Stats Made Easy

Practical Tools for Effective Experimentation

Monday, February 12, 2007

Weather to be or not to be, that is the question

Last night I got a panicked call from my host for a talk scheduled tomorrow night to a group of quality professionals and their student section at Purdue University. Predictions had just firmed up for a major winter storm that might dump up to a foot of snow in parts of Indiana. Which parts would get snow was hard to forecast, but it seemed likely to be rain south of Indianapolis – my flight destination, icy there and snowy to the north in Lafayette – home of Purdue. Thus, given I’d be driving through the middle of this wintry mess, my host's bias toward canceling the meeting met with little resistance from me. At the moment, based on tonight’s weather reports, it appears that we made the right decision. However, I’ve seen plenty of dire weather predictions fizzle over the years, particularly for snow and/or ice, which often end up precipitating as relatively benign rain due to unexpected warmth.

North American winter storms can wreak havoc on a grand scale, for example, when ice builds up to a point where power lines come down over broad areas. However, hurricanes like Katrina really strike fear in the hearts of insurance underwriters. Richard Mullins of the “Tampa Tribune” reports* on the use of simulations for predicting the financial scale of disasters like this. According to him, some storm models sell for as much as $10 million! For that price, one would assume the results would be unbiased. However, non-profit and privately-funded researchers interviewed by Mullins agree that results from studies underwritten by insurance companies naturally fall to the high side, whereas ones done for the public interest tend to the low end. The range went from $2 billion to $12 billion for 2005’s Hurricane Wilma!

Things really get wacky when one tries to assess risks of buying a vacation property in Florida to escape the wretched winter weather of the northern USA, from Indiana on up. Where would one be safest in a beach home – a place like Jacksonville that’s experienced no category 4 hurricane in 150 years? Maybe they are ‘due’ for one. A contrarian might take an opposite tack – buy where the most recent horrific hurricane hit, such as the surprisingly robust Wilma that tracked in to Florida after clobbering Cancun.

My idea is to simply rent a haven in Florida during the winter – the season when there are no hurricanes. I would leave at the first sign of snow up north and not go back until it melted. I wonder if any fellow northerners have thought of this?

*“Calculating Disaster,” Sunday, 2/11/07


Post a Comment

<< Home