1. Introduction
a. Tornado hazards, social vulnerability, and Latinx/e populations in the Southeast United States
The United States ranks the highest among the countries with the most frequent occurrences of tornadoes globally due to meteorological and topographical conditions that create an environment that is conducive to the formation of severe thunderstorms and tornadoes (Guo et al. 2016; Lu et al. 2015). Within the United States, previous studies on the geographic distribution of tornado events have found the Southeast region, in particular, has a high risk for tornadoes and tornado-related losses (e.g., fatalities, injuries, damages) due to a combination of physical and social variables, such as low visibility (e.g., harder to see tornadoes due to forest cover and occurring at night) and inadequate shelters (e.g., fewer basements and tornado shelters, a high percentage of mobile homes; Sutter and Simmons 2010; Strader et al. 2018, 2022), and the high frequency of nocturnal tornadoes which have been found to be 2.5 times as deadly (Ashley et al. 2008). Furthermore, in the last decade, Latinx/e1 populations (immigrant and native-born) have grown in the U.S. Southeast, with some counties tripling in the last 10 years (U.S. Census 2020; see Fig. 1). This growth also brings attention to the diversity of this population, particularly in terms of culture, language, backgrounds, and experiences. Coming from various backgrounds and experiences, Latinx/e individuals may immigrate from various places in Latin America and the Caribbean and may be exposed to severe weather that they may have never experienced in their home country (e.g., tornadoes, winter weather).
(a) Percent of Latinx/e population per capita from the 2020 census and (b) relative risk of a strong (EF2+) tornado in Southeast counties from 1950 to 2019.
Citation: Weather, Climate, and Society 17, 1; 10.1175/WCAS-D-24-0010.1
Prior research has found that Latinx/e populations, in particular, encounter various social, cultural, and political barriers that increase their vulnerability to harm during tornado hazards (First et al. 2021; Gaviria Pabon 2022; Trujillo-Falcón et al. 2021, 2024). A population’s vulnerability to a tornado hazard is not based solely on its proximity to the event, but rather, the impact varies significantly depending on the vulnerability of various groups within the affected population. Social vulnerability is a multidimensional issue that is shaped by physical, social, cultural, economic, and political factors that determine an individual’s or community’s ability to prepare for, respond to, and adapt to hazards (e.g., Cutter et al. 2003; Strader et al. 2017; Morss et al. 2011). This perspective recognizes that people are not born vulnerable, but rather, they are made vulnerable by interactions between individual and contextual factors.
Regarding social vulnerability, language access has been found to significantly influence the vulnerability of Latinx/e populations during severe weather. For instance, the majority of severe weather forecasts and warnings in the United States are in English, and therefore, individuals with limited English proficiency (LEP) are less likely to receive and understand tornado warnings and, therefore, may be prevented from taking life-saving protective actions. One tragic example occurred during the 2013 tornado outbreak in El Reno, Oklahoma; a Guatemalan, Spanish-speaking family of seven was swept away by flash floods and killed after taking shelter from the tornado in a storm drain (NWS 2014). A NOAA/NWS assessment of El Reno event found there were insufficient forecast/warning resources available in Spanish which prevented individuals and groups from taking life-saving protective measures (NWS 2014). Several studies have focused on understanding the levels of tornado warning and watch comprehension among English and Spanish-speaking populations, highlighting challenges and disparities experienced by Spanish-speaking populations (Aguirre 1988; Ahlborn et al. 2012; Donner et al. 2012; First et al. 2021; Gaviria Pabon 2022; Senkbeil et al. 2014; Trujillo-Falcón et al. 2021, 2022). For example, Ahlborn et al. (2012) examined the effectiveness of tornado warning systems for Spanish-speaking populations in Oklahoma, and their study revealed that only 25.6% of native Spanish speakers reported being fluent in English, significantly lower than U.S. Census estimates. Additionally, Spanish speakers had less access to critical warning systems, including the internet and local telephone warning programs (Ahlborn et al. 2012). In a national-level study, Trujillo-Falcón et al. (2022) surveyed U.S. English (N = 1550) and Spanish (N = 1050) speakers to assess their ability to interpret weather watch and warning translations. They found significant language disparities in the reception and comprehension of severe weather warnings for Spanish speakers in comparison to English speakers. Furthermore, their study revealed that Spanish-speaking individuals often faced barriers in understanding the urgency and instructions conveyed in weather warnings, which could be attributed to language differences and the lack of culturally relevant communication strategies (Trujillo-Falcón et al. 2022).
Additional research has also underscored the importance of assessing primary weather information sources for Latinx/e communities. For example, Bitterman et al. (2024) conducted a nationwide survey to assess differences in severe weather information sources between English and Spanish speakers in the United States. Their study found that Spanish-speaking participants predominantly relied on local Spanish-language television (TV) and radio, social media, weather apps, and community networks for severe weather information. In contrast, English-speaking respondents used a broader mix of traditional media, official websites, and digital platforms (Bitterman et al. 2024). Furthermore, prior studies have found that trust in the sources providing hazard information is a fundamental factor in determining whether individuals and communities take protective actions during tornado events. In particular, Latinx/e immigrants who have undocumented legal status face a persistent risk of immigration enforcement activity (immigrant detention or deportation) and are disproportionately affected by racial discrimination, exploitation, and economic hardships—each of which may limit the ability to trust public officials’ warnings (Hatzenbuehler et al. 2017; Méndez et al. 2020; Trujillo-Falcón et al. 2024). Due to fear of deportation and discrimination in the Southeast (Held et al. 2022), many Latinx/e communities may be hesitant to trust weather information from public officials and government agencies, and this fear can result in delayed responses to seek protection during hazardous events. For instance, the U.S. Southeast has some of the most exclusionary immigrant policies and has experienced the two largest immigration raids in the previous ten years (Burke 2018; Hatzenbuehler et al. 2017; U.S. Immigration and Customs Enforcement 2019). Prior work conducted by Senkbeil et al. (2012) in the U.S. Southeast identified significant differences in tornado risk perception, preparedness, and shelter lead time among Latinx/e and White respondents. They also noted that following the Tuscaloosa tornado in 2011, many Latinx/e residents declined to be interviewed about their experiences due to fears about current immigration laws (Senkbeil et al. 2012). This reluctance highlights the significant impact of immigration laws on exacerbating vulnerability, as fear of deportation or detention can prevent undocumented immigrants from seeking assistance, utilizing shelters, and sharing their experiences, which further leads to underreporting of their experiences, essentially making them an “invisible” population group in weather hazard preparedness and response planning (Méndez et al. 2020).
Finally, many Latin American countries do not experience tornado hazards as frequently as the United States, and when Latinx/e immigrants might arrive in the United States, they may struggle to relate to and personalize weather information within their new environment and could perceive their personal risk to be lower than those who have prior experience (Klockow et al. 2014). For example, Gaviria Pabon (2022) found following the 10 December 2021 tornado in Mayfield, Kentucky, that immigrant Guatemalan residents indicated having a lack of experience with tornadoes in their home country of Guatemala alongside language barriers limited their ability to take protective action. Trujillo-Falcón et al. (2024) also found that Latinx/e immigrant populations were disproportionately affected by the 10 December 2021 tornado due to social and systemic vulnerabilities, such as English proficiency and varying cultures of disaster preparedness, which created barriers to inclusive disaster response.
b. Study goals
While Latinx/e populations are one of the fastest-growing minoritized populations in the Southeast, few studies have examined various social, cultural, and political factors that increase their vulnerability to harm during tornadoes (Gaviria Pabon 2022; Trujillo-Falcón et al. 2021). Furthermore, there is a historical tendency to analyze ethnicity as broad, monolithic categories in social vulnerability and tornado hazard research. For instance, the labels “Hispanic,” “Latino,” or “Latinx/e” have often been employed as an umbrella term to refer to a diverse group of people who share a common cultural, linguistic, or ancestral connection to Spanish-speaking countries or Latin America. While this approach can offer important broad insights, it sometimes inadvertently assumes a level of homogeneity within a population group that does not necessarily reflect the complex reality of various groups of cultures and dialects (Gaviria Pabon 2022; Held and Lee 2017; Trujillo-Falcón et al. 2021). People from the same ethnic or racial background may face distinct challenges and exhibit varying levels of preparedness and awareness in the face of tornado hazards. Intraethnic disparities or vulnerabilities refer to axes of difference in experiences, outcomes, or access to resources within a single ethnic group that may impact an individual’s or community’s ability to prepare for, respond to, and adapt to hazards (Grineski et al. 2013).
For Latinx/e populations in tornado-vulnerable areas, receiving warning information and knowing what protective action to take is essential for seeking protection during a tornado event. The current study examined tornado warning and watch comprehension, severe weather information sources, and barriers in tornado warning communication and response among 820 Latinx/e adults (immigrant and native-born) across 11 states in the U.S. Southeast and examined intraethnic heterogeneity that could potentially exacerbate tornado vulnerability. This study adds to research previously conducted by Ahlborn et al. (2012), Gaviria Pabon (2022), First et al. (2021), and Trujillo-Falcón et al. (2021, 2022, 2024), which has examined the connections between ethnicity, race, immigration status, nativity (where one was born—to indicate immigrant or native born), geographical location, and tornado reception and response. In addition, this study builds upon work conducted by Smith et al. (2023), who conducted a nationally representative study using data from the University of Oklahoma’s Severe Weather and Society Survey and data from the annual Federal Emergency Management Agency’s (FEMA) National Household Survey to examine racial and ethnic disparities in severe weather response and preparedness and risk perceptions in the United States. Their study identified important severe weather risk prevalence statistics for adults by race, ethnicity, and socioeconomic characteristics but identified the need for additional finer-scale and heterogeneous studies to go beyond analyzing race and ethnicity as monolithic categories in social vulnerability and hazard research. Our study, in turn, examined intraethnic disparities in tornado hazard understanding, reception, and response among Latinx/e adults in the Southeast. As the weather enterprise works to understand information ecosystems in support of weather-related decision-making among vulnerable and underrepresented populations, understanding intraethnic differences in human vulnerability in the face of tornado threats can help inform better disaster planning and response among vulnerable and underrepresented populations in the U.S. Southeast. Our research questions include the following:
-
(i) What are the levels of tornado warning and watch comprehension among Latinx/e adults in the U.S. Southeast, and (ii) to what extent do differences exist in tornado warning and watch comprehension based on factors such as language, nativity, and citizenship status?
-
(i) What are the primary sources of severe weather information for Latinx/e adults in the U.S. Southeast, and (ii) to what extent do differences exist in information sources based on language, nativity, and citizenship status?
-
(i) What barriers in tornado warning communication and response do Latinx/e adults in the U.S. Southeast encounter, and (ii) to what extent do differences exist in barriers based on language, nativity, and citizenship status?
2. Methods
a. Participant demographics
Latinx/e adults age 18 and older living in the Southeast United States [Alabama (AL), Florida (FL), Georgia (GA), Kentucky (KY), Mississippi (MS), North Carolina (NC), South Carolina (SC), Tennessee (TN), Virginia (VA), Los Angeles (LA), and West Virginia (WV)] were recruited from the National Opinion Research Center’s (NORC) LatinoSpeak panel. NORC uses a multistage probability sampling frame to recruit a panel that is representative of the U.S. household population. To recruit English- and Spanish-speaking survey participants for the LatinoSpeak panel, NORC contacts U.S. households by mail, telephone, email, and using field interviewers that conduct in-person follow-up recruitment in both English and Spanish for reaching harder-to-reach population groups (e.g., immigrant, low-income households). NORC’s 2147 Latinx/e panelists were invited to complete the survey by telephone or internet, and a total of 820 completed the survey for a 38.2% response rate.
Prior to the current survey being fielded, panelists were first asked a screening question to self-report their English language proficiency. LEP is a phrase officially defined by the federal government to classify “individuals who do not speak English as their primary language and who have a limited ability to read, speak, write, or understand English” (U.S. Department of Justice 2011). Panelists who indicate they speak English “very well” were then invited to take the survey in English. Panelists who did not indicate they speak English very well are then asked a follow-up question: “Do you speak Spanish?” Panelists who indicated they speak Spanish very well were then invited to complete the survey in Spanish. According to established definitions (U.S. Department of Justice 2011), participants who completed the survey in Spanish were classified as having LEP, as they self-reported speaking English less than very well. However, it is important to note that our sample includes only a specific subset of LEP Latinx/e participants—those who are highly proficient in Spanish.
The survey was fielded from 8 to 21 May 2023. Of these completed surveys, 820 were collected online and 8 by phone, and 625 were collected in English and 195 in Spanish. Email reminders were sent to panelists on 8, 9, 11, 16, 21, 25, and 28 May to encourage participation. The median completion time for the survey was 14 min, and participants received cash or points for prizes from NORC for completing the survey. The study was conducted in accordance with the ethical guidelines set forth by the University of Tennessee. Informed consent was obtained from all participants, ensuring their voluntary participation, anonymity, and the right to withdraw at any point during the study.
A majority of participants were identified as female (55.6%), were born in the United States (66.0%), lived in the South Atlantic (89.9%), were English speaking (76.3%), and a U.S. citizen (89.9%). The majority were identified as Puerto Rican (24.7%), followed by Cuban (21.8%), and Mexican, Mexican–American, Chicano (18.7%), South American (10.5%), Central American (8.0%), and Caribbean (4.3%). See Table 1 for participant demographics.
Participant demographic information.
b. Survey design
The survey contained several questions aimed at understanding Latinx/e respondents’ awareness of tornado alerts (tornado watch and warning), their sources of weather information during severe weather situations, potential barriers to seeking protection during a tornado, and protective decision-making processes. The current study focuses on questions pertaining to tornado watch/warning comprehension, sources of severe weather information, and barriers to protective action during a tornado. Building upon prior work conducted by Trujillo-Falcón et al. (2022), the first survey question (Q1A and 1B) asked respondents about their objective comprehension of NWS tornado alerts, specifically whether they can correctly differentiate between a tornado watch and a tornado warning. Next, survey questions 2A–2K inquired about the sources from which respondents typically receive weather information during severe weather events. These sources included options such as television, radio, wireless emergency alerts, landline phones, social media, outdoor warning sirens, NOAA Weather Radio, family and friends, school staff, churches/faith-based leaders, and community organizations. Last, survey questions 3A–3E focused on identifying potential barriers that might prevent Latinx/e individuals from seeking protection during a tornado threat. These barriers include a lack of information in Spanish, knowledge of the tornado’s location, understanding of what actions to take, access to transportation to reach shelter, and the availability of a safe shelter location. Each question asked respondents to rate the extent to which these barriers affect their decision-making on a scale from 1 to 3 (1 = not at all; 2 = somewhat; 3 = a great deal). The survey was developed in English and then was translated to Spanish by two graduate research assistants through a back-translation process. First, a native Spanish-speaking research assistant who has extensive research experience with Latinx/e immigrants in the Southeast translated all materials from English to Spanish. Following the initial translation, a second research assistant, who is also a native Spanish speaker, independently translated the Spanish materials back into English. Prior to data collection, both the English and Spanish surveys were pilot tested by NORC for clarity with 30 participants. See Table 2 for survey questions and response options in English and Spanish.
Survey questions.
c. Analyses
All data analyses were conducted using International Business Machines Corporation (IBM) Statistical Package for the Social Sciences (SPSS), version 29. First, a descriptive analysis of response frequencies and percentages was conducted to examine tornado warning comprehension, tornado information sources, and barriers to seeking protection from tornado threats across all participants (N = 820). Second, for explanatory variables with ordinal categories, we employed the nonparametric Wilcoxon–Mann–Whitney (WMW) test to evaluate two independent ethnic group categories (Steele et al. 2022) that included participants’ English language proficiency, place of birth (born in the United States or not born in the United States), and citizenship (U.S. citizen or non-U.S. citizen) to examine whether there were statistically significant differences in tornado warning comprehension, tornado information sources, and barriers to seeking protection from tornado threats. The Mann–Whitney U test was chosen because it does not assume a normal distribution or homogeneity of variances within the data, which are required by parametric tests. Statistical significance was set at p < 0.05 for all analyses and we rejected the null hypothesis if the p value was less than 0.05.
3. Results
a. Tornado watch/warning comprehension
In response to research question 1i (i.e., identifying tornado warning and watch comprehension among Latinx/e adults in the U.S. Southeast), the results found that 71.4% of the total sample demonstrated the ability to correctly identify a tornado warning when presented with the provided descriptions. Additionally, 65.2% of the total sample accurately identified a tornado watch based on the provided descriptions. These results indicate moderate recognition and comprehension of tornado-related warnings and watches within the surveyed population of Latinx/e participants. However, in response to research question 1ii (i.e., identifying differences in a tornado warning and watch comprehension based on factors such as language, nativity, and citizenship status), our results found several differences across groups or intraethnic heterogeneity in tornado watch and warning comprehension among respondents. In terms of tornado watch comprehension, the results found the highest watch comprehension difference between LEP and non-LEP respondents. The results found that non-LEP respondents demonstrated a higher watch comprehension rate (73.7%) compared to LEP respondents (40.4%). U.S. citizens also exhibited higher watch comprehension (68.9%) than non-U.S. citizens (50.0%), and individuals born in the United States displayed higher watch comprehension (71.8%) than those born outside the United States (57.0%). For tornado warning comprehension, non-LEP respondents demonstrated a higher warning comprehension rate (75.3%) compared to LEP respondents (54.1%). Similarly, U.S. citizens exhibited better warning comprehension (72.1%) than non-U.S. citizens (62.1%). When considering place of birth, individuals born in the United States also displayed higher warning comprehension (72.9%) than those born outside the United States (66.7%). See Fig. 2 for percentages of correct tornado watch and warning responses across groups.
Percentages of correct tornado watch and warning responses across groups.
Citation: Weather, Climate, and Society 17, 1; 10.1175/WCAS-D-24-0010.1
b. Sources of severe weather information
Sources of severe weather information were analyzed in response to the second survey question (Q2 A–K). First, in response to research question 2i (what are the primary sources of severe weather information for Latinx/e adults in the U.S. Southeast?), the results found the top three information sources included wireless emergency alerts (smartphone alerts) with the highest mean score of 2.65 (±0.59), followed by television (e.g., TV news, emergency broadcast system announcements) with a mean score of 2.40 (±0.68), and family, friends, and neighbors, received a mean score of 2.14 (±0.70). The results found the lowest mean for information sources included school staff who received a mean score of 1.74 (±0.77), community organizations and churches/faith-based leaders, which had mean scores of 1.66 (±0.78) and 1.56 (±0.73), and landline phone (e.g., reverse—911) which received the lowest mean score of 1.54 (±0.78). See Table 3 for severe weather information sources across all participants.
Severe weather information sources across all participants. Mean scores and ±standard deviations (±SD) are indicated for sources of information.
Next, when we analyzed intraethnic heterogeneity using the Wilcoxon–Mann–Whitney test differences across survey language, nativity, and citizenship status, we observed significant variations in the mean ranks between LEP and non-LEP respondents in the context of severe weather information sources in response to research question 2ii (i.e., identifying differences in information sources based on English language proficiency, nativity, and citizenship status). Specifically, LEP respondents ranked television (p = 0.031), amplitude modulation (AM)/frequency modulation (FM) commercial radio (p < 0.001), landline phone (p < 0.001), NOAA Weather Radio (p = 0.018), school staff (p < 0.001), churches/faith-based leaders (p < 0.001), and community organizations (p = 0.003) higher compared to their non-LEP counterparts. In addition, when analyzing the data by U.S. citizenship status, non-U.S. citizens had significant differences in the rankings of school staff (p < 0.001) and churches/faith-based leaders (p = 0.002) as important severe weather information sources compared to U.S. citizens. Furthermore, a statistically significant difference in the preference for obtaining weather information from NOAA Weather Radio was found between individuals born outside the United States and those born in the United States, with a higher preference among those born in the United States compared to those born outside the United States. See Tables 4a–c for intraethnic heterogeneity among Latinx/e sources of severe weather information.
Comparison of LEP and Non-LEP responses for sources of information using the WMW test. Only significant differences are listed with a p value < 0.05. Confidence intervals (CIs) are provided in paratheses.
Comparison by U.S. citizenship for sources of information using the WMW test. Only significant differences are listed with a p value < 0.05. CIs are provided in paratheses.
Comparison by birthplace for sources of information using the WMW test. Only significant differences are listed with a p value < 0.05. CIs are provided in paratheses.
c. Barriers to tornado threat communication and response
Finally, potential barriers that might prevent Latinx/e individuals from seeking protection during a tornado threat were analyzed. First, in response to research question 3i (i.e., identifying barriers in tornado warning communication and response that Latinx/e adults in the U.S. Southeast encounter), results found identified barriers included lack of transportation to get to shelter with the highest mean score of 1.70 (±0.75), followed by lack of a safe place to shelter with a mean score of 1.66 (±0.76), lack of knowledge on where the tornado is located with a mean score of 1.61 (±0.68), lack of knowledge on what to do with a mean score of 1.55 (±0.70), and lack of information in Spanish with a mean score of 1.44 (±0.68). See Table 5 for identified barriers to tornado threat communication and response across all participants.
Identified barriers to tornado threat communication and response across all participants. Mean scores and ±standard deviations (±SD) are indicated for sources of information.
In response to research question 3ii (i.e., identifying differences in barriers based on language, nativity, and citizenship status), Wilcoxon–Mann–Whitney test results found heterogeneity across English proficiency and U.S. citizenship status. First, we observed significant differences in the mean ranks between LEP and non-LEP responses in terms of identified barriers. Specifically, LEP respondents identified a lack of information in Spanish (p = 0.019), a lack of knowledge on what to do (p = 0.041), and a lack of a safe place to shelter (p = 0.011) at a greater rate in comparison to non-LEP respondents. In addition, when analyzing the data by U.S. citizenship status, non-U.S. citizens had significant differences in the rankings of a lack of knowledge on what to do (p = 0.032) and a lack of a safe place to shelter (p = 0.006) as identified barriers in comparison to U.S. citizens. However, no significant differences were found in barriers between participants born in the United States and outside the United States. See Tables 6a and 6b for differences in identified barriers across language and citizenship.
Comparison between LEP and non-LEP responses for identified barriers to tornado threat communication and response using the WMW test. Only significant differences are listed with a p value < 0.05. CIs are provided in paratheses.
Comparison by U.S. citizenship for barriers to tornado threat communication and response using the WMW test. Only significant differences are listed with a p value < 0.05. CIs are provided in paratheses.
4. Discussion
Prior research in risk communication highlights that the credibility of warning messages is based on trust through sustained relationships between the receiver and the sender of messages (Campbell et al. 2020). Historically, the approach to communicating hazard risk has been one-way or top-down, with technical experts sending messages to the public, with the goal of the public taking the expert’s recommendations and changing their behavior and actions (Yasui 2017). As shown in this study and others, challenges pertaining to language barriers, cultural differences, and trust in public officials have been found to inhibit the effectiveness of hazard warning communication with Latinx/e populations. Given that prior research has found that Latinx/e populations encounter various social, cultural, and political barriers that increase their vulnerability to harm during tornado hazards, examining intraethnic differences among Latinx/e populations in a region with a high risk for tornadoes and tornado-related losses may provide important insights for disaster planning and response for vulnerable and underrepresented populations. Social vulnerability is a dynamic and evolving state of being that is influenced by various dimensions, including physical, social, cultural, economic, and political factors, which collectively determine an individual’s or community’s capacity to ready themselves for, cope with, and adjust to various hazards (such as those discussed in Cutter et al. 2003; Strader et al. 2017; Morss et al. 2011). Our study specifically examined how differences in English language proficiency, nativity, and U.S. citizenship status may contribute to increased vulnerability to harm during tornado hazards among Latinx/e communities in the Southeast.
First, we examined tornado watch and warning comprehension among Latinx/e adults in the U.S. Southeast. Overall, the results indicate a moderate recognition and comprehension of tornado-related warnings (71.4%) and watches (65.2%) within the surveyed population. However, our intraethnic analyses revealed significant heterogeneity in tornado watch and warning comprehension when considering English language proficiency, nativity, and U.S. citizenship status. Latinx/e respondents who were English proficient exhibited higher comprehension rates for both tornado watches (73.7% vs 40.4%) and warnings (75.3% vs 54.1%) compared to respondents with LEP. U.S. citizens also demonstrated better comprehension for tornado watches and warnings than non-U.S. citizens, and individuals born in the United States displayed higher watch and warning comprehension than those born outside the United States (Fig. 2). These results build upon prior studies that have found differences between non-Latinx/e and Latinx/e respondents on tornado watch and warning comprehension, where Latinx/e respondents scored on average a bit lower on the scale than non-Latinx/e groups (Ripberger et al. 2019), and Spanish speakers scored lower than English speakers in their understanding of the definitions of tornado watches and warnings (Trujillo-Falcón et al. 2022). However, the current study provides additional insight into important differences in how Latinx/e populations in the Southeast interpreted tornado watch and warning terminology and how various social vulnerability factors, such as English language proficiency, nativity, and U.S. citizenship status, impact watch and warning comprehension. Furthermore, in comparing our results with national-level analyses on tornado watch and warning comprehension (e.g., Trujillo-Falcón et al. 2022), we want to highlight some important differences in findings. In a national-level study, Trujillo-Falcón et al. (2022) found 79% of English speakers and 60% of Spanish speakers were able to correctly identify a tornado warning. Our study, which concentrated on the U.S. Southeast and only within Latinx/e populations, found even lower tornado warning comprehension rates with 75% of English proficient respondents and 54% of LEP respondents able to correctly identify a tornado warning. These results highlight the need for multilingual weather hazard campaigns, particularly in a region that has a high risk for tornadoes and tornado-related losses (e.g., fatalities, injuries, damages). Together, these findings highlight how monolingual communication barriers can prevent LEP populations in the Southeast from understanding the information in severe weather forecasts and warnings when they are issued by the Storm Prediction Center and local NWS offices and warnings (Ripberger et al. 2019).
We also explored the primary sources of severe weather information for Latinx/e adults in the U.S. Southeast. We found wireless emergency alerts (smartphone alerts), television, and information from family, friends, or neighbors emerged as the top three sources in our results. However, LEP respondents ranked several traditional information sources, such as television, AM/FM commercial radio, landline phones, NOAA Weather Radio, and informal sources, such as school staff, churches/faith-based leaders, and community organizations, higher compared to their English-proficient counterparts. In addition, non-U.S. citizens also had distinct preferences for obtaining weather information from informal sources, specifically school staff and churches/faith-based leaders, in comparison to U.S. citizens (Tables 4a–c). These findings support prior work by Aguirre (1988) and Ahlborn et al. (2012) in highlighting the importance of various informal information sources within Spanish-speaking and immigrant Latinx/e communities.
Finally, in addition to uncovering heterogeneity in sources of severe weather information comprehension and sources, this study identified several barriers that may prevent Latinx/e individuals from seeking protection during tornado threats in the Southeast. Specifically, LEP respondents identified a lack of knowledge regarding what actions to take during a tornado threat and not having access to safe shelter, as particularly prominent barriers compared to their non-LEP counterparts. Furthermore, non-U.S. citizens expressed greater difficulties in terms of knowing what actions to take during a tornado and a lack of options for secure shelter when compared to U.S. citizens (Tables 6a,b). These findings support prior qualitative work conducted by Gaviria Pabon (2022) and Trujillo-Falcón et al. (2024) in identifying how inequities in weather risk communication, various cultural and disaster subcultures, and immigration policies disproportionally increased harm for Latinx/e immigrant populations in the Southeast during tornado the 10–11 December 2021, tornado outbreak.
Our results yield several important findings that can inform NWS severe weather planning and response with Latinx/e communities in the Southeast. Moving forward, multilingual severe weather forecasts and warnings are needed, along with culturally and linguistically appropriate outreach campaigns (e.g., educational factsheets, social media kits, etc.; Eisenman et al. 2009). To date, multiple NWS groups are working to improve severe weather communication within Spanish-speaking communities, including the Spanish Outreach Team (SOT) and the Multimedia Assistance in Spanish (MAS) team. The SOT, composed of bilingual meteorologists, translates and distributes various outreach materials, while the MAS team supports active weather operations with Spanish-language messaging, ensuring effective communication before, during, and after weather emergencies. Additionally, the American Meteorological Society (AMS) Committee for Hispanic and Latinx Advancement focuses on organizing community efforts and improving resource accessibility for bilingual meteorologists (see Trujillo-Falcón et al. 2021 for an overview on NWS resources for Spanish-speaking communities). These groups provide an essential role in bridging severe weather communication gaps for Spanish-speaking communities and their continuation and expansion can further enhance NWS commitments to ensuring equitable weather information across U.S. communities.
Our study’s findings also support the need for integrating and partnering with local community organizations and Latinx/e leaders to further culturally and linguistically appropriate risk communication strategies within Latinx/e populations (Phillips and Morrow 2007; Tripati et al. 2024). Partnership strategies could include working with trusted community members (e.g., Latinx/e leaders, Spanish media outlets) to provide cultural and linguistically appropriate education on determining tornado risk and protective actions (e.g., how to understand radar images on a map, how to understand probabilistic outlook information, how to get a variety of alerts) and identifying tornado shelter options, especially among undocumented immigrants who are disproportionately affected by political discrimination and fear of deportation in their daily lives (Méndez et al. 2020). Building relationships and communicating warning messages through trusted messengers has been shown to maximize message receipt and the taking of appropriate and recommended protective actions (Campbell et al. 2020). Efforts to alleviate hazard vulnerabilities and reduce harm for marginalized communities require establishing relationships with trusted messengers in the community and addressing their needs. However, building trust within these communities is not an overnight, top-down process but rather requires ongoing mutual engagement, a genuine commitment to understanding their specific needs and concerns, and the investment of time and resources. We recommend dedicated funding and resources be made available to build and sustain severe weather communication with Latinx/e leaders and communities. This investment could include hiring and training bilingual community liaisons to foster trust and develop programs that promote ongoing community engagement (e.g., community meetings, workshops, meals, etc.) to create a feedback loop of knowledge transfer and participation between the weather enterprise and Latinx/e communities. We believe investing additional resources for establishing and maintaining these partnerships with Latinx/e communities can better inform disaster planning and response for vulnerable and underrepresented populations.
5. Limitations
As with any research study, this study is not without limitations and challenges. First, our geographical concentration of the Southeast states and Florida is important to note when comparing our sample to other national-level Latinx/e samples (see Fig. 1 and Table 1). In terms of recruiting survey data that is representative of the Latinx/e population in the U.S. Southeast region, we encountered complexities in weighting our sample for creating a Latinx/e-representative sample for the region (see Brown 2015). A significant portion of our respondents were from Florida and the South Atlantic. Florida’s Latinx/e populations are particularly diverse compared to the greater U.S. Latinx/e population per capita from the U.S. Census (2020). For example, across the greater United States, Mexicans represent over 50% of the Latinx/e population, while in Florida and the South Atlantic United States, Cubans (28%) and Puerto Ricans (21%) represent the highest rates for countries of origin with Mexicans comprising only 14%. Thus, the diversity within Florida’s Latinx/e population may influence the generalizability of our findings to other regions with different demographic compositions.
Second, while the current study examined differences in how LEP and non-LEP Latinx/e populations in the Southeast receive, understand, and act upon severe weather information, we recognize that Latinx/e populations consist of a wide range of Spanish dialects and cultural backgrounds. For example, Gaviria Pabon (2022) has examined the role of culture (e.g., country of origin and language) and hazard experience (e.g., prior experience with tornadoes) and how disaster subcultures play a significant role in influencing how Latinx/e immigrants perceive and understand severe weather threats. Future research is needed to further consider the diversity of Latinx/e representativeness and the role of various cultures and dialects across U.S. regions and climatic zones.
Third, to assess English language proficiency, our approach utilized a binary subjective response measure (very well or not), which did not capture the full spectrum of English language proficiency. Participants with moderate proficiency, who might still provide valuable insights, were not included. In addition, while Latinx/e adults mostly speak English and Spanish, they can also speak other languages (e.g., indigenous languages). Our language proficiency approach did not account for multilingual participants who might have varying degrees of proficiency in multiple languages. Our study design also did not explore if literal English translations of NWS warnings are sufficient and clear for Spanish-speaking and bilingual populations. Future research should focus on the interpretation of messages and the intended response of Spanish-speaking and bilingual populations (Trujillo-Falcón et al. 2022). Furthermore, while we focused on intraethnic disparities, we did not focus on additional demographic and socioeconomic characteristics such as gender, income, age, and disability status. Additional studies are needed to examine these intersecting factors to illustrate and address the underlying mechanisms that contribute to vulnerability and disparities during tornado events (First 2024). This includes examining how gender, income, age, and disability intersect with intraethnic disparities to receive, understand, and act upon severe weather information.
6. Conclusions
The Southeast region of the United States faces a high risk of tornadoes and related losses, making effective warning communication critical. However, Latinx/e households in this area often encounter barriers due to limited English proficiency, cultural differences, and mistrust of public officials, which inhibit the effectiveness of these warnings. This study surveyed a sample of Latinx/e adults across 11 states in the Southeast to examine the impact of various social vulnerability factors (e.g., English language proficiency, nativity, and immigration status) on their reception and understanding of severe weather information. The results highlight significant intraethnic disparities in tornado hazard comprehension and response based on identified social vulnerability factors. These findings suggest that the NWS and its partners should collaborate with Latinx/e communities and organizations to refine communication strategies and address vulnerability factors in future hazard research, alert systems, and outreach campaigns.
We define Latinx/e individuals as having cultural, ancestral, or national ties to countries in Latin America and the Caribbean, which include individuals from Mexico, Central America (e.g., Guatemala, Honduras, El Salvador), South America (e.g., Colombia, Argentina), and the Spanish-speaking Caribbean (e.g., Puerto Rico, Cuba, Dominican Republic). We recognize that selecting an encompassing and inclusive term is a challenge and we chose to follow a modified version of the National Institutes of Health’s ethnic categorization of Latino/a but to strengthen inclusivity of nonbinary individuals with the use of Latinx/e for this article.
Acknowledgments.
We would like to express our gratitude to the three anonymous reviewers for their valuable comments and constructive feedback, which significantly improved the quality of this manuscript. This study was funded by the National Oceanic and Atmospheric Administration, VORTEX U.S.A Award (NA220AR4590520). This study was funded by the National Oceanic and Atmospheric Administration (NA220AR4590520). The funding source did not participate in study design, data collection and summary, drafting, or submission of this manuscript.
Data availability statement.
The authors support the commitment made by AMS to open and reproducible research. The survey data used in this study are securely stored on University of Tennessee servers and are accessible with reasonable request and data confidentiality approval, in accordance with University of Tennessee Institutional Review Board policies and the protection of study participants.
REFERENCES
Aguirre, B. E., 1988: The lack of warnings before the Saragosa tornado. Int. J. Mass Emerg. Disasters, 6, 65–74, https://doi.org/10.1177/028072708800600105.
Ahlborn, L., J. M. Franc, and D. S. Med, 2012: Tornado hazard communication disparities among Spanish-speaking individuals in an English-speaking community. Prehospital Disaster Med., 27, 98–102, https://doi.org/10.1017/S1049023X12000015.
Ashley, W. S., A. J. Krmenec, and R. Schwantes, 2008: Vulnerability due to nocturnal tornadoes. Wea. Forecasting, 23, 795–807, https://doi.org/10.1175/2008WAF2222132.1.
Bitterman, A., J. Ripberger, M. Krocak, C. Silva, H. Jenkins-Smith, J. Trujillo-Falcón, A. Gaviria Pabón, and S. Ernst, 2024: WXS23: A nationwide survey of severe weather information sources among English and Spanish speakers in the U.S. Harvard Dataverse. U.S. Harvard Dataverse, accessed 16 August 2023, https://doi.org/10.7910/DVN/JFLKOJ.
Brown, A., 2015: The unique challenges of surveying U.S. Latinos. Pew Research Center, accessed 21 July 2024, https://www.pewresearch.org/race-and-ethnicity/2015/11/12/the-unique-challenges-of-surveying-u-s-latinos/.
Burke, S., 2018: Immigration raid takes 97 into custody at Tennessee plant. U.S. News, https://apnews.com/general-news-04869e32e439490fb27a229fea7a69b1.
Campbell, N., K. Roper-Fetter, and M. Yoder, 2020: Risk Communication Involving Vulnerable Populations: An Annotated Bibliography. Natural Hazards Center, University of Colorado Boulder, 61 pp.
Cutter, S. L., B. J. Boruff, and W. L. Shirley, 2003: Social vulnerability to environmental hazards. Soc. Sci. Quart., 84, 242–261, https://doi.org/10.1111/1540-6237.8402002.
Donner, W. R., H. Rodriguez, and W. Diaz, 2012: Tornado warnings in three southern states: A qualitative analysis of public response patterns. J. Homeland Secur. Emerg. Manage., 9, 1547–7355, https://doi.org/10.1515/1547-7355.1955.
Eisenman, D. P., D. Glik, L. Gonzalez, R. Maranon, Q. Zhou, C.-H. Tseng, and S. M. Asch, 2009: Improving Latino disaster preparedness using social networks. Amer. J. Prev. Med., 37, 512–517, https://doi.org/10.1016/j.amepre.2009.07.022.
First, J. M., 2024: Examining tornado exposure, post-tornado distress, and gender following the March 2020 tornado in Nashville, Tennessee. Environ. Hazards, 1–14, https://doi.org/10.1080/17477891.2024.2379895.
First, J. M., K. Ellis, M. L. Held, and F. Glass, 2021: Identifying risk and resilience factors impacting mental health among Black and Latinx adults following nocturnal tornadoes in the U.S. Southeast. Int. J. Environ. Res. Public Health, 18, 8609, https://doi.org/10.3390/ijerph18168609.
Gaviria Pabon, A. R., 2022: The role of disaster subcultures in Spanish-speaking Hispanic/Latino immigrants in the United States. M.S. thesis, Dept. of Geography and Environmental Sustainability, University of Oklahoma, 76 pp., https://shareok.org/server/api/core/bitstreams/ba2666c9-e22f-43ef-827d-2de8709fc318/content.
Grineski, S. E., T. W. Collins, and J. Chakraborty, 2013: Hispanic heterogeneity and environmental injustice: Intra-ethnic patterns of exposure to cancer risks from traffic-related air pollution in Miami. Popul. Environ., 35, 26–44, https://doi.org/10.1007/s11111-012-0184-2.
Guo, L., K. Wang, and H. B. Bluestein, 2016: Variability of tornado occurrence over the continental United States since 1950. J. Geophys. Res. Atmos., 121, 6943–6953, https://doi.org/10.1002/2015JD024465.
Hatzenbuehler, M. L., S. J. Prins, M. Flake, M. Philbin, M. S. Frazer, D. Hagen, and J. Hirsch, 2017: Immigration policies and mental health morbidity among Latinos: A state-level analysis. Soc. Sci. Med., 174, 169–178, https://doi.org/10.1016/j.socscimed.2016.11.040.
Held, M. L., and S. Lee, 2017: Discrimination and mental health among Latinos: Variation by place of origin. J. Ment. Health, 26, 405–410, https://doi.org/10.1080/09638237.2016.1207220.
Held, M. L., J. M. First, M. Huslage, and M. Holzer, 2022: Policy stress and social support: Mental health impacts for Latinx adults in the southeast United States. Soc. Sci. Med., 307, 115172, https://doi.org/10.1016/j.socscimed.2022.115172.
Klockow, K. E., R. A. Peppler, and R. A. McPherson, 2014: Tornado folk science in Alabama and Mississippi in the 27 April 2011 tornado outbreak. GeoJournal, 79, 791–804, https://doi.org/10.1007/s10708-013-9518-6.
Lu, M., M. Tippett, and U. Lall, 2015: Changes in the seasonality of tornado and favorable genesis conditions in the central United States. Geophys. Res. Lett., 42, 4224–4231, https://doi.org/10.1002/2015GL063968.
Méndez, M., G. Flores-Haro, and L. Zucker, 2020: The (in)visible victims of disaster: Understanding the vulnerability of undocumented Latino/a and indigenous immigrants. Geoforum, 116, 50–62, https://doi.org/10.1016/j.geoforum.2020.07.007.
Morss, R. E., O. V. Wilhelmi, G. A. Meehl, and L. Dilling, 2011: Improving societal outcomes of extreme weather in a changing climate: An integrated perspective. Annu. Rev. Environ. Resour., 36 (1), 1–25, https://doi.org/10.1146/annurev-environ-060809-100145.
NWS, 2014: May 2013 Oklahoma tornadoes and flash flooding. NWS Service Assessment, 63 pp., www.weather.gov/media/publications/assessments/13oklahoma_tornadoes.pdf.
Phillips, B. D., and B. H. Morrow, 2007: Social science research needs: Focus on vulnerable populations, forecasting, and warnings. Nat. Hazards Rev., 8, 61–68, https://doi.org/10.1061/(ASCE)1527-6988(2007)8:3(61).
Ripberger, J. T., M. J. Krocak, W. W. Wehde, J. N. Allan, C. Silva, and H. Jenkins-Smith, 2019: Measuring tornado warning reception, comprehension, and response in the United States. Wea. Climate Soc., 11, 863–880, https://doi.org/10.1175/WCAS-D-19-0015.1.
Senkbeil, J. C., M. S. Rockman, and J. B. Mason, 2012: Shelter seeking plans of Tuscaloosa residents for a future tornado event. Wea. Climate Soc., 4, 159–171, https://doi.org/10.1175/WCAS-D-11-00048.1.
Senkbeil, J. C., D. A. Scott, P. Guinazu-Walker, and M. S. Rockman, 2014: Ethnic and racial differences in tornado hazard perception, preparedness, and shelter lead time in Tuscaloosa. Prof. Geogr., 66, 610–620, https://doi.org/10.1080/00330124.2013.826562.
Smith, M. D., J. E. Ten Hoeve, C. Lauer, and V. Brown, 2023: A weather-ready nation for all? the demographics of severe weather understanding, reception, and response. Wea. Climate Soc., 15, 229–262, https://doi.org/10.1175/WCAS-D-22-0056.1.
Steele, L. G., A. Bostic, S. M. Lynch, and L. Abdelaaty, 2022: Measuring ethnic diversity. Annu. Rev. Sociol., 48, 43–63, https://doi.org/10.1146/annurev-soc-030420-015435.
Strader, S. M., W. S. Ashley, T. J. Pingel, and A. J. Krmenec, 2017: Projected 21st century changes in tornado exposure, risk, and disaster potential. Climatic Change, 141, 301–313, https://doi.org/10.1007/s10584-017-1905-4.
Strader, S. M., W. S. Ashley, T. J. Pingel, and A. J. Krmenec, 2018: How land use alters the tornado disaster landscape. Appl. Geogr., 94, 18–29, https://doi.org/10.1016/j.apgeog.2018.03.005.
Strader, S. M., W. S. Ashley, A. M. Haberlie, and K. Kaminski, 2022: Revisiting U.S. nocturnal tornado vulnerability and its influence on tornado impacts. Wea. Climate Soc., 14, 1147–1163, https://doi.org/10.1175/WCAS-D-22-0020.1.
Sutter, D., and K. M. Simmons, 2010: Tornado fatalities and mobile homes in the United States. Nat. Hazards, 53, 125–137, https://doi.org/10.1007/s11069-009-9416-x.
Tripati, A., and Coauthors, 2024: Centering equity in the nation’s weather, water, and climate services. Environ. Justice, 17, 45–53, https://doi.org/10.1089/env.2022.0048.
Trujillo-Falcón, J. E., O. Bermúdez, K. Negrón-Hernández, J. Lipski, E. Leitman, and K. Berry, 2021: Hazardous weather communication en español: Challenges, current resources, and future practices. Bull. Amer. Meteor. Soc., 102, E765–E773, https://doi.org/10.1175/BAMS-D-20-0249.1.
Trujillo-Falcón, J. E., and Coauthors, 2022: ¿Aviso o Alerta? Developing effective, inclusive, and consistent watch and warning translations for U.S. Spanish speakers. Bull. Amer. Meteor. Soc., 103, E2791–E2803, https://doi.org/10.1175/BAMS-D-22-0050.1.
Trujillo-Falcón, J. E., A. R. G. Pabón, J. Reedy, and K. E. Klockow-McClain, 2024: Systemic vulnerabilities in Hispanic and Latinx immigrant communities led to the reliance on an informal warning system in the December 10–11, 2021, tornado outbreak. Nat. Hazards Rev., 25, 04023059, https://doi.org/10.1061/NHREFO.NHENG-1755.
U.S. Census Bureau, 2020: The Hispanic population in the United States: 2020. Accessed 27 August 2023, https://www.census.gov/data/tables/2020/demo/hispanic-origin/2020-cps.html.
U.S. Department of Justice, 2011: Commonly asked questions and answers regarding the Limited English Proficiency (LEP) individuals 042511, LEP.gov., 8 pp., https://www.lep.gov/sites/lep/files/media/document/2020-03/042511_QA_LEP_General_0.pdf.
U.S. Immigration and Customs Enforcement, 2019: Fiscal year 2019 Enforcement and Removal Operations Report. U.S. Department of Homeland Security, 32 pp., https://www.ice.gov/sites/default/files/documents/Document/2019/eroReportFY2019.pdf.
Yasui, E., 2017: Communicating risk for vulnerable groups: A case study of the Mano community’s strategies for collective knowledge to action. Local Environ., 22, 478–491, https://doi.org/10.1080/13549839.2016.1229762.
