Rising Tides and Changing Attitudes: Community Adaptive Planning and Behavior Change in North Kingstown, Rhode Island

The consequences of global climate change in the form of sea level rise and more frequent intense storms are likely to cause significant impacts on coastal ecosystems and critical infrastructure in vulnerable coastal municipalities. This could result in major economic losses and social disruption unless these communities proactively plan for the impacts of a changing climate. As a small state with a large coastal population, Rhode Island is highly vulnerable to impacts from climate change, thus the state is positioned to act as a potential leader in a national movement towards proactive adaptation. It is clear that actions must be taken, however implementing effective policy changes requires significant political will as well as the support of decision makers and communities. This case study assesses municipal officials’ perceptions of the risks sea level rise and increased storminess pose to North Kingstown, Rhode Island and analyzes the relationship between town decision makers’ understanding of climate change risk and adaptive planning behavior. Evaluation of local decision makers’ mental models concerning climate change knowledge and perceptions of risk will provide insights for those working with town decision makers to proactively incorporate adaptation actions in town comprehensive and capital improvement planning. Results	  of	  this	  research	  display	  no	  correlation	  between	  local	  decision	  makers’ levels	  of	  climate	  change	  knowledge	  and	  their	  individual	  preparatory	  behavior	  or between	  personal	  adaptive	  behavior	  and	  levels	  of	  support	  for	  proactive municipal	  adaptation.	  A	  strong	  correlation	  was	  found	  between	  individuals with mental models closely matching the expert model of climate change knowledge and levels of support for municipal adaptation planning and actions.	  Additionally,	  this study	  found	  a	  moderate correlation between subjects’ exposure	  to	  climate	  change information and levels of support for municipal adaptation.	  Increasing awareness of the risks associated with impacts of climate change through improved communication, educational programs, and public outreach is likely to be an effective way of promoting proactive adaptation in vulnerable coastal communities.


Introduction
The consequences of global climate change in the form of sea level rise and more frequent intense storms have the potential to cause significant impacts on vulnerable coastal ecosystems and critical infrastructure in coastal communities around the world (Alley et al., 2007;Ashton, Donnelly, and Evans, 2008;Bender et al., 2010;Douglas, 2001). On October 29, 2012, the devastating power of increasingly severe storms  (Titus et al., 2009;Field et al., 2007;Frumhoff et al., 2007).
As a small state with a large coastal population, Rhode Island is highly vulnerable to impacts from climate change. This gives the state a leadership opportunity in a national movement towards proactive climate change adaptation (Frumhoff et al., 2007). The age of Rhode Island's many historic coastal communities means that sea level rise is already impacting some infrastructure and will become an even greater concern as the rate of sea level rise accelerates. It is clear that actions must be taken.
However, implementing state-wide and municipal policy changes requires significant political will as well as the capacity to change the behaviors of individuals and communities. This case study focusing on the coastal community of North Kingstown, Rhode Island, provides insights into the current behaviors and attitudes of municipal decision-makers and suggests methods of increasing proactive climate change adaptation behaviors and actions.

Objectives of study
Damages incurred from sea level rise and intense storms will cause significant economic losses in local communities unless these communities plan to mitigate the impacts of global climate change (Field et al., 2007;Douglas, 2001;Titus et al., 2009). In order to start the process of planning for climate change impacts, municipal officials need to make a shift in their beliefs and attitudes and engage in adaptive behavior change (Doppelt, 2008). As decision makers' beliefs and behaviors related to climate change adaptation shift, municipalities can more effectively plan and implement adaptive actions to minimize damages from the impacts of climate change.
The objective of this study is to assess local town officials' perceptions of the risks that sea level rise and increased storminess pose to the town of North Kingstown, Rhode Island, and analyze the relationship between town decision makers' perceptions of climate change risk and community adaptive planning behavior. Evaluation of decision makers' knowledge of climate change and behavior related to adaptation and preparation will provide insights for coastal managers and policy makers working to incorporate adaptive actions in town comprehensive and capital improvement planning.

Research questions
This Rhode Island case study focuses on five core research questions: (1)

Research hypothesis
The research hypotheses for this study are: (1) Individuals whose mental models most closely fit the expert model of climate change and risk will be at a more advanced stage in the five-stage model of personal adaptation behavior change; (2) Individuals whose mental models most closely fit the expert model of climate change knowledge will be more supportive of municipal adaptation planning and actions; (3) Individuals who have attended educational programs or seminars presenting information on climate change impacts and risks will be more supportive of municipal adaptation planning and actions; and (4) Individuals who are at a more advanced stage in the five-stage model of personal adaptation behavior change will be more supportive of municipal adaptation planning and actions.  (Stake, 1995;Yin, 1994 and Capital Improvement Plans, however they must overcome numerous challenges and obstacles before definitive proactive adaptation actions may be implemented (Reiner, 2012).

Climate Change in Rhode Island
Global climate change presents numerous challenges to Rhode Island's coastal communities, including accelerating sea level rise, increased storminess, and changing ocean conditions (Alley, 2007;Rhode Island Coastal Resources Management Council, 2009;Douglas, 2001;Smith et al., 2010). Inundation of coastal areas caused by increased sea levels and storm surge resulting from more intense storms is a major threat to private and public buildings as well as important infrastructure such as waste water management facilities, power substations, transportation networks, wetlands, agricultural lands, and historic and cultural sites (Douglas, 2001;Field, 2007).
Damages incurred from sea level rise and storm surge inundation will cost local communities significant economic losses unless these communities plan for the impacts of global climate change expected in their area (Field et al., 2007;Douglas, 2001;Titus et al., 2009).
The latest report by the Intergovernmental Panel on Climate Change (IPCC) predicts that global mean sea level will rise between 0.18 and 0.49 meters by 2100, however the regional expected sea level rise may vary depending on the particular circumstances affecting a specific location (Alley, 2007;Sallenger, 2012). The IPCC estimates are also arguably conservative as they do not include the uncertain contributions from melting of the Greenland and Antarctic ice sheets (Bamber et al., 2009). Semi-empirical models of sea level rise that include contributions from melting ice sheets predict increases of up to 1.4 meters in global sea level by 2100 (Rahmstorf et al., 2007).
Although predictions of future global sea level vary depending on numerous factors, there is unequivocal evidence that indicates that increases in "global average air and ocean temperatures, widespread melting of snow and ice, and rising global average sea level" will impact vulnerable coastal ecosystems and human infrastructure (Gidley et al., 2009, p. 1 (Farmer, 2012;Natural Hazards Observer, 2012;Smith, 2013). These three major storms occurring in less than three years provide a possible preview of the type of extreme weather events that many climate models indicate will become more frequent due to global climate change (Seelye, 2010;Cooper, 2011). In order to minimize the risks and costs associated with sea level rise and increasing storm intensity, local communities must begin the process of proactively adapting to the impacts of global climate change through individual and municipal behavior change.

The Transtheoretical Model of Behavior Change
The Transtheoretical Model of behavior change (TTM) theorizes that behavior change must progress through five stages of change before resulting in a permanent and lasting change in behavior (Prochaska et al., 1997). This research tests whether the five-stage model can be applied to a climate change adaptation scenario, thereby allowing researchers and planners to assess the level of individuals' preparedness for climate change adaptation. Furthermore, this research tests whether an individual decision maker's level of preparedness at home is correlated to his or her support for community-wide adaptation measures. This information can provide insights to coastal managers and decision makers resulting in the development of more effective outreach and educational materials.
The TTM is based on the theory that change is a process encompassing many steps, not a single event. This model identifies five essential stages that individuals must pass through in order to make a permanent behavior change: (1) Precontemplation, (2) Contemplation, (3) Preparation, (4) Action, and (5) Maintenance. The TTM then uses these stages to develop a stage-based intervention designed to move individuals through the five stages of change, a method that has proven to have a 10 to 15 times greater impact than traditional behavior change models in studies focused on changing health behaviors (Prochaska, 2008). This model is widely used as a leading approach to changing nearly 50 types of risky behaviors in the field of public health and has been used successfully to address smoking cessation and changes in diet, exercise, and medication compliance (Prochaska and DiClemente, 1983). This stage-based approach to changing behavior is frequently used in the field of healthcare and has provided exceptional results for the last twenty-five years, however there is increasing interest in the application of this methodology to the area of environmental and climate studies. Recently the TTM has been identified as a potential tool for implementing change related to environmental and climate change behaviors (Semenza et al., 2008;Gertner, 2009;Doppelt, 2008;Doppelt, 2010); however little work has been done yet on these topics. Thus, this thesis will test the appropriateness of using the TTM for understanding adaptation to climate change.
A recent study conducted by the University of Rhode Island's Cancer Prevention Research Center expanded the use of the TTM model beyond health behaviors, utilizing TTM methods to encourage climate change mitigation behaviors such as driving less, biking to work, and recycling (Mundorf et al., 2013). Although the TTM was developed as a method of changing health behaviors, this thesis hypothesizes that its stage-based approach would have the potential to assist coastal managers and municipal decision makers in evaluating current climate change adaptation behaviors and implementing efforts to promote proactive adaptation behavior changes.
Climate change is often an overwhelming subject for managers and planners to understand and begin adaptation planning for, but the wide success seen through TTM methods in changing high risk health behaviors suggested these same methods might be utilized as a powerful model in understanding and attempting to alter beliefs and behavior related to the risks posed by climate change (Doppelt, 2008). There are numerous challenges in applying health behavior-based methods to climate change adaptation, such as difficulty in identifying concrete and achievable adaptation actions that individuals can take, and uncertainty in gauging when individuals have reached adaptation objectives. However, the TTM's fundamental building block, the stagebased approach to change, provides coastal managers, policy makers, and educators with a new model and tool for altering thinking about climate change risks and motivating proactive adaptation actions among municipal decision makers (Doppelt, 2008).

Mental Model Analysis
Analyzing decision makers' mental models and perceptions of their communities' risks from sea level rise and increased storm intensity can provide important insights for guiding adaptation planning and regulation implementation (see, Mozumder et al., 2011;Lowe & Lorenzoni, 2007). Mental models provide insight into an individual's internal understanding and perception of external problems or phenomenon, such as sea level rise and climate change. They are vitally important because they influence the way individuals make decisions and resolve problems (Genter & Stevens, 1983;Jones et al., 2011). Understanding audiences' mental models is thus critical to developing more effective communication and decision-making pertaining to risk mitigation and proactive adaptation planning (Morgan et al., 2002;Steelman and McCaffrey, 2013).
An additional consideration pertaining to the mental models analysis of North Kingstown town officials is that individuals participating in team decision-making function more effectively as a team and have better team processes when they share task-based mental models (Cannon--Bowers et al., 1990;Mathieu et al., 2000). Mental model analysis has previously been used in numerous studies conceptualizing how people understand coastal management problems, ocean and coastal processes, wildfire mitigation, and the risks of sea level rise and storm surge (Kempton and Falk, 2000;Thompson, 2005Thompson, , 2007Stocker and Kennedy, 2009;Marcucci et al., 2012;Champ et al., 2012;Hulst, 2012). Analysis of the mental models of North Kingstown town decision makers will build on these previous studies and will be useful to foster convergence in their understanding of climate change science and their approach to adaptation, leading to more effective team decision-making.

Selection of Research Subjects
Research subjects for this case study were selected from the pool of municipal officials, town employees, town council members, and numerous board members in North Kingstown. "Purposive sampling" in which interview subjects are chosen based on the purpose of the research is a good method of sampling for intensive and critical case studies as well as when focusing on a specific population, such as decisionmakers in a particular municipality (Bernard, 2011 Research subjects were initially identified by reviewing public listings of board and council members on the municipal government website. I contacted potential subjects by email and by phone, introduced and explained my general field of inquiry, and invited them to participate in the research through an interview.

The Expert Model
The "expert model" in this case study consists of the knowledge base against which all subject mental models were compared. Comparison between the expert model and subjects' mental models demonstrates the municipal decision-makers level of climate change knowledge. The study also determined the disparity between experts' perceptions of North Kingstown's risk of climate change impacts and town officials' perceptions, as this discrepancy is an important factor to consider in developing and implementing proactive adaptation planning on a municipal level. The mental models methodology used by Morgan et al. (2002) contains four distinct steps: "(1) developing an "expert model"; (2) conducting mental model interviews; (3) coding and analyzing the interviews; and (4) evaluating the differences and gaps between the expert model and the mental models of the interview subjects" (Smythe, 2011 This case study utilized the general mental model methodology outlined by Morgan et al. (2002), however I also used an innovative manner of coding responses using numerical values to represent expert and subject answers. The experts' answers to each question were coded and each answer was given a numeric score, then the "expert model" value of each response was determined based on the mean value of the five experts' responses. For example, the experts were asked how likely they thought it was that the town of North Kingstown would experience any impacts from future climate change on a scale of 1 to 10 with 1 representing "highly unlikely" and 10 representing "highly likely." Each expert then answered by identifying a number within the 1 to 10 range and the mean of the five responses was recorded as the expert model value of that specific question. During the interviews the experts were also asked numerous open-ended questions, such as "What risks are there to North Kingstown from extreme weather?" and their responses were incorporated in the development of the final survey instrument used in conducting subject interviews.
The finalized expert mental model consisted of the numerical sum of the mean score of all the questions asked during the coastal management experts' interviews.
This quantitative score describes the knowledge base and mental model of experts pertaining to the topic of climate change and coastal risks and provides a base for comparison with the mental models of municipal decision makers.

Interviews
Data for this research was collected through interviews conducted with municipal decision-makers in North Kingstown using a semi-structured interview script consisting of a combination of closed-and open-ended questions (Bernard, 2011). I chose to use a semi-structured interview format due to time limitations, which necessitated only meeting once with each subject, and because the case study consisted of interviews with professionals and municipal officials accustomed to efficient use of time in meetings (Bernard, 2011). The mix of closed-and open-ended questions was used in order to obtain subjects' responses to nearly identical questions while retaining aspects of a mental models survey approach (Bernard, 2011;Morgan et. al., 2002). The semi-structured interview format used open-ended questions to determine interviewees' current beliefs regarding climate change and adaptation in general combined with closed-ended questions to ascertain interviewees' understanding of the likely impacts and risks posed by climate change. Open-ended questions designed to elicit subjects' initial reactions and perceptions were asked first, followed by multiple choice and closed-ended questions to determine subjects' responses to specific queries. Prior to interviewing subjects, academic advisors and experts in the field of coastal zone management reviewed the interview questions and suggested changes and improvements, resulting in alterations to improve the clarity and flow of the questions. Review of the interview instrument by experienced researchers and experts in the field prior to subject interviews served as a form of pretesting and was vital to ensuring the interview instrument used clear, concise wording and format and did not contain questions that were too broad or misleading (Bernard, 2011). Interviews with subjects were conducted between June and September 2012 at a convenient location of his or her choosing, usually a home, office, or nearby coffee shop. The duration of interviews varied from approximately half an hour to over an hour. I conducted all interviews in person and recorded all interviews for later transcription.
The interview instrument used for data collection followed a "funnel design" beginning with general questions concerning coastal zone management issues and narrowing to focus on climate change impacts (see Appendix 1) (Bernard, 2002). This manner of interviewing developed by Morgan et al. (2002) for use in risk analysis allows the researcher to determine the subject's understanding and perception of the general topic of inquiry without influencing the interviewee's answers, thus determining what is most important to the interviewee. The semi-structured design of the interview instrument started with broad questions to assess the subjects' level of agreement with statements regarding the severity and likelihood of local impacts from sea level rise and increased storm intensity (Mozumder et al., 2011). Subjects were then asked through structured questions how concerned they are about local risks posed by these climate change impacts, narrowing the interview to address the specific topic of local climate change risk. Following these questions, subjects were asked if their town, committee, or board is considering adaptive planning in their decisionmaking. For the purposes of this study I defined "adaptive planning" to mean plans and preparatory actions developed and/or implemented by town decision makers in order to minimize or reduce damages from projected future sea level rise and increased storminess (Lausche, 2009). Depending on the subjects' answers to initial questions, they were asked follow-up questions pertaining to why their decisionmaking body was not considering adaptive planning and what impediments or obstacles they face to including sea level rise adaptation in their future planning.
During the interviews I recorded whether responses were prompted or unprompted, which contributed to the development of each subjects' mental model score during the analysis stage.
Interviewees' mental models were explored through the interview in order to determine their perceptions and understanding of sea level rise and climate change risks in their local area (Jones et al., 2011;Morgan et al., 2002). Initial questions in the interview instrument were purposefully broad and open-ended in order to elicit the interviewees' most important thoughts and responses to the subject matter, per the methods used to elicit a subject's mental model (Morgan et al., 2002). These initial interview questions (Questions #1-14, see Appendix 1) were used in the development of the Municipal Adaptation Level and Individual Behavior Change metrics, described in detail below. Later questions focusing more specifically on climate change allowed me to determine the interviewees' mental models more precisely as well as gauge their knowledge and understanding of current climate change science and projections of future risks and impacts. Using mental model methods, I also determined what topics were most important and significant to interview subjects as opposed to simply what information they knew or were familiar with. The responses from subjects obtained through individual interviews were collected and analyzed, as will be explained in detail below, in order to determine the mental models and climate change knowledge of North Kingstown's municipal decision-makers. This data was then compared through statistical analysis of quantitative scores to the "expert model" of knowledge pertaining to projected climate change impacts, coastal risk from sea level rise and storm surge, and current adaptation projects in North Kingstown.

Data Preparation
All interviews were recorded and transcribed in preparation for analysis of the Municipal Adaptation, Education, and State Involvement Scores, described fully below, qualitative open-ended questions were transcribed and organized for coding.
These five metric scores obtained from each subject's interview responses were then coded and analyzed in order to answer each of the four primary research questions.

Metric Development and Scoring
The first step in the analysis of this case study's collected data was the development of five separate metrics designed to represent each subject's behaviors, 3. If you have taken previous actions, are there any follow-up actions you will make in the future?
The coding format for this metric followed the methodology of the Transtheoretical Model of behavior change in which each subject was identified as being in one of the five stages of behavior change based on their responses: (1) Pre-contemplation, (2) Contemplation, (3) Preparation, (4) Action, and (5) Maintenance (Prochaska and Velicer, 1997). IBC responses were assigned points according to this predetermined scoring rubric.

Stage of Behavior Change
Subject Response 1 Pre-contemplation Has not previously considered taking measures to prevent damage to home or property from storms. 2 Contemplation Has considered taking measures to prevent damage to home or property from storms but has not started to take action yet. 3 Preparation Has decided on what measures they will take to prevent damage to their home or property from storms; is conducting research or consulting professionals but has not taken definitive actions yet. 4 Action Has taken definitive actions (such as trimming trees, re-grading yard, installing a sump pump) to prevent damage to their home or property from storms. 5 Maintenance Has previously taken definitive actions to prevent damage to their home and property from storms and is actively maintaining their actions (such as continued tree trimming) while planning for future protective actions.
Each subject's IBC score was recorded in conjuncture with their identifying number so that a subject's scores on each of the five metrics could be compared. The mean score of the fifteen subjects was also determined for comparison with the experts' mean score on the same Individual Behavior Change metric.

Development of the Mental Model Score
In order to determine the subjects' quantitative Mental Model score for comparison with the Expert Model, responses to ten interview questions (Questions Unprompted answers that agreed with the Expert Model were each given 1 point and prompted answers received 0.5 points. Several subjects responded that specific risks "might" be a problem for North Kingstown and these "maybe" answers were given 0.5 points if unprompted and 0.25 points if prompted.
Three questions asking subjects about the likelihood that North Kingstown would experience impacts from climate change in general, and sea level rise and increasing storm intensity specifically, were designed on a ten-point scale with 1 representing "highly unlikely" and 10 representing "highly likely" (Questions #17, #21, #24, see Appendix 1). These questions' points were awarded 1 through 10 for each question according to the subject's response. For example, a subject that responded with a 7 on the 1 to 10 scale concerning likelihood of sea level rise impacts would receive 7 points for his answer. The use of numerical scoring to evaluate interviewees' mental models was an independently developed approach that departed from typical mental model analysis as outlined by Morgan et al. (2002) and Jones et al. (2011). In order to facilitate quantitative statistical analysis of interviewees' responses, I used defined scoring rubrics to assign numerical values to the subjects and experts' answers. Interviewees' responses were coded and calculated to determine their scores for each of the five metrics, which were compared using statistical analysis methodology. Through the development and use of quantitative scores I was able to determine the statistical correlations and significance of relationships between interviewees' responses as displayed by the five metrics. This statistical approach provides a quantitative overview of the correlations between decision makers' mental models and the four other metrics.

Development of the Municipal Adaptation Score
The Municipal Adaptation score was obtained through analysis of six questions (Questions #2, #4, #7, #17, #20, #26, see Appendix 1) designed to elicit subjects' level of preparedness for undertaking climate change adaptation actions on a municipal level, and to determine how important they considered proactive planning and actions.
Subjects' qualitative responses were coded and assigned quantitative values between 1 and 5 based on defined scoring rubrics corresponding to each question. Questions and their scoring rubrics are as follows: • Subject believes the town should plan 3-5 years in advance 2 Subject believes the town should plan 5-10 years in advance 3 Subject believes the town should plan 11-20 years in advance 4 Subject believes the town should plan 21-50 years in advance 5 Subject believes the town should plan 50-100 years in advance • What changes should the town decision makers implement to minimize North Kingstown's risk of damage from climate change-driven impacts?
Points Subject Response 0 Subject does not think that any changes should be made 1 Subject has no clearly conveyed ideas of changes decision-makers should implement 2 Subject has at least one idea but it is off topic and/or not related to the question 3 Subject has vague ideas but does not clearly convey ideas for specific changes 4 Subject conveys specific ideas for changes that are related to minimizing impacts 5 Subjects conveys specific and detailed ideas for changes to minimize damages After all subjects' responses were assigned values based on these five rubrics, the average value of each subject's answers was determined by adding the values of all five questions and dividing by five. The resulting mean of each subject's responses determined their Municipal Adaptation score.

Development of the Education Score
Subjects' Education scores were obtained through responses to two questions (Questions #29, #30, see Appendix 1) concerning their previous participation in any conferences, symposiums, meetings, or educational programs that involved climate change or sea level rise information. Subjects who had not participated or attended any discussions or meetings concerning climate change or sea level rise received a score of 0. Subjects who had participated in general discussions or meetings involving climate change around the state of Rhode Island were given a score of 1. Subjects who had attended discussions or meetings specifically in North Kingstown regarding local sea level rise and climate change initiatives or problems as well as other meetings on the subject in the state received a score of 2.

Development of the State Involvement Score
The State Involvement metric represents subjects' perspectives concerning the level of guidance and oversight the state should have over towns in regards to developing and enforcing regulations, policies, and programs designed to assist towns in climate change adaptation. Subjects' transcribed responses were evaluated and divided into four categories based on the level of state involvement each subject preferred. Each category was then assigned a numeric value between 0 and 3 according to the following rubric:

0
• The state should stay out of the town's business, the state should leave the town alone and have no involvement with municipal planning of this kind • The state's current level of involvement is fine, they do not interfere much and their only role should be to provide information 1 • The state should assist towns in climate change planning but towns should take primary role • The primary role the state should play is in providing funding for adaptation planning that is developed and decided on by the town 2 • The state should take the lead in developing policy and regulations with the consultation and assistance of towns 3 •   (Zar, 1996). Inasmuch as the data collected in this case study did not uniformly have equal variances (as determined by the F-Max Test) or show normal distributions (as determined by the

Shapiro Wilks Test), I used the non-parametric statistical tests Spearman Rank
Correlation and the Mann-Whitney Two Sample Test for hypothesis testing.
Correlation analysis was used to test the null hypothesis that r = 0; i.e., that there was not a linear correlation between specific metrics as identified in the following hypotheses: a) Individuals whose mental models most closely fit the expert model of climate change and risk will be at a more advanced stage in the five-stage model of In addition, scatter plots of the metric scores were used to qualitatively evaluate relationships between variables. The Mann-Whitney test was used to test the null hypothesis that there was no difference in means between the two groups for each metric.

Research Questions 1 and 2: Municipal Preparedness for Climate Change Impacts
The first two research questions explored how town decision makers conceptualize their town's level of municipal preparedness for climate change impacts and compared their conceptualizations to the experts' views of the town's preparedness.
The views of experts and subjects regarding North Kingstown's level of preparedness for impacts from severe weather related to climate change were markedly different. Experts and subjects were asked to rank the town's current level of preparedness for ten different impacts on a scale from 1 to 5 with 1 representing "not prepared" and 5 representing "very prepared." The overall current preparedness level was determined by obtaining the mean of the ten different impacts. The same process was followed to determine the optimal preparedness score from responses to the question of how prepared North Kingstown should be for the same ten impacts as the previous question. Although treating each of the ten impacts as equally important for preparedness presents issues regarding the weight of each impact's severity and potential for damage, developing weighted variables for the ten impacts was beyond the scope of this case study, so for simplicity the impacts were all treated as having an equal weight. Future research in this area would benefit from using weighted variables that may provide greater insight into decision makers' understanding of the importance of different impacts and the potential risks they pose.
Experts viewed North Kingstown as currently ranking (mean + SD) 2.7 + 0.7 out of 5 on the preparedness scale and believed that the town should optimally be 4.5 + 0.7 out of 5 for the listed severe weather impacts. In contrast, subjects viewed the town as currently ranking 3.1 + 1.0 out of 5 and thought that they should rank 4.4 + 0.7 out of 5 in a best-case scenario. These initial results indicate that subjects consider North Kingstown currently to be better prepared than the experts believe it is and that the subjects do not think they need to improve the town's preparedness level as much as the experts do. However, the results of a Mann Whitney test used to compare the differences in the experts' and subjects' responses did not show statistically significant disparity between the two groups (Mann Whitney Test, W = 151, p < 0.6). This nonsignificant result indicates that the Mann Whitney test is likely not valid in this case due to the small sample size of experts (n=5). The low sample size of experts does not provide great statistical power for comparison, thus the results of the Mann Whitney test analyzing the disparity between experts and subjects are not a reliable indication of the true relationship between the two groups.

Figure 6: Experts' and Subjects' scores pertaining to North Kingstown's current and optimal levels of preparedness for climate change impacts. The scores are based on a 1 to 5 scale with 5 representing optimal preparedness.
The disparity between experts and subjects regarding North Kingstown's current and optimal levels of preparedness is an important factor to take into consideration when working on municipal preparation and adaptation goals in the town. This disparity indicates that in order to encourage town decision makers to place high priority on increasing municipal preparedness they must first recognize that their understanding of their current level of preparedness does not coincide with experts' views of the town's preparedness. Furthermore, experts believe the town needs to be more prepared than the town decision makers think it does, thus the decision makers are likely to cease preparation efforts before reaching the experts' perceived optimal preparedness level.

Research Question 3: Decision Makers' Behavior Change Correlation to Severe Weather Events
The third research question explored whether the stage of adaptive behavior change that town decision makers have reached as individuals related to previous personal impacts from severe weather events. Results from the 15 subjects interviewed indicate that 93.3% (14/15) of the subjects sustained some form of damage to their home or property from storms in the last ten years including flooding, loss of power for extended periods of time, and damage from falling trees. Among the interviewees, 1 subject was in Pre-contemplation, 1 was in contemplation, 2 were in action, and the remaining 11 subjects were in the Maintenance phase of behavior change regarding taking measures to prevent or minimize future damage to their home and property from storms.

. Subjects' current stages of individual behavior change and percentage of subjects that have sustained damage to their home or property from storms in the past ten years
I was unable to test the correlation between a subject's stage of behavior change and whether he or she was taking measures to prevent future damage because there was insufficient variation with respect to the behavior change variable. Only one subject in the sample did not sustain damage from storms in the last ten years and that individual was in the Maintenance stage of change while both subjects in Precontemplation and Contemplation had sustained damage. Thus, I was unable to draw any definitive conclusions regarding the correlation between decision makers' individual stages of behavior change and whether they had sustained previous damage from severe weather events due to lack of variation in the data collected.

Research Question 4: Correlation Between Individual Behavior Change and Support for Municipal Adaptation
The fourth question in this case study investigates whether town decision makers' individual levels of adaptive behavior change are related to their level of willingness to support implementation of adaptive actions at a municipal level. This question corresponds to my fourth hypothesis, that individuals who are at a more advanced stage in the five-stage model of personal adaptation behavior change will be more supportive of municipal adaptation planning and actions.
In order to answer this question and test the hypothesis, quantitative scores were developed as described in Section 3.5 for each of five metrics: Individual

The hypothesis that a positive correlation existed between subjects' Individual
Behavior Change score and their Municipal Adaptation Level score was tested using a Spearman rank correlation (r) analysis to test the null hypothesis that r=0 (no relationship). The results of the correlation analysis showed there to be no relationship between the two variables (r=0.08, p=0.77) ( There is a moderately positive correlation (r=0.49) between the Education and Municipal Adaptation Level metrics that approaches statistical significance (p= 0.06).
This correlation indicates that town decision makers who have participated in educational programs or meetings that included climate change topics both within the state and specifically in North Kingstown are somewhat more likely to place municipal adaptation as a higher priority and consider the town to be at risk from climate change impacts. Given this positive correlation, coastal managers and hazard mitigation experts may want to use climate change educational seminars and programs as a method of increasing municipal decision makers' attention to adaptation needs and promote proactive adaptation through local town-oriented programs as well as state-wide seminars.

Figure 7. Scatterplot displaying positive correlation between Education and Municipal Adaptation Level metric scores.
There is no correlation (r=0. 25, p=0.4

) between the Education and Individual
Behavior Change metrics, however the small sample size used in this case study and the fact that 73% of the subjects were already in the Maintenance stage means that this lack of correlation may not be a reliable result. Similarly, the lack of correlation (r=0.3, p=0.3) between the Education and Mental Model metrics may be due to the small sample size and the two-point Education metric scale. Future research including more extensive educational background analysis may provide a more accurate understanding of the correlation between education and mental models regarding climate change.

Research Question 5: Correlation Between Decision Makers' Mental Models and Support for Municipal Adaptation
The fifth research question explored whether the decision makers' conceptualizations of climate change, risk and adaptation were related to their level of support for municipal adaptive actions. This question corresponds to my second research hypothesis, that individuals whose mental models most closely fit the expert  Table 4

. Spearman correlations between climate change impact likelihood, Mental Model and Municipal Adaptation Level scores of subjects, p-values in parentheses.
Both subjects and experts were asked to estimate the likelihood that North Kingstown would experience impacts from climate change-related sea level rise over a period of years in order to determine whether a variation existed between experts' and decision makers' perceptions of when the town should expect impacts from sea level rise. Experts uniformly believed that North Kingstown should anticipate impacts from sea level rise far sooner than decision makers did. On a 1 to 10 scale with 1 representing "highly unlikely" and 10 representing "highly likely," experts believed that it was "very likely" (7.08 out of 10) that North Kingstown would experience impacts from climate change-related sea level rise in the next ten years (by 2022) while subjects believed it was "possible" (4.82 out of 10) that impacts might occur. Experts unanimously agreed there was "highly likely" (10 out of 10) that sea level rise would impact North Kingstown by 2100, however subjects thought that it was "moderately likely" (7.68 out of 10) that their town would experience impacts in the same time frame. This comparison also reveals that subjects believe that impacts from sea level rise will pose a greater risk to North Kingstown than increased storminess.
Future research is needed to explore why decision makers perceive sea level rise as a greater threat than increased storminess; however, I hypothesize that regular flooding of the downtown Wickford parking lot during spring tides and storm events (see The Spearman test results display a highly significant correlation (r=0.95, p<0.001) between subjects' beliefs regarding the likelihood that North Kingstown will experience impacts from future climate change (Question #17, see Appendix 1) and their level of support for municipal adaptive actions. These statistical results further support my findings that decision makers who perceive climate change impacts such as sea level rise and increased storm intensity as serious threats to North Kingstown are more willing to support proactive adaptation efforts on the municipal level. Thus, in order to increase proactive adaptive planning and actions, coastal managers, educators, and policy makers must make municipalities more aware of the reality and severity of risks posed by climate change impacts in the near future.  Table 5. Likelihood that NK will experience impacts from climate change-related sea level rise over a period of years. Scores are the average of a 0-10 ranking where 0 = highly unlikely and 10 = highly likely.

Figure 8. Subjects and Experts' estimates of the time frames in which North
Kingstown may experience impacts from sea level rise measured on a 1 to 10 scale with 1 representing "highly unlikely" and 10 representing "highly likely." This wide discrepancy between experts' and subjects' perceptions of the immediacy of sea level rise impacts in North Kingstown poses a significant challenge to coastal managers and policy makers trying to encourage proactive municipal adaptation. When the municipal decision makers believe that sea level rise and climate change are distant threats that do not pose direct risks to their town in the near future (within the next 10-20 years), they are unlikely to make the difficult, costly, and potentially unpopular decisions that would move the town towards proactive adaptation. This marked difference between experts' and municipal decision makers' perceptions of sea level rise risk must be taken into consideration by policy makers and coastal managers working to improve towns' hazard mitigation and climate change adaptation planning. An additional factor that must be considered is the fact that many municipal decision makers are elected and serve 3-to 5-year terms. These elected officials are often not willing to take unpopular actions that may jeopardize their political future; planning for climate change impacts that are not currently affecting their town is not a high priority, especially in the current economic environment. As one interviewee stated, "because of the economy and because of financial issues [North Kingstown] is in a pendulum swing causing these political issues [like climate change] to get pushed to the back burner." Planning for sea level rise and climate change necessitates looking farther ahead than the next election cycle and may entail difficult, costly, and unpopular decisions. Coastal managers and policy makers working with municipal officials must be aware of the decision makers' often limited time frames and political cycles. This creates the potential for conflict because long-term planning looking more than 10 or 20 years into the future must include climate change adaptation actions.

Significance of Study Results
The strongest correlations between metrics in this case study was between Subjects' Municipal Adaptation Levels, Mental Models, and State Involvement metrics. The single strongest correlation was between Municipal Adaptation Level and Mental Model scores with a Spearman correlation of 0.73 (p-value = 0.002). Several results of this case study stand out as particularly important factors for consideration by municipal decision makers, coastal managers, and policy makers.

Divergence Between Expert and Subject Mental Model Scores
The Mental Model metric in this case served as a measure of experts' and subjects' understanding and knowledge of climate change projections and likely impacts in the state of Rhode Island. This metric was also used as a method of gauging the level of importance subjects placed on preparing and planning for climate change individually and as a community, as well as indicating what level of risk they anticipate climate impacts pose in North Kingstown. Experts were in close agreement with each other that climate change impacts are a serious threat to the town, particularly the highly vulnerable historic village of Wickford, with a mean Mental Model score of 180 (SD = 10.2). In contrast, municipal subjects' mean Mental Model score was 156.7 (SD = 22.7). These results display a surprisingly wide divergence between experts and subjects, indicating that coastal managers working with towns to introduce climate change adaptation planning are thinking about climate change risks and impacts on a very different level than municipal decision makers. Additionally, there is a large variance within subjects' Mental Model scores which means town officials who need to work together to develop long-term planning goals and adaptive actions also have divergent understandings and perceptions of climate change, potentially resulting in uncoordinated, inefficient, or directly opposing approaches to dealing with the challenges presented by impacts such as sea level rise. In order to develop coordinated, effective planning focused on proactive adaptation in North Kingstown, town decision makers must work together. Coastal managers can assist in this effort by facilitating meetings and presentations and keeping officials up to date on the most current climate change science in an effort to reduce the variance within the decision makers' mental models. Decision makers who share similar mental models pertaining to their common task (in this case study, climate change adaptation) are far more effective at resolving problems, work more cooperatively, and produce better final decisions and plans as well as working as a more effective team. Thus, reducing the wide variability between North Kingstown's decision makers' mental models regarding climate change must be a high priority (see, Cannon--Bowers et al., 1990;Mathieu et al., 2000)

Variability in Subject Scores
Subjects' responses on all five metrics displayed much higher variance than experts' responses, indicating that decision makers in North Kingstown vary widely in their levels of knowledge, understanding, education, and beliefs concerning climate change impacts. The coefficient of variance values for all metrics clearly shows the differing levels of agreement between Subjects and Experts. In order to determine whether the level of variance between subjects' and experts' responses was statistically significant, the variance of each group (Subjects and Experts) was calculated for each metric and compared using the F-Max test of equal variance (Zar, 1996).

Metrics
Individual  Proactive adaptation planning and actions designed to mitigate the risks climate change impacts pose on a municipal level require the cooperation and coordination of the town's major decision makers and officials. However, cooperation between decision-making groups, councils, stakeholders, and state agencies is difficult to achieve in the best of times, and the politically charged topic of climate change makes collaboration even more challenging (Beratan, 2007). North Kingstown's decision makers currently have widely varying attitudes and beliefs regarding the risks posed by climate change impacts, the role the state should play in regulating adaptation actions and policies, and what should be done to prepare for likely impacts.
In order to effectively implement proactive adaptive planning efforts and adaptation actions at a municipal level the town decision makers must reach a closer agreement on what the risks are, who should be addressing them, and what needs to be done.
Improved communication between the numerous decision-making bodies within the municipal government combined with increased awareness of current climate change science is likely to reduce variation in the town officials' attitudes and beliefs, a necessary first step in movement towards proactive adaptive planning.

Correlation Between Education and Other Metrics
The results of this case study support previous studies' findings that education on a particular issue does not correspond to changes in individual behavior regarding the problem (Beratan, 2007;Andreasen, 2006;Kotler and Zaltman, 1971;McKenzie-Mohr and Smith, 1999); however, the statistical analysis does indicate a correlation between education and support for municipal adaptation. the need for "top down" education, starting with the decision makers, was also cited as a way of making decision makers "more receptive to adaptation ideas." The moderately positive correlation between the Education and Municipal Adaptation Level metrics indicates that improved access to climate change information and educational presentations or programs may result in increased support for municipal adaptation planning and actions among town decision makers.

Key Findings
The results of this case study indicate that municipal decision makers are largely unaware of current planning efforts in North Kingstown related to sea level rise and hazard mitigation. Although the Planning Department has been working with the URI Coastal Resources Center on a pilot project mapping sea level rise vulnerability and the project is currently entering its second phase, only 6 of the 15 decision makers interviewed mentioned anything related to this pilot project. Since a modified mental model methodology was used in the interview instrument, the fact that less than half of the decision makers mentioned the current planning project indicates either that they are unaware of these efforts or they may be aware of them but do not consider the project to constitute a significant action that the town is already taking towards proactive adaptive planning.

Conclusions and Recommendations
The results of this case study provide valuable insight into the attitudes, Planning is also involved in developing long-term plans for the state and could take a leading role by incorporating climate change projections in their work. Due to the divergent beliefs and attitudes within the largely elected municipal decision making body, the role of the State in encouraging and mandating adaptation actions should increase in order to provide continuity and free the adaptation process from the political election cycle. The results of this case study show that local decision makers do not have the knowledge (and possibly not the political will) to implement adaptation actions that may be unpopular and impact their ability to get reelected.
However, many of the decision makers do not want to turn responsibility for climate change adaptation over to the State because they are opposed to increased regulations and development restrictions. State agencies, such as the CRMC, consist primarily of appointed decision makers, thus they have greater ability to do long-term big-picture planning.
Although members of the CRMC are political appointees, the laws and regulations developed by CRMC provide limits to politics by establishing rules pertaining to the management of coastal resources and presenting policy recommendations. Municipal decision makers that serve in a short-term elected capacity potentially lack the knowledge, political will, and skill for long-term planning that is necessary for the development and implementation of proactive adaptation actions. Thus state-level decision makers and agencies must play the key role in preparing Rhode Island's coastal communities for the impacts of climate change.
However, it is important to note that a collaborative approach must be used to effectively overcome the challenges posed by climate change. In order for adaptation planning to be effectively implemented and enforced, state agencies and municipal decision makers must build trusting and cooperative working relationships (Beratan, 2007). While the state must be responsible and take the leading role in setting an agenda and providing tools for implementing proactive adaptation, Rhode Island's coastal towns need to be involved in the decision making process to increase community knowledge and understanding of climate change risks and thus improve cooperation and compliance.
(Highly unlikely) 1 2 3 4 5 6 7 8 9 10 (Highly likely) 21) Now I'm going to ask you how likely you think it is that NK will experience impacts from climate change-related sea level rise over a period of years. For the purposes of these questions 1 = highly unlikely and 10 = highly likely.
23) How high do you think SLR will be by 2050?

Impacts
How prepared do you think NK is for the following impacts?