1. Introduction
Storms and floods are the most frequent and costly weather-related disasters in the United States, accounting for 71.1% of the damage caused by extreme weather events between 1980 and 2011 (Smith and Katz 2013) and an estimated $626.9 billion (U.S. dollars in 2011) in economic losses. Those impacts are expected to increase with climate change due to more frequent intense storms (Grinsted et al. 2013; Holland and Bruyère 2014) and sea level rise (Kopp et al. 2014; IPCC 2013), along with continued development in flood-prone areas (Crossett et al. 2013).
The management of flood risk has long focused on large-scale engineering projects, such as sea walls and levees, designed and implemented by government agencies (Lonnquest et al. 2014). Recently, there has been a shift toward a more integrated approach, which includes flood prevention and damage alleviation through small-scale measures taken by communities and households, such as flood protection devices (e.g., flood vents), adaptive building uses, and flood insurance (Interagency Climate Change Adaptation Task Force 2011; McDaniels et al. 1999; Samuels et al. 2006). The success of these programs depends on local residents’ willingness and ability to undertake those measures. Although individual adaptation measures have demonstrated ability to reduce flood damage (Kreibich et al. 2005; Schanze 2008), relatively few people take them voluntarily (Kunreuther 1996).
A recent review found that individuals’ willingness and ability to take such measures was unrelated to their perceptions of flood risk (Bubeck et al. 2012). The authors speculated that once individuals have adopted some adaptation measures, however effective, they may treat the risks as under control and do no more. The review found less willingness to act among individuals who estimated higher costs for these measures, preferred public flood defense measures, or saw government as responsible (see also Kellens et al. 2013). Conversely, and echoing findings with respect to climate change perceptions (Lee et al. 2015), the review found greater stated willingness to adopt individual adaptation measures among people who viewed them as effective, who knew more about flooding hazards, and who had experienced flooding directly.
Indeed, there is a growing body of evidence that experience, both direct and observed, influences acceptance of climate change (Reser et al. 2014) and the related events of flooding (Taylor et al. 2014) and extreme weather (Capstick and Pidgeon 2014; Howe and Leiserowitz 2013; Lujala et al. 2015). For example, people who have experienced damage that they attribute to climate change see greater future risks (Akerlof et al. 2013), such as flooding and landslides (Lujala et al. 2015). In a national survey of U.K. residents, Taylor et al. (2014) found that self-reported “heat-wave discomfort” was associated with greater acceptance of climate change. Rudman et al. (2013) found greater support for politicians who supported climate change among New Jersey residents after Hurricane Irene and Superstorm Sandy. Before the storms, acceptance of climate change was the best predictor of that support; afterward experience was.
Although studies find consistent results regarding the role of experience in personal adaptation and risk perceptions, the same cannot be said regarding the role of social support, such as having a sympathetic friend to hear problems or family members able to help with transportation or childcare. Some studies find that people with greater social support are more likely to take protective measures before, during, and after disasters (Riad et al. 1999; Kaniasty 2012). For example, Riad et al. (1999) found perceived social support predicted whether individuals evacuated in advance of Hurricane Hugo (1989) and Hurricane Andrew (1992). That support might include having a place to go, receiving needed information, and having someone to talk to (Kaniasty and Norris 1995). Other studies, though, have found that people sometimes take risks when they can treat others as a safety net (Hsee and Weber 1999; Weber and Hsee 1998; Schneider et al. 2014). It is unclear how these processes will balance out with respect to coastal flooding risk.
The present work asks how risk perceptions and social support are related to the adaptation behavior of residents of two New Jersey counties (Monmouth and Ocean) that were devastated by Superstorm Sandy in late October 2012 (Blake et al. 2013). Sandy was the most costly U.S. storm since 1990, other than Hurricane Katrina (Crossett et al. 2013), with a damage estimate of $68 billion (U.S. dollars in 2013) (Sullivan and Uccellini 2013). At the time of our study (summer 2014), these communities were still dealing with the storm’s aftermath. These counties continue to be at risk, as they are predicted to be increasingly vulnerable to future coastal flooding and storm surge risks because of sea level rise, given their location and topography (Strauss et al. 2012; Hauer et al. 2016). We report an initial exploratory study with qualitative open-ended interviews, followed by a structured survey assessing how social support and risk perceptions are related to the adaptation behavior of these individuals, who have experienced coastal flooding directly.
2. Qualitative open-ended interviews
a. Methods
We recruited 14 New Jersey residents from Highlands and Sea Bright, in Monmouth County, and Little Egg Harbor and Tuckerton, in Ocean County (Fig. 1)1 using snowball sampling methods (Goodman 1961). Participants were recruited with the help of New Jersey Future (www.njfuture.org), a nonprofit, nonpartisan organization that brings together residents, community leaders, and public officials to promote responsible land-use policies. New Jersey Future introduced the first author to key informants in the community, who then helped set up interviews with community members. The interviews were conducted in May 2014 and took place at community centers or other locations convenient for participants (e.g., a café) and were conducted by the first author. The interviews lasted approximately an hour and were audio recorded for later transcription. They began with general questions and continued to more specific ones related to decisions facing community members, allowing respondents to direct the flow and express themselves in their own terms (Morgan et al. 2002; Bruine de Bruin and Bostrom 2013). Carnegie Mellon’s Institutional Review Board for the Protection of Human Subjects approved the research protocol. Informed consent was secured from participants, who were not compensated.
Map of communities included in this study.
Citation: Weather, Climate, and Society 9, 2; 10.1175/WCAS-D-16-0042.1
1) Participants
All participants were full-time residents of one of the four communities, with most having lived there at least 20 years (and some all their lives). According to self-reports, their average age was 62.4 years [standard deviation (SD) = 8.3]; 69.2% had at least a college degree; 30.1% worked in public service (government, police, local government), 23.1% in education (high school, college), 23.1% in other professions (speech therapist, building owner), 15.4% in real estate, and 15.4% in service industries; and 41.6% were female.
2) Interview protocol
The interview protocol was informed by nine informal interviews with emergency planning and preparedness experts working in coastal communities in New Jersey and New York, eliciting their perceptions of the issues to address with residents. These interviews provided an informal version of an “expert model” for structuring the interviews around topics potentially relevant to adaptation behavior. The interview protocol was pilot tested with Carnegie Mellon University (CMU) students and resiliency planners familiar with coastal NJ communities. It had three parts, eliciting beliefs about 1) coastal flooding, 2) responsibility for preparing for those risks, and 3) the costs and benefits of possible protective measures.
The interview began with six open-ended questions eliciting respondents’ beliefs about coastal flooding: Please tell me about sea level here in [community]. Tell me what might cause sea level here to become lower/higher in the future than it is today. Tell me what causes coastal flooding in [community]. Tell me about the types of weather events that could result in coastal flooding in [community]. How will sea level affect coastal flooding, when [weather event] happens? What would be different if sea level were lower/higher than it is today? Next, participants rated three statements (1 = completely disagree, 7 = completely agree) about who should be responsible for preparing for the risks, then explained their answers: I believe that state and local government are responsible for helping me prepare for the risk of coastal flooding, I believe that the federal government is responsible for helping me prepare for the risk of coastal flooding, and I believe that preparing for the risk of coastal flooding is entirely my own responsibility. Finally, they were asked what do you think is the best way to respond before a coastal flood due to [weather event]? That weather event was the interviewee’s response to the earlier question tell me about the types of weather events that could result in coastal flooding in [community].
At the end of the interview, participants answered demographic questions.
3) Analytical approach
All interviews were digitally recorded and transcribed. In analyzing them, we read each transcript to identify key themes and developed a master list of codes. Two coders (the first author and a trained CMU undergraduate) independently coded the transcripts into those themes. The results reflect our interpretation of participants’ narratives. We used it to inform the design of our structured survey, including which questions to ask and what language to use. (A complete list of codes and frequency of mentions is in Table A1 in appendix A.) We first report emergent themes regarding the risks [section 2b(1)] and responsibility for managing them [section 2b(2)], then follow with themes related to place and social support [section 2b(3)], and conclude with the measures that participants identified, for short- and long-term protection [section 2b(4)].
b. Results
1) Understanding of coastal flooding
(i) Residents see themselves as at risk for major coastal flooding
Participants distinguished major and minor (nuisance) flooding, reporting that major flooding happens when “perfect storm” factors come together (8 = number of interviewees mentioning the topic), as when nor’easters bring heavy precipitation (9), high winds (northwesterly) cause increased wave action (3), or high tides occur during a full moon (7). Many see themselves as facing risk for coastal flooding (6 s). In the terms of one participant, “You live in a coastal town, so it’s kind of expected that if you can afford to live near the ocean you have to know that there’s certain risks” [participant (p) 3].
(ii) Residents see flooding risk as increasing, which may or may not be due to climate change
Many noted what they have read about sea level rise (10) and climate change (8). When talking about increased risk (8), some cited direct personal experience (6) during time spent on or near the water (e.g., fishing, swimming). One long-time resident (p4) said, “visual appearance says to me that the water is higher than it had been before. It doesn’t get as low as it used to be. So, I’m not saying that it’s always higher, but I think on average we have less beach to look at than we did [before].”
Others invoked aspects of the built environment (5), such as increased building in areas vulnerable to coastal flooding. A few mentioned subsidence (2), poor or aging infrastructure (e.g., sea walls, sewer systems; 1), or erosion (1), whereby sand or dirt underneath homes is sucked out by wave action.I don’t remember the winters being this crazy when I was younger. Not that they didn’t snow, but it just didn’t—I feel like it snows much earlier and much later and much more frequently. It was so cold that we actually had snow on the ground for months, which hasn’t happened in a really long time. So I feel like patterns are definitely changing, and I would imagine that we would have more nor’easters in about 30 years or before then. (p2)
(iii) Residents see dire consequences of flooding
Indeed, it is such psychological trauma that made some residents (3) say that they would consider moving should another Sandy-like storm happen again.We packed our mementos in plastic bins…and when the water came in, it tumbled over the shelving, which tumbled over the tubs and the tubs’ lids fell open…what was worse, I think, is the fact that her father was killed in the war. And, all his mementos were there. So she never knew her father other than what was left by her mother. And, it’s all gone. (p5)
So I think it’s going to bring more income into the area. Sea Bright has always been somewhat of a wealthy town where the people have the resources to rebuild. But some of the areas in town, these are older families that have been there and maybe they cannot afford to stay. But I think money will come in, eventually rebuild the town to where it's more of [an] affluent area…I think it is part of a progression, I think it was the economic stimulus, if you would, has done. If you look at Long Bridge, I do not [want] to compare Sea Bright to Long Bridge, but before Pier Village was in there, that was just a local community. They weren’t really damaged by higher tides, but somebody came in and bought the whole area and now it’s more desirable. It may not be fair that people were forced to leave, but as an economic thing, they added a lot of money to the tax base for the town and there’s a lot of nice housing. (p3)
2) Views on the responsibility for preparing for the risks of coastal flooding
Most participants reported feeling personally responsible for preparing their homes and family for coastal flooding (9), with some explicitly saying that the responsibility comes with their decision to live in a place at risk (7). They saw state and local authorities as sharing responsibility, mentioning tasks such as providing information (6), creating building codes and making sure they are enforced (1), performing local preparedness measures (e.g., building and maintaining major infrastructure; 3), helping citizens (1), and providing shelter for those who need it (1). They saw federal authorities [e.g., Federal Emergency Management Agency (FEMA)] as being responsible for providing resources for recovery but leaving the actual work to local authorities (2). When asked to rate responsibility for preparing on a scale of 1 = not at all responsible to 7 = completely responsible, residents rated themselves as being most responsible [n = 9, mean (M) = 5.55, SD = 1.62], followed by state and local authorities (n = 9, M = 4.82, SD = 0.78) and federal authorities (n = 9, M = 3.59, SD = 1.46).
3) Views on place and social support
But we had to…we had to…move on after three days of crying. You just, I just got up on the fourth day and just said, “Okay, it’s [my possessions are] gone. You have to move on.” You just...and fortunately, our children were very supportive. Our children came right away, they took off from work. My daughter- and son-in-law came up from Maryland. My daughter’s father-in-law came up. And we just started the cleanup. And fortunately, my daughter told me not to go downstairs until they got some of the bad stuff out. And my son-in-law set up a generator. And I kind of fed the neighborhood. I had a grill and I kinda made grilled cheese sandwiches and made soup for the people on the block that didn’t have food and didn’t have electricity. And I kept busy that way until I was able to handle going downstairs. And my son and his boys came on the weekend. And I just, we just started as a family cleaning up and the neighbors helped neighbors, and our block…our block was really wonderful. (p6)
Such support appeared to help residents feel and perhaps be more resilient: “I have a good family structure. I have a good structure of friends. I think this is all very important. There are many people who don’t have that option and I think they would be the ones who would fall through the crack. I think I would be able to recover, personally. Sometimes you have to take life with what it throws at you. It’s the old adage of turning lemons into lemonade” (p9). It may have contributed to their commitment to stay (9) and rebuild at the same location if needed (3) because they love the place (3): “I’m committed to stay here. I see people that keep waders on the porch in case that day happens. I want to be here” (p1). If not able to stay, one resident would “move to another beach community” (p2).
4) Views on ways to prepare for the risks of coastal flooding
Participants distinguished measures aimed at both imminent and long-term threats. For imminent threats, they mentioned diverse things that they could acquire as ways to prepare for imminent threats, including a small boat (3), candles (1), an electric generator (5), gasoline or other fuel (2), lamps (1), radio (1), sandbags (1), wood (1), woodstove (1), safe-boxes for valuables (1), extra cash (for when ATMs, computers, and the Internet go out; 1), and a landline (1). Some described actions such as subscribing to a service that provides current risk information (6), copying important documents (1), creating an emergency evacuation plan (9), putting valuables in higher places at home (8), and moving cars and boats to higher ground (7). For long-term threats, they suggested purchasing flood insurance (1), putting homes on pilings (11), flood proofing (1), and protective landscaping (1). Several mentioned moving away as a way to deal with increasing flood risk (9): “You know, I’d probably move, I think, inland, you know, to an area where I felt that flooding wasn’t going to be an issue” (p14).
c. Discussion
We interviewed 14 long-term residents of coastal communities strongly affected by Sandy. As summarized in Fig. 2, they saw themselves as at risk for coastal flooding, with that risk increasing, which some attributed to climate change and others to changes in the natural or built environment, often citing personal experience. Most saw the risks as posing both tangible threats (financial damages) and intangible ones (social viability of their community). They expressed strong attachment to the place and their lives there, which included responsibility for preparing for the risks. Nearly all pointed to social support as playing an important role in evacuation decisions as well as in their ability to cope with the aftermath of the storm. Most offered practical measures for addressing both imminent and long-term threats. We designed a survey to assess the prevalence of these beliefs and the relationships among them.
Major themes identified during the interviews and how they informed the design of the survey.
Citation: Weather, Climate, and Society 9, 2; 10.1175/WCAS-D-16-0042.1
3. Quantitative approach: Surveys
a. Methods
Participants were again recruited from Highlands, Sea Bright, Little Egg Harbor, and Tuckerton. After discussions with local public officials, community leaders and researchers working in the region, we decided to conduct the survey completely online in order to have the best chance of reaching dislocated residents. Given high Internet penetration in the area, most residents should have access. We posted recruitment advertisements on town homepages and town Facebook pages. Nonetheless, we cannot claim a representative sample.
b. Participants
We recruited 224 residents. According to their self-reports, they were 49.7% female, with an average age of 56.5 (SD = 12.9), 50.0% holding at least a college degree (BA, BS), and 50.0% having an annual household income of at least $76 000 (U.S. dollars). They reported their political affiliation as Independent (33.5%), Republican (31.4%), Democrat (15.4%), or other (2.7%), with 17.0% preferring not to answer. Many (42.9%) reported having at least one person over the age of 64 living at home, a few (19.1%) reporting at least one child (17 years old or younger) living at home. Most reported being Caucasian (90.5%), followed by Native American (2.1%), other (2.1%), and Latino (1.1%), with 2.1% preferring not to answer. The most popular source of information about events or news was the Internet (76.2%), followed by newspapers (63.5%), friends or family (61.9%), radio (37.6%), other media such as television (19.6%), work (18.5%), and school (5.3%). The median time in their current home was 13 years. Nearly all are homeowners (97.0%), with most reporting that their local residence is their primary (69.6%) or secondary home (20.0%).
c. Survey protocol
The survey was pilot tested with Carnegie Mellon University students and resiliency planners at New Jersey Future (www.njfuture.org) familiar with New Jersey shore residents. The design of the survey was informed by the literature and the themes that emerged from the interviews. Figure 2 shows how each such theme is represented in the survey. The survey instrument had seven sections in the following order: personal adaptation behavior, risk perception, tolerance for flooding risk, views about community adaptation, social support, climate change acceptance, and tenure of residence. The survey concluded with demographic questions, including ones related to flooding experience.
The seven sections were as follows:
Personal adaptation. We assessed personal adaptation in two ways: Sandy actions and intentions to act. For the former, respondents first read the following: On October 29, 2012, Sandy made landfall near Brigantine, New Jersey. Think back to the days leading up to Sandy’s landfall, and the days that followed. Think about the flooding that resulted, and what happened to you because of it. Residents then read the following: In 1 or 2 sentences, tell us any ways you would prepare for a future flood that you didn’t do before Sandy. Responses were coded as 0 = no, will not take action, for responses such as no or taking action will not make a difference, and 1 = yes, will take action, if respondents listed any action (e.g., move vehicle, raise home on pilings).
Risk perception. A composite score (Cronbach’s α = 0.86) for risk perception was created by averaging residents’ responses to the following questions (1%–100%): What were the chances of such [Sandy-like] flooding in a typical year, 30 years ago? What do you think the chances are of such flooding happening in the next year? What are the chances of such flooding in a typical year, 30 years from now? What do think the chances are of such flooding happening at least once in the next 30 years?
Tolerance. A composite score [r(209) = 0.62, p < 0.001] for tolerance was created by averaging residents’ responses to the following questions (1%–100%): What would the yearly [Sandy-like] flood risk have to be before you and your family decide to move from this area? What would the chances of a flood over the next 30 years have to be before you and your family decide to move from this area?
Community adaptation. A composite score (Cronbach’s α = 0.88) for feelings about adaptation was created by averaging residents’ agreement with the following statements (1 = very strongly disagree, 7 = very strongly agree): preparing for flooding risk reduces property damage if a flood were to happen, preparing for flooding risk makes the community more resilient, preparing for flooding risk reduces mental health problems if a flood were to happen, preparing for flooding risk makes people want to move away from here, preparing for flooding risk makes people not want to move here, preparing for flooding risk decreases my property value, preparing for flooding risk hurts the local economy, and preparing for flooding risk causes my flood insurance costs to increase. The negative framing of the last five statements reflects how these concepts were expressed in the interviews. They were reversed coded so that higher agreement indicates positive feelings about adaptation, and lower agreement indicates negative feelings.
Social support. Respondents completed an adapted version of Krause’s (2001) social support scale, measuring social networks (family, friends, confidants), received support (emotional, tangible, informational), satisfaction with support, and negative interactions, responding to 14 questions. Example questions include the following: If you were sick in bed, how much could you count on the people around you to help out? Told you what they did in a stressful situation that was similar to the one you were experiencing? The response options were 1 = never, 2 = once in a while, 3 = fairly often, and 4 = very often. We created a mean index of these responses as a measure of social support, with a Cronbach’s alpha of 0.94.
Residence tenure. As an assessment of depth of place attachment, we asked residents in total, how many years have you lived in this community?
Climate change acceptance. We used one question from Maibach et al.’s (2009) Six Americas survey: “Recently you may have noticed that global warming has been getting some attention in the news. Global warming refers to the idea that the world’s average temperature has been increasing over the past 150 years, may be increasing more in the future, and that the world’s climate may change as a result. What do you think? Do you think that global warming is happening?” (1 = yes, 2 = no, 3 = I do not know; Maibach et al. 2009, p. 77).
d. Analytic approach
Statistical analyses were conducted using Stata (version 14; Stata Corp, College Station, Texas). Descriptive statistics were used to characterize respondents’ experience with catastrophic flooding and personal adaptation behavior. They were also used to examine reported risk perception, tolerance for flooding, social support, community adaptation views and tenure of residence. Paired sample t tests were conducted to investigate views on whether residents see the chances of major flooding as increasing over time, comparing past versus present, past versus future, and present versus future. We also compared their reported tolerance for flood with their estimates for the chances of flooding to examine whether residents see the risks as above or below their level of tolerance.
A logistic regression analysis was conducted with personal adaptation behavior as the dependent variable, and risk perception, tolerance, social support, community adaptation views, tenure, and climate change as explanatory variables, controlling for sex, age, income, and education. We also conducted correlational analyses to identify any significant relationships between perceptions of risk, tolerance, views on community adaptation, social support, and climate change perceptions.
e. Results
1) Prevalence of beliefs, attitudes, and behaviors
(i) Experience and personal adaptation behavior
Nearly all respondents reported that they or someone they knew had experienced a flood (91.1%), most often Superstorm Sandy (71.8%) or Hurricane Andrew (7.0%). Few reported that they or someone they know had experienced physical or mental injury (13.1%), while almost all (94.6%) reported that they or someone they know had experienced a financial loss. Most reported that they would take personal adaptation measures to protect themselves and their families (191 out of 219 responses).
(ii) Risk perceptions and tolerance
On balance, respondents saw the annual chances of Sandy-like flooding as having been relatively high (M = 28.1%) in the past and believed those chances to be about the same today [M = 26.8%; Table 1, rows 2–3; t(209) = 0.71, se = 1.64, p = 0.48; Cohen’s d = 0.05]. Compared to the scientific estimates shown in Table 2, these judgments overestimate the probability for the past (10%–20% chance of flooding) but underestimate it for the present (16%–60% chance of flooding). Respondents saw the probability of catastrophic flooding 30 years from now (M = 54.4%; Table 1, line 4) as significantly higher than that of 30 years ago, t(208) = 14.7, se = 1.81, p < 0.001, and d = 1.01, or today, t(212) = 18.3, se = 1.52, p < 0.001, and d = 1.25. Nonetheless, that estimate was still lower than the scientific ones (53%–100%) seen in Table 2.
Descriptive statistics of the survey variables.
Annual flood risk observed or estimated for the past (1985), present (2015) and future (2045). Note that, as shown in Table 2, an increase in annual risk of floods is defined as having 10% or 20% annual risk in 1985. Increases through 2015 come from observed local sea level trends; projected increases through 2045 come from local sea level projections from Kopp et al. (2014). Highlands and Sea Bright values are based on data from and projections at a tide gauge at the Battery in New York City, roughly 32 km away; Tuckerton and Little Egg Harbor values are based on a tide gauge at Atlantic City, roughly 27 km away. Flood height reference is the mean higher high water line defined over the 1983–2001 tidal epoch. Sea level/flood height/risk relationships computed following Tebaldi et al. (2012).
The distributions of these judgments (Figs. 3a–c) show a spike at 50%, which previous research has found to reflect using 50 in the sense of 50–50, to express uncertainty rather than as a numeric value (e.g., Fischhoff and Bruine de Bruin 1999; Bruine de Bruin et al. 2000). Table 1 (line 5) and Fig. 3c show that respondents’ judgments for the probability of at least one flood in the next 30 years (M = 59.1%) was only slightly higher than their judgment of the annual risk at that time. This contrast replicates a familiar research result: the difficulty of estimating how risks compound over time (e.g., Cohen et al. 1971; Shaklee and Fischhoff 1990).
Distribution of estimates for the chances of flooding in the past (30 years ago), present (next year), future (in 30 years), and at least once in the next 30 years.
Citation: Weather, Climate, and Society 9, 2; 10.1175/WCAS-D-16-0042.1
On average, respondents reported that they would tolerate ~58% annual chance of catastrophic flooding and ~66% chance of such flooding at least once in 30 years before deciding to move from the area (Table 2, rows 7 and 8). At the individual level, 85.5% implicitly found the current annual risk tolerable (comparing this judgment to their estimate of current annual risk), and 57.2% found the cumulative risk over the next 30 years tolerable.
Respondents who saw higher flooding risks also reported greater tolerance for it [r(213) = 0.24, p < 0.001].
(iii) Climate change acceptance
Most respondents reported that they accepted that climate change is happening (n = 124), with minorities reporting that they did not know (n = 39) or thought that it was not (n = 33). We found that climate change accepters, on average, give 18.4% higher probability estimates of the risk of flooding than did non-accepters {R2 = 0.11, B = 16.7, t(170) = 3.35, p < 0.001, 95% confidence interval [7.07, 26.35]}, controlling for age, sex, household income, and education.
(iv) Social support and tenure
Respondents see themselves as having strong social support of various forms (Table 2, rows 10–24). On average, they reported having lived in their community for about 20 years (Table 2, row 9).
(v) Judged effectiveness of community adaptation measures
Overall, respondents were unsure about the effects of community adaptation measures, with means falling below the midpoint of the scale eliciting their agreement with statements asserting the effectiveness of the various measures (Table 1, rows 25–33). They expressed the least agreement with the claim that preparing for flooding risk could reduce mental health problems if a flood were to happen (Table 2, row 28). Their median judgment was at the scale midpoint for the claims that preparing for flooding risk reduces property damage if a flood were to happen and makes the community be more resilient (Table 2, row 27).
2) The roles of risk perceptions and social support in adaptation behavior
(i) Predicting personal adaptation behaviors
Given that almost all respondents reported taking some personal adaptation measures, there was little variance to predict. Indeed, we found only one significant predictor, in a logistic regression predicting reported intentions from risk perceptions, risk tolerance, social support, tenure, acceptance of climate change, and beliefs about the effectiveness of community adaptation measures, controlling for basic demographics (Table 3). Respondents were about 2 times more likely to report personal adaptation measures as their reported social support increased by 1 unit (e.g., going from once in a while to fairly often).
Summary of logistic regression analysis for variables predicting personal adaptation behavior (n = 171), controlling for demographics. Note that the controls are sex, age, income, and education (omitted from the table); eB = exponentiated B, and B represents the unstandardized regression coefficient. Climate change predictors (yes, it is happening, no, it is not happening, and I do not know) coded as 1 for yes and 0 for no. Yes, it is happening is the reference category. A Hosmer–Lemeshow chi-square value of 7.12, df = 5, and p value of 0.52 suggests a good fit.
(ii) Correlates of social support
Respondents reporting greater social support also held more positive views about the effectiveness of community adaptation measures [r(211) = 0.16, p = 0.03] and were somewhat more tolerant of flooding risks [r(211) = 0.13, p = 0.06]. Moreover, those with greater social support were also more likely to see community adaptation as resulting in greater community resilience [r(211) = 0.14, p = 0.02], less property damage [r(211) = 0.17, p = 0.02], and fewer mental health problems should flooding occur [r(211) = 0.18, p = 0.01].
4. Discussion and conclusions
In both in-depth interviews and a structured survey of individuals in communities deeply affected by Superstorm Sandy, we found strong awareness of the risk of coastal flooding and the expectation that it would get worse. In the interviews, most respondents cited climate change as a factor. In the survey, most respondents reported believing that climate chance was happening. Those who did not (16.8% of respondents) had appreciably lower perceptions of the risk. However, nonbelievers reported the same tolerance for risk and judged the effectiveness of community adaptation measures similarly.
Compared with scientific estimates (Table 2), survey respondents tended to overestimate the annual probability of flooding 30 years ago, correctly estimate that risk today, and underestimate the risk 30 years hence. They greatly underestimated the probability of experiencing at least one flood in the next 30 years (seen as a certainty by only one-third of respondents; Fig. 3d), capturing the difficulty of estimating cumulative risk.
Nearly all respondents in the interviews reported personal experiences, such as direct observations of rising sea level and changing weather patterns that they interpreted as evidence of increasing risk. For example, “What is causing it? I can’t really say. It could be just a natural cycle. It could be contributed to global—ice melting or whatever. But either way, it is happening…something is different in the last 52 years in this area” (p8). In the survey, one respondent in seven reported that the current annual flooding risk was higher than their tolerable level; about 43% thought that it would be higher than their tolerable level in 30 years. Most reported that the annual chance of flooding would have to be at least 50% before deciding to move (Fig. 4a).
Distribution of estimates of annual chances and chances of at least one flooding event happening in the next 30 years before residents would consider moving.
Citation: Weather, Climate, and Society 9, 2; 10.1175/WCAS-D-16-0042.1
Respondents’ tolerance was marginally higher for those who reported greater social support. It was unrelated to how long they had lived in the area or their acceptance of climate change. Those reporting higher social support also reported taking more personal protective measures and seeing community adaptation measures as more effective. In contrast, responses to the questions about personal and community measures were unrelated to acceptance of climate change or how long they had lived in the community. Thus, people with stronger social support may be more tolerant of flooding risk because of the “cushion” that it provides (Hsee and Weber 1999; Weber and Hsee 1998) but also because of their engagement with personal and community actions. For example, respondents who reported greater social support also rated community protective action as increasing community resilience, while reducing mental health impacts and property damage. These complex roles of social support (less willing to leave, more willing to adapt in place) and attachment to place warrant future study, as do potentially related factors such as social norms and feelings of self-efficacy.
Limitations
One limitation of our results is that the sample is not representative but was recruited from individuals with direct experience with the event through personal contacts, town websites, and web-based newsletters. We did not send surveys by mail as many residents were not living in their homes, which were still being repaired from Sandy damage, or had their primary residence elsewhere. Thus, we obtained a sample of interested and affected parties (Stern and Fineberg 1996). We can only speculate on how the kinds of individuals not represented here might have responded. For example, residents who have left the area, or are not connected to the community sources that we used for recruiting, might report less social support, lower risk tolerance, and less belief in the effectiveness of community action.
A second limitation is our relatively small sample size (n = 224), whose members varied little on some critical measures (e.g., personal action), thereby precluding more complex statistical models (which require variance to predict). A third limitation is that we do not know respondents’ exact location and hence could not calculate a scientific risk estimate for each to compare with their judgments (which was part of our original plan), leaving us with the ranges shown in Table 2. Given the overall patterns observed here, more detailed study of how local conditions affect judgments is warranted (e.g., are they related to local terrain, views of the ocean?).
A final observation is that participants’ reported political ideology was unrelated to any attitude or behavior reported here, despite being been an important predictor of climate change acceptance and, to a lesser extent, of mitigation-related behaviors in other studies (Hornsey et al. 2016). Speculatively, political ideology may matter less when it comes to practical matters of adaptation behavior than it does on principled matters of mitigation policies (Wong-Parodi and Fischhoff 2015).
Our findings suggest that local residents are aware of the risk of flooding and expect it to get worse, although without fully appreciating how bad it might be. They feel responsibility for preparing for that risk but are uncertain about the value of most community protective measures. Social support appears central to their thinking, leading some residents to take measures because those are the right things to do but without confidence in their efficacy. During the interviews, participants who had little difficulty listing things that they could do often wondered about the options for people with fewer financial or social resources [“people who don’t have as much as other people” (p9)] and other constraints (elderly, nonnative English speakers, one-story homes, etc.). Providing all residents with practical options (Margolis and McCabe 2003) is a challenge for planners and officials (Armitage et al. 2011). Those efforts might consider the role of social processes in facilitating in individual and community actions—both to reduce the risk of flooding and to speed recovery from it in the face of increasing risk due to climate change.
Acknowledgments
This work was supported by a research grant from the Connecticut Sea Grant Program (R/CSAP-9-CT) and the Center for Climate and Energy Decision Making (SES-0949710).
APPENDIX A
Codebook
The counts and frequency of mentions per code TA1 identified in the interviews are listed in Table A1.
Shows counts and frequency of mentions per code identified in the interviews.
APPENDIX B
Summary Correlations
A summary of correlations between community adaptation measures, flooding risk, and risk tolerance is presented in Table B1.
Summary of correlations between community adaptation beliefs, flooding risk, and risk tolerance.
REFERENCES
Akerlof, K., E. W. Maibach, D. Fitzgerald, A. Y. Cedeno, and A. Neuman, 2013: Do people “personally experience” global warming, and if so how, and does it matter? Global Environ. Change, 23, 81–91, doi:10.1016/j.gloenvcha.2012.07.006.
Armitage, D., F. Berkes, A. Dale, E. Kocho-Schellenberg, and E. Patton, 2011: Co-management and the co-production of knowledge: Learning to adapt in Canada’s Arctic. Global Environ. Change, 21, 995–1004, doi:10.1016/j.gloenvcha.2011.04.006.
Blake, E. S., T. B. Kimberlain, R. J. Berg, J. P. Cangialosi, and J. L. Beven II, 2013: Tropical cycne report: Hurricane Sandy, 22–29 October 2012. National Hurricane Center Rep. AL182012, 157 pp. [Available online at http://www.nhc.noaa.gov/data/tcr/AL182012_Sandy.pdf.]
Bruine de Bruin, W., and A. Bostrom, 2013: Assessing what to address in science communication. Proc. Natl. Acad. Sci. USA, 110, 14 062–14 068, doi:10.1073/pnas.1212729110.
Bruine de Bruin, W., B. Fischhoff, S. G. Millstein, and B. L. Halpern-Felsher, 2000: Verbal and numerical expressions of probability: “It's a fifty–fifty chance.” Organ. Behav. Hum. Decis. Processes, 81, 115–131, doi:10.1006/obhd.1999.2868.
Bubeck, P., W. J. W. Botzen, and J. C. J. H. Aerts, 2012: A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Anal., 32, 1481–1495, doi:10.1111/j.1539-6924.2011.01783.x.
Capstick, S. B., and N. F. Pidgeon, 2014: Public perception of cold weather events as evidence for and against climate change. Climatic Change, 122, 695–708, doi:10.1007/s10584-013-1003-1.
Cohen, J., E. I. Chesnick, and D. Haran, 1971: Evaluation of compound probabilities in sequential choice. Nature, 232, 414–416, doi:10.1038/232414a0.
Crossett, K., B. Ache, P. Pacheco, and K. Haber, 2013: National coastal population report: Population trends from 1970 to 2020. NOAA State of the Coast Rep. Series, U.S. Department of Commerce, 22 pp.
Fischhoff, B., and W. Bruine De Bruin, 1999: Fifty–Fifty=50%? J. Behav. Decis. Making, 12, 149–163, doi:10.1002/(SICI)1099-0771(199906)12:2<149::AID-BDM314>3.0.CO;2-J.
Goodman, L., 1961: Snowball sampling. Ann. Math. Stat., 32, 148–170, doi:10.1214/aoms/1177705148.
Grinsted, A., J. C. Moore, and S. Jevrejeva, 2013: Projected Atlantic hurricane surge threat from rising temperatures. Proc. Natl. Acad. Sci. USA, 110, 5369–5373, doi:10.1073/pnas.1209980110.
Hauer, M. E., J. M. Evans, and D. R. Mishra, 2016: Millions projected to be at risk from sea-level rise in the continental United States. Nat. Climate Change, 6, 691–695, doi:10.1038/nclimate2961.
Holland, G., and C. L. Bruyère, 2014: Recent intense hurricane response to global climate change. Climate Dyn., 42, 617–627, doi:10.1007/s00382-013-1713-0.
Hornsey, M. J., E. A. Harris, P. G. Bain, and K. S. Fielding, 2016: Meta-analyses of the determinants and outcomes of belief in climate change. Nat. Climate Change, 6, 622–626, doi:10.1038/nclimate2943.
Howe, P. D., and A. Leiserowitz, 2013: Who remembers a hot summer or a cold winter? The asymmetric effect of beliefs about global warming on perceptions of local climate conditions in the U.S. Global Environ. Change, 23, 1488–1500, doi:10.1016/j.gloenvcha.2013.09.014.
Hsee, C. K., and E. U. Weber, 1999: Cross-national differences in risk preference and lay predictions. J. Behav. Decis. Making, 12, 165–179, doi:10.1002/(SICI)1099-0771(199906)12:2<165::AID-BDM316>3.0.CO;2-N.
Interagency Climate Change Adaptation Task Force, 2011: Federal actions for a climate resilient nation: Progress report of the Interagency Climate Change Adaptation Task Force. Interagency Climate Change Adaptation Task Force, 32 pp. [Available online at https://www.whitehouse.gov/sites/default/files/microsites/ceq/2011_adaptation_progress_report.pdf.]
IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp.
Kaniasty, K., 2012: Predicting social psychological well-being following trauma: The role of postdisaster social support. Psychol. Trauma, 4, 22–23, doi:10.1037/a0021412.
Kaniasty, K., and F. H. Norris, 1995: Mobilization and deterioration of social support following natural disasters. Curr. Dir. Psychol. Sci., 4, 94–98, doi:10.1111/1467-8721.ep10772341.
Kellens, W., T. Terpstra, and P. De Maeyer, 2013: Perception and communication of flood risks: A systematic review of empirical research. Risk Anal., 33, 24–49, doi:10.1111/j.1539-6924.2012.01844.x.
Kopp, R. E., R. M. Horton, C. M. Little, J. X. Mitrovica, M. Oppenheimer, D. J. Rasmussen, B. H. Strauss, and C. Tebaldi, 2014: Probabilistic 21st and 22nd century sea-level projections at a global network of tide-gauge sites. Earth’s Future, 2, 383–406, doi:10.1002/2014EF000239.
Krause, N., 2001: Social support. Handbook of Aging and the Social Sciences, 5th ed., R. H. Binstock and L. K. George, Eds.Academic Press, 272–294.
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, doi:10.5194/nhess-5-117-2005.
Kunreuther, H., 1996: Mitigating disaster losses through insurance. J. Risk Uncertainty, 12, 171–187, doi:10.1007/BF00055792.
Lee, T. M., E. M. Markowitz, P. D. Howe, C. Y. Ko, and A. A. Leiserowitz, 2015: Predictors of public climate change awareness and risk perception around the world. Nat. Climate Change, 5, 1014–1020, doi:10.1038/nclimate2728.
Lonnquest, J., B. Toussaint, J. Manous Jr., and M. Ertsen, Eds., 2014: Two centuries of experience in water resources management: A Dutch-U.S. retrospective. Institute for Water Resources, U.S. Army Corps of Engineers, and Rijkswaterstaat, Ministry of Infrastructure and the Environment, 377 pp.
Lujala, P., H. Lein, and J. K. Rød, 2015: Climate change, natural hazards, and risk perception: The role of proximity and personal experience. Local Environ., 20, 489–509, doi:10.1080/13549839.2014.887666.
Maibach, E., C. Roser-Renouf, and A. Leiserowitz, 2009: Global warming’s six Americas 2009: An audience segmentation analysis. Yale Project on Climate Change and the George Mason University Center for Climate Change Communication Rep., 140 pp.
Margolis, H., and P. P. McCabe, 2003: Self-efficacy: A key to improving the motivation of struggling learners. Prev. Sch. Failure Altern. Educ. Child. Youth, 47, 162–169, doi:10.1080/10459880309603362.
McDaniels, T. L., R. S. Gregory, and D. Fields, 1999: Democratizing risk management: Successful public involvement in local water management decisions. Risk Anal., 19, 497–510, doi:10.1111/j.1539-6924.1999.tb00424.x.
Morgan, M. G., B. Fischhoff, A. Bostrom, and C. Atman, 2001: Risk Communication: A Mental Models Approach. Cambridge University Press, 351 pp.
Reser, J. P., G. L. Bradley, and M. C. Ellul, 2014: Encountering climate change: ‘Seeing’ is more than ‘believing.’ Wiley Interdiscip. Rev.: Climate Change, 5, 521–537, doi:10.1002/wcc.286.
Riad, J. K., F. H. Norris, and R. B. Ruback, 1999: Predicting evacuation in two major disasters: Risk perception, social influence, and access to resources. J. Appl. Soc. Psychol., 29, 918–934, doi:10.1111/j.1559-1816.1999.tb00132.x.
Rudman, L. A., M. C. McLean, and M. Bunzl, 2013: When truth is personally inconvenient, attitudes change the impact of extreme weather on implicit support for green politicians and explicit climate-change beliefs. Psychol. Sci., 24, 2290–2296, doi:10.1177/0956797613492775.
Samuels, P., F. Klijn, and J. Dijkman, 2006: An analysis of the current practice of policies on river flood risk management in different countries. Irrig. Drain., 55, S141–S150, doi:10.1002/ird.257.
Schanze, J., 2008: Systematisation, evaluation and context conditions of structural and non-structural measures for flood risk reduction: FLOOD-ERA Joint Report. CRUE Research Rep. I-1, 207 pp.
Schneider, C. R., D. D. Fehrenbacher, and E. U. Weber, 2014: Catch me if I fall: Cross-cultural differences in willingness to take financial risks as a function of social and state ‘cushioning.’ Cross-National Data Center in Luxembourg, LWS Working Paper Series 16, 37 pp. [Available online at http://www.lisdatacenter.org/wps/lwswps/16.pdf.]
Shaklee, H., and B. Fischhoff, 1990: The psychology of contraceptive surprises: Cumulative risk and contraceptive effectiveness. J. Appl. Psychol., 20, 385–403, doi:10.1111/j.1559-1816.1990.tb00418.x.
Smith, A., and R. Katz, 2013: US billion-dollar weather and climate disasters: Data sources, trends, accuracy and biases. Nat. Hazards, 67, 387–410, doi:10.1007/s11069-013-0566-5.
Stern, P. C., and H. V. Fineberg, Eds., 1996: Understanding Risk: Informing Decisions in a Democratic Society. National Academy Press, 249 pp.
Strauss, B. H., R. Ziemlinski, J. L. Weiss, and J. T. Overpeck, 2012: Tidally adjusted estimates of topographic vulnerability to sea level rise and flooding for the contiguous United States. Environ. Res. Lett., 7, 014033, doi:10.1088/1748-9326/7/1/014033.
Sullivan, K. D., and L. Uccellini, 2013: Service assessment: Hurricane/post-tropical cyclone Sandy, October 22–29, 2012. U.S. Department of Commerce, NOAA, and NWS, 66 pp. [Available online at http://www.weather.gov/media/publications/assessments/Sandy13.pdf.]
Taylor, A., W. Bruine de Bruin, and S. Dessai, 2014: Climate change beliefs and perceptions of weather‐related changes in the United Kingdom. Risk Anal., 34, 1995–2004, doi:10.1111/risa.12234.
Tebaldi, C., B. H. Strauss, and C. E. Zervas, 2012: Modelling sea level rise impacts on storm surges along US coasts. Environ. Res. Lett., 7, 014032, doi:10.1088/1748-9326/7/1/014032.
U.S. Census Bureau, 2014: American Community Survey (ACS). Accessed 31 March 2016. [Available online at https://www.census.gov/programs-surveys/acs/.]
Weber, E. U., and C. Hsee, 1998: Cross-cultural differences in risk perception, but crosscultural similarities in attitudes towards perceived risk. Manage. Sci., 44, 1205–1217, doi:10.1287/mnsc.44.9.1205.
Wong-Parodi, G., and B. Fischhoff, 2015: The impacts of political cues and practical information on climate change decisions. Environ. Res. Lett., 10, 034004, doi:10.1088/1748-9326/10/3/034004.
Highlands, New Jersey, is a coastal community of approximately 5000 fulltime and seasonal residents, who are mostly white (93%), with a median household income of approximately $75 000 (U.S. Census Bureau 2014). Sea Bright, New Jersey, is a smaller coastal community adjacent to Highlands, with approximately 1400 people, who are mostly white (95%) with a median household income of about $74 000. Little Egg Harbor, New Jersey, is the largest coastal community in our study, with approximately 20 000 residents, who are mostly white (94%) with a median household income of about $59 000. Tuckerton, New Jersey, is adjacent to Little Egg Harbor and has approximately 3400 residents, who are mostly white (94%) with a median household income of about $53 000.
