1. Introduction
Nearly 22 million people living along the East Coast and the Gulf of Mexico in the United States are at risk from storm surge flooding (Zachry et al. 2015). The 2022 Atlantic hurricane season was a clear reminder of the devastating effects of storm surge on personal safety and economic well-being (Masters 2022). Recent hurricanes have also illustrated that communicating storm surge risks is challenging due to the complexity of the phenomenon (Zachry et al. 2015), the inherent uncertainty in forecasting (Fossell et al. 2017), and the public’s interpretation of information (Wilhelmi et al. 2023). Risk communication challenges have also been attributed to the types of visual information products and messages that are being used to warn the public about storm surge risks (Zachry et al. 2015; Lazrus et al. 2020). While recent research produced new knowledge on weather warning messages (Sutton and Fischer 2021; Sutton and Woods 2016), hurricane evacuation behavior (Huang et al. 2016; Lazo et al. 2015; Lindell and Perry 2012; Morss et al. 2016; Tanim et al. 2022), and risk perception (Lemée et al. 2022; Roy et al. 2022), there is still a knowledge gap in understanding how storm surge information presented in maps and other types of visualizations affect one’s understanding of storm surge risk.
Geovisualizations play a central role in communicating risks and hazards (Severtson and Vatovec 2012; Sherman-Morris et al. 2015; Lindner et al. 2018; MacPherson-Krutsky et al. 2020; Lindell 2020; Clive et al. 2021; Simpson et al. 2022). Literature that investigates people’s perceptions of information presented on maps is growing and spans multiple disciplines and hazard domains (MacPherson-Krutsky et al. 2020; Lindell 2020). Empirical studies on wildfires (Cao et al. 2016), volcanoes (Clive et al. 2021), floods (Houston et al. 2019), storm surge (Sherman-Morris et al. 2015; Wilhelmi et al. 2023), and sea level rise (Retchless 2018) illustrate that map design, dimensionality and realism of the visualized data, and cartographic elements, such as color and scale, play an important role in visual information processing and understanding of risk. Studies have shown that consideration of the audiences and their map-reading abilities in map design and thematic content could improve the effectiveness of hazard and risk communication (Meyer et al. 2014; Retchless 2014; Chen et al. 2016; Lindell 2020; Stempel and Becker 2021).
Prior research demonstrated that personalizing risk information can increase understanding of risk and improve protective behaviors (Severtson and Vatovec 2012; Maidl and Buchecker 2015; Markwart et al. 2019). A recent study on storm surge risk communication (Wilhelmi et al. 2023) illustrated that being able to “relate,” “connect to,” or “personalize” visual information helped people understand their storm surge risks. This was particularly helpful for people with lower map-reading abilities. While cartographic studies explored approaches of the representation of hazards and proposed frameworks for visualizing risks (Kostelnick et al. 2013), the knowledge of what aspects of visual representation increase map effectiveness, relatability, and risk understanding is still limited.
Studies that investigate different ways to visualize and communicate risks and hazards are growing (Kostelnick et al. 2013; Lindell 2020). Prior research, combined with the advancements in geovisualization technologies, produced new knowledge on the role of scale, recognizable context (Appleton and Lovett 2003), and dimensionality and realism (Chassin et al. 2022) in environmental representation. Specifically, regarding visual risk communication, Wilhelmi et al. (2023) found that storm surge maps presented at multiple scales and with familiar places increased the effectiveness of surge forecast maps in understanding risk. Research on climate change (Sheppard 2015; Sheppard et al. 2008; Retchless 2014) also demonstrated that the depiction of familiar features and recognizable contexts, especially at the local scale, could lead to greater personal engagement. The imagery of flood impacts accompanying flood projections (Keller et al. 2006) and hurricane storm surge forecasts (Rickard et al. 2017; Wilhelmi et al. 2023) have been shown to increase information comprehension and risk perception. Evaluation of three-dimensional (3D) visualizations, four-dimensional (4D) time-enabled animations, and immersive augmented and virtual reality (Chamberlain et al. 2015, 2016; Unreal Engine 2018; Tomkins and Lange 2019; Stempel and Becker 2021; Bernhardt et al. 2019) also showed improvement in information comprehension and risk perception and generated new knowledge on how people perceive space and place (Bruns and Chamberlain 2019). These examples illustrate the potential to improve risk communication through geovisualizations, especially when cartographic elements and visual representations of risk connect people to their environment or, broadly speaking, to a place.
Geovisualizations are one of several mechanisms for connecting people to information about the environment. To understand how people perceive their connection to an environment using geovisualizations, we can draw from prior work about sense of place (Tuan 1974, 1977; Masterson et al. 2017; Raymond et al. 2017). The concept of sense of place broadly describes how humans connect to places, through meaning, functions, and emotional bond (Cresswell 2004, 2014; Stedman 2008; Trentelman 2009; Masterson et al. 2017; Raymond et al. 2017). Considering values, meanings, and feelings that people associate with a specific place (Steele 1981; Williams and Stewart 1998), sense of place scholarship generally encompasses concepts of place meaning and place attachment, where place attachment includes place identity and place dependence (Williams and Vaske 2003; Raymond et al. 2010; Kudryavtsev et al. 2012; Brown et al. 2015; Raymond et al. 2017), although nuanced conceptualization, analytical approaches, and measures vary (Bukvic et al. 2022). Prior theoretical and empirical work showed that the sense of place concept is complex, where socially constructed boundaries of a place and associated place meanings (Raymond et al. 2017) interact with specific geographic locations, geographic scales, and sociocultural and biophysical elements of place (Williams 2014; Raymond et al. 2017; Ardoin et al. 2019; Quinn et al. 2019). These latter representations of a place have direct implications for incorporating sense of place constructs into geovisualizations, although research in this area remains limited.
Just as sense of place can influence feelings of connection, the scale of place (Lewicka 2011; Raymond et al. 2017; Quinn et al. 2019; Ardoin et al. 2019) and visual representations of biophysical and sociocultural meaningful elements of place (Newell and Canessa 2018; Wilhelmi et al. 2023) also offer implications for storm surge risk communication. Brown et al. (2020) demonstrated that mapping place values using participatory approaches provides an important insight into place meaning at different scales. Studies also linked measures of place attachment to proenvironmental behaviors (Kudryavtsev et al. 2012), resource planning (Cantrill and Senecah 2001; Newell and Canessa 2018), climate adaptation (Quinn et al. 2015), and postdisaster behaviors (Chamlee-Wright and Storr 2009), thus illustrating a potential to explore the role of place attachment in the context of storm surge risk information.
This study builds on previous work and contributes to the intersection of sense of place, geovisualization, and risk communication scholarship. Recognizing that numerous conceptual frameworks of sense of place exist, here, we conceptualize that place meaning (i.e., valued elements of biophysical and sociocultural environments), place attachment (i.e., place identity and place dependence), and scale of place (e.g., community, city, region) interact to form a sense of place. For the purpose of geovisualizations, we consider place, more broadly as an intersection of geographical space and human perceptions and meanings (Merschdorf and Blaschke 2018).
We explore hazardous weather communication (Morss et al. 2017; Lazrus et al. 2020; Wilhelmi et al. 2023) by investigating the following research questions: 1) How do residents of coastal communities perceive a sense of place in their everyday lives and when hazardous weather threatens? 2) How can place and sense of place constructs be integrated into geovisualizations of storm surge threats? 3) What relationships exist between different ways of representing a place in storm surge risk geovisualizations and people’s ability to connect to a place, and how are these relationships influenced by sociodemographic characteristics, map-reading ability, and sense of place measures? First, we empirically explore the concepts of place meaning, scale of place, and place attachment in the context of a storm surge risk. Second, we identify and implement cartographic and visual design approaches for visualizing places. Third, we conceptualize how geovisualizations enable people to connect to a place and develop measures of connecting to a place through visuals. And finally, we integrate these concepts and measures to assess the effect of geovisualizations in a large population survey.
2. Methods
We employ an interdisciplinary approach that combines a cognitive mapping participatory process, development of storm surge risk geovisualizations (also referred to here as visuals), and testing these visuals in a survey with members of the public in at-risk communities in the southeastern United States.
a. Study area
This study was conducted in the coastal communities of South Carolina and Georgia (Fig. 1)—an area vulnerable to storm surge. This vulnerability is due to its relatively flat coastal geography, growing population, and aging infrastructure. The study area included urban and rural communities, located on the mainland and on barrier islands. It also included a range of population sizes and demographics. Prior to our study, this region had experienced multiple tropical storms, including Hurricanes Dorian (2019), Michael (2018), Florence (2018), Irma (2017), and Matthew (2016), several of which produced significant coastal flooding.
Map depicting the study domain overlaid with areas vulnerable to hurricane storm surge. Storm surge flooding data source: Zachry et al. (2015).
Citation: Weather, Climate, and Society 15, 4; 10.1175/WCAS-D-23-0044.1
b. Development of storm surge risk visuals for testing in a survey
1) Exploring place meaning
The first step in designing storm surge geovisualizations was conducting the cognitive mapping focus groups. Cognitive maps bring together personal narratives and a spatial cognitive schema to provide a representation of those narratives as they relate to spatial elements of characteristics (Rapoport 2016). Building on prior research on place meaning and place value mapping (Brown et al. 2015, 2020), we conducted a series of focus groups in coastal Georgia communities within our study area (Fig. 1). Focus groups were intended to facilitate a grounded theory approach (Thornberg 2012) to help us theorize categories of places, features, and experiences that are meaningful and identifiable on a map across a range of residents. To accomplish this, we merged a cognitive mapping participatory approach with focus groups to form a grounded visualization process (Forrester et al. 2015; Kellams 2006; Knigge and Cope 2006). Through this work, we aimed to learn about a scale of place, meaningful locations, and valued features of coastal biophysical and sociocultural environments. We focused on characterizing the everyday places and activities, as well as places individuals associated with a storm surge hazard.
The six focus groups, conducted in November 2019, included urban populations (Savannah and Garden City) and populations who live outside of urban areas (Pin Point), on barrier islands (Tybee Island and Skidaway Island), and in rural communities (Crescent, McIntosh County). Focus groups were composed of 6–10 individuals with a total of 44 participants. Nearly half of the participants (48%) lived in their community for more than 30 years, with the majority owning their homes. Among the participants, more than half experienced storm surge and 95% have seen storm surge maps in the past. The participants’ age was between 25 and 85 years old, and 52% identified as female. Racial composition included 36% Black and 55% White. The majority had an annual income greater than $50,000.
In the cognitive mapping activity, participants were given a blank sheet of paper and two blank maps at local (town) and regional (county) scales, specific to each focus group location. First, the participants were invited to write down what they valued in their communities, including people, activities, areas, or specific places that make them feel that they are part of this community. The participants were then asked to identify where these places, activities, or people are located using one or both blank maps (town or county) based on the spatial extent that best illustrates the scale of their community. These individual-level activities were followed by group discussions about what participants valued the most about their community and what they perceived as being at risk from storm surge. Figure 2 shows examples of the listing and mapping outputs, including nonspatial place meaning features (e.g., culture) as well as those that have specific geographic locations (e.g., church).
Examples of output from the cognitive mapping activity: (a) lists from the Skidaway, Pin Point, Garden City, and Savannah focus groups; (b) regional-scale map from the Crescent focus group, and (c) local-scale map from the Tybee Island focus group. The red color on the maps indicates places perceived to be vulnerable to storm surge.
Citation: Weather, Climate, and Society 15, 4; 10.1175/WCAS-D-23-0044.1
The different types of meaningful places identified in the focus groups informed the development of geovisualizations for testing in the survey. To analyze the meaningful places identified, spatial data from the participatory mapping activity were digitized in Esri ArcGIS software. A total of 325 geographically identifiable locations across all focus groups were inductively coded into 16 categories: economy, recreation, heritage and culture, environment, food, social, spiritual, essential services, aesthetics, transportation, entertainment, learning and research, health, leisure, governance, and safety. Following Brown et al. (2015) and the literature on ecosystem services (La Rosa et al. 2016), these inductively created categories were aggregated or further revised for consistency with prior studies to create a place meaning typology. The final typology included nine categories representing people and social connections, natural environment, heritage and culture, economy, recreation, religion/spirituality, food, aesthetics, essential services, and infrastructure. Examples of these categories were used in the visuals described below.
2) Visual content, design, and representation of place
The focus groups informed the development of geovisualizations. Given our large study domain, it was prohibitive to generate visuals for each location where a survey respondent lives. To control for differences in geographic locales, as well as existing knowledge of a specific area, we generated a hypothetical (fictional) place (Baker 1995; Newcombe and Chiang 2007) for the design of storm surge risk visuals to be tested in the survey. Informed by the focus group data, we chose a scale of place of approximately 20 mi × 20 mi (32 km × 32 km). Within that scale, we created a hypothetical coastal town, named Keel, with a typical Georgia and South Carolina coastline, hydrography, and road network. Keel’s geography was derived from a real coastal location outside of our study domain. We modified the geography to eliminate major portions of the terrain, customized major roadways, and created fictitious locations representing features of the biophysical and sociocultural environments per our place meaning typology.
For two-dimensional (2D) visuals, we used Esri ArcGIS Pro software augmented by Adobe Photoshop and Illustrator for producing high-quality cartographic outputs and images of flooded places with varying levels of water heights. Images of storm surge impacts accompanying the locations of select places (e.g., church, grocery store) were designed in Adobe Photoshop using the Flood 2 plugin software (Flaming Pear 2023). We customized light, color, and textures for a more realistic representation of the storm surge inundation. For the 3D images, we used Esri ArcGIS Pro for an abstract representation of an urban neighborhood and Google Earth software for a more realistic 3D view. The storm surge data were obtained from NOAA’s National Hurricane Center (i.e., ensemble product of maximum storm surge heights for category-3 hurricanes; Zachry et al. 2015) and were held constant across all visuals. We used three incrementally darker gradations of a single (blue) color to symbolize flooding above ground value ranges (Brewer 2006). We simplified and smoothed the inundation boundaries for the purpose of testing in the survey across multiple digital devices (e.g., desktop computer, smartphone).
Table 1 describes the six visuals with different cartographic elements and visual representations along with the rationale for selecting these design and mapping approaches. For each of the six visuals, two versions were created: one with a pinpoint locating a hypothetical home inside the area with 3–6 ft of forecasted flooding and one locating the home outside of the flood area. Thus, there were 12 visuals in total. The final geovisualizations are shown in the results, section 3a.
Description of geovisualizations; the final geovisualization figures are provided in section 3a.
c. Survey with members of the public
1) Survey instrument
The survey used for testing the geovisualizations is a part of a larger project and related research on sense of place and risk perception. The survey instrument was structured into three major parts. Part 1 (preexperiment) began with several sociodemographic questions used in respondent screening and quotas (i.e., age, home zip code, gender, race, ethnicity, and income). This was followed by questions about respondents’ perceived residence in a hurricane evacuation zone, past hurricane experiences, hurricane information sources, interpretation of storm surge forecast water levels, and meaningful community characteristics (not used in this article). Part 1 also included questions about respondents’ sense of place, including scale of place and place attachment. The scale of place question was based on research by Brown et al. (2015) and Ardoin et al. (2019). The place attachment measures were adapted from Williams and Vaske (2003), which included place identity and place dependence measures, with their sixth place dependence item removed because other empirical studies found that it was not (or only weakly) statistically related to place dependence (see also Raymond et al. 2010).
In part 2 (experiment), all respondents were presented with the following scenario: “Imagine that you live in a place called Keel, which is near the coast of the Atlantic Ocean. Keel has a population of about 100 000 people. Imagine that it is Monday morning, and a hurricane is forecast to affect your area in about 48 h (2 days), on Wednesday morning. Below is a figure that shows the forecast of flooding from storm surge for your area. The location of your home is shown by a yellow marker.” Following this description, the respondents were randomized into 1 of 12 experimental conditions, in which they were presented with 1 of the 12 visualizations described in sections 2b(2) and 3a.
Each respondent was then asked the same series of questions to measure their understanding of, perceptions of, and responses to the information presented in the geovisualization. This included several questions not reported on here, about cognitive and affective risk perceptions, efficacy beliefs, and intended behavioral responses, adapted from recent and concurrent survey research (Lazo et al. 2015; Morss et al. 2016, 2018; Demuth 2015; Demuth et al. 2016, 2023; Walpole and Wilson 2021). It included measures of map-reading ability (i.e., use of scale) and of the effect of geovisualizations on enabling people to connect to a place, which we developed building on prior research (Brown et al. 2003; Kianicka et al. 2006; Scannell and Gifford 2010; Wilhelmi et al. 2023).
In part 3 (postexperiment), all respondents were asked questions about additional sociodemographic characteristics (e.g., education, housing type, years of residence in the area), respondents’ access to basic needs, cultural worldviews, and stated preferences among the visuals (not included in this article).
Here, we present results from a subset of questions focusing on understanding respondents’ sense of place and the effect of different geovisualizations on connecting people to a place. Additionally, we consider how these effects vary with sociodemographic characteristics, measures of place attachment, scale of place, and map-reading abilities. The survey questions and response scales used to measure the latter three sets of variables are shown in Table 2.
Survey measures for the sense of place, map-reading ability, and connecting to place variables used in this study.
2) Survey data collection
Survey data were collected by a national survey company, Qualtrics, using available panels in the region. All questionnaires were self-administered by respondents in a web-based environment. Our target survey population was residents (18 years or older) living within zip codes in Georgia and South Carolina at risk of storm surge inundation (Fig. 1). Initially, we defined these zip codes based on potential inundation areas from a category-3 hurricane, using the same NOAA storm surge data used to create the visuals [section 2b(2)]. Additional zip codes were added to decrease time in the field, including those with potential inundation of category-4 hurricane, resulting in a total of 81 zip codes, with 50 zip codes from South Carolina and 31 from Georgia. Among these, 36 were considered urban (per U.S. Census urban classification) and 45 were considered rural representing communities located on the mainland and on barrier islands. We administered the survey between 21 July and 3 September 2020. To reduce the potential for overweighting of certain populations in the respondent population, we used a quota-based sampling strategy in which Qualtrics recruited respondents in their panel while also filling quotas in certain categories, including urban and nonurban zip codes, age, household income, gender, and race. A total of 4394 people opened the survey with 2103 starting the survey after acknowledging their informed consent. From these, 1442 provided complete and quality-controlled responses. All participants, volunteers who sign up to participate in research studies, were recruited by Qualtrics through opt-in databases, registration websites, or panel management companies.
d. Data analysis
The survey data were analyzed in the Statistical Package for the Social Sciences (SPSS). Frequency analyses were conducted to provide descriptions of demographic characteristics and quantify responses to survey questions. We then conducted a series of linear regression analyses to investigate the effects of different potential predictor variables on the measures of sense of place and connecting to place shown in Table 4. Each of the regression analyses included as predictors the set of sociodemographic variables shown in Table 3. Income was recoded as two binary variables, comparing those with high (>$100,000) and low (<$50,000) annual household incomes to those with medium incomes (between $50,000 and $100,000).
Sociodemographic and map-reading characteristics of the survey population.
The first set of linear regressions examined the effect of these predictor variables on each of the two sense of place measures. In the second set of linear regressions, each of the four connecting to place measures was used as the dependent variable. Because these connecting to place questions were asked in the experimental section of the survey, this set of regressions included as additional predictor variables respondents’ hypothetical home location in the visual (outside or inside the flooded area) and the visual they received. The six visuals were included in the regression as five binary variables, using the Control visual as the reference category. We also added map-reading ability (measured as shown in Table 1) as a predictor, with respondents who selected the correct distance (25 mi) coded as correct, and all other responses (other distances or I do not know/I cannot tell) coded as incorrect. The third set of linear regressions was the same as the second set, with two more predictor variables added: the two sense of place measures.
We also explored whether the effect of any visual on people’s connection to a place differed based on their levels of sense of place. To do this, we ran a series of regressions using the same variables as the third set of linear regressions, but with visual statistically interacted with either scale of place or place attachment. However, none of the interactions in these models were significant. Thus, we present the more parsimonious set of regressions without interactions.
In the second and third sets of regressions, as noted above, binary variables were used to compare each of the other five visuals in Table 1 to the Control visual. To compare differences between each of the visuals, we also ran versions of the same linear regressions, but with pairwise comparisons between each of the visuals substituted, one at a time, for the full set of visual binary variables. We refer to these as pairwise visual comparison models, without and with sense of place predictors.
3. Results
a. Geovisualizations
The visualizations shown in Fig. 3 correspond to the descriptions in Table 1. Figures 3a–f illustrate all six versions of the visuals for a hypothetical home location (shown by a yellow pinpoint mark) inside the flood area (referred to as an “in group” in the survey experiment). Figure 3g shows an example of the visuals representing a hypothetical home location outside the flood area (referred to as an “out group” in the survey experiment). The visuals show different ways of representing place meaning (Figs. 3b,c), geographic scale (Figs. 3d–f), dimensionality (Figs. 3e–g), and realism (Figs. 3c,g).
(a)–(f) Visualizations for the six different variations described in Table 2 showing storm surge flooding for a hypothetical home location (yellow mark) inside the flood area. (g) As in (f), but for a hypothetical home location outside the flood area.
Citation: Weather, Climate, and Society 15, 4; 10.1175/WCAS-D-23-0044.1
Compared to the Control visual (Fig. 3a), 2D Labels (Fig. 3b) aims to increase personal relevance and ability to connect to a place, by showing locations of meaningful places [discussed in section 2b(1)]. The 2D Photos (Fig. 3c) builds on 2D Labels and aims to create a more affective connection to a place and help the audience to visualize different water heights above ground. Three images were created to represent examples of our place meaning typology: 1) people, social connections, and religion/spirituality are represented by a photo of a flooded church, 2) essential services/infrastructure and food categories are represented by a flooding near a supermarket, and 3) natural environment, economy, and recreation categories are represented by a damaged fishing/shrimping boat. Each photo corresponds with the location on the map with different classifications of water heights. The 2D Inset visual (Fig. 3d) aims to provide greater personal relevance and connection to a place by localizing risk information through geographic scale. By “zooming in” on a neighborhood area with street labels and building footprints, the hypothetical home location is clearly visible relative to the boundary of the forecasted flood area. The 3D Abstract visual (Fig. 3e) is included to explore the effect of dimensionality and localization with a simple building geometry, emphasizing building shapes and water height above ground, relative to street level. The 3D Realistic visual (Fig. 3g) aims to explore the effect of realism in 3D models in addition to dimensionality. This visual includes a photorealistic representation of a landscape with a variety of colors, textures, and geometries.
b. Descriptive statistics
1) Characteristics of the survey respondents
The sociodemographic characteristics of the survey respondents are shown in Table 3. Slightly more than half of the 1442 respondents were female. The majority of the respondents reported their race as White (70.6%), whereas 29.4% of respondents reported their race as Black or African American (23.4%), American Indian or Alaska Native (2.6%), Asian (3.3%), Native Hawaiian or Pacific Islander (0.3%), or Other (4%). The respondents’ age range was 18–88 years old, and they had lived in their communities ranging from less than a year to 76 years.
For map-reading ability, 62.5% of respondents answered the question about map scale incorrectly, with 44.3% selecting an incorrect numerical response about the distance between two map locations and 18.2% saying “I don’t know/I can’t tell.” Additional analysis (not shown) indicated that respondents who were younger, White, of at least medium income, and those with a bachelor’s degree or more were more likely to answer this question correctly. Similar to Montello et al. (1999), we did not see differences between male and female respondents. These findings suggest that there can be potentially important relationships between population characteristics and interpretations of map-based information. There was no difference in the map-reading ability measure by hypothetical home location or geovisualization.
2) Distributions of dependent variables
Summary statistics (Table 4) and distributions (Fig. 4) are shown below for the primary dependent variables. For scale of place, the majority of respondents said that the region they most strongly identify with and depend on is within a 29-mi range of their home, with only about 10% of respondents reporting a range of 70 mi or greater. For place attachment, although some prior research has treated measures similar to those we used as two conceptually different scales, place identity and place dependence (e.g., Williams and Vaske 2003; Raymond et al. 2010), other studies have found that these components of place attachment are highly correlated (see, e.g., Kudryavtsev et al. 2012). We performed a statistical common factor analysis to explore whether our measures represent one or more underlying latent constructs by assessing the common variance that is shared among them (DeVellis 2011). Our data indicated that the 11 question measures loaded onto a single factor rather than clustering into two factors of place identity and place dependence (see in the online supplemental material), and thus, we treated place attachment as a single scale, calculated as the mean of the 11 measures. We found place attachment to be high, with approximately two-thirds of respondents reporting a value higher than 3 on the 1–5 scale.
Summary statistics for the two sense of place and four connecting to place variables.
Frequency distribution for the two sense of place and four connecting to place variables.
Citation: Weather, Climate, and Society 15, 4; 10.1175/WCAS-D-23-0044.1
For the four connecting to place variables, the distributions are left-skewed, suggesting that overall respondents reported that they were able to relate to the information, envision themselves in the fictional community of Keel, make sense of the information, and use it to understand the risk of flooding from storm surge. For relatability and perceived sensemaking and understanding, approximately three-quarters of respondents selected “agree” or “strongly agree.” Envisionability was somewhat lower, although still overall high, with approximately two-thirds of respondents selecting “agree” or “strongly agree.”
c. Linear regressions predicting scale of place and place attachment
Table 5 shows the results of the first set of linear regressions, predicting the scale of place and place attachment from the sociodemographic variables. We see statistically significant effects for age for both variables such that older respondents reported a greater attachment to their communities, and they also reported considering their community to occupy a smaller area. We see significant effects for race such that White respondents reported a considerably smaller scale of place than non-White respondents. For income, those with annual household incomes above $100,000 (compared to those with incomes between $50,000 and $100,000) reported higher place attachment and larger scales of place. Those who had resided for longer in a community and with more years of formal education also reported a greater attachment, whereas those with more years of formal education reported smaller scales of place. Neither gender nor community type, defined here by residence in a rural or urban zip code, were significant predictors of either of these dependent variables.
Results of linear regression predicting scale of place and place attachment using sociodemographic characteristics. One asterisk (*) indicates p < 0.05, two asterisks (**) indicate p < 0.01, and three asterisks (***) indicate p < 0.001. Bold denotes statistically significant results.
d. Linear regressions predicting connecting to a place
Table 6 shows the results for the second set of linear regressions, predicting relatability, envisionability, perceived sensemaking, and perceived understanding of risk from the sociodemographic, map-reading, and visual-related variables. For the sociodemographic variables, Table 6 shows that White respondents reported a greater ability to relate to the information, greater ability to envision themselves in Keel, and greater ease in making sense of the information. Respondents with incomes below $50,000 reported a lower ability to relate to the information and to envision themselves in Keel. In addition, those with a bachelor’s degree reported a greater ability to envision themselves in Keel than those without a bachelor’s degree. Age, gender, length of residence in community, and community type (urban or rural zip code) had no significant effects on any of these four dependent variables. Whether respondents were depicted as inside or outside of the flooded area also had no significant effects on connecting to place.
Results of linear regression predicting four measures of connecting to a place (each on a 1–5 scale) using sociodemographic characteristics, visuals, hypothetical home location, and map-reading ability. One asterisk (*) indicates p < 0.05, two asterisks (**) indicate p < 0.01, and three asterisks (***) indicate p < 0.001. Bold denotes statistically significant results.
The measure of map-reading ability was a statistically significant predictor for all four connecting to place measures (Table 6). Specifically, those who were able to correctly identify the distances between two map locations reported a greater ability to relate to the information, greater ability to envision themselves in Keel, greater ease in making sense of the information, and greater perceived understanding of storm surge risks using the information in the visual they received. This finding supports prior research on the use of scale as a basic map-reading ability and its effect on map comprehension (MacPherson-Krutsky et al. 2020).
Comparisons of each visual with Control are shown in Table 6, and results are summarized across the pairwise visual comparisons in Table 7. Across these analyses, only 3D Abstract exhibited different effects from the other visuals, contributing to higher reported connection to a place. Respondents who received the 3D Abstract visual (Fig. 3e) reported higher relatability than those who received the 2D Inset (Fig. 3d) or 3D Realistic visual (Fig. 3f). Those who received 3D Abstract also reported higher perceived sensemaking than those who received Control or 2D Inset (Fig. 3d). And those who received the 3D Abstract visual reported higher perceived understanding than those who received several of the other visuals. The visuals had no significant effects on respondents’ reported ability to envision themselves in Keel.
Results from linear regressions predicting four measures of connecting to a place using pairwise comparisons of visuals, along with the other predictors in Table 6 [without sense of place (SoP) variables] and Table 8 (with SoP). Only statistically significant (p < 0.05) results for visual comparisons are shown; none of the pairwise comparisons between Control, 2D Label, 2D Photos, or 2D Inset were statistically significant. One asterisk (*) indicates p < 0.05, two asterisks (**) indicate p < 0.01, and three asterisks (***) indicate p < 0.001. Bold denotes statistically significant results.
e. Linear regressions predicting connecting to a place, with scale of place and place attachment as added predictors
Table 8 shows the results of adding the scale of place and place attachment measures to the regression models discussed in section 3d. The results for the predictor variables included in both sets of models are similar to those shown in Table 6; the only change is that education no longer has a statistically significant effect on envisionability. The results from most of the pairwise comparisons between visuals are also similar to those in section 3d: the 3D Abstract visual contributed to greater relatability, perceived sensemaking, and perceived understanding compared to several of the other visuals, but it had no effect on envisionability (Table 7).
Results of linear regression predicting four measures of connecting to a place using sociodemographic characteristics, visuals, hypothetical home location, map-reading ability, scale of place, and place attachment. One asterisk (*) indicates p < 0.05, two asterisks (**) indicate p < 0.01, and three asterisks (***) indicate p < 0.001. Bold denotes statistically significant results.
For the sense of place variables, respondents with smaller (more local) scales of place reported a greater ability to relate to the information in the visual they received. Notably, there are significant effects of place attachment for all four dependent variables: respondents with greater place attachment reported a greater ability to relate to the information, envision themselves in Keel, make sense of the information, and understand risks associated with storm surge using the information.
4. Discussion
a. Understanding sense of place in coastal communities
The results show that the sense of place in our studied population is nuanced and multifaceted. Our six focus groups, representing different locales near Georgia’s coast, indicated that place meaning is constructed from an intersection of environment, culture, social connections, and economy. In addition, nature-based recreation, food, religion and spirituality, aesthetics, and infrastructure have a role in the people–place relationships. While the full analysis of the focus group data is outside the scope of this paper, the results shown here illustrate that place meaning includes both biophysical and sociocultural dimensions with many location-specific features that can be mapped and visualized across spatial scales.
Across our survey sample, we found place attachment to be relatively high. We also found that place attachment tended to be higher among respondents who were older, had higher incomes, had lived in their community longer, and had more years of formal education. These results are consistent with other studies (e.g., Bukvic et al. 2022), showing that coastal residents form a positive bond with their surroundings due to proximity to the natural amenities, such as beaches and water, and related industries (Brown et al. 2015; Maguire et al. 2011). Our findings are also consistent with prior research that associated higher place attachment with higher income (Kick et al. 2011) and longer lengths of residence (Lewicka 2011). There has been less consensus about the role of age and education in place attachment (Jamali and Nejat 2016), and thus, the associations we found offer interesting insights that could be explored in the context of other sociocultural variables.
Our survey included a commonly used place attachment questionnaire (Williams and Vaske 2003; Raymond et al. 2010). However, we discovered the responses did not produce the differentiation between place identity and place dependence found in some prior work, with a statistical factor analysis showing that all 11 measures of the place attachment scale loaded onto a single factor. In the context of this research, we can see one possible explanation. Contrary to previous studies that measure place attachment in the context of a specific location (e.g., reservoir, national park), we consider place in a broader sense of a coastal community, thus combining both functional (i.e., place dependence) and emotional (i.e., place identity) aspects of place attachment.
This study offers insights about the scale of place variability within our population sample. Our results suggest the majority of our survey respondents perceived their scale of place to be within a 29-mi range. Larger (more regional) scale of place was associated with younger and non-White respondents and those with fewer years of formal education. Those with higher income also reported a larger geographic area for their scale of place, which is similar to the findings from Ardoin et al. (2019). These findings provide a foundation for further investigation, especially in the context of the coastal environment. Previous studies did not find age (Brown et al. 2015) and race (Ardoin et al. 2019) to be significantly associated with the scale of place.
Interestingly, we did not see a significant difference in place attachment or scale of place between urban and rural residents in our sample. This could be explained by the proximity of coastal recreational and natural amenities that residents of urban areas find meaningful and use in their everyday lives. This also could stem from our classification of rural and urban zip codes (i.e., per U.S. Census population density classification of urban areas). We recognize that “urban” and “place” concepts can be contextual and lack precise spatial boundaries compared to the predefined boundaries of the zip codes. Future research could explore how the scale of place and place attachment vary along the urban–rural continuum using different approaches for defining “urban” communities.
b. Integrating sense of place into geovisualizations of place
This research presents a novel way to blend the functional purpose of a map while simultaneously trying to facilitate an emotional connection to a place. This was achieved by combining qualitative social science research methods, data visualization techniques, cartographic principles, and storm surge forecast information. Earlier studies showed that cartographic research has benefited from the use of qualitative methods (Suchan and Brewer 2000), and in this study, we explicitly integrated focus group methodology as a first step in geovisualization development. Creating a collective cognitive map of meaningful places was instrumental in identifying a scale of place and generating the place meaning typology. The methods developed in our research on storm surge risk can be applied to other hazard risks.
We recognize there are numerous ways to visually represent a hazard and a place (Lindell 2020). In this study, we developed geovisualizations for testing in an online survey; therefore, the visuals had to be adaptable for different digital devices (e.g., desktop, tablet, phone) and aligned with products currently used by agencies and mainstream media. In addition, we had to balance the number of different conditions (visuals) along with the survey size and a host of related questions. For this reason, we settled on six geovisualizations that varied along an array of visual representations and aspects of place and sense of place. Additional geovisualizations can be explored in the future, but we would advocate these should be practical and easily and widely distributable and the study be robust.
This study highlighted the challenges of creating maps for communicating hazard risks to the public (Dransch et al. 2010). As our survey respondents represented multiple communities within the study area, one of the challenges was to create a fictional community that resembled real-world locales and contained emotionally interesting and functionally relevant content based on our place meaning typology. We carefully considered which features to label and selected layouts, colors, and text sizes to draw attention to map features representing a place while avoiding visual clutter. The survey question on envisionability helped to validate our approach.
The geovisualizations’ design also emphasized the challenge of representing spatially uncertain storm surge forecasts at a local scale. While local scale may be more suitable for the personalization of information and the use of 3D visual techniques, the flood boundaries at this scale may appear to have higher spatial precision than the forecast models can produce. The intersection of weather forecast uncertainty and visualization of risks at a local scale is an important area for future research.
c. Connecting to a place through visuals
Our findings suggest that all visuals resulted in relatively high scores for all four measures of connecting to place. This is an interesting result, as we hypothesized that visuals that aimed to provide higher personal relevance (e.g., 2D Photos) or dimensionality and realism (3D Realistic) would be more effective compared to the Control visual. One explanation is that all our maps, including the Control, were developed following best practices from cartography and risk communication research (e.g., Dransch et al. 2010) and therefore were generally effective. Another explanation is that the respondents who received the Control visual may have been familiar with this type of representation of storm surge risk, as similar visuals are commonly used in mainstream media and by government agencies.
When we look at the relative impacts of different visuals on the ability to connect to a place, we found that the 3D Abstract visual was associated with increased relatability, perceived sensemaking, and perceived understanding, compared to other visuals. Although these effects are small, Tables 6–8 together suggest that they are robust. We can explain and contextualize these findings in a number of ways. First, there is evidence that simplifying the representation of buildings through abstraction may improve cognition (Chassin et al. 2022). Second, it is possible that the 3D Abstract visual allows respondents to identify location-based risks (Severtson and Vatovec 2012) while providing a connection to a place through a simplified 3D urban scene (Sheppard 2012). And third, specifically regarding the understanding of risk, we recognize that the survey question did not directly assess respondents’ understanding of the risk information; rather, it asked them to self-assess their understanding. Given this, it is possible that the 3D Abstract visual provides a sense that the information is simpler or easier to accurately interpret than it actually is. While we cannot conclude that survey respondents actually understood the risk information, our findings can lead to future research that will objectively measure the comprehension of risk information.
One noteworthy result from this study is the effect of sociodemographic factors on several measures of connecting to place. Specifically, non-White respondents reported a lower relatability, envisionability, and perceived sensemaking than White respondents, and low-income respondents reported a lower relatability and envisionability. These findings are important because they indicate the potential for more systemic issues. They suggest that regardless of how the visual information is presented, we may see substantial impacts of factors such as race and income on connecting the risk to place, which may then have potential downstream effects on subsequent behavior in response to the risk.
Map-reading ability was associated with differences in all four measures of connecting to place, with respondents who correctly used map scale to answer the distance question, reporting higher levels. Our results also suggest that the ability to correctly identify distance using map scale was generally low in our population sample and associated with race, age, income, and education. These findings and significant relationships between population characteristics and interpretations of map-based information could have practical implications in communicating risks to the public and can be further investigated, using a larger set of map comprehension questions (MacPherson-Krutsky et al. 2020). Future research could also investigate whether digital devices (e.g., viewing information on a smartphone versus a desktop computer) play a role in how people interact with maps and use scale and distance information.
When we add place attachment and scale of place into the regression models, we see that place attachment is a significant predictor of all four measures of connecting to place. This is notable for a couple of reasons. First, our place attachment measure asks respondents about their own home communities, whereas the experimental condition visuals pertain to a fictional community (Keel), yet we still see consistent associations. It is interesting to consider why we might be seeing such effects. There seem to be two main possibilities: either the kinds of people who are more attached to their communities are also likely to respond positively to the connecting to place measures even in a hypothetical context, or the community of Keel is sufficiently similar to the communities that people live in that they “transfer” their attachment to this similar place that stands in for their home community in the experiment. Regardless of the means of transfer, it is interesting that higher place attachment leads to greater reported ability to relate to, make sense of, and understand risk information conveyed visually and to envision oneself in the place shown in the map.
Another noteworthy finding is that survey respondents with more local scales of place reported a greater ability to relate to the information shown in the map, regardless of the visual they received. This finding is consistent with the prior qualitative research by Wilhelmi et al. (2023) where study participants connected the relatability of information with maps presented at local scales; the combination of the two helped them better understand storm surge risk.
This study makes a methodological contribution through developing and empirically testing measures of connecting people to a place through visuals. Future research can build on these measures and connect them to information comprehension, risk perception, and protective behaviors. In future research, it may also be important to evaluate these questions in the real-world communities with the visuals representing those communities.
5. Conclusions
With the overarching goal of understanding how geovisualizations enable coastal residents to understand and respond to risk, this study used an interdisciplinary approach to create new knowledge about the effectiveness of geovisualizations in storm surge risk communication. We empirically explored the concepts of place meaning, scale of place, and place attachment in the context of coastal communities of Georgia and South Carolina. We conceptualized how geovisualizations may facilitate connection to a place, and thus increase the personal relevance of risk information, and developed measures of connecting to a place through visuals. We then created a series of geovisualizations incorporating aspects of place meaning, scale of place, dimensionality, and realism and tested the effectiveness of these visuals in a large population survey.
The geovisualizations developed in this study were generally effective in enabling people to connect to the place shown in the visuals; however, a 2D regional-scale map displayed together with a 3D abstract representation of a neighborhood was the most helpful in enabling people to relate to a place, quickly make sense of the information, and understand the risk. Our results showed that while geovisualizations of storm surge risk can be effective generally, they appear to be less effective in several important and vulnerable sociodemographic groups. We saw substantial impacts of race, income, map-reading ability, place attachment, and scale of place on how people connected the storm surge risk shown in the visual to a place. These findings have implications for how weather forecasters and emergency managers communicate storm surge risks and other hazards with diverse audiences through geovisualizations.
Acknowledgments.
Sadly, Dr. Heather Lazrus, an environmental anthropologist and a coauthor of this paper, passed away in February 2023 after a long battle with cancer. We had the incredible fortune to work with Heather, and we integrated her invaluable contributions to this study throughout the paper. Heather was a wonderful friend and colleague who will be missed dearly.
This research was supported by the National Science Foundation Award 1853699. The National Center for Atmospheric Research is sponsored by the National Science Foundation.
Data availability statement.
The geospatial data and deidentified survey data will be available to other researchers from the corresponding author upon reasonable request. The authors are in the process of developing a publicly available data archive, for data that can be shared in accordance with human subjects guidelines, which will be completed within two years of the project end date.
REFERENCES
Appleton, K., and A. Lovett, 2003: GIS-based visualisation of rural landscapes: Defining ‘sufficient’ realism for environmental decision-making. Landscape Urban Plann., 65, 117–131, https://doi.org/10.1016/S0169-2046(02)00245-1.
Ardoin, N. M., R. K. Gould, H. Lukacs, C. C. Sponarski, and J. S. Schuh, 2019: Scale and sense of place among urban dwellers. Ecosphere, 10, e02871, https://doi.org/10.1002/ecs2.2871.
Baker, E. J., 1995: Public response to hurricane probability forecasts. Prof. Geogr., 47, 137–147, https://doi.org/10.1111/j.0033-0124.1995.00137.x.
Bernhardt, J., J. Snellings, A. Smiros, I. Bermejo, A. Rienzo, and C. Swan, 2019: Communicating hurricane risk with virtual reality: A pilot project. Bull. Amer. Meteor. Soc., 100, 1897–1902, https://doi.org/10.1175/BAMS-D-17-0326.1.
Brewer, C. A., 2006: Basic mapping principles for visualizing cancer data using geographic information systems (GIS). Amer. J. Prev. Med., 2, S25–S36, https://doi.org/10.1016/j.amepre.2005.09.007.
Brown, B., D. D. Perkins, and G. Brown, 2003: Place attachment in a revitalizing neighborhood: Individual and block levels of analysis. J. Environ. Psychol., 23, 259–271, https://doi.org/10.1016/S0272-4944(02)00117-2.
Brown, G., C. M. Raymond, and J. Corcoran, 2015: Mapping and measuring place attachment. Appl. Geogr., 57, 42–53, https://doi.org/10.1016/j.apgeog.2014.12.011.
Brown, G., P. Reed, and C. M. Raymond, 2020: Mapping place values: 10 lessons from two decades of public participation GIS empirical research. Appl. Geogr., 116, 102156, https://doi.org/10.1016/j.apgeog.2020.102156.
Bruns, C. R., and B. C. Chamberlain, 2019: The influence of landmarks and urban form on cognitive maps using virtual reality. Landscape Urban Plann., 189, 296–306, https://doi.org/10.1016/j.landurbplan.2019.05.006.
Bukvic, A., A. Whittemore, J. Gonzales, and O. Wilhelmi, 2022: Understanding relocation in flood-prone coastal communities through the lens of place attachment. Appl. Geogr., 146, 102758, https://doi.org/10.1016/j.apgeog.2022.102758.
Cantrill, J. G., and S. L. Senecah, 2001: Using the ‘sense of self-in-place’ construct in the context of environmental policy-making and landscape planning. Environ. Sci. Policy, 4, 185–203, https://doi.org/10.1016/S1462-9011(01)00023-5.
Cao, Y., B. J. Boruff, and I. M. McNeill, 2016: Is a picture worth a thousand words? Evaluating the effectiveness of maps for delivering wildfire warning information. Int. J. Disaster Risk Reduct., 19, 179–196, https://doi.org/10.1016/j.ijdrr.2016.08.012.
Chamberlain, B. C., M. J. Meitner, and R. Ballinger, 2015: Applications of visual magnitude in forest planning: A case study. For. Chron., 91, 417–425, https://doi.org/10.5558/tfc2015-071.
Chamberlain, B. C., R. Liu, and J. Canfield, 2016: Using 3D visualizations to investigate perceptions of streetscape design. J. Digital Landscape Archit., 2016, 84–91.
Chamlee-Wright, E., and V. H. Storr, 2009: “There’s no place like New Orleans”: Sense of place and community recovery in the ninth ward after Hurricane Katrina. J. Urban Aff., 31, 615–634, https://doi.org/10.1111/j.1467-9906.2009.00479.x.
Chassin, T., J. Ingensand, S. Christophe, and G. Touya, 2022: Experiencing virtual geographic environment in urban 3D participatory e-planning: A user perspective. Landscape Urban Plann., 224, 104432, https://doi.org/10.1016/j.landurbplan.2022.104432.
Chen, Y.-H., S. E. Zick, and A. R. Benjamin, 2016: A comprehensive cartographic approach to evacuation map creation for Hurricane Ike in Galveston County, Texas. Cartogr. Geogr. Inf. Sci., 43, 68–85, https://doi.org/10.1080/15230406.2015.1014426.
Clive, M. A. T., J. M. Lindsay, G. S. Leonard, C. Lutteroth, A. Bostrom, and P. Corballis, 2021: Volcanic hazard map visualisation affects cognition and crisis decision-making. Int. J. Disaster Risk Reduct., 55, 102102, https://doi.org/10.1016/j.ijdrr.2021.102102.
Cresswell, T., 2004: Defining place. Place: A Short Introduction, Blackwell Ltd,. 1–18.
Cresswell, T., 2014: Place: An Introduction. John Wiley and Sons, 240 pp.
Demuth, J. L., 2015: Developing a valid scale of past tornado experiences. Ph.D. dissertation, Colorado State University, 195 pp., https://api.mountainscholar.org/server/api/core/bitstreams/df796588-a9b8-4957-b703-7683e65488e0/content.
Demuth, J. L., R. E. Morss, J. K. Lazo, and C. Trumbo, 2016: The effects of past hurricane experiences on evacuation intentions through risk perception and efficacy beliefs: A mediation analysis. Wea. Climate Soc., 8, 327–344, https://doi.org/10.1175/WCAS-D-15-0074.1.
Demuth, J. L., R. E. Morss, G. Wong-Parodi, A. B. Schumacher, J. J. Alland, D. Smith, N. Herbert, and H. D. Walpole, 2023: Understanding people’s evolving risk assessments and decisions during tropical cyclone threats: Design and implementation of a novel longitudinal, real-time survey methodology. NOAA Final Rep., 79 pp., https://repository.library.noaa.gov/view/noaa/50908.
DeVellis, R. F., 2011: Scale Development: Theory and Applications. 3rd ed. Sage Publications, 205 pp.
Dransch, D., H. Rotzoll, and K. Poser, 2010: The contribution of maps to the challenges of risk communication to the public. Int. J. Digital Earth, 3, 292–311, https://doi.org/10.1080/17538941003774668.
Flaming Pear, 2023: Flood 2 software. Flaming Pear Software, http://www.flamingpear.com/flood-2.html.
Forrester, J., B. Cook, L. Bracken, S. Cinderby, and A. Donaldson, 2015: Combining participatory mapping with Q-methodology to map stakeholder perceptions of complex environmental problems. Appl. Geogr., 56, 199–208, https://doi.org/10.1016/j.apgeog.2014.11.019.
Fossell, K. R., D. Ahijevych, R. E. Morss, C. Snyder, and C. Davis, 2017: The practical predictability of storm tide from tropical cyclones in the Gulf of Mexico. Mon. Wea. Rev., 145, 5103–5121, https://doi.org/10.1175/MWR-D-17-0051.1.
Houston, D., and Coauthors, 2019: The influence of hazard maps and trust of flood controls on coastal flood spatial awareness and risk perception. Environ. Behav., 51, 347–375, https://doi.org/10.1177/0013916517748711.
Huang, S.-K., M. K. Lindell, and C. S. Prater, 2016: Who leaves and who stays? A review and statistical meta-analysis of hurricane evacuation studies. Environ. Behav., 48, 991–1029, https://doi.org/10.1177/0013916515578485.
Jamali, M., and A. Nejat, 2016: Place attachment and disasters: Knowns and unknowns. J. Emerg. Manage., 14, 349–364, https://doi.org/10.5055/jem.2016.0299.
Kellams, T. R., 2006: The mind, the narrative, and the city: How narratives of space make place in cognitive maps. M.S. thesis, Dept. of Landscape Architecture and Regional and Community Planning, Kansas State University, 189 pp., https://krex.k-state.edu/handle/2097/35517.
Keller, C., M. Siegrist, and H. Gutscher, 2006: The role of the affect and availability heuristics in risk communication. Risk Anal., 26, 631–639, https://doi.org/10.1111/j.1539-6924.2006.00773.x.
Kianicka, S., M. Buchecker, M. Hunziker, and U. Müller-Böker, 2006: Locals’ and tourists’ sense of place. Mt. Res. Dev., 26, 55–63, https://doi.org/10.1659/0276-4741(2006)026[0055:LATSOP]2.0.CO;2.
Kick, E. L., J. C. Fraser, G. M. Fulkerson, L. A. McKinney, and D. H. De Vries, 2011: Repetitive flood victims and acceptance of FEMA mitigation offers: An analysis with community–system policy implications. Disasters, 35, 510–539, https://doi.org/10.1111/j.1467-7717.2011.01226.x.
Knigge, L., and M. Cope, 2006: Grounded visualization: Integrating the analysis of qualitative and quantitative data through grounded theory and visualization. Environ. Plann., 38A, 2021–2037, https://doi.org/10.1068/a37327.
Kostelnick, J. C., D. McDermott, R. J. Rowley, and N. Bunnyfield, 2013: A cartographic framework for visualizing risk. Cartographica, 48, 200–224, https://doi.org/10.3138/carto.48.3.1531.
Kudryavtsev, A., R. C. Stedman, and M. E. Krasny, 2012: Sense of place in environmental education. Environ. Educ. Res., 18, 229–250, https://doi.org/10.1080/13504622.2011.609615.
La Rosa, D., M. Spyra, and L. Inostroza, 2016: Indicators of cultural ecosystem services for urban planning: A review. Ecol. Indic., 61, 74–89, https://doi.org/10.1016/j.ecolind.2015.04.028.
Lazo, J. K., A. Bostrom, R. E. Morss, J. L. Demuth, and H. Lazrus, 2015: Factors affecting hurricane evacuation intentions. Risk Anal., 35, 1837–1857, https://doi.org/10.1111/risa.12407.
Lazrus, H., O. Wilhelmi, R. Morss, J. Henderson, and A. Dietrich, 2020: Information as intervention: How hurricane risk communication interacted with vulnerability and capacities in Hurricane Sandy. Int. J. Mass Emerg. Disasters, 38, 89–120, https://doi.org/10.1177/028072702003800106.
Lemée, C., and Coauthors, 2022: Impact of risk experience and personal exposure on coastal flooding and coastal erosion risk perception and coping strategies. J. Risk Res., 25, 681–696, https://doi.org/10.1080/13669877.2021.1962952.
Lewicka, M., 2011: Place attachment: How far have we come in the last 40 years? J. Environ. Psychol., 31, 207–230, https://doi.org/10.1016/j.jenvp.2010.10.001.
Lindell, M. K., 2020: Improving hazard map comprehension for protective action decision making. Front. Comput. Sci., 2, 27, https://doi.org/10.3389/fcomp.2020.00027.
Lindell, M. K., and R. W. Perry, 2012: The protective action decision model: Theoretical modifications and additional evidence. Risk Anal., 32, 616–632, https://doi.org/10.1111/j.1539-6924.2011.01647.x.
Lindner, B. L., F. Alsheimer, and J. Johnson, 2018: Assessing improvement in the public’s understanding of hurricane storm tides through interactive visualization models. J. Coastal Res., 35, 130–142, https://doi.org/10.2112/JCOASTRES-D-17-00223.1.
MacPherson-Krutsky, C. C., B. D. Brand, and M. K. Lindell, 2020: Does updating natural hazard maps to reflect best practices increase viewer comprehension of risk? Int. J. Disaster Risk Reduct., 46, 101487, https://doi.org/10.1016/j.ijdrr.2020.101487.
Maguire, G. S., K. K. Miller, M. A. Weston, and K. Young, 2011: Being beside the seaside: Beach use and preferences among coastal residents of south-eastern Australia. Ocean Coastal Manage., 54, 781–788, https://doi.org/10.1016/j.ocecoaman.2011.07.012.
Maidl, E., and M. Buchecker, 2015: Raising risk preparedness through flood risk communication. Nat. Hazards Earth Syst. Sci., 15, 1577–1595, https://doi.org/10.5194/nhess-15-1577-2015.
Markwart, H., J. Vitera, S. Lemanski, D. Kietzmann, M. Brasch, and S. Schmidt, 2019: Warning messages to modify safety behavior during crisis situations: A virtual reality study. Int. J. Disaster Risk Reduct., 38, 101235, https://doi.org/10.1016/j.ijdrr.2019.101235.
Masters, J., 2022: The bizarre and destructive 2022 hurricane season ends. Yale Climate Connections, accessed 30 November 2022, https://yaleclimateconnections.org/2022/11/the-bizarre-and-destructive-2022-hurricane-season-ends/.
Masterson, V. A., R. C. Stedman, J. Enqvist, M. Tengö, M. Giusti, D. Wahl, and U. Svedin, 2017: The contribution of sense of place to social-ecological systems research: A review and research agenda. Ecol. Soc., 22, 49, https://doi.org/10.5751/ES-08872-220149.
Merschdorf, H., and T. Blaschke, 2018: Revisiting the role of place in geographic information science. ISPRS Int. J. Geo-Inf., 7, 364, https://doi.org/10.3390/ijgi7090364.
Meyer, R. J., J. Baker, K. Broad, J. Czajkowski, and B. Orlove, 2014: The dynamics of hurricane risk perception: Real-time evidence from the 2012 Atlantic hurricane season. Bull. Amer. Meteor. Soc., 95, 1389–1404, https://doi.org/10.1175/BAMS-D-12-00218.1.
Montello, D. R., K. L. Lovelace, R. G. Golledge, and C. M. Self, 1999: Sex‐related differences and similarities in geographic and environmental spatial abilities. Ann. Assoc. Amer. Geogr., 89, 515–534, https://doi.org/10.1111/0004-5608.00160.
Morss, R. E., J. L. Demuth, J. K. Lazo, K. Dickinson, H. Lazrus, and B. H. Morrow, 2016: Understanding public hurricane evacuation decisions and responses to forecast and warning messages. Wea. Forecasting, 31, 395–417, https://doi.org/10.1175/WAF-D-15-0066.1.
Morss, R. E., and Coauthors, 2017: Hazardous weather prediction and communication in the modern information environment. Bull. Amer. Meteor. Soc., 98, 2653–2674, https://doi.org/10.1175/BAMS-D-16-0058.1.
Morss, R. E., C. L. Cuite, J. L. Demuth, W. K. Hallman, and R. L. Shwom, 2018: Is storm surge scary? The influence of hazard, impact, and fear-based messages and individual differences on responses to hurricane risks. Int. J. Disaster Risk Reduct., 30, 44–58, https://doi.org/10.1016/j.ijdrr.2018.01.023.
Newcombe, N. S., and N. C.-R. Chiang, 2007: Learning geographical information from hypothetical maps. Mem. Cognit., 35, 895–909, https://doi.org/10.3758/BF03193464.
Newell, R., and R. Canessa, 2018: From sense of place to visualization of place: Examining people-place relationships for insight on developing geovisualizations. Heliyon, 4, e00547, https://doi.org/10.1016/j.heliyon.2018.e00547.
Quinn, T., I. Lorenzoni, W. N. Adger, K. Obrien, and E. Selboe, 2015: Place attachment, identity, and adaptation. The Adaptive Challenge of Climate Change, K. O’Brien and E. Selboe, Eds., Cambridge University Press, 160–170.
Quinn, T., F. Bousquet, and C. Guerbois, 2019: Changing places: The role of sense of place in perceptions of social, environmental and overdevelopment risks. Global Environ. Change, 57, 101930, https://doi.org/10.1016/j.gloenvcha.2019.101930.
Rapoport, A., 2016: Human Aspects of Urban Form: Towards a Man—Environment Approach to Urban Form and Design. Elsevier, 448 pp.
Raymond, C. M., G. Brown, and D. Weber, 2010: The measurement of place attachment: Personal, community, and environmental connections. J. Environ. Psychol., 30, 422–434, https://doi.org/10.1016/j.jenvp.2010.08.002.
Raymond, C. M., M. Kyttä, and R. Stedman, 2017: Sense of place, fast and slow: The potential contributions of affordance theory to sense of place. Front. Psychol., 8, 1674, https://doi.org/10.3389/fpsyg.2017.01674.
Retchless, D. P., 2014: Sea level rise maps: How individual differences complicate the cartographic communication of an uncertain climate change hazard. Cartogr. Perspect., 77, 17–32, https://doi.org/10.14714/CP77.1235.
Retchless, D. P., 2018: Understanding local sea level rise risk perceptions and the power of maps to change them: The effects of distance and doubt. Environ. Behav., 50, 483–511, https://doi.org/10.1177/0013916517709043.
Rickard, L. N., J. P. Schuldt, G. M. Eosco, C. W. Scherer, and R. A. Daziano, 2017: The proof is in the picture: The influence of imagery and experience in perceptions of hurricane messaging. Wea. Climate Soc., 9, 471–485, https://doi.org/10.1175/WCAS-D-16-0048.1.
Roy, K. C., S. Hasan, O. I. Abdul-Aziz, and P. Mozumder, 2022: Understanding the influence of multiple information sources on risk perception dynamics and evacuation decisions: An agent-based modeling approach. Int. J. Disaster Risk Reduct., 82, 103328, https://doi.org/10.1016/j.ijdrr.2022.103328.
Scannell, L., and R. Gifford, 2010: Defining place attachment: A tripartite organizing framework. J. Environ. Psychol., 30, 1–10, https://doi.org/10.1016/j.jenvp.2009.09.006.
Severtson, D. J., and C. Vatovec, 2012: The theory-based influence of map features on risk beliefs: Self-reports of what is seen and understood for maps depicting an environmental health hazard. J. Health Commun., 17, 836–856, https://doi.org/10.1080/10810730.2011.650933.
Sheppard, S. R. J., 2012: Visualizing Climate Change: A Guide to Visual Communication of Climate Change and Developing Local Solutions. Routledge, 528 pp.
Sheppard, S. R. J., 2015: Making climate change visible: A critical role for landscape professionals. Landscape Urban Plann., 142, 95–105, https://doi.org/10.1016/j.landurbplan.2015.07.006.
Sheppard, S. R. J., A. Shaw, D. Flanders, and S. Burch, 2008: Can visualization save the world? Lessons for landscape architects from visualizing local climate change. Digital Design in Landscape Architecture, E. Buhmann et al., Eds., Herbert Wichmann Verlag, 29–31.
Sherman-Morris, K., K. B. Antonelli, and C. C. Williams, 2015: Measuring the effectiveness of the graphical communication of hurricane storm surge threat. Wea. Climate Soc., 7, 69–82, https://doi.org/10.1175/WCAS-D-13-00073.1.
Simpson, M., L. Padilla, K. Keller, and A. Klippel, 2022: Immersive storm surge flooding: Scale and risk perception in virtual reality. J. Environ. Psychol., 80, 101764, https://doi.org/10.1016/j.jenvp.2022.101764.
Stedman, R. C., 2008: What do we mean by place meanings? Implications of place meanings for managers and practitioners. Understanding Concepts of Place in Recreational Research and Management, L. E. Kruger, T. E. Hall, and M. C. Stiefel, Eds., Vol. PNW-GTR-744, USDA Forest Service, Pacific Northwest Research Station, 61–82.
Steele, F., 1981: The Sense of Place. CBI Publishing, 216 pp.
Stempel, P. J., and A. Becker, 2021: Is it scientific? Viewer perceptions of storm surge visualizations. Cartographica, 56, 120–136, https://doi.org/10.3138/cart-2020-0004.
Suchan, T. A., and C. A. Brewer, 2000: Qualitative methods for research on mapmaking and map use. Prof. Geogr., 52, 145–154, https://doi.org/10.1111/0033-0124.00212.
Sutton, J., and C. Woods, 2016: Tsunami warning message interpretation and sense making: Focus group insights. Wea. Climate Soc., 8, 389–398, https://doi.org/10.1175/WCAS-D-15-0067.1.
Sutton, J., and L. M. Fischer, 2021: Understanding visual risk communication messages: An analysis of visual attention allocation and think-aloud responses to tornado graphics. Wea. Climate Soc., 13, 173–188, https://doi.org/10.1175/WCAS-D-20-0042.1.
Tanim, S. H., B. M. Wiernik, S. Reader, and Y. Hu, 2022: Predictors of hurricane evacuation decisions: A meta-analysis. J. Environ. Psychol., 79, 101742, https://doi.org/10.1016/j.jenvp.2021.101742.
Thornberg, R., 2012: Informed grounded theory. Scand. J. Educ. Res., 56, 243–259, https://doi.org/10.1080/00313831.2011.581686.
Tomkins, A., and E. Lange, 2019: Interactive landscape design and flood visualisation in augmented reality. Multimodal Technol. Interact., 3, 43, https://doi.org/10.3390/mti3020043.
Trentelman, C. K., 2009: Place attachment and community attachment: A primer grounded in the lived experience of a community sociologist. Soc. Nat. Resour., 22, 191–210, https://doi.org/10.1080/08941920802191712.
Tuan, Y.-F., 1974: Space and place: Humanistic perspective. Prog. Hum. Geogr., 6, 233–246.
Tuan, Y.-F., 1977: Space and Place: The Perspective of Experience. Edward Arnold, 226 pp.
Unreal Engine, 2018: The weather channel breaks new ground with immersive mixed reality unreal engine. YouTube, https://www.youtube.com/watch?v=x2aCSV5zYlA&t=123s.
Walpole, H. D., and R. S. Wilson, 2021: A yardstick for danger: Developing a flexible and sensitive measure of risk perception. Risk Anal., 41, 2031–2045, https://doi.org/10.1111/risa.13704.
Wilhelmi, O. V., R. E. Morss, H. Lazrus, J. Boehnert, and J. Gambill, 2023: Examining the roles of visualizations in people’s understanding of uncertain storm surge forecasts, interpretation of risk, and decision-making. SSRN, 4442115, https://doi.org/10.2139/ssrn.4442115.
Williams, D. R., 2014: Making sense of ‘place’: Reflections on pluralism and positionality in place research. Landscape Urban Plann., 131, 74–82, https://doi.org/10.1016/j.landurbplan.2014.08.002.
Williams, D. R., and S. I. Stewart, 1998: Sense of place: An elusive concept that is finding a home in ecosystem management. For. Sci., 96, 18–23.
Williams, D. R., and J. J. Vaske, 2003: The measurement of place attachment: Validity and generalizability of a psychometric approach. For. Sci., 49, 830–840.
Zachry, B. C., W. J. Booth, J. R. Rhome, and T. M. Sharon, 2015: A national view of storm surge risk and inundation. Wea. Climate Soc., 7, 109–117, https://doi.org/10.1175/WCAS-D-14-00049.1.