I have written several blogs on preparedness over the past couple of months. I wanted to get most of that covered before the peak of the hurricane season. During the next few weeks, I plan to focus more on forecasting and, in particular, on the limitations of the forecasts. We will take a look at forecasting track, intensity and structure. Today, we will start with track forecasting.
One of the success stories with tropical cyclone (TC) forecasting is with track forecasting. Observations are better, computer models are more sophisticated, and computers are faster. The National Hurricane Center (NHC) verifies every TC forecast it makes. The trends lines in the below graph clearly show the improvement over the years.
The above graph as well as most of the information in this and upcoming blogs comes from the excellent verification page on the NHC web site at http://www.nhc.noaa.gov/verification/. I thank Hurricane Specialist James Franklin for the detailed information on individual TCs in recent annual verification reports. The NHC has been very honest in reporting its errors and wants everyone to understand the uncertainties involved with forecasting.
I worked at the NHC from 1972 to 2007 and have seen numerous track forecasting records set. But there is a natural volatility in TC track characteristics and annual errors can vary significantly from year to year. For this reason, representative error characteristics must be obtained using a longer period of record.
The NHC now uses a 5-year sample to define its long-term track forecast error characteristics. The official NHC 5-year average track forecast errors (2003-2007) are as follows: 67 miles at 24 hours, 122 miles at 48 hours, 177 miles at 72 hours, 239 miles at 96 hours, and 313 miles at 120 hours. That is approximately 60 miles of track error per day.
We should be thankful that track forecasting is improving. But it is critical that we understand the limitations. Even today with all the advances in forecasting, the 24-hour track forecast error is over 60 miles. That means that a hurricane forecast to make landfall over Miami Beach in 24 hours could easily hit up the coast near Palm Beach or down the coast near Tavernier in the Florida Keys.
I’ll share more on the NHC’s Cone of Uncertainty and Probabilistic Wind Speed Graphics in upcoming blogs.