No sooner has severe weather ripped through a region, as a series of 305 tornadoes did the American South from Apr. 25-28, than the rush begins to announce estimates of the cost—in lives lost and property damage. Often the property estimates, in particular, arrive before it's humanly possible for anyone to make a fact-based assessment. That's because the first numbers used to shorthand the extent of the destruction come not from people but computer models, and draw on data from the storm itself, not from observed effects.
"As an engineer, just knowing the severity of the winds helps us understand the damage," says Tom Larsen, senior vice-president for Eqecat in Oakland, Calif. Eqecat builds software to help insurance companies "quantify exposure" to natural and man-made disasters. Based on how powerful a tornado was, how long it stayed on the ground, and how wide it was, Eqecat estimates what it did to buildings or crops in its path. "We can't exactly predict [that] there is a mailbox that is going to go through a window, but we can look at claims data in the past and say on average for this level of winds this is the kind of insured losses that you will see."
The National Weather Service dispatches teams immediately after storms to survey the damage on the ground; if it's really bad, as it was in Alabama on Apr. 26, the Federal Emergency Management Agency will send inspectors to determine if enough damage has been caused to justify use of federal funds. So far in 2011, FEMA has made 24 such declarations—12 for winter storms, 8 for tornadoes.
None dispute the increased frequency of tornadoes: As of May, there have already been more destructive twisters in 2011 than typically occur in a year. Larsen and other catastrophic risk modelers do, however, debate how much of a factor climate change is, and dispute the suburban myth that tornadoes are attracted to trailer parks. "It's just that trailers are more easily damaged at low wind speeds," he says, "and make for better damage photos in the aftermath."