LJWorld.com weblogs Lawrence Weather Watch

“Extreme” Hurricane Season Coming?


Sounds a bit extreme, don’t you think? There are some out there in the private forecasting community who think the 2010 Atlantic hurricane season will be one for the record books. Of course, it’s not impossible. However, the problem with this line of thinking is that predictions for extreme outcomes — like for ‘extreme’ hurricane seasons — are often not useful when the predictions themselves are only slightly better than climatology.

To be sure, our science has progressed to the level at which it is indeed possible to make a forecast for next year’s Atlantic hurricane activity that is superior to the forecast for just an average season. The team of hurricane forecasters from Colorado State University has demonstrated this in recent years. Their predictions provide a nontrivial advantage over the “average” forecast. This is quite an achievement, especially when you consider the fact that the Atlantic basin has the largest year-to-year variability of any of the global tropical cyclone basins. But we still need to be wary of forecasts for tail events (ones on the edges of the probability distribution), particularly when it comes to hurricanes.

As far as this upcoming season is concerned, there are some early signs that suggest there will be more hurricanes than normal. One is the unusually warm sea surface temperatures in the main development region in the tropical Atlantic. Another is the expected decline of El Nino (and the reduction of vertical wind shear associated with its demise). Since wind shear tends to tear hurricanes apart, less wind shear can help promote a more favorable environment for storms to grow. A wild card in all of this, and perhaps one of the stronger predictors, is the MJO, or Madden Julian Oscillation. The MJO is a tropical disturbance that can modulate the likelihood of hurricane genesis for weeks at a time. Unfortunately, there is no clear indication yet how the MJO might set up during the upcoming season. I think we’ll have to wait until early summer before hurricane forecasters will have a reasonably confident picture of how extreme this season will be.


riverdrifter 8 years ago

"The team of hurricane forecasters from Colorado State University has demonstrated this in recent years." In 2009 Bill Gray's team forecast an extreme hurricane season didn't they? How did that work out for them?

I agree with your last sentence 100% Greg, BTW.

gpostel 8 years ago

Hi Riverdrifter,

This group, yes, provides skillful hurricane frequency forecasts (i.e., better than climatology). In fact, between 1950 and 2007, their DEC forecasts for the following year have shown a correlation with observations at 0.73 (which is an r^2 of 0.53). All from a forecast nearly a year out. Outstanding! That means their long-range forecasts account for over 1/2 the variability in the observations. In the game of hurricane forecasting, that's really, really good. Now, year by year, there are misses such as the one you mention in 2009. One needs to understand, though, that these kinds of errors in the predictions are not only expected in the nonperfect world of seasonal hurricane forecasts, but also that their existence does not mean that the progs are not useful and skillful. And that's all that I was getting at. And for the record, their DEC 2008 prediction for season 09 was for a more active season (and depending on which metric you choose - only slightly more active. For example, they forecasted 7 hurricanes and the 50 year average is 6). By no means did they call for an extreme season.

gpostel 8 years ago

Hi LarryNative

totally agree. it's almost like you have to come up with a new vocabulary to catch someone's attention these days. For example, when you go to the store, and the clerk asks if you want paper or plastic, and you say paper. Why is the clerk's response, "awesome" ? is it really awesome that I chose paper instead ? kinda funny.

gpostel 8 years ago

Hi thebcman,

yep. 09 was a bust. 3 hurricanes and 9 named storms.

Commenting has been disabled for this item.