The association of extreme heat and PM_2.5 with preterm birth and low birthweight in Florida, United States

Abstract

Introduction:

The increase in global temperature due to climate change are leading heatwaves to become more frequent and intense and air quality to worsen. Extreme heat and air pollution have been linked to adverse birth outcomes, which could be mitigated by air conditioning (AC). This study examined associations between heat and PM_2.5 with preterm birth (PTB) and low birthweight (LBW) as well as indicators for AC availability.

Methods:

We used Florida birth certificates to conduct a retrospective cohort study of 4,514,067 births (2000 to 2022). Weather data were obtained from the Parameter-elevation Regressions on Independent Sloped Model and exposure based on mother's residential address. Three heatwave definitions were evaluated based on consecutive days with maximum temperature exceeding 90th percentile. PM_2.5 concentrations (µg/m³) were extracted from the North America Chemical Reanalysis dataset and averaged by trimester. AC exposure was estimated using real estate data. Logistic regression was used to estimate adjusted risk ratios (aRR).

Results:

We found an increased risk of PTB with increasing heatwave days occurring within a heatwave episode (aRR 1.001, 95% CI 1.001–1.002) during the first trimester and an inverse association in the second (aRR 0.996, 95% CI 0.995–0.997) and third trimesters (aRR 0.921, 95% CI 0.920–0.922). An inverse association with term LBW was observed in the third trimester (aRR 0.981, 95% CI 0.979–0.984). Higher PM_2.5 exposure during the first trimester was associated with increased PTB risk (aRR 1.010, 95% CI 1.008–1.012) and an inverse association with LBW (aRR 0.988, 95% CI 0.984–0.992) with increasing PM_2.5 concentrations in the first trimester. There was no mitigation effect of AC.

Discussion:

These results suggest the importance of considering duration and magnitude in defining heat exposure to understand the effects of extreme heat on pregnancy. There is a need for improved comparability of heatwave definitions as climate change continues to accurately assess their public health impacts.

1 Introduction

Climate change is altering weather patterns and the environment, and in turn adversely impacting human health. The increase in global temperatures alters the probability of hurricanes, floods, and heatwaves (1). In the United States, heatwaves are expected to become more intense and more frequent due to global climate change, according to the Fifth National Climate Assessment Reports (2).

While climate change affects everyone, the impacts are uneven across the country (3). Communities in the Gulf of Mexico region are projected to experience the largest heat exposure increase compared to others around the nation (4), with regional average temperatures projected to increase up to 3 °C in 2041–2060 and up to 6 °C in 2080–2099 (5). With higher temperatures due to climate change, air quality may worsen as a result of accelerating photochemical reactions that create atmospheric conditions favoring pollutant accumulation (6, 7). Furthermore, the changing climate has the potential for severe consequences in several states along the Gulf Coast, as a result of the concentration of many socially disadvantaged populations. With increased temperatures and longer, more frequent, and intense heatwaves in the Gulf Coast, those individuals are at the highest risk of experiencing heat-related mortality, along with infants under one year of age (8). Compared to 1975–1995, the number of extreme heat days and associated mortality is projected to increase throughout the twenty-first century in several large cities, including Jacksonville and Tampa, Florida (9).

One particularly vulnerable population is pregnant women and infants. Increased exposure to extreme heat may have a negative effect on birth outcomes. Exposure to higher ambient temperatures (10, 11) and frequent heatwaves (12) has been associated with the risk of preterm delivery and lower birthweight. Several biological mechanisms have been proposed to explain these associations, including maternal dehydration, reduced uterine blood flow, and increased systemic inflammation during periods of thermal stress (13–15). Similarly, numerous studies have demonstrated an association between air pollution and adverse birth outcomes. Fine particulate matter with an aerodynamic diameter smaller than 2.5 micrometers (PM_2.5) is of particular concern due to its ability to penetrate deep in the lungs and the bloodstream (16). Because of this ability to enter systemic circulation, PM_2.5 exposure during pregnancy may interfere with placental function and fetal development (17, 18). A 2023 meta-analysis found that higher maternal PM_2.5 exposure, modeled per interquartile range (IQR) increase, was associated with preterm birth (RR 1.05, 95% CI 1.03–1.07) and low birthweight (RR 1.02, 95% CI 1.01–1.04) (19). Together, these findings suggest that both extreme heat and air pollution may contribute to adverse birth outcomes.

Although numerous studies have linked extreme heat and air pollution to adverse birth outcomes, findings have varied across populations and study designs. Some studies report stronger associations than others, suggesting that differences in how heat exposure is measured may influence estimated effects, or that populations vary in susceptibility. Heatwave definitions are inconsistent across studies as there is not a universally applicable definition. Heatwave definitions tend to differ in the temperature metric used (e.g., maximum temperature, mean temperature, humidity, and apparent temperature), the threshold of exceedance (e.g., absolute (> 35 °C) or relative (percentile based), and duration of heat (20). In general, most definitions use different combinations of intensity and duration of heat to define a heatwave (20, 21). Lan et al. (2012), for example, used maximum temperatures ≥ the 98th percentile for ≥ three days in their study examining mortality during heatwaves (22), while Wang et al. (2012) defined a heatwave as maximum temperatures ≥ 35 °C for ≥ two days when examining the impact of heatwaves on mortality and hospital admissions (23). Furthermore, prior research has also demonstrated that slight changes in heatwave definitions may result in appreciable differences in the magnitude of the associations between heat and health outcomes (11, 24), for example, observed varying associations between heatwaves and health outcomes depending on the heatwave index used to define heatwave days, with stronger positive associations observed for definitions based on higher temperature thresholds, particularly when heatwaves were defined as mean daily temperatures exceeding the 98th percentiles for at least two consecutive days

In addition, exposure to extreme heat may be modified by individual- and household-level factors that influence indoor conditions, including housing characteristics, access to cooling resources such as residential air conditioning (AC), and socioeconomic status which may influence the aforementioned conditions as well as job exposure to heat and transportation possibilities. Residential AC can substantially reduce indoor temperatures and has been found to mitigate heat-related morbidity and mortality by limiting personal exposure to extreme ambient heat (25–27).

Therefore, in this analysis of the relationship between heat and PM_2.5 with the risk of preterm birth (PTB) and term low birthweight (LBW), we investigated the impact of various definitions of heat waves as well as indicators of AC availability. We hypothesized that exposure to extreme heat and higher PM_2.5 concentrations would be associated with an increased risk of preterm delivery and lower birthweight. Furthermore, we hypothesized that definitions with longer streaks of consecutive heatwave days would be associated with an increased risk of preterm delivery and lower birthweight. We also assessed effect modification by type of AC on the relationship between extreme heat and adverse birth outcomes, hypothesizing that higher AC availability would reduce the association between heatwaves and the risk of PTB and LBW.

2 Materials and methods

2.1 Study population

We conducted a retrospective cohort study of mothers who gave birth in Florida between 2000 and 2022 using birth records from the Florida Department of Health. Our analysis was restricted to observations with accurate latitude and longitude (N = 4,662,195) and singleton births (N = 4,514,345) during the study period, and we excluded births with missing data on birthweight (N = 278). Furthermore, in our low birthweight analysis, we restricted to term births (≥ 37 weeks gestation) (N = 4,125,271). The final total analytical sample included 4,514,067 observations, which was used for the heatwave analyses. Analyses of PM_2.5 exposure and AC type were conducted on smaller subsets of the cohort due to data availability and are described in the relevant exposure sections below.

2.2 Outcome

The outcomes of interest for our analysis were PTB and term LBW. PTB was defined as gestational weeks less than 37 weeks. Term LBW was defined as a birthweight less than 2500 grams at gestational weeks greater than or equal to 37 weeks. Gestational weeks and birthweight were obtained from the birth records.

2.3 Exposure

2.3.1 Heatwave definitions

Heatwave exposure was individually assigned based on the mother's residential address. Weather data were obtained from the Parameter-elevation Regressions on Independent Sloped Model (PRISM), which provides spatially and temporally consistent raster datasets. The data are available at a daily resolution from 1981 to the present, with a 4 km spatial resolution covering the contiguous United States. For the present analysis data from 2000 to 2022 was used. Heatwave exposure was defined using a relative temperature threshold based on trimester-specific maximum daily temperatures exceeding the 90th percentile. Each address was converted into latitude and longitude coordinates, and the corresponding PRISM grid was extracted for each location. Due to the relatively long study period, the maximum daily temperature time series was linearly detrended prior to identifying heatwave days to account for long-term temperature increases. Without detrending, heatwave classifications would have been disproportionately concentrated in more recent years due to climate-driven warming. For each day, a linear regression model was used to estimate expected temperatures, and deviations from this model were calculated. Days with detrended maximum temperatures exceeding the 90th percentile were classified as heatwave days. We did not consider February 29th and those values were excluded from analysis. Three heatwave exposure metrics were evaluated based on the duration of consecutive heatwave days: (a) the total number of days exceeding the 90th percentile threshold, (b) the number of days occurring within episodes of at least two consecutive heatwave days, excluding single days, and (c) the number of days occurring within episodes of at least three consecutive heatwave days, excluding single days and days within episodes of two consecutive days.

2.3.2 PM_2.5 concentrations

PM_2.5 concentrations (µg/m³) were extracted from the North America Chemical Reanalysis (NACR) dataset developed by the Air Quality Laboratory at George Mason University with 12 km spatial gridding from 2002 to 2021 (28). Concentrations were linked to the mother's latitude and longitude information on the birth record based on grid cell containment, where each birth record was assigned to the PM_2.5 concentration of the grid cell containing its reprojected coordinates. The average PM_2.5 concentration (µg/m³) was calculated for the duration of each trimester. The analytic sample for the PM_2.5 analysis included 2,253,882 births (2,931,650 for term LBW) due to availability of NACR PM_2.5 data.

2.3.3 Air conditioning type

The air conditioning residential estimate was constructed using real estate data provided by The Warren Group, distributed through Dewey Data with multiple contextual covariates, including Cooling Degree Days, sociodemographic indicators (e.g., median income, race) and urban development metrics from the Historical Settlement Data Compilation for the United States (HISDAC-US) (29) (Ahn et al., 2024). Details are further described in Ahn & Uejio (2025) (30). The accuracy of the predicted household level AC data was central AC 98%, no AC 79%, and other AC 79%. To merge with the birth records, we implemented a two-step matching process to link birth data with AC data using Mutual Nearest Neighborhood Matching, followed by Nearest Neighborhood Matching for non-mutual matches, utilizing metrics such as distance and mutual matching indicators for filtering. Additionally, a filter of less than ten meters distance to household AC was used to account for AC geocoding to the wrong household. The AC subanalysis included 951,196 (871,576 for term LBW), reflecting the years for which AC information was available.

2.4 Covariates

Covariates were determined a priori based on their relation to extreme heat and PM_2.5 or PTB and term LBW. Maternal age, race/ethnicity, education, insurance status, pre-pregnancy BMI, previous PTB, and behavioral factors such as alcohol use have been shown to influence the risk of PTB and LBW (31–33). Maternal and paternal characteristics [mother's age, mother and father's race/ethnicity, mother and father's education, pre-pregnancy body mass index (BMI), previous preterm infant, alcohol use, Women, Infant, and Children (WIC) program participation, and principal source pay] were ascertained from birth records. We categorized pre-pregnancy BMI into the following categories: underweight (BMI < 18.5), normal weight (BMI 18.5–24.9), overweight (BMI 25–29.9), and obesity (BMI > 30). Furthermore, to account for temporal patterns in ambient temperature and air pollutant concentrations, the season of exposure was included (34, 35). We created a season variable for each trimester, categorized as follows: spring (March 20th—June 20th), summer (June 21st—September 21st), fall (September 22nd—December 20th), and winter (December 21st—March 19th). The variables were defined as the season into which the majority of the trimester fell.

Missing covariate data was handled using multiple imputation by chained equations (MICE). The number and proportion of missing observations for each variable are reported in Table 1; missingness was highest for mother's race (n = 1,000,082, 22.2%), WIC participation (n = 716,841, 15.9%), and mother's ethnicity (n = 619,601, 15%). A complete-case analysis would have included 2,742,642 births (60.8%) of the full analytic sample. The imputation model included all covariates included in the analytic models. With this approach, a regression model is fit for each covariate with missing values conditioning on all other covariates, and missing values are then imputed sequentially (36). We generated 5 multiply imputed datasets, with 5 iterations of the chained equations performed for each dataset.

2.5 Statistical analysis

Study population characteristics were summarized descriptively. To assess the association between heatwaves, PM_2.5 concentrations, and PTB and term LBW, we constructed separate logistic regression models and calculated odds ratios to estimate the relative risk (RR) and 95% confidence intervals (CI). Additionally, to control for seasonality, we restricted births in which most of the days in a given trimester occurred between May through October and estimated the relative risk and 95% CIs to assess relationships between heatwaves, PM_2.5 concentrations, and PTB and term LBW. To control for confounding, models were adjusted for the mother and father's race/ethnicity, mother and father's education, and alcohol use. Race/ethnicity and education may influence where and the level of extreme heat and PM_2.5 exposure experienced, as well as an increased risk of PTB and term LBW. Models were also adjusted for the following covariates as they are related to the risk of PTB and term LBW and can increase precision of the estimates: mother's age, pre-pregnancy BMI, previous preterm infant, and principal source of payment.

We conducted a subanalysis to assess potential effect modification by AC type. Due to AC data availability, the subanalysis was only carried out over the years 2016 to 2021. We included an interaction term between heatwaves and type of AC into the aforementioned logistic regression models. An overall p-value for the interaction was obtained using a likelihood ratio test comparing models with and without the interaction term. Stratified effect estimates were also calculated to examine associations within each category of AC type using the model with the interaction term between AC type and heatwave definitions. All analyses were conducted using R version 4.3.2 and the “emmeans” package.

2.5.1 Sensitivity analysis

First, to ensure that we are obtaining consistent results regarding the associations between heatwaves and PTB and term LBW, we applied alternative heatwave definitions (e.g., 95th and 98th percentile) in the logistic regression models. Using the alternative heatwave definitions, we considered only trimesters occurring in the hot seasons, restricting births in which the majority of time within each trimester was in May through October. Although term LBW is fairly often used as an approximation for reduced fetal growth in the absence of shortened gestation, restricting births to those greater than or equal to 37 weeks gestation may result in collider bias on the pathway between PM_2.5 exposure and birthweight. To assess the robustness of our term LBW analysis, we examined the association between PM_2.5 and LBW without restricting by gestational age. To understand the effects of removing the long-term temporal temperature trend on the estimated associations between heatwave exposure and PTB and LBW, we repeated our primary heatwave analyses using non-detrended daily maximum temperature. Lastly, as humidity may influence physiological heat stress, we conducted an additional analysis using heat index as the exposure. Heat index incorporates both temperature and relative humidity, providing a combined measure of heat and humidity exposure. Average heat index was calculated over the 7, 14, 21, and 28 days preceding delivery and evaluated in separate logistic regression models.

3 Results

3.1 Descriptive analysis

Of the 4.5 million births in our study, 8.6% were delivered preterm. When the study population was restricted to term births, 2.5% had a low birthweight. Among the study population, 55.8% were White, 56.2% were non-Hispanic, 31.3% had a high school diploma or GED, 37.6% had a normal pre-pregnancy weight, and 39.4% and 37.6% had Medicaid or private insurance, respectively (Table 1). On average, women were 28.2 years old. Among births to Black women, 24.0% were preterm and 29.0% were term low birthweight births, compared with 16.6% that were full term and 16.2% that were normal birthweight (Table 1).

3.2 Associations between heatwaves, PM_2.5 and PTB/term LWB

Each heatwave day occurring within a heatwave episode (defined as ≥ 1 consecutive heatwave day) during the first trimester was associated with an increased risk of PTB (aRR 1.002, 95% CI 1.001–1.002) (Table 2). In the second and third trimester, each heatwave day within these episodes was inversely associated with risk of PTB (aRR 0.996, 95% CI 0.995–0.997; aRR 0.921, 95% CI 0.920–0.922, respectively). In the third trimester, each heatwave day within a heatwave episode was inversely associated with risk of LBW (aRR 0.981, 95% CI 0.979–0.984). Similar patterns were observed when heatwave exposure was defined using episodes of ≥ 2 or ≥ 3 consecutive heatwave days (Table 2). Results were largely consistent after controlling for seasonality, with the same magnitude and direction of associations observed across outcomes and trimesters (Table 3).

Each unit (µg/m³) increase in PM_2.5 concentration during the first trimester was associated with a higher risk of PTB (aRR 1.010, 95% CI 1.008–1.012) (Table 4). Similar positive associations were observed for second and third trimester exposure and remained after controlling for seasonality (Table 5). In contrast, each unit (µg/m³) increase in PM_2.5 concentration during the first trimester was inversely associated with risk of LBW (aRR 0.988, 95% CI 0.984–0.992) (Table 4). This pattern was also observed for second and third trimester exposure and remained after controlling for seasonality (Table 5).

3.3 Effect modification by AC type

We did not observe significant effect modification by AC type for one, two, or three consecutive heatwave days in any trimester on PTB and LBW (Supplementary Table 1 and 2). Overall, the direction and magnitude of associations observed in AC-stratified analyses were generally consistent with the primary results. In the first trimester, we observed an increased risk of PTB (aRR 1.002, 95% CI 1.001–1.002) and LBW (aRR 1.002, 95% CI 1.000–1.003) for two consecutive heatwave days among participants with central AC, which was consistent with the increased PTB risk observed in the overall models; no corresponding association was observed among participants without central AC (aRR 1.000, 95% CI 0.982–1.018). While these results were also seen when examining three consecutive heatwave days, we observed an inverse association with PTB and LBW for one, two, and three consecutive heatwave days in the second trimester, similar to the associations observed in the overall results. In the third trimester, we observed a similar inverse association with PTB among those with no and other AC, as well as an inverse association with LBW among those with other AC, again consistent with the decreased risks observed in the primary analyses. Results were similar when the distance to AC was filtered to less than ten meters.

3.4 Sensitivity analyses

The results from the first sensitivity analysis assessing associations between heatwave days at the 95th and 98th percentile and PTB and LBW were consistent with that of our primary analysis (Supplementary Table 3 and 4). Similar to the main analysis, heatwave exposure in the first trimester was associated with a higher risk of PTB, while heatwave exposure in the second and third trimesters was inversely associated, and heatwave exposure in the third trimester was inversely associated with risk of LBW. Results from the second sensitivity analysis controlling for seasonality were also largely consistent with the primary analysis, with similar increased risk of PTB in the first trimester and decreased risks of PTB and LBW in later trimesters (Supplementary Table 5 and 6). However, for heatwaves defined using the 98th percentile, we observed an increased risk of PTB in the second trimester when examining two consecutive heatwave days (aRR 1.005, 95% CI 1.000–1.009) and three consecutive heatwave days (aRR 1.008, 95% CI 1.003–1.014) (Supplementary Table 6), whereas the primary analysis showed an inverse association with PTB in the second trimester when controlling for seasonality. In our third sensitivity analysis assessing associations between PM_2.5 and LBW regardless of gestational age, results were not consistent with the primary analysis. Positive associations were observed for all trimesters and remained after controlling for seasonality (Supplementary Table 7 and 8). When heatwave exposures were defined using the non-detrended temperature, associations were generally attenuated compared to those obtained using detrended temperature (Supplementary Table 9). For PTB, small inverse associations were observed during the first and second trimesters across the three-heatwave definitions. Third trimester associations were substantially attenuated towards the null compared to the stronger inverse associations observed in the primary analysis. Associations with LBW were also generally small, with estimates near the null in the third trimester. In analyses incorporating both temperature and humidity, we observed no evidence of an association between heat index and PTB across all exposure windows (Supplementary Table 10). Each 1 degree Celsius increase in average heat index was associated with an increased risk of LBW during the 28 days preceding delivery (aRR 1.003, 95% CI 1.001–1.005) (Supplementary Table 10). Estimates were similar across the 7, 14, and 21 days preceding delivery.

4 Discussion

We examined the associations between extreme heat and PM_2.5 concentrations and the risk of PTB and term LBW. For PTB, we observed an increased risk associated with heatwave days during the first trimester and a decreased risk in the second and third trimesters across all heatwave definitions. Similar patterns were observed when seasonality was controlled. For term LBW, heatwave days had no effect when examining one, two, or three consecutive heatwave days overall or during the first and second trimesters. A decreased risk of term LBW was observed in the third trimester across all heatwave definitions. Furthermore, we observed an increased risk of PTB and a decreased risk of LBW with increased PM_2.5 concentrations. No evidence of heatwave effect modification by AC type was observed.

Our findings for extreme heat exposure demonstrates that early pregnancy may represent a possible window of susceptibility. Epidemiologic studies have identified early gestation as a potential window of susceptibility for PTB, specifically from heat exposure (37). The first trimester is a critical stage for placental development, and environmental stressors such as exposure to extreme heat and air pollution may disrupt placentation and maternal vascular adaptation through pathways involving oxidative stress, systemic inflammation, and altered vascular function (38). Evidence from previous studies has indicated extreme heat exposure is associated with markers of placental function, including lower placental growth factor concentrations (39, 40). Similarly, PM_2.5 exposure has been shown to impair placental vascularization and induce oxidate stress in experimental studies (41). Perturbations occurring during this early developmental window could have consequences that persist later in pregnancy and contribute to adverse outcomes such as preterm delivery.

As previous research has been inconsistent, some of our findings align with prior studies, particularly the increased risk of PTB associated with heatwave exposure in the first trimester, while others differ from previously reported associations. Numerous studies have explored associations between extreme heat and adverse birth outcomes, with many reporting elevated risks of PTB and term LBW (12, 42, 43). A recent meta-analysis concluded that exposure to heatwaves throughout pregnancy increased the odds of PTB (OR 1.05, 95% CI 1.03–1.07) (44). A 2019 study explored susceptible windows in the association between heatwave and PTB and observed positive associations between heatwave exposure and all gestational months. Similar to the present study, heatwaves were defined using daily maximum temperature at the 90th and 95th percentile for 2, 3, or 4 consecutive days (45). While these findings are generally consistent with ours, other studies have reported differing results. A 2018 cohort study observed that apparent temperature exposure was positively associated with term LBW in the second (1.7%, 95% CI 0.3–3.2) and third trimester (15.8%, 95% CI 5.0–27.6) (46). However, this study's temperature metric considered relative humidity in addition to temperature. An Australian cohort study found positive associations between extreme daily mean temperature at the 95th percentile and PTB in the third trimester (OR 1.61, 95% CI 1.55–1.67) and no associations in the first or second trimester (47). Heatwaves in this study were defined as an average of the daily mean temperature during each trimester, whereas our study used daily maximum temperature. A cohort study in California found an increased risk of PTB in those who experienced extreme heat during their last week of gestation during the warm months (May through September) at the 90th, 95th, and 98th percentiles (48). This study defined heatwaves as at least 2, 3, or 4 consecutive days during the last week of gestation, while the present study considered the entire third trimester. These studies highlight that while relationships between heatwaves and adverse birth outcomes are frequently observed, differences in exposure metrics and regional climate conditions likely contribute to the mixed findings.

When examining relationships between PM_2.5 exposure and PTB and term LBW, our findings for PTB are supported by a broad base of epidemiologic evidence, whereas evidence for term LBW was less consistent. Previous research has generally reported positive associations between maternal PM_2.5 exposure and adverse birth outcomes, including increased risks of PTB and LBW in large population-based and systematic review studies (49–53). However, in the present study we observed a decreased risk of term LBW with increasing PM_2.5 concentrations, which differs from many previously reported findings.

The magnitude of the observed associations was relatively small, consistent with studies relying on population-level exposure measures. However, several findings differed from those reported in prior studies, particularly the inverse associations observed between heatwave exposure and PTB in later trimesters and the decreased risk of term LBW with increasing PM_2.5 concentrations. Several factors may contribute to the inverse associations observed between heatwave exposure and PTB in later trimesters. One possible explanation is the limited opportunity for pregnancies that end in preterm birth to experience exposures later in gestation. Because preterm delivery shortens the duration of pregnancy, pregnancies that deliver earlier may have less opportunity to be exposed to heatwaves in the second and third trimesters, which could bias estimates toward an apparent inverse association. However, the observed associations during the second and third trimesters may reflect differences in opportunities for exposure rather than a protective effect of extreme heat. Our results should be interpreted with caution and not as evidence that heatwave exposure reduces the risk of preterm birth. Analytic approaches that account for gestational time at risk and treat heat as time-varying may be needed to evaluate associations between second and third trimester extreme heat exposure and time of delivery Additionally, exposure misclassification may be greater in later pregnancy if mothers alter daily routines or spend more time indoors. The inverse association observed between PM_2.5 exposure and term LBW may also reflect the relatively low levels and limited variability of PM_2.5 concentrations observed in this population, which may differ from regions with higher pollution levels where positive associations have more commonly been reported. Residual confounding by unmeasured socioeconomic characteristics, healthcare utilization, and seasonal factors may also influence these associations.

One reason for differences in our results compared to other studies could be inconsistencies in heatwave definitions. Heatwaves are commonly defined using combinations of temperature metrics (e.g., maximum, mean, or apparent temperature), percentile-based thresholds, and minimum durations of consecutive days, and these methodological choices can influence estimated health effects. A 2023 systematic review of the impact of heatwave definition on cardiovascular health noted heterogeneity in not just the definition of heat, but also in the meteorological data used, the sample size, and methodology for analysis (54). A 2015 study in China showed that heatwave definitions based on daily mean temperature generally had higher mortality risks that definitions based on daily maximum temperature, suggesting that mean temperature was more predictive than maximum temperature (55). As the present study used daily maximum temperature, this may partially explain why our results differed from the Australian cohort study that utilized daily mean temperature. If mean temperature better reflects sustained thermal exposure across the day and night, this difference in exposure metric could influence observed associations with birth outcomes. Previous research has also observed conflicting results with regards to the temperature threshold used in defining a heatwave. In the present study, risk of PTB was found to increase with temperature threshold (90th, 95th, and 98th), similar to other studies that have observed increased health effects as the threshold increased (55–57). However, a few studies found lower effects with higher thresholds (98th: 13.5%, 95% −0.1–28.9; 99th: 0.1%, 95% −14.3–16.9) (58), which the authors suggested may reflect the small number of heatwave events occurring at very high temperature thresholds, resulting in reduced statistical power and more unstable effect estimates.

The absence of effect modification by residential AC type warrants further consideration. Although we observed no evidence of effect modification by estimated AC type, this finding should not be interpreted as evidence that AC does not mitigate the health effects of extreme heat. AC was measured using residential AC type rather than AC use or indoor temperature nor captured factors such as frequency of use, affordability of cooling, or time spent outside the home. Additionally, access to and use of residential cooling may reflect broader socioeconomic circumstances. Previous studies have reported greater heat-related vulnerability among socioeconomically disadvantaged populations, suggesting a complex relationship between AC access, socioeconomic resources, and heat-related health effects (12, 44). Future research using more precise measures of indoor temperature and AC use is needed to determine whether residential cooling modifies the effects of extreme heat during pregnancy.

This study has a few limitations. Ambient temperature and PM_2.5 exposure were assigned based on the mother's residential address rather than individual exposure levels, which may result in exposure misclassification and attenuation of effect estimates. Factors such as personal activities, air conditioning prevalence, and occupation can influence an individual's level and duration of exposure to heat and PM_2.5. Additionally, although models were adjusted for several maternal and socioeconomic characteristics, residual confounding by unmeasured or unavailable maternal and obstetric factors remains a possibility. Information on mother's residential address was only available at time of delivery; therefore, we could not account for the possibility of a mother relocating during her pregnancy. Previous research as indicated between 12% to 40% of mothers move at least once during pregnancy (59–61), and using a mother's address at time of delivery may not accurately reflect her exposure during pregnancy. However, several studies addressing the impact of mobility on environmental exposures have reported that even among mothers who moved during pregnancy, there are not significant differences in exposure if using residential address at delivery rather than address at conception (61–63). Additionally, information on the AC type was only available for the year 2016 to 2021, whereas the study period spanned from 2000 to 2022. AC classification may also not accurately reflect household cooling conditions throughout pregnancy. Mothers may have installed or upgraded AC systems or relocated to residence with different cooling characteristics. The absence of observed effect modification by AC type should therefore be interpreted cautiously. Finally, most of Florida lies within the subtropical climate zone and experiences long hot and humid summers and mild and wet winters (64). Average monthly temperatures tend to range between 50°F and 85°F throughout the year (65). Therefore, our results may not be generalizable to other geographic areas with differing climates.

In this study, heatwave exposure was associated with an increased risk of PTB in the first trimester and an inverse association in later trimesters. Heatwave exposure was also associated with an inverse association of term LBW during the third trimester. PM_2.5 exposure was associated with an increased risk of PTB. Our findings also suggest that estimated associations between heat exposure and birth outcomes may depend on how heatwaves are defined, including the temperature metric and threshold used to characterize extreme heat events. These results highlight the importance of carefully considering heatwave definitions when evaluating health impacts, as differences in exposure metrics may contribute to variability in findings across studies. As climate change continues to increase the frequency, intensity, and duration of extreme heat events, improving the comparability of heatwave definitions will be critical for accurately assessing their public health impacts. Future research should also investigate how these associations vary across geographic regions and populations to better identify vulnerable groups and inform targeted public health interventions.

References

• 1.

  LindwallC. What Are the Effects of Climate Change? Natural Resources Defense Council (2022). Available online at:https://www.nrdc.org/stories/what-are-effects-climate-change#weather. (Accessed January 26, 2026).

  • Google Scholar
  • View reference in article

• 2.

  National Aeronautics and Space Administration. The Effects of Climate Change (2024). Available online at:https://science.nasa.gov/climate-change/effects/. (Accessed January 26, 2026).

  • Google Scholar
  • View reference in article

• 3.

  National Oceanic and Atmospheric Administration. Climate change impacts (2025). Available online at:https://www.noaa.gov/education/resource-collections/climate/climate-change-impacts. (Accessed January 26, 2026).

  • Google Scholar
  • View reference in article

• 4.

  First Street Foundation. Heat Model Methodology (2024). Available online at:https://firststreet.org/research-lab/published-research/heat-simplified-methodology/. (Accessed January 26, 2026).

  • Google Scholar
  • View reference in article

• 5.

  AlmazrouiMIslamMNSaeedFSaeedSIsmailMEhsanMAet al. Projected changes in temperature and precipitation over the United States, Central America, and the Caribbean in CMIP6 GCMS. Earth Syst and Environ. (2021) 5:1–24. 10.1007/s41748-021-00199-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 6.

  EbiKLMcGregorG. Climate change, tropospheric ozone and particulate matter, and health impacts. Environ Health Perspect. (2008) 116:1449–55. 10.1289/ehp.11463

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7.

  LeeYCShindellDTFaluvegiGWenigMLamYFNingZet al. Increase of ozone concentrations, its temperature sensitivity and the precursor factor in south China. Tellus B. (2022) 66:23455. 10.3402/tellusb.v66.23455

  • CrossRef
  • Google Scholar
  • View reference in article

• 8.

  BasuRSametJM. Relation between elevated ambient temperature and mortality: a review of epidemiologic evidence. Epidemiol Rev. (2002) 24:190–202. 10.1093/epirev/mxf007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9.

  GreeneSKalksteinLSMillsDMSamenowJ. An examination of climate change on extreme heat events and climate-mortality relationships in large U.S. Cities. Wea Climate Soc. (2011) 3:281–92. 10.1175/WCAS-D-11-00055.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 10.

  BasuRMaligBOstroB. High ambient temperature and the risk of preterm delivery. Am J Epidemiol. (2010) 172:1108–17. 10.1093/aje/kwq170

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 11.

  KentSTMcClureLAZaitchikBFSmithTTGohlkeJM. Heat waves and health outcomes in Alabama (USA): the importance of heat wave definition. Environ Health Perspect. (2014) 122:151–8. 10.1289/ehp.1307262

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 12.

  DarrowLAHuangMWarrenJLStricklandMJHolmesHANewmanAJet al. Preterm and early -term delivery after heat waves in 50 US metropolitan areas. JAMA Netw. Open. (2024) 7(5):e2412055. 10.1001/jamanetworkopen.2024.12055

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13.

  BonellASonkoBBadjieJSamatehTSaidyTSossehFet al. Environmental heat stress on maternal physiology and fetal blood flow in pregnant subsistence farmers in the Gambia, West Africa: an observational cohort study. Lancet Planet Health. (2022) 6:e968–76. 10.1016/S2542-5196(22)00242-X

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 14.

  SamuelsLNakstadBRoosNBonellAChersichMHavenithGet al. Physiological mechanisms of the impact of heat during pregnancy and the clinical implications: review of the evidence from an expert group meeting. Int J Biometeorol. (2022) 66(8):1505–13. 10.1007/s00484-022-02301-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 15.

  JiaoASunYSacksDAvilaCChiuVMolitorJet al. The role of extreme heat exposure on premature rupture of membranes in southern California: a study from a large pregnancy cohort. Environ Int. (2023) 173:107824. 10.1016/j.envint.2023.107824

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16.

  Environmental Protection Agency. Particulate Matter (PM) Basics (nd). Available online at:https://www.epa.gov/pm-pollution/particulate-matter-pm-basics. (Accessed January 31, 2026).

  • Google Scholar
  • View reference in article

• 17.

  KaurKLesseurCDeyssenrothMKloogISchwartzJDMarsitCJet al. PM2.5 Exposure during pregnancy is associated with altered placental expression of lipid metabolic genes in a US birth cohort. Environ Res. (2022) 211:113066. 10.1016/j.envres.2022.113066

  • CrossRef
  • Google Scholar
  • View reference in article

• 18.

  FussellJJauniauxESmithRBurtonGJ. Ambient air pollution and adverse birth outcomes: a review of underlying mechanisms. BJOG. (2024) 131:538–50. 10.1111/1471-0528.17727

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 19.

  WangWMuSYanWKeNChengHDingR. Prenatal PM2.5 exposure increases the risk of adverse pregnancy outcomes: evidence from meta-analysis of cohort studies. Environ Sci Pollut Res. (2023) 30:106145–97. 10.1007/s11356-023-29700-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 20.

  SmithTTZaitchikBFGohlkeJM. Heat waves in the United States: definitions, patterns and trends. Clim Change. (2013) 118:811–25. 10.1007/s10584-012-0659-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21.

  XuZFitzGeraldGGuoYJalaudinBTongS. Impact of heatwave on mortality under different heatwave definitions: a systematic review and meta-analysis. Environ Int. (2016) 89-90:193–203. 10.1016/j.envint.2016.02.007

  • CrossRef
  • Google Scholar
  • View reference in article

• 22.

  LanLCuiGYangCWangJSuiCXuGet al. Increased mortality during the 2010 heat wave in Harbin, China. EcoHealth. (2012) 9:310–4. 10.1007/s10393-012-0790-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 23.

  WangXYBarnettAGWeiweiYFitzGeraldGTippettVAitkenPet al. The impact of heatwaves on mortality and emergency hospital admissions from non-external causes in Brisbane, Australia. Occup Environ Med. (2012) 69:163–9. 10.1136/oem.2010.062141

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24.

  TongSWangXYBarnettAG. Assessment of heat-related health impacts in Brisbane, Australia: comparison of different heatwave definitions. PLoS One. (2010) 5:e12155. 10.1371/journal.pone.0012155

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 25.

  JayOCaponABerryPBroderickCDe DearRHavenithGet al. Reducing the health effects of hot weather and heat extremes: from personal cooling strategiesto green cities. Lancet. (2021) 398:709–24. 10.1016/S0140-6736(21)01209-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 26.

  KravchenkoJAbernethyAPFawzyMLyerlyK. Minimization of heatwave morbidity and mortality. Am J Prev Med. (2013) 44:274–82. 10.1016/j.amepre.2012.11.015

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 27.

  NunesBPaixãoEDiasCMNogueiraPMarinho FalcãoJ. Air conditioning and intrahospital mortality during the 2003 heatwave in Portugal: evidence of a protective effect. Occup Environ Med. (2011) 68:218–23. 10.1136/oem.2010.058396

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 28.

  CampbellPCTongDQChangSMaSLiYDingJet al. Increased contributions of climate-driven wildfires to nitrogen deposition in the United States. Commun Earth Environ. (2026) 7:254. 10.1038/s43247-026-03279-4

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 29.

  AhnYLeykSUhlJHMcShaneCM. An integrated multi-source dataset for measuring settlement evolution in the United States from 1810 to 2020. Sci Data. (2024) 11:275. 10.1038/s41597-024-03081-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30.

  AhnYUejioCK. A comprehensive dataset of residential air conditioning prevalence in the continental United States. Sci Data. (2025) 12:1717. 10.1038/s41597-025-06104-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31.

  AlcockSLealMBeukesJNyatiLHThompsonUNorrisSA. Maternal age and parity influences on health outcomes: a multivariable regression analysis of mothers and infants. BMC Pregnancy Childbirth. (2025) 25:1094. 10.1186/s12884-025-08194-8

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 32.

  SuDSamsonKHansonCAnderson BerryALLiYShiLet al. Racial and ethnic disparities in birth outcomes: a decomposition analysis of contributing factors. Prev Med Rep. (2021) 23:101456. 10.1016/j.pmedr.2021.101456

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 33.

  CowellWHsuHLJustAC. Air pollution exposure and birth weight in the ECHO cohort. JAMA Network. (2025) 8(12):e2551459. 10.1001/jamanetworkopen.2025.51459

  • CrossRef
  • Google Scholar
  • View reference in article

• 34.

  BeltranAJWuJLaurentO. Associations of meteorology with adverse pregnancy outcomes: a systematic review of preeclampsia, preterm birth and birth weight. Int J Environ Res Public Health. (2014) 11:91–172. 10.3390/ijerph110100091

  • CrossRef
  • Google Scholar
  • View reference in article

• 35.

  JardelHMartinCLHoyoCRappazzoKM. Interplay of gestational parent exposure to ambient air pollution and diet characteristics on preterm birth. BMC Public Health. (2023) 23:822. 10.1186/s12889-023-15676-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 36.

  BeesleyLJBondarenkoIElliottMRKurianAWKatzSJTaylorJMG. Multiple imputation with missing data indicators. Stat Methods Med Res. (2021) 30:2685–700. 10.1177/09622802211047346

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 37.

  RenMZhangCDiJChenHHuangAJiJSet al. Exploration of the preterm birth risk-related heat event thresholds for pregnant women: a population-based cohort study in China. Lancet Reg Health West Pac. (2023) 37:100785. 10.1016/j.lanwpc.2023.100785

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 38.

  ParisiFFeniziaCIntroiniAZavattaAScaccabarozziCBiasinMet al. The pathophysiological role of estrogens in the initial stages of pregnancy: molecular mechanisms and clinical implications for pregnancy outcome from the periconceptional period to end of the first trimester. Hum Reprod Update. (2023) 29(6):699–720. 10.1093/humupd/dmad016

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 39.

  ChenQWangYLiJGuAZhaiX. Cumulative effects of extreme ambient temperatures on placental perfusion and function markers in early pregnancy: analysis from a birth cohort study. Environ Res. (2025) 269:120930. 10.1016/j.envres.2025.120930

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 40.

  ChenQWangYLiJGuAZhaiX. Association of ambient temperature in early pregnancy with placental markers and HDP risk. Eur J Public Health. (2025) 35(Suppl 4):ckaf161.238. 10.1093/eurpub/ckaf161.238

  • CrossRef
  • Google Scholar
  • View reference in article

• 41.

  YueHJiXZhangYLiGSangN. Gestational exposure to PM2.5 impairs vascularization of the placenta. Sci Total Environ. (2019) 665:153–61. 10.1016/j.scitotenv.2019.02.101

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 42.

  YuGYangLLiuMWangCShenXFanLet al. Extreme temperature exposure and risks of preterm birth subtypes based on a nationwide survey in China. Environ Health Perspect. (2023) 131:87009. 10.1289/EHP10831

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 43.

  MohammadiDNaghshinehESarsangiASakhvidiMJZ.. environmental extreme temperature and daily preterm birth in sabzevar, Iran: a time-series analysis. Environ Health Prev Med. (2019) 21:5. 10.1186/s12199-018-0760-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 44.

  ChersichMFPhamMDArealAHaghighiMMManyuchiASwiftCPet al. Associations between high temperatures in pregnancy and risk of preterm birth, low birth weight, and stillbirths: systematic review and meta-analysis. Br Med J. (2020) 371:m3811. 10.1136/bmj.m3811

  • CrossRef
  • Google Scholar
  • View reference in article

• 45.

  WangJTongSWilliamsGPanX. Exposure to heat wave during pregnancy and adverse birth outcomes: an exploration of susceptible windows. Epidemiol. (2019) 30:S115–21. 10.1097/EDE.0000000000000995

  • CrossRef
  • Google Scholar

• 46.

  BasuRRauRPearsonDMaligB. Temperature and term low birth weight in California. Am J Epidemiol. (2018) 187:2306–14. 10.1093/aje/kwy116

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47.

  YeTGuoYHuangWZhangYAbramsonMJLiS. Heat exposure, preterm birth, and the role of greenness in Australia. JAMA Pediatr. (2024) 178:376–83. 10.1001/jamapediatrics.2024.0001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48.

  IlangoSDWeaverMSheridanPSchwarzLClemeshaRESBrucknerTet al. Extreme heat episodes and risk of preterm birth in California, 2005–2013. Environ Int. (2020) 137:105541. 10.1016/j.envint.2020.105541

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 49.

  SongSGaoZZhangXZhaoXChangJZhangJet al. Ambient fine particulate matter and pregnancy outcomes: an umbrella review. Environ Res. (2023) 235:116652. 10.1016/j.envres.2023.116652

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50.

  ParasinNAmnuaylojaroenTSaokaewS. Prenatal PM2.5 exposure and its association with low birth weight: a systematic review and meta-analysis. Toxics. (2024) 12:446. 10.3390/toxics12070446

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51.

  GuoTWangYZhangHZhangYZhaoJWangQet al. The association between ambient PM2.5 exposure and the risk of preterm birth in China: a retrospective cohort study. Sci Total Environ. (2018) 633:1453–9. 10.1016/j.scitotenv.2018.03.328

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52.

  ZhouWMingXYangYHuYHeZChenHet al. Association between maternal exposure to ambient air pollution and the risk of preterm birth: a birth cohort study in Chongqing, China, 2015–2020. Int J Environ Res Public Health. (2022) 19:2211. 10.3390/ijerph19042211

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53.

  BekkarBPachecoSBasuRDeNicolaN. Association of air pollution and heat exposure with preterm birth. Low birth weight, and stillbirth in the US: a systematic review. JAMA Netw Open. (2020) 3:e208243. 10.1001/jamanetworkopen.2020.8243

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54.

  NawaroJGianquintieriLPagliosaASechiGMCaianiEG. Heatwave definition and impact on cardiovascular health: a systematic review. Public Health Rev. (2023) 44:1606266. 10.3389/phrs.2023.1606266

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55.

  ChenKBiJChenJChenXHuangLZhouL. Influence of heat wave definitions to the added effect of heat waves on daily mortality in Nanjing, China. Sci Total Environ. (2015) 506-507:18–25. 10.1016/j.scitotenv.2014.10.092

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 56.

  ZengWLaoXRutherfordSXuYXuXLinHet al. The effect of heat waves on mortality and effect modifiers in four communities of Guangdong Province, China. Sci Total Environ. (2014) 482-483:214–21. 10.1016/j.scitotenv.2014.02.049

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 57.

  AndersonBGBellM. Weather-Related mortality: how heat, cold, and heat waves affect mortality in the United States. Epidemiology. (2009) 20:205–13. 10.1097/EDE.0b013e318190ee08

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58.

  SonJLeeJAndersonBBellML. The impact of heat waves on mortality in seven Major cities in Korea. Environ Health Perspect. (2012) 120(4):566–71. 10.1289/ehp.1103759

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59.

  SaadehFBClarkMARogersMLLinkletterCDPhippsMGPadburyJFet al. Pregnant and moving: understanding residential mobility during pregnancy and in the first year of life using a prospective birth cohort. Matern Child Health J. (2013) 17:330–43. 10.1007/s10995-012-0978-y

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60.

  FellDBDoddsLKingWD. Residential mobility during pregnancy. Paediatr Perinat Epidemiol. (2004) 18:408–14. 10.1111/j.1365-3016.2004.00580.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 61.

  HodgsonSLurzPWWShirleyMDFBythellMRankinJ. Exposure misclassification due to residential mobility during pregnancy. Int J Hyg Environ Health. (2015) 218:414–21. 10.1016/j.ijheh.2015.03.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 62.

  ChenLBellEMCatonARDruschelCMLinS. Residential mobility during pregnancy and the potential for ambient air pollution exposure misclassification. Environ Res. (2010) 110:162–8. 10.1016/j.envres.2009.11.001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 63.

  LupoPJSymanskiEChanWMitchellLEWallerDKCanfieldMAet al. Differences in exposure assignment between conception and delivery: the impact of maternal mobility. Paediatr Perinat Epidemiol. (2010) 24:200–8. 10.1111/j.1365-3016.2010.01096.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64.

  WinsbergMD. Climate of Florida. National Climatic Data Center (2011). Available online at:https://climatecenter.fsu.edu/images/fcc/climateofflorida.pdf. (Accessed February 12, 2026).

  • Google Scholar

• 65.

  Florida State University. Temperature. Florida Climate Center (n.d.). Available online at:https://climatecenter.fsu.edu/products-services/data/statewide-averages/temperature. (Accessed February 12, 2026).

  • Google Scholar