Risk: Storm tracking - Forecasting success.

(From Reinsurance)

Byline: Mark Saunders, lead scientist on the TropicalStormRisk.com

The active 2003 Atlantic hurricane season has already seen two hurricanes strike the US. Hurricane Claudette struck Texas on 15 July and hurricane Isabel hit North Carolina on 18 September (Figure 1). In addition, damaging hurricanes have struck Bermuda (hurricane Fabian on 5 September) and Halifax, Nova Scotia (hurricane Juan on 29 September). However, it was Isabel which most captured the attention of (re)insurers.

After early estimates placed Isabel's expected insurance loss at $3bn, business professionals throughout the world were glued to their computer screens for several days as the giant storm tracked towards US landfall.

Although Isabel's predicted industry loss steadily came down as the storm weakened, eventually making landfall as a weak category 2 hurricane with maximum sustained winds of around 100mph (85 knots), the threat remained high enough for federal government buildings to be shut down in Washington DC, 300 miles from the point of landfall.

A feature of hurricane Isabel was the success of its track forecasts.

Landfall was predicted correctly at a lead of 72 hours to within 20km and 1 hour. Accurate hurricane track and intensity forecasts have been used traditionally to issue evacuation warnings and to save lives. However, increasingly they are also being employed as an important financial risk tool for insurers, reinsurers and risk managers. The purpose of this article is to describe how hurricane forecasts are produced, to highlight the remarkable success of the track forecasts for hurricane Isabel, and to outline the value of such forecasts to (re)insurers.