I grew up racing sailboats on the Great South Bay of Long Island in New York and became intrigued by the weather, especially the winds and waves. We experienced a few hurricanes and I was always glued to the windows (not a good idea!) to see how far the trees could bend in the gusts. I raced in college at Florida State and developed a keen interest in the behavior of winds...to try to get a competitive advantage out on the race course. I still compete in sailing today and am a member of the 2006 U. S. Sailing Team in the RS:X Windsurfing class. Hurricane winds are fascinating on a variety of time and space scales; there are few things so beautiful (or maybe I should say sickening) as the sea surface as observed out the window of a NOAA P3 flying 250 m above the waves in 100 kt winds!

“...I don’t really pay much attention to whether it’s expected to be an active year or not—I want to know whether Andrew is coming.”

A few years ago we were asked by the state of Florida to assist in developing a public insurance loss projection model. We have a team of scientists from several universities representing Atmospheric Science, Wind Engineering, Statistics, Computer Science, and Finance. The models available today for rate-making are proprietary, so it’s important that the state have the ability to look at results and evaluate risk with a baseline model that is transparent, independent, and open to public scrutiny.

Starting with the characteristics of the hurricane climate record, we simulate thousands of years of hurricane activity, look at the probabilities of a given location having winds above various levels, and compute the damage and the resulting insurance losses. These models have been around since the mid-1980s and as computing power has improved they’ve become very sophisticated, with the ability to simulate the complete lifecycle of storms from genesis through several days of movement, to landfall, and then decay. The challenge is to continually evaluate and improve the model by bringing in new findings as the research progresses in several fields. At the same time we are required to undergo a vigorous model-certification process each year which expends an incredible amount of time and resources; it should probably be every two years or so.

Landfall forecasts are a difficult concept. Specific hurricane landfall forecasts are not issued because forecasters and emergency managers do not want the public to focus on a particular location when the uncertainty of the forecast is such that in the next six hours the forecast will shift to a new location, and then another one or perhaps back again in another six hours. However, it’s still pretty easy for the public to draw the "thin line" and see where the forecast crosses the coast, effectively making it a landfall forecast. In my opinion, these are the most important forecasts and need to be evaluated after each event.

Our paper evaluated the errors and found no significant improvement in forecasts of landfall position over a 25-year period, compared with improvements of ~1% per year for general position forecasts in the Atlantic basin. Landfall forecasts comprise only about 13% of all forecasts issued so our sample size was relatively small but we are continuing this work and hope to have updated accuracy estimates for the period of 1976-2006 by next year. The new NOAA strategic plans no longer specifically discuss improving hurricane landfall forecasts, but the USWRP landfall forecast goals are still listed in announcement of federal funding opportunity for the Joint Hurricane Testbed.

  How bad do you think this year's hurricane season will be? Are we prepared?

The seasonal forecasts of hurricane activity seem to capture the media’s attention and everybody and their brother are doing press conferences about their forecast or forecast update. In 1992 Hurricane Andrew struck during a year with a seasonal forecast of "lower than normal" activity. Speaking as a Floridian affected by Andrew and other storms, I don’t really pay much attention to whether it’s expected to be an active year or not—I want to know whether Andrew is coming. The people of Florida have experienced eight hurricanes over the last two years, North and South Carolina had a bunch of storms from 1996-2003, and Alabama, Mississippi, Louisiana, and Texas dealt with Ivan, Katrina, and Rita recently, so the people in those states are probably well prepared. The places to worry about are where there have not been storms in folks’ recent memory—probably Georgia, the southern Texas coast, and the Atlantic coast from Maryland to Maine.

  If you are free to do so, would you tell us about your current projects?