The 2010 NOAA seasonal hurricane forecast is out. The prediction is for a banner year or in government speak: “active to extremely active.”
14 to 23 named storms (top winds of 39 mph or higher), including:
8 to 14 hurricanes (top winds of 74 mph or higher), of which:
3 to 7 could be major hurricanes (category 3, 4 or 5; winds of at least 111 mph)
This would be a very valuable thing if
- There was a history of accuracy in this forecast
- The numbers meant anything
The number of storms is inconsequential unless one hits you. Even then only one counts. Human tragedy doesn’t parallel the number of storms. 1992 was a very light season and the year of Hurricane Andrew!
Beyond that top wind speed is only one piece of the equation. As with real estate ‘location, location, location’ is the critical factor. A poorly placed tropical storm can do lots more damage than a monster out to sea.
A nameless friend, well known in the small world of tropical weather experts, likens NOAA’s wide range of potential storms to a gambler betting every number at roulette then bragging when he hits! He went on:
“14-23 storms are you kidding me? That must be 75% of the full distribution. I thought the point of a forecast was to narrow the possibilities from the full range.”
Usually an ensemble of computer models is run to produce a forecast like this. Then the outputs are pared to the most likely results. He’s complaining too many possible scenarios have been left in. The forecast deck is stacked so the forecaster can’t lose!
The Numbers Guy at the Wall Street Journal is similarly complaining.
It is possible that NOAA is keeping its estimates loose to avoid having egg on its face later: In the past nine years, NOAA’s May storm-count predictions have proved accurate about 40% of the time.
Why do they even produce these forecasts? Who benefits?
NOAA and the Hurricane Center do great public good when they forecast individual storms. However, these dubious long range projections only cause skepticism down the road. Just what we don’t need!