of conflict in need of mediation and agreement. When mediating and seeking agreement WDF builds on diplomacy, for which it emphasizes as a starting point the establishment of good relations and skills of the planner to deal with conflict situations.

The development of the California bay-delta programme illustrates this approach (Islam & Susskind, 2012). The Sacramento-San Joaquin river delta, discharging in the San Francisco Bay (hence the acronym bay-delta) is a fragile, ecosystem while also the storeroom for water throughout California (Kallis, Kiparsky, & Norgaard, 2009). In the bay-delta, conflict arose since the 1940s about the diversion of water to particular users (Lacan & Resh, 2016). This conflict arose due to fundamentally opposite uses and complexity of the system (Kallis et al., 2009). The conflict was resolved with the development of a programme for which the stakeholders (State and federal agencies and over 30 different represented stakeholder groups) worked out a mutually advantageous solution based on the principles underlying WDF.

3.1.1. Uncertainty and handling mechanisms

WDF acknowledges that uncertainty remains a challenge in all water management efforts. Uncertainty is conceptualized in WDF based on the insight that water networks behave in unpredictable, and hence uncertain ways. This opposes the notion of forecasting in contemporary water management (Kiang, Olsen, & Waskom, 2011). Related to the reasonably far forecasting of water supplies, WDF instead ‘assumes that the supply and quality of water are more unpredictable than that (and becoming even more so, for example, because of a changing climate)’ (Islam & Susskind, 2012, p. 272). Consequently, WDF argues that uncertainty must explicitly be addressed in water network characterization.

On a meta-level water networks are characterized by the use of a diagnostic framework which builds on two variables: the amount of uncertainty and the amount of disagreement. These variables determine if water networks and associated problems are simple (small uncertainty, low disagreement), complicated (either high uncertainty or large disagreement) or complex (high uncertainty and large disagreement). The diagnostic framework implies that planners or water managers can plot uncertainty on a scale to determine the complexity of a water situation (with the other determinant being disagreement) (Islam & Susskind, 2012, p. 91). From this perspective it is clear that setting up the Californian bay-delta programme was undertaken in a particularly complex situation with high disagreement and multiple contrasting uncertainties (Lacan & Resh, 2016).

In contrast to this diagnostic framework, addressing uncertainty is made subject to the perspective of actors who convene to address a problem. Joint fact-finding should allow for mutual agreement concerning the amount of uncertainty and ways to proceed, but not necessarily solve uncertainty nor disagreement. This is derived from the complexity of water problems due to which ‘we cannot talk about finding optimal or engineered solutions unless a great many non-objective assumptions are imposed’ (Islam & Susskind, 2012, p. 8). Such non-objective assumptions undermine the credibility of water managers when they claim sole reliance on scientific or technical judgements. The amount of disagreement may very well be about uncertainty or scientific certainty. This makes uncertainty and the depending diagnostic of a water network highly subjective. Before uncertainty can be addressed by WDF it should be discursively constructed, which allows for disagreement about what uncertainty is and how it should be handled.

The tension between conceptualizing uncertainty as a variable for determining the character of a water network (simple, complicated, complex), while being subject to the perspective of stakeholders, becomes visible in the role of the planner. To characterize a network, which the planner has to do before deciding about the appropriate management strategy, the planner has to identify the amount of uncertainty and disagreement. When stakeholders convene, uncertainty depends on frames of reference or the ‘Weltanschauung’ of involved actors and is subject to consensus among the convened group of stakeholders. WDF does not indicate ambiguity as a separate aspect of water problems. Instead, it emphasizes sufficient representation of different societal frames when convening stakeholders. The consensus reached after thorough selection of stakeholders allows for incorporation of ambiguity as much as possible due to co-construction of knowledge within the process. WDF is consequent in equipping planners to take on a neutral, mediating role and managing processes, but by doing so uncertainty is strongly embedded in the agreement/disagreement continuum. Thus, the approach does not equip planners to systematically choose between management strategies for specific types of problems or networks. This is part of the consensus-seeking process, such as the process in the bay-delta example. WDF hoovers between a meta-level analysis of what a network might be and equipping planners to handle complex water networks, through open-ended interpretation of problems and provision of guidance on the procedure to develop strategies. Where does this leave planners who need tools to handle uncertainty? For this, WDF proposes collaborative adaptive management (CAM).

3.1.2. Adaptiveness

WDF argues for ‘new tools … required to model emerging water concerns’ (Islam & Susskind, 2012, p. 273). These tools include collaborative decision-making and joint fact-finding, combined in CAM. The use of CAM is proposed in order to allow for the recalibration of policies and plans and to ensure the possibility of mid-course corrections of such policies and plans, based on what WDF typifies as ‘careful’ monitoring (Islam & Susskind, 2012, p. 202). CAM is embedded in an adaptive learning orientation which ‘takes advantage of the unexpected’ (Islam & Susskind, 2012, p. 16), contrasting corrective actions through monitored change.

Adaptiveness is connected to changes on the certainty–uncertainty continuum and is tightly connected to flexibility:

Any strategy can only be optimum under certain conditions, and when those conditions change, the strategy may no longer be ideal. To survive an organization needs to be flexible and adaptive. Flexible adaptation also requires new connections and new ways of seeing things. (Islam & Susskind, 2012, p. 68)

A new way of seeing things is the reconceptualization of systems into networks where different nodes are in constant interplay. Planners can influence such nodes to resolve disagreement by creating more value than previously perceived. The negotiation turns the problem from a zero-sum game into a non-zero-sum or win–win situation. This may be done by finding the best way to interact and adapt to other players in a network to juggle conflicting constraints and achieve the best possible outcome, feasible in the network’s circumstances. To do so, WDF argues that adaptive learning (in contrast to adaptive management) is necessary. Adaptive learning intents the long-term co-production of explicit and tacit water knowledge. In this co-production, the role of planners is to recognize and understand cross-scale and cross-level dynamics to ensure management of complex networks. This management is a long-term effort based on contingent agreement which should incorporate adaptive management principles, but which is reactive to occurring change. WDF is strongly normative about the adaptiveness strived for. One of the core assumptions of WDF is: ‘the management of water networks ought to be adaptive and negotiated using a “non-zero-sum” approach’ (Islam & Susskind, 2012, p. 10; our emphasis). Adaptiveness is, thus, central to the approach.

The adaptive management scheme in the bay-delta is often praised as a successful exemplar of adaptive management (Kallis et al., 2009). This particular scheme enabled the stakeholders to manage currently uncertain future change by adapting to changing circumstances when they occur (Vlieg & Zandvoort, 2013). Adaptive management refers in this case to increasing site-specific information to propose informed adjustments, for example, by monitoring effects of interventions on biological parameters. Such adaptive management, as also proposed in the WDF is reactive to uncertain change, instead of forward-looking.

3.1.3. Outcomes

WDF promises specific outcomes or products when the approach is congruently adopted by planners. WDF claims to develop agreements which are fair, efficient and wise. These three characteristics are emphasized because:

Unless agreements are viewed as fair (by those affected), efficient (by those who have to pay for them), and wise (by those with the expertise to judge), one or more parties, even if they reluctantly sign an agreement, will look for opportunities to reopen negotiations or to ‘get even’ later. (Islam & Susskind, 2012, p. 135)

These agreements, and the approach to come to agreement in the first place, are not embedded in an institutional context. The approach deliberately ‘rejects the unquestioned authority of hierarchical governance structures’ (Islam & Susskind, 2012, p. XII), making its position highly critical regarding existing authority structures, without offering steps to transition from existing governance structures towards the proposed situation of ongoing diplomacy. It also does not offer insight in transfer options to different planning cultures or governance settings.

WDF offers not only agreements about specific interventions or resource allocations. To handle uncertainty and allow for contingent steps, sensitivity to initial configurations is just as important. Therefore, ‘the parties [in an agreement] may specify what will happen if various future events occur. Final agreements should also include dispute resolution provisions indicating how parties who fear that something has gone wrong are expected to proceed’ (Islam & Susskind, 2012, p. 147). Thus, agreements set initial steps to proceed in specific water systems, and rules for contingent steps based on believable forecasts and decisions about how to handle data gaps and interpretative differences. With such agreements, WDF claims to offer the necessary flexibility, based on adaptive learning and continuous adjustments, which should make long-term planning irrelevant.

3.2. Adaptive delta management

ADM (van Rhee, 2012) combines insights from scenario planning to explore future developments (Haasnoot, Middelkoop, Offermans, van Beek, & van Deursen, 2012), flexibility studies in economic and engineering studies (Scholtes & de Neufville, 2009) and the concept of ‘mainstreaming’ developed in climate adaptation studies (Gersonius et al., 2016). ADM is developed to better handle uncertainty in the complex context of managing water in the Netherlands. Planners are equipped to solve complex water problems and are offered tools to handle future uncertainties (van Rhee, 2012). The approach aids strategy-making with elements such as the construction of pathways, formal assessment of flexibility and scanning possible options to keep open. It also aims at combining investment and policy agendas of stakeholders.

ADM originated in the politically perceived necessity to integrate and enhance the core values solidarity, flexibility and sustainability for water management in the Netherlands (I&M, 2009). Enhancement of these values was connected to fit ADM in the historically developed institutional setting, planning culture and the set of tools in use. The integration of the above-mentioned domains of insight (e.g. scenario planning, flexibility, mainstreaming) reflect the three core values. Solidarity (which explicitly relates to intergenerational and intra-generational justice) is embedded in systematically mapping the externalities of choices. The mode of working which enhances flexibility is to anticipate future change in a transparent way through adaptation pathways (Haasnoot et al., 2012). Adaptation pathways are a tool with which planners can combine a portfolio of measures to assess their timely use under different scenarios. Sustainability relates to the mainstreaming of agendas through involvement of companies and inhabitants affected by choices and to the integration of spatial quality into strategy formulation (I&M, 2009, pp. 263–266).

ADM embraces a system perspective and emphasizes the implications for monitoring and optimizing a system relative to what are perceived unchangeable goals for safety and fresh water supply. Its premise is that the natural system is complex and dynamic, but society’s goal setting is not. Consequently, the approach builds on a notion of different trajectories or pathways which a developing water system might follow in the future and follows several steps to create strategies. Strategies consist of ‘goals, related measures and one or more development trajectories’ (van Rhee, 2012, p. 18) assuming that clear goals can be determined to design subsequent measures and development trajectories. Moreover, ADM builds on a contextual understanding of complex water systems and only very implicitly indicates the agency of water managers and planners to alter course.

ADM can be illustrated by the case of replacing seven interlinked hydraulic structures in the river Meuse (van Rhee, 2012). During the 1920s the Meuse was canalized by building shipping locks and moveable weirs to improve its functionality for transport, drainage and flood prevention (Disco & van Vleuten, 2002). In the coming decades, the structures need replacement, reason to reconsider them in the broader functionality of the river (van der Vlist, Ligthart, & Zandvoort, 2015). Determining their future use is, however, a complex problem. Regional water uses are adapted to the historic configuration of the river and future functionality depends on where and how structures are designed. ADM intends to offer insight in how to deal with uncertainty in such complex water problems.

3.2.1. Uncertainty and handling mechanisms

Uncertainty in ADM is described by using the scheme of Courtney, Kirkland, and Viguerie (1997) and distinguishes four types of uncertainty. ADM aims to tackle uncertainty about the future as described by Courtney et al.’s (1997) type two, which is uncertainty with a limited and distinct set of possible outcomes and type three which is uncertainty within a bandwidth when looking further in time (van Rhee, 2012, p. 26).

Uncertainty is not only explicitly characterized but it is also described at multiple places in the planning approach. Enhancing flexibility, for example, builds on the idea that uncertainty diminishes over time. Moreover, uncertainty is described for different types of unknown developments to which one might want to be flexible. The approach stresses that uncertainty has to be made as explicit as possible to choose the most appropriate tool to handle it. Pertaining to the nature of uncertainty, ADM primarily relates to ontic, irreducible uncertainty. ADM indicates the presence of ontic uncertainty in climate change and socio-demographic change. An understanding of these types of change will always be incomplete and predictability will always falter. Coping with such change is, however, possible by waiting, since the uncertainty at this moment might diminish over time. In the Meuse example, the current strategy is to renovate the weirs for some years to monitor changes to better predict (but still not fully) how, for example, climate change may impact river discharges.

The key connection between handling uncertainty and adaptiveness are tipping points (Kwadijk et al., 2010). These moments in time are described as tipping points ‘because going on in the current fashion becomes too expensive, technically impossible or societally unacceptable’ (van Rhee, 2012, p. 18). Tipping points indicate an expiration date at which a policy or measure is deemed not feasible anymore and translate uncertainty into questions of timing. When planners connect alternate policy strategies to scenarios of future change, tipping points can make uncertainty explicit. This offers an advantage since tipping points can be made measurable more easily compared to scenarios. Multiple uncertain variables can, thus, be compared and offer a time frame in which expiration dates will fall: ‘What are the first and last moments at which a measure or strategy does not suffice anymore and, thus, additional measures need to be taken?’ (van Rhee, 2012, p. 24).

In the Meuse, the seven weirs which manage the water level each have different tipping points with specific uncertainty about when these are reached (van der Vlist et al., 2015). Timing depends on deterioration of the structures and the effects of climate change on river discharges, but also on the demand for shipping and fresh water use along the river banks. While there are complications to determine the exact tipping points (van der Vlist et al., 2015), first use of ADM informed the particular timing of necessary replacements and the possible alternative replacement strategies regarding societal demands (van Rhee, 2012).

3.2.2. Adaptiveness

Adaptiveness in ADM intends to ameliorate the lack of predictive capacity and uncertainty about how the future unfolds. Adaptiveness helps to take adequate measures. Adequate measures are taken at the precise moment when they are needed for their ameliorating effects. To do so planners need to be flexible in time (that is to be able to postpone or advance measures) and assess measures to the wider physical and institutional context. Adequacy relates to the avoidance of lock-in or lock-out situations or a deliberate and transparent choice to get into such a situation. ADM equips planners to take short-term decisions and interventions only in light of their possible long-term impact on the system.

Building on the concept of tipping points, possible trajectories or adaptation pathways (Haasnoot et al., 2012) are made applicable for achieving medium and long-term societal goals. These trajectories are projected under different scenarios and their adequacy regarding possible but uncertain change. Strategies in ADM can be developed on different scales which has implications for the contribution of adaptiveness to handle uncertainty. On a project level, for example, an investment decision for a weir, adaptiveness is enhanced through knowing the costs and benefits with respect to scenarios of future change. This allows for seizing option value for long-term developments. For the weirs in the Meuse, an option value is the possible buy-in of additional headway to handle future draught. Such option values create adaptiveness to handle currently unknown future change. On a strategic or regional spatial scale, ADM addresses the possibility to keep options open and seize opportunities for flexibility by coupling multiple stakes. An important function is to gain insight in the possibilities to alter course if a tipping point for the current policy strategy comes close.

In addition to uncertainty, a reason for adaptiveness may be to couple agendas of stakeholders. Adaptiveness offers the opportunity to couple strategies with agendas, by alternating the decision about the use and timing of measures to seize short-term opportunities or to strategically wait until coupling might become advantageous (van Rhee, 2012, p. 35). For the Meuse, combining a weir with a bridge might create synergy. Planners may gain synergy, but need to avoid unnecessary interdependencies between agendas. This can be done by, for example, cost–benefit analysis of measures both independently and combined to see if measures can be executed separately without losing the advantage of coupling agendas.

ADM uses two types of adaptiveness. First, adaptiveness with regard to lock-in/lock-out mechanisms and the coupling of short-term interventions with long-term consequences and, second, to enable planners to enhance adaptive capacity of the managed system, by enlarging decision space and enabling flexibility. These are tightly interwoven because the first type of adaptiveness enables and creates the second. Thus, there is an intricate relationship between different types of adaptiveness (i.e. adaptive planning and adaptive capacity) underlying ADM.

In the Meuse, ADM enabled planners to handle uncertainty about replacement based on the functionality of individual structures and in relation to key nodes in the water network. For each structure prevalent uncertainties and their interdependencies on a network level were determined (van Rhee, 2012). This enabled planners to relate functionality-dependent choices (for example, retaining the river for transport or not) to uncertainties for specific structures. It did not, however, abate ambiguity due to, for example, the choice about prioritizing specific functions of hydraulic structures over others.

3.2.3. Outcomes

ADM develops ‘covenants for first and subsequent decisions’ (van Rhee, 2012, p. viii). The emphasis is put on first decisions ‘which take into account conceivable future decisions; the art of leaving decision-space open for later choices and to maintain or increase flexibility of the water- and spatial system’ (van Rhee, 2012, p. 5). These are embedded in strategies for water management and institutionally embedded in monitoring and yearly updated planning programmes. ADM’s outcome is a set of pathways which aid decision-makers in optimal timing and phasing of measures with respect to uncertain drivers of change. ADM ‘leads to a composite strategy, or a set of alternative strategies with intermediate possibilities for revisions’ (van Rhee, 2012, p. 14). Within such a composite strategy, four types of measures are distinguished: measures which are (still) effective as (parts of) the current strategy, measures that are part of an improved strategy, measures that are profitable through coupling to other agendas and are optimized regarding their timing, and measures to keep options open for future choices or different strategic directions. The outcomes are not only these measures based on societal consensus as to their desirability, ADM also equips planners to assess the effectiveness of measures and determine the best possible strategy in light of different scenarios.

3.3. Comparing the two approaches

WDF and ADM both start from the insight that predictability about the future is impossible due to the interlinked character of water systems. Linkages exist across different spatial and organizational scales, different stakes and values and non-linear causality of actions. A second common starting point is the difficult handling of uncertainty as a result of this interlinked character of water systems. WDF presumes that modelling is fraught with uncertainty and variability which leads to the impossibility to optimize water allocations. In ADM, uncertain climate change and variability in weather and climatic extremes necessitates reassessment of measures. The outcomes of both approaches are also comparable. Where WDF develops ‘agreements’, ADM develops ‘covenants’ in which both opt for reaching agreement between multiple contending parties about water management issues. The way to come to these outcomes, however, differs.

The differences can be summarized as the coupling of uncertainty to claims about information and knowledge; the conceptualization of adaptiveness in each of the approaches and the description of the fit with institutional contexts in which adaptiveness needs to be embedded. We discuss these differences in more detail below.

3.3.1. Uncertainty, information and knowledge claims

WDF starts from the premise that knowledge is essentially contested and stakeholder involvement leads to the most optimal agreements for managing water networks. ADM positions the technical production of water management strategies first and starts from the premise that more stakeholder influence might hinder the construction of adequate strategies, because stakeholders lack detailed insight in the functioning of water systems. This premise neglects other types of knowledge such as tacit, local knowledge of inhabitants of a region. WDF explicitly engages with such tacit, local knowledge to develop win–win solutions.

As shown above, uncertainty is an important concept in both approaches, while they handle uncertainty differently. In WDF uncertainty is primarily an issue for meta-level decisions about what a water problem entails regarding its degree of complexity. Planners’ decisions are based on uncertainty involved in the encountered type of problem and the related handling perspectives. Uncertainty as conceptualized in WDF cannot be resolved, but also not anticipated through specific interventions. Uncertainty arising out of stakeholder frames might be handled through joint fact-finding. Joint fact-finding is proposed to handle, not necessarily reduce uncertainty, and is about believability and mutual acceptance of information. This translates in contingency settlements which allow for renegotiation and alteration of agreements. ADM starts from the assumption that uncertainty sometimes can be resolved, or time will offer additional insight in currently uncertain issues. Sometimes uncertainty cannot be resolved. ADM assumes that some uncertainties demand different (timing of) measures than others and that not all measures are adequate. Therefore, ADM proposes to use expert groups to establish detailed insight in the adequacy of measures and to decide on issues of timing in the preparation phase of strategies. This role of expert knowledge does not align with WDF which emphasizes the use of tacit knowledge and mutual fact-finding wherein all stakeholders are engaged.

Compared to ADM, WDF seems to be better equipped to deliver policy under uncertainty. ADM focusses on ontic uncertainty, but mostly neglects uncertainties with an epistemic or ambiguous character. WDF focusses on epistemic and ambiguous natures of uncertainty while ontic uncertainty is, to some extent, reactively handled by adaptive management principles. Adaptive management is, however, made subject to consensus, which may lead to inadequate solutions with respect to ontic or epistemic uncertainty. Important to note is that, while ADM as approach lacks attention for ambiguity, ambiguity is dealt with in the Dutch institutional context in which ADM is proposed, even to the extent that this may hinder sufficient attention for the implications of ontic uncertainty. Externalizing ambiguity from ADM may produce outcomes perceived as illegitimate by some actors, especially when their frame of reference and tacit knowledge are not incorporated into the process. When confronted with complex problems both ambiguity and ontic uncertainty are in need of careful handling of planners. The two approaches do not offer specific safeguards to ensure such handling.

3.3.2. Adaptiveness to handle uncertainty

While WDF builds on a general conception of adaptive management to handle uncertainty, ADM anticipates future adaptations through a set of tools which allows for seizing opportunities and avoiding lock-in and lock-out. To handle uncertainty in WDF the notion that neither collaborative management nor adaptive management are fully suited to plan for networked situations led to the proposal for CAM (Caves et al., 2013). CAM, however, neglects the dynamic nature of complex systems, resulting in ontic uncertainty. A statement such as ‘CAM assumes that water network managers will never get everything right on the first try’ (Islam & Susskind, 2012, p. 202) and the language concerning CAM (i.e. calibration, experiments) indicate a perspective wherein searching for optimizing management solutions, such as a specific allocation of water rights, is essential. Optimization is, however, rejected by WDF (Islam & Susskind, 2012, p. 8). This rejection is necessary because good relations and diplomatic skills, essential for finding mutual gains, do only thrive by not foreclosing the solution space.

An issue less developed in both approaches is how to account for adaptiveness. Both approaches build on ongoing monitoring (for ADM building on tipping points and drivers of change; WDF by monitoring systems in negotiated agreements), but do not specify how this should be done. WDF sees monitoring as part of ‘experiments’ to recalibrate policies and reconsider long-term goals and objectives. Two issues are at stake. First, a perception of agreements as experiments may result in problems already identified for trial-and-error approaches (Moore & Hockings, 2013). For example, not properly taking into account the unforeseen effects of agreements can lock-in the system on an undesired development trajectory. ADM explicitly accounts for this issue of unforeseen effects by taking an anticipatory adaptive planning view, while WDF is partly embedded in adaptive management principles and ongoing adjustments instead of planned headroom for unforeseen corrections of policy. Second, monitoring becomes problematic, since complex situations render mid-course corrections unpredictable. Both approaches do not specify which variables need to be monitored (if possible at all) to effectuate adaptiveness. Even if some insights about what variables need to be monitored might be available, the assessment of such efforts might necessitate calibration of agreements, rendering monitoring itself obsolete.

3.3.3. Institutional context for adaptiveness

Both approaches are created in a specific institutional context. WDF builds implicitly on aspects derived from the US institutional context. This may hinder application in other situations, because it may not immediately be clear how the approach fits with local perceptions on uncertainty and adaptiveness (and other aspects of the approach) and which institutional presumptions influence the approach. For example, the vision on long-term planning, which might be rendered obsolete according to WDF, may reflect the US context in which a strong libertarian political climate obstructs long-term planning endeavours and the availability of sufficient resources for such activities. Moreover, WDF ‘rejects the unquestioned authority of hierarchical governance structures’ (Islam & Susskind, 2012, p. xii) even though WDF claims to offer a universally applicable approach for water problems. This necessitates the tailoring of the approach to assumptions and claims about specific institutional contexts, for example, related to the outspokenness of stakeholders and democratic institutions for representation implied in the approach. ADM is explicitly linked to the Dutch context and elaborates on how the approach diverges from contemporary Dutch practice and on transition steps for implementing ADM. This offers transparency as to the practical implications of its situated implementation for other contexts as well. The Dutch context, however, is characterized by long-term planning efforts and an egalitarian organized institutional structure. This may need to be accounted for when transferring ADM to other situations.

One issue related to handling uncertainty is the inclusion of stakeholders. WDF proposes to resolve the influence of different knowledge claims and perceptions on uncertainty by stakeholder inclusion throughout planning processes. ADM proposes a funnelling approach in which three groups have different functions. A large stakeholder group with representation of all stakes needs to establish and accord the objectives, principal drivers of change to be discussed and possible options for a strategy. A small group of experts does in-depth research, synthesizes details, executes uncertainty and risk analysis and keeps the process on track to develop specific and adequate strategies. Decision-makers need to adopt a strategy into the formal decision-making processes. Thus, both approaches assume outspoken stakeholder groups and the institutionalization of deliberative principles.

ADM deliberately engages with how strategies become embedded in existing institutional structures. It does not claim any overhaul is needed per se. Instead ADM offers steps to integrate the approach into an existing situation. WDF does signal the case dependency of planning water systems: ‘Consequently, one needs to be cognizant of the applicability and limitations of a given framework while analyzing case studies to derive generalizable principles that can be applied in other regions, domains, and situations’ (Islam & Susskind, 2012, p. 31). Signalling case dependency does not lead to proposing specific steps for implementing the approach. Therefore, ADM seems to be more specific about its case dependency and institutional context which allows for seeing the fit and possible misfit with other contexts.

4. Discussion and conclusions

Owing to the difficulty of planning water situations fraught with unpredictability and uncertain change, the authors of the WDF and ADM critique existing planning approaches as being not fully up to the task of handling complexity and uncertainty. Consequently, they developed planning approaches to get grip on such situations and to operationalize adaptiveness to handle uncertainty. Because we wanted to know what adaptiveness was used in planning approaches to handle uncertainty in complex water problems we compared these two approaches.

We found that the alternatives offered by WDF and ADM differ, even though both approaches propose to develop agreements and covenants after signalling that variability, uncertainty and non-linear causality problematize planning for water systems. While WDF advances CAM as a set of ongoing adjustments and continuous learning, ADM attempts to anticipate uncertain change through its adaptive planning view. We found that different types of adaptiveness are used in the approaches. This reflects the diffuse understanding of adaptiveness in scientific literature divided among something to either reactively adapt to altering circumstances, experiment to optimize solutions or agreements and proactively adapt to possible change by deliberate anticipation. A possible explanation for the difference between the approaches might have to do with the signalled weak points of existing planning schemes. WDF focusses mainly on the lacking attention for stakeholder inclusion and negotiation, fitting with more reactive styles of adaptiveness. ADM elaborates on the insufficient inclusion of long-term uncertain consequences in planning decisions, better fitting with adaptiveness through deliberate, planned anticipation. Thus, both aim to ameliorate planning of water systems, but do so on different aspects, which is reflected in the type of adaptiveness they apply.

Our results show that the approaches differ in their conceptualization of uncertainty. In both approaches frameworks are offered for addressing multifaceted situations, but these frameworks do not account for the three different natures of uncertainty (ontic, epistemic, ambiguous). This result supports previous studies where it is argued that elaboration of uncertainty characterized by its three natures simultaneously is as yet undeveloped (van den Hoek, Brugnach, Mulder, & Hoekstra, 2014). An implication of this may be that neither of the approaches congruently offers tools to handle all types of uncertainty associated with complex water problems. This can become problematic when planners want to address situations wherein uncertainties of different natures intermingle.

When contrasting the different types of uncertainty with the two frameworks, our results indicate that WDF offers a more coherent account of each of the three natures compared to ADM. However, in making adaptiveness subject to knowledge claims and consensus WDF opens up the possibility of inadequate solutions for specifically uncertainty with an ontic nature. In ADM the handling of ontic uncertainties is emphasized, while for congruently handling all uncertainties with the approach, it is less suitable. ADM lacks attention for primarily the ambiguous nature of uncertainty and some elaboration of the interrelation between different uncertainties and the proposed role of adaptiveness.

Connected to this, both approaches should better scrutinize how to transfer them to other institutional settings. To implement them in existing situations requires additional insight in possible transitory steps to embed the approaches in existing legal frameworks and local planning cultures. More clearly than WDF, ADM is already embedded in an institutional context, which corroborates the conclusion of Dewulf and Termeer (2015, p. 768) that ‘[o]verall the current institutional context for ADM in the Netherlands is quite favourable’. A lack of attention for the possible fit and misfit of approaches with the institutional context seems to be an important omission to account for uncertainty, next to limited resources, high costs, long duration and large technical requirements for managing uncertainty (Woodruff & Stults, 2016). Further study into the different design choices involved when using planning approaches and tools for handling complex situations fraught with uncertainty may alleviate some of these omissions.

Although this comparative analysis helps to identify how planners may use approaches to handle uncertainty by adaptiveness in the planning of water systems, a limitation is that we studied only two approaches based on their conceptualization of uncertainty and adaptiveness. Our study focused on the use of adaptiveness to handle uncertainty, not on the overarching operational design of the approaches or their fine-tuning in day-to-day planning practice. Each of these limitations offers themes for future research, whereof an important direction is to study approaches in empirical cases to draw out insight in the use of approaches to address uncertainties and propose adaptive interventions adequate for the consequences of prevalent uncertainties. This should, however, be studied alongside the way of operationalizing concepts into adequate planning approaches as such.

Adaptiveness to handle uncertainty is expected to garner increasing attention as the malign consequences of decisions under uncertainty become more prevalent. The attention and development of adequate planning approaches to account for adequate interventions and adaptiveness amid uncertainty will probably only increase due to uncertain climate change. Our results suggest that planning approaches such as WDF and ADM can enable the handling of uncertainty through adaptiveness, but need to be improved based on specific insights in what adaptiveness and uncertainty entail for the development of water management plans.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Mark Zandvoort is researcher in adaptive planning and water management at Wageningen University, the Netherlands.

Maarten van der Vlist is special Associate Professor at Wageningen University and principal expert adaptive delta management and land use planning at Rijkswaterstaat, the Netherlands.

Adri van den Brink is Professor and Chair of Landscape Architecture at Wageningen University.

ORCID

M. Zandvoort http://orcid.org/0000-0001-6207-7562

A. van den Brink http://orcid.org/0000-0002-8403-7470

Appendices

Appendix 1. Interviews

1. Policy advisor and member ADM team staff delta commissioner

2. Author ADM

3. Strategist DPIJ (1)

4. Strategist DPIJ (2)

5. Policy advisors Rijkswaterstaat (2 persons)

6. Policy advisor Ministry of Infrastructure and the Environment

7. Manager and strategist Delta programme Rivers (2 persons)

8. Senior policy advisors from DPIJ (2) and staff delta commissioner (1)

9. Author 1, WDF (formally spoken with on three separate occasions)

10. Author 2, WDF

Appendix 2. Coding scheme

The coding scheme including themes, coded variables and number of occurrences in the formal descriptions of adaptive delta management and the water diplomacy framework.

Table

Themes	Variables (codes)	Number of Occurrences
		ADM	WDF
Uncertainty	Uncertainty	53	44
	Uncertainty characterization	18	3
	Predictability	3	12
	Complexity	6	19
Adaptiveness	Adaptiveness	36	31
	Adequate measures	26	12
	Flexibility	15	7
	Link with uncertainty	15	10
	Lock-in/out	10	2
Context	Institutional context	25	20
	Problem description	7	23
	Stakeholders	16	28
	Decision-making	29	8
	Framing	5	20
	Transparency	4	3
	Responsibility	2	10
	Normativity	3	9
	Planning context	14	9
	Role of the planner	11	40
	Building blocks	22	43
	Products	16	26
	Interdependencies	13	17
	Physical context	8	2
	Scales	2	5
	Time	28	7
Descriptors	Description of the approach	13	27
	Information	1	20