Why Did the National Hurricane Center Blow It on Earl?
Businesses gearing up for a profitable Labor Day weekend in a recession economy were devastated, but it wasn’t Hurricane Earl that did the damage. It was the inaccurate predictions from the National Hurricane Center.
From Cape Cod to Maine thousands of would-be vacationers cancelled their hotel reservations, as hurricane warnings were issued for the Cape and tropical storm warnings were posted all along Maine's coast. Reacting to these warnings local media sprung into action -- mighty Earl was on the way! Reporters waited along the beaches from New York to Maine.
Amtrak cancelled service from New York to Boston. Continental Airlines cancelled 60 flights.
The strongest sustained wind on Nantucket Island was 41 miles per hour -- just above tropical storm strength of 39 miles per hour. The peak wind gust there was 54 miles per hour -- far short of the minimum hurricane force of 74 miles per hour. Rainfall totaled 2 to 4 inches.
In a part of the country accustomed to dealing with Nor’easters, this was just another windy, rainy Friday night. By Saturday morning Earl was gone, along with profit opportunities for hotels and restaurants across southern and eastern New England.
Is the National Hurricane Center guilty of yet another example of manufactured panic by the government that brought you Y2K, West Nile virus, and the bird flu scare? Is there a systemic process in government that tends to blow events and potential events out of proportion?
The Hurricane Center did accurately predict the path of the hurricane as it passed south of New England. It was forecast to pass south of Nantucket Island Friday night as it was weakening to a less potent tropical storm.
If they got the storm track right, why were the other elements of the forecast so bad?
The winds of a hurricane blow counter-clockwise in the northern hemisphere, making the strongest winds of a hurricane on the storm's right side -- where the forward motion of the hurricane and the winds around the center are generally in the same direction. The opposite effect is realized on the left side of the storm. As Earl was moving to the northeast, the wind on the north side of the storm was blowing opposite to the storm's forward motion. This has the effect of reducing the winds on the left.
Long Island, Connecticut, Rhode Island, Cape Cod and the islands, coastal New Hampshire, and Maine were accurately predicted to be on the weaker side of Earl. But in spite of this, tropical storm warnings (meaning winds of 39 to 73 miles per hour) would hit coastal New Hampshire and Maine as Earl passed well to the south Friday night.
Some hotels at Old Orchard Beach, Maine, lost 80 percent of their reservations when vacationers heard the scary forecasts. Saturday, Sunday, and Labor Day Monday’s weather was spectacular on the coast of Maine, but many vacationers weren’t there to enjoy it.
I was a television meteorologist for 25 years, seven of them working at the NBC station in West Palm Beach, Florida. I hosted and produced a dozen one-hour hurricane preparedness programs while there. I worked closely with the National Hurricane Center. Today, I am an independent meteorologist, and I was monitoring the forecasts from the National Hurricane Center when Earl was coming across the Atlantic. As it approached New England, one element of the predictions from the Hurricane Center nagged at me -- the overly broad area of tropical storm winds being predicted.
The Hurricane Center has a large number of dynamical and statistical computer models that they use to help them predict tropical storms and hurricanes. The forecasts they get from these models can have a fairly wide range of solutions both for the track of the storm and its strength. It was a consensus of these various models, and the experience of the forecasters themselves, on which the Hurricane Center based its forecasts for Earl.
I was monitoring the forecasts of the GFS (Global Forecasting System) model that is produced and distributed over the internet four times a day from NOAA. The GFS is not a hurricane model. It was not developed to specifically predict the movement and behavior of hurricanes, as many of the models at the Hurricane Center are. However, I have found through the years that the GFS does a credible job at forecasting the track and strength of hurricanes.
As Hurricane Earl was 1,200 miles southeast of Cape Hatteras, North Carolina, the storm was getting stronger. The maximum sustained winds would reach 140 miles per hour by Wednesday night. Even this far from New England, the GFS model was predicting that the strongest winds of a weakening Earl would be restricted to Cape Cod and the islands Friday night and early Saturday morning. There was never any forecast that indicated strong winds of tropical storm strength would blow anywhere near coastal New Hampshire or Maine.
As Earl moved closer, each successive forecast from the GFS continued to indicate no strong winds at all across southern Connecticut, Rhode Island, northeastern Massachusetts, New Hampshire, or Maine. I called my sister, who lives in coastal southeastern Rhode Island, to tell her not to worry about the storm -- I assured her there would be no damaging winds or significant impact. I talked to a friend who owns a marina in Groton, Connecticut, and assured him that his marina would suffer little or no impact from Earl.
In Maine, the Portland Yacht Club pulled some 100 boats from the water to prepare for the “tropical storm force winds” predicted for the coastal waters Friday night and early Saturday morning. The winds never came, but boat owners looking to enjoy time on the water during the holiday weekend were certainly impacted. Many coastal restaurants, shopkeepers, and other tourist-related businesses suffered the impacts of the forecast as well -- all at a time when many are struggling to stay afloat financially.
Overly cautious warnings by the Hurricane Center may help to protect them from legal liability should deviations in their forecasts occur. However, causing businesses to close and lose money when there were strong indications that tropical storm conditions would not strike is neither good forecasting nor good policy.