Over the last 30 years, the United Nations, climate scientists, and governments around the world have claimed that climate change is making natural disasters including hurricanes, floods, and heatwaves more frequent.
“Climate change has helped drive a fivefold increase in the number of weather-related disasters in the last 50 years,” reported National Public Radio last fall, citing a report by the U.N.’s World Meteorological Organization.
But the data also show that the number of climate related disasters actually declined over the last 20 years by about 10 percent.
The best-available data comes from International Disaster Database, or EM-DAT, which is collected and made public by the Center for Research on the Epidemiology of Disasters (CRED) in Belgium.
The discovery that disasters have declined since 2000 was done by Roger Pielke, Jr., a University of Colorado, Boulder political scientist. Pielke said the finding “is very good news and completely contrary to conventional wisdom.”
Some people on Twitter objected to Pielke’s analysis. “You're using a regression line for a data set with a high degree of natural variation,” wrote one person, “to draw a very broad conclusion. And the slope of that line is barely inverted. This does not show anything.” Said another, “This is a textbook example of not knowing what standard deviation or regression mean!” Other people noted that the frequency of climate-related disasters rose from 1900 to 2019.
But Pielke said that no tests of statistical significance were required to show the decline. “If my bank account had $100 last week and today has $90, then it has 10% less money,” said Pielke. “Period. You don’t need to apply tests of statistical significance to see that. In technical terms the data here is the population, not a sample from a population.”
And Pielke doesn’t disagree with the trend showing increasingly frequent disasters before the year 2000 but noted that it didn’t contradict the data showing a decline since 2000. “The period since 2000 is viewed as the most reliable for data reliability,” Pielke said, “but it is safe to say that even since 2000, coverage has improved. So the 10% decline is possibly an underestimate.“ Moreover, noted Pielke, “the trends are consistent with independent, peer-reviewed research (e.g., this and this).”
So the two trends, one of rising climate-related natural disasters for a century until the year 2000, and then declining disasters since 2000, are both true. Why is that?
What do you mean by “disaster”?
CRED says its EM-DAT database includes disasters since 1900 that meet one of the following criteria
10 or more people dead;
100 or more people affected;
The declaration of a state of emergency
A call for international assistance
EM-DAT’s finding of less frequent disasters is consistent with long-term trends. There has been 92% decline in the decadal death toll from natural disasters since its peak in the 1920s. In that decade, 5.4 million people died from natural disasters. In the 2010s, 400,000 did.
The 92% decline in deaths over the last century occurred during a period when the global population nearly quadrupled, and the global temperature rose 1.3 degrees centigrade.
Pielke finds that the cost of natural disasters globally also declined as a share of GDP between 1990 and 2020. In a 2020 review of 54 studies over the last 22 years, and published in the field’s leading scientific journal, Pielke found “little evidence to support claims that any part of the overall increase in global economic losses documented on climate time scales is attributable to human-caused changes in climate.”
Pielke, who is one of the world’s leading authorities on natural disasters and climate change, has been widely published in the peer-reviewed literature and was the first scholar to show the need to “normalize” natural disaster data, which has been accepted by the Intergovernmental Panel on Climate Change across multiple reports.
Natural disasters are much more expensive in 2022 than they were in 1922. But, as research by Pielke and others show, the increase can be explained entirely by rising wealth. Normalizing the data by accounting for rising wealth is thus essential when looking for a climate change signal from the noise of rising prosperity.
Over the last four decades, poor nations like Bangladesh have reduced death tolls by over 90% thanks to simple measures like cyclone warning systems and storm shelters. “Look at Cyclone Ampham in India and Bangladesh earlier this year,” said Pielke. “It killed about 120 people. Fifty years ago, it would have killed thousands.” A cyclone in Bangladesh killed 135,000 people in 1991 while another in 1970 killed 300,000.
In other words, hurricanes, floods, and other natural disasters resulting from extreme events aren’t getting worse. They’re getting better. Much better. Given the flood of alarming news about climate change, many will be surprised to learn that hurricanes aren’t increasing in frequency, and that deaths from natural disasters are at their lowest point in 120 years. “A total of 2,900 people lost their lives in natural disasters in the first half of the year,” announced Munich Re on in 2020, “much lower than the average figures for both the last 30 years and the last 10 years.”
What, then, explains increasingly frequent disasters from 1900 to 2000?