Wong Roger, Zhang Lingling, Zhang Kai
Department of Public Health and Preventive Medicine, Norton College of Medicine, SUNY Upstate Medical University, Syracuse, NY, USA; Department of Geriatrics, Norton College of Medicine, SUNY Upstate Medical University, Syracuse, NY, USA.
Department of Mathematics & Statistics, University at Albany, State University of New York, Albany, NY, USA.
Sci Total Environ. 2025 Mar 1;967:178761. doi: 10.1016/j.scitotenv.2025.178761. Epub 2025 Feb 13.
Millions of U.S. residents experience increasingly more prevalent weather events due to climate change, however, there is limited research exploring the vulnerability to multiple extreme weather events using a national U.S.
Identify patterns in exposures to climate events, and examine sociodemographic factors associated with increased climate event vulnerability.
Data was retrieved from the May 2022 American Trends Panel, a nationally representative sample of 10,282 United States adults. We performed a latent class analysis, a statistical method used to identify unobserved subgroups (latent classes) within a population, to group respondents by patterns in five climate event experiences (heatwave, intense storm, wildfire, drought, and sea level rise), and analyzed variables associated with vulnerability to climate events using weighted multinomial logistic regression, a statistical method that models the probability of membership in one of several outcome categories (climate vulnerability groups) relative to a reference category, while accounting for survey weights to ensure generalizability to the U.S.
Respondents were categorized into four latent classes, which are unobserved subgroups identified through patterns in exposures to five climate events (heatwave, intense storm, wildfire, drought, and sea level rise). These subgroups were based on exposures to heatwave (42.5 %), intense storm (43.2 %), wildfire (21.3 %), drought (30.8 %) and sea level rise (15.8 %): high (9.8 %), heat-storm (22.2 %), heat-drought (13.4 %), and low (54.6 %) climate event vulnerability. Relative risk for high climate event vulnerability refers to the likelihood of belonging to the "high vulnerability" group compared to the "low vulnerability" group. It is assessed using the relative risk ratio (RRR), which is a measure of the association between a particular sociodemographic factor (e.g., age, gender, region) and the likelihood of being in a specific vulnerability group relative to the reference group. For instance, an RRR <1 indicates a reduced risk, while an RRR >1 indicates an increased risk compared to the reference category. Relative risk for high climate event vulnerability was lower for older adults (RRR = 0.39, p < 0.001), potentially reflecting a greater capacity to cope with certain climate events, such as access to stable housing or resources. However, this finding should not be interpreted as older adults being universally less vulnerable. Numerous studies have shown that older adults are at significantly higher risk during heatwaves due to physiological and social factors, which our analysis may not fully capture. Relative risk for high vulnerability was higher for females (RRR = 1.42, p = 0.01) and residents in the South (RRR = 2.05, p = 0.003) and West (RRR = 9.31, p < 0.001) geographic regions. Relative risk for heat-drought was higher for Hispanic adults (RRR = 1.51, p = 0.03), but lower for high school graduates (RRR = 0.40, p = 0.01) compared to those who did not complete high school.
We identified several underlying climate event exposure subpopulations, ranging from low to high vulnerability. As climate-related events become more frequent, our results provide critical insights for stakeholders to identify high-risk individuals and prioritize resources for disaster management.
由于气候变化，数以百万计的美国居民经历着越来越普遍的天气事件，然而，利用美国全国性数据探索对多种极端天气事件的脆弱性的研究有限。
确定气候事件暴露模式，并研究与气候事件脆弱性增加相关的社会人口因素。
数据取自2022年5月的美国趋势调查小组，该小组是对10282名美国成年人的全国代表性样本。我们进行了潜在类别分析，这是一种用于识别总体中未观察到的亚组（潜在类别）的统计方法，通过五种气候事件经历（热浪、强烈风暴、野火、干旱和海平面上升）的模式对受访者进行分组，并使用加权多项逻辑回归分析与气候事件脆弱性相关的变量，加权多项逻辑回归是一种统计方法，用于模拟相对于参考类别的几个结果类别（气候脆弱性组）之一的成员概率，同时考虑调查权重以确保对美国具有普遍性。
受访者被分为四个潜在类别，这些类别是通过对五种气候事件（热浪、强烈风暴、野火、干旱和海平面上升）的暴露模式识别出的未观察到的亚组。这些亚组基于对热浪（42.5%）、强烈风暴（43.2%）、野火（21.3%）、干旱（30.8%）和海平面上升（15.8%）的暴露情况：高气候事件脆弱性（9.8%）、热风暴（22.2%）、热干旱（13.4%）和低气候事件脆弱性（54.6%）。高气候事件脆弱性的相对风险是指与“低脆弱性”组相比属于“高脆弱性”组的可能性。它使用相对风险比（RRR）进行评估，RRR是衡量特定社会人口因素（如年龄、性别、地区）与相对于参考组处于特定脆弱性组的可能性之间关联的指标。例如，RRR<1表示风险降低，而RRR>1表示与参考类别相比风险增加。老年人高气候事件脆弱性的相对风险较低（RRR = 0.39，p < 0.001），这可能反映出他们应对某些气候事件的能力更强，例如拥有稳定的住房或资源。然而，这一发现不应被解释为老年人普遍更不易受影响。许多研究表明，由于生理和社会因素，老年人在热浪期间面临的风险显著更高，而我们的分析可能并未完全涵盖这些因素。女性（RRR = 1.42，p = 0.01）以及南部（RRR = 2.05，p = 0.003）和西部（RRR = 9.31，p < 0.001）地理区域的居民高脆弱性的相对风险较高。西班牙裔成年人热干旱的相对风险较高（RRR = 1.51，p = 于未完成高中学业的人。
我们确定了几个潜在的气候事件暴露亚群体，其脆弱性从低到高不等。随着与气候相关的事件变得更加频繁，我们的结果为利益相关者识别高风险个体并为灾害管理确定资源优先级提供了关键见解。 03）,但与未完成高中学业的人相比，高中毕业生热干旱的相对风险较低（RRR = 0.40，p = 0.01）。
