Cycle of Cyclones or Something Else?

August 28, 2005 – Vol.10 No.23

A CYCLE OF CYCLONES, OR SOMETHING ELSE?

Ninety degree waters fueled Katrina to a Category 5 hurricane. Forecasters doubted that Katrina could maintain that status and would likely diminish a notch before hitting land. They were correct.

Those same hurricane experts from the U.S. National Hurricane Center (NHC) predicted landfall for Katrina close enough to New Orleans, or other heavily populated Gulf coast areas, to cause a catastrophe. Right again.

Now they’re also saying that worst of the season is yet to come. Tropical Depression Lee has come and gone. Hurricane Maria is wandering in the central Atlantic as I write this. (Maria is no threat to land other than perhaps Greenland)

Officially the NHC says that we are in the midst of a multi-decadal signal for increased hurricane activity in the Atlantic that could last a decade or more longer.

Hurricane activity, they say, goes up and down with a few decades of normal activity switching to a few decades of increased activity. Plotted on a curve, the up and down multi-decadal signal shows up as a wave. A full up and down cycle should be about 50 years or so, more or less.

(However, years with El Nino events in the Pacific would be down years for Atlantic hurricanes even in the middle of upward hurricane activity swing.)

The NHC, and thus the U.S. government, says that main contributing factor of this year’s busy season is that multi-decadal signal, which includes those high sea surface temperatures (SSTs) which fueled Katrina.

(Keep in mind though that strength of Katrina can be attributed to high SSTs, but the landfall near New Orleans is matter of steering currents in the atmosphere not the multi-decadal signal. A similar hurricane could have just as easily hit the same area in a down decadal period as well as in an up cycle. But, with a greater number of hurricanes landfall will, obviously, be more frequent. Hurricanes could strike in the same place more than once in a year. New Orleans could get hit again anytime. )

The NHC doesn’t give a reason for the increased SSTs. Outside the agency some say the warmer than expected water is a result of global warming. Most scientists disagree saying that we’re too early in the warming process for this to happen.

But is there something flawed in the theory of the multi-decadal signal?

That flaw, if there is one, could be in the full-view, big picture analysis of all the hurricane data. That flaw may be this. In order to determine if there are cycles of hurricane activity doesn’t it have to be proven that there have been many cycles in the past? Like dozens of cycles perhaps?

From my admittedly somewhat small exposure to Statistics I remember one thing. The bigger the sample set of data the more accurate the result will be with the smallest margin of error. (Which is why pollsters need a thousand or more samples to come to a reasonably accurate conclusion)

As I recall, the minimum sample set is about 30, but even with that the margin of error is large. That class was years ago, I may be wrong, but I’m sure the minimum sample set must be larger than 2.

Looking at the graph of hurricanes per year there are certainly enough years to show that there has been an up - down - up curve in the number of hurricanes since about 1950. There was a peak in the 1950’s, a valley in the 70’s, 80’s and early 90’s, now we’re climbing the hill again.

The graphed data shows about a cycle and a half, but is that enough to say that there is and has always been a multi-decadal signal in hurricanes. You’d think you’d need more cycles - not yearly hurricane seasons - to come to that conclusion.

Using that minimum of 30 cycles, each with an average of say 50 years, scientists would have to look back 1500 years of Atlantic hurricanes to come to that conclusion. Of course they can’t.

True, some say global warming may be a cyclical thing too, but scientists have looked back hundreds if not thousands of years, as well as looked at a wide range of information, to come to that conclusion, not just 50-plus years of hurricane activity.

So should the theory of the multi-decadal signal be looked at again? In terms of statistical analysis, yes.

Visit the National Weather Service Climate Prediction Center NOAA: August 2005 Update to Atlantic Hurricane Season Outlook http://www.cpc.noaa.gov/products/outlooks/hurricane.html