Nitpicking the WSJ Numbers Guy:
My print column this week examines the tricky science of projecting hurricane damage ahead of landfall, as the east coast of the U.S. braced itself for the expected arrival of Hurricane Irene.
Roger Pielke Jr., a political scientist at the University of Colorado, Boulder, who has studied hurricanes’ impact, posted on his blog Friday several analogues to Irene, with the caveat, “None of the storms is really a good analogue.” He noted he’d been getting many calls after Nate Silver, in a blog post on the New York Times website, projected possible Irene damage based on a study Pielke and colleagues conducted of prior hurricanes’ economic damage. ...
Because there are scant recent data on deaths from storms, most forecasters focus on economic damage, which Pielke found has been proportional to the fourth power of wind speed. Willoughby says that storm intensity is proportional to the cube of wind speed. So why the higher exponent? One possibility is that storms of a certain intensity pass a threshold beyond which they “breach buildings’ envelopes,” meaning they begin to cause damage within buildings, ramping up damage quickly. Other studies have found that storm damage may vary with the seventh power or even the ninth power of wind speed.
Pielke’s damage calculations don’t include inland flooding, which could be a big problem with Irene. His team wasn’t able to find a single number that predicts flooding damage, unlike speed, for wind. In the Midwest, six-week rainfall totals were most important, while in the southwest, rainfall over a few hours better predicted the biggest danger, of flash flooding, according to Pielke.
For these and other reasons, including the usual uncertainties of any climate forecast, projecting hurricane damage is no sure thing. “It’s really uncertain at this point what the damage might be exactly,” Pielke said. “We could get lucky or we could get really unlucky.”
via blogs.wsj.com
The simple rule of thumb above seems a little too simple, which is the point of the question mark in this post title. Rules of thumb, like one million dollars per mile and wind speed to the Xth power, are OK in terms of giving orders of magnitude estimates but they are usually wrong. In the hurricane case, it really matters where the hurricane lands. Landing near Miami, New York and New Orleans will generate lots of damage. Landings near less populated areas will generate less damage.
Also, I'm not sure what this line is all about: "Because there are scant recent data on deaths from storms, most forecasters focus on economic damage, which Pielke found has been proportional to the fourth power of wind speed." What does "scant recent data on deaths" as motivation for economic damages all about? Deaths and economic damage are two measures of the anthropocentric severity of a storm. The death data can be found at the National Hurricane Center website.
Note: My favorite line from today's hurricane coverage was something like "Hurricane Irene landed near Cape Lookout as a powerful ... Category 1 storm (gasp)." The Weather Channel seems particularly disappointed that the wind speed isn't a bad as forecast.
