Abstract

Previous studies have explored the consequences of flood events for exposed households and companies by focusing on single flood events. Less is known about the consequences of experiencing repeated flood events for the resilience of households and companies. In this paper, we therefore explore how multiple floods experience affects the resilience of exposed households and companies. Resilience was made operational through individual appraisals of households and companies’ ability to withstand and recover from material as well as health and psychological impacts of the 2013 flood in Germany. The paper is based on three different datasets including more than 2000 households and 300 companies that were affected by the 2013 flood. The surveys revealed that the resilience of households seems to increase, but only with regard to their subjectively appraised ability to withstand impacts on mobile goods and equipment (e.g., cars, TV, and radios). In regard to the ability of households to withstand overall financial consequences of repetitive floods, evidence for nonlinear (quadratic) trends can be found. With regard to psychological and health-related consequences, the findings are mixed but provide tentative evidence for eroding resilience among households. Companies’ resilience increased with respect to material assets but appears to decrease with respect to ability to recover. We conclude by arguing that clear and operational definitions of resilience are required so that evidence-based resilience baselines can be established to assess whether resilience is eroding or improving over time.

1. Introduction

The immediate experience of a flood event is often associated with enormous financial losses and can have considerable negative health-related as well as psychologically distressing consequences resulting in anxiety, depression, or posttraumatic stress disorder, among others (Fernandez et al. 2015). For companies, floods can cause considerable costs due to business interruption as well as direct economic damage to buildings and equipment (Chinh et al. 2016). In this sense, flood experience may undermine a companies’ economic prosperity or the resilience of flood-exposed people, which includes not just the immediate pathological consequences but also the wider perception of individual agency, emotions attached to the place of residency, social ties and networks, and trust in responsible institutions (Kuhlicke et al. 2011; Tapsell and Tunstall 2008; Walker-Springett et al. 2017).

At the same time, there is a rapidly growing body of research showing that flood events also offer an opportunity for change: in the aftermath of flood events, households, companies, and authorities seem to be more inclined to invest in adaptation measures in order to mitigate the consequences of future flood events (Kreibich et al. 2011; Thieken et al. 2007) and by doing so are often able to reduce the monetary damage caused by a subsequent flood event quite effectively (Bubeck et al. 2012b; Kreibich et al. 2017; Kreibich and Thieken 2009; Thieken et al. 2016b).

However, most studies treat single flood events in isolation. By relying on a cross-sectional survey design, empirical studies usually try to better understand how individuals withstand and recover from the impact of single flood events (Kuhlicke et al. 2011). Such event-centered studies, however, are hardly able to capture feedback effects (Bubeck et al. 2012b) or legacy effects due to the experience of frequent flood events. Therefore, little is known about how a household’s or company’s flood history shapes its ability to withstand and recover from a flood event. With this study, we would like to address this gap in research, which is relevant in view of future flood risk. Although there are profound regional differences and uncertainties in numerical projections of future flood risks (Kundzewicz et al. 2014), it is expected that flood risk tends to increase in the future. The reasons behind this are related to climate change (Hirabayashi et al. 2013; Milly et al. 2008) as well as the continuous increase in the population and economic assets in flood-prone areas (Jongman et al. 2012).

This study explores the cumulative effects of multiple flood experience on the resilience of affected households and companies. Resilience is understood here as a household’s or a company’s ability to withstand and recover from the impacts of a natural hazard. More specifically, we focus on how household members as well as companies perceive their own ability to withstand damage to immobile and mobile objects as well as to their psychological and health-related integrity. With regard to recovery processes, we focus on how long it took households to recover after the flood event (in months) and how their household situation changed as a result of the flood (i.e., worsened or improved). Companies were asked how long their business activities were interrupted or hampered as a consequence of the flood.

Empirically, the current research focuses on the 2013 flood in the Elbe and Danube catchments of Germany, which was a record-breaking event in hydrological terms (Schröter et al. 2015). With about EUR 6–8 billion it caused considerably lower direct losses than the 2002 flood, which resulted in damage costs amounting to EUR 11.6 billion (DKKV 2003, 2015). However, for many households and companies, the 2013 flood was the second or third flood event that they had experienced since 2002. In addition to the 2002 flood, floods also occurred in 1999, 2005, 2006, 2010, and 2011 in the Danube and Elbe catchments (Kienzler et al. 2015). We therefore regard the 2013 flood as a possibility for better understanding the interrelation of flood experience and resilience.

The current research is based on three different survey samples, all of them conducted and collected independently of each other. Acknowledging the differences in the ways the surveys were conducted and the different interests underlying the surveys, as well as the slightly different ways in which questions were asked and how answer categories were provided in the surveys, we do not strive for a comparison in the strict sense but rather look for common trends in our data.

Starting with a thorough literature review focusing on the interrelation between flood experience, adaptation, and resilience (chapter 2), we further specify our conceptual approach to resilience and adaptation as well as the methodology underlying the three different samples in chapter 3. In chapter 4, we present the empirical results and then summarize and discuss the methodological and policy-relevant implications of our findings in chapter 5. More detailed information on the statistical analyses is provided in the  appendix.

2. Literature review

a. Toward an operational definition of flood resilience

Resilience is an opalescent term that has multiple meanings in different scientific communities. Most of the contemporary work on resilience refers to the definition of ecological resilience developed by Holling (1973) who integrated three different stability properties under the unifying umbrella term resilience; that is, recovery (return to status quo after disturbance), resistance (buffering the impact of a disturbance), and persistence (staying intact as an identifiable object/subject over time) (Grimm and Wissel 1997).

This definition and its uncountable variations (Brand and Jax 2007; Fisher 2015) have become quite popular, not only among ecologists but also among social scientists (Cote and Nightingale 2012; Hutter et al. 2013; Kuhlicke 2013; Weichselgartner and Kelman 2015) as well as among scholars working at the boundary between society and ecology (Adger 2000; Folke et al. 2002; Thieken et al. 2014; Walker et al. 2004).

As an implication, there is no agreed-upon understanding let alone definition of resilience both in science and practice (Alexander 2013; Weichselgartner and Kelman 2015). To give just a few examples, it is debated whether resilience is a descriptive or a prescriptive concept (Berkes 2007; Kuhlicke 2013, 2019) and whether and how resilience differs from vulnerability (Cutter et al. 2008; Fekete et al. 2014) or adaptive capacity (Gallopín 2006). In this sense, resilience has become an all-encompassing concept overloaded with quite diverse stocks of meanings (Jones and Tanner 2017). Brand and Jax are therefore concerned that “both conceptual clarity and practical relevance of the concept of resilience are critically in danger” (Brand and Jax 2007, p. 1). Similarly, Weichselgartner and Kelman plea for more precisely defining and operationalizing resilience in order to be able to establish a clear baseline “against which to make decisions regarding the resilience level” (Weichselgartner and Kelman 2015, p. 10). They consider this a relevant step forward in order to overcome the vagueness of current resilience-based analysis and to better connect the concept with “people’s experience on the ground” (Weichselgartner and Kelman 2015, p. 15).

In this paper, we aim to take on this challenge by employing an operational definition of resilience, which draws on debates in natural hazards research (Cutter et al. 2008). It focuses on how the capacity of households’ and companies’ to cope with and recover from the impact of a flood is changing in consequence of experiencing multiple floods. Furthermore, we are interested in subjective interpretations of different dimensions of resilience (Jones and Tanner 2017), that means how households and companies evaluate their own ability to withstand and recover from flood-related impacts to immobile and mobile objects as well as to their psychological and health-related integrity.

The relevance of taking such a comprehensive, actor-centered perspective on resilience is highlighted by recent studies. It is apparent that flood events are associated with a range of tangible and intangible consequences (Bubeck et al. 2017; Kuhlicke et al. 2011; Thieken et al. 2016b). Next to monetary impacts (Kreibich et al. 2014; Meyer et al. 2013), this can include negative effects on physical and mental health, including injuries but also psychiatric symptoms (Ahern et al. 2005; Alderman et al. 2012): an increase in the cases of depression, anxiety, and psychosomatic symptoms (headache, bodily pain) and a higher probability of posttraumatic stress disorder (Bonanno et al. 2010; Stein et al. 2007). Furthermore, results indicate that flooding can negatively affects people’s psychological well-being (Hudson et al. 2019) and leads to an increased use of medication (Fernandez et al. 2015). In addition, Thieken et al. (2016a) and Reiter et al. (2018) found that German households affected by the 2013 Elbe flood perceived psychological stress and recovery activities more seriously than financial losses. These findings underline that an approach to resilience is needed that focuses on households’ and companies’ perception of their own ability to withstand and recover from the different kinds of negative consequences of flood events, including not just material aspects but also psychological and health-related consequences. At the same time, resilience of households and companies depends on both prior flood experience as well as on their previous adaptive actions.

b. Multiple flood experience, adaptation, and resilience

In 1962 Robert W. Kates used the expression “prison of experience” (Kates 1962, p. 140) to highlight how future actions are guided by past experiences (Kuhlicke 2010, 2015). Since then, studies have increasingly focused on the interrelation of flood experience, adaptation, and resilience (Begg et al. 2017; Wachinger et al. 2013). Previous studies showed that direct flood experience can have a positive effect on risk awareness (Felgentreff 2003; Gotham et al. 2018; Lawrence et al. 2014; Plapp and Werner 2006), can result in a higher awareness of the potential negative consequences of climate change (Ogunbode et al. 2017; Spence et al. 2011), and can increase the motivation to take adaptive actions (Bubeck et al. 2012a; Grothmann and Reusswig 2006; Lawrence et al. 2014; Zaalberg et al. 2009). However, flood experience should not be understood as the sole driving factor for adaptive behavior; adaptive behavior is influenced by various variables including descriptive norms, negative affect, and outcome as well as self-efficacy (for an overview, see van Valkengoed and Steg 2019).

At the same time, research has shown that personal adaptive measures (e.g., measures taken to protect and prepare one’s house before a flood event) have positive effects on flood related damage as they reduce flood-related financial damage up to 50% (Kreibich et al. 2015; Poussin et al. 2015). Furthermore, there is an emerging field of research researching the nonmaterial benefits of individual-level adaptation strategies (Lamond et al. 2018) by focusing on how adaptive actions are able to prevent negative psychological impacts or to provide a greater sense of security (Hudson et al. 2019; Joseph et al. 2015).

However, in most cases flood experience is treated in isolation. The dominant practice of conducting cross-sectional surveys with a focus on single flood events usually neglects that adaptive behaviors, flood experience, and resilience are not just developing dynamically over time, they are also mutually influencing each other. Although it is meanwhile acknowledged that the interrelation between adaptive behavior and risk awareness is shaped by feedback effects and recursive processes (Siegrist 2013, 2014), little is known about the feedback of flood experience on resilience. A study conducted by Begg et al. (2017) suggests that perceived severity of an actual flood experienced has a significant influence on relevant motivational factors shaping the willingness to take adaptive actions. Households that perceived health and stress-related consequences due to the 2013 flood in Germany and that took adaptive measure before the 2013 flood report significantly lower response efficacy as well as lower individual responsibility after the flood than respondents who implemented measures but perceived the consequences to be less severe.

However, to our knowledge, there is not a single study published that focuses on how the frequency of flood experience influences the ability of households and companies to withstand and recover from flood events. Based on the literature, we can only put forward the hypothesis that flood experience might be an important factor that enhances the ability of households to withstand flood related impacts and to recover quicker as they might have developed some kind of routine-based knowledge to cope more effectively with the immediate flood situation. At the same time, each flood event might also be associated with an enormous physical, mental, and collective encumbrance resulting in possibly negative effects in the long run. In this sense, multiple flood experiences can also be understood as a burden for affected households. For companies, the situation might be different, as they might systematically improve their coping and adaptive capacities. However, these are only speculative thoughts, as there is no empirical evidence available so far. With this study, we address this gap in knowledge by providing empirical insights on the interrelation of multiple flood experience and resilience.

3. Theoretical constructs and methodology

a. Theoretical constructs

This study is based on three different theoretical constructs (Fig. 1). Based on Cutter et al. (2008), we distinguish conceptually between 1) adaptive actions and emergency actions on the one hand and 2) resilience on the other hand. 3) The third relevant construct was respondents’ prior flood experience. Figure 1 provides a model underlying the analysis.

Fig. 1.

Theoretical constructs and their relationship.

Fig. 1.

Theoretical constructs and their relationship.

Individual adaptation and emergency actions are understood as actions that aim at preventing or minimizing the anticipated or actual negative consequences of future flood events. They comprise long-term precautionary measures such as informative measures (e.g., gathering information on flood loss mitigation, joining neighborhood flood networks), constructional measures (e.g., adapted interior fittings, sealing the cellar) and taking out natural hazard insurance. They also include emergency measures that are undertaken just before or during a flood event such as moving contents upstairs or protecting the building against inflowing water (Kreibich and Thieken 2009). Early warning is an important prerequisite for undertaking emergency measures; however, it was not part of this study.

Resilience is understood as a household’s or a company’s ability to withstand and/or recover after a disturbing event. While resistance describes the ability to buffer the negative consequences of a disturbing event, recovery is understood as the ability to return to daily life after a disturbing event. More specifically, we operationalized the construct “withstand” how households and companies rated the perceived impacts on immobile and mobile objects as well as to households’ respondents psychological and health-related integrity. This includes information about monetary losses (e.g., overall, building, household contents, equipment, goods, business interruption) as well as information about perceived consequences associated with a flood event indicated on ordinal scales (overall, building, household contents, material objects, sentimental objects, health, mental consequences/stress, personal strain, business interruption). The construct “recovery” was made operational by asking households how long it took them to recover after the flood event (in months) and how their situation changed as a result of the flood, whereas companies were asked how long their business activities were interrupted or hampered.

Flood experience was taken into account by focusing on households’ and companies’ prior flood experience. However, we did not analyze the total number of floods experienced, but rather used the 2013 flood as the reference event in all samples and grouped respondents for which the 2013 flood was the first, second or third flood event or more.

We included additional variables, such as sociodemographic variables (i.e., age, tenure, gender, and building type), companies’ characteristics (size, sector, tenure, spatial situation) as well as flood characteristics.

b. Case study areas

Empirically, the paper is based on three different surveys. The surveys were collected in the German states most affected by the 2013 flood, among these, the states of Saxony (EUR 1.9 billion direct losses in 2013) and Bavaria (EUR 1.3 billion direct losses) (DKKV 2015; see Fig. 2).

Fig. 2.

Areas affected by the 2013 flood and the number of questionnaires completed.

Fig. 2.

Areas affected by the 2013 flood and the number of questionnaires completed.

The 2013 flood caused with about EUR 6–8 billion lower direct losses than the 2002 flood, which resulted in damage costs amounting to EUR 11.6 billion (DKKV 2003, 2015). Damage on the household level usually results from demolished building structures as well as repair work, including replacing broken windows, repairing heating systems, etc. In addition, households needed to repair or replace damaged contents, such as domestic appliances, telephone and computer systems, furniture or carpets. Losses for companies related also to different aspects refer above all to building losses; equipment losses; losses to goods, products, and stock; and losses due to business interruption (Thieken et al. 2016b).

In Germany, private adaptation measures are requested by law from flood prone property owners in accordance with their resources and capabilities (Begg et al. 2017). Flood insurance penetration in Germany has increased gradually over recent years (Surminski and Thieken 2017), with large regional differences due to past compulsory flood insurance in the former German Democratic Republic and in the federal state of Baden-Wuerttemberg (Seifert et al. 2013).

c. Methodology

The different surveys were conducted independently of each other. Table 1 provides an overview of the distribution of the three surveys and their sampling procedures.

Table 1.

Overview of the main commonalities and differences of the samples.

Overview of the main commonalities and differences of the samples.
Overview of the main commonalities and differences of the samples.

Table 2 gives a detailed overview of how the single theoretical constructs were made operational in the different samples. All surveys addressed the overarching conceptual constructs of “adaptation” and “resilience,” however, there are differences in the items and scales used to operationalize the constructs; there are also differences in how the single questions were asked and what kind of answer categories the surveys provided. While in sample 1 households, respondents were asked to report about the consequences of all flood events experienced, respondents in samples 2 households and 3 companies were asked to report only about the 2013 flood. Prior flood experience was taken into account by asking respondents whether they had experienced a flood before and if yes, how many and when (before 2013). Furthermore, the extent to which adaptation and emergency measures were enquired about differed between sample 1 households on the one hand and sample 2 households and 3 companies on the other. The implications of the outlined differences are discussed more in depth in chapter 5).

Table 2.

Detailed overview on questions and answer categories provided in the samples.

Detailed overview on questions and answer categories provided in the samples.
Detailed overview on questions and answer categories provided in the samples.

Due to the differences in the design of the questionnaires and the data collection processes as well as the different operationalization of the central constructs/measures, we chose not to combine the samples but applied a sample-based approach to look for converging evidence or discrepancies across the three samples.

Sample 1 households was collected by employees of the Helmholtz Centre for Environmental Research and includes households that experienced repetitive flooding between 2002 and 2013. The survey was distributed personally and conducted in the German states of Saxony and Bavaria. All towns surveyed have experienced at least one flood event, and most of them two or even three. Flood risk maps from Zürs Public (since 2014 renamed Kompass Naturgefahren; www.kompass-naturgefahren.de) and flood risk maps produced by the Saxon government were used to identify the affected areas in Saxony. For Bavaria, local councils were approached to identify areas that experienced flooding in 2013. These councils provided us with detailed lists of those streets that had been affected by flooding in 2013 and before.

A total of 6502 questionnaires were distributed personally between six and eight months after the 2013 flood (see Begg et al. 2017 for details). This included a prepaid reply envelope and a letter explaining the background of the survey and asking residents to return the survey to the Helmholtz-Centre for Environmental Research (Zentrum fur Umweltforschung) (UFZ) within the next days. In total, 1380 completed surveys were returned, which provided an overall response rate of 21.2%. Of the 1380 surveys, 990 were completed by residents in Saxony (response rate of 21.7%), while 390 were completed by residents in Bavaria (response rate of 20%). 65% of respondents were affected by the 2013 flood (N = 889). Therefore, only these surveys were used in this analysis. The questionnaire included detailed questions about every single flood event experienced since 2002 in a chronological order to better understand the cumulative effects of repeated flood experience (1999 was neglected as it was considered as a rather minor flood even in comparison to the 2002 and subsequent flood events). This included questions about financial losses, financial support, and perceived consequences. The survey also contained questions about adaptation actions and about individual attitudes to the distribution of responsibility as well as about decision-making processes in flood risk management (Begg et al. 2017).

The dataset sample 2 households was collected under the responsibility of the University of Potsdam, Geoforschungs Zentrum (GFZ) German Research Centre for Geosciences and the reinsurance company Deutsche Rückversicherung (DKKV 2015; Thieken et al. 2016b). It was based on computer-aided telephone interviews (CATI) with residents that suffered from (financial) property damage due to flooding in 2013 in Germany. The interviews were conducted around nine months after the flood event. Based on information from affected districts and municipalities, flood reports, press releases, or flood maps [e.g., ZÜRS or flood masks derived from satellite data provided by German Aerospace Center (Deutschen Zentrum für Luft- und Raumfahrt) (DLR), Centre for Satellite Based Crisis information; http://www.zki.dlr.de/], street lists were compiled, which served as a basis for looking up the telephone numbers of potentially affected residents from the public telephone directory. All 43 281 searched telephone numbers were contacted, from which 16 554 numbers were either not recognized or available during the campaign. Another 16 721 households stated that they had not been affected by the flood, with “affected” being defined as having suffered (financial) flood damage. In total 1652 interviews were completed between 18 February and 24 March 2014.

The questionnaire represented a modified version of previous surveys (Kienzler et al. 2015; Kreibich et al. 2005; Thieken et al. 2017). Overall, it addressed a broad range of topics such as flood impact characteristics, physical and financial flood losses to buildings and household contents, flood experience and flood awareness, precautionary measures, early warning, emergency measures, evacuation, contamination of the flood water, cleanup, recovery of the affected household, general characteristics of the building and household contents, nonmaterial flood effects (on health), aid and financial compensation, and sociodemographic variables. At the beginning of the interview, the interviewer asked to question the person in the household with the best knowledge about the flood event.

The dataset sample 3 companies focused on small companies and was also collected by a joint venture of the University of Potsdam, GFZ, and Deutsche Rückversicherung (DKKV 2015; Thieken et al. 2017, 2016b). The list of affected streets that was compiled for sample 2 households formed the basis for generating property-specific random samples of companies (i.e., their telephone numbers). Additional effort was undertaken to increase the number of large companies in the sample. Therefore, expert interviews were undertaken and flood and press reports were analyzed to identify flood-affected large companies. All companies irrespective of their size were interviewed between May and July 2014 using a standardized survey, which resulted in 557 completed interviews. However, for these analyses only small companies with ≤10 employees and ≤EUR 1 million damage (n = 328) were selected since we assume that smaller companies behave similarly to households. The person with the most knowledge about the flood damage was interviewed.

The questionnaire addressed the following topics: flood impact (e.g., water depth, contamination), flood warning, emergency measures, evacuation, cleaning up, characteristics of and damage to company assets (buildings, equipment, goods, products, and stock, etc.), recovery, precautionary measures, flood experience and awareness, and characteristics of the company (sector, number of employees, etc.). Further details about the approach, partly based on preceding surveys, are published in Kreibich et al. (2007), Thieken et al. (2016a, 2017), and Sieg et al. (2017).

d. Statistical analysis

The surveys were coded and analyzed using the statistics software SPSS and R Version 3.3.2. For statistical analysis, we used correlational analysis (Spearman correlations) to test for bivariate relationships between flood experience and our outcome variables (uptake of adaptation and emergency actions, resilience indicators). For multivariate relationships, we applied multiple linear regression analysis or logistic regression analysis for dichotomous outcome variables. We included the adoption of adaptation measures (constructional adaptation actions, insurance), the adoption of emergency measures and resilience indicators (overall flood consequences, financial damage, property damage, loss of material objects, health- and stress-related consequences) as dependent variables as well as respondents’ sociodemographic characteristics (age, tenure, building type) and their previous flood experience as independent variables in the analysis. To test for possible nonlinear (i.e., quadratic) effects of previously experienced floods on household resilience, we included the number of previous flood experiences and its quadratic term in the regression (Miller et al. 2013). This strategy has been applied in previous research (Klein et al. 2011). Following Miller et al. (2013), we mean centered the data for the number of previously experienced floods prior to computing the quadratic term.

4. Results: Flood experience, adaptation, and resilience

a. Sociodemographic characteristics

The comparison of the main sociodemographic characteristics of the two different samples related to households (samples 1 households and 2 households) revealed quite profound differences (see  appendix for details, Table A2). Sample 2 households is characterized by a higher mean age, a higher proportion of female respondents, a higher share of homeowners, a higher share of detached houses, and a higher share of people with lower flood experience. Some of the differences can be traced back to differences in the sampling and screening procedures of the surveys, as well as the techniques of the survey. Telephone surveys are increasingly hampered by the substitution of landlines by mobile phones. Younger adults (<30 years), especially, tend to live in wireless-only households, leading to an underrepresentation of this age group in landline-only surveys (Greenberg and Weiner 2014). As a result, the average age in sample 2 households is higher than in sample 1 households; the latter was based on a paper-and-pencil questionnaire (see Table 1). Furthermore, the use of the term “flood-affected” differs across the three samples: while sample 1 households uses a broad interpretation of “flood-affected,” including residents that (only) suffered from no power supply or traffic disruptions, sample 2 households and 3 companies define “flood-affected” as having suffered (financial) flood damage. This resulted in high screen-out rates of almost 40% (see description of sample 2 households above) and probably led to a higher share of homeowners—and consequently to a higher share of detached and terraced houses—in sample 2 households compared to sample 1 households (Table 3). Finally, the design of the surveys differs across the three samples: sample 1 households was based on a lumped approach focusing on hot spot areas that have experienced repeated flooding, while samples 2 households and 3 companies aimed at surveying all areas that were flooded in 2013 including regions that had not been affected by the flood of 2002, for example, in the federal state of Baden-Wuerttemberg, Thuringia, most of Lower Saxony, and some regions in Saxony-Anhalt (along the river Saale). Consequently, the share of residents that experienced flooding for the first time is higher in sample 2 households than in sample 1 households (Table 3). Despite these differences, we think that the conceptual commonalities across the three samples are sufficient to jointly explore the effect of repeated flooding on adaptation and resilience.

Table 3.

Main sociodemographic characteristics of household samples.

Main sociodemographic characteristics of household samples.
Main sociodemographic characteristics of household samples.

b. Multiple flood experiences and adaptation

Based on previous findings (see section 2), the study is based on the hypothesis that the experience of (multiple) flood events leads to an increasing capacity to act before, during and after an event. This assumption was tested by analyzing how the number of floods that respondents experienced between 2002 and 2013 correlates with adaptation actions taken before or during the 2013 flood (see Tables A1A3 for details).

It is noteworthy, that the general percentage of interviewees who implemented adaptation actions and thus responded positively to the questions is higher in sample 2 households (see Table 4). We attribute this difference predominantly to two factors: first, the considerably higher share of homeowners in sample 2 households—homeowners usually have more possibilities (particularly in respect to constructional measures) and a higher motivation to implement adaptation measures (Bubeck et al. 2012a,b); second, to the way in which the questions were asked in the survey. While in sample 1 households, households were openly asked about whether and which adaptation actions they had taken, sample 2 households provided respondents with a prepared list of specific measures they could choose from. This influences the response behavior. For example, in a survey distributed in 2002, water pumps and backflow preventers were not included and therefore only a few household (less than 5%) mentioned these measures. In subsequent surveys, these measures were included and more than 20% (backflow preventer) and even 50%–70% (water pumps) stated they had these measures in place (Kienzler et al. 2015).

Table 4.

Adaptation/emergency actions and previous flood experience (2013 as reference flood).

Adaptation/emergency actions and previous flood experience (2013 as reference flood).
Adaptation/emergency actions and previous flood experience (2013 as reference flood).

Generally, the empirical results indicate a rather robust, but modest, significant positive correlation between flood experience and the adaptation/emergency actions of households and companies (for more details see Tables A1A3). There are significant correlations in all three samples for almost all of the tested variables indicating that the more often households experienced a (damaging) flood event, the more likely they were to take action in order to adapt. In addition, there is also a significant positive correlation with regard to taking out insurance and flood experience among the surveyed households. However, while households tend to take out insurance after experiencing their first flood, their opportunity to do so is decreased after the second flood (see Table 3). It is often the case that insurance companies do not insure buildings with prior flood damage, or only with high premiums (Seifert et al. 2013), although the assessment of insurability according to the hazard zoning system ZÜRS is gaining more importance than the number of previous claims (DKKV 2015).

In addition, we conducted a regression analysis to investigate which of the tested variables influence adaptation actions (see Table 5; for more details see Tables A4A6). For households, this included age, tenure, and building type in addition to flood experience. For companies, this included flood experience, size of the company, sector, spatial situation, and tenure.

Table 5.

Summary results of regression analysis—adaptation of households and companies. Note: * p < 0.05; ** p < 0.01.

Summary results of regression analysis—adaptation of households and companies. Note: * p < 0.05; ** p < 0.01.
Summary results of regression analysis—adaptation of households and companies. Note: * p < 0.05; ** p < 0.01.

The regression analysis for sample 1 households revealed that all tested adaptation and emergency measures show a positive relationship between flood experience and adaptation. There is also a positive relationship between tenure and adaptation (i.e., homeowners adopted adaptation measures more often than renters). The analysis for sample 2 households showed that being a homeowner and having experienced (multiple) floods before 2013 increases the likelihood for households to take adaptation and emergency measures, with the exception of taking out flood insurance, which is only influenced by tenure. In addition, both in sample 1 households and 2 households, age was negatively correlated with undertaking emergency measures during the flood and implementing constructional adaptation measures, respectively (sample 2 households).

Results for sample 3 companies revealed that flood experience has a significant positive influence on adaptation and emergency actions taken by companies, but again not on insurance. The only significant influence on the decision to take out natural hazard insurance was “tenure,” which seems to be plausible, since tenants are not likely to insure the respective building of their company (although they can insure their contents).

c. Households: Multiple experience and resilience

In a first step, we analyzed the ability of households to withstand damage to immobile objects (i.e., buildings). Figure 3 reveals that financial damage to a building fell considerably between the first and the second flood. The third flood was again associated with higher damage than the second flood. A drop in the amount of financial damage thus pointed toward an increased ability of households to withstand damage to buildings and other rather immobile objects. We attributed the higher financial damage in sample 2 households to the higher share of homeowners, who usually experience and report higher financial damage.

Fig. 3.

Financial damages, their perception, and previous flood experience (2013 reference event).

Fig. 3.

Financial damages, their perception, and previous flood experience (2013 reference event).

The perceived consequences of damage to buildings corresponded with reported financial damages. However, while the financial damage for the third flood were considerably lower than the first flood, the perceived consequences reach similar (sample 2 households) or even higher values (sample 1 households) with regard to the building damage (see Fig. 3).

There are different reasons for this nonlinear, U-shaped pattern. To exclude this U shape from being a result of flood characteristics, we tested whether and how the financial damage correlate with flood characteristics (i.e., water level and duration). While damage to the building correlated positively with both water depth and the duration of flooding (see Table A3), the results also indicated that respondents with more flood experience were affected by higher water levels, but still reported lower damage. This suggests that the observed pattern for financial damage with regard to flood experience cannot simply be explained by flood characteristics, but that other variables also influence the financial damage for companies and households—a point we return to in the discussions section in chapter 5.

In a second step, we analyzed the ability of households to withstand damage to mobile objects (contents). The ability of households to withstand losses to mobile objects increased significantly between the first and the second flood experienced (see Fig. 4): when the 2013 flood was the first flood experienced, this again resulted in the highest values with regard to the loss of or damage to mobile objects—whether of financial or sentimental value. However, thereafter values remained on a similar (sample 2 households) or slightly lower level (sample 1 households) if the 2013 flood was experienced as the third flood.

Fig. 4.

Perceived average consequences and previous flood experience (2013 as reference flood).

Fig. 4.

Perceived average consequences and previous flood experience (2013 as reference flood).

In a third step, we analyzed the ability of households to withstand negative psychological and health-related consequences. The ability to withstand the negative psychological and health-related consequences was again different to the previously reported ability to withstand damage to immobile and mobile objects (see Fig. 3), as it seems to decline with flood experience. However, there are also profound differences between both samples. In sample 1 households, there was evidence that health-related and—descriptively—psychological consequences follow a nonlinear pattern (i.e., consequences are perceived as more severe in the case of a third flood) indicating an exponential relationship between multiple flood experiences and stress- and health-related consequences. However, in sample 2 households, it was the first flood that was associated with the highest values, followed by the third flood.

In a fourth step, we analyzed how households were able to recover: the time it took for households to return to daily life and the relative change of the household situation compared to before the flood were positively correlated to the number of floods experienced, at least in sample 1 households. Among respondents who experienced 2013 as their third flood, 72% (n = 161) reported that it took them three months or longer to return to daily life, compared to 67.1% (n = 325) who experienced 2013 as their first and 52.7% (n = 366) as their second flood. Similarly, the share of respondents who perceived their household situation to be worse or considerably worse compared to their preflood situation was higher (43.9%, n = 157) for respondents that experienced 2013 as their third flood in comparison to the answers provided by respondents from households that had experienced less floods (first flood: 29.5%, n = 312; second flood: 29.5%, n = 353).

The results of sample 2 households indicate a somewhat different pattern regarding postimpact recovery. The results revealed that respondents who experienced floods for the third time needed significantly less time on average to completely restore their buildings compared to respondents who were affected for the first or second time. Results on recovery time with respect to household contents showed a similar, albeit nonsignificant trend. At this point, however, it must be taken into account that the corresponding question regarding respondents’ recovery was slightly different between sample 1 households and 2 households (see Table 2). In contrast to sample 1 households, the question in sample 2 households primarily related to the recovery time of material values and was evaluated separately for damage to building and damage to household contents.

To further investigate household resilience, we submitted the different measures of resilience (financial damage, perceived consequences, time to recover) to multiple regression analysis, including the number of flood events experienced as well as the ownership status (owner vs tenant), the type of building (detached/terraced/multifamily house), age (as covariates), and constructional adaptation (yes/no) in the regression. To additionally test for nonlinear effects of flood experience we also included the squared number of flood events experienced (for a similar strategy, see Klein et al. 2011; Masson and Barth 2019).

For sample 1 households, results of the regression analyses generally supported the findings described above (see Table 6; for more detailed information see Table A5), that is, the relationship between the number of floods experienced and the resilience measures (exception: psychological consequences) are nonlinear, corroborating the patterns displayed in Figs. 3 and 4. For instance, our findings indicated a quadratic (U shaped) relationship between the number of floods experienced on the one hand and the financial damage to households and consequences to the building on the other. The regression analysis conducted for sample 2 households revealed that there is a negative relationship between constructional adaptation and reported financial damage to household contents, indicating that more adapted households also reported lower damage. A similar pattern could be observed in sample 1 households with regard to the loss of sentimental objects and adaptation.

Table 6.

Summary results of regression analysis—adaptation of households. Note: * p < 0.05; ** p < 0.01.

Summary results of regression analysis—adaptation of households. Note: * p < 0.05; ** p < 0.01.
Summary results of regression analysis—adaptation of households. Note: * p < 0.05; ** p < 0.01.

d. Companies: Multiple flood experiences and resilience

Among the companies interviewed, the distributional pattern for financial damage with regard to buildings was different to households (see Fig. 5): companies that experienced the 2013 flood as their second flood event reported the highest damage on average. Financial damage due to business interruptions also resulted in the highest values, if 2013 was experienced as the second flood. However, both financial damage to the building and due to business interruption decreased with the third flood event. This seems to point toward a weak trend that asset damage decrease with increasing experience. This trend becomes more robust with respect to financial damage to mobile equipment and goods. Both correlated negatively with the number of floods experienced as damage decreased linearly with each flood event. This seems to point to a learning effect after each flood event resulting in lower damage.

Fig. 5.

Consequences for companies (financial damage and business interruption).

Fig. 5.

Consequences for companies (financial damage and business interruption).

Business interruption was shortest for those companies who experienced the 2013 flood as their first flood. The longest business interruptions were suffered by companies experiencing the 2013 flood as their second flood, whereby the difference between the first and the second flood were significant and the difference between the second and third were not. Similar to private households, it was also the case that companies with more flood experience were affected by higher water levels. Thus, the potentially positive effects of more experience are counteracted by the negative effects of higher hazard magnitudes.

For this sample, regression analysis revealed no significant relationship between experience and resilience (see Table 7). It was mainly the companies’ size that had an effect on their resilience.

Table 7.

Summary results of regression analysis—adaptation of companies. Note: * p < 0.05; ** p < 0.01.

Summary results of regression analysis—adaptation of companies. Note: * p < 0.05; ** p < 0.01.
Summary results of regression analysis—adaptation of companies. Note: * p < 0.05; ** p < 0.01.

5. Discussion

The three different surveys revealed that there is a strong positive correlation between flood experience and adaptation among all samples: households and companies that had experienced repeated flooding prior to 2013 were more likely to invest in adaptation actions and were also more capable of mitigating damage during the 2013 flood (also by taking emergency actions) compared to respondents with less or no previous flood experience. Our findings thus confirm previous research on flood experience and adaptation, both for affected households and companies.

Based on our empirical data and the analyses we conducted, it is difficult to judge whether increased adaptation also leads to greater resilience of households and companies. Our findings rather underline that resilience is influenced by many characteristics.

Prior flood experience shapes the resilience of households and companies. Generally speaking, between households’ first and second experienced flood, an increase in resilience can be detected. When 2013 was experienced as the second, the ability to withstand financial damage was higher, as well as the ability to withstand damage to immobile and mobile objects. Furthermore, the ability to withstand negative psychological and health-related consequences increased or remained on a similar level. We attribute the difference between the first and the second flood to a learning effect that results in an increased capacity to act upon and to cope with the consequences of flooding.

When households experienced the 2013 flood as their third flood, their ability to withstand damage to immobile objects decreased compared to the second flood. Experiencing a third flood seemed to act as a tipping point; a preliminary finding that needs further exploration in future research. There are different reasons for this nonlinear, U-shaped pattern. While damage to the building correlated positively with both water depth and duration of flooding (see Table A3), the results also indicated that respondents with more flood experience were affected by higher water levels, although they still reported lower damage. This suggests that the observed pattern for financial damage with regard to flood experience cannot simply be explained by flood characteristics. Even if somewhat speculative, the reported higher damage for the third flood might also be a result of learning effects among respondents of how to account for financial damage. This could therefore represent a reporting bias due to a more realistic assessment of the financial damage compared to households with less flood experience.

When households experienced the 2013 flood as their third flood, their ability to withstand negative psychological and health-related consequences also decreased. While sample 2 households show a similar U-shape pattern as the financial damage, the findings of sample 1 households reveal an exponential relationship. There is evidence that households’ capacity to cope with a flood event emotionally and healthwise is undermined, if respondents have experienced their third flood. It also took households longer to recover and their situation was worse or considerably worse compared to before the flood (sample 1 households). In our interpretation this supports the assumption that experiencing three floods within a relatively short time span can be understood as a chronically distressing situation that results in a deterioration of human well-being and thus undermines the resilience of households (see also research on learned helplessness for similar results; Seligman and Peterson 2001).

If households experienced their third flood, their ability to withstand damage to mobile objects increased. Such objects can be moved without great effort and residents report lower losses with each flood event experienced. Another finding of sample 2 households points in a similar direction, revealing that respondents who experienced the 2013 flood as their third flood, required significantly less time on average to restore their buildings compared to respondents who were affected for the first or second time. With each flood event experienced, households develop certain flood-related capacities that help them to reduce damage and restore buildings and interiors quicker than for previous floods. However, the result might also indicate that households simply had less mobile objects that were exposed to the risk of flooding. This, however, is difficult to capture in a survey.

Companies’ ability to withstand direct financial losses increases when they experience multiple flood events, particularly with regard to mobile goods and equipment. However, the negative consequences of business interruption increase with flood experience. This indicates that companies are able to reduce damage to their material assets, but they have difficulties to recover from the flood event and to return to the same state of business that they were in before the flood occurred.

With regard to losses of valuable objects, findings for companies were quite similar to the households’ results. Furthermore, companies seem to build up experience-based capacities with each flood event, reducing the vulnerability of mobile equipment and goods. Again, bivariate correlation analysis with respect to sample 3 companies also supports this relationship as companies with multiple flood experiences have significantly taken adaptation and emergency actions more often. However, such companies also experienced higher damage, which might have been even higher without adaptation, pointing again to the contingent relationship between adaptation and resilience. Whether this finding also holds for medium-size or large companies still needs to be tested.

Implications for future research

Based on our findings, we first suggest that there is a need to establish more systematic and evidence-based, resilience-related baselines against which the various interventions of local, regional, and national initiatives or the implementation of private (i.e., at the property level) or public adaptation measures can be evaluated (Weichselgartner and Kelman 2015). In many countries private actors are asked to take on greater responsibility in flood risk management and more generally in terms of climate adaptation, by reducing their individual risks through their own adaptation actions (Begg et al. 2015). It is usually assumed that such actions have positive effects (Begg et al. 2017). While our findings support such effects, particularly with regard to mobile objects (e.g., furniture, cars, and sentimental objects), such simple relationships are harder to detect with regard to building damage, but also with regard to perceived psychological and health-related consequences. Therefore, the establishment of evidence-based resilience baselines is needed in order to understand whether the redistribution of responsibility to the individual level would have the desired effects (i.e., an increase in individual resilience) or whether the capacity of individuals is limited to increase their resilience.

Second, our findings suggest that greater efforts should be undertaken to better understand the relationship between adaptation and resilience. Previous research has often focused on the factors that shape people’s motivation to invest in property-level adaptation measures. The implementation of adaptation measures is usually equated to an increase in households’ or companies’ resilience. Our findings at least question this simple, linear relationship (correlation) and hint at a positive, but by far more complicated and nonlinear relation. The reasons for this are manifold. Property-level adaptation measures often have a design level that might be overtopped by a flood and to be effective they need to be properly in place before premises are flooded; if homeowners (or employees) are not available or not capable of installing the measure in time, then the measure cannot fulfill its full potential. In addition, the effects of adaptation measures to reduce nonmonetary consequences are probably not as significant as one would expect. However, in order to better understand these effects more research is required to assess the relation between adaptation/emergency actions and resilience in a more systematic manner.

Third, more research is needed on the framing effects and how they shape respondents’ answers in survey-based research. We attribute some of the differences in our findings, particularly with regard to health-related and psychological consequences, but also with respect to the greater perceived severity of the third flood event, to the way we addressed households’ flood experience in the survey. While sample 2 households focused on the 2013 flood and addressed prior flood experience just as a lumped variable, sample 1 households focused on all floods experienced since 2002. Respondents were asked to report detailed about the consequences of each flood event they experience chronologically. This means that if the 2013 flood was experienced as the third flood event by households, then they had to answer several questions about previous floods before turning their attention to the 2013 flood. In this sense, sample 1 households primed the cumulative flood experience of households. Although our results hint at possible priming effects, more experimental research is needed to better understand how the framing/cognitive processing of previous flood events (in surveys) translates into appraisals of these events. For example, experimental designs that manipulate respondents’ processing of flood-related information (i.e., the time given to think about previous flood events or the number of questions asked about previous events) may provide an opportunity to investigate such framing effects more systematically.

Finally, more comparative research is needed. Comparing case studies in order to generate converging evidence or discrepancies across the three samples was only possible as we shared some of the theoretical constructs. We are convinced that more systematic comparisons across case studies is necessary, which requires researchers to agree upon, share, and apply some basic theoretical constructs and their operationalization across different case studies. Future research, however, might also use moderator analysis (as a part of meta-analysis) to investigate possible boundary conditions of (our) research findings.

6. Summary and conclusions

The key question addressed in this paper was whether the resilience of households and companies increases or decreases as a consequence of multiple flood experiences and how this is influenced by respondents’ emergency and adaptation actions. Our analysis shows robust evidence for a positive relationship between flood experience and the adoption of adaptation and emergency measures across all three samples. The relationship between flood experience and resilience is not as straightforward: The effect of multiple flood experiences depends upon the aspects of resilience that one is focusing on (i.e., financial losses, the subjective perception of such losses or perceived psychological and health-related impacts) as well as on the number of floods experienced. While our findings suggest that the resilience of households increases from the first to the second flood, the third flood event seems to be a tipping point: households’ ability to withstand damage to immobile objects decreases as does their ability to withstand negative psychological and health-related consequences. Our findings furthermore suggest in this respect that multiple flood experiences are a burden for affected households. Similarly, companies’ resilience in terms of their ability to withstand direct financial losses increases when they have experienced multiple flood events, particularly with regard to mobile goods and equipment. However, companies also have some difficulties recovering from the flood event and reassuming their preflood state of business if they have experienced multiple flood events. However, there are clear differences between samples that cannot easily be explained by the data.

We therefore conclude that while it is relevant and politically desirable to establish evidence-based resilience baselines for households and companies, it seems equally relevant to be more consistent when defining and operationalizing resilience in quantitative studies and to have a better methodological understanding of the effects from how flood experience is framed in surveys as well as how sampling and screening procedures affect respondents’ answers. As the results of this study indicate, following diverse sets of sampling and framing approaches can be very positive, particularly if results point to similar patterns, even when they have relied on different sampling and framing strategies. However, differences in the results of empirical research also need to be better understood and assessed as to whether they can be explained by the underlying methodology. If this challenge is ignored, there is the risk of both overestimation and misrepresentation of the effectiveness of property-level adaptation on the resilience of households and companies.

APPENDIX

Correlation Tables and Results of the Regression Analysis

The appendix shows the results of the correlation as well as of the regression analysis (Tables A1A9).

Table A1.

Correlations (Corr) sample 1 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.

Correlations (Corr) sample 1 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Correlations (Corr) sample 1 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Table A2.

Correlations (Corr) sample 2 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.

Correlations (Corr) sample 2 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Correlations (Corr) sample 2 households. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Table A3.

Correlations (Corr) sample 3 companies. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.

Correlations (Corr) sample 3 companies. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Correlations (Corr) sample 3 companies. Note: * p < 0.05; ** p < 0.01. Significant correlations are in bold.
Table A4.

Linear and logistic regressions of adaptive and emergency actions sample 1 households. DV: dependent variable. Note: * p < 0.05; ** p < 0.01.

Linear and logistic regressions of adaptive and emergency actions sample 1 households. DV: dependent variable. Note: * p < 0.05; ** p < 0.01.
Linear and logistic regressions of adaptive and emergency actions sample 1 households. DV: dependent variable. Note: * p < 0.05; ** p < 0.01.
Table A5.

Regressions of adaptive and emergency actions sample 2 households. Note: * p < 0.05; ** p < 0.01.

Regressions of adaptive and emergency actions sample 2 households. Note: * p < 0.05; ** p < 0.01.
Regressions of adaptive and emergency actions sample 2 households. Note: * p < 0.05; ** p < 0.01.
Table A6.

Linear and logistic regressions of adaptive and emergency actions sample 3 companies. Note: * p < 0.05; ** p < 0.01.

Linear and logistic regressions of adaptive and emergency actions sample 3 companies. Note: * p < 0.05; ** p < 0.01.
Linear and logistic regressions of adaptive and emergency actions sample 3 companies. Note: * p < 0.05; ** p < 0.01.
Table A7.

Regressions of resilience sample 1 households. Note: * p < 0.05; ** p < 0.01.

Regressions of resilience sample 1 households. Note: * p < 0.05; ** p < 0.01.
Regressions of resilience sample 1 households. Note: * p < 0.05; ** p < 0.01.
Table A8.

Regressions of resilience sample 2 households. Note: * p < 0.05; ** p < 0.01.

Regressions of resilience sample 2 households. Note: * p < 0.05; ** p < 0.01.
Regressions of resilience sample 2 households. Note: * p < 0.05; ** p < 0.01.
Table A9.

Regressions of resilience sample 3 companies. Note: * p < 0.05; ** p < 0.01.

Regressions of resilience sample 3 companies. Note: * p < 0.05; ** p < 0.01.
Regressions of resilience sample 3 companies. Note: * p < 0.05; ** p < 0.01.

REFERENCES

REFERENCES
Adger
,
W. N.
,
2000
:
Social and ecological resilience: Are they related?
Prog. Hum. Geogr.
,
24
,
347
364
, https://doi.org/10.1191/030913200701540465.
Ahern
,
M.
,
R. S.
Kovats
,
P.
Wilkinson
,
R.
Few
, and
F.
Matthies
,
2005
:
Global health impacts of floods: Epidemiologic evidence
.
Epidemiol. Rev.
,
27
,
36
46
, https://doi.org/10.1093/epirev/mxi004.
Alderman
,
K.
,
L. R.
Turner
, and
S.
Tong
,
2012
:
Floods and human health: A systematic review
.
Environ. Int.
,
47
,
37
47
, https://doi.org/10.1016/j.envint.2012.06.003.
Alexander
,
D. E.
,
2013
:
Resilience and disaster risk reduction: An etymological journey
.
Nat. Hazards Earth Syst. Sci.
,
13
,
2707
2716
, https://doi.org/10.5194/nhess-13-2707-2013.
Begg
,
C.
,
G.
Walker
, and
C.
Kuhlicke
,
2015
:
Localism and flood risk management in England: The creation of new inequalities?
Environ. Plann.
,
33C
,
685
702
, https://doi.org/10.1068/c12216.
Begg
,
C.
,
M.
Ueberham
,
T.
Masson
, and
C.
Kuhlicke
,
2017
:
Interactions between citizen responsibilization, flood experience and household resilience: Insights from the 2013 flood in Germany
.
Int. J. Water Resour. Dev.
,
33
,
591
608
, https://doi.org/10.1080/07900627.2016.1200961.
Berkes
,
F.
,
2007
:
Understanding uncertainty and reducing vulnerability: Lessons from resilience thinking
.
Nat. Hazards
,
41
,
283
295
, https://doi.org/10.1007/s11069-006-9036-7.
Bonanno
,
G. A.
,
C. R.
Brewin
,
K.
Kaniasty
, and
A. M. L.
Greca
,
2010
:
Weighing the costs of disaster: Consequences, risks, and resilience in individuals, families, and communities
.
Psychol. Sci. Public Interest
,
11
,
1
49
, https://doi.org/10.1177/1529100610387086.
Brand
,
F. S.
, and
K.
Jax
,
2007
:
Focusing the meaning(s) of resilience: Resilience as a descriptive concept and a boundary object
.
Ecol. Soc.
,
12
,
23
, https://doi.org/10.5751/ES-02029-120123.
Bubeck
,
P.
,
W. J. W.
Botzen
, and
J. C. J. H.
Aerts
,
2012a
:
A review of risk perceptions and other factors that influence flood mitigation behavior
.
Risk Anal.
,
32
,
1481
1495
, https://doi.org/10.1111/j.1539-6924.2011.01783.x.
Bubeck
,
P.
,
W. J. W.
Botzen
,
H.
Kreibich
, and
J. C. J. H.
Aerts
,
2012b
:
Long-term development and effectiveness of private flood mitigation measures: An analysis for the German part of the river Rhine
.
Nat. Hazards Earth Syst. Sci.
,
12
,
3507
3518
, https://doi.org/10.5194/nhess-12-3507-2012.
Bubeck
,
P.
,
A.
Otto
, and
J.
Weichselgartner
,
2017
: Societal impacts of flood hazards. Oxford Research Encyclopedia on Natural Hazards Science, Oxford University Press, https://doi.org/10.1093/acrefore/9780199389407.013.281.
Chinh
,
D. T.
,
P.
Bubeck
,
N. V.
Dung
, and
H.
Kreibich
,
2016
:
The 2011 flood event in the Mekong Delta: Preparedness, response, damage and recovery of private households and small businesses
.
Disasters
,
40
,
753
778
, https://doi.org/10.1111/disa.12171.
Cote
,
M.
, and
A. J.
Nightingale
,
2012
:
Resilience thinking meets social theory
.
Prog. Hum. Geogr.
,
36
,
475
489
, https://doi.org/10.1177/0309132511425708.
Cutter
,
S. L.
,
L.
Barnes
,
M.
Berry
,
C.
Burton
,
E.
Evans
,
E.
Tate
, and
J.
Webb
,
2008
:
A place-based model for understanding community resilience to natural disasters
.
Global Environ. Change
,
18
,
598
606
, https://doi.org/10.1016/j.gloenvcha.2008.07.013.
DKKV
,
2003
: Hochwasservorsorge in Deutschland—Lernen aus der Katastrophe 2002 im Elbegebiet (Flood prevention in Germany—Learning from the 2002 catastrophe in the Elbe area). Deutsches Komitee für Katastrophenvorsorge e.V., 152 pp.
DKKV
,
2015
: Das Hochwasser im Juni 2013: Bewährungsprobe für das Hochwasserrisikomanagement in Deutschland (The flood in June 2013: A test for flood risk management in Germany). DKKV Series 53, 207 pp.
Fekete
,
A.
,
G.
Hufschmidt
, and
S.
Kruse
,
2014
:
Benefits and challenges of resilience and vulnerability for disaster risk management
.
Int. J. Disaster Risk Sci.
,
5
,
3
20
, https://doi.org/10.1007/s13753-014-0008-3.
Felgentreff
,
C.
,
2003
:
Post-disaster situations as “window of opportunity”? Post-flood perceptions and changes in the German Odra River region after the 1997 flood
.
Erde
,
134
,
163
180
.
Fernandez
,
A.
,
J.
Black
,
M.
Jones
,
L.
Wilson
,
L.
Salvador-Carulla
,
T.
Astell-Burt
, and
D.
Black
,
2015
:
Flooding and mental health: A systematic mapping review
.
PLOS ONE
,
10
, e0119929, https://doi.org/10.1371/journal.pone.0119929.
Fisher
,
L.
,
2015
:
Disaster responses: More than 70 ways to show resilience
.
Nature
,
518
,
35
–35, https://doi.org/10.1038/518035a.
Folke
,
C.
,
S.
Carpenter
,
T.
Elmqvist
,
L.
Gunderson
,
C. S.
Holling
, and
B.
Walker
,
2002
:
Resilience and sustainable development: Building adaptive capacity in a world of transformations
.
Ambio
,
31
,
437
–440, https://doi.org/10.1579/0044-7447-31.5.437.
Gallopín
,
G. C.
,
2006
:
Linkages between vulnerability, resilience, and adaptive capacity
.
Global Environ. Change
,
16
,
293
303
, https://doi.org/10.1016/j.gloenvcha.2006.02.004.
Gotham
,
K. F.
,
R.
Campanella
,
K.
Lauve-Moon
, and
B.
Powers
,
2018
:
Hazard experience, geophysical vulnerability, and flood risk perceptions in a postdisaster city, the case of New Orleans
.
Risk Anal.
,
38
,
345
356
, https://doi.org/10.1111/risa.12830.
Greenberg
,
M. R.
, and
M. D.
Weiner
,
2014
:
Keeping surveys valid, reliable, and useful: A tutorial
.
Risk Anal.
,
34
,
1362
1375
, https://doi.org/10.1111/risa.12250.
Grimm
,
V.
, and
C.
Wissel
,
1997
:
Babel, or the ecological stability discussions: An inventory and analysis of terminology and a guide for avoiding confusion
.
Oecologia
,
109
,
323
334
, https://doi.org/10.1007/s004420050090.
Grothmann
,
T.
, and
F.
Reusswig
,
2006
:
People at risk of flooding: Why some residents take precautionary action while others do not
.
Nat. Hazards
,
38
,
101
120
, https://doi.org/10.1007/s11069-005-8604-6.
Hirabayashi
,
Y.
,
R.
Mahendran
,
S.
Koirala
,
L.
Konoshima
,
D.
Yamazaki
,
S.
Watanabe
,
H.
Kim
, and
S.
Kanae
,
2013
:
Global flood risk under climate change
.
Nat. Climate Change
,
3
,
816
821
, https://doi.org/10.1038/nclimate1911.
Holling
,
C. S.
,
1973
:
Resilience and stability of ecological systems
.
Annu. Rev. Ecol. Syst.
,
4
,
1
23
, https://doi.org/10.1146/annurev.es.04.110173.000245.
Hudson
,
P.
,
W. J. W.
Botzen
,
J. K.
Poussin
, and
J. C. J. H.
Aerts
,
2019
:
The impacts of flooding and flood preparedness on subjective well-being: A monetisation of the tangible and intangible impacts
.
J Happiness Stud.
,
20
,
665
682
, https://doi.org/10.1007/s10902-017-9916-4.
Hutter
,
G.
,
C.
Kuhlicke
,
T.
Glade
, and
C.
Felgentreff
,
2013
:
Natural hazards and resilience: Exploring institutional and organizational dimensions of social resilience
.
Nat. Hazards
,
67
,
1
6
, https://doi.org/10.1007/s11069-011-9901-x.
Jones
,
L.
, and
T.
Tanner
,
2017
:
‘Subjective resilience’: Using perceptions to quantify household resilience to climate extremes and disasters
.
Reg. Environ. Change
,
17
,
229
243
, https://doi.org/10.1007/s10113-016-0995-2.
Jongman
,
B.
,
P. J.
Ward
, and
J. C. J. H.
Aerts
,
2012
:
Global exposure to river and coastal flooding: Long term trends and changes
.
Global Environ. Change
,
22
,
823
835
, https://doi.org/10.1016/j.gloenvcha.2012.07.004.
Joseph
,
R.
,
D.
Proverbs
, and
J.
Lamond
,
2015
:
Assessing the value of intangible benefits of property level flood risk adaptation (PLFRA) measures
.
Nat. Hazards
,
79
,
1275
1297
, https://doi.org/10.1007/s11069-015-1905-5.
Kates
,
R. W.
,
1962
: Hazard and choice perception in flood plain management. Department of Geography, University of Chicago Research Paper 78, 157 pp.
Kienzler
,
S.
,
I.
Pech
,
H.
Kreibich
,
M.
Müller
, and
A. H.
Thieken
,
2015
:
After the extreme flood in 2002: Changes in preparedness, response and recovery of flood-affected residents in Germany between 2005 and 2011
.
Nat. Hazards Earth Syst. Sci.
,
15
,
505
526
, https://doi.org/10.5194/nhess-15-505-2015.
Klein
,
O.
,
L.
Licata
, and
S.
Pierucci
,
2011
:
Does group identification facilitate or prevent collective guilt about past misdeeds? Resolving the paradox
.
Br. J. Soc. Psychol.
,
50
,
563
572
, https://doi.org/10.1111/j.2044-8309.2011.02028.x.
Kreibich
,
H.
, and
A. H.
Thieken
,
2009
:
Coping with floods in the city of Dresden, Germany
.
Nat. Hazards
,
51
,
423
436
, https://doi.org/10.1007/s11069-007-9200-8.
Kreibich
,
H.
,
A. H.
Thieken
,
T.
Petrow
,
M.
Müller
, and
B.
Merz
,
2005
:
Flood loss reduction of private households due to building precautionary measures—Lessons learned from the Elbe flood in August 2002
.
Nat. Hazards Earth Syst. Sci.
,
5
,
117
126
, https://doi.org/10.5194/nhess-5-117-2005.
Kreibich
,
H.
,
M.
Müller
,
A. H.
Thieken
, and
B.
Merz
,
2007
:
Flood precaution of companies and their ability to cope with the flood in August 2002 in Saxony, Germany
.
Water Resour. Res.
,
43
, W03408, https://doi.org/10.1029/2005WR004691.
Kreibich
,
H.
,
I.
Seifert
,
A. H.
Thieken
,
E.
Lindquist
,
K.
Wagner
, and
B.
Merz
,
2011
:
Recent changes in flood preparedness of private households and businesses in Germany
.
Reg. Environ. Change
,
11
,
59
71
, https://doi.org/10.1007/s10113-010-0119-3.
Kreibich
,
H.
, and Coauthors
,
2014
:
Costing natural hazards
.
Nat. Climate Change
,
4
,
303
, https://doi.org/10.1038/nclimate2182.
Kreibich
,
H.
,
P.
Bubeck
,
M.
Van Vliet
, and
H.
De Moel
,
2015
:
A review of damage-reducing measures to manage fluvial flood risks in a changing climate
.
Mitig. Adapt. Strategies Global Change
,
20
,
967
989
, https://doi.org/10.1007/s11027-014-9629-5.
Kreibich
,
H.
, and Coauthors
,
2017
:
Adaptation to flood risk: Results of international paired flood event studies
.
Earth’s Future
,
5
,
953
965
, https://doi.org/10.1002/2017EF000606.
Kuhlicke
,
C.
,
2010
:
The dynamics of vulnerability: Some preliminary thoughts about the occurrence of ‘radical surprises’ and a case study on the 2002 flood (Germany)
.
Nat. Hazards
,
55
,
671
688
, https://doi.org/10.1007/s11069-010-9645-z.
Kuhlicke
,
C.
,
2013
:
Resilience: A capacity and a myth: Findings from an in-depth case study in disaster management research
.
Nat. Hazards
,
67
,
61
76
, https://doi.org/10.1007/s11069-010-9646-y.
Kuhlicke
,
C.
,
2015
: Vulnerability, ignorance and the experience of radical surprises. Routledge International Handbook of Ignorance Studies, M. Groß and L. McGoey, Eds., Routledge, 239–246.
Kuhlicke
,
C.
,
2019
: Risk and resilience in the management and governance of natural hazards. Oxford Research Encyclopedia on Natural Hazards Science, Oxford University Press, https://doi.org/10.1093/acrefore/9780199389407.013.299.
Kuhlicke
,
C.
,
A.
Scolobig
,
S.
Tapsell
,
A.
Steinführer
, and
B.
De Marchi
,
2011
:
Contextualizing social vulnerability: Findings from case studies across Europe
.
Nat. Hazards
,
58
,
789
810
, https://doi.org/10.1007/s11069-011-9751-6.
Kundzewicz
,
Z. W.
, and Coauthors
,
2014
:
Flood risk and climate change: Global and regional perspectives
.
Hydrol. Sci. J.
,
59
,
1
28
, https://doi.org/10.1080/02626667.2013.857411.
Lamond
,
J.
,
C.
Rose
,
N.
Bhattacharya-Mis
, and
R.
Joseph
,
2018
: Evidence review for property flood resilience phase 2 report. Flood Re and UWE Bristol, 60 pp., https://www.floodre.co.uk/wp-content/uploads/UWE-report_Evidence-review-for-PFR_Phase-2-report.pdf.
Lawrence
,
J.
,
D.
Quade
, and
J.
Becker
,
2014
:
Integrating the effects of flood experience on risk perception with responses to changing climate risk
.
Nat. Hazards
,
74
,
1773
1794
, https://doi.org/10.1007/s11069-014-1288-z.
Masson
,
T.
, and
M.
Barth
,
2019
: Solving the paradox—(Further) evidence for a quadratic relationship between in-group centrality and group-based guilt. Br. J. Soc. Psychol.,
58
,
917
937
, https://doi.org/10.1111/bjso.12310.
Meyer
,
V.
, and Coauthors
,
2013
:
Review article: Assessing the costs of natural hazards—State of the art and knowledge gaps
.
Nat. Hazards Earth Syst. Sci.
,
13
,
1351
1373
, https://doi.org/10.5194/nhess-13-1351-2013.
Miller
,
J. W.
,
W. R.
Stromeyer
, and
M. A.
Schwieterman
,
2013
:
Extensions of the Johnson-Neyman technique to linear models with curvilinear effects: Derivations and analytical tools
.
Multivar. Behav. Res.
,
48
,
267
300
, https://doi.org/10.1080/00273171.2013.763567.
Milly
,
P. C. D.
,
J.
Betancourt
,
M.
Falkenmark
,
R. M.
Hirsch
,
Z. W.
Kundzewicz
,
D. P.
Lettenmaier
, and
R. J.
Stouffer
,
2008
:
Stationarity is dead: Whither water management?
Science
,
319
,
573
574
, https://doi.org/10.1126/science.1151915.
Ogunbode
,
C. A.
,
Y.
Liu
, and
N.
Tausch
,
2017
:
The moderating role of political affiliation in the link between flooding experience and preparedness to reduce energy use
.
Climatic Change
,
145
,
445
458
, https://doi.org/10.1007/s10584-017-2089-7.
Plapp
,
T.
, and
U.
Werner
,
2006
: Understanding risk perception from natural hazards. Examples from Germany. RISK21—Coping with Risks due to Natural Hazards in the 21st Century, W. J. Ammann, S. Dannenmann, and L. Vulliet, Eds., CRC Press, 101–108.
Poussin
,
J. K.
,
W. J.
Wouter Botzen
, and
J. C. J. H.
Aerts
,
2015
:
Effectiveness of flood damage mitigation measures: Empirical evidence from French flood disasters
.
Global Environ. Change
,
31
,
74
84
, https://doi.org/10.1016/j.gloenvcha.2014.12.007.
Reiter
,
J.
,
B.
Wenzel
,
C.
Dittmer
,
D. F.
Lorenz
, and
M.
Voss
,
2018
: The 2013 flood in the community of Elbe-Havel-Land in the eyes of the population. Research Rep. of the Quantitative Survey, KFS Working Paper 08, 44 pp., https://refubium.fu-berlin.de/bitstream/handle/fub188/21920/Working_paper_Elbe_flood_English_final.pdf?sequence=1.
Schröter
,
K.
,
M.
Kunz
,
F.
Elmer
,
B.
Mühr
, and
B.
Merz
,
2015
:
What made the June 2013 flood in Germany an exceptional event? A hydro-meteorological evaluation
.
Hydrol. Earth Syst. Sci.
,
19
,
309
327
, https://doi.org/10.5194/hess-19-309-2015.
Seifert
,
I.
,
W. J. W.
Botzen
,
H.
Kreibich
, and
J. C. J. H.
Aerts
,
2013
:
Influence of flood risk characteristics on flood insurance demand: A comparison between Germany and the Netherlands
.
Nat. Hazards Earth Syst. Sci.
,
13
,
1691
1705
, https://doi.org/10.5194/nhess-13-1691-2013.
Seligman
,
M.
, and
C.
Peterson
,
2001
:
Learned helplessness
. International Encyclopedia of Social & Behavioral Sciences, 1st ed. Science Direct,
8583
8586
.
Sieg
,
T.
,
K.
Vogel
,
B.
Merz
, and
H.
Kreibich
,
2017
:
Tree-based flood damage modeling of companies: Damage processes and model performance
.
Water Resour. Res.
,
53
,
6050
6068
, https://doi.org/10.1002/2017WR020784.
Siegrist
,
M.
,
2013
:
The necessity for longitudinal studies in risk perception research
.
Risk Anal.
,
33
,
50
51
, https://doi.org/10.1111/j.1539-6924.2012.01941.x.
Siegrist
,
M.
,
2014
:
Longitudinal studies on risk research
.
Risk Anal.
,
34
,
1376
1377
, https://doi.org/10.1111/risa.12249.
Spence
,
A.
,
W.
Poortinga
,
C.
Butler
, and
N. F.
Pidgeon
,
2011
:
Perceptions of climate change and willingness to save energy related to flood experience
.
Nat. Climate Change
,
1
,
46
49
, https://doi.org/10.1038/nclimate1059.
Stein
,
D. J.
,
S.
Seedat
,
A.
Iversen
, and
S.
Wessely
,
2007
:
Post-traumatic stress disorder: Medicine and politics
.
Lancet
,
369
,
139
144
, https://doi.org/10.1016/S0140-6736(07)60075-0.
Surminski
,
S.
, and
A. H.
Thieken
,
2017
:
Promoting flood risk reduction: The role of insurance in Germany and England
.
Earth’s Future
,
5
,
979
1001
, https://doi.org/10.1002/2017EF000587.
Tapsell
,
S. M.
, and
S. M.
Tunstall
,
2008
:
“I wish I’d never heard of Banbury”: The relationship between ‘place’ and the health impacts from flooding
.
Health Place
,
14
,
133
154
, https://doi.org/10.1016/j.healthplace.2007.05.006.
Thieken
,
A. H.
,
H.
Kreibich
,
M.
Müller
, and
B.
Merz
,
2007
:
Coping with floods: Preparedness, response and recovery of flood-affected residents in Germany in 2002
.
Hydrol. Sci. J.
,
52
,
1016
1037
, https://doi.org/10.1623/hysj.52.5.1016.
Thieken
,
A. H.
,
S.
Mariani
,
S.
Longfield
, and
W.
Vanneuville
,
2014
:
Preface: Flood resilient communities—Managing the consequences of flooding
.
Nat. Hazards Earth Syst. Sci.
,
14
,
33
39
, https://doi.org/10.5194/nhess-14-33-2014.
Thieken
,
A. H.
,
T.
Bessel
,
S.
Kienzler
,
H.
Kreibich
,
M.
Müller
,
S.
Pisi
, and
K.
Schröter
,
2016a
:
The flood of June 2013 in Germany: How much do we know about its impacts?
Nat. Hazards Earth Syst. Sci.
,
16
,
1519
1540
, https://doi.org/10.5194/nhess-16-1519-2016.
Thieken
,
A. H.
, and Coauthors
,
2016b
:
Review of the flood risk management system in Germany after the major flood in 2013
.
Ecol. Soc.
,
21
,
51
, https://doi.org/10.5751/ES-08547-210251.
Thieken
,
A. H.
,
H.
Kreibich
,
M.
Müller
, and
J.
Lamond
,
2017
: Data collection for a better understanding of what causes flood damage—Experiences with telephone surveys. Flood Damage Survey and Assessment: New Insights from Research and Practice, D. Molinari, S. Menoni, and F. Ballio, Eds., Wiley, 95–106.
van Valkengoed
,
A. M.
, and
L.
Steg
,
2019
:
Meta-analyses of factors motivating climate change adaptation behaviour
.
Nat. Climate Change
,
9
,
158
163
, https://doi.org/10.1038/s41558-018-0371-y.
Wachinger
,
G.
,
O.
Renn
,
C.
Begg
, and
C.
Kuhlicke
,
2013
:
The risk perception paradox: Implications for governance and communication of natural hazards
.
Risk Anal.
,
33
,
1049
1065
, https://doi.org/10.1111/j.1539-6924.2012.01942.x.
Walker
,
B. H.
,
C. S.
Holling
,
S. R.
Carpenter
, and
A.
Kinzig
,
2004
:
Resilience, adaptability and transformability in social-ecological systems
.
Ecol. Soc.
,
9
,
5
, https://doi.org/10.5751/ES-00650-090205.
Walker-Springett
,
K.
,
C.
Butler
, and
W. N.
Adger
,
2017
:
Wellbeing in the aftermath of floods
.
Health Place
,
43
,
66
74
, https://doi.org/10.1016/j.healthplace.2016.11.005.
Weichselgartner
,
J.
, and
I.
Kelman
,
2015
:
Geographies of resilience: Challenges and opportunities of a descriptive concept
.
Prog. Hum. Geogr.
,
39
,
249
267
, https://doi.org/10.1177/0309132513518834.
Zaalberg
,
R.
,
C.
Midden
,
A.
Meijnders
, and
T.
McCalley
,
2009
:
Prevention, adaptation, and threat denial: Flooding experiences in the Netherlands
.
Risk Anal.
,
29
,
1759
1778
, https://doi.org/10.1111/j.1539-6924.2009.01316.x.

Footnotes

© 2019 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).