Part One
This effort started out as an exercise to determine if the Atlantic Ocean indeed had a strong multi-decadal signal in its hurricane activity based on multi-decadal variations in the strength of the Atlantic Ocean thermohaline circulation (THC) or was the whole concept of this decades long pattern either nonexistent or was it something that could not be proven (at least not without conjecture, i.e. a lack of firm definition). However the effort eventually ended up with better insight into the likelihood of a very active season (or not) since the verification of a multi-decade pattern of increased hurricane activity must look to the number of named storms for confirmation that a pattern indeed exists.

If there is roughly a 25 year pattern of increased tropical cyclone activity in the Atlantic Basin, then the seasonal numbers should reflect that pattern. If the current strong signal started in 1995 and has continued through at least 2008 do the number of storms reflect that? The answer, of course, is Yes with 9 of the 15 seasons having produced 14 named storms or more. Does the climatological record easily define the previous strong signal? This time the answer is Probably Yes. The period seems to be 1932 to 1945 with 8 of the 14 seasons producing above normal activity but only two seasons with 14 or more named storms (1933 and 1936). CSU defines the period from 1950 to 1969 as â€˜activeâ€™, but only two seasons had 14 or more named storms in that 20-season period (1953 and 1969), and only 9 of the 20 seasons had above average activity â€“ which is nothing more than average activity over a 20 season period and not indicative of a strong THC. With 13 named systems in 1949, Iâ€™m at a loss as to why CSU didnâ€™t start their defined â€˜activeâ€™ period in 1949 vice 1950.

Does the climatological record hint at another 10 to 15 year period of increased activity that might suggest another earlier strong THC? This time the answer is No. Is there a well-defined multi-decadal pattern of increased hurricane activity in the historical record? Again the answer is No, although with a little imagination it might be easier to define a 25 year period of inactivity between shorter periods of activity. The current period of activity is not in question â€“ it exists, but it seems to be linked to something that cannot be proven. Some additional unanswered questions include: What defines the â€˜endâ€™ of an active period? How many years are needed to meet a multi-decadal definition? How many years in an active period must produce above-average storm totals in order to define â€˜activeâ€™? It is one thing to state that periods of increased Atlantic Basin storm activity are associated with a strong THC â€“ but it remains nothing more than a theory until these questions (and Iâ€™m sure there are others) are answered to the satisfaction of the scientific community as a whole. This is not a criticism â€“ rather more of a concern for additional definition.

Part Two
While plowing through the seasonal totals to develop my thoughts reflected above in Part One, it provided the data to examine the probabilities for active seasons â€“ something that NOAA, CSU, TSR, WSI and many of you believe will happen this season. Our annual early season outlook for expected storm totals generated 30 useable responses with an average expectation of 16 named storms, 9 hurricanes and 5 major hurricanes. The inputs ranged from a low of 9/5/2 to a high of 29/16/7.

Since 1885 there have been 8 seasons with 16 or more tropical storms or hurricanes â€“ or 6.4% of the 125 seasons. Those highly active seasons were 1887, 1933, 1936, 1969, 1995, 2003, 2005 and 2008. During these seasons, an average of 5 storms made landfall on the U.S. coastline (the range was from 3 to 7). Between June 1st and November 30th there are 26 weeks that define the hurricane season. Sometimes storms form earlier than June (but not this season) and, with less frequency, the season can sometimes extend into December. Using the 26-week season as a standard baseline, if you forecast a seasonal total of 26 storms you are forecasting the development of an average of one storm per week for the entire season and thus far, 2005 was the only season to achieve that level of ultra-high activity. Looking at the dates of the first and second storms of these high-activity seasons provides some insight into the odds for having another season with 16 or more named storms. If the first storm doesnâ€™t develop by mid June and the second storm doesnâ€™t develop by the first week in July, the chances of having a season with 16 or more named storms is one in 125 or 0.8% (eight tenths of one percent).

Since 70% of the responses predicted at least 14 named storms, I also reviewed the climatological record for that level of moderately high activity. Since 1885 there were 8 additional seasons with at least 14 or 15 tropical storms or hurricanes so a season with 14 or more storms has occurred during 16 of the 125 seasons â€“ or 12.8%. These moderately active seasons with 14 or 15 storm totals were 1916, 1953, 1990, 1998, 2000, 2001, 2004 and 2007. These seasons also had an average of 5 storms making U.S. landfall (the range was from 1 to 10). Once again looking at the dates of the first and second storms for these additional moderately active seasons suggested that if the first storm does not develop by the last week in July and the second storm does not develop by the second week in August, the chances of having a season with 14 or 15 named storms is the same, i.e., one in 125 or 0.8%.

Of course any new season can change climatology and the statistical record, but these dates are at least a good starting point for you to determine the likelihood of success for your own personal seasonal forecast.
Good Luck,
ED