Flood history and community engagement in flood risk
management in the UK: a clash of paradigms
Kseniia Puzyreva (Id #: 11127635)
kseniiapuzyreva@gmail.com
Thesis supervisor: Dr. D.H. (Danny) de Vries
Fieldwork supervisor: Dr. Sue Tapsell
Second reader: Dr. R.J. (Rob) van Ginkel
Word count: 13734
Research Masters Social Sciences
Graduate School of Social Sciences
University of Amsterdam
Amsterdam
2
Abstract
Following the general trend in approaches to hazard mitigation, the English system of flood risk management has recently experienced a social turn involving the participation of multiple stakeholders, including local communities, in the different stages of flood management. Achieving local engagement is however challenging. By looking at the example of community flood action groups (CFAGs) in the South-East of England, this paper aims to understand how the flood history of an area influences the emergence and sustainability of community engagement in flood risk management in the UK. The data used in this study was collected during a four-month period of fieldwork conducted in four communities within the River Thames Catchment. This data was obtained through semi-structured interviews, field observations, and archival research, and analysed within the framework of a constructivist strain of grounded theory. The findings show that flood histories are situated in social time and may have different readings across and within the communities. The local flood histories inform the residents’ expectations about future flood hazards. In situations where the established expectations fail to coincide with the new realities of a particular hazardscape, a condition of surprise is likely to arise among the inhabitants that creates the context for the emergence of the CFAGs. The sustainability of these groups, however, largely depends on the extent to which they will be accepted by and involved in existing networks of flood risk management. This research sheds light on the issues associated with the social turn in flood management and offers insights that may help to ensure a more inclusive system of flood risk management.
Keywords
3
1. Introduction
1.1. Community flood action groups in the Thames Catchment
What is a flood? As a person who has never experienced a flood, I used to think of floods as any instance of water in a ‘wrong’ or unusual place triggered by certain unstable weather patterns (extensive rainfall, ice/snow melting, the increase in the river/sea levels, etc.). While talking with the representatives of four community flood groups that operate within four different flood-prone areas in the southeast UK, I started to grasp that the understanding of floods and flood history that belongs to those who are used to living with water that can be very different. The local flood histories I encountered in these areas were not historically comprehensive: certain floods were ‘glorified’, while others were overlooked. Of course, the exclusion of some of the floods that took place in the 17th and 18th centuries might be explained by the
faults of memory or the lack of historical sources passing this information down to today’s residents. However, how can we explain the overlooking of a flood that took place just one year ago, or the overlooking of a flood that is ‘officially’ considered to be the most recent? What are the logics that stand behind multiple readings of the ‘same’ flood history?
The contemporary system of flood hazard management in the UK has recently experienced what policy documents label as a ‘social turn’ (Nye et al., 2011), following an overarching shift from flood protection to flood risk management (Hartmann & Spit, 2016). A move towards societal Flood risk management (henceforth FRM) implies the devolution of responsibilities to actors previously uninvolved in expert decision-making, an approach that advocates for public involvement. In England, participation in FRM is part of a broader political rhetoric (Cameron, 2010), which was conveyed into regulation through the Localism Act (2011). Even before the introduction of the European Floods Directive in 2007, England had increasingly emphasized the government’s limited capacity to fully protect its citizens from flood damage and stressed the need for the devolution of responsibilities and public involvement. However, despite the strong calls for public participation and inclusive decision-making, research reports limited the involvement and influence of the citizens and communities on hazard-related decision-making (Begg et al., 2017; Hartmann & Spit, 2016; Wehn et al., 2015). Wehn and colleagues (2015) argue that the public has a generally limited impact on and involvement in decision-making, although there are certain differences in the roles citizens play during the phases of the hazard management cycle. In the example of Doncaster, in Yorkshire, UK, the researchers show how the role of the public transformed from being mere recipients of information during the preparation and response phase to a position of consultants during the recovery and mitigation phase. It is however unclear how to interpret these results if the analytical boundaries between different phases of the hazard management cycle are omitted. Studying the involvement of local stakeholders in the decision-making related to flood-defence planning and implementation, Begg and colleagues (2017) and Kelman (2001) also report rather uneven signs of the social turn within English FRM. Finding that financial constraints hamper local stakeholders’ involvement in decision-making, this study supports the research by Thaler and Priest (2014), who argue that partnership funding is likely to increase the number of the stakeholders but also creates the context for the reproduction of environmental inequalities and an inequitable distribution of risk.
In an attempt to understand the conditions that may influence public involvement in FRM, flood management professionals and researchers often invoke the concept of flood history as a contextual factor that influences public engagement (Orr et al., 2015; Bradford et al., 2012; Burningham et al., 2008; Pagneux et al., 2011; Thaler & Levin-Keitel, 2016). The researchers claim that flood history and the presence of recent flood events serve as important triggers for the formation of community-level action and engagement in FRM practices. These claims, however, are bound by a rationalist understanding of flood history and do not take into account the fact that the readings of the ‘same’ flood history may drastically differ across and even within affected communities. Moreover, flood management literature
4 most often approaches flood history from the perspective of flood management professionals, rather than from the perspectives of the population at risk of flooding. Such an approach promotes desituated understanding of flood history and its influence on community engagement in flood management. This thesis counters the rationalist conceptualization of flood histories prevalent in hazard management literature, arguing that rather than having one possible objective and chronological sequence of events, flood histories should instead be regarded as cultural constructions (Climo & Cattell, 2002; Crumley, 2000; De Vries, 2011b).
The purpose of this research was to examine how flood history of an area may influence the emergence and sustainability of community engagement in flood risk management (FRM) in the UK looking at the example of four community flood action groups (CFAGs) – the groups of flood-prone residents aimed at managing flood risks in their local environments. During a four-month qualitative study of four areas in the Thames Catchment, UK I looked at how CFAGs bring back past ecological conditions, how multiple readings of flood history interact with each other within and outside flood-prone communities and argue that the extent to which flood management stakeholders accept various readings and consequent strategies of flood management that the CFAG pursues influences the group’s ability to sustain over time.
1.2. Dominant perspectives on the emergence and sustainability of flood action groups
Scholars address the emergence and sustainability of community engagement in FRM from two main perspectives: either through an institutional perspective stressing the importance of institutional support structures, or through a sociological approach to risk perception. The most vivid examples of the first approach are studies that emphasize the importance of political inclusiveness and that argue for new forms of knowledge and practices of hazard management and decision-making. Bäckstrand (2003), for example, argues that for community engagement to emerge and be sustainable, appeals for public involvement are insufficient if the very structure, logic, and legitimation practices of decision-making – mainly based on scientific knowledge – are not changed as well. A similar argument is made by Brown and Damery (2002), who argue that community engagement in hazard management is constrained by the still-dominant institutionalization of technical perspectives on flood management within regulatory bodies such as the Environment Agency - a public body established in 1996 under the U.K. Department for Environment, Food and Rural Affairs - that make local communities mere recipients of technical knowledge about flooding, leaving little space for their equal participation and contribution. Other barriers for community engagement in FRM include the complexity of institutional arrangements (Buchecker et al., 2013; Nikitina et al., 2011) and the confusion caused by the fragmented division of responsibilities, as well as public distrust towards government agencies (Harries & Penning-Rowsell, 2011). At the same time, the lack of clear boundaries for demarcating responsibilities for flood risk management may become the basis for the emergence of community engagement in the form of political pressure groups. These groups aim to make sure that the social contract between the people and the state is not breached (Geaves & Penning-Rowsell, 2015, 2016). The sustainability of these forms of community engagement is, however, questionable, since bridging the gap in the social contract leaves such groups without any other purpose for which to continue their activities. More sustainable and long-lasting engagement is more like to result from collaboration and communication with other stakeholders, through the support of external (Geaves & Penning-Rowsell, 2014; Shaw, 2006) and internal (Buckland & Rahman, 1999; Coates, 2010; Maskrey, 1989; Murphy, 2007) structures. Building on Miller’s (2007) assumption that pre-disaster social organization is the community’s primary resource in mitigating threats, Coates (2010) argues that a community’s ability to respond collectively to a hazard depends on the local social structures and the networks existing within and outside that community. The type of local structures (i.e. absent, causal, organized, or formal) in turn largely influences the potential of community engagement in becoming systematic and sustainable as opposed to the immediate provision of help. Shaw (2006) furthermore stresses the importance of external support structures, claiming that for continuous
5
community involvement, a conducive policy environment requires collaboration with the local governmental institutions and the incorporation of community-level activities in policy matters.
The second approach is a sociological approach to risk perception, which derives from studies that focus on how the experience of a hazard can influence community-level involvement in hazard mitigation. This approach is largely informed by a rationalist risk-perception paradigm, which presumes that the experience of a hazard can prompt an individual or a group to make a rational assessment of a risk and then undertake actions in response. Experience with hazards is often assumed to facilitate the recognition of risk and consequently to increase the willingness to take mitigating action. These studies often employ concepts such as hazard experience, the knowledge of a hazard, and hazard memory or history as a backdrop against which to examine other factors influencing the emergence and sustainability of the local initiatives. Contradicting the social-ecology assumption that community action is likely to emerge only after a major flood has disrupted daily life (Lichterman, 2009), Geaves and Penning-Rowsell (2014) indicate that experience of any type of flood hazard, regardless of scale, leads to some form of public engagement. The scale of a flood, however, may influence the form that civic engagement might take, with smaller floods being likely to trigger the formation of local initiatives through the activation of already-existing networks, while big floods are likely to trigger the formation of new flood action groups. At the same time, the influence of flood experience on people’s preparedness and willingness to take action can hardly be regarded as straightforward, and the so-called risk-perception paradox is likely to occur (Bubeck et al., 2012; Wachinger et al., 2013). This line of research suggests that the temporal characteristics (i.e., frequency and timing) of flood hazards and the histories they produce may strongly affect community engagement. The frequent occurrence of floods is often regarded as a reminder of existing risks and tends to increase public recognition that mitigative action is required (Thaler & Levin-Keitel, 2016; Viglione et al., 2014). Brody and colleagues (2010) suggest that when flooding occurs on a repetitive basis and produces a long history of flood events and associated losses, the need for non-structural, soft FRM approaches becomes more obvious and the context for public involvement emerges. Comparing three sites in the UK, Thaler and Levin-Keitel (2016) conclude that a lack of recent flood events in the community served as one of the main obstacles for the emergence of community-level engagement in flood management. Thus, while recent or frequent experience of disaster tends to increase knowledge and sensitivity levels (Alexander, 2000; Parker & Harding, 1979) the absence of a ‘fresh’ flood experience is considered to lessen awareness of risks and to have detrimental consequences for the emergence and sustainability of local initiatives.
While scholars may employ the concepts of hazard history, collective hazard memory, and a shared hazard experience when conducting flood hazard research, these concepts rarely constitute the focus of the study (De Vries, 2008; McEwen et al., 2012). Serving as a backdrop, the exact influence of history and memory, or temporality in general, is often left unaddressed. This results in the absence or poor definition of these concepts and an inability to include these factors in more comprehensive explanatory models. Additionally, when investigating the influence of flood history as a scene within which other factors are studied, researchers tend to implement an oversimplified understanding of flood history and its influence on community engagement in flood management. The majority of the studies approach the concept of hazard history as objective, chronological and linear, embedded in mathematical time, external and independent of social life. The studies tend to undervalue the influence the relativity of flood history (its situatedness in time and human ecological experience) may have on people’s understanding of flood hazards and actions required for their mitigation. As tempting as it may be to demonstrate causation between a local flood and the actions of flood-affected residents, the relationship between these variables appears to be more complex when looking at flood history as a social construct that may have multiple interpretations.
6 To complement a rationalist understanding of flood history, I propose a third perspective, which follows the viewpoint of historical ecology looking at the interaction between humans and environments where they reside going back in the history (Balée, 2002; Crumley, 2000; De Vries, 2008; Swetnam et al., 1999). Examining how flood history influences community engagement in FRM I investigate how through their memories, experiences and practices flood groups situate floods, reproduce flood histories and chose particular strategies of action and also look at how particular flood events revise human activity, creating the context where the emergence of the CFAGs becomes possible. To do this I combine the discussion of history-memory relations (Archibald, 1999; Climo & Cattell, 2002; Halbwachs & Coser, 1992; Kammen, 2011; Nora, 1996) with the notions of hazard temporality and surprise (De Vries, 2008; Lupton, 1999).
With respect to the history-memory debate, my study draws on the constructivist understanding of history (Climo & Cattell, 2002; Halbwachs & Coser, 1992) and contributes to the two existing strains of research. The first strain considers history as a static, positivistic, and scientific protocol aimed at objective, evidence-based, true representation of the past (Nora, 1996; Yerushalmi, 2011). The second, more recent, strain regards history not as a quest for truth, but as a quest for meaning (Archibald, 1999; Climo & Cattell, 2002; De Vries, 2011; Kammen, 2011). The combination of two perspectives, allows to assume that flood history is relative and the way it is reproduced is context and purpose dependent (Olick & Robbins, 1998). Flood history may take a form of a positivistic scientific protocol (Halbwachs & Coser, 1992; Nora, 1989) if the framework where it is reproduced strives for scientific objectivity as pertains to knowledge and trustworthy evidence-based data or may become faculty for defining the meaning as opposed to a faculty of defining the truth (Archibald, 1999; Kammen, 2011). The proposition that the distance between the meaning and the truth is not that broad and in certain context meaning should be given priority over objectivity (Archibald, 1999; Climo & Cattell, 2002; Kaufman, 1994) is particularly overlooked in hazard management literature but deserves significant attention. Similar to therapeutic settings, where memories are elicited not to develop a true story but to use reminiscing as a healing process (Hendricks, 1995; Kenyon et al., 2001), it appears crucial for the field of flood risk management to pay attention to the subjective meaning-based reconstructions of a flood history. What is meaningful for communities at risk may have prevalence over the ‘objective’ representation of the past, as people’s understandings of flood history is made on the basis of culturally transmitted knowledge and experience, rather than scientific projections striving for factual representations of past ecological realities (Crumley, 1994).
Considering flood history as a social construct – as an activity of deriving meaningful patterns from particular meaningful pasts (De Vries, 2008) – requires looking at history as being situated in a social time (Sorokin & Merton, 1937), an idea usually ignored in the flood management literature (De Vries, 2008, 2011). According to Sorokin and Merton (1937), social time is qualitative and heterogeneous, as opposed to objective, quantitative, and chronological, deriving its qualities from meanings common to particular social groups. The quality, meaning, or social significance of certain events make them the referents on the social time scale. Being produced within particular socio-environmental settings (De Vries, 2008), referents that compose local flood histories orient people’s expectations about future flood hazards and coordinate their flood-related activities. However, when perceived reality departs qualitatively from expectations the condition of collective surprise is likely to emerge (De Vries, 2011; Holling, 1986). As De Vries (2008) points out, being a result of interaction between the uncertain hazards and environment-related human activities floodplains are places where surprise is thought to be a defining element. The condition of surprise that emerges out of human-environment interactions in particular geographical location forms the context where human activities can be revised and new organizational forms, such as CFAGs can emerge. This required looking at the characteristics of the ecological situation that preceded the surprising flood (Balée, 2002; Crumley, 1994; De Vries, 2011).
7 This thesis applies a cultural-constructivist approach (Lupton, 1999) to the understanding of flood history and its influence on the sustainability of community engagement in flood risk management. My research shows the relativity of flood history and the possibility of multiple readings, pays attention to the processes and logics that stand behind the reproduction of different local flood histories, and takes into account the influences of hazard temporality (De Vries, 2011). This study demonstrates not only how flood histories can obtain different readings depending on who produces them and under what conditions, but moreover shows how these different readings may influence the sustainability of community engagement in FRM. I will use one of the four case studies, the village of Datchet, to illustrate this point and shed light on the issues associated with community-engagement and the ‘social turn’ in English flood risk management.
2. Methodology
2.1. Research setting and methods
As historical ecology suggests, the human-environment relations in particular geographical location may be investigated using archival materials, archaeology, oral tradition and history, and proxy measures drawn from the Earth Sciences (Crumley, 2007: 5). In this research, I used qualitative methodology combining semi-structured interviews with field observations and archival research.
The interviews were conducted with the following groups of informants: members of the CFAGs, representatives of the Environment Agency of the UK responsible for the areas under study, and local community volunteers who were neither affiliated with the local CFAGs nor with the flood authorities. Sampling was organized into two stages. During the first stage, the sampling of the respondents from the CFAGs was done through the official websites of the local authorities (parish councils and county councils), and through the CFAGs’websites webpages on Facebook. The sampling of representatives from the Environment Agency was also organized through the official website of the organization. The second stage of sampling was done by means of the ‘snowball’ technique, in which I asked my informants to suggest people who might in their view be helpful in discussing the topics addressed in the research. In total, I conducted 21 interviews, each approximately 1.5 hours long, which were coded and analysed. To capture the informant’s understandings and meanings, the interview questions were designed to be as concept-neutral and open-ended as possible. Interviews were organized around three main topics: how CFAGs bring back past ecological events, emergence and development of the CFAG, and the CFAG’s embeddedness in cooperative networks with other stakeholders involved in FRM within and outside the community.
I also looked at how flood memories and experiences become ingrained in the local landscape and material practices of the CFAGs and the Environment Agency (as an institution responsible for collecting, storing and transmitting information about the ecological past). The core observation sites were the “offices” of the CFAGs (i.e. the Parish Office, the CFAG chairman’s house); the local libraries where the meetings of the CFAGs, the Environment Agency and the wider community were organized; the Parish Council meetings; and locations in the community that had any artifacts translating flood experience of a community.
Working with archival data, I sought out pertinent documents, such as copies of the rural district council meeting minutes, historical books, Environment Agency flood reports in the form of spreadsheets, and visual data such as locally produced flood maps, photographs and the Environment Agency Historic Flood Outlines in the form of GIS-layered maps. The archival data was obtained from local libraries or the borough archives as well as through personal connections with local historians. The Environment Agency Flood Outlines dataset was obtained electronically from a UK governmental website providing
8 open-access data published by governmental departments, public bodies, and local authorities (Environment Agency, 2015).
Data was gathered during a four-month period of fieldwork, which focused on stakeholders within densely-populated flood-prone areas of the Thames Catchment, at the following sites: Datchet (Royal Borrow of Windsor and Maidenhead), Teddington (Richmond Borough), Purley (Croydon Borough), and Manor Park (Slough Borough). The criteria for choosing these sites were 1) the presence of local CFAGs, 2) historical exposure to flood risks, and 3) the absence of the relatively recent research on flood perceptions conducted in the area. I wished to avoid having informants abused by excessive research and to decrease the possible difficulties associated with entering in the field.
2.2. Data analysis
This research methodology focused on how local flood histories were produced within communities and to what extent this affected the emergence and sustainability of CFAGs. First, I looked at which events were included in the local readings of flood history and why. Second, I looked at the extent to which different readings of the flood history were accepted by flood management stakeholders within and outside the community. Conducting interviews and field observations and engaging in archival data analysis resulted in a collection of textual (transcripts of interviews, flood reports, historical accounts of flooding instances) and visual (flood maps, photos of flood artefacts, photos of CFAGs meetings) data across all four fieldwork sites. The data was then entered into the qualitative data analysis software ATLAS.ti, which helps to organize various types of data, lowers the chance of random selectivity, and helps with tracking the emergence and development of the research concepts and arguments.
The constructivist strain of grounded theory was used to approach data analysis (Charmaz, 2006). This choice was informed by two reasons. The first reason was theoretical scarcity as the field is dominated by perspectives that do not take into account the relativity of flood history and do not allow for an understanding of how flood history influences community engagement in FRM. Grounded theory allows for core issues and concepts to emerge naturally from the data instead of forcing preconceived theoretical categories and concepts (Glaser, 1978). In the case of theoretical scarcity, this seems crucial since the implementation of already-existing concepts and assumptions could result in a theory that reproduces blind spots and bears limited explanatory power. Implementing grounded theory, in turn, allows for the pursuit of ‘unanticipated directions of inquiry’ and thus generates new theoretical reasoning (Charmaz, 2006, p. 155). Second, grounded theory provides a set of research strategies for studying the experience and understanding of floods and flood history, as it allows for the deconstruction of definitions and concepts that are often taken for granted. Flood plain residents’ constructions of what constitutes a local flood history are ‘neither convenient fabrications nor idiosyncratic inventions’ (Charmaz, 1990: 1161). Rather these constructions reflect people’s understandings of their experiences, which often contradict professional conceptions.
2.3. Coding scheme
The analysis of the data was organized into several phases: coding, analytic writing accompanied with extensive memo writing, and theoretical sampling. The coding scheme was developed in two steps. For the first step, I began with the initial coding as, according to Charmaz, it allows for openness to ‘any theoretical directions proposed by my data’ (2006: 46) and allows the core themes to emerge naturally. I implemented a mixture of descriptive and process coding (Saldaña, 2015) and coded the interviews paragraph by paragraph. The analytic level of codes ranged from descriptive codes (e.g., ‘reporting unusual causes of the 2014 flood’) to potential analytic categories (‘surprise’). I also used in vivo coding (i.e., ‘the great flood’) as it allows for the observation of the respondent’s meanings and understandings of certain phenomena without pushing them into the conceptual frameworks of the researcher. After the initial cycle of coding and determining which codes were most frequent and significant, I engaged in a second step: focused coding. This stage of analysis allows for assessing which codes have the best
9 interpretive and explanatory capacities. It presupposes the constant comparison of codes to codes, codes to categories, categories to categories, and categories back to empirical data. As the result of this process of reorganization and categorization, some codes merged and the higher-level analytic categories were created. The core categories were the following: hydrological reading of the flood history, community-preparedness reading, clash between readings, flood temporality, surprise, and CFAG cooperation with other stakeholders. The process of coding was accompanied by extensive memo writing, as it provided the basis for exploring and developing not only descriptions but also analytic ideas, which formed the basis for further theoretical reasoning.
2.4. Reflexivity
Working in the field, I distinguished three dimensions of my self. Entering the field, I introduced myself as a researcher with certain experience in conducting research on community engagement in flood management. Doing this, I showed my familiarity with the field, the topics I was about to discuss, and my ability to speak the same language as my informants. This allowed me to conduct conversations in a natural way, without turning each respondent’s narrative into a lecture. My second dimension of self was as a researcher from the Netherlands. The informants immediately connected their understandings of the Netherlands as a state-of-the-art country in FRM and saw me as a potential conveyor of knowledge. My informants often asked me to bring examples of how similar issues are regulated in the Dutch context, which I was luckily prepared to do. At certain moments, being a researcher from the Netherlands allowed me to be somewhat of an expert, able to share my knowledge and give advice when needed. This significantly assisted me in establishing rapport with informants, as giving advice and sharing knowledge made discussions a reciprocal activity. It also allowed me avoid being an ‘intellectual tourist’ who grabbed information and left. My third dimension of self was as a person who has never been affected by floods. The lack of this personal experience played one of the most important roles in the collection and analysis of data as it allowed me to stay open to all possible conceptions of floods and flood histories. Being open to multiple interpretations paved the way for a rigorous quest for meaning that formed the basis of this study.
2.5. Informed consent
The ethics of this research proposal were reviewed and approved by the supervisor and director of University of Amsterdam’s Research Master’s Social Science program. Due to the character and purpose of the study, gaining signed informed consent was considered too formal and potentially intrusive to the fieldwork, where bridging the gap between the observer and the observed was highly desirable. This however, did not exempt me from ensuring that informants’ decisions to participate in the research were knowledgeable and voluntary (Thorne, 1980, p. 287). Thus, entering the field required me to clearly present myself as a researcher who was studying the emergence and sustainability of community engagement in flood management. This did not involve disclosing that the focus of the research was on the relation between flood histories and community engagement, as that might have biased the informants’ narratives. To ensure the voluntary aspect of participation, before the interview the informants were reminded about their ability to end the conversation if they felt uncomfortable with the topic or the character of a discussion. For analytical purposes the interviews were audiotaped and the interviewees were asked for permission to record the conversations and informed that they could stop the recording if they wanted. I also notified the informants that the recordings would only be used for research purposes and that, to insure the anonymity of the narrator, interviews’ transcripts would not contain any personal information. Moreover, since the research included field observations as a method of data collection, my participation in meetings (such as CFAG and local council meetings) included my being introduced as a researcher by the meetings’ moderators. The meetings were not audiotaped; instead they are reflected in my written and visual field notes. Photos of the meetings, if taken, do not show faces, to preserve the anonymity of the attendees.
10
Although with important nuances, data analysis of the four case studies allowed for the identification of three general observations. First, residents of flood-prone areas reproduced local flood histories as event-centred as opposed to narrating a chronological sequence of events. Local flood histories are devoid of floods that are unimportant to the residents, and are instead composed of flood events that are worth attention and action and serve as the baselines for the assessment of past and future flood hazards. These situated experience- and meaning-based readings strongly differ from the flood histories in the official flood reports that aim at factual representation of the past. Second, apart from the case of Manor Park, where the CFAG was initiated by a nongovernmental organization, the emergence and activation of flood groups can be associated with the surprise that arises when people’s experience-based expectations do not coincide with the realities of a new hazardscape. An experience or, as the case of Teddington suggests, being close to experience a surprising flood creates a context where the transformation of the local system becomes possible, i.e., where the local CFAGs can emerge. Third, the sustainability of community engagement in flood management strongly depended on the extent to which a flood group became accepted by and involved in existing FRM networks. Three of four groups studied lacked contacts with FRM stakeholders inside and outside the community and found it particularly difficult to continue functioning over time. The village of Datchet has been chosen as a case study to illustrate these findings in greater detail and explain the relationship between flood history and community engagement in FRM.
3.1. A protocolled-scientific reading of Datchet’s flood history
The official history of flooding in the village of Datchet can be accessed through the flood reports produced by the Environment Agency. Among other duties, the agency ‘is responsible for taking the strategic overview of the management of all sources of flooding and coastal erosion, which includes providing evidence and advice to inform Government policy and support others’ (Environment Agency, 2014). The agency provides three types of open-access datasets that can be used to obtain the history of flooding in Datchet: the Historic Flood Map, Flood Outlines, and Aerial Images. In order to acquire information on each flood registered in the territory of Datchet, I used the recorded Flood Outlines dataset (recorded instances of flooding in a specific area), which covered the territory of the south of England. The records go back to 1946 when predecessors of the Environment Agency started to collect detailed information about flood incidents. In contrast to other types of the maps (such as surface water maps, fluvial hazard maps, etc.) produced on the basis of predictive flood modelling, the Historic Flood Outlines are generated from maps, sketches, aerial photographs, and water height measurements taken or obtained either by the Environment Agency or by its predecessors at the time of the actual flood event. The Flood Outlines dataset is comprised of a GIS layer that shows all the agency records of historic flooding from rivers, the sea, groundwater, and surface water. The GIS files are also accompanied by spreadsheets that contain information on the start and end dates of flooding, and indicate the sources of flooding, the type of flood, and information on whether the flood is to be included in the Historical Flood Maps.
The GIS files allowed me to examine the flood events that took place in the territory of Datchet from 1947 until 2014. According to the dataset, during this period, the village was flooded seven times: in 1947, 1974, 1990, 2000, 2003, 2007, and 2014. The maps include outlines for each flood with a different colour, showing the areas within and outside the village that were covered by a particular flood, and allowing for the superimposition of layers to visualize the differences in the extents of the floods (Figure 1).
11 Figure 1. Historic Flood Outlines for Datchet. Period from 1947 to 2014. Source: Environment Agency
The Historic Flood Maps and the Flood Outlines constitute a formal, positivistic flood history, reflecting a ‘factual past’, a past established through hydrological evidence. The combination of GIS maps and spreadsheets illustrate a scientific, protocolized history, or a chronological time series of sporadic events embedded in visual and textual representation (Halbwachs & Coser, 1992). Such a representation of history corresponds with what I refer to as a ‘non-event-based’ production of flood history and the Environment Agency’s purposes for data collection emphasizing the chronological continuity of data. This scientific flood history is used in the validation of the predictive flood models to ensure the credibility of inputted data and is employed by various risk management authorities such as the Lead Local Flood Authorities, Local Planning Authorities, and the Local Resilience Forums. It is also used by commercial and non-commercial users for the risk assessment and development of planning applications, as well as by the staff of the Environment Agency itself.
Despite the fact that the Historic Flood Outlines serve as a reminder of the continuous flood risk, they do not provide any insights into the meaning that these floods had or still have for the areas affected. Moreover, it is not clear to what extent the production of the Flood Outlines include local knowledge, and to what extent the additions or corrections to existing flood outlines are allowed if new information becomes available. Their adherence to a chronological time scale allows for going forward, but does not presuppose going back and rethinking the past. As an important risk management tool, mapping – a chronological, sequential representation of historical data – produces an objectified understanding of risk, which turns risk into data. This leaves little room for the acknowledgement of the alternative understandings and readings of local flood histories and associated flood risks. As my research will show, the residents of the flood-prone communities rarely build local history on the time-series official historical data, but appeal more to personal experience, memories and meanings, which particular floods have for them and their local environments.
3.2. Surprise and the emergence of the Datchet hydrological CFAG
Despite the Environment Agency’s attempts to provide as detailed and as comprehensive a history of floods as possible, the making sense of the risks and the construction of local flood history by those who
12
used to live with water appeared to work differently. A foreword to the ‘Datchet Past’ illustrates the historical memory of the local residents I talked to. It reads: Described in 1724 as a low and watery place, water is a dominant theme in Datchet’s story, with disastrous floods, both within and beyond living memory, and with a large pond occupying the middle of the village before the familiar Greens existed (Kennish, 1999). This historian’s account of Datchet’s flood history goes back not to 1947, as the Historical Flood Outline provides as the baseline, but instead as far back as the 16th century, when the village became residential and the first flood alleviation barriers found their place in the village. Oral histories and local historical accounts suggest that minor floods were a common fact of life in the past, and until the late 1800s dwellings were not built in the flood-prone areas. In the 1800s, Datchet experienced a series of floods, in 1774, 1809, 1822, 1852, and 1894. Fifty-three years later, in 1947, the village was affected by a flood that continues to be one of the core baselines in local flood history as well as the baseline for the EA dataset. In the parish’s magazine of March 1947, the vicar of Datchet started his note referencing the Windsor Express: the greatest flood of the century and the worst catastrophe in living memory that engulfed the Thames valley over a weekend (quoted in Kennish, 2014).
The socio-ecological features of the 1947 flood qualitatively stood out from its predecessors. The evidence suggests that the water was rising at an extremely high speed and reached the top of the riverbank in Datchet with a further threat of breaking the bank. The flood management activities were described as a ‘war with floods’ (Eton Rural District Council, 1947). The floodwater inundated houses and businesses, some of which had to be relocated. The railway line from Staines to Windsor was disrupted and the platform in Datchet became a canal. Telephone lines were out of order and trees obstructed the roads, having been felled by 70-100 mile an hour wind gusts. Despite a huge collective effort at flood mitigation and to assist those in need, local organizations were not sufficient for handling the flood and the military became involved. The military ‘ducks’ (amphibious vehicles) were used to transport people, food, and other supplies.
Along with the socio-ecological features of 1947 flood, the temporal situatedness of the event significantly heightened its importance and meaning. Apart from following a 53-year-long lull, the event of 1947 co-occurred with the end of wartime, when the recovery from World War II had scarcely begun. In our conversation, the Datchet local historian recalled stories of those who witnessed the flood and described the context in the following way:
It flooded at a very bad time and that was a huge drama. It was so dramatic, it was such an event that everybody before and after the flood had heard and was quite keen to tell you about all the water coming down to Slough Road and to the village, and how long they could not get to school, and how people actually dealt about it. I haven’t talked to many people but I know the context well; it was just after the war…all supplies showed that there was rationing of food and that was the lowest point you could possibly be.
The difficult times of gradual post-war recovery were complicated even more by unstable and unexpected weather patterns that resulted in one of the coldest winters in the UK. The Meteorological Office accounts suggest that from ‘late January until mid-March easterly winds brought a series of snowstorms and a freeze to post-war Britain’ (Met Office, 2017). Like the majority of the areas in Britain at the time, the village of Datchet was covered with snow and ice. When the flood occurred, the rivers were bloated with the snow and the frozen soil was unable to absorb the water.
The 1947 flood effected a qualitative transformation of reality and solidly fixed new knowledge and experience within the landscape, which allowed the following generation to internalize this event as one of the baselines in the community’s flood history. According to the oral histories, the flood of 1947 introduced new arrangements into people’s daily lives and new rules in Datchet development and planning. For instance, all buildings now must be built nine inches above the level of what the water reached during the 1947 flood event. That flood was a surprise for people, as their previous historical
13
ecological experience could not correspond to the new flood realities. It created both physical (embedded in the landscape) and mental benchmarks in the local flood history of Datchet, which today’s residents referred to constantly while speaking about other floods that affected the community.
So that how we all knew about the floods, because of the history. Just search 1947 flood and you will see all the photos.
Mike, flood affected resident I think as I mention 1947 was the big one and then there was 2002 or 3 and also the last one [2014], the EA have the bigger picture probably.
Janet, flood marshal
Talking to Mr Smith, a local parish councillor, spokesperson on floods, and one of the first residents willing to talk about the Datchet CFAG, I discovered that for the local residents, the period from 1947 to 2003 was another lull that produced a state of ‘temporal vulnerability’ – or a condition of surprise – for the community (De Vries, 2011). The following quote from my interview with Mr Smith offers a vivid example of a somewhat ragged reproduction of the local flood history, full of temporal holes and loops:
In 1947 we had weather conditions when we had ice and ground frozen for a long time and then we had a large amount of rainfall … Since then we had floods in 2003 and that [2003] was caused by misoperation of the Jubilee River.
The lack of relatively meaningful floods between 1947 and 2003 formed the pool of experience and knowledge out of which local residents made projections about future flood risks. Within this period, smaller floods became something of a cultural normalcy, which was disrupted by the relatively big flood of 2003. According to local oral histories, persistent rain at the end of 2002 produced flood conditions in the Thames catchment and the removal of part of the flood bank contributed to the beginning of the flood. As result, the Datchet Village Greens, the golf course, and several roads were flooded and the school suffered badly. The water stayed in gardens and created a pond in the middle of the village. Although the flood of 2003 was not as big as the one in 1947 it brought new complications to Datchet since new infrastructure and housing arrangements had developed in the village since the 1950s.
Apart from the temporal situatedness of the 2003 flood (a relatively big flood following a period of smaller floods) and the unexpected ecological properties of the flood, the flood’s co-occurrence with the introduction of the new flood alleviation scheme shaped the meaning some of the local residents assigned to that flood. In early January 2003, the Environment Agency’s new flood alleviation scheme – the Jubilee River – was put into operation and tested for the first time. Whatever the real cause of the flood, some residents of the local community with whom I talked tended to associate the ‘abnormal’ amount of water with the introduction of the new flood scheme, blaming the Environment Agency for erroneous calculations and failures in the construction of the channel. One of the local community volunteers recalls:
The Jubilee River was put in to prevent flooding in Maidenhead, Eaton and protect all those areas from flooding, that was the intention … unfortunately, it wasn’t made properly, the design was incomplete, it had lots of faults and consequently the gates were opened and they lost control over the water flow and the Datchet was flooded.
The surprising flood of 2003 and the search for an answer – an attempt to bridge the gap between the historical ecological experience of previous flood events and the realities of the new environment – created a context in which a CFAG started to emerge. The group began as a branch of the Parish Council, at the initiative of a local engineer who replaced the former spokesperson on floods, who, according to the current chair of the CFAG did not have enough understanding of how to deal with the flooding and
14
possible solutions for flood mitigation. While trying to discover the cause of the flood in hydrological terms, the group began its communications with the Environment Agency, pressing it to release information on what caused the flood of 2003 and providing it with hydraulic data pointing out the miscalculations that led to the failure of the channel. In an effort to find an answer for what caused the unusual flood and a strong willingness to ‘do something about it’, accompanied by the CFAG chair’s background in engineering, the CFAG read the local flood history through a hydrological lens. Everything – from the language that the CFAG members used to describe floods to the content of their descriptions and material practices – contributed to what the group considered meaningful about the floods: hydrology. Their descriptions of floods were full of technical details, such as systemic explanations of water flows and the principles of water dynamics, as well as an emphasis on the course of the flood and the technical aspects of flood elimination. The quote from the interview with the chairman of the CFAG brings an understanding:
We now got an entirely different way of operating the JR [Jubilee River] and this should stop the hydraulic surges. The reason for the hydraulic surges is because the JR does not have a steel base – it should have this when it was designed … Because it now operates without a steel base, the hydraulic surge took out the foundations of some of the bridges and some of the other measures like weirs.
The CFAG’s flood history was centred on events that were considered of particular importance to the group. From the oral histories of the group’s chair, and from talking to a person who called himself an ‘independent volunteer’ but stayed in close contact with the CFAG, it became clear that the selecting of events to memorize often had a functional character. Reading the local flood history from the hydrological perspective, the CFAG selected to memorize those flood events that it considered ‘relevant’, events that could contribute to their understanding of how to alleviate future flood risks by improving existing flood defences or developing new ones. The following quote from the CFAG chair illustrates how certain flood events become baselines while others are considered functionally irrelevant:
But if we look over the last years from 2003 to 2014 ... yeah … [2007] is an event, underneath we recognized and that’s it. But it didn’t give us much information regarding dealing with major floods…it didn’t really affect us here… it is not a big deal.
The flood event of 2007, although reported by the flood authorities, did not attract much attention from the CFAG. Described as a quick summer flood caused by extensive rainfall, the flood of 2007 was considered relatively unimportant, compared to the existing baseline of the 2003 event since it did not bring new additional knowledge to managing flood risk in the village. The situation with the 2014 flood, however, was different. The hydrological properties of the 2014 flood such as the high speed of water, leakages in the bunds, pop-up flooding in unexpected areas, and an anticipated risk of sewer pipe failure, created a new state of surprise for the CFAG. It made the 2014 flood worth remembering, making it the second baseline in recent local flood history. The following quote from an interview with the head of the CFAG shows how the 2014 flood breached the community’s expectations by contradicting existing historical-ecological experience:
There is another issue in Datchet, which we didn’t know about. When the water levels come into Datchet common brook, they had backflow into the recreation ground. The water level pressurized the ground – which is all gravel – pressurized and caused second flooding in roads and houses that were not in any direct path of any flooding. Within 10 minutes it was on the front door of houses … funny enough, the river was dropping, and we had flooding increasing in other areas of Datchet where we weren’t expecting any flooding.
15
The attentiveness to the hydrological properties of floods can be found not only in the oral accounts of the group members but also in the ways the CFAG ‘materializes’ the flood history in flood mapping and record keeping. During one of our conversations in the Datchet Parish Council office, the head of the CFAG agreed to show me the maps the group had produced during and after the 2003 and 2014 floods (Figure 2 and 3), explaining which ‘major’ flood events deserved to be included in the flood records:
I took the records of the 2003, that was a major one and I have got 2013-14. I have got that all on full analysis. Some of the minor floods that were coming up, I have got some records of those, but they went away, the ones that really count are 2003 and 2014, I have got information and photos of those events. I was able then to calculate the height of water in that [2003] … and we know from the EA what the flow was in 2014 when the river level was higher in Datchet; I was able to compare it with 2003.
Figure 2. Flood map produced by Hydrological CFAG during and after the flood of 2003
16
The big lull between 1947 and 2003, accompanied with the development of the infrastructure in the area, introduction of a new flood alleviation scheme and unexpected hydrological properties of the 2003 flood breached experience-based expectations that Datchet residents acquired through dwelling on a floodplain. This resulted in a condition of surprise that revised flood-related activities of some of the local residents and paved the way for a formation of the local CFAG. Having its initial aim to understand the causes of unusual flood and then to make sure that village would be protected from the future flooding the group started to focus on the hydrological properties of floods and the technical aspects of FRM. This focus is reflected in the event-focused local flood history, in which the group makes the baselines out of flood events that contributed to understanding how to defend the community from future threats. Such a reading largely corresponds to the flood-defensive approach that currently remains dominant in the FRM field in the UK (Begg et al., 2017).
3.3. Emergence of the community-preparedness CFAG
The surprising flood of the 2014 also created the context for the formation of the second local CFAG within the village of Datchet. The temporal situatedness of the 2014 largely contributed to the surprising affect. While the 2003 event was particularly meaningful to the group of concerned residents forming the hydrological CFAG, the new flood group that emerged later in the community did not share this reading. The members of the new CFAG who I talked to recall that after 1947, Datchet was affected by floods several times: in 1990, 2000, 2003, and 2007, yet only the 2013-14 flood became the event that shifted historical baseline from 1947 flood. The collection of minor flood events between 1947 and 2014 made up the pool of experiences, memories, and attitudes towards floods, which in the most general sense can be characterized as a state of relative indifference. Thus, for instance, the events of the 2000s were described by one of the new CFAG members as minor flooding that affected no private properties, and that subsided after a short period of time:
I live here for 31 years. We have had one minor flood, not in a sense that we had in 2014, I cannot remember…in 1990 something, it was on a village greens, no properties damaged, lasted for 44 no 48 hours…if we go back, there were floods in 1940s, have you seen all this? All these photos about 1947?
The 2003 event was regarded as big but relatively unimportant compared to the event of 2014, as it did not cause much disruption to the community and any significant internal flooding. A village resident who volunteered during the 2014 flood described the event of 2003 in the following way:
2003 was one of 24-hour events, very quick. If you were at the end of the village and not coming to the centre, you would not know it was flooded, and no houses were flooded, the railway line wasn’t disrupted, it wasn’t as impactful [as 2014].
The temporal situatedness of the 2014 flood was also accompanied by the unusual socio-ecological properties of the event, which did not correspond to existing historical-ecological experience. The fluvial flooding of 2014 was accompanied with groundwater flooding, which caused the inundation of parts of the village that had not been affected before. Several private properties were flooded, local business premises were closed for several days, the railway lines were disrupted, and some parts of the village were inaccessible without special vehicles. The flood disrupted the normal functioning of the village and caused chaos. Oral histories suggest that the organization of sandbag distribution was lacking. People were trying to acquire as many sandbags as possible regardless of their personal level of exposure to risk, thus leaving those in need without protection against the water. When it became clear that the civil authorities did not have enough capacity to manage the crisis, the Royal Borough of Windsor and Maidenhead called in military troops to coordinate and assist in the flood management. The troops took the lead in flood mitigation activities including community engagement, installing water pumps, arranging the delivery of sandbags, and monitoring the areas under threat. A woman who organized a 24/7 ‘sandbag watch’ recalls:
17
There was a complete chaos. As soon as the army came in, they sort of organized all the essential things and we managed to get some more sandbags.
The involvement of the army and the military troops in FRM is a rare event that contributed to the surprise experienced by the new CFAG. According to English legislation, military forces cannot be involved in national government or local council duties until requested through a system known as Military Aid to Civil Authorities (MACA) (Ministry of Defence, 2015). Activation of MACA is, however, a last resort, used when the local authorities prove themselves unable to manage emergency situations. But if the involvement of the army is, at least, provided for the English system of emergency management, the involvement of the members of the Royal Family is certainly not included in a hazard management protocol. The arrival of the princes to Datchet and their physical involvement in the management of the flood became one more event that contributed to the surprising effect, making the 2014 flood a new baseline. The organizational void – chaotic distribution of FRM responsibilities within the community- created a context in which a group of local volunteers made a decision to create a Datchet CFAG with the aim to increase Datchet’s resilience by making the community aware of possible flood risks and ready to withstand future calamities.
Although the community-preparedness CFAG was composed of people who had lived in Datchet for 25-30 years and had experienced the same number of floods as the members of the hydrological CFAG, the local flood histories reproduced by the two groups appeared to be very different. Members of the community-preparedness CFAG assigned different meanings to the 2014 flood than the members of the hydrological group. For example, according to the hydrological CFAG, the involvement of the army represented an example of an organized collaboration between the parish and the higher authorities, while for the community-preparedness CFAG their arrival was proof of the community’s unpreparedness for future floods. Talking about this incident in a local cafe that had stayed open during the 2014 flood, a former parish councillor who was one of the initiators of the community-preparedness CFAG describes his attitude to the involvement of the army in the following way:
It is not about calling in a force. And when the army came they have been here for about 36 hours when they said, ‘we can’t help the community’, so the army called the community meeting saying ‘if you don’t come together we cannot help’. So, that was a lesson.
Along with the involvement of the army as proof of community unpreparedness, the need for community organization was seen in the lack of a network of volunteers who would be able to provide assistance, in different parts of the village, especially those that were difficult to access due to inundation; the lack of concern about vulnerable people, such as the elderly, who were unable to acquire sandbags during the flood; and the lack of communication between the local flood authority (Parish Council) and the village residents. Talking to the representatives of the community-preparedness CFAG, I learned that the Datchet community flood plan, which was supposed to provide people with all the necessary information on how to prepare for a flood, was abundant in technical details but was seen as unhelpful for the residents in terms of what to do and where to go before, during, and after the flood. The wife of the former councillor who initiated the community-preparedness CFAG described this problem in the following way:
I think that most of the people on the outside feel the same way, that there is a huge gap between the technical flood plan and the people. They [Parish Council] know all the organizations involved and know when to step in, good luck to them I'd say, but people just don’t read flood plans and even if they try to, the most people are not friendly to the documents. And so what is missing is very clear notification that when this happens … we are keeping an eye on that, so don’t panic… these are the numbers you call, or something like this. We need to have something that clearly states when people [should] worry, when they don’t, and what they need to do.
18
The community-preparedness reading of the local flood history can also be found in the artefacts that the CFAG produces, such as the Datchet Flood Advice website.1 The homepage refers to the events of the 2014 flood as the reason behind the CFAG, and the website gives flood advice ranging from how to prepare a household flood plan to how to deal with insurance issues after the flood. The idea and the content of the website clearly reflect the logic that stands behind a community-preparedness reading of the flood history: ‘helping to make the Datchet Community aware of the issues and to point to helpful sources of advice on matters flooding’.
Figure 4: The print screen of the Datchet Flood Advice website launched by the Community-preparedness CFAG after the flood of 2014
To summarize, the temporary lull between 1947 and 2014, unusual ecological features of the 2014 flood together with the uncoordinated responses of the Parish Council and the hydrological CFAG, which triggered the involvement of the Army, created a condition of surprise from which the second CFAG started to evolve. Emphasizing the importance community preparedness for future threats, the new CFAG developed as a counterpart to the hydrological flood group and as an interface between the public and the official structural efforts of flood mitigation. As the following section will show, although the community-preparedness CFAG clearly followed a desired shift towards more socially-oriented FRM proclaimed by the flood management policies, the group ran into challenges in finding support among other stakeholders involved in flood management in Datchet.
3.4. Multiple readings of flood history and the sustainability of the CFAGs
The Environment Agency held a drop-in session on 8 November 2017 between 4 and 7 p.m. in the local library, a cozy old building in the centre of the village. The building itself serves as a physical reminder of
19
the historical ecological experience of Datchet, with a flood warning board on the wall outside the building, a painting of the 16th-century flood, and a couple of folders labelled ‘Flood’ inside the library. The meeting was organized around a stand displaying various materials about the Thames Scheme: its project timeline, the budget, and the map with ‘propositions’, the three or four options where newly engineered construction could emerge. The meeting was chaired by the Environment Agency Flood Resilience Officer, who was assisted by two other members of the team who took notes of the meeting, writing down people’s questions, and proposing that the attendees subscribe for quarterly updates on the project. The meeting was casual, with people gathering around the display, talking to each other and to the representatives of the Environment Agency. The attendees mostly discussed the technical aspects, the options of where to install new flood-related construction. People stepped up to the map and pointed out certain places on the river, where, according to their knowledge and experience, the new constructions should or should not be deployed.
Talking to the EA Officer after the meeting, I learned that this meeting format was chosen purposefully, from experience, as it created an informal atmosphere that allows for more community engagement, according to Environment Agency. Choosing an informal drop-in session over a formal conference-like official meeting, the meeting in the library illustrated the Environment Agency’s attempt to follow a ‘social turn’. According to the agency, this type of meeting allows people to contribute their historical ecological knowledge to the development of a scheme, and is more conducive towards community engagement in flood management, as it does not exaggerate the potential differences in status of knowledge and experience in the room.
Figure 5: Display used by the Environment Agency during the drop-in session in Datchet Library in
November 2017
Yet, the space that the Environment Agency leaves for community input in flood management mimics the logic that stands behind the dominating official reading of the flood history: a focus on the hydrological properties of floods, the technical aspects of FRM, and the managing of flood risk in the long-term with the help of engineering solutions. The community here is regarded as a source of the local ecological
