Introduction

Tropical cyclones, also known as typhoons and hurricanes, have caused an estimated 1.33 million deaths since the beginning of the 20th century and affected more than 629 million people in this timeframe. A tropical cyclone is a non-frontal storm system that is characterized by a low pressure center, spiral rain bands and strong winds. Usually it originates over tropical or sub-tropical waters and rotates clockwise in the southern hemisphere and counter-clockwise in the northern hemisphere. Depending on their location and strength, tropical cyclones are referred to as hurricanes (western Atlantic/eastern Pacific), typhoons (western Pacific), and cyclones (southern Pacific/Indian Ocean) ¹. Approximately half of tropical cyclones recorded and more than 90% of cyclone-related deaths originate in Asian waters ¹. Cyclones are large organized storms with well-defined cores that begin over tropical or subtropical waters, often as a result of monsoon troughs and easterly waves ². An average of 37 tropical storms occur each season and they range in size from 100 to more than 1,000km in diameter and are known for strong winds and bands of torrential rain that revolve around the center or eye of the storm ³. In the Eastern Pacific basin the season begins two weeks early on May 15th. The levels of intensity of these storms range from tropical depression (winds <17 meters per second), to tropical storm (winds 18-32 m/s), and cyclone (>33 m/s) ⁴. Once formed, cyclones maintain strength by pulling heat and moisture from warm ocean waters ³. The damage and deaths related to cyclones are the result of three major forces: winds in excess of 155 miles per hour; storm surge where the level of the sea rises as much as 10 meters and move ashore; and secondarily due to floods resulting from torrential rains. Storm surges and floods are the primary causes of death in cyclones ⁵.

The impacts from cyclones are concentrated in coastal areas of South and East Asia, Madagascar, the east coast of North and Central America and the Caribbean. Mortality is concentrated in Asia, economic losses follow a similar pattern; however, total economic losses are greatest in affluent countries with developed infrastructure ⁶. Future vulnerability to cyclones will increase due to factors including population growth, urbanization, increasing coastal settlement and changing weather patterns. The objectives of this review are to describe the impact of cyclones in terms of mortality, injury, and displacement and, to the extent possible, identify risk factors for associated with these outcomes. This is one of five reviews on the human impact of natural disasters, the others being volcanoes, floods, tsunamis, and earthquakes.

Methods

Data on the impact of cyclones were compiled using two methods, a historical review of cyclone events and a systematic literature review of publications relating to the human impacts of cyclones.

Historical Event Review

A historical database of significant cyclones from 1980 to mid-2009 was created from publicly available data. Multiple data sources were sought to ensure a complete listing of events and to allow for cross checking. The two primary data sources were the Centre for Research on the Epidemiology of Disasters (CRED) International Disaster Database (EM-DAT) ⁷ and the National Hurricane Center (NHC)⁸ because they included information on human impacts. The events in the EM-DAT database include one or more of the following criteria: 10 or more people killed or injured; 100 people affected; declaration of a state of emergency; or a call for international assistance. The NHC database included information on all cyclones in the Atlantic, Caribbean and Gulf of Mexico.

The EM-DAT event list was downloaded in August 2009 and NHC data downloaded in February 2010. Event lists were reconciled to create a combined list of events from both data sources which were then tabulated and summarized for 1980 through 2009. See https://www.jhsph.edu/refugee/natural_disasters/_Historical_Event_Review_Overview.html for the database of tropical cyclone events. A total of 948 events were retained from EM-DAT and 331 from the NHC. For cyclone impacts reported by EM-DAT, zeroes were treated as missing values because they were used as placeholders and their inclusion in the analysis could contribute to the under estimation of tsunami impacts.

To assess risk factors for cyclone-related mortality the following categories were used: none (0 deaths), low (1-9 deaths), medium (10-99 deaths) and high (≥100 deaths). Bivariate tests for association were performed using chi-square for categorical and ANOVA for continuous measures. All covariates, with the exception of World Bank developmental level which was highly correlated with per capita GDP were subsequently included in a multinomial logistic regression model to assess relative risk of mortality at a given level as compared to events with no deaths. Analyses were performed using Stata Statistical Software, Version 11.0 ⁹.

Systematic Literature Review

Key word searches in MEDLINE (Ovid Technologies, humans), EMBASE (Elsevier, B.V., humans), SCOPUS (Elsevier B.V., humans), and Web of Knowledge/Web of Science (Thomson Reuters) were performed to identify articles published in July 2007 or earlier that described natural hazards and their impact on human populations. Following the systematic review, a further search was conducted to identify relevant articles published through October 2012. One search was done for all the five natural hazards described in this set of papers. This paper describes the results for cyclones. The systematic review is reported according to the PRISMA guidelines. Key words used included natural hazard(s), natural disaster(s), volcano(es), volcanic, volcanic eruption, seismic event, earthquake(s), cyclone(s), typhoon(s), hurricane(s), tropical storm(s), flood(s), flooding, mudslide(s), tsunami(s), and tidal wave(s). Key words included for impact on humans were affected, damage(d), injury, injuries, injured, displaced, displacement, refugees, homeless, wounded, wound(s), death(s), mortality, casualty, casualties, killed, died, fatality, fatalities and had to be used in either the title, abstract or as a subject heading/key word. The search resulted in 2,747 articles from MEDLINE, 3,763 articles from EMBASE, 5,219 articles from SCOPUS, and 2,285 articles from ISI Web of Knowledge. Results from the four databases were combined and duplicates were excluded to yield a total of 9,958 articles.

Title screening was performed to identify articles that were unrelated to natural disasters or human populations. Each title was screened by two independent reviewers and was retained if either or both reviewers established that inclusion criteria were met. To ensure consistent interpretation of inclusion criteria, percent agreement was assessed across reviewers for a small sample of articles, and title screening began after 80% agreement on inclusion was achieved. A total of 4,873 articles were retained for abstract review. Articles were excluded if they met one or more of the following criteria: language other than English; editorial or opinion letter without research; not related to human populations; individual case report/study; focus on responders; and not related to human or environmental vulnerabilities or impacts of hazards. Each abstract was then screened by two reviewers and retained if either or both established that inclusion criteria were met. Included abstracts were coded for event type, timeframe, region, subject of focus, and vulnerable population focus. A total of 558 cyclone articles were retained for article review; 193 articles focusing on the impacts of cyclones on human populations in terms of mortality, injury, and displacement were prioritized for abstraction. Upon full review, 49 articles were retained including 48 that underwent dual review, standard data abstraction and one that was identified as a review article (Figure 1). The additional review then identified eleven articles through October 2012 that met the inclusion criteria for abstraction in the mortality and injury review. A summary of the final 58 abstracted articles is presented in Table 1.

Results

Historical Event Review

During the 30-year observation period (1980-2009), 1,080 cyclones were recorded with an average of 32 (range 16-66) annually. The number of events reported annually by NHC and EM-DAT increased over time as did the total number of events; NHC reported a lower number of events because their focus area is the Americas (Figure 2). Both the frequency of cyclones and affected population size increased over time; cyclone related mortality did not follow a similar trend and mortality peaks were associated with infrequent high-impact events such as cyclone Gorky (Bangladesh, 1991) and cyclone Nargis (Myanmar, 2008) (Figure 3).

By decade, 42,5% (n=459) of events occurred in the 2000s, 29.6% (n=320) in the 1990s, and 27.9% (n=301) in the 1980s. The impact of cyclone events across regions is summarized in Figure 4. The World Health Organization regions of the Western Pacific (WPRO) and the Americas (AMRO) accounted for more than 80% of all reported events. The mortality was greatest in the SEARO regions while only 8% of deaths occurred in the AMRO region despite accounting for 37% of all events occurring there. Although the SEARO region accounted for only 9% of all events, it had 53% of the affected population and 80% of all deaths.

The overall impact of cyclones on human populations is summarized in Table 2. An estimated 466.1 million people were reported to be affected by cyclones between 1980 and 2009, including 20.1 million that were rendered homeless. These figures likely substantially underestimate the true impact because estimates of the total affected population and the homeless population were reported in 78.7% (n=799) and 26.2% (n=283) of events, respectively. The distribution of the affected population was highly skewed: when reported on average 592,830 people affected per storm, but the calculated median was only 20,000 per event. Monetary damages were reported by EM-DAT in 15.4% of events and evacuation was reported by NHC in only 2.3% events so these outcomes were not assessed because of insufficient reporting.

Mortality and Injury. Mortality data had the most complete reporting. When the two sources were combined, information on deaths was reported in 97.1% of events and deaths occurred in 89.4% of cyclones between 1980 and 2009. Overall, 412,644 deaths were reported in the historical event review, the vast majority from EM-DAT. For cyclones where mortality was reported, there was a median of 14 deaths per event (mean = 430, 5% trimmed mean= 41, range 1-138,849) when using the highest reported death toll. The two deadliest storms, Cyclone Gorky (Bangladesh, 1991; 138,866 deaths) and Cyclone Nargis (Myanmar, 2008; 138,366 deaths) accounted for two-thirds of cyclone deaths between 1980 and 2009. The next order of high mortality events with 10,000-15,000 deaths included Hurricane Mitch (Honduras, 1998) and a cyclone in Bangladesh (1985) and there were 16 events with 1,000-9,999 deaths over the 30 years. In total, the 1.9% (n=20) of events with mortality >1000 accounted for 83.9% of all cyclone deaths, suggesting that cyclone mortality is concentrated in infrequent and extreme events.

Injury data were available in 397 (36.8%) events, with a total of 290,654 cyclone-related injuries documented. When reported, there was a median of 46 injuries per storm (mean=775, 5% trimmed mean=200, range 1-138,849) when the highest reported figure was used. To estimate the total number of injuries, it was presumed that injuries occurred in events with reported deaths. There were 965 cyclones with reported fatalities. When the median and 5% trimmed mean for injuries were applied to the remaining 568 events with fatalities but no injury reporting, it was estimated that between 28,400 and 113,600 unreported cyclone related injuries may have occurred between 1980 and 2009.

Ordinal logistic regression was used to assess country-level characteristics associated with storm mortality categories (Table 3). All country-level variables were found to be significantly associated with mortality. The mean GDP per capita was $13,191 (SD 17,709, range 1433-99,383) and the mean Gini index score ¹⁰, which is a measure of equality, was 41.3 (SD 7.3, range 24.9-64.3). When assessed by mortality category, the highest per capita GDP occurred in the no deaths category while the lowest was observed in the ≥100 death category (p<.001), suggesting that the cyclone mortality risk is concentrated in lower income countries.

Relative risk ratios for cyclone mortality from the final multinomial logistic regression model, using events with no deaths as the reference category, are presented in Table 4. Per capita GDP, WHO region and event decade were significantly associated with excess mortality, in particular for the mid- and high level mortality categories (10-99 deaths and ≥100 deaths, respectively). The proportion of events with mid- to high mortality levels decreased in the 1980s and 1990s, but during the 2000s, the relative risk ratios of mid- and high level mortality events were statistically similar to the 1970s. The Western Pacific region, where the highest proportion of mid- and high level mortality events occur, was used as the reference category for regional comparisons. Relative risk ratios for mid-level mortality events were significantly lower in the European/Eastern Mediterranean, Americas, and Southeast Asia regions as compared to the Western Pacific. Relative risk ratios for high level mortality events were statistically similar to the Western Pacific for all regions except the Americas with significantly lower risk. No significant relationship between GINI index and mortality risk was observed whereas GDP was inversely associated with risk of high mortality events.

Systematic Literature Review

Mortality. Among articles meeting inclusion criteria for full review (n=56), 34 reported mortality data including 16 that provided information on direct or indirect causes of death (Table 5) and 10 that reported sex-specific mortality counts or risk (Table 6) ^11–66,80. Most articles provided some information about the distribution of deaths across population subgroups or an individual’s location at the time of the event; with one exception, all articles reported on hurricane impact in the United States. When aggregated, 54% of US hurricane deaths were classified as direct deaths and 43% as indirect deaths. Among direct deaths, drowning was the most common cause of death, accounting for 59% of direct deaths followed by trauma, which accounted for 39% of direct deaths. Among indirect deaths cause of death was less frequently reported however trauma was the most common cause of indirect death followed by motor vehicle accidents, carbon monoxide poisoning, fires or burns, and electrocution. When examined by sex, an increased mortality risk among men was observed in the eight of the ten studies that reported deaths by sex; males accounted for 59.4% of US hurricane deaths reported (Table 6). Studies that reported sex-specific mortality in the 1991 Bangladesh cyclone observed a higher mortality rates among females (71/1000) compared to males (15/1000) ^20,21,22. Age was associated with increased mortality risk in numerous studies where both children ^20,21,22, and older adults ^22,57,58,60, experienced disproportionate mortality. Other risk factors for mortality included residence type ^22,35, not reaching shelter ^14,21, geographic location ^20,36, race ^57,58, flood level ⁵⁷, and deforestation ^28,61.

Injury. Injury data were reported in 28 of the 58 articles, 15 of which provided information on injury type (Table 7). Most articles reported that the majority of injuries were minor but it was not possible to aggregate injury data due to the different study designs, reporting methods, and data sources. Lacerations, wounds, contusions, blunt trauma, animal/insect bites, and motor vehicle injuries were among the most frequent types of injuries reported. The three population based surveys estimated injury rates between 3.8 and 4.5% ^37,62,66. When assessed by age, injury rates were highest among middle age adults in numerous studies ^{13,25,32,40,53}. Males were at higher risk for injury ^13,24,25,54; in all studies but one in the studies that reported gender ²². Location was a risk factor for injury in many reports, including administrative unit or location relative to storm path ⁴², within a city ^24,25,44, and being outdoors ²⁵. Race was associated with increased injury risk in several studies, however, the race most at risk for injury varied ^24,25,26,56. Only three articles presented injury data on storms outside the United States ^11,62,65, reflecting the paucity of information from less developed countries and a need for additional research in cyclone injury epidemiology in these regions.

Discussion

Main Findings

In the 30 year period between 1977 and 2009, approximately 466 million people were affected by cyclones; 20.1 million left homeless, 412,000 people died and 290,000 were injured, excluding an estimated 28,000 to 114,000 unrecorded injuries. The mortality estimate presented in this study is consistent with recent estimates in other studies ⁶⁷, but the numbers injured and displaced are likely gross underestimates given the low frequency with which these figures are reported. Findings from the historical event review are also consistent with previous observations that cyclone mortality varies by region, economic development level, and event severity.Cyclone impacts were concentrated in the Asia-Pacific region and the majority of fatalities occurred in developing nations. High mortality events, with death tolls in excess of 1000, occur in less than 2% of events and more than two-thirds of all cyclone deaths between 1980 and 2009 occurred in two events (Cyclone Gorky, Bangladesh 1991 and Cyclone Nargis, Myanmar 2008). The number of cyclones and deaths increased each decade but the average number of deaths per storm decreased. Human vulnerability to cyclones will increase in future years due to population growth, urbanization, increased coastal settlement, poverty, and changing weather patterns which is associated with an increase in the number of high intensity cyclones.

A significant disparity between cyclone mortality in developing and developed nations persists, though apart from simple casualty counts there is little information available on the epidemiology of cyclone morbidity and mortality in less developed countries. This indicates a need for additional research outside of the US. The United Nations Development Program (UNDP) identified 29 developing nations and four developed nations that are at risk for cyclones ⁶⁷ but 42% and 27% of cyclone deaths in the past two centuries have occurred in Bangladesh and India, respectively ⁶⁸. Additionally, the majority of high-fatality storms occurred in the latter half of the 20th century though no developed nation sustained more than 1000 deaths from a cyclone in this time period^67,69. The leading explanations for regional differences in mortality is the size of the at risk population and the capacity for pre-event evacuation. Higher population densities in the Western Pacific and South East Asia where dense settlements in low lying areas are associated with increased risk of death in from storm surge ^70,71. A higher economic development among the countries in the Americas is associated with lower regional mortality rates because of increased capacity to evacuate. Prior to the implementation of early warning, evacuation, and shelter systems an estimated 90% of cyclone mortality was attributed to storm surge drowning ^{72,73,74,75,76}. Improvements in forecasting, and early warning systems and in evacuation and shelter procedures, particularly in developed countries, have reduced storm-surge related mortality and increased proportional morbidity and mortality in the post-impact period ^16,19,69.

Comparison with Previous Reviews

Findings of this review were contrary to the conclusions of other recent reviews which concluded that most storm-related mortality in developed countries occurs in the post-impact period ⁶⁹. In the systematic literature review, 79% of the 946 included deaths where storm phase was reported occurred during the impact period. Direct and indirect deaths, respectively, accounted for 56% and 44% of deaths (n=1450 deaths where cause was reported) and the primary causes of death were drowning (33%), direct injuries or trauma (21%), and indirect injuries or trauma (7%). Studies that included gender breakdown for cyclone-related deaths, most of which are accounts of cyclone events in the United States, consistently reported greater proportion of male as compared to female deaths; when aggregated, males accounted for 59% of reported US hurricane deaths. In less developed countries females face a greater mortality risk ^21,22. An increased risk of death in younger ^20,21,22, and older populations ^22,57,58,60, was also observed which is consistent with broader natural disaster mortality trends. However, it is important to note that primary research on cause and timeframe of death as well as demographic and other factors associated with increased mortality risk is limited almost completely to the United States. Future studies on the human impacts of cyclones should be focused in Southeast Asia and the Western Pacific, the regions where the majority of cyclone impacts occur.

While minimal data on cyclone-related injuries and mortality is available from less developed settings, it can be presumed that developing nations also bear the burden of cyclone-attributable injury where the frequency and severity of injuries are inversely related to degree of physical protection ⁶⁹. Minor trauma is common among the injured, including lacerations, abrasions and contusions, puncture wounds, and sprains and fractures ^{11,13,24,25,29,32,40,41,51,53,62,63}. The majority of those injured in cyclones can be treated on an outpatient basis and do not require sophisticated surgical or inpatient care ^5,77. Reported causes of death and injury in more developed countries during the post-impact phase include blunt trauma, vehicle accidents, carbon monoxide poisoning, burns and fires, electrocution, and chain-saw injuries. As compared to needs for food, water, shelter, and sanitation, injuries are not usually a major public health problem in the post-impact phase ⁷⁸ This suggests and that mobile field hospitals and specialized surgical teams may be ineffective responses and that non-medical relief may be a more appropriate strategy for morbidity and mortality reduction in the aftermath of cyclones ⁵.

Limitations

The availability and quality of data has likely increased and improved over time, however, in many events deaths, injuries, and affected population size are unknown or unrecorded. For most events no data were reported for injured, displaced, and affected populations, contributing to underestimation of impacts. Inconsistencies and errors were common in data from different sources that called into question the reliability of available data. In some cases inclusion criteria and definitions were not ideal which created difficulties in reconciling event lists. Challenges were encountered when modeling cyclone mortality including a non-normal distribution, which necessitated analysis with a categorical outcome. Information on 2007-2009 GDP and 2009 GINI index were used for analysis regardless of the event year, but these values may have been different for events in the 1980s. Additionally, some countries did not exist or have merged with other nations since the 1980s and many of the smaller island countries in the Caribbean are territories of European countries which necessitated the use of GDP, GINI, and development levels which may not be representative of realities in the cyclone affected area and/or time period. When combined with uncertainty in the historical record and the relative paucity of primary research focusing on cyclone impacts in heavily affected Asian region, conclusions that can be drawn about cyclones impacts on human populations are limited. Other principal limitations of the literature review are 1) that an in-depth quality analysis of all reviewed articles was not undertaken, and 2) the fact that only English language publications were included which likely contributed to incomplete coverage of studies published in other languages originating from low and middle income countries.

Conclusions

Analysis of the impact of cyclones on human populations is challenging given the paucity of data from the most affected regions, the occasional occurrence of extreme high mortality events, and the reporting inconsistencies including both lack of standardized definitions and temporal changes in collection procedures, completeness and accuracy of data. However, even with this under-representation the impact of cyclones is huge, with 466 million people affected, 412,644 deaths and 290,654 injuries were reported as a result of cyclones between 1980 and 2009. The primary cause of cyclone-related mortality in both developed and less developed countries was storm surge drowning. In more developed countries an increased proportion of deaths and injuries were observed in the aftermath of cyclones as a result of improved early warning systems and evacuation. Male gender was associated with increased mortality risk in developed countries, whereas female gender was linked to higher mortality risk in less developed countries. Both older and younger population sub-groups also face an increased mortality risk.

Cyclones have significantly impacted populations in Southeast Asia, the Western Pacific, and the Americas regions over the past quarter of a century with less developed nations in Asia bearing the majority of the mortality and injury burden. Additional preparedness and mitigation strategies, particularly in less developed countries where the majority of cyclones occur, can lessen the impact of future events. In particular, improvements in forecasting, early warning systems, evacuation and shelter procedures, and public education on safety precautions and post-impact hazards could reduce cyclone-related morbidity and mortality in future decades.

Competing Interests

The authors have no competing interests to declare.

Correspondence

Shannon Doocy, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St, Suite E8132, Baltimore, MD 21230. Tel. 410-502-2628. Fax: 410-614-1419. Email: [email protected].