Musings: 2009 Hurricane Forecast Dampened by Cooling Waters

The team of Professors Philip Klotzbach and William Gray of the Department of Atmospheric Science at Colorado State University (CSU) released their first revision to their initial forecast for the 2009 hurricane season. They now see this year as an average season, down from their prior assessment of it being an active season. The new forecast calls for 12 named storms, down from 14 in their December 10, 2008, initial forecast. These storms will produce six hurricanes and two intense ones, each category lower by one. While the number of named storms is above the 50-year average spanning 1950-2000, the number of hurricanes and intense hurricanes is in line with the historical averages for that period.


While the reduction in the number of tropical storms and hurricanes does not appear overly dramatic, there is a marked reduction in the number of storm days. What is most notable about the new forecast is how much lower it is than the storm record of the past two years. In fact, the latest 2009 forecast would compare more closely with the 2006 storm season than almost any other year in the past six years.

The new April forecast was prepared based on the professors' new projection methodology that was used for the first time last year and proved quite good at forecasting storm activity for 2008. Without trying to explain the exact methodology, which is based on late-winter predictors and an early December hindcast, the forecasters show that their new forecasting methodology has significantly improved the hindcast record. The new methodology has correctly predicted above- or below-average tropical storm seasons in 45 out of 58 hindcast years for a 78% accuracy rate. The predictions have had a smaller error than climatology in 37 out of 58 years for a 64% performance rating. The average hindcast error is 26 NTC (net tropical cyclone) units compared with 44 NTC for climatology. The new methodology also showed improved stability when the study time period was broken in half with it explaining 59% of the variance from 1950-1978 and 72% of the variance from 1979-2007. Maybe equally as important is the new methodology correlated with observations at 0.80 for the years from 1995-2007 and at 0.85 for the years from 2002-2007.

Another step involved in preparing the forecast is to examine past hurricane seasons for ones that had weather factors similar to those that exist at the present time or have demonstrated similar patterns in the months leading up to the forecast time. The forecasting team identified five years - 1951, 1968, 1976, 1985 and 2001 - as being appropriate analog years for 2009. The CSU forecasters say that they anticipate 2009's tropical storm season will have activity in line with what was experienced in the average of these five years.


While the public's curiosity in knowing what the upcoming hurricane season will be like is one reason for the preparation of the forecast, the ultimate goal is to try to develop a system that can better forecast where storms will make landfall to help in storm preparedness. The effort at developing this system has been ongoing for a handful of years, and this year it will be expanded from the U.S. East and Gulf coasts to several of the islands in the Caribbean, but it is still a very inexact science. The highlights of the projections call for marginally higher than average landfall percentage chances for the U.S. coasts and average chances in the Caribbean.

The probabilities of at least one major (Category 3,4 or 5) hurricane making landfall on each of the following coastal areas is:

  1. Entire U.S. coastline - 54% (average for last century is 52%)
  2. U.S. East Coast including Florida - 32% (average for last century is 31%)
  3. Gulf Coast from Florida westward to Brownsville - 31% (average for last century is 30%)
  4. Average major hurricane landfall risk in the Caribbean.


If the accuracy of tropical storm landfalls could be improved, it would prove to be of great value for the people at risk, local governments, first responders and insurance companies. We will be interested to see the new Caribbean island forecasts that will be unveiled in the CSU tropical storm forecast update to be issued in early June.

The four major storms that landed on the U.S coast in 2005 (Dennis, Katrina, Rita and Wilma) and caused extensive destruction, coupled with the four storms that hit the Southeast in 2004 (Charley, Frances, Ivan and Jeanne) raised questions about the possible role that global warming played in these two destructive hurricane seasons. The three Category 2 hurricanes (Dolly, Gustav and Ike) that hammered the Gulf Coast last year causing extensive damage further fueled the global warming concerns.

The global warming arguments have been given considerable attention among the media who reference recent scientific papers claiming to show the linkage. Despite the global warming of the sea surface that has taken place over the last three decades, the global numbers of hurricanes and their intensity have not shown increases in recent years except for the Atlantic basin. It is significant that while global surface temperatures have increased over the last century and over the last 30 years, there is no reliable data to indicate increased hurricane frequency or intensity in any of the earth's other tropical cyclone basins.

In the Atlantic basin, there has been a very large increase in major hurricanes during the 14-year period of 1995-2008 (average 3.9 per year) compared to the prior 25-year period of 1970-1994 (average 1.5 per year). This increase appears to be primarily due to the multi-decadal increase in the Atlantic Ocean thermohaline circulation (THC) that is not directly related to global sea surface temperatures or CO2 increases. The THC changes seem to be primarily driven by changes in the ocean's salinity.

As the hurricane forecasting team points out, in a global warming or global cooling world, the atmosphere's upper air temperatures will warm or cool in unison with the sea surface temperatures. This does not appear to bear any relationship to the number or intensity of hurricanes. For the period 1945-1969 when the world was experiencing a weak cooling trend, the Atlantic basin experienced 80 major (Cat 3-4-5) hurricanes and 201 major hurricane days. In contrast, during 1970-1994 when the world was undergoing a general warming trend, there were only 38 major hurricanes (48% as many as in the earlier period) and 63 major hurricane days (31% as many). CO2 amounts in the later period were approximately 18% higher than in the earlier period.


Another database the CSU forecasters compared activity and global warming against is the measurements of U.S. landfalling tropical cyclones since 1899. Global mean ocean and Atlantic sea surface temperatures have increased by about 0.4°C between the two 55-year periods studied - 1899-1953 and 1954-2008 - yet the frequency of U.S. landfall numbers has shown a slight downward trend for the later period.


The downward trend is particularly noticeable for the U.S. East Coast and Florida where the difference in landfalls of major Category 3,4 and 5 hurricanes between the 43-year periods of 1923-1965 and 1966-2008 of 24 and seven is particularly large. For the entire U.S. coastline there were 38 major hurricanes that made landfall in the 1923-1965 period compared to only 26 in 1966-2008 despite the fact that CO2 averaged approximately 365 parts per million (ppm) during the latter period and only 310 ppm in the earlier one. The chart in Exhibit 15 is of personal interest as we grew up in the region of Southern New England visited by several of the 1954 and 1955 hurricanes.


The catalyst for the global warming concern is the record year for storms in 2005 and the spectacular destruction of New Orleans when the levees failed. The 28 storms that year, however, are not necessarily an indication of something beyond natural processes. In 1933 there were 21 named storms when there was no satellite or aircraft surveillance data. The records show that all 21 named storms had tracks west of 60°W where surface observations were more plentiful. If all the 2005 named storms whose tracks were entirely east of 60°W and therefore may have been missed by the 1933 technology were eliminated, it would reduce the total named storms that year by seven, bringing the year's adjusted total to 21, the same number of storms as observed in 1933.

The CSU forecasters also point out that when compared to other hurricane seasons contained in the National Hurricanes Center database extending back to 1875, there were six previous seasons with more hurricanes than 2005. Those years were 1878, 1893, 1926, 1933, 1950 and 1995. There were also five prior seasons with more major hurricane days (1893, 1926, 1950, 1961 and 2004). So while 2005 was a very active hurricane year, it is not as much of an outlier as many media and scientists have proclaimed.

The damage caused by the increase in hurricanes making landfall is also not a surprising fact. With so much of the nation's population located along the East Coast and Gulf Coast regions, any tropical storm will cause significant damage. This possibility has been forecasted as far back as 1989. The CSU forecasting team pointed out that the country was fortunate that during the early part of the current strong THC period, only 3 of 32 major hurricanes that formed in the Atlantic basin between 1995 and 2003 made U.S. landfall. The long-term average is that approximately 1 in 3.5 major hurricanes that form in the Atlantic basin makes U.S. landfall. This luck failed to hold beginning with the 2004 hurricane season.

It is encouraging that maybe 2009 will have a hurricane season without note. Of course, the success of the United States in avoiding a devastative hurricane landfall depends on the number of storms that form, their intensity and their track. While the U.S. might avoid storms, they could wreck havoc on our neighbors to the south. The domestic natural gas industry might welcome a hurricane this season that did selective physical damage to offshore gas producing facilities, but that would be too good to be true. Let's hope that when the CSU professors update their 2009 forecast next, it is tamer. We deserve a year with minimal natural catastrophes after the economic trauma of the past nine months.

Parks Paton Hoepel & Brown
  Reprinted with permission from PPH & B