Anticipating flash-floods: Multi-scale aspects of the social response
Introduction
The general goal of this research is to describe how people behave in front of a flash flood with a specific reference to the dynamics of their physical and human environment. Flash flooding is a short-fuse natural hazard demanding a prompt response to avoid casualties and damages (Ruin et al., 2008, Borga et al., 2011). Like other research groups over the last decade (Drobot and Parker, 2007), we are interested in how individuals behave in such situations, under what physical conditions and how they are warned and communicate together. We focus on the period between the onset of the generating storm (precursor) and the flood outbreak (peak of danger) that we name the anticipation period. Previous papers presented different aspects of the data collected (Creutin et al., 2009) and the specific methodology used to collect them during post flood investigations (Ruin et al., 2014). The space–time framework defined to appreciate the timeliness of human anticipation of flash floods has already been presented in Creutin et al. (2013). This paper uses the same set of field observations complemented with a new case. These observations were collected through a sort of snowball sampling: based on testimonies from emergency services, we start the interviews with people who experimented a critical situation, but got out alive. During the interviews, we collect the interpersonal network carried out during the event. The following interviews are taken inside this interpersonal network with the aim at covering the entire network. Therefore, the interviewees can be people affected by the flood, their relatives, or neighbors, emergency agents, or people who were by chance in the same place at the same moment during the flood. This method enable us a cross validation of information given the different interviewees.
Overall, the database counts 137 interviews made after three events: the 2002 Gard event in France (30 interviews), the Fella River event occured in Italy in 2003 (42 interviews) and the 2010 Var event, France (65 interviews). The two first events have already been described in Creutin et al. (2013). The Var event has already been presented by Ruin et al. (2014). If the grid for qualitative interviews evolved from one investigation to another, the method first developed in Creutin et al. (2009) has been used for coding the responses into actions undertaken during the events. Thus the set of data analyzed counts 165 anticipated actions, with descriptive information about the place and time of the actions, their social context (improvisation alone, with other people, or action following a planned strategy). The categorization of the actions regarding their objectives is made following the proposed classification, with a distinction between information, organization and protective actions.
The current paper elaborates more on the multi-scale and multi-level construction of the human response to flash-floods using terminology and concepts proposed in multi-disciplinary environmental problems to achieve some common understandings about cross scale issues (like in Gibson et al., 2000, Cash et al., 2006).
As for all extreme situations, the exceptional nature of flash flooding makes the observation and study of human response complicated. First, we need to analyze real-life experiences to bridge the gap between the intended behaviors and actual actions. The consequence is that post-event survey is the main mode of investigation with its inherent limitations in terms of sampling and reliability of narratives. Second, the framework used to study social responses when facing exceptional conditions needs to be quite open to a combination of approaches, as it is the case in an increasing number of social studies (Punch, 2013). Third, qualitative methodologies appear as a compulsory first step given the above-mentioned limitations in term of sampling and novelty of the type of environmental observation.
In our studies, the social response is addressed in terms of flash flood coping practices (as opposed to representation) through the narration of individuals who performed those actions, that we call stories. Given the scarcity of the narratives and the richness of their contents describing a variety of situations, we closely associated the data collection with its analysis by using a pragmatic approach. These methodological choices pertain respectively to the “practice based approach” (see for instance Corradi et al. (2010)) and the “grounded theory” (see for instance Corbin and Strauss (1990)).
Rooting human actions in the space–time dynamics of the socio-physical environment by integrating the sensitivity to environmental clues and understanding the use of available information invite to understand better the links that tie ”knowledge” to “action” – the goal of the Practice Theory. “Practice allows researchers to investigate empirically how contextual elements shape knowledge and how competence is built around a contingent logic of action” (Corradi et al., 2010). This theory considers “practices” as “embodied, materially mediated arrays of human activity centrally organized around shared practical understanding” (Schatzki et al., 2001). Looking at our problem through this prism transfers the question of the social response time to the question of a learning process in which a human “trial and error process” competes with the environment dynamics. Nevertheless, we face two limitations. One is related to the exceptional character of the investigated situations, a break of the daily activities in a narrow time–space window that corresponds to crisis definition (Reilly, 1993). The second is to be more interested in the performance of the response than in its motor. As stated by for science and technology studies, “a practice approach has often turned into an injunction to study performances rather than representations, doing rather than sense-making, the production of stability rather than the enactment of change” (Araujo et al., 2008). We believe that our approach is consistent with the view of the domain of social sciences given by as “neither the experience of individual actors nor the existence of any form of social totality but social practices ordered across space and time”.
The Grounded Theory, defined as “the discovery of theory from data” (Glaser and Strauss, 1999), aims at using data to derive theories that fits empirical situations (Strauss and Corbin, 1997, Glaser and Strauss, 1999). It fits well with the need of practical solutions in front of unusual situations. Our field observation method through interviews and the collection of narratives describing actions performed to cope with a rapidly changing situation shares many of the “canons and procedures” of this theory (Corbin and Strauss, 1990). We conduct in parallel data collection and analysis, progressively elaborating “concepts” able to label types of activity that we code into “categories”. Although we did not keep track through “memos” of the various adaptations of our interview format, we must admit retrospectively that our analysis “patterns and processes” evolved progressively according to the progression of our “theory”. Our sampling is directly led by the happenings we learn about along the field investigation.
This paper presents the current stage of our flash flood response “theory” relying on a set of narrated practices collected during past field investigations. As for many coupled human and natural systems (CHANS), the dynamical interactions between humans and rivers during flash floods deserves multidisciplinary attention should we want to define better warning and prevention policies (Liu et al., 2007). The problem presents at least three main dimensions: (i) the spottiness of the hydro-meteorological drivers, (ii) the gradation of human actions in this type of exceptional situations and (iii) the multi-level nature of the activated social networks. This paper is mainly devoted to the two last points that are dealt in Sections 4 An analytical scale of anticipatory actions, 5 Scale of human response networks. Section 2 examines the multi-scale nature of flash-flood anticipation, using a classical set of physical and human dimensions and introduces the notion of timeliness scale developed in Section 3. At the end, Section 6 opens the discussion concerning the cross levels and cross scales opportunities for better understanding the anticipation process facing flash floods.
Section snippets
Flash flood crisis dynamics
First, it is useful to recall the way we reduce the space–time dynamics of the flood drivers to a single dimension of “anticipation time” that summarizes the physical embedment of human activity.
Physically speaking, individual response to flooding is first considered to be a mobility issue: to move or to stay put. In fact when facing hazardous conditions individuals, being at home, at work, shopping or travelling between such places, have to decide whether it is safer to stay where they are or
Timeliness scale: the essence of flood space–time dynamics
Our hypothesis to appreciate the timeliness of anticipatory actions is threefold: (i) During a flood, the peak of danger is considered to be the moment when invading waters reach their highest level – we consider this moment in time as marking the end of the anticipation period; (ii) The remaining time before the peak of danger is obviously important to appreciate the timeliness – the same action, like taking protection measure, has not the same adequacy if realized before or after the peak of
An analytical scale of anticipatory actions
For flash floods as for the other foreseeable hazards, the question of the available time to prepare after the occurrence of the first precursors and before the danger is central. Carsell et al. (2004) defines the “timeline” of a flood crisis from the “storm onset” toward “recovery”. For them, the aim of a flood warning system is “to increase the mitigation time available prior to the onset of flooding”. We use here the term “flood anticipation” to describe this period of time including actions
Scale of human response networks
According to the terminology and concepts proposed by Gibson et al., 2000, Cash et al., 2006, flood crisis is the place of interaction of a number of human scales. We saw in the previous section that human activity during a flood can be coded following an analytical scale, and, in turn, that the evolution of the system states along the “anticipatory actions” axis provides action patterns and stories, that we can analyze. We examine in this section how to describe various levels of interpersonal
Cross-scale and cross-level question
What is the interest of the proposed scale analysis beyond a pure theoretical framework? We propose here to address several questions where a cross-scale approach may be a useful.
Question 1: Who is responsible for anticipatory actions in our societies governed by welfare states, where risk is “under the control” of ad-hoc official organizations? Is it only the highest level in the hierarchy of the crisis network? Table 2 shows that, based on narratives collected after different events, it
Conclusion
As a conclusion we would underline three points.
Firstly, the use of timeliness, anticipatory actions and crisis network scales enable to understand differently what happens both physically and socially during an extreme hydro-meteorological event. Unlike common view where the protection phase is the result of an organized chain of actions involving specific actors, this approach contributes to the awareness of a more complex process that evolves for each actor from the perception of the first
Acknowledgements
This study was partly funded through the project “AdaptFlood” funded by the French National Research Agency (ANR). The field experiments that provided the data set used benefited from the international program MISTRALS (Mediterranean Integrated Studies at Regional and Local Scales) and from the European Project HYDRATE (FP6 Project no. 037024). The authors are very grateful to Eve Gruntfest for enriching discussions and comments about the preliminary versions of the paper.
References (32)
- et al.
Flash flood forecasting, warning and risk management: the HYDRATE project
Environ. Sci. Policy
(2011) - et al.
Supplementing flash flood reports with impact classifications
J. Hydrol.
(2013) - et al.
A space and time framework for analyzing human anticipation of flash floods
J. Hydrol.
(2013) - et al.
Advances and challenges in flash flood warnings
Environ. Hazards
(2007) - et al.
The concept of scale and the human dimensions of global change: a survey
Ecol. Econ.
(2000) - et al.
Characterisation of selected extreme flash floods in Europe and implications for flood risk management
J. Hydrol.
(2010) The technology of effective crisis management: more than the daily routine
J. High Technol. Manage. Res.
(1993)- et al.
Human exposure to flash floods – relation between flood parameters and human vulnerability during a storm of September 2002 in Southern France
J. Hydrol.
(2008) - et al.
Market practices and forms: introduction to the special issue
Mark. Theory
(2008) - et al.
Quantifying the benefit of a flood warning system
Nat. Hazards Rev.
(2004)
Scale and cross-scale dynamics: governance and information in a multilevel world
Ecol. Soc.
A multi-instrumented approach to observing the activity rescheduling decision process
Transportation
Grounded theory research: procedures, canons, and evaluative criteria
Qual. Sociol.
Through the practice lens: where is the bandwagon of practice-based studies heading?
Manage. Learn.
Catchment dynamics and social response during flash floods: the potential of radar rainfall monitoring for warning procedures
Meteorol. Appl.
Cited by (19)
Forecasting extreme weather events and associated impacts: case studies
2023, Extreme Weather ForecastingInterdisciplinary Travel
2020, Mobilities Facing Hydrometeorological Extreme Events 2: Analysis of Adaptation RhythmsA data-driven analysis of flash flood hazard, fatalities, and damages over the CONUS during 1996–2017
2019, Journal of HydrologyCitation Excerpt :Flash floods are mainly caused by intense rainfall and they can be extremely dangerous to property, infrastructure, and human lives since they can happen with little or no warning (Hardy et al., 2016; Yin et al., 2016). The rapid onset of flash floods makes flood risk management a challenge and restricts the opportunity for timely communication and decision making, which in turn results in abundant damages and casualties (Creutin et al., 2013; Lutoff et al., 2016). Therefore, flash floods are known to be the most dangerous types of flooding (Gourley et al., 2017; Schroeder et al., 2016; Terti et al., 2017).
Establishing information seeking pathways in slow and flash floods
2018, International Journal of Disaster Risk ReductionCitation Excerpt :The value of effective communication with the community during a disaster is well established as a central factor in successful disaster management. It has been attributed to saving lives and livelihoods [1–8], while communication done badly has contributed to failed response [9,10] and even to deaths [11]. In Australia between 2003 and 2016, reviews of incidents and exercises, including the 2009 Black Saturday bushfires, found that between 19% and 21% of all recommendations for improvement related to communication with the community [12,13].
Exposure to flash floods: The conflict between human mobility and water mobility
2018, Mobility in the Face of Extreme Hydrometeorological Events 1: Defining the Relevant Scales of Analysis