Roger Pielke Jr: The Decline of Tornado Devastation
Despite what you might have heard about ‘extreme weather events,’ damage and loss of life from twisters is in retreat.
So far in 2014, the United States has experienced fewer tornadoes than in any year since record-keeping began in 1953, or even before. Greg Carbin, a meteorologist with the Storm Prediction Center of the National Oceanic and Atmospheric Administration, has called this “likely the slowest start to tornado activity in any year in modern record, and possibly nearly a century.” But just because tornado activity has declined doesn’t mean that we can let down our guard, as potentially large impacts are always a threat.
Overall, however, the good news for residents of the Midwest’s “Tornado Alley” and elsewhere is that over the past six decades America has witnessed a long-term decrease in both property damage and loss of life. That’s the finding that I and Kevin Simmons and Daniel Sutter, two of the nation’s leading tornado experts, have gleaned from studying the data on almost 58,000 tornadoes observed since 1950.
Using estimates collected by the NOAA’s Storm Prediction Center, we used several approaches, including taking inflation into account, to “normalize” historical losses to 2014 dollar values in order to estimate how much damage would occur if tornadoes of the past occurred with today’s levels of population and development.
The nearby chart shows our central estimates. How do we know if they are any good? We performed several independent checks. For instance, we already know which years had exceptionally large property losses: 1953, 1965, 1974 and 2011—and these four years show up clearly in our data set. A more sophisticated check is to compare trends in the incidence of tornadoes with trends in damage. Counts of tornadoes at different strengths can serve as an independent basis for evaluating our methods. Logically, these trends would match up. And once we break the overall data set into a series of shorter periods to take into account changes in the way meteorologists have tracked tornadoes over time, the trends do match, supporting our approach.
Certainly the potential for tornado damage in the U.S. remains strong. Just three years ago, the country was wracked by a series of particularly destructive storms, including a tornado outbreak in late April 2011 that killed more than 300 people across seven states, and one in May that devastated Joplin, Mo. That year was one of just three since 1950 with more than $25 billion in damage, and the 560 deaths in 2011 were the most fatalities since 1925, when 794 people died.
The average annual U.S. property losses caused by tornadoes, from 1950 to 2013, is $5.9 billion in today’s dollars. However, for the first half of the data set (1950-81), the annual average loss was $7.6 billion, and in the second half (1982-2013), it was $4.1 billion—a drop of almost 50%.
Does the substantial decline in average annual damage mean that there have actually been fewer tornadoes? Not necessarily. The U.N.’s Intergovernmental Panel on Climate Change concluded recently that the quality of the available data on tornado frequency and intensity makes drawing conclusions about long-term trends problematic: “There is low confidence in observed trends in small spatial-scale phenomena such as tornadoes and hail.”
What we can say with some certainty is that the number of years with very large tornado losses has actually decreased. Consider that from 1950 to 1970 the U.S. saw 15 years with tornado damage in excess of $5 billion a year. From 1993 to 2013 there were only four such years, with three since 2008.