Flood risk management in the Canadian prairie provinces: Defaulting towards flood resistance and recovery versus resilience

Abstract

Major flood events are likely to happen more frequently and be more severe under changing land use and climatic conditions. Adapting to floods using resilience-based flood risk management (FRM) policies and initiatives is a more appropriate solution than relying solely on flood defence structures or disaster recovery programmes. The primary authority for FRM in Canada is the provinces, but in practice, the policies and responsibilities are distributed across complex, multi-departmental, multi-scalar systems of government. To examine the extent to which institutional arrangements for FRM facilitate or constrain FRM resilience, this paper uses a case study of FRM policies, instruments and practices in the three Canadian prairie provinces where floods have been particularly severe in recent years. Document analysis provided insights into current FRM policies and instruments while semi-structured interviews with 34 individuals working in an FRM capacity informed the roles and responsibilities for FRM implementation. Results indicate that the current FRM policies and instruments across the prairie region have the basic requirements for flood resilience. However, flood resilience is inherently challenged by institutional fragmentation, lack of clarity of FRM roles and responsibilities, and policy layering and competing mandates, which biases FRM towards resistance and recovery solutions. To go beyond solely coordinating flood emergency response and recovery, the paper suggests an overarching regional or national FRM strategy and boundary organisation to coordinate roles and responsibilities for the specific purpose of flood resilience. This requires an agency with the mandate to manage FRM policy instruments, and clearly allocate decision authority amongst the multiple levels and layers of FRM governance.

Le changement climatique et le changement d’usage des terres risquent d’augmenter la probabilité que des désastres naturels, tels quel les inondations, surviennent. Au lieu de dépendre uniquement sur la construction de structures de prévention et des plans de reprises après désastres, un plan de gestion de risques d’inondations (PGRI) basé sur la résilience est une solution plus efficace en cas de désastres pour les environnements affectés. Les autorités chargées des PGRIs au Canada sont les gouvernements provinciaux, où les PGRIs sont distribués à travers une multitude de départements gouvernementaux. Afin d’analyser l’impact des PGRIs au Canada, une étude de cas démontrant les politiques et pratiques des PGRIs dans trois provinces canadiennes récemment affectées par de grandes inondations, a été utilisée. De plus, une analyse de documents a donné un aperçu des pratiques courantes dans la gestion de risques d’inondations, et 34 entretiens avec des responsables des PGRIs ont aidé à expliquer les rôles est responsabilités du processus d’implémentation. Les résultats indiquent que les pratiques actuels des PGRIs sont adéquates pour assurer un fonctionnement basique à travers les prairies. Cependant, la répartition interdépartementale et le manque de clarté des rôles et responsabilités des PGRIs, ainsi que les demandes de différentes plateformes institutionnelles nuisent à l’efficacité de la résilience aux inondations. Au lieu d’implémenter seulement des projets d’intervention d’urgence et de relèvement, cet article suggère une stratégie principale des PGRIs pour mieux coordonner les rôles et responsabilités au niveau provincial et fédéral. Ceci demande l’établissement d’un group responsable à l’implémentation et la gestion des PGRIs, ainsi que la répartition claire des pouvoirs décisionnels à travers les départements responsables des programmes.

Introduction

The frequency and severity of major flood events is expected to increase globally under changing climatic conditions (Whitfield 2012; Winsemius et al. 2016; Vitousek et al. 2017), with extreme flood losses expected to more than double in some regions by 2050 (Jongman et al. 2014). In Canada, between 1970 and 1990, more than 10 flood disasters were recorded, costing 27 lives, the evacuation of over 34,000 people, and damages of ~$4.6 million CAD; between 1990 and 2017, 130 flood disasters were recorded, killing 21 people, evacuating 215,458 people and costing an estimated $8.1 billion CAD (Public Safety Canada 2017). Changes in flood frequency, severity and timing will have significant implications for flood management, infrastructure, land use, the insurance industry and the Canadian economy (El-Jabi et al. 2016).

The traditional approach to flood risk management (FRM) has been to control or manage flood events, typically through flood control structures designed to contain or divert flood waters (Klijn et al. 2008; Shrubsole 2013), and to control or manage human behaviours through laws and regulations that restrict certain land uses in flood-prone areas (Zevenbergen and Gersonius 2007; Morrison et al. 2017). Given the stochastic nature of floods and their increasing severity, coupled with rapidly changing land uses (Henstra and Thistlethwaite 2017) and anthropogenic climate change (Vincent et al. 2015), efforts to control floods have proven insufficient for protecting lives and infrastructure (Klijn et al. 2008; Park et al. 2013; Rouillard et al. 2015). In response, scholars and government agencies are lobbying for resilience-based approaches (EU Flood Directive 2007/60/EC; McEwan and Jones 2012; Rosner et al. 2014; Gilissen et al. 2016; Morrison et al. 2017). Flood resilience, defined simply as the ability of society to avoid, cope with, recover from and adapt to flood events (see Schelfaut et al. 2011), is based on a combination of resistance-based strategies that seek to prevent or control extreme flood events with adaptive strategies that focus on mitigating, coping with and recovering from floods when they occur (Morrison et al. 2017).

There is a growing body of research on resilience across multiple natural resource sectors (Cosens and Williams 2012; Bakker and Morinville 2013, Clarvis et al. 2014). A consistent theme is the importance of institutional arrangements – policies and programmes and the network of agencies and actor–structure relations responsible for their implementation – as a requisite for creating more resilient social and ecological systems (Cook and Bakker 2012; Ferguson et al. 2013; Clarvis et al. 2014; United Nations Framework Convention on Climate Change 2014). However, despite a growing understanding of resilience and the importance of supportive institutional arrangements, the formulation of policies or programmes that promote resilience, the mechanisms for coordinating interests, and the processes for implementation and management actions pose enduring challenges (Lautze et al. 2011; Schmidt 2013; Chilima et al. 2017; IFAD 2017).

Considerable research has focused on understanding the hydrology of flooding (e.g. Fang and Pomeroy 2008; Modrick and Georgakakos 2015), developing better predictive tools (e.g. de Castro et al. 2013; Ward et al. 2015) and looking back on major flood events (e.g. Szeto et al. 2015; Ahmari et al. 2016; Shook 2016). Research has also addressed the challenges of fragmented governance, and the need for greater coordination, capacity and responsiveness (e.g. Gober and Wheater 2014), but there has been limited detailed analysis of whether and how existing institutional arrangements influence the implementation of FRM and support FRM resilience (Morrison et al. 2017). This paper examines the extent to which institutional arrangements for FRM facilitate or constrain FRM resilience. The focus is on Canada’s prairie provinces, which have been subject to major flood events in recent years, of increasing severity (Whitfield 2012; Gober and Wheater 2014), but the policy approaches to FRM, and the reported experiences of those engaged in FRM practice, pose important learning opportunities for other regions and jurisdictions across Canada and, perhaps, internationally.

Methods

Study area

Our study focuses on Canada’s prairie provinces – Manitoba, Saskatchewan and Alberta. The costs associated with disaster spending under the federal Disaster Financial Assistance Arrangements (DFAA) programme have more than quadrupled over its 45-year history (IBC 2015), with nearly 60% of all payments due to extreme weather events issued to the prairie provinces (OPBO 2016). The majority of these payments were due to flood events, and the proportion of disaster spending due to flooding on the prairies has increased from 57% over the lifetime of the programme to 82% between 2005 and 2014 (OPBO 2016). A recent special issue of the Canadian Water Resources Journal on floods in Canada highlighted many of the recent major flood events in the prairie region, including their physical attributes, impacts and management responses. In it, Pomeroy et al. (2016), for example, describe 2013 flood events in the South Saskatchewan and Elk River basins, with water levels nearing historic events of the late 1800s, resulting in more than a dozen communities declaring local states of emergency. Flooding of the Bow and Elbow rivers in the southern Alberta was described as Alberta’s worst-ever natural disaster (Brun and Whitfield 2016). The Assiniboine River flood one year later, caused by heavy precipitation in mid-summer, resulted in states of emergencies being declared in both Saskatchewan and Manitoba. Ahmari et al. (2016) report that the summer flood was unprecedented in the previous 130 years of observation of the Assiniboine River; similar claims were made about the record flooding of the Assiniboine that occurred only 3 years earlier (Brimelow et al. 2015).

Flood management in Canada is largely carried out at the provincial level. However, municipalities, watershed agencies and the federal government each play a role in FRM policy, programme implementation, land-use zoning and disaster recovery assistance. The prairie provinces exemplify the complex, multi-level, multi-actor governance environment (see Gober and Wheater 2014) through which flood risk is managed in Canada, and so provide a good case study for understanding the challenges that exist in implementing policies aimed at resilience.

Data collection and analysis

Data were collected via semi-structured interviews and an analysis of FRM policies and instruments. A total of 60 individuals working in a FRM capacity in the three prairie provinces were contacted for interviews, of which 34 participated (Table 1). An initial group of participants was identified through organisational websites for people or departments that were likely to be involved in FRM, with further participants identified through a snowball sampling design (Gifford et al. 2009). Most interviews were conducted over the phone, with a small number carried out in person. Two interviews involved multiple people – one with a group of two, and one with a group of four. Interviews were on average 60 min long. One provincial government department requested to provide written responses as they felt it was the best way to coordinate among the four people in their organisation engaged in FRM. For confidentiality, the identity of interview participants is withheld in the reporting of results and participants are instead identified as belonging to one of the stakeholder groups and the region in which they work.

Table 1. Study participants by stakeholder group and location.

	Federal government	Provincial government	Municipal government	Other stakeholders^a	Total
Non-province specific	4	–	–	7	11
Alberta	–	2	5	1	8
Saskatchewan	–	2	1	4	7
Manitoba	–	5	2	1	8
Total	4	9	8	13	34

^a Academia, consultancy, non-governmental organisations (NGOs), private business.

Interviews were semi-structured based on key themes in the FRM resilience literature (Morrison et al. 2017) and explored the perceived distribution of roles and responsibilities for FRM, how different institutions are involved in FRM, how institutions interact, policy influences, what influences the effectiveness of policy, and what data, information and knowledge are needed to support effective FRM resilience. Participants were broadly asked the same questions; however, some differences occurred among groups in recognition of their differing levels and types of involvement in FRM. For example, questions to non-government participants about monitoring the effectiveness of FRM policies were less likely to be relevant. Interviews were transcribed and coded thematically using QSR NVivo v. 10 software based initially on the structured interview topics. Within each topic, results were analysed and coded to identify similar themes over seven rounds of coding, and then used to develop a series of more broad, generic statements describing commonly held FRM perspectives.

Interviews were supplemented by an analysis of FRM policies and strategies. A web-based search was used to identify flood-related policy instruments from federal, provincial and selected municipal government departments. For the purposes of this study a ‘policy instrument’ is any document intended to influence the practice of FRM, including: acts or legislative requirements; regulations; by-laws that relate to permitting, infrastructure or land-use zoning; funding programmes for flood mitigation, preparedness, and disaster recovery; and guidance documents for government departments, businesses, individuals or others on flood preparation or risk management. The policy instruments examined were current to the start of 2017.

The policy search was done prior to the interview process and again following the interviews as participants were asked to identify any policies they were aware of that influence FRM. This two-pronged approach allowed us to review what is available to the public, as well as to capture those key policies that were identified by FRM professionals as influential in practice. Not included in the documents reviewed were any policies that deal with technical aspects of water management, such as storm water treatment operations or reservoir and dam operational guidelines. Although important for how floodwaters are managed, technical operating guidance was considered more relevant to FRM management than governance.

Policy instruments were categorised based on type: strategies, legislation, programmes, and guidance and information. Strategies outline a set of guiding principles, set broad goals and objectives and provide a framework for a coordinated set of FRM programmes or actions. Legislation enshrines FRM principles in law, assigning legal responsibility and allowing enforcement of those FRM principles. Programmes capture actions and activities that are funded to address well-bounded and specified aspects of FRM, such as erosion control, community defence or relocation programmes. Guidance and information provide standard approaches to FRM and information. Guidance is both technical and lay, such as guidance on flood mapping, building design or emergency planning, as well as public guidance on flood preparation and recovery. Information includes data that are shared or available for assisting FRM, such as flood maps, hydrological measurements and other data.

Each policy instrument was then categorised based on its primary focus, using the five FRM approaches described by Gilissen et al. (2016): (1) prevention; (2) defence; (3) mitigation; (4) preparation; and (5) response and recovery. Prevention seeks to minimise flood risk to people, property and infrastructure. Defence aims to increase resistance to flood waters through structures such as dykes or dams. Mitigation seeks to accommodate flood waters to reduce peak flows, and improve the ability of buildings and infrastructure to cope with floods. Preparation aims to reduce the impact of floods and protect people, property and infrastructure during an emergency. Recovery seeks to return the socio-ecological system to its pre-flood state. Collectively, these approaches capture the proactive and reactive nature of FRM resilience.

Results

Results are presented below in three sections: (1) policy instruments and their approach to FRM; (2) the division of FRM roles and responsibilities as perceived by interview participants; and (3) commonly expressed strengths and challenges to current FRM across the prairie provinces.

FRM policy instruments

There were 65 FRM provincial and federal policy instruments identified that apply to the prairie provinces. Ten of these are strategies, 14 are legislative instruments, 19 are programmes, and 22 are guidance and information documents (Table 2). Of the five broad approaches to FRM identified by Gilissen et al. (2016), the two most common instruments were preparation and response (43%), and prevention (40%). Defence was addressed by 32% of policy instruments, mitigation by 25% and flood recovery by 23%.

Table 2. Flood risk management (FRM) policy instruments applicable in Canada’s prairie provinces by type, owner and approach.

				FRM approach supported
Policy instruments (n = 65)		Owner^a	Flood prevention	Flood defence	Flood mitigation	Flood prep. and response	Flood recovery
Strategies (n = 10)
	An Emergency Management Framework for Canada	F	x	x	x	x	x
	National Disaster Mitigation Strategy	F	x	x	x	x
	Resilience and Mitigation Framework for Alberta Floods	AB	x	x	x	x
	Land use framework	AB	x
	Water Security Agency 25-year plan	SK	x	x	x	x
	Manitoba Emergency Plan & Flood Annexe	MB				x	x
	Area Structure Plans (AB)	M	x
	Inter-municipal Development Plans (AB)	M	x
	Municipal Development Plans	M	x
	Emergency response plans	M				x
			8	4	4	6	2
Legislation (n = 14)
	Municipal Government Act	AB	x
	Water Act	AB	x	x	x
	Emergency Management Act	AB				x	x
	The Planning and Development Act	SK	x			x	x
	Provincial Drainage Regulations	SK		x	x
	Statement of Provincial Interests Regulations	SK	x
	Water Security Agency Act	SK	x	x	x
	The Emergency Planning Act	SK				x	x
	Designated Flood Areas Regulations	MB	x
	Dyking Authority Act	MB		x
	Emergency Measures Act	MB				x	x
	The Planning Act	MB	x
	Water Resources Administration Act	MB	x	x
	Land use bylaws	M	x
			9	5	3	4	4
Programme (n = 19)
	Build Canada Fund	F		x
	National Disaster Mitigation Program	F		x	x
	Growing Forward 2	F			x
	Disaster Financial Assistance Arrangements	F				x	x
	Flood Hazard Identification Program	AB	x	x	x	x
	Provincial Mitigation Program	AB		x	x	x
	Agricultural Watershed Enhancement Program	AB			x
	Alberta Community Resilience Program	AB		x	x
	Erosion Control Program	AB		x			x
	Floodway Relocation Program	AB	x
	Watershed Resiliency and Restoration Program	AB		x
	Disaster Recovery Program	AB					x
	Emergency Flood Damage Reduction Program	SK				x
	Provincial Disaster Assistance Program	SK					x
	Erosion Control Assistance	SK		x	x
	Rural Water Control Assistance	SK		x	x
	Community Flood Protection Program	MB		x
	Disaster Financial Assistance (program)	MB					x
	Individual Flood Protection Program	MB		x
			2	11	8	4	5
Guidance and information (n = 22)
	National Building Codes	F	x
	National Flood Mapping Guidelines	F	x			x
	Risk-based land-use guidance	F	x
	Water Survey of Canada Flood Information	F				x
	Emergency preparedness guide	F				x
	FRM Guidelines for the Location of New Facilities	AB	x
	Guidelines on extreme flood analysis	AB	x			x
	Stormwater Management guidelines	AB		x	x
	Flood Choices for Albertans	AB	x
	Flood hazard map application	AB	x
	Alberta Municipal Recovery Toolkit	AB					x
	Personal and Family Preparedness	AB				x
	Nine steps for Emergency Plan Development	SK				x
	Flood Information	MB				x
	Flood Recovery	MB					x
	Preparing for a flood	MB				x
	Health and Safety	MB				x
	Evacuation	MB				x
	Managing Stress	MB					x
	Prepare	MB				x
	Recover	MB					x
	Respond	MB				x
			7	1	1	12	5
			Flood prevention	Flood defence	Flood mitigation	Flood prep. and response	Flood recovery
	Total number of documents: 65	Total	26	21	16	28	15
			40%	32%	25%	43%	23%

^a F = federal government; AB = Alberta government; SK = Saskatchewan government; MB = Manitoba government; M = municipal government.

Individual FRM policy instruments often address different approaches, with 40% of instruments addressing three or more of the five approaches. The most common FRM approaches identified were flood prevention and flood preparation. Legislation was most strongly concerned with: (1) enabling flood prevention (64%) via acts that support land-use planning and the relocation of people and property from areas of high flood risk; and (2) flood defence (36%) via the construction and maintenance of water control structures. Of the programmes identified, most were focused on flood defence (58%), mitigation (42%) or both (31%). Guidance and information instruments tended to focus primarily on flood prevention (26%) and flood preparation and response (46%).

FRM roles and responsibilities

Many stakeholders are involved in FRM, including government departments and agencies at federal to municipal levels, as well as non-governmental organisations. For example, Public Safety Canada and Infrastructure Canada are federal ministries responsible for administering various flood funding programmes; Environment and Climate Change Canada and Natural Resources Canada are federal ministries engaged in hydrological and meteorological data collection and the development of FRM guidance; and the Department of National Defence is a federal ministry that provides emergency response measures to flood-affected regions. Provincially, multiple departments and agencies are involved in land-use planning, emergency preparedness and response, water management, infrastructure design, and public health and safety in the prairie provinces of Canada. At the municipal level, depending on the size of the municipality, governments are engaged in setting land-use bylaws, and flood emergency preparedness and response. External to government, organisations such as watershed agencies, conservation groups and land developers also play a role in flood management and protection measures at local and regional scales, although their roles are less well defined.

Thirty different FRM roles and responsibilities were identified by study participants, and assigned to one or more of three levels of government, as well as to organisations external to government (Table 3). Fifteen of these roles and responsibilities were identified by more than 25% of participants. The most frequently identified roles and responsibilities broadly reflected matters relating to FRM implementation, funding and information management. The most prominent role, identified by 68% of participants, was the development, implementation and enforcement of land-use planning – said to be a municipal responsibility. The second most frequently identified role was the provision of funding for flood mitigation, response and recovery, identified by 62% of participants – consistently identified as a federal responsibility. The implementation of FRM initiatives (e.g. flood mitigation projects; construction and maintenance of water control structures) was the third most frequently identified, by 56% of participants, and believed to be a shared responsibility of provincial and municipal governments. Interestingly, participants from municipal governments did not assign this responsibility to themselves, indicating that they may not see themselves as on the front lines of policy implementation.

Table 3. Perceived ownership of flood risk management (FRM) roles and responsibilities, reported by number of participants assigning the role or responsibility.

		Number of participants assigning the role to federal, provincial, municipal, or other stakeholder groups (total n = 34)^a
FRM roles and responsibilities identified by study participants	Total	Federal	Provincial	Municipal	Other
Developing, implementing and enforcing land use planning	23	0	6	17	0
Provision of funding for FRM mitigation, response and recovery	21	17	4	0	0
Implementation of FRM policies and actions	19	0	9	10	0
Assess flood hazards and risks, carry out flood mapping	17	5	9	3	0
Public and stakeholder education on flood risk	15	2	6	1	6^b
Support municipalities with technical expertise, flood fighting and other FRM resources	13	0	13	0	0
Coordinate policy and action across jurisdictional boundaries	12	9	2	0	1
Data collection and dissemination	12	7	5	0	0
Ensure emergency preparedness	12	0	1	11	0
Provide national guidance and standards	10	10	0	0	0
Research and development of practice, tools, alternatives or FRM-related science	10	7	1	1	1
Development of an FRM strategy	10	6	3	1	0
Flood forecasting	10	2	8	0	0
Oversight and enforcement of FRM policies and actions	9	5	4	0	0
Coordination of emergency response	9	0	9	0	0
Provide emergency support when required	5	5	0	0	0
Responsible for First Nations FRM	5	4	1	0	0
Administer federal funding programmes	5	0	5	0	0
Regulate drainage	5	0	2	3	0
Develop provincial guidance and standards	4	0	4	0	0
Coordination between levels of government, ministries or departments	3	0	2	1	0
Responsibility for major rivers	2	1	1	0	0
Water resource management	2	0	2	0	0
Support government to reduce the financial and human burden	2	0	0	0	2
Translate science for practical use	2	0	0	0	2
Support provinces and municipalities through a central resource of expertise	1	1	0	0	0
Watershed management within province	1	0	1	0	0
Coordination of evacuation of First Nations	1	0	0	0	1
Responsible drainage	1	0	0	0	1
Determine socially acceptable levels of risk	1	0	0	1	0

^a Bold formatting indicates most frequently identified key roles and responsibilities.

^b Most commonly associated with non-governmental stakeholders.

The assessment of flood hazards and risk and flood mapping was identified by half of participants. Of those participants who assigned this responsibility, half assigned it as a provincial responsibility and the other half to other groups. For example, federal, provincial and other participants saw hazard and risk assessment and flood mapping as either a federal or provincial responsibility, whereas municipalities mostly considered it the responsibility of either the provincial or municipal governments. This variation indicates a lack of clarity about the responsibility for flood hazard and risk assessment and flood mapping, or a perceived overlap of responsibility.

Public and stakeholder education on flood risk, identified by 44% of participants, was not clearly associated with any single stakeholder group. Approximately 18% equally assigned this responsibility to each of the provincial government and other stakeholder groups, whereas federal participants did not identify this as a federal responsibility. These results suggest flood risk education is not a major thrust of their agencies’ FRM activities. Provincial government participants assigned it equally between the province and other stakeholders.

Responsibility for the development of FRM strategies also varied. This task was most often assigned to the federal government, but was thus assigned by provincial and municipal participants, not federal government participants. There were also some notable roles and responsibilities rarely assigned. Except for emergency response, roles that involve coordination of actors were not clearly assigned to any level of government. Only 9% identified coordination of strategies, goals, policies and actions between government levels, ministries and departments to be anyone’s responsibility. Coordination of data and information exchange between researchers and practitioners, such as the research needs of practitioners or data needs of researchers, was also unclear, with only 6% of participants assigning responsibility.

Perceptions of how FRM governance supports practice

Participants perceive there to be both strengths and challenges to FRM in the Canadian prairies. These are synthesised in Table 4, along with examples of some representative or common observations shared by study participants. Emergency response was perceived as working effectively and a major strength in current institutional arrangements. Emergency response plans are common in municipalities, and their development and implementation was reported to be well supported by the provincial governments. Roles and responsibilities are also well understood, with clear lines of communication in place regarding flood emergencies. The availability of physical resources, such as sandbags, pumps and earth moving equipment was commonly known, as well as procedures to mobilise these resources when needed. From the federal level, there was said to be secure funding from the DFAA, which does not appear to be limited during or after an event. According to interview participants, current institutional arrangements facilitate the effective implementation of emergency procedures through preparation, coordination of action and the guarantee of finance from DFAA.

Table 4. Perspectives on the strengths and challenges of current flood risk management (FRM) roles and responsibilities.

Key strengths	Supporting statements from interviewees
Ability to manage and recover from flood events: • Emergency response is effective ○ Response and recovery is well funded. ○ Emergency preparedness is improving.	• We’ve always concentrated on response, and we do that really well. We have a clear emergency management procedure…. It identifies lead roles and responsibilities of all partners within the province • It’s a well-rehearsed cycle of processes each year in flood season. It’s all coordinated through a lead organisation which brings all the stakeholders together • Generally, when you get into an emergency, you know spending control is going to be lifted…haven’t had the experience where we’ve been limited in what we can do to protect life and safety by lack of financial resources • We have increased the activities that we undertake from a preparedness standpoint. Municipal emergency response plans are now a legal requirement and are now much more common in municipalities than they were 20 years ago • Provinces are providing guidance on response plan development, ensuring that municipalities are aware of roles and responsibilities and resources that are available
Commitments to collaborative action: • Collaborations between river basin organisations, watershed groups and conservation districts • Cost sharing of FRM project funding • Data sharing	• The watershed group incorporates planning and advice on a watershed-wide scale. It is a beneficial collaboration of municipal and provincial governments, conservation groups and other stakeholders. We can pool our technical and financial resources to plan on a more regional scale • Municipalities get a lot of support from the province; they really try to help us and push forward our projects and collaborate on our funding proposals. The province has a grant programme to support community and individual flood protection. Then there’s the disaster financial assistance programme • Government organisations do try to work together…. We [watershed organisation] have a contract with all of our partners which determine how we’ll handle the sharing of information and data • A lot of the data we get is shared more informally through professional contacts. I know some people in the provincial government who might have data I need; they know I might have data they need
Key challenges	Examples of statements from interviewees
Conflicts between policy goals and policy implementation: • Financial resources are biased towards response and recovery • Conflicts of interest exist between prevention of flood risk and economic development • Conflicts exist between regulatory regimes • Funding can be hard to access	• We [municipalities] see a lot of support for emergency relief after the fact, when we haven’t been able to get our hands on funding related to the prevention of flooding…we saw hundreds of millions of dollars in this area for emergency flood relief, but we feel that if we had seen a fraction of that to go towards prevention, the recovery costs would have been much smaller • Municipal governments are dependent on property taxes and thus have an incentive to allow development in flood-prone areas where often they can collect higher taxes • We [municipalities] are trying to mitigate for increased drainage in the neighbouring province, but because we have very restrictive regulations here, it is very difficult. We are trapped between a lack of regulation and inflexible regulations • A $20 million CAD event is what is needed to trigger disaster financial assistance federally. In a rural municipality, it’s easy to have significant damage that is under $20 million CAD • It can be hard to convince governments to invest in mitigation to address flood risk when there are obvious immediate needs in health care, and education
Inconsistent focus on FRM: • Uncertainty of financial support for mitigation • Policy cycles make long-term FRM planning difficult • Policy windows influence resources available for FRM policy and action	• We [the province] need to say that there’s going to be annual funding available to help improve flood protection levels or help in understanding that flood risk. But the NDMP is a 4-year programme. We don’t know what will be there after that • The time horizon over which policy is expected to make a difference is extremely important and it needs to also have time to unfold. But we are not good at planning in cycles that are longer than election cycles • When resources are only available after a flood, it becomes hard. What we end up doing is band-aid fixing issues when a disaster occurs, because that’s the funding that’s available to us – rather than working towards mitigation efforts, which is what we should be doing
Availability and coordination of resources: • FRM knowledge and expertise are too widely distributed to be used in the most effective manner • Existing tools for FRM are inadequate • Lack of data to support decisions, assess risks and enforce legislation • Data is not coordinated	• Human resources are spread over three provinces, three levels of government and many departments and organisations • Municipal authorities often don’t have a technical background and cutbacks in provincial and federal departments mean that the support for effective FRM isn’t there • Those of us working in FRM are split up and under-resourced and have a great deal of difficulty doing a good job, despite our best efforts • We have some unique geography. And while there are tools that are available out there, that are being exploited in other countries, we’re not sure how well we could use them or what their effectiveness would be in Canada • We need data on our hazards, the hydrology, meteorology, topography. We need the data of where are our elements at risk, people, property and infrastructure. But we don’t really have this • During a flood, we needed cadastral data and GIS datasets that were held by another department. Although they were willing to share it, they had to sanitise some of the data, such as people’s names and personal information, which took too long to organise • There is no leadership or coordination of data gathering. As a result, we end up with many groups saying, ‘Well, I’m going to have to start collecting my own data’, and then these data failed to get merged

NDMP = National Disaster Mitigation Program.

There is also a strong commitment to collaborative action in addressing flood risk, with many examples reported – the most common concerning financial arrangements. A strength identified by many participants was the commitment to cost sharing of FRM projects in recognition that municipalities often do not have the financial capacity to fully fund local flood risk reduction projects, while provincial governments lack the capacity to fully fund large-scale projects. River basin organisations, watershed groups and conservation districts were said to facilitate effective stakeholder engagement, thus allowing coordinated action and more efficient use of resources across municipalities. The willingness to share data and other flood information, internally in government, between governments, or between organisations, was also identified as a strength of the current governance structure, including data for provincial flood forecasting, meteorological information, and federal and provincial hydrological monitoring networks. Data sharing between organisations was most commonly identified as an informal process through professional contacts, but some participants know of formal data sharing agreements.

Participants also identified some enduring challenges, including potential FRM policy conflicts. The federal government’s implementation of its major role, namely funding FRM mitigation and response and recovery, was said to be heavily biased towards response and recovery – both in terms of the amounts of money available, and the time scales over which it is secured. For example, the estimated federal expenditure through DFAA is $900 million CAD per year; it does not have any defined upper limit and it appears to be an ongoing programme with no end or review date. In contrast, the National Disaster Management Program (NDMP) is comprised of $40 million CAD investment per year, to be reviewed after 5 years. This distribution of funding was found to conflict with participants’ desire for a more balanced approach, a conflict also supported by the review of policy instruments (Table 2).

A second policy conflict concerned municipal governments. The major FRM responsibility attributed to municipalities was developing and enforcing land-use planning to reduce flood risk (Table 3); however, this was perceived as conflicting with municipal economic development priorities. This conflict is exacerbated by a lack of oversight compelling municipalities to enforce land-use planning practices. In general, provincial governments have no jurisdiction over municipal land-use planning, or authority to enforce municipal administrations. The result is that provincial legislation supporting flood prevention has reduced influence on municipal FRM, as it competes with local economic needs. Measures have been recently introduced to strengthen land-use planning. For example, the governments of both Alberta and Manitoba have passed legislation that gives them authority over land-use planning in designated high flood risk zones. The effectiveness of this change might not be seen for several years.

A third policy conflict was the differing regulatory regimes across provinces, reducing local ability to reduce flood risk. A municipal participant living close to a provincial border gave a practical example of this. It was perceived that run-off from the upstream province from agricultural drainage is increasing municipal flood risk. However, the municipality and the province in which it is located have no influence over drainage in the neighbouring province. With the near-border location, there are few options available to mitigate flood risk. One option suggested by a study participant is to ensure that the municipality has an adequate drainage structure capable of providing protection. In practice, however, interview responses indicate that provincial conservation laws make this difficult, as they aim to preserve wetlands by discouraging drainage and do not consider options tailored to local conditions that may help reduce flood risk.

Participants also expressed concern that support for FRM is irregular, with more political resources available after major events or with certain governments. The participants described policy windows – after a flood event, there is greater awareness of flood risk and resources are more readily available. However, participants expressed frustration over the lack of consistency in support prior to flood events, when resources would be most valuable for FRM planning and prevention. Time-bounded funding programmes, such as the 5-year NDMP, were also said to cause variations in the financial resources available over more extended time periods. When these policies reach the end of their lifespan, they may be continued, reformed or scrapped, which hinders the ability for municipalities to plan for longer term FRM. In provincial governments, FRM was said to be in constant competition for funding with other policy areas. Participants perceive the outcome of this stochastic attention to flooding is uncertainty over political and financial resources and a reduced ability to commit to longer-term strategies.

A lack of strategic coordination of human, physical and technical resources emerged as a final theme, contributing to FRM capacity constraints. Concern was expressed about a lack of national FRM coordination, causing inefficiencies through duplication of efforts, poor integration of policies inter-provincially, and gaps in roles and responsibilities. The federal government has no legislative mandate for FRM strategy or policy. Results from interviews suggest that the governance environment provides no opportunity for strategic coordination of provincial FRM strategies and policies, creating variations in FRM across provinces and municipalities, thus resulting in duplication of effort and a lack of standardisation of practice.

Interviewees alluded to the impact of this lack of coordination most clearly in two ways. First, data availability, acquisition and management were reported as a big challenge due to the lack of agreed-upon data needs and standard procedures or protocols for collecting, interpreting and disseminating data – e.g. data related to stream flow, topography, meteorology, snowmelt, soil moisture, and housing and infrastructure location and value. The result, according to participants, is reduced effectiveness of data sharing and reduced ability to understand flood risk regionally, provincially and nationally. Second, the human capacity for FRM was said to be impacted by a lack of strategic coordination. Although knowledge and experience and numbers of staff are considered strong in certain aspects of FRM (e.g. emergency response, flood forecasting), the distribution of this expertise is patchy and uneven across provinces and government levels, ministries and departments. Expertise is thus effectively siloed. The perceived outcomes are that there are difficulties in developing more diverse and integrated FRM options, difficulties in accurately assessing the effectiveness of FRM initiatives, and a lack of capacity to develop more effective tools.

Discussion

Results of the assessment of FRM policy instruments suggest that, in principle, FRM across the prairie provinces embraces a diversity of approaches – prevention, defence, mitigation, preparation, and response and recovery (Gilissen et al. 2016) – reflecting the basic requirements for flood resilience (Rouillard et al. 2015; van Herk et al. 2015). However, results also revealed considerable challenges to implementing this diverse policy agenda, including lack of clarity of roles and responsibilities, policy conflicts, limited capacity and data availability – which collectively biases FRM towards reactive solutions. Given international FRM experience, this is not unexpected. Challenges with conflicting and fragmented policy, policy instruments that do not adequately support FRM goals, limited capacity, and no overall coordinating FRM strategy are problematic elsewhere, including Sweden, the Netherlands, Belgium, France and the United Kingdom, to name a few (Rouillard et al. 2015; Ek et al. 2016; Hegger et al. 2016).

Government institutions and processes and how they are coordinated have considerable influence on flood resilience (Clarvis et al. 2014). The current FRM environment in the prairie provinces is characteristic of fragmented governance – many organisations with differing priorities, where no one regime is dominant, but all are involved in managing the same resource; and policy layering – where policy goals and instruments are added to existing ones, incrementally, often in an ad hoc way over time, and in an environment characterised by multiple levels of authority and multiple actors (Rayner and Howlett 2009). Provinces are the primary authority for FRM in Canada, but in practice the policies and responsibilities are distributed across complex, multi-departmental, multi-scalar systems of government including federal programmes and agencies, provinces, and more than 1,000 municipal governments. The environment produced is one typified by mismatched priorities, goals, capacities and resources across levels of government, jurisdictions and departments (Bakker and Cook 2011; Cook 2014; Gober and Wheater 2014), and higher FRM financial costs (OPBO 2016).

For example, federal FRM funding policies do not always match the FRM priorities of provincial and municipal governments, skewing FRM actions towards response and recovery objectives. There are three major federal funding streams in Canada which, in principle, support FRM: the DFAA, aimed at public safety through response and recovery; the Build Canada Fund (BCF), which finances infrastructure projects aimed at the public good; and the NDMP, which supports disaster prevention and mitigation. Although the BCF could be used to fund flood prevention and mitigation projects, to the knowledge of the interviewees it has not yet been used to do so. The typical way federal funds support FRM is as an insurer in the event of flood damage, financing the restoration of local infrastructure to pre-flood conditions using DFAA resources. There is a perception that DFAA is the main source of federal funding for FRM, even though it is not designed to be applied proactively. The more recent introduction of the NDMP, focused on prevention and mitigation, may further discourage the use of BCF for flood mitigation and defence; however, NDMP functions more in an investigatory manner, trying to understand what is required to support effective prevention and mitigation. Thus it is not, in its current form, an appropriate source of support for major proactive flood mitigation projects. Mismatched FRM policies and priorities are exacerbated due to limited coordinating influences such as strategies, frameworks or organisations established to deal holistically with FRM; different aspects of FRM are often subsets of other policy areas, such as infrastructure design or land-use planning. The result is constraints on the implementation of FRM policies or initiatives that go beyond defence structures or damage recovery.

Evidence from this study suggests that the challenges to FRM may not be insurmountable, and there are lessons to be learned from certain FRM strategies that can be transferred to other aspects. For example, emergency management was almost unanimously considered by interviewees to function effectively. Emergency management in Canada is comprised of a coordinated strategy that spans all three levels of government (federal, provincial, municipal) and, in the event of an emergency, defines clear roles and responsibilities, establishes the management procedures and processes, and provides the necessary resources. This is representative of an incident command system, or a hierarchical command structure, and pre-planned framework, that is put in action during an acute emergency event (Moynihan 2009) to ensure coordination and cooperation of the many actors involved in emergency response. This approach has proven useful in other countries beyond emergency response. For example, van Herk et al. (2015) explains the transition of Dutch FRM from a focus on structural defence to ‘room for the river’ as being attributed in large part to a dedicated strategy with clearly distributed roles and decision-making power.

A national strategy and distributed, but better coordinated, FRM oversight is needed for the prairie provinces and, arguably, across the country. This may require the establishment of a national agency (Pomeroy et al. 2016), providing a governance architecture and serving as a boundary organisation (Meuleman 2014; Prager 2015) to organise and mediate between different stakeholders; facilitate interactions; and organise policy instruments, procedures, and roles and responsibilities to manage for resilience. The authors are cognisant that the solution to fragmented water governance is not simply to create more centralised, hierarchical systems (Lemos and Agrawal 2007; Gober and Wheater 2014; Hegger et al. 2016). At the same time, the authors acknowledge that localised or fragmented organisations are often ill equipped to deal with system-wide and transboundary problems, or lack the capacity, strategy and authority to deal with uncertain, complex or surprise events (Craig 2008; Gober and Wheater 2014). This paper thus argues for a hub-and-spoke model for FRM, where a national FRM strategy, setting out broad principles and overarching policy direction, is rolled out through more regionalised centres of action across the country.

In the prairie provinces, a foundation already exists for this type of architecture, with the Prairie Provinces Water Board (PPWB). The PPWB was established in 1948, through a joint agreement among Alberta, Saskatchewan, Manitoba and Canada to recommend the best use of interprovincial waters and water allocations between the provinces. Its primary mandate is to ensure that transboundary waters are equitably apportioned and protected in accordance with a master agreement, but it also provides an important forum for information exchange and promoting cooperation in water management (PPWB 2012). Board members are senior officials engaged in water resources administration in each of the provinces, with additional members from the federal government – a co-location arrangement that provides the opportunity for mutual learning (Feldman and Ingram 2009). Of course, adopting an FRM role would require an expansion of the PPWB’s existing mandate and the agency and power to function as its own organisation – implementing a national FRM strategy, yet ensuring regional interests are coordinated, policy layering reduced, and thus local interests protected.

Conclusion

A recent analysis of trends in Canada’s climate indicates significant increases in winter and spring temperatures, increased April streamflow, and an earlier start of the spring high-season flow (Vincent et al. 2015). Coupled with changing land use, increasing urbanisation of watersheds, and floodplain development, the frequency and severity of major flood events is likely to increase. Adapting to floods, through resilience-based FRM policies and initiatives, rather than relying solely on flood defence structures or disaster recovery programmes, is the most appropriate, long-term solution. Results of this research show that FRM policy instruments in the prairie provinces are diverse, and include prevention, defence, mitigation, preparation, and response and recovery. However, flood resilience is inherently challenged by institutional fragmentation, lack of clarity of FRM roles and responsibilities, policy layering and competing mandates. As a result, FRM in the prairie provinces tends to default towards flood resistance and/or recovery, rather than flood resilience. What is needed in the prairie region, and arguably nationally, is an overarching regional or national FRM strategy and boundary organisation to coordinate roles and responsibilities for the specific purpose of flood resilience – rather than solely coordinating flood emergency response and recovery. This requires an agency with the mandate to manage FRM policy instruments, and clearly allocate decision authority amongst the multiple levels and layers of FRM governance.

Funding

This research was funded by the Social Sciences and Humanities Research Council of Canada [grant number 430-2014-01116].

Acknowledgements

We wish to acknowledge Andrea Carroll and Claire Crowley for assistance with data collection, and the contributions of interview participants.