Assessing the Capacity to Govern Flood Risk in Cities and the Role of Contextual Factors

: Sea level rise and increased storm events urge cities to develop governance capacity. However, a cohesive conceptual and empirical-based understanding of what governance capacity implies, how to measure it, and what cities can learn, is largely lacking. Understanding the inﬂuence of context is critical to address this issue. Accordingly, we aim to identify crosscutting contextual factors and how they prioritise different elements of governance capacity to address urban ﬂood risk. In doing so, a framework of nine conditions and 27 indicators is applied in two Dutch cities and two cities in the United Kingdom. Three crosscutting contextual factors are identiﬁed that may explain differences in capacity-development priorities: (1) ﬂood probability and impact; (2) national imposed institutional setting; and, (3) level of authority to secure long-term ﬁnancial support. Capacity-priorities include, the recent political devolution in the UK, which emphasizes the role of citizen awareness, stakeholder engagement, entrepreneurial agents, and the overall necessity for local capacity-development. The Dutch focus on ﬂood safety through centralised public coordination reduces ﬂood probability but inhibits incentives to reduce ﬂood impacts and lowers public awareness. In conclusion, the three identiﬁed contextual factors enable a better understanding of capacity-building priorities and may facilitate learning between cities.

between cities, practitioners, technology facilitators, and scientists. These city platforms often aim to achieve a transition or transformation towards sustainable urban practices. Examples include C40, 100 Resilient Cities, Europe's Covenant of Mayor Aadapt and IWA's Principles for Water-Wise Cities [8][9][10][11]. Despite these valuable and highly necessary attempts, it remains difficult to utilise the large potential gain of mutual learning between cities because each city has to operate in a different contextual setting and may have different priorities. Accordingly, it is a scientific challenge to develop new approaches that can facilitate meaningful exchange of knowledge, experiences, and learning practice in order to enable these cities to develop the governance capacity necessary to transform themselves towards sustainable practices [12,13].
The scientific literature on climate change adaptation, water governance, and adaptive capacity has identified a plethora of barriers, enablers, and conditions that influence the problem-solving capability of the public and private sectors (e.g., [14][15][16]). However, despite the valuable insights provided by these studies, the body of literature has arguably not yet established sufficient conceptual coherence to empirically validate the large number of theoretical premises. First, many identified conditions are predominantly based on theoretical considerations that are not sufficiently supported by empirical findings [16][17][18]. Second, the definitions of these identified barriers, enablers, and conditions are often neither made explicit, nor is it clear how they are operationalized, measured, and how they relate to one another [15,17]. Third, many barriers or enablers are applied within the context of individual case studies, often without considerable efforts to identify general patterns or transferable lessons [16,19]. These limitations call for a more coherent diagnostic analysis regarding the capacity of cities to govern flood risk.
Conceptual definitions generally establish themselves through repeated testing, evaluations, and refinements. Accordingly, the notion of governance capacity can be considered as the product of a conceptual integration of empirical knowledge from many different case studies and theoretical debates. Therefore, the concept of governance capacity is, in an implicit manner, a rationale that has to be derived from comparative analyses. In order to do so, a coherent, empirical-based diagnostic framework is required that can be used to assess existing governance capacities in different urban contexts. In this way, a database can be developed that can be used to identify overarching patterns and transferable lessons across a range of case studies, while, at the same time, it can improve our understanding of cause-effect relations in detail through individual cases. Such a mutual purpose requires a thorough understanding of how contextual factors influence the governance capacity. However, it is not well-understood in what way governance capacity is influenced by context and which contextual factors account for differences between urban flood risk governance [15,16,20]. Such an understanding can however be considered as a precondition for learning, in particular city-to-city learning.
In this paper, we aim to contribute to a better understanding of the context-specific capacity development priorities of urban flood risk governance in the United Kingdom (UK) and the Netherlands. Our second aim is to explore crosscutting contextual factors that may explain the observed differences. In order to fulfil this research aim, a comparative case-study approach will be applied in four cities in both countries (see Section 3). Section 2 provides the conceptual framework behind the analysis of governance capacity that will be applied in this paper and introduces the definition of crosscutting contextual factors. Section 3 first provides a rationale for the selection of the four case-study cities. Second, the data collection methods are provided. In Section 4, the case study results are presented and crosscutting contextual factors that may impede, enhance, or prioritise different elements of governance capacity are explored. Section 5 discusses other potential context factors and the implications of the identified contextual factors for interpreting governance capacity analyses in other cities. We end with the conclusions in Section 6.

A Framework to Analyse Context-Specific Governance Capacities
Like other consensus concepts-such as adaptive capacity, resilience, or sustainability-concrete definitions for governance capacity are strongly debated and differ considerably. Nevertheless, a few communalities can be identified [21]. First, governance capacity is related to the ability of actors to address collective problems across organisations, and often includes multiple levels of governance [22]. Second, governance capacity is formed through actor's interactions, which are shaped by social-institutional settings and allocation of resources [23]. Third, actors' frame of reference, which includes their interests, values, and culture, shape interactions, and thereby the capacity to address joint problems. Since actors' interactions are by definition complex, unpredictable, and susceptible to external social, economic, and environmental developments, governance capacity in itself does not guarantee effective change, but can rather be considered as a precondition or enabler. The institutional setting, rules, and regulations that enable actors to collaborate and address shared problems is generally referred to as institutional capacity (e.g., [24]). Governance capacity is broader as it also includes the role of resources and discourses.
Due to this broad scope, there is little agreement on which indicators or proxies are most valid to assess governance capacity. Consequently, the researchers' normative assumptions of what governance capacity entails leads to implicit accenting of certain issues over others. However, both impasses and opportunities result from many interconnected processes with often unforeseen effects that cannot be explained by a single barrier or event [20]. Hence, in order to understand dynamic governance mechanisms, it is necessary to consider a broader range of conditions that together form the capacity to address flood risk. Based on an extensive literature review, Koop et al. [21] developed a diagnostic framework that coherently assesses the most important conditions that together determine the capacity of cities to govern water challenges. This empirical-based Governance Capacity Framework (GCF) consists of nine conditions (Table 1) and three dimensions: knowing, wanting, and enabling. The "knowing" dimension refers to the need to involve actors to be aware, understand, and learn about the risks and impacts of strategic choices and policy. In this dimension, the level of awareness (condition 1), existence of useful knowledge (condition 2) and the network's ability to continuously learn (condition 3) are being assessed. The "wanting" dimension relates to the necessity that actors commit, cooperate, act upon ambitions, and use their skills to find solutions. More specifically, the stakeholder engagement process (condition 4), management ambition (condition 5), and the role of agents of change (condition 6) are being assessed. Finally, the "enabling" dimension analyses the network potential to address water challenges (condition 7), the financial viability (condition 8), and existing policy instruments and action plans that actors can use to implement policies that address flood risk (condition 9). Table 1 provides an overview of the nine conditions and 27 indicators that form the capacity to govern flood risk.
The framework is designed with the aim to consistently analyse the governance capacity of cities across various world regions in order to identify opportunities for mutual learning between them. However, the governance capacity of cities is a result of multi-level governance processes and it is shaped by the local contextual setting. In order to enable meaningful exchange of knowledge, experiences and learning practices, contextual factors have to be taken into account. According to the principle of parsimony, we aim to identify the most influential reoccurring contextual factors that need to be considered. These contextual factors are referred to as 'crosscutting contextual factors'. Table 1. Overview of nine governance conditions and 27 indicators that comprise the governance capacity analysis [21]. The Likerts scoring system for each indicator is included in the supplementary information.

Indicator
Pre-Defined Question

Condition 1: Awareness
Awareness refers to the understanding of causes, impact, scale and urgency of flood risk.

Community knowledge
To what extent is knowledge regarding the current and future risks, impacts, and uncertainties of flood risk dispersed throughout the community and local stakeholders which may results in their involvement in decision-making and implementation?

Local sense of urgency
To what extent do actors have a sense of urgency, resulting in widely supported awareness, actions, and policies that address flood risk?

Behavioural internalization
To what extent do local communities and stakeholders try to understand, react, anticipate and change their behaviour in order to contribute to solutions regarding flood risk?

Condition 2: Useful knowledge
The availability, transparency and cohesiveness of information that actors can use.

Information availability
To what extent is information on the water challenge available, reliable, and based on multiple sources and methods, in order to meet current and future demands so as to reveal information gaps and enhance well-informed decision-making?

Information transparency
To what extent is information on the water challenge accessible and understandable for experts and non-experts, including decision-makers?

Knowledge cohesion
To what extent is information cohesive in terms of using, producing and sharing different kinds of information, usage of different methods and integration of short-term targets and long-term goals amongst different policy fields and stakeholders in order to deal with the water challenge?

Condition 3: Continuous learning
Continuous learning refers to the level of social learning ranging from refining current practices, critical investigation of fundamental beliefs or questioning underlying norms and values.

Smart monitoring
To what extent is the monitoring of process, progress, and policies able to improve the level of learning (i.e., to enable rapid recognition of alarming situations, identification or clarification of underlying trends)? Or can it even have predictive value?

Evaluation
To what extent are current policy and implementation continuously assessed and improved, based on the quality of evaluation methods, the frequency of their application, and the level of learning?

Cross-stakeholder learning
To what extent are stakeholders open to and have the opportunity to interact with other stakeholders and deliberately choose to learn from each other?

Condition 4: Stakeholder engagement process
Stakeholder engagement is required for common problem framing, gaining access to a wide variety of resources and creating general support that is essential for effective policy implementation.

Stakeholder inclusiveness
To what extent are stakeholders interact in the decision-making process interaction (i.e., are merely informed, are consulted or are actively involved)? Are their engagement processes clear and transparent? Are stakeholders able to speak on behalf of a group and decide on that group's behalf?

Protection of core values
To what extent (1) is commitment focused on the process instead of on early end-results? (2) do stakeholders have the opportunity to be actively involved?
(3) are the exit procedures clear and transparent? (All three ensure that stakeholders feel confident that their core values will not be harmed.)

Progress and variety of options
To what extent are procedures clear and realistic, are a variety of alternatives co-created and thereafter selected from, and are decisions made at the end of the process in order to secure continued prospect of gain and thereby cooperative behaviour and progress in the engagement process?

Indicator
Pre-Defined Question

Condition: 5 Management ambition
Policy ambitions assesses if current policy is ambitious, feasible, well-embedded in local context and if it forms a cohesive set of long-term and short-term goals within and across sectors.

Ambitious and realistic management
To what extent are goals ambitious (i.e., identification of challenges, period of action considered, and comprehensiveness of strategy) and yet realistic (i.e., cohesion of long-term goals and supporting flexible intermittent targets, and the inclusion of uncertainty in policy)?

Discourse embedding
To what extent is flood risk management policy interwoven in historical, cultural, normative and political context?

Management cohesion
To what extent is policy relevant for flood risk management and coherent regarding (1) geographic and administrative boundaries; and (2) alignment across sectors, government levels, and technical and financial possibilities?

Condition 6: Agents of change
In order to drive change, agents of change are required to show direction, motivate others to follow and mobilize the resources required.

Entrepreneurial agents
To what extent are the entrepreneurial agents of change enabled to gain access to resources, seek and seize opportunities, and have influence on decision-making?

Collaborative agents
To what extent are actors enabled to engage, build trust & collaborate and connect business, government & sectors in order to address flood risk in an unconventional & comprehensive way?

Visionary agents
To what extent are actors in the network able to manage and effectively push forward long-term and integrated strategies which are adequately supported by interim targets?

Condition 7: Multi-level network potential
Urban flood risk governance involves a plethora of actors and interests from all levels of government, organizations and (private) stakeholders. For sustainable solutions, working in networks is an essential determinant for effective solutions.

Room to manoeuvre
To what extent do actors have the freedom and opportunity to develop a variety of alternatives and approaches (this includes the possibility of forming ad hoc, fit-for-purpose partnerships that can adequately address existing or emerging flood risk challenges)?

Clear division of responsibilities
To what extent are responsibilities clearly formulated and allocated, in order to effectively address the flood risk challenges?

Authority
To what extent are legitimate forms of power and authority present that enable long-term, integrated and sustainable solutions for flood risk challenges?

Condition 8: Financial viability
The continuation of flood risk funding is crucial which needs to be supported by affordable flood risk related services and an overall willingness to pay for floor risk management 8. 1 Affordability To what extent are flood risk related services and climate adaptation measures available and affordable for all citizens, including the poorest? 8.2 Consumer willingness to pay How is expenditure related to flood risk perceived by all relevant stakeholders (i.e., is there trust that the money is well-spent)?

Financial continuation
To what extent do financial arrangements secure long-term, robust policy implementation, continuation, and risk reduction?

Condition 9: Implementing capacity
Implementing capacity is about the effectiveness of policy instruments. The effectiveness is also related to the compliance to policy and regulation and the familiarity with flood emergency plans.

Policy instruments
To what extent are policy instruments effectively used (and evaluated), in order to stimulate desired behaviour and discourage undesired activities and choices?

Statutory compliance
To what extent is legislation and compliance, well-coordinated, clear and transparent and do stakeholders respect agreements, objectives, and legislation?

Preparedness
To what extent is the city prepared (i.e., there is clear allocation of responsibilities, and clear policies and action plans) for both gradual and sudden uncertain changes and events?

Case Study Selection
In order to better understand the context-specific governance capacity development priorities of urban flood risk governance, two countries with (1) considerable flood vulnerability and (2) with a long track record in flood risk management, were selected: the UK and the Netherlands. In both countries, two cities were selected that resemble the flood risk challenges and exemplify the policy efforts of their country. In the UK, Milton Keynes (261,800 inhabitants) and Leicester (394,000 inhabitants) were selected. Like most other flood prone cities in the UK, both selected cities are vulnerable for river flooding and inundation of urban areas due to rainfall runoff. In the Netherlands, Rotterdam (638,000 inhabitants) and Amsterdam (833,624 inhabitants) were selected. As for at least 25% of the Netherlands, both cities are situated below sea level and potentially exposed to sea flooding, river flooding, and inundation due to rainfall runoff. For both countries, an 'old' city and a 'new' city were selected to better represent the diversity within both countries. The 'old' cities of Leicester and Amsterdam both have an historic city centre and high population densities (respectively, 4494 people km −2 and 5042 people km −2 ) that complicate the implementation of physical flood risk measures. On the other hand, the 'new' cities of Milton Keynes and Rotterdam have lower population densities (respectively 2584 people km −2 and 2920 people km −2 ), which results in greater flexibility to adapt urban infrastructure in order to cope with flood risks. As experiences with recent flood events is an important determinant for adaptation efforts (e.g., [25,26]), all four selected cities did not experience any significant flood events in the last decades. The European Union (EU) Water Framework Directive (WFD) applies to all four cities and the cities are more or less equally affected by the 2008 financial crisis. Furthermore, climate change mitigation and adaptation have become a central theme in the national policy in both countries [27].

Data Collection
Each of the 27 indicators that are listed in Table 1 have a pre-defined question and a Likert scoring system that ranges from very encouraging (++) to very limiting (−−) the overall governance capacity (the Likert scale indicator scoring is provided in the supplementary information). A three-tiered approach was applied to score each indicator in all four case studies in order to validate findings by different sources:

Preparatory Desk Study
A desk study of primary sources, such as governmental reports, legal documents, and secondary sources, such as scientific literature, websites, newspapers, and etcetera, resulted in a substantiated preliminary score for each indicator.

Stakeholder Selection & Interviews
Stakeholders were identified, categorized, and their responsibilities, relations, and main interactions were mapped in order to select a range of stakeholders that include the largest variety of interests. At least one stakeholder that, respectively, represents the government, the market, and civil society was included, as suggested by Lange et al. [28]. We also included at least one knowledge institute or university per city. The following organisations were selected for interviews. Leicester: University of Leicester, Trent Rivers Trust, Leicester City Council, local consultancies, Flood Warden, and the Environment Agency. Milton Keynes: Milton Keynes City Council, Cranfield University, local consultancies, the Highways Agency, the Environment Agency, and Bedford Group of Internal Drainage Boards. Amsterdam: Waternet, Municipality of Amsterdam, local consultancies, University of Amsterdam, and the Municipal Health Service (GGD); Rotterdam: Evides (drinking water company), Delta Programme Climate adaptation, local consultancies, Rotterdam University of Applied Science, Municipality of Rotterdam, and the water board Hoogheemraadschap Schieland en de Krimpenerwaard. In many cases, multiple persons from the identified key stakeholders were selected with different roles, areas of expertise, and responsibilities in order to minimize the risk of bias and unravel socially desirable responses. In total, 30 in-depth interviews were conducted. The face-to-face interviews lasted for about an hour and were made anonymous in order to reduce socially biased answers. The pre-defined questions that are listed in Table 1, formed the common thread for the researcher. In the interviews, questions were reformulated in alignment with the person's background and expertise. In fact, some interviews focused only on a few indicators. For each indicator, at least five different persons were interviewed as a minimum requirement. The scores were determined based on the preparatory desk study, interviews, and additional collection of reliable and accessible information that may ratify or provided nuances to the interview findings.

Verifying Findings
The overall results were checked by asking feedback from all interviewees. After the incorporation of this feedback, the final scores were determined. By providing arguments and references, the indicator's scores and pre-defined questions were substantiated.

Exploring Crosscutting Contextual Factors
Based on the GCF analyses, the indicator scores that mostly differed between the four cities, were examined more closely through in-depth interviews and document analyses in order to explore crosscutting contextual factors that could explain these differences. In each of the previously mentioned interviews, national contextual differences between the UK and the Netherlands were explored as well. In addition, flood management related documentation in both countries were studied (documents such as the Pitt review [29] in the UK and the Delta Programme [30] in the Netherlands) and a detailed analysis of local, regional, national, and international policies that are related to flood risk management was performed. Table 2 shows the overall results of the GCF analyses. Section 4.1 will present the national flood risk management context in the UK, followed by the results of the city of Milton Keynes and Leicester. In Section 4.2, the national flood risk management context of the Netherlands is provided, followed by the results of the cities of Rotterdam and Amsterdam. In Section 4.3, crosscutting contextual factors are explored that may impede, enhance, or prioritise different elements of governance capacity.

Flood Risk Governance in the UK
The UK has a long history in flood events, mainly with respect to rivers bursting their banks, as well as local water nuisance due to unsustainable drainage systems. The UK has a risk-based approach with a recognition that floods cannot be fully prevented [31]. In this context, citizens and property owners are considered to have a fundamental role. For example, the UK's private flood risk insurance system appeals to individuals being aware of their flood risk and taking action to protect their property [32,33]. The summer floods of 2007 proved to be a determining event. Between May and July, the highest rainfall intensity was recorded since records started in 1766 [34]. About 55,000 properties were inundated along the major rivers Severn, Don, and Thames. The total damage amounted to £4 billion [35].
A review led by Sir Michael Pitt [29] identified that a lack of clarity over which organisations were responsible for surface water flooding and emergency action led to contrasting and uncoordinated messages. Moreover, flood impact data was found to be fragmented and replicated. Consequently, the Pitt review called for a permanent centralised coordinating body. As a result, the Government's Department for Environment, Food, and Rural Affairs (DEFRA) took coordinated action [36]. First, the Environment Agency (EA), responsible for the main rivers, coasts and reservoirs, got more authority to maintain a strategic overview [37]. Second, Regional Flood and Coastal Committees (RFCCs) were established. At the regional level, the EA is required to consult RFCC and get approval for implementing and expending revenues raised [38]. Third, more power was directed to local authorities through the Lead Local Flood Authority (LLFA). The LLFAs are obliged to develop, maintain, apply, and monitor a strategy for local flood risk management of surface runoff, groundwater, and ordinary watercourses within its area [37,38]. The LLFA is required to register flood risk assets, investigate flood incidents, and promote Sustainable Drainage systems (SuDS) through close cooperation with the highway authority, local risk management authorities, the regional Internal Drainage Boards (IDBs), and through public consultation. The sewer system is management by private companies who fund their operations through water bills under the supervision of Ofwat, the economic regulator for water, and sewerage services in England and Wales.

Milton Keynes
Milton Keynes is a designed city and national frontrunner in sustainability and innovation. Of the 261,800 inhabitants, 1753 properties in the older, lower areas have a high flood risk [39]. The citizens, who have recently experienced water nuisance, apply mitigation measures, such as walls and airbricks, to protect their property (indicator 1.3 behavioural internalisation). However, most citizens have little knowledge or commitment with flood risk challenges (indicators 1.1 community knowledge and 1.2 local sense of urgency; [39]) and are not willing to pay more than they already do (indicator 8.2 consumer willingness to pay). Flood risk information is provided transparently through real-time warning systems that monitors river and sea levels: https://www.gov.uk/check-flood-risk. The EA flood risk maps are complemented with detailed assessments from the local authorities providing accurate information about current and future flood risks (indicator 2.1 information availability).
Moreover, local flood risk prediction maps are available for periods of five days and further ahead, and a free flood warning service for house owners and businesses is available: https://www.gov.uk/ sign-up-for-flood-warnings (indicator 2.2 information transparency).
Milton Keynes' City Council has a coordinative task as LLFA. In their local flood risk management strategy [39], the suggested agreements and goals are realistic and moderately ambitious, taking into account funding limitations. Each objective is translated into three measures, each with intermittent targets that have clear deadlines (indicator 5.1 ambitious and realistic management; [38]). All stakeholders are free to engage in projects and have the opportunity to express their concerns and provide feedback on draft plans through workshops and on paper (indicator 4.1 stakeholder inclusiveness). Different departments within the City Council, are responsible for narrowly defined flood-related tasks, each with different drivers and budget allocations. For example, interests for spatial development can contest flood risk management goals. The division of responsibilities (indicator 7.2) with respect to SuDS is somewhat complex. Highways departments are tasked with ensuring good traffic flow through, amongst others, SuDS. Other actors outside City Council also have an important role. Anglian water is tasked with the sewer system. Parks Trust is a charity who takes care of Milton Keynes parks and green spaces and the Bedford Group of IDBs manages the regional drainage system. The jurisdiction of the IDB is inconsistent with the river catchment area, leading to difficulties in both upstream and downstream coordination. Despite many efforts, the fragmented organisational structure leads to limited knowledge cohesion (indicator 2.3) and suboptimal management cohesion (indicator 5.3).
Due to recent devolution of political power and cuts in national funding [40], local authorities have to find new funding sources, such as local investors. The City Council reviews the funding opportunities every six months. However, this system may reduce financial security for long-term more proactive measures (indicator 8.3 financial continuation). Likewise, the window of opportunity for entrepreneurial agents (indicator 6.1) to innovate, is also financially limited. Hence, the national imposed institutional setting and financial constraints are considered to be important contextual factors that impact the local governance capacity.

Leicester
Leicester is situated in the wide flat River Soar valley and it is particularly vulnerable to heavy downpours and prolonged periods of rain. From the East and West hills, a number of large watercourses flow quickly towards the River Soar through urban areas. Major developments projects are expected upstream, which could increase the surface water runoff and flood vulnerability. It is estimated that about 1915 residential and commercial properties are at risk of a 1 in 75 years flood [41]. Leicester, therefore, recognises the importance that residents understand the causes, risk and impact of their properties being flooded, how to respond to emergencies, and how they can be involved in local decision-making. This is also reflected in the high availability, transparency, and cohesion of flood risk information (condition 2 useful knowledge; [42]).
In general, differences in culture and language form barriers for people's awareness, preparedness, and their ability to recover from flood events [43]. As one of the most culturally diverse cities in the UK, Leicester allocates much resources to effectively communicate with multicultural communities. For example, flood risk information is translated in multiple languages. It also poses extra challenges to include citizens in the stakeholder engagement process (condition 4). Consultation procedures are clear, adhere to national requirements, and all relevant stakeholders and responsible risk authorities are involved. For example, everyone could provide feedback on the draft local flood risk management strategy through consultation meetings and online platforms. However, citizen involvement heavily depends on a select group of well-informed, non-transient, and often highly educated citizens. Stakeholders can also raise issues themselves in council or ward meetings. However, the level of influence that stakeholders have on the end-result is somewhat limited. Engagement is mainly via consultation and not that frequently through focus groups that co-produce knowledge to explore optimal solutions. Leicester is open for this type of stakeholder engagement, but is also restricted by financial resources.
As a result of diminished national funding, Leicester needs to fund flood-defence projects through their own resources, partnership funding, or local developers. However, these funding mechanisms may limit the long-term financial security necessary to pro-actively adapt to challenges of climate change and land-use change. Consequently, the role of local individuals that provide a long-term vision, promote initiatives, bring actors together, and mobilize the required local resources, has become critical (indicator 6.1 entrepreneurial agents). The room to manoeuvre (indicator 7.1) that these agents of change need to effectively seize opportunities for new projects and innovations is somewhat limited. The division of responsibilities, interests, and tasks are divided over many actors (indicator 7.2 clear division of responsibilities) and they pose constraints for the coordination of different policies (indicator 5.3 management cohesion). The EA is responsible for the main river Soar while Leicester City Council (LCC) is the LLFA and is responsible for flood risk and spatial planning. The Leicestershire county and district councils carry out flood risk management works on minor watercourses surrounding the municipality. The water and sewerage company Severn Trent has the duty to maintain the sewer system and drain their area. Successful attempts have been made to integrate the development of green space with flood risk management [44]. However, the flood risk goals are much less synergetic with respect to national targets for housing development that the LCC has to comply with. In terms of contextual factors, the city's vulnerability to heavy downpours and prolonged periods of rain requires a high level of citizen awareness to adapt, anticipate, and cope with 'unavoidable' flood events. Moreover, limited national and local flood management authority leads to uncertainties with respect to local financial resources.

Flood Risk Governance in the Netherlands
About 55% of the Netherlands is flood prone and about 25% lies below sea level, including the country's main economic district and largest cities. In 1953, the levees in the south-eastern delta region were unable to withstand a major storm surge. In total, 1836 people died, over 2000 square kilometres of land was flooded, and the damage was estimated at €5.2 billion [45]. In response to this disaster, the national Delta Plan was initiated which included the embankment of the estuaries in the south-eastern areas. The Delta works were designed to prevent such a large-scale disaster to ever happening again. The Delta Plan has a strong emphasis on engineering flood defence structures and it is controlled by the national authority 'Rijkswaterstaat'. The Dutch government fully compensates flood damages. However, a number of river flood incidences in the 1990s changed this dominant discourse. In 1993, the river Meuse burst through its banks and flooded one-fifth of the southern province of Limburg. Moreover, the risk of river flooding led to the largest post-war evacuation of about 80,000 people in 1995. As a consequence, the 'Room for the Rivers' policy was developed, which emphasized the role of spatial planning in dealing with climate change induced extreme weather events. Accordingly, the focus shifted from reducing flood probabilities to reducing flood impacts [46].
The national government's Delta Commission formulates the strategic goals and provides a financial structure for flood defence and freshwater provision within the Delta Programme. The provinces are responsible for the regional operationalization of these national strategies and to manage groundwater bodies. The Ministry of Infrastructure and Water Management is responsible for protection against the sea, main rivers, and canals. The water boards manage the regional water systems. The water boards raise their own public taxes to recover their costs and have their own elected representatives. Finally, the municipalities have a duty of care regarding the sewer system, rainwater, and groundwater, which is financed through a specific tax. Due to significant damages of recent storm events and new more extreme climate predictions, the Delta Programme [30] announced extra investments for urban areas. Municipalities are required to perform a 'water stress test' in order to map their long-term resilience and to identify the required action with respect to flooding, drought, and heat stress [30].

Rotterdam
Rotterdam is home to 638,000 people and has Europe's biggest port. The city has large flood risk challenges as it is situated below sea level in the delta of the river Rhine and Meuse, and its main canal 'de Nieuwe Waterweg' is openly connected to the North Sea. Rotterdam has a long tradition of flood risk management and the city's water safety is highly dependent on the national dikes and sea barriers (indicator 5.2 discourse embedding). Rotterdam is one of the safest delta cities in the world. The main responsibilities lie with the municipality, the water boards, and the Dutch government through the Delta Programme. Local policies strictly adhere to national flood risk policies and safety standards [47]. At the local level, the division of responsibility (indicator 7.2) between public and private actors is more determined on a case-by-case fashion. Regular inter-sectorial meetings, in particular, between the city's three water boards, the municipality, and main businesses in the harbour enable cohesive water management. Nevertheless, inter-sectorial exploration of synergies can be further enhanced (indicator 3.3 cross-stakeholder learning and 5.3 management cohesion). For example, through smarter, more integrated monitoring, and evaluation of the city's drainage system (indicator 3.1 smart monitoring and 3.2 evaluation).
Overall, long-term goals are ambitious and implemented by intermittent targets that are embedded in various policies (indicator 5.1 ambitious and realistic management; [47,48]). Within Rotterdam's Climate Proof programme, climate change adaptation is considered to be a means for creating employment, social cohesion, and citizen engagement, which enhances economic growth. Delta technology and knowhow are referred to as the blue-green economy and are considered as export products. Investments in a safe and vital infrastructure is expected to attract international investments, ensure long-term growth, and prosperity. There is much room for experimentation. Many small-scale pilots and experiments such as floating houses, farms, hotels and forests are supported by local authorities. Initiatives, such as collective city gardens, nature-based playgrounds, and child friendly neighbourhoods, are combined with climate change adaptation goals and improve the attractiveness and liveability of Rotterdam. Important climate change adaptation projects are realised, such as a water square, which combines water storage with the improvement of the quality of urban public space. Another example is the a large underground parking lot that can also store excess stormwater. In some cases, more unconventional collaboration with the private sector has been established leading to for example the development of tidal parks along the main canals: the 'Nieuwe Maas' and the 'Nieuwe Waterweg' (indicator 6.2 collaborative agents). Visionary agents (indicator 6.3), who push for long-term, integrated, and climate adaptive flood risk management at the political or strategic level are most active at the national scale. In Rotterdam, entrepreneurial agents (indicator 6.1), who initiate new concepts, innovations, and ideas, are most active.
Whilst Rotterdam is not a rich city according to Dutch standards, the municipal and regional taxes are generally being considered affordable (indicator 8.1 affordability). The High Water Protection Programme, which is a part of the Delta fund, ensures funding for water safety measures. An additional benefit of this financial structure is that the municipality saves time and resources otherwise spend on acquisition (indicator 8.3 financial continuation). In the context of high potential flood impact and substantial future challenges, the city is forced to be proactive. In these efforts, Rotterdam is strongly supported by the authority of national and regional policy and institutions, which is an important contextual factor that provides the necessary financial resources for Rotterdam to be innovative and proactive.

Amsterdam
Amsterdam is an old and densely populated city with a complex hydrological setting. The city is protected from the North Sea, the Lek River, lake Marker, and regional water bodies through a complex system of dikes, dams, and sluices. The city is connected to the North Sea via a canal running through the city centre. North and south of the canal, the city is protected by dikes under national authority which ensure high safety standards (flood risk of once in 10,000 years). In some areas, a lower safety standard of once every 1250 years is operational. Amsterdam has applied a flood safety approach. However, recently the national Delta Programme has initiated a 'Multi-layer Safety' (MLS) approach, which is being piloted in six voluntary cases, one of them is Amsterdam. MLS consists of three layers [30]: 1.
Reducing flood probability through flood defence infrastructure 2.
Reducing flood impact through adaptive spatial planning 3.
Reducing flood impact by preparing flood response strategies A thorough scanning according to the MLS revealed that suboptimal spatial planning, in particular, with respect to vital infrastructure, largely increases the potential flood impact. In particular, the harbour area Westpoort is vulnerable. The Westpoort area is vital for the supply of electricity for a third of the city. Furthermore, flooding of its wastewater treatment plants, data centres and the chemical industry would lead to large scale damage. Despite a long-term preparedness to anticipate flood risks, the flood response strategy was found to be insufficiently equipped (indicator 9.3 preparedness; [49]). Cost-efficiency of spatial adaptation is low due to the low flood probability. However, the city is pioneering with a strategy that combines flood adaptation with projected infrastructure refurbishments and with measures to reduce heat stress, air pollution, or water nuisance. For example, soil remediation can be combined with measures to increase the ground-level. This integral cost-efficient approach can slowly reduce the flood impact over the next decades. However, the implementation of this strategy is hampered by limited awareness and knowledge of flood risks beyond the water authorities (indicator 1.1 community knowledge; [49]). Amsterdam is served by one municipally owned utility for water, called 'Waternet'. Waternet manages the entire urban water cycle, including drinking water, water safety, surface water and wastewater transport, and treatment (indicator 7.2 clear division of responsibilities). This integrated approach and close ties with the municipality and regional water board is unique and it provides sufficient authority (indicator 7.3 authority) and room to manoeuvre (indicator 7.1 room to manoeuvre) for individual agents to coordinate the long-term implementation of cost-efficient synergies (condition 7).
By 2040, the city expects an additional 70,000 houses within its borders, resulting in further densification of sealed areas, such as roofs, streets, and parking spaces. A climate-induced increase in the frequency and intensity of storm events will further increase the city's vulnerability for water nuisance. The Amsterdam Rainproof programme was created to make Amsterdam resistant to the increasingly common downpours. Adaptive measures, such as constructing green roofs, urban gardening, and the use of rainwater for toilet flushing are being mainstreamed into urban planning. However, it is challenging to do this in a city with limited awareness. Hence, there is not yet a fully shared ambition, rather low commitment, and a lack of an inter-sectorial policy and coordination. More understandable and cohesive knowledge provision (indicator 2.2 information transparency and 2.3 knowledge cohesion) may facilitate a better integration between urban planning, climate adaptation, and flood impact reduction objectives. The new national imposed obligation to perform a water stress test may provide an opportunity for improvement [50]. The national imposed institutional setting leads to a clear division of responsibility but it also limits the involvement of citizens and stakeholders beyond the water management authorities. Moreover, a strong focus on reducing flood probability complicates local efforts to reduce the impact of potential flooding.

Crosscutting Contextual Factors for Cities
Reflecting on the indicator scores within the four case studies (Table 2), we can identify a few key differences ( Table 3). The root cause of these differences may be explained by crosscutting contextual factors. Moreover, the relative importance of some indicators differed between cities in the two case-studies in the UK and the Netherlands. In particular, indicators 1.2 local sense of urgency, conditions 2 useful knowledge, condition 4 stakeholder engagement process, and indicator 6.1 entrepreneurial agents, were found to be relatively important within the context of the UK. Whereas, condition 1 awareness, indicator 8.2 consumer willingness to pay, and indictor 9.3 preparedness, may require most attention in the Dutch context. The observed differences in relative importance, and key differences in indicator scoring, were examined in more detail through in-depth interviews, document analyses, and a multi-level policy analyses. The aim of this explorative exercise was to find crosscutting contextual factors that can explain these differences. As a result of this explorative exercise, we have identified three crosscutting contextual factors: (1) flood probability and impact; (2) national imposed institutional setting; and, (3) level of authority to secure long-term financial support.

Flood probability and impact
The hydro-physical setting largely determines the probability and impact of flood events and pre-selects the viable solutions. The Netherlands faces flood challenges that are characterised by a low probability but high impacts and a short warning time [51]. As a consequence, high safety standards are required that are accomplished through structural flood prevention measures. Improving flood defences in vulnerable reclaimed areas-the 'Polders'-increases feelings of safety that stimulate investments and further economic development in these areas. It becomes a logical choice to invest in flood defence of increasingly valuable social and economic assets. Due to the large investments that have already been made, each additional investment receives an increasing return creating a path-dependency. This self-reinforcing phenomenon has also been described as the 'levee effect' [52]. On the contrary, in the UK most flood risks tend to have a relatively lower impact but higher probability and it is possible to predict and prepare for floods well in advance. In this hydro-physical setting, not all floods can be prevented.
These diverging processes that are related to flood probability and impact have important repercussions for the role and responsibility of individuals, their expectations and trust that they have in water management authorities [53]. In the Netherlands, the government monopoly on flood safety has greatly reduced the involvement of citizens and the private sector. Moreover, the government has compensated damages that are caused by major floods, while flood insurances do not exist. Consequently, most people are rather disconnected from flood challenges and take their safety for granted. Measures to reduce the flood impact (indicator 9.3 preparedness) are typically not cost-efficient and are difficult to implement due to limited awareness beyond the water sector. On the contrary, the UK does not apply a minimum safety standard. A cost-benefit analysis fully determines the optimal protection at the local scale. Individuals can therefore be exposed to a great variety of flood probability and impact.
The UK's private flood risk insurance system requires individuals to be aware of their own flood probability and impact, and act accordingly by protecting and insuring their property [32,33]. In this context, the availability and transparency of local flood risk information for citizens becomes critical (indicator 2.1 information availability and 2.2 information transparency). Accordingly, these indicators score high for both Milton Keynes and Leicester. Citizen awareness also becomes critical, and in fact, an interesting difference was found between the analysed cities in the UK and in the Netherlands. Awareness of local authorities was found to be highest in the Netherlands, whereas citizen awareness was low (indicator 1.1 community knowledge and 1.2 local sense of urgency). In the English cities, this was rather different. Leicester has a higher citizen behavioural internalisation, which can be explained as a necessity given the high flood probabilities. In fact, knowledge, awareness, and behavioural internalisation of citizens can be considered as key indicators in the UK context of considerable food probability and appeal on individual to take their responsibility. The flood probability and impact characteristics largely explains the different management pathways that both countries have taken. Overall, this contextual factor lead to differences in condition 1 awareness, condition 2 useful knowledge, and in particular, to indicator 9.3 preparedness. These elements are particularly relevant for cities in the UK to help citizens to cope with higher flood risks and particularly relevant for Dutch cities to reduce their flood impact in the long run.

National imposed institutional setting
The existing institutional context defines, to a large degree, which actors will be involved, how they act, and which new initiatives emerge. In fact, new initiatives are likely to be discussed in already existing coordination bodies in order to avoid large transaction costs that are involved in setting up new bodies [54]. The institutional setting in both countries is rather different as a result of the 1953 Dutch catastrophe and the social-political reactions afterwards. The 1953 event has led to a strong discourse of flood safety coordinated by the central government through the Delta Programme. Flood risk is considered as a matter of public safety and national priority. Regional water management is controlled by the water boards. Water boards have separate elections and raise their own tax. In turn, municipalities are responsible for urban drainage, spatial planning, and the sewer system. Urban flood risk management is strongly driven by national legislation and policy leading to cohesive knowledge production (indicator 2.3 knowledge cohesion) and largely integrated management practices (indicator 5.3 management cohesion). The 'water stress test', that was mandated by the Delta Programme for all Dutch cities, is a clear illustration of this national coordination. Hence, the institutional setting is characterised by a national monopoly on flood safety and public institutions responsible for regional and urban flood safety. The dominance of a single public governance arrangement in the Netherlands seems to limit the scope, interaction and learning of other actors, in particular with respect to the reduction of flood impacts (indicator 1.2 local sense of urgency, 1.3 behavioural internalisation, and 3.3 cross-stakeholder learning, [54]).
Traditionally flood risk management in the UK was focused on drainage of mainly agricultural land, organised through IDBs in the lowlands. From the 1970s onwards, priorities shifted from agriculture towards urban flood protection and ecological non-structural measures within a river basin approach [27]. The central government incrementally changed from almost completely decentralised towards more centralised decision-making and funding. However, a process of political devolution can be observed throughout the last decade. In particular, the Cameron administration's 'National Flood and Coastal Erosion Risk Management Strategy' [40] ushers for a shift in responsibility towards the LLFA while the central government continues to coordinate policy.
The government has no statutory duty of care to protect land or property from flooding, but has only permissive powers. Private insurance companies provide cover against floods for residential properties and the majority of commercial buildings since the 1950s. Due to technological progress that increases knowledge about flood risk exposure, the pricing system has become much more accurate in revealing previously unknown cross-subsidy [55]. The risk-based approach led to a great variety in premiums including some 'uninsurable' properties where governmental rehousing programmes or higher safety standards are necessary. This national imposed institutional setting may jeopardise stakeholders' core values, such as their house prices, insurance, and liveability (indicator 4.2 protection of core values). Finally, the responsibilities for emergency planning, spatial planning, and emergency response may be allocated to different local risk management authorities. The flood response strategy therefore tends to be fragmentally organised, whereas proactive flood mitigation is impeded due to a lack of a resourceful authorities that can address SuDS measures at the river basin scale (indicator 9.3 preparedness). In fact, flood risk management largely depends on SuDS measures, particularly in upstream areas which are managed by a variety of local organisation. At the city scale, SuDS also involve the LLFA and water companies that are responsible for sewer drainage system. Therefore, the division of responsibilities, roles, and tasks is rather dispersed and each actor is only accountable for their own often narrowly defined tasks (indicator 7.2 clear division of responsibilities), leading to suboptimal use of policy instruments (indicator 9.1 policy instruments).
The differences in national imposed institutional setting between the UK and the Netherlands also emphasizes different governance conditions. In the Netherlands, city's need to pay extra attention with respect to condition 1 awareness and condition 3 continuous learning, particularly with respect to learning together with stakeholders outside the water sector. In the UK, condition 4 stakeholder engagement process is of particular important because many decisions that impact them are made locally. In addition, the city's multi-level network potential requires priority (condition 7), given the rather fragmented division of tasks and roles.

Level of authority to secure long-term financial support
The liberal governance style in the UK regards the task of the government as ensuring the most cost-effective outcome for taxpayers' money [56]. Accordingly, the allocation of funding to LLFAs is done on a case-by-case cost-benefit evaluation based on nine Outcome Measures [57]. Since the general election in 2010, a process of political devolution also introduced significant cuts in national funding in response to the state budget deficits. In order to fill this funding gap, local projects that do not or only partially qualify for central funding require a significant financial contribution from local communities, industries, or governmental agencies; a process that is referred to as 'partnership funding' [40]. Different studies have indicated that these local investments are often initiated after recent flood events through active engagement of Flood Action Groups (FAGs) (e.g., [25,43]). In these cases, awareness, knowledge, and learning (conditions 1, 2 and 3) have to be well-established in order to successfully apply for funding. In particular, cross-stakeholder learning (indicator 3.3) is key in these cases. However, in most cases, such a financial structure tends to be erratic and may inhibit investments in more holistic and proactive measures [58].
In our four case-study cities, we recognised a well-known pattern that shows that often the non-transient well-educated citizens tend to organise themselves and gain a significant influence on local decision-making through various forms of co-management and knowledge co-production (e.g., [59,60]). Therefore, the partnership funding system tends to work appropriately in well-educated middle class communities but not so well in more deprived areas [25,43]. Hence, flood protection levels are diverging in the UK and the local agents of change (condition 6), the stakeholder engagement process (condition 4) and overall capacity-building have become critical for ensuring adequate flood safety standards throughout the country. This is particularly challenging in cities with more deprived communities that lack significant recent floods-experiences (e.g., [25,43,61]), such as the city of Leicester.
In the Netherlands, differences in flood safety between cities is rather limited. The Delta Programme ensures long-term funding of flood defence through the Delta Fund. In period 2017-2031, about €17 billion is reserved, for which 66% is to ensure water safety measures. Every year, the funding programme is extended with one year [30]. Regional flood risk is managed by the water boards. Water boards have elected representatives and have the authority to raise tax for regional flood risk management. Likewise, the municipality raises public taxes to finance the sewer system. In this way, long-term financial security is largely ensured, irrespective of political turns, and to some extent, financial crises. Hence, the level of authority to ensure financial support is largely related to condition 6 agents of change. In the Netherlands, visionary agents have secured long-term financial support. In the UK, local entrepreneurial and collaborative agents are essential to gain access to resources, seek and seize opportunities, and to have a significant impact on local decision-making.

Discussion
In this paper, we found three crosscutting contextual factors that may explain the root cause of the observed differences (Table 3) between the analysed cities: (1) flood probability and impact; (2) national imposed institutional setting; and, (3) level of authority to secure long-term financial support. The three crosscutting contextual factors provide a useful narrative to interpret the impeding and enhancing element of governance capacity and help cities in the UK and the Netherlands to identify priorities for flood risk management. Moreover, these contextual factors may enable a better interpretation and prioritisation of activities to increase the capacity to govern flood risk in European cities, because they also adhere to the EU WFD, have comparable levels of wealth and demographics, and they all apply democratic principles. Further research may help us to understand whether these contextual factors also apply to other world regions, or that other contextual factors need to be considered as well to understand the most important impeding and enhancing elements that determine the urban capacity to govern flood risk. Such crosscutting contextual factors may include social-cultural factors. For example, the way that different cultures experience time, frame problems and conceive solutions [62], public-private arrangements [63], or the perception and communication of environmental risk [64][65][66], may influence the capacity-building process [67]. Moreover, urban feedback loops may be different in other world regions. For example, Rahmasary et al. [68] found that insufficient solid waste collection and treatment led to sewer clogging and urban flooding in Bandung, Indonesia. Such urban flooding also poses the risk of malaria outbreaks and substantial traffic congestion [69]. Moreover, ground subsidence as a consequence of groundwater over-abstraction aggravates urban flooding. Hence, in order to address flood risks in such an urban context requires an improved provision of basic services, such as drinking water, sanitation, and solid waste collection.
The type of assessment being applied in this research provides a snapshot and is merely an indication of what might be expected in the long-term urban transformation process. In order to identify the overarching lessons and to provide applicable and effective knowledge to individual cities, the consistent empirical-based analysis of governance capacity in combination with the identification of influential contextual factors are required. The three crosscutting contextual factors emphasize different elements of governance capacity in both countries. The higher flood probability, decentralised institutional setting, and recent political devolution in the UK emphasizes the responsibility of individuals to pursue their own interests. National budget cuts requires local public and private stakeholders to financially contribute to flood management measures. Within this context, condition 1 awareness, 2 useful knowledge, and 4 stakeholder engagement process appear to be essential to empower citizens and local actors and to create the right conditions for them to fulfil the responsibility that is given to them. In particular, the role of individuals that gain access to resources, seek and seize opportunities, and influence decision-making becomes critical (indicator 6.1 entrepreneurial agents).
In the Netherlands, the low flood probabilities, national monopoly on flood safety, and long-term financial continuation of flood safety programmes, results in low awareness beyond the water authorities (condition 1). However, sea level rise, increased river discharges, storm events, and ground subsidence necessitate alternative approaches focussed on flood preparedness and the reduction of flood impact (indicator 9.3 preparedness). This will be challenging because these measures are often not cost-efficient due to low flood probabilities and prioritises condition 1 awareness. In both countries, the role of local digital social platforms may provide interesting opportunities to engage citizens and local stakeholders in decision-making and improve awareness and behavioural change [70].
At large, individual case studies can provide a deep contextual understanding of the policy process at the cost of generalizability [71]. On the other hand, comparative studies parallel several case studies across contexts in order to explain variation, but at the cost of contextual understanding [72]. The identified contextual factors may serve both purposes. It helps to understand the variation between cities and countries and provides guidelines for individual cities to understand their capacity-development priorities. The literature related to variables limiting or enabling the governance capacity to address water challenges in the context of climate change lack coherence [15,17]. Conceptual rationales and individual case studies provide a plethora of plausible suppositions that are, however, not sufficiently tested on generalizability [16][17][18][19]. This study illustrates a diagnostic framework to measure governance capacity consistently across various urban contexts in order to identify overarching patterns, test conceptual presumptions, and validate results that are obtained from individual case studies in order to allow for theory-building. A precondition for such an exercise is to properly account for key crosscutting contextual factors. Based on the principle of parsimony, we limited the number of contextual factors to the minimum required to interpret differences in governance capacity between cities with varying national backgrounds. The crosscutting contextual factors can identify capacity-development opportunities and priorities, and provides a heuristic to pool cities with similar contexts or challenges to enhance mutual learning between them. In this way, the learning potential between cities can be utilised by providing a concrete, applicable, and empirical-based assessment frame that bundles scientific insights regarding the main contextual factors, barriers, and enablers that determine the capacity of cities to govern water-related challenges. Such a coherent assessment in dozens of cities would provide a unique opportunity to test hypotheses' and identify transferable lessons for cities to better address water-related challenges that are based on structured empirical research. Such an approach may facilitate and accelerate the necessary transformation of cities in the face of the unprecedented challenges of water, waste, and climate change in a rapidly urbanizing world [1].

Conclusions
Urbanization, sea level rise, and extreme rainfall urge cities to further develop their capacity to govern their flood risks. A cohesive conceptual understanding of what governance capacity implies, how it can be measured, and what cities can learn from existing practices, is largely lacking. Empirical studies may be key to better understand capacity-development opportunities with respect to the various contextual factors at play. Accordingly, we aimed to contribute to a better understanding of the context-specific capacity development priorities of urban flood risk governance in the UK and the Netherlands, and explore crosscutting contextual factors that may explain the observed differences in both countries. We found that the institutional setting in the UK and recent political devolution and national austerity measures enlarged differences in flood safety standards. In this context, the role of citizen awareness, useful knowledge, stakeholder engagement process, and entrepreneurial agents of change become critical components of governance capacity. On the contrary, the Dutch focus on flood safety through centralised public coordination with long-term financial continuity results in high flood safety standards. However, this approach also inhibits incentives to reduce flood impacts and lowers awareness, as most citizens take flood protection for granted. The three crosscutting contextual factors that we have identified provide a useful narrative for cities in the UK and the Netherlands to improve their capacity to govern flood risk. These contextual factors may also apply in other European cities. More research is necessary to gain a deeper understanding of governance capacity development in other world regions and for other urban environmental challenges. We found that urban governance capacity is, to a large extent, a product of multi-level governance processes and therefore cities have to respond to broader national or international contextual factors in order to identify and seize available opportunities to improve their capacity to govern flood risk. A thorough understanding of the different elements of governance capacity and how broader national and international contextual factors influence them is therefore indispensable.