Are natural disasters and industrial accidents that involve hazardous substances becoming more frequent and severe? This is one of the questions researchers at Resources for the Future attempted to answer by developing a database on natural disasters and industrial accidents that occurred worldwide during the years 1945 to 1986. Trends evidenced by the database have implications for policies and resource allocations to prevent, prepare, and respond to these events.
During the years 1945 to 1986, more than 1,200 natural disasters, each with 25 or more associated fatalities, occurred worldwide. These disasters resulted in more than 2.3 million deaths. During the same period, industrial accidents, each with 5 or more fatalities, that involved the release of hazardous materials were much less frequent and claimed far fewer lives. Nearly 300 of these accidents occurred worldwide, resulting in the deaths of almost 15,000 people.
These statistics were derived from a database recently developed by researchers at Resources for the Future (RFF). The database makes it possible to discern whether the natural disasters and industrial accidents noted above became more frequent, or severe, or both over the years 1945 to 1986; to discover which regions of the world suffered most as a result of their occurrence; and to determine which types of these events were most common and which were most devastating. All this information has implications for the development of policies to prevent, prepare for, or respond to natural disasters and industrial accidents and for the targeting of resources to efforts that will garner the greatest life-saving benefits.
The RFF database improves on previous surveys and compilations of data on natural disasters and industrial accidents by providing a compendium of information drawn from two dozen publicly available sources of statistics on these events. It contains a record of every natural disaster and every industrial accident in which a hazardous material caught fire, exploded, or was released in a toxic cloud that (1) occurred between 1945 and 1986, (2) was mentioned by at least one of the above sources, (3) resulted in enough fatalities to be defined as "major," and (4) was characterized by sudden onset and immediate fatalities. The last criterion excluded certain natural disasters—namely, droughts, epidemics, and famines—and certain industrial catastrophes of a chronic nature—such as the long-term discharge of hazardous substances into the environment.
The setting of a lower fatality threshold for industrial accidents—that is, a minimum of only 5 fatalities as opposed to 25 fatalities—reflects two facts. First, individual industrial accidents generally claim fewer lives than individual natural disasters, although death tolls alone do not indicate the severity of the social and economic repercussions of these accidents. Second, risks due to technological hazards are perceived by the public to be higher than risks due to natural hazards.
To show the geographical distribution of the frequency and severity (as measured in fatalities) of major natural disasters and industrial accidents, the Resources for the Future database divided the world into regions. With respect to natural disasters, it was divided into seven regions: North America, Latin America and the Caribbean, Europe and the former Soviet Union, sub-Saharan Africa, the Middle East and North Africa, Southern Asia, and East Asia and the Pacific. With respect to industrial accidents, it was divided into three regions: the United States and Canada, Europe and the former Soviet Union, and all other countries.
In the RFF database, natural disasters were divided into three types and industrial accidents into four types. Natural disasters were broadly classified as meteorological, geological, and other. Meteorological disasters were further divided into floods, tropical cyclones (including hurricanes, cyclones, and typhoons), other storms (including tornadoes), and heat and cold waves. When the majority of deaths that data sources attributed to a tropical cyclone or other storm was caused by a flood, the event was classified as a flood. Geological disasters were divided into earthquakes, volcanic eruptions, and tsunamis (tidal waves produced by sub-marine earth movements or volcanic eruptions). Other disasters included landslides and naturally occurring fires. The RFF database classified industrial accidents as being transportation-related, occurring at a fixed facility, involving a pipeline, or unknown. It did not include accidents that occurred during mining and demolition activities because these activities involve hazards of a highly specific nature.
Number of major natural disasters worldwide
Major natural disasters, 1945–1986
The database reveals that, during the years 1945 to 1986, the annual number of natural disasters that resulted in 25 or more deaths generally increased until the 1980s. There is a similar though less pronounced upward trend in the annual number of disasters that resulted in 100 or more deaths. The annual number of disasters that resulted in 1,000 or more deaths remained relatively constant (see top figure, p. 10).
One possible explanation for the upward trend in those disasters that resulted in at least 25 fatalities is increasingly better reporting by news sources and government agencies of events of this kind happening in more remote regions of the world. However, the most likely explanation is the substantial rise in world population and the increasing vulnerability of economically disadvantaged people. Population growth, increasing urbanization, land shortages, and economic hardships often force people in less developed countries to migrate to locations having geological and meteorological hazards, while endemic poverty limits emergency preparedness and response and the ability of individuals to cope with natural disasters.
During the years 1945 to 1986, East Asia and the Pacific suffered the greatest number of major natural disasters and the greatest number of associated deaths.
On average, 30 natural disasters killed 56,000 people per year and 1,850 people per disaster during the years 1945 to 1986. These figures may be misleading, however. The average number of fatalities per year and per disaster was vastly inflated by a few disasters resulting in unusually high death tolls. The three deadliest disasters—the 1948 earthquake in the former Soviet Union (110,000 deaths), the 1970 cyclone in Bangladesh (500,000 deaths) and the 1976 earthquake in China (700,000 deaths)—accounted for more fatalities than all the other natural disasters combined (see bottom figure, p. 10). If the deaths due to these three events are excluded, the average number of fatalities per year and per disaster would drop to about 25,000 and 800, respectively.
The frequency and severity of major natural disasters were unevenly distributed across the globe. During the years 1945 to 1986, three regions bore the brunt of these disasters: East Asia and the Pacific, Latin America and the Caribbean, and Southern Asia. East Asia and the Pacific suffered both the largest number of natural disasters (401) and the most deaths (977,000). The number of deaths, however, is inflated by the earthquake that occurred in China in 1976. When the fatalities from this and the USSR and Bangladesh disasters noted above are excluded from the total death toll of each region, Southern Asia is found to have the highest total number of deaths and the highest number of deaths per disaster.
During the same time period, North America had a relatively large number of natural disasters (149) that resulted in relatively few deaths, both in total (15,000) and on average per disaster (104). Europe and the former Soviet Union had slightly fewer disasters (127). Of all the regions of the world, sub-Saharan Africa had the smallest number of disasters (45) and the smallest number of deaths (9,000); these numbers would have been much higher if epidemics, droughts, and famines had been included in the RFF database. Although the remaining region, the Middle East and North Africa, had the second smallest number of disasters (86), it had the highest number of deaths per capita.
These seven regions vary considerably in size. Although it might be expected that the largest regions would experience the most disasters, size appears to be of less importance than location in relation to the principal areas of geological and meteorological activity. For instance, Southern Asia comprises only 4 percent of the world's land area but suffered 18 percent of the disasters between 1945 and 1986.
The annual number of deaths due to natural disasters was more variable than the annual number of natural disasters. The substantial drop in the number of these deaths regionally and worldwide from 1980 to 1986 is contrary to expectations, given the increase in the number of disasters during this period. The drop may have resulted from the normal fluctuations in the magnitude of natural disasters, or it may reflect improved abilities to prepare for and respond to these events.
During the years 1945 to 1986, 917 meterological disasters killed about 1,072,000 people. Although less frequent, geological disasters, which numbered 225, claimed approximately 1,240,000 lives—about 168,000 more than meteorological disasters. The number of meteorological disasters increased dramatically over time, but the number of associated deaths did not. In contrast, the number of geological disasters remained relatively stable while the number of associated deaths grew substantially. Other types of natural disasters were much less common and claimed far fewer lives than either meteorological or geological disasters.
The most common types of natural disasters were floods, tropical cyclones, earthquakes, and other storms. Floods accounted for nearly one-third (395) of the total number of disasters (1,267). Compared with floods, tropical cyclones claimed a larger number of lives in total (791,000) and per disaster (2,907). Earthquakes caused about half the deaths attributable to natural disasters (1,198,000) and resulted in the most deaths per disaster (6,272).
Worldwide fatalities due to major natural disasters
Major industrial accidents, 1945–1986
During the years 1945 to 1986, there were 293 major industrial accidents worldwide, or an average of 7 per year. The number of these accidents that resulted in 25 or more deaths was consistently smaller than the number that resulted in 5 or more deaths (see top figure, p. 12 ).
The RFF database reveals an upward trend in the number of these accidents through 1979—the year in which the greatest number of them occurred—and a steep downward trend thereafter. The extent to which these trends are related to the level of industrial activity is unclear, given the difficulty of measuring such activity worldwide; but the level of industrial activity certainly did not fall off dramatically after 1979. The explanation could be that the rising number of accidents throughout the 1960s and 1970s led to organizational, operational, and regulatory changes that reduced the potential for accidents to occur.
Despite the decline in the number of disasters between 1980 and 1986, there was a substantial increase in the number of deaths during that period. This increase is primarily attributable to two accidents that resulted in very high fatalities: a fuel truck crash in the Salang Tunnel in Afghanistan in 1982 that killed 2,700 people, and a chemical plant accident in Bhopal, India, in 1984 that killed 3,849 people (see bottom figure, p. 12). On average, there were 350 deaths per year and 50 deaths per accident.
As was the case with major natural disasters, the frequency and severity of major industrial accidents were unevenly distributed across the globe. During the years 1945 to 1986, the United States and Canada together suffered the largest number of major industrial accidents. However, the greatest number of fatalities due to these accidents occurred in countries outside North America (excluding Mexico) and Europe (including the former Soviet Union) in the period 1980–1986.
The number of major industrial accidents peaked in the period 1973–1979. Many of the accidents during this period occurred in the southern United States and may be attributed to increased oil drilling and other industrial activities associated with the 1973 oil embargo. Until 1980, however, the total number of deaths due to major industrial accidents increased only marginally compared with the substantial increase in the number of these accidents.
During the years 1945 to 1986, there were 124 major industrial accidents in the United States and Canada, 86 in Europe and the former Soviet Union, and 83 in all other nations. Associated fatalities numbered 2,016, 2,680, and 10,280, respectively. Thus, the average number of deaths per major industrial accident was approximately 16 for the United States and Canada, 31 for Europe and the former USSR, and 124 for all other nations.
Accidents of all four types classified in the RFF database increased throughout most of the years from 1945 to 1979. Although they decreased from 1980 to 1986, the number of associated deaths during that period rose dramatically due to two fixed-facility accidents (in Bhopal and Mexico City), one transportation accident (in Afghanistan), and one pipeline accident (in Sao Paulo, Brazil), each of which had a very high number of fatalities.
During the years 1945–1986, the number of transportation-related accidents roughly equalled that of fixed-facility accidents; the number of fatalities associated with each of these types of accidents was also roughly the same. Transportation accidents accounted for 4,108 (46 percent) of the major industrial accidents and 31(45 percent) of the resulting deaths during this period. Fixed-facility accidents accounted for 118 (40 percent) of the accidents and 7,063 (47 percent) of the deaths. If not for the accidents in Afghanistan and Bhopal, transportation accidents would have surpassed fixed-facility accidents in number of total deaths and deaths per accident.
The United States and Canada suffered the largest number of major natural disasters between 1945 and 1986; but the largest number of deaths due to these accidents occurred outside North America and Europe.
The RFF database reveals that 35 percent of the major industrial accidents at fixed facilities and 68 percent of the associated fatalities occurred in the chemical industry. If not for the Bhopal accident, however, the chemical industry would have been responsible for fewer deaths than the petroleum refining industry, which accounted for 27 percent of the accidents at fixed facilities and 15 percent of the resulting deaths. The manufacturing industry accounted for 20 percent of all the fixed-facility accidents but less than 3 percent of the total fatalities. The remaining fixed-facility accidents were attributed to other industries or to unidentified industries.
The database shows that maritime accidents accounted for 47 percent of all the major industrial accidents that occurred during the transport of hazardous materials and 33 percent of the fatalities due to industrial accidents. However, roadway accidents, which accounted for 29 percent of the transportation-related accidents, were responsible for 53 percent of the deaths. If not for the Salang Tunnel accident in Afghanistan, maritime accidents would have been responsible for a greater number of deaths than roadway accidents. Railway accidents accounted for 16 percent of the transportation-related accidents and 10 percent of the resulting deaths. Road/rail accidents—that is, accidents involving the collision of road vehicles with trains—and accidents involving unidentified means of transport accounted for the remaining transportation-related accidents.
Pipeline accidents accounted for 33 (11 percent) of all major industrial accidents and 860 (6 percent) of all the deaths due to these accidents. With the exception of the oil pipeline explosion that killed 508 people in Sao Paulo in 1984, most of the pipeline accidents resulted in relatively few deaths. Overall, gas pipeline accidents were almost five times more common than liquid pipeline accidents, but the latter killed almost twice as many people (557 deaths compared with 303).
Comparative impacts of disasters and accidents
During the years 1945 to 1986, major natural disasters were more frequent and severe than major industrial accidents. The RFF database reveals that natural disasters occurred four times more often than industrial accidents and claimed more than 150 times as many lives each year. It also shows that each natural disaster resulted in more than 30 times as many deaths as each industrial accident.
The majority of deaths were caused by a relatively small number of disasters and accidents with particularly high numbers of fatalities. The ten worst natural disasters (in terms of fatalities) accounted for 66 percent of all deaths attributable to natural disasters, and the ten worst industrial accidents accounted for 63 percent of all deaths due to industrial accidents.
The annual number of deaths worldwide due to major natural disasters and industrial accidents was approximately 56,000. This number is slightly greater than the annual number of deaths due to highway accidents in the United States.
Information needs
The development of the RFF database revealed serious shortcomings in the reporting of data on major natural disasters and industrial accidents, ranging from omissions to errors and ambiguities. None of the sources used in the database had a complete record of reports on either natural disasters or industrial accidents, and there were often gaps in the information that was given. Moreover, reports of the same event sometimes differed so much from one source to another that it was difficult to determine whether a disaster or accident described by one source was the same disaster or accident described by another source.
Practices for recording the dates, locations, conditions, contributing factors, and consequences of major natural disasters and industrial accidents all need to be standardized. More complete and consistent information is also needed about how and why these events occur, and about the consequences (other than fatalities) of these events—such as the number of people injured and the magnitude of damages to property and the environment. The fulfillment of these needs is critical to the improved understanding and management of natural disasters and industrial accidents.
Theodore S. Glickman is a senior fellow in the Center for Risk Management at RFF. Dominic Golding is a fellow in the center.
A version of this article appeared in print in the June 1992 issue of Resources magazine.
A version of this article appeared in print in the June 1992 issue of Resources magazine.
