1. Introduction
Lightning is one of the deadliest and most destructive hydrometeorological phenomena that seriously threatens humans and animals, causing fatalities and property damage (Cummins and Murphy 2009; Gutiérrez-Anguamea et al. 2023). It can also disrupt economic and social activities across various sectors, including health, insurance, forestry, electricity, agriculture, telecommunications, transportation, tourism, and recreation (Dlamini 2009; Mills et al. 2010; Holle and Cooper 2016; Yair 2018; Gutiérrez-Anguamea et al. 2023; Rahman et al. 2023). Lightning is responsible for many deaths worldwide, with estimates ranging from 6000 to 24 000 deaths per year and 10 times as many injuries globally (Holle and Cooper 2016; Cooper and Holle 2019a). This makes it a significant public health issue worldwide (Biswas et al. 2016; World Meteorological Organization 2021; Rahman et al. 2023).
Despite the many deaths and economic losses attributed to lightning, lightning is often underestimated as a hazard, especially in developing countries, where communities may lack awareness and infrastructure to mitigate risks (Dlamini 2009; Kabir and Jakariya 2021; Rahman et al. 2022). While lightning-related deaths are usually considered minor-scale climate-related disasters due to their lesser impact on individual events compared to major disasters like hurricanes (Sabur 2012; Raga et al. 2014; Kabir and Jakariya 2021), they accumulate over time, exerting a substantial impact (Kabir and Jakariya 2021).
Misinformation and lack of awareness about lightning safety measures further exacerbate the social and economic damage caused by lightning. Recognizing lightning as a severe threat to human life and property is crucial (Kabir and Jakariya 2021). Being struck by lightning is not a random occurrence or a matter of luck. It requires the presence of lightning at the same location as the person, who must also be vulnerable to lightning (Holle and Zhang 2023). Previous studies have highlighted that many lightning-related deaths may result from a mixture of natural and anthropological characteristics of the population. These studies have evaluated the hazard posed by lightning through diverse approaches, such as regional analysis of lightning behavior, examination of fatalities, and consideration of the social and physical conditions of the population (e.g., Raga et al. 2014; Roeder et al. 2015; Holle and Cooper 2016; Biswas et al. 2020; Islam and Schmidlin 2020; Farfán et al. 2023; Gutiérrez-Anguamea et al. 2023; Rahman et al. 2023).
Developed and developing countries exhibit notable differences regarding lightning-related fatalities and injuries. Developed nations exhibit lower rates of such incidents compared to their developing counterparts, owing to advancements in lightning detection technology, lightning-safety infrastructure, heightened public awareness, improved socioeconomic conditions, and educational initiatives (Rakov 2012; Yair 2018; Gomes and Gomes 2021; Holle et al. 2021; Nazri et al. 2021; Islam and Schmidlin 2020). In contrast, developing countries face distinct challenges, such as high population density and the prevalence of labor-intensive agriculture, which expose the rural population to lightning hazards throughout the day, unsafe dwellings, and inadequate public awareness regarding lightning hazards (Holle and Cooper 2016; Dewan et al. 2017; Holle et al. 2019; Islam and Schmidlin 2020). It is essential to address these challenges to improve safety and reduce lightning-related incidents in developing countries.
Mexico has a wide range of climates, from tropical to subtropical. Mexican climate is characterized by a distinct rainy season marked by frequent thunderstorms (Christian et al. 2003; Kucieńska et al. 2010). This exposure to lightning makes Mexico notably vulnerable, especially when considering the higher risk faced by developing countries in tropical and subtropical regions with rural populations (Biswas et al. 2016). However, despite this exposure, there are limited comprehensive studies on the hazards of lightning in the region. Previous studies, such as Raga et al. (2014) and Gutiérrez-Anguamea et al. (2023), have examined the geographical distribution of lightning-related deaths in Mexico. These studies have found that the occurrence of such deaths is not necessarily correlated with areas of high population density or lightning density. Instead, they are linked to underlying vulnerabilities (Raga et al. 2014; Gutiérrez-Anguamea et al. 2023).
This insight highlights the need to comprehend the factors contributing to lightning-related risks within the Mexican context. Previous approaches to estimate lightning risk in Mexico have primarily focused on hazard considerations, such as the number of thunderstorm days to measure risk [e.g., Centro Nacional de Prevención de Desastres (CENAPRED) 2012]. However, risk emerges from the interplay between physical variables and social conditions, demanding a comprehensive understanding that encompasses both the interaction of hazard and vulnerability (Aven 2016; Oliver-Smith et al. 2017; Alcántara-Ayala 2019; Sutton 2019; Dominguez et al. 2021). Given Mexico’s significant exposure to lightning and the heightened vulnerability of its population, especially in rural areas, a comprehensive understanding of lightning risk is essential for informed decision-making and the development of targeted prevention measures.
The primary aim of this study is to determine the lightning risk at the municipal level across Mexico, integrating both physical and social dimensions. We consider the hazard represented by the average number of days with lightning activity in each region. At the same time, a social index is used to characterize social vulnerability. This index incorporates various socioeconomic conditions, such as education level, access to essential services, housing quality, and poverty rates. To ensure the accuracy of the risk estimation methodology, we compare the estimated lightning risk with the officially documented lightning-related fatalities by the Mexican government on a per-municipality basis.
By combining physical and social dimensions, our research aims to offer insights into understanding lightning risk, which could result in more effective and targeted risk mitigation strategies. Integrating demographic data in assessing lightning risk can help target the most vulnerable population (Rahman et al. 2023). Moreover, analyzing lightning risk at the municipality level can provide detailed insights for policymakers and emergency responders, aiding in the development of measures to minimize the impact of lightning.
2. Data and methods
a. Lightning-related fatalities database
This study focuses exclusively on lightning fatalities due to the absence of official reports on lightning-related injuries in Mexico. Data on lightning fatalities from 1998 to 2021 were obtained from the Department for Health Information (DGIS, by its acronym in Spanish), which is operated by the Mexican Secretariat of Health (DGIS 2023). This database systematically compiles death reports, primarily documented by hospitals and healthcare centers throughout Mexico. Lightning-related fatalities are identified in the official Mexican records using codes X330–X339 (the meanings of these codes are provided in Table 1). This database not only includes details on the locations of the incidents, specifying the state and municipality, but also provides additional information concerning the demographic profile of the victims, encompassing their gender, age, educational background, and economic activity. This database provides relevant information for understanding the various social factors associated with lightning strike fatalities in Mexico.
Codes associated with deaths from lightning in the official records of the DGIS database.
Lightning fatalities are generally underreported, especially in rural areas where people frequently underestimate the associated risks (López et al. 1993; Biswas et al. 2016; Holle et al. 2019; Rahman et al. 2023). Moreover, a single lightning strike can cause multiple victims, leading to underestimating lightning’s potential damage and consequences (Roeder et al. 2015; Gutiérrez-Anguamea et al. 2023). This presents a significant challenge when relying on official records to evaluate the risk of lightning strikes. For instance, in some municipalities, the number of deaths attributed to a single lightning strike can overestimate the risk. In others, the actual number of fatalities might be underreported, leading to an underestimation of the risk. To address this issue, we considered the reported deaths per municipality and estimated the incidence of lightning strike reports per municipality (incidents per municipality). To make this estimation, we considered reported deaths occurring on the same day in the same municipality as potentially resulting from the same lightning event. This approach allows us to distinguish between municipalities where lightning fatalities are more frequent, regardless of the number of deaths per incident, and those where such incidents are less common but involve multiple fatalities. By doing so, we can better understand and address the risk associated with lightning strikes across different regions.
Furthermore, to highlight the long-term cumulative impact of lightning fatalities compared to other disasters categorized as major catastrophes with higher visibility and risk, we examine reported deaths associated with meteorological-related disasters in Mexico during the study period from the Emergency Events Database (EM-DAT). The EM-DAT database was created in 1988 by the Centre for Research on the Epidemiology of Disasters at the Université Catholique of Louvain. This database defines disasters as unexpected and overwhelming events that cause harm to humans. EM-DAT includes disasters with at least one of the following: 10 or more deaths, a minimum of 100 people affected, or a call for international assistance or emergency declaration. Additionally, secondary criteria are applied for past events lacking quantitative data, such as being the worst disaster in a region or causing significant damage (EM-DAT 2023).
b. Lightning density and lightning days
Lightning data for our study were obtained from the World Wide Lightning Location Network (WWLLN) Global Lightning Climatology (WGLC) developed by Kaplan and Lau (2021). This free, gridded database, adjusted for detection efficiency, is based on data from the (WWLLN) and spans from 2010 to 2021. The dataset provides daily cloud-to-ground lightning strokes per square kilometer at a spatial resolution of 0.5° (Kaplan and Lau 2021).
While acknowledging potential biases in lightning detection (e.g., detection efficiency and ground base station dependency), the WGLC offers notable advantages for our study compared to alternative databases. These advantages include free access, an extended data period, improved spatial resolution, and broader geographical coverage. This distinguishes WGLC from databases characterized by shorter observation periods, limited spatial precision, or associated costs (Kaplan and Lau 2021). Furthermore, the underlying WWLLN database, upon which WGLC relies, has undergone evaluation in our study region. WWLLN has low detection efficiency, but it exhibits consistent monthly variability, aligning with findings from other databases (Kucieńska et al. 2010, 2012; Raga et al. 2014).
Using daily data from the WGLC, we estimated the average annual number of days with lightning activity (lightning days), counting the days with lightning density different from zero. Since the annual number of days is a discrete variable, it possesses the advantage of being less sensitive to deficiencies in the lightning density data from WGLC. Moreover, we consider it a crucial variable, as demonstrated in previous studies where deaths in Mexico were not associated with lightning density (e.g., Raga et al. 2014; Gutiérrez-Anguamea et al. 2023). However, we hypothesize that locations experiencing more lightning days, even with low lightning density, present a greater risk to the population engaged in outdoor activities (i.e., people living in rural communities).
c. Census data and vulnerability index per locality
Population data from 1998 to 2021 were obtained from the official census and intercensal records in Mexico conducted by the National Institute for Statistics and Geography (INEGI 2020). The data were used to estimate a linear increase in population based on information gathered from official population censuses conducted in 2000, 2010, and 2020 and intercensal surveys conducted in 1995, 2005, and 2015.
Vulnerability encompasses the social, economic, and physical conditions that increase the susceptibility of communities to the impacts of natural hazards, reflecting their preparedness to face such events (Dominguez et al. 2021). Based on the suggestions made by Raga et al. (2014) and Gutiérrez-Anguamea et al. (2023), which imply a connection between lightning fatalities in Mexico and underlying vulnerabilities, such as education levels and socioeconomic activities, we constructed a social vulnerability index based on the normalized marginalization index (NMI) computed by Consejo Nacional de Población (CONAPO 2020). The marginalization index measures the extent to which the population has access to education, suitable housing, and medium-to-high labor incomes (CONAPO 2020). Since this work is focused on the municipality level, it is essential to note that lightning-related fatalities are specifically localized at small localities within a municipality. This implies a potential for significant biases when employing social indices (used for depicting social vulnerability) at the municipality level, as fatalities may concentrate in localities within the municipalities with elevated levels of marginalization. However, municipalities can exhibit, on average, a lower degree of marginalization when compared to single localities. This tendency is particularly noteworthy in rural regions surrounding urban centers within municipalities.
To address this issue, we use the NMI at the locality level, encompassing 108 144 localities nationwide. The NMI ranges from 0 to 1 and differentiates localities based on the overall impact of deficiencies experienced by the population due to the lack of access to primary education, residing in inadequate housing, and the consequences of lacking essential goods. Higher values of NMI indicate the lowest marginalization levels in a locality (e.g., highest access to adequate housing and public services) (CONAPO 2020). Since higher values of an index representing vulnerability are expected to correspond to the highest marginalization levels in a locality, we define the vulnerability index to counter this trend as 1 − NMI. Consequently, the vulnerability is expressed on a scale from 0 to 1. A zero index denotes low social vulnerability, while one represents high social vulnerability.
d. Methodology for quantifying lightning risk at locality and municipality level
Fatalities caused by lightning incidents often arise from a complex interplay of both natural and social factors within the population (Raga et al. 2014; Biswas et al. 2020; Islam and Schmidlin 2020; Gutiérrez-Anguamea et al. 2023; Rahman et al. 2023). Some studies conceptualize climate risk as a combined effect between atmospheric variables and social conditions (e.g., Oliver-Smith et al. 2017; Alcántara-Ayala 2019; Sutton 2019; Dominguez et al. 2021). Consequently, we hypothesize that the assessment of lightning risk can be enclosed as a function dependent on a natural hazard component, represented by the normalized number of lightning days, and a social vulnerability, representing the inherent vulnerabilities of the population to such atmospheric phenomena.
We calculated the locality level risk (LR) using the 108 144 localities where the NMI is available. We represent the LR as the normalized product of the hazard by the vulnerability, where LR ranges from 0 to 1. A value of 0 denotes very low risk, while a value of 1 indicates the highest level of risk. The hazard at the locality corresponds to the normalized mean lightning days for the pixel where the locality is centered, and the vulnerability is estimated from the NMI value at the given locality. We categorized the LR into five ordinal levels, ranging from “very low” to “very high” risk. The classification was achieved using the Dalenius and Hodges method (Dalenius and Hodges 1959), which aims to categorize data to minimize variance within each level while maximizing it between levels (CONAPO 2020). This procedure ensures that the risk categories exhibit as much homogeneity as possible, following the same methodology of categorization used for the NMI by CONAPO (2020).
To categorize the risk at the municipality level (R), we have selected the highest LR of all the localities in the same municipality, representing the worst-case risk in the municipality. This approach accounts for the most critical conditions within each municipality, capturing the maximum risk observed at the locality level. This methodology takes a conservative approach by considering the most critical scenario. It offers a nuanced understanding of the risk distribution across municipalities and highlights the importance of addressing localized social vulnerabilities to develop effective risk management strategies.
3. Results
a. Temporal evolution and spatial distribution
Between 1998 and 2021, a total of 2573 deaths resulting from lightning incidents were reported by the Mexican government. In the same period, the EM-DAT database documented 1256 deaths in Mexico attributed to major disasters, such as flooding, tropical cyclones, and extreme temperatures. These disasters often lead to a high number of fatalities but happen in episodic events. While acknowledging that both databases may underreport deaths, it is important to emphasize the remarkable difference in the magnitude of both reports. The different orders of magnitude highlight that lightning-related incidents over an extended period have more accumulated deaths than flooding, tropical cyclones, and extreme temperatures. This high number of deaths emphasizes the need for understanding both lightning activity and social factors that conduct these deaths over time.
In Mexico, the lightning fatality rate has seen a notable decline, dropping from approximately 5.6 fatalities per million people in the early 1980s to about 0.3 fatalities per million people in 2021 (Fig. 1). Figure 1 shows the time series of lightning-related deaths per million people from 1998 to 2021, derived from official records (black line). Additionally, it includes data for the 1979–97 period, as estimated by Raga et al. (2014) (dashed gray line). The figure displays a significant reduction in reported deaths, with plausible explanations ranging from underreporting deaths to broader social transformations that have diminished the population’s exposure to the phenomenon. While it remains challenging to pinpoint specific causative factors from the available data, similar patterns have been observed in studies in countries like the United States (e.g., López and Holle 1998; Holle and Cooper 2016; Holle and Zhang 2023).
Mexico’s lightning-related deaths per million people from 1979 to 2021. The dashed gray line represents the data presented by Raga et al. (2014), and the black line represents the DGIS data for the 1998–2021 period. A dashed green line shows the percent of rural population interpolated from official population censuses and intercensal surveys. The horizontal red line marks the 0.5 deaths per million threshold.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
The horizontal line in Fig. 1 represents the 0.5 deaths per million threshold, considered a typical upper limit for more developed nations (Holle 2016; Cooper and Holle 2019a; Holle and Zhang 2023). Mexico achieved this threshold around the end of the 2010–20 decade. One key demographic aspect contributing to the decline in deaths is the diminishing percentage of the population residing in rural areas. In recent decades, there has been a discernible trend toward urbanization, with a notable concentration of most of the population in cities. This urban shift is accompanied by improved infrastructure, particularly in terms of lightning protection. Consequently, this shift in population distribution will mean a progressive decrease in lightning-related fatalities over time (Holle et al. 2005; Holle and Cooper 2016; Gutiérrez-Anguamea et al. 2023). The dashed green line in Fig. 1 (on the right axis) shows this decline in the overall population residing in rural areas.
According to census data, in the early twentieth century, 70% of Mexico’s population lived in rural areas, contrasting with the present figure of approximately 20%. This significant demographic transition underscores a substantial decrease in the proportion of the population that is most exposed to the hazards of lightning, offering a plausible explanation for the observed decline in fatalities.
The spatial distribution of accumulated fatality cases for the 1998–2021 period, both by state and municipality, is shown in Figs. 2a and 2b. Figure 2a indicates that the states registering the highest incidence are in central, northwestern (specifically in Chihuahua), and southeastern Mexico. Estado de Mexico leads with the highest number of reports (539), followed by Oaxaca (206), Michoacán (168), Guerrero (133), Veracruz (133), Puebla (131), Guanajuato (125), Chiapas (80), Chihuahua (79), and San Luis Potosí (77). Figure 2b provides a detailed view of the accumulated fatality cases at the municipal level, highlighting subregions within states where cases are notably concentrated. For instance, there is a notable concentration of cases in municipalities within the Sierra Madre Occidental, the central Mexican region, and the southern areas of Veracruz and Tabasco.
Spatial distribution of accumulated fatality incidents per (a) state and (b) municipality during 1998–2021. States with more than 50 accumulated fatality incidents are labeled in (a).
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
Additionally, states such as Oaxaca or Chiapas, which have a high number of cases at the state level, appear to be less affected at the municipality level. However, this interpretation can be misleading. These states are characterized by their lower population density and higher number of municipalities (Oaxaca with 570 and Chiapas with 125). Consequently, the low values of cases per municipality in those states suggest that the fatalities are evenly distributed across their territories.
In the subsequent sections, we comprehensively explore the primary factors contributing to this geographic preference. We will focus on understanding the underlying hazard elements and social vulnerabilities contributing to the observed concentration of fatality cases in these regions.
b. General characteristics of the affected population and vulnerability
Figure 3 provides a comprehensive overview of key demographic characteristics, shedding light on the inherent vulnerabilities of the population in Mexico when exposed to lightning. Figure 3a shows that a significant majority (81.4%) of reported lightning-related deaths are males. Notably, the highest proportion of fatalities (52.1% of total deaths) is observed among men residing in rural regions. This pattern is consistent with observations in countries with marked rural populations engaged in labor-intensive agricultural practices (e.g., Bangladesh), where males are disproportionately affected. The latter is likely due to their increased involvement in outdoor agricultural activities (e.g., Raga et al. 2014; Al-Amin Hoque et al. 2019; Kabir and Jakariya 2021), where the availability of secure spaces to avoid the lightning threat is infrequently encountered (Cooper and Holle 2019b).
Overview of death distribution in the database for the 1998–2021 period by gender and as a function of (a) type of residence (urban or rural), (b) age, (c) education level, and (d) occupation. Blue bars represent the male population, and the red bars represent the female population.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
Figure 3b shows the distribution of deaths by age range and gender. The results highlight a striking concentration of deaths within the productive age groups. Approximately 51% of fatalities occur in the range between 10 and 35 years, supporting the fact that the agricultural workforce is commonly these ages. Notably, a higher incidence of deaths is observed among males aged 5–30, while the distribution among females remains relatively uniform between the ages of 5 and 50. This observation aligns with a recognized trend wherein young males constitute a disproportionate number of lightning casualties, potentially attributed to their risk-taking behavior and engagement in outdoor employment (Cooper and Holle 2019a).
Figure 3c shows the distribution of fatalities categorized by both education level and gender. The data expose that 57.3% of recorded deaths are associated with individuals with an elementary level of school education. Among these fatalities, men constitute 46.9% of the total at the elementary school education level. This education level is highly related to the frequency of fatalities and injuries. Lower levels of education are identified as a significant factor contributing to a lack of awareness regarding lightning risks, especially in rural areas (Raga et al. 2014; Dewan et al. 2017; Islam and Schmidlin 2020; Gomes and Gomes 2021; Rahman et al. 2022, 2023).
Figure 3d shows how people’s occupation and gender influence the distribution of lightning fatalities. The figure reveals that men involved in fieldwork account for more than half (53%) of the recorded deaths. This observation is consistent with previous studies, which have reported that agricultural activities and other open-area activities are the most common scenarios for lightning-related fatalities (Al-Amin Hoque et al. 2019; Kabir and Jakariya 2021). In the figure, “other activities” refer to occupations not primarily associated with outdoor activities in rural areas, particularly those related to administrative, commercial, professional, and technical domains.
In general, Fig. 3 suggests how lightning-related deaths in Mexico are linked to outdoor work, especially for agricultural activities. Men with low education levels predominantly do these outdoor activities in rural areas. This observation implies that the population’s social vulnerability is crucial for estimating the risk of death by lightning in Mexico. Figure 4a illustrates the maximum vulnerability level per municipality based on the NMI at the locality level. The degree of social vulnerability of a municipality was determined by assigning the social vulnerability level of the most marginalized locality, thereby highlighting the highest level of social vulnerability within each municipality (see data and methods section). This figure highlights that a significant portion of Mexico’s municipalities have localities with high social vulnerability, particularly in the rural localities of the municipalities. Comparing Fig. 4a with Fig. 2b reveals that most municipalities with high lightning-related fatalities incidents are located in regions with high or very high vulnerabilities. This is illustrated in Fig. 4b, which shows the distribution of lightning-related fatalities incidents with the social vulnerability levels of the municipalities where these events were reported. Remarkably, 38.7% and 50.5% of incidents occurred in municipalities categorized as having “high” and very high social vulnerability, respectively.
(a) Maximum vulnerability level per municipality (based on vulnerability per locality) and (b) distribution of lightning-related fatality incidents based on social vulnerability levels at a municipality level.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
c. Lightning activity and fatalities
The mean annual lightning density climatology for Mexico is represented in Fig. 5a using WGLC. This climatology aligns with previous studies based on WWLLN data in Mexico (e.g., Kucieńska et al. 2010, 2012). Significantly, regions exhibiting the highest lightning density are located near the southern region of the Gulf of Mexico, along the Eastern Pacific coast, and the central region, particularly at the joining borders of Estado de México, Michoacán, and Guerrero. However, it is worth noting that Murphy and Holle (2015) pointed out a limitation of WWLLN data, indicating a lack of recorded lightning activity over the Sierra Madre Occidental in the states of Durango, Sinaloa, Sonora, and extending into the U.S. states of Arizona and New Mexico. Additionally, there is relatively little lightning activity over the Yucatan Peninsula. In contrast, lightning density is notably high along the Pacific coast near Mazatlán, offshore from Jalisco, extending to Guatemala. The discrepancy in WWLLN data stands in contrast with lightning climatologies derived from the optical transient detector (OTD) and Lightning Imaging Sensor (LIS) (Christian et al. 2003; Cecil et al. 2014), as well as the ground-based global lightning dataset GLD360 (Holle 2016).
Climatology of WGLC for 2010–21, including (a) mean annual lightning density and (b) mean annual lightning days for Mexico. The borders of Estado de México, Michoacán, and Guerrero are highlighted in white. The distribution of lightning-related fatality incidents is based on (c) lightning density and (d) lightning days.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
In contrast, Fig. 5b illustrates the spatial distribution of mean annual lightning days in Mexico using WGLC. While the pattern mirrors the locations of maximum density, the distribution is more widespread, encompassing extensive regions along the Eastern Pacific coast, the Yucatan Peninsula, and Central Mexico, each experiencing over 100 lightning days per year on average. Remarkably, the southwestern part of Estado de México stands out, experiencing over 175 lightning days yearly. This region is depicted by intense deep convection (Novo and Raga 2013), where most fatalities in Estado de México and Michoacán occur, as previously discussed by Raga et al. (2014). While the discrepancies between lightning density and lightning day distribution might be partially explained by the limitations of the lightning data grid and detection sensitivities, the overall pattern aligns well with established maps of thunderstorm days based on station data used by Mexican government agencies (e.g., CENAPRED 2012). Therefore, despite these limitations, this map provides a valuable representation of the general trends in thunderstorm day distribution across Mexico.
The spatial distribution of lightning days resembles the spatial distribution of incidents by municipality, shown in Fig. 2b, suggesting that the mean number of lightning days is a key ingredient in estimating the risk of death by lightning in Mexico. To substantiate this hypothesis, Figs. 5c and 5d present the cumulative fatality cases based on mean lightning density and lightning days per municipality, respectively. While it is acknowledged that an excessive occurrence of lightning alone is not the predominant factor influencing lightning casualties (Holle et al. 2021), a discernible pattern emerges. Specifically, more fatalities tend to occur in municipalities with more lightning days.
However, this relation does not hold for lightning density. This finding is supported by Gutiérrez-Anguamea et al. (2023), particularly for Northwestern Mexico, where fatalities do not consistently occur in regions with higher lightning density. Therefore, using lightning density as a measure of lightning risk to humans may be inadequate. Furthermore, some studies indicate that areas with high lightning density often exhibit a clustered distribution of strikes, with specific locations such as mountaintops, topographic features, and bodies of water repeatedly targeted by lightning (Pfeiffer et al. 2013, and references therein). In Mexico, regions with high lightning density are typically associated with rugged topography. Consequently, this spatial autocorrelation makes valley bottoms and other areas suitable for agriculture less prone to lightning. Additionally, individuals residing in regions with higher lightning density might be more aware of the associated hazards due to the intense lightning activity in the region.
Conversely, regions with frequent days with lightning, even with lower lightning density, present challenges for agricultural workers who must care for fields regularly during the growing season. The scarcity of natural features exacerbates this risk, leaving fewer shelters from lightning strikes. Given the potential for injury or death from a single lightning strike, the overall risk remains significant.
d. Risk of death by lightning
By examining Figs. 4 and 5d, we propose that two key components for understanding the risk of death by lightning in Mexico are intricately tied to 1) the occurrence of lightning days, representing the physical hazard, and 2) the social vulnerability of the communities exposed to such atmospheric phenomena. Employing these two variables, we have constructed a comprehensive map illustrating the risk of death by lightning per municipality in Mexico, as presented in Fig. 6.
Risk of death by lightning per municipality in Mexico.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
This figure captures the synergy between the physical hazard and social vulnerability, revealing widespread regions across Mexico characterized by high and very high risk levels. Notably, these risk-prone areas encompass extensive regions along the Eastern Pacific coast, the Yucatan Peninsula, and the central part of the country, aligning with the spatial distribution of the hazard and the social vulnerability of municipalities in those regions. The synthesis of hazard and vulnerability data accentuates the necessity for considering both the geographical distribution of lightning days and the social vulnerability of the population to evaluate and address the risk of lightning-induced fatalities precisely.
Figure 7 shows the aggregation of fatality incidents based on their corresponding risk levels, which helps evaluate the alignment between the risk map and lightning-related fatalities. As depicted in the figure, 25.6% and 56.5% of incidents were observed in municipalities categorized as having high and very high risk, respectively. Moreover, this figure reveals a consistent upward trend in incidents corresponding to increasing risk levels, displaying discernible differences among each risk category. For example, there is a 31% difference between areas labeled as high and very high risk. This pattern contrasts with the vulnerability-focused analysis presented in Fig. 4b. In the figure, incidents were similarly distributed in areas with high and very high vulnerability, with only a 13% difference between these two vulnerability levels. The clear differentiation between the aggregated fatality incidents per risk level and the consistent increase in cumulative fatalities per risk level suggests that the risk map can better approximate the actual risk levels.
Distribution of lightning-related fatality incidents based on municipality risk levels.
Citation: Weather, Climate, and Society 16, 3; 10.1175/WCAS-D-23-0151.1
As shown in Fig. 4a, Mexico is characterized by high social vulnerability, particularly in rural localities. This pronounced vulnerability, coupled with high exposure to hazard, favors very high risk in municipalities across the country, particularly in regions located in the Sierra Madre Occidental, the central Mexican region, and the southern areas of Veracruz and Tabasco. A similar pattern is evident in Fig. 2b. Additionally, states such as Oaxaca and Chiapas exhibit high-risk levels, coinciding with their high cumulative case numbers per state. Due to their sociocultural characteristics and limited access to medical services in remote rural regions, this high-risk value may imply that these states are more susceptible to underreporting lightning deaths.
4. Conclusions
Lightning significantly threatens human life, property, and diverse economic sectors. Despite its frequently underestimated impact, lightning-related fatalities, injuries, and associated economic losses demand serious consideration. This study evaluated the risk of death by lightning at the municipality level across Mexico, recognizing the intricate interplay between atmospheric variables and social conditions in estimating this risk.
The temporal analysis revealed a declining trend in lightning-related fatalities in Mexico, from 5.6 fatalities per million people in the early 1980s to about 0.3 fatalities per million in 2021. Mexico achieved the 0.5 deaths per million threshold by the late 2010s, aligning with lightning-related fatalities levels similar to the levels of more developed nations in previous decades. This decline can be attributed to demographic shifts and improved infrastructure, particularly in urban areas. Urbanization in Mexico has increased significantly, with about 70% of the population residing in rural areas in the early twentieth century compared to just 20% today. This shift from rural to urban has reduced the population’s exposure to lightning hazards, offering a plausible explanation for the observed decline in fatalities. However, certain demographic groups, like rural populations engaged in labor-intensive agricultural practices and possessing lower levels of education, continue to face significant vulnerability. Targeted interventions and risk mitigation strategies are crucial.
Spatial analyses highlighted the geographical clustering of lightning-related fatalities, emphasizing the intricate interplay between natural hazards and social vulnerability. Regions along the Sierra Madre Occidental, the Eastern Pacific coast, the Yucatan Peninsula, and central and southeastern Mexico are remarkably high-risk areas due to their elevated frequency of lightning days and high social vulnerability. This result underscores the key significance of understanding the correlation between lightning days and social vulnerability in comprehending the broader spatial distribution of risk. Here, we demonstrated that both the natural hazard (lightning days) and the social vulnerability (social marginalization) play a critical role in unraveling the complexity of the risk scenario in Mexico.
To the best of our knowledge, the risk map presented in this study is the first estimation of lightning-related fatality risk in Mexico that considers both hazard and vulnerability. This risk map successfully integrates both natural hazard and social vulnerability data, proving adequate skill in capturing the distribution of fatality incidents and depicting a consistent upward trend in incidents corresponding to increasing risk levels. Notably, 82.1% of incidents occurred in municipalities categorized as high and very high risk. In contrast, prior risk maps, such as CENAPRED (2012), focused solely on hazard considerations.
Our result underscores the importance of considering both the natural hazard and social vulnerability for a thorough assessment of lightning risk. However, limitations include the underreporting of lightning-related fatalities, particularly in rural areas, and potential biases in lightning detection data. This study focuses on fatalities due to the lack of readily available data on lightning-related injuries, which could provide a more complete risk assessment. Nevertheless, analyzing lightning-related deaths offers an initial insight into risk assessment.
Acknowledgments.
This study was financially supported by UNAM-PAPIIT under Grant IA103222. We thank Dr. Jed O. Kaplan and two anonymous reviewers for their helpful comments on the early version of this manuscript.
Data availability statement.
Lightning fatality data from the official records of the Mexican government are available through the Department for Health Information operated by the Mexican Secretariat of Health at http://www.dgis.salud.gob.mx/. EM-DAT data are available through the Centre for Research on the Epidemiology of Disasters (CRED) at the Université Catholique of Louvain at www.emdat.be. Daily gridded lightning density data are described in Kaplan and Lau (2021) and available at https://doi.org/10.5281/zenodo.4774528. The table containing the risk level for each municipality is available in the online supplemental material. The table contains the following fields: “State_Code” and “Municipality_Code,” which are numerical labels assigned by the Mexican Government to the states and municipalities. These codes facilitate comparisons with other datasets regarding Mexico. Additionally, the table includes the names of the states and municipalities, as well as the latitude and longitude of each municipality’s centroid for easy spatial localization. The table also provides numerical values for the normalized risk and associated risk level. For reference, it includes the normalized vulnerability and vulnerability level, along with the accumulated reported fatality incidents for the period 1998–2021. All other referenced data are available by contacting the corresponding author.
REFERENCES
Al-Amin Hoque, M., M. M. Billah, and B. Pradhan, 2019: Spatio-temporal and demographic distribution of lightning related casualties in northeastern part of Bangladesh. Int. J. Disaster Risk Reduct., 38, 101197, https://doi.org/10.1016/j.ijdrr.2019.101197.
Alcántara-Ayala, I., 2019: Desastres en México: Mapas y apuntes sobre una historia inconclusa (in Spanish). Invest. Geogr., 100, e60025, https://doi.org/10.14350/rig.60025.
Aven, T., 2016: Risk assessment and risk management: Review of recent advances on their foundation. Eur. J. Oper. Res., 253 (1), 1–13, https://doi.org/10.1016/j.ejor.2015.12.023.
Biswas, A., K. Dalal, J. Hossain, K. Ul Baset, F. Rahman, and S. Rahman Mashreky, 2016: Lightning injury is a disaster in Bangladesh?—Exploring its magnitude and public health needs. F1000 Res., 5, 2931, https://doi.org/10.12688/f1000research.9537.1.
Biswas, R. N., M. N. Islam, M. J. Mia, and M. N. Islam, 2020: Modeling on the spatial vulnerability of lightning disaster in Bangladesh using GIS and IDW techniques. Spat. Inf. Res., 28, 507–521, https://doi.org/10.1007/s41324-019-00311-y.
Cecil, D. J., D. E. Buechler, and R. J. Blakeslee, 2014: Gridded lightning climatology from TRMM-LIS and OTD: Dataset description. Atmos. Res., 135–136, 404–414, https://doi.org/10.1016/j.atmosres.2012.06.028.
CENAPRED, 2012: Mapas de índices de riesgo a escala municipal por fenómenos hidrometeorológicos. Accessed 17 April 2024, http://www.atlasnacionalderiesgos.gob.mx/descargas/Metodologias/Hidrometeorologico.pdf.
Christian, H. J., and Coauthors, 2003: Global frequency and distribution of lightning as observed from space by the Optical Transient Detector. J. Geophys. Res., 108, 4005, https://doi.org/10.1029/2002JD002347.
CONAPO, 2020: Índice de marginación por localidad. Accessed 5 December 2023, https://www.gob.mx/conapo/documentos/indices-de-marginacion-2020-284372.
Cooper, M. A., and R. L. Holle, 2019a: Gender, age, and casualties per incident. Reducing Lightning Injuries Worldwide, M. A. Cooper and R. L. Holle, Eds., Springer Natural Hazards, Springer International Publishing, 91–98.
Cooper, M. A., and R. L. Holle, 2019b: Current global estimates of lightning fatalities and injuries. Reducing Lightning Injuries Worldwide, M. A. Cooper and R. L. Holle, Eds., Springer Natural Hazards, Springer International Publishing, 65–73.
Cummins, K. L., and M. J. Murphy, 2009: An overview of lightning locating systems: History, techniques, and data uses, with an in-depth look at the U.S. NLDN. IEEE Trans. Electromagn. Compat., 51, 499–518, https://doi.org/10.1109/TEMC.2009.2023450.
Dalenius, T., and J. L. Hodges Jr., 1959: Minimum variance stratification. J. Amer. Stat. Assoc., 54, 88–101, https://doi.org/10.1080/01621459.1959.10501501.
Dewan, A., M. F. Hossain, M. M. Rahman, Y. Yamane, and R. L. Holle, 2017: Recent lightning-related fatalities and injuries in Bangladesh. Wea. Climate Soc., 9, 575–589, https://doi.org/10.1175/WCAS-D-16-0128.1.
DGIS, 2023: Dirección General de Información en Salud: Defunciones. Accessed 5 December 2023, http://www.dgis.salud.gob.mx/.
Dlamini, W. M., 2009: Lightning fatalities in Swaziland: 2000–2007. Nat. Hazards, 50, 179–191, https://doi.org/10.1007/s11069-008-9331-6.
Dominguez, C., A. Jaramillo, and P. Cuéllar, 2021: Are the socioeconomic impacts associated with tropical cyclones in Mexico exacerbated by local vulnerability and ENSO conditions? Int. J. Climatol., 41, E3307–E3324, https://doi.org/10.1002/joc.6927.
EM-DAT, 2023: EM-DAT, CRED/UCLouvain. Accessed 5 December 2023, https://www.emdat.be.
Farfán, L. M., L. Brito‐Castillo, M. Z. F. López, and J. Cortés‐Ramos, 2023: Deep convection, lightning activity and population impact over northwest Mexico: Coastal hotspots from 2009 to 2018. Int. J. Climatol., 43, 5620–5639, https://doi.org/10.1002/joc.8164.
Gomes, C., and A. Gomes, 2021: Lightning: Public concepts and safety education. Lightning, C. Gomes, Ed., Lecture Notes in Electrical Engineering, Vol. 780, Springer, 275–300.
Gutiérrez-Anguamea, G. A., C. M. Minjarez-Sosa, X. G. Torres-Carrillo, and G. E. Vázquez-Becerra, 2023: Lightning fatalities and its correlation with social variables in northwestern Mexico. Geomatics Nat. Hazards Risk, 14, 2184673, https://doi.org/10.1080/19475705.2023.2184673.
Holle, R. L., 2016: A summary of recent national-scale lightning fatality studies. Wea. Climate Soc., 8, 35–42, https://doi.org/10.1175/WCAS-D-15-0032.1.
Holle, R. L., and M. A. Cooper, 2016: Lightning occurrence and social vulnerability. Atmospheric Hazards—Case Studies in Modeling, Communication, and Societal Impacts, J. S. M. Coleman, Ed., InTech, 3–19.
Holle, R. L., and D. Zhang, 2023: Human impacts, damages, and benefits from lightning in Arizona. Flashes of Brilliance, R. L. Holle and D. Zhang, Eds., Springer International Publishing, 97–152.
Holle, R. L., R. E. López, and B. C. Navarro, 2005: Deaths, injuries, and damages from lightning in the United States in the 1890s in comparison with the 1990s. J. Appl. Meteor., 44, 1563–1573, https://doi.org/10.1175/JAM2287.1.
Holle, R. L., A. Dewan, R. Said, W. A. Brooks, M. F. Hossain, and M. Rafiuddin, 2019: Fatalities related to lightning occurrence and agriculture in Bangladesh. Int. J. Disaster Risk Reduct., 41, 101264, https://doi.org/10.1016/j.ijdrr.2019.101264.
Holle, R. L., M. A. Cooper, and N. Navarrete-Aldana, 2021: Lightning injury: Occurrence and medical treatment. Lightning, C. Gomes, Ed., Lecture Notes in Electrical Engineering, Vol. 780, Springer, 263–273.
INEGI, 2020: Censo de Población y Vivienda 2020. Accessed 5 December 2023, https://www.inegi.org.mx/.
Islam, M. S., and T. W. Schmidlin, 2020: Lightning hazard safety measures and awareness in Bangladesh. Nat. Hazards, 101, 103–124, https://doi.org/10.1007/s11069-020-03864-6.
Kabir, F., and M. Jakariya, 2021: Climate change and lightning risk in Bangladesh. Climate Change in Bangladesh, M. Jakariya and M. N. Islam, Eds., Springer Climate, Springer International Publishing, 183–203.
Kaplan, J. O., and K. H.-K. Lau, 2021: The WGLC global gridded lightning climatology and time series. Earth Syst. Sci. Data, 13, 3219–3237, https://doi.org/10.5194/essd-13-3219-2021.
Kucieńska, B., G. B. Raga, and O. Rodríguez, 2010: Cloud-to-ground lightning over Mexico and adjacent oceanic regions: A preliminary climatology using the WWLLN dataset. Ann. Geophys., 28, 2047–2057, https://doi.org/10.5194/angeo-28-2047-2010.
Kucieńska, B., G. B. Raga, and R. Romero-Centeno, 2012: High lightning activity in maritime clouds near Mexico. Atmos. Chem. Phys., 12, 8055–8072, https://doi.org/10.5194/acp-12-8055-2012.
López, R. E., and R. L. Holle, 1998: Changes in the number of lightning deaths in the United States during the twentieth century. J. Climate, 11, 2070–2077, https://doi.org/10.1175/1520-0442(1998)011%3C2070:CITNOL%3E2.0.CO;2.
López, R. E., R. L. Holle, T. A. Heitkamp, M. Boyson, M. Cherington, and K. Langford, 1993: The underreporting of lightning injuries and deaths in Colorado. Bull. Amer. Meteor. Soc., 74, 2171–2178, https://doi.org/10.1175/1520-0477(1993)074<2171:TUOLIA>2.0.CO;2.
Mills, B., D. Unrau, L. Pentelow, and K. Spring, 2010: Assessment of lightning-related damage and disruption in Canada. Nat. Hazards, 52, 481–499, https://doi.org/10.1007/s11069-009-9391-2.
Murphy, M. J., and R. L. Holle, 2015: Comment on “Deaths by lightning in Mexico (1979–2011): Threat or vulnerability?”. Wea. Climate Soc., 7, 273–274, https://doi.org/10.1175/WCAS-D-14-00046.1.
Nazri, K., S. C. Lim, and C. Gomes, 2021: Lightning safety awareness level in Malaysia. F1000 Res., 10, 921, https://doi.org/10.12688/f1000research.73064.2.
Novo, S., and G. B. Raga, 2013: The properties of convective storms in central Mexico: A radar and lightning approach. Atmósfera, 26, 461–472, https://doi.org/10.1016/S0187-6236(13)71088-9.
Oliver-Smith, A., I. Alcántara-Ayala, I. Burton, and A. Lavell, 2017: The social construction of disaster risk: Seeking root causes. Int. J. Disaster Risk Reduct., 22, 469–474, https://doi.org/10.1016/j.ijdrr.2016.10.006.
Pfeiffer, M., A. Spessa, and J. O. Kaplan, 2013: A model for global biomass burning in preindustrial time: LPJ-LMfire (v1.0). Geosci. Model Dev., 6, 643–685, https://doi.org/10.5194/gmd-6-643-2013.
Raga, G. B., M. G. de la Parra, and B. Kucienska, 2014: Deaths by lightning in Mexico (1979–2011): Threat or vulnerability? Wea. Climate Soc., 6, 434–444, https://doi.org/10.1175/WCAS-D-13-00049.1.
Rahman, M. M., and Coauthors, 2022: Knowledge, attitude, and practices towards lightning in Bangladesh. Sustainability, 14, 448, https://doi.org/10.3390/su14010448.
Rahman, M. M., I. A. Shobuj, M. M. H. Santo, and M. T. Hossain, 2023: Public perception toward lightning in a lightning-prone area of Bangladesh. Int. J. Disaster Risk Reduct., 89, 103638, https://doi.org/10.1016/j.ijdrr.2023.103638.
Rakov, V. A., 2012: Lightning discharge and fundamentals of lightning protection. J. Light. Res., 4, 3–11, https://doi.org/10.2174/1652803401204010003.
Roeder, W. P., B. H. Cummins, K. L. Cummins, R. L. Holle, and W. S. Ashley, 2015: Lightning fatality risk map of the contiguous United States. Nat. Hazards, 79, 1681–1692, https://doi.org/10.1007/s11069-015-1920-6.
Sabur, A. K. M. A., 2012: Disaster management system in Bangladesh: An overview. India Quart., 68, 29–47, https://doi.org/10.1177/097492841106800103.
Sutton, R. T., 2019: Climate science needs to take risk assessment much more seriously. Bull. Amer. Meteor. Soc., 100, 1637–1642, https://doi.org/10.1175/BAMS-D-18-0280.1.
World Meteorological Organization, 2021: WMO atlas of mortality and economic losses from weather, climate and water extremes (1970–2019). Tech. Rep. WMO 1267, 90 pp., https://uncclearn.org/wp-content/uploads/library/1267_Atlas_of_Mortality_en.pdf.
Yair, Y., 2018: Lightning hazards to human societies in a changing climate. Environ. Res. Lett., 13, 123002, https://doi.org/10.1088/1748-9326/aaea86.
