The last couple of blogs have been on the National Hurricane Center’s (NHC) “Cone of Uncertainty.” Yesterday, we looked at an example from Hurricane Ike in 2008 when the cone first moved over South Florida at the 5-day time period. Today, let’s look at a couple of other examples.
The graphic below shows the Cone of Uncertainty issued at 11:00 am EDT Monday, 8 September 2008.
Let’s say that we lived on Galveston Island, Texas. Although Ike was centered over eastern Cuba, the cone at the longer time periods clearly included the Galveston area with the 5-day center of Ike forecast to be somewhere around 100 miles southeast of Galveston. Folks in the Galveston area had been within the cone for 24 hours and should have already been thinking about what they would do if the threat continued increasing.
The graphic below shows the Cone of Uncertainty issued at 11:00 pm EDT Monday, 8 September 2008.
By this time, the NHC had shifted the track southward with a 5-day center position forecast over southern Texas. Recall that Ike made final landfall on the northern end of Galveston Island very early on 13 September. But the cone at this time (11 pm EDT 8 September 2008) included all of the Texas coast and much of the Louisiana coast. Given that a hurricane is not a point but a large circulation (especially in the case of Ike), those in the Galveston area should not have interpreted this forecast as anything resembling an “all clear.” Most people understood this, but I suspect some did not.
Remember, the Cone of Uncertainty is created such that two thirds of the time, the actual track of the tropical cyclone stays within the cone based on the past 5 years of forecast errors. The NHC eventually shifted the forecast track farther to the north which was closer to the actual track. But the fact that the center of Ike did not exactly follow the “skinny black line” shown on the cone graphic was not a surprise to the forecasters. It should not have been a surprise to anyone.