What is vital to remember about the accuracy of hurricane forecasts is that not everyone shares the same view as to what an accurate forecast is. Citizens of an affected area will often feel as though they didn’t get an accurate advance warning of what they encountered, statistical analysts that perform best-track reviews will then often have their own view as to what is accurate, and of course the forecast teams that issued the warnings will also have their own take on the matter. What this often leads to is a lot of confusion and a fair dose of double speak. In order to honestly look at what is considered accurate, the best way to start is by examining what some of the various agencies which issue hurricane forecasts consider accurate, the process they use to arrive at those forecasts as it is a factor in accuracy, and finally the way accuracy is viewed by the individual players that produce and use the forecast.
As it is wholly unfair to use only one agency to discuss the accuracy of hurricane forecasting, the NHC (National Hurricane Center) they need to be examined alongside the PWC (Pacific Warning Center), and JTWC (Joint Typhoon Warning Center). For anyone unaware of the difference between a hurricane and typhoon, it is nothing more than a matter nomenclature related to the storms position in relation to the International dateline. In this manner a fair view can be presented as to what standards are in place and how each actually stacks up in comparison to each other. While Japan has an established warning center for typhoons as well, they will not be used because they follow a different model for forecast issuance which is very misleading. It should also be noted television meteorologists are not going to be used as a standard for comparison as the vast majority do nothing more than read these agencies forecasts in their own words.
For starters a hurricane forecast is based on three distinct but related sources of data; satellite imagery, computer generated forecast models, and good old fashioned hard chart data based on land, sea, and upper air readings divided into two separate charts. The Surface chart creates a picture of events at the 850 millibar level while the upper air chart serves the same purpose at the 200 millibar level. The upper air chart of course does not employ land and sea based data, but rather pilot reports (Pireps), and upper air soundings (UJ reports). Computer models are only so good as the information they were programmed with so in some cases a fair dose of the human element of experience and gut instinct can come into play depending on the individual knowledge possessed by the members of each forecast team.
To define what is considered accurate, between the three major warning centers there are two different standards not wholly known to people that have not worked in the specific field. In the case of JTWC and PWC, the standard for success is to have forecasts measured at the 24, 48, and 72 hour intervals. The goal is to have a forecast which provides an eye position (Fix) which is no more than 100, 200, and 300 nautical miles awry respectively at those designated points of measure. While that sounds quite simple, it is a big planet. The NHC however uses a broader criteria which is less well defined to measure their accuracy. The use more of a geologic marker system. For instance if they were to predict landfall for a hurricane in the southernmost portion of the state of Georgia, but it instead made its landfall in the northernmost portion of the state of North Carolina, that would be considered a successful accurate forecast. The same would be true in reverse as well.
The second area accuracy is to be examined lies within intensity. The NHC does not really have a defined parameter for success in this area. Loosely speaking the aim is to be within 25 mph. That is a fairly large window to work with. JTWC and PWC however aim for a 15 mph window. Ten miles per hour agin may not seem like much of a difference, but that is also the baseline used to predict average and maximum wind gusts so it does make a difference.
In terms of statistical analysis performed by each center which is examined and reported after the fact by Best Track officers or similarly titled persons, the historic pecking order for accuracy has been JTWC, PWC, and NHS bringing up the rear in comparison to each other under the most strictly defined parameters for success. JTWC is still the only agency to have met and surpassed the 100, 200, 300nm measuring stick at the 24,48, 72 hour intervals over the course of an entire year. The others have met this in part, but never in full. While NHS boasts that they do an excellent job, they barely manage to meet the most base level of criteria for success they define and can amend of their own accord.This has been statistically true in comparison to the other warning centers over a period of decades.
With the knowledge of what the forecasters hold themselves to as a standard for accuracy, the next thing to look at is how the public views their forecasts. For the most JTWC and to a slightly lesser degree PWC have enjoyed a tremendous amount of respect from the people whom are the end users of their forecasts. NHC on the other hand is taken with a bit of a grain of salt, ask people going all the way back to Hurricane Belle to Hurricane Hugo, and more recently Hurricane Katrina how good the information they received was. While NHC is not terrible, they tend to lag a bit as their performance is not tied to the money make. NHC is comprised of civilians, while JTWC and PWC are military centers that overlap to serve the population as a whole.
The general trend is that NHC is improving with the accuracy of their forecasts against their peer agencies, but the public does view much of what they provide with a bit of skepticism. While public perception is not a true indicator of accuracy as no group of people will ever be truly happy with any forecast regarding a force of nature like a hurricane after they have felt its wrath, it is something which indicates a established trend of publicly busted forecasts. As the measuring for accuracy is an internal process which is also subject to the possibility of post analytic “smoothing” of the numbers, skepticism can also be present in the viewing of those.
For the most part each center does a serviceable job trying to perform an impossible task. There is a reason meteorologists are referred to as weather guessers, no matter how much data you have, no matter how good that data is, predicting the future in regards to a violent force of nature is really nothing more than guessing. The data and experience makes it more of an educated guess, but as people are fond of saying, “It is what it is.”
