Unusual 2002 Atlantic Hurricane Season

Meteorologists Jill F. Hasling and Dr. John C. Freeman of Weather Research Center [WRC] makes an annual hurricane forecast each year which is based on the Orbital Cyclone Strike Index [OCSI]. The OCSI was developed to predict which section of the United States has the highest risk of experiencing a landfall of a tropical storm or hurricane. In November of 2001, meteorologist Jill F. Hasling gave the Center's prediction that the West Coast of Florida had the highest risk with an 80% chance of experiencing a tropical storm or hurricane. She went on to state that the second highest risk coasts were Texas to Alabama coast, each with a 60% chance. Also that the Gulf of Mexico oil leases had an 81% chance of experiencing a tropical storm or hurricane this year. With five storms (Bertha, Fay, Hanna, Isidore and Lili) over the Gulf of Mexico leases this year, this forecast verified. Hanna verified the forecast of the west coast of Florida and Bertha, Fay, Isidore and Lili verified for the second highest probability forecasts of Texas to Alabama.

Mexico 20% 39%
Texas 60% 49% Fay
Louisiana to Alabama 60% 59% Bertha,Hanna, Isidore, Lili
West Florida 80% 70% Hanna
East Florida 40% 39% Edouard
Georgia to N. Carolina 30% 54% Kyle
East Coast of US 10% 32%
Gulf oil Blocks 81% 78% Bertha, Fay, Hanna, Isidore, Lili

Interesting events of the 2002 Atlantic Hurricane season were:

  • 7 named storms made landfall on the US Coast [with Gustav as a near miss]
  • 4 storms made landfall from Louisiana to Alabama
  • 9 storms occurred during September
  • Kyle lasted for 23 days.

Kyle was the third longest storm since 1871 behind Ginger in 1971 which lasted 31 days and Igna in 1969 which lasted 25 days. Carrie in 1957 lasted 23 days as well. Since 1871 there have been eight storms which lasted 20 days more.

  • 1893 – September 25 – October 15 21 days
  • 1899 – Augusts 3 –24 22 days
  • 1900 - August 27- September 15 20 days
  • 1915 - August 2 –23 22 days
  • 1957 - September 2 – 24 [Carrie] 23 days
  • 1966 - September 21 – October 11 [Inez] 21 days
  • 1969 - September 20 – October 14 [Igna] 25 days,/li>
  • 1971 - September 4 – October 5 [Ginger] 31 days

The secondary prediction from the Orbital Cyclone Strike Index [OCSI] model is the prediction of the number of named storms expected in the Atlantic and Gulf of Mexico. This forecast did not do very well this year. However meteorologists at the Center question if all of the named storms were truly tropical in nature. The questionable storms were Cristobal, Gustav, Josephine and Kyle. If these storms were sub-tropical in nature then there would have only been 8 tropical named storms in the Atlantic and the forecast would have been more realistic. The OCSI forecast is based on tropical system and not the naming of subtropical systems. The outlook, released in November 2001, called for a below normal season for the number of storms. The prediction was for 6 named systems with 3 of these storms intensifying into hurricanes. On average there are 9.7 named tropical storms per year with 5.7 of the storms intensifying into hurricanes. This year there were 12 named storms with 4 of them intensified into hurricanes. Two of these hurricanes became very strong hurricanes.

The OCSI predicted a 50% chance that one of the 2002 hurricanes predicted would intensify into a strong hurricane of category three intensity or higher. This verified with Hurricane Isidore which had maximum sustained winds of 110 knots and Hurricane Lili which had maximum sustained winds of 125 knots. Lili was a Category 4 Hurricane on the Saffir/Simpson Scale as it moved over the Eugene Island Gulf of Mexico oil lease. Fortunately, Lili weakened to a Category 2 hurricane prior to making landfall along the Louisiana Coast.

The OCSI predicted a 70% chance that at least three of the named storms would make landfall somewhere along the United States coast. Seven named storms made landfall along the United States Coast. Bertha, Edouard, Fay, Hanna, Isidore, Kyle and Lili all impacted the United States Coast. Gustav was a near miss to the North Carolina coast and if it had made landfall would have been the eighth storm along this US coast this year. There have been ten other years since 1871 when seven or more storms made landfall along the United States coast. 1879 [7], 1886 [7], 1893 [7], 1916 [9], 1933 [7], 1936 [7], 1947 [7], 1953 [7], 1959 [7], and 1985 [ 8].

Cuba had a 40% chance of a storm making landfall along those coasts according to OCSI. Isidore and Lili both made landfall on Cuba's western tip.

The OCSI predicted a 77% chance of an August Storm and 100% chance for September and October storms. These predictions verified with the storms listed below.

This year there were nine named tropical storms in the Atlantic basin in September.

Tropical Storm Dolly August 29- September 4
Tropical Storm Edouard September 1- 6
Tropical Storm Fay September 5-7
Hurricane Gustav September 8-12
Tropical Storm Hanna September 11-14
Hurricane Isidore September 14-26
Tropical Storm Josephine September 17-19
Hurricane Kyle September 20 – October 12
Hurricane Lili September 21 – October 4

There have been nine other years with 9 or more named storms or hurricanes in the Atlantic during the month of September. They were 1889 [12], 1932 [9], 1933 [11], 1949 [11], 1950 [12], 1953 [12], 1961 [10], 1969 [10] and 1984[12].

While the 2002 hurricane season might have started out slow with only 3 storms, Arthur, Bertha and Cristobal, things changed on August 29th. First, Dolly formed, followed by Eduoard, Fay, Gustav, Hanna, Isidore, Josphine, Kyle, and Lili. The tropics continued to be active with Tropical Depression Thirteen close to being a storm and a vigorous tropical wave in the far Atlantic.

In the past eleven years where the phase of the OSCI was the same as this year, eight of the years had six or less named storms. Six of the eleven years were El Niño Years. This statement has verified with the El Niño establishing itself this summer. Dr. John Freeman's long range forecast, released in 1998, called for a 40% chance of an ENSO event year in 2002. The Orbital Cyclone Strike Index [OCSI] model, developed at Weather Research Center by Dr. Freeman and Ms. Hasling, is based on the premise that there are orbital forces that influence the global circulation pattern on the sun and subsequently, the global circulation pattern of the earth. The sun's circulation pattern is tracked by the sun spot cycle. Using this solar cycle to make an index, hurricane climatology has been summarized into an index called the OCSI. This index has been used by WRC since 1985 to make annual forecasts of the section of North American coast that has the highest risk of experiencing a tropical storm or hurricane. Meteorologists at the Center can make their hurricane outlook for the next ten years when the year of the sunspot minimum is determined.

Weather Research Center meteorologists never adjust their hurricane forecast during the hurricane season. In November of 2001, Jill Hasling gave the Center's prediction of 6 named storms, with 3 becoming hurricanes. Once the forecast is issued in January or February the forecast does not change. Unlike their colleagues, their original forecast called for a below average year.

When you compare WRC's OCSI forecast of the number of tropical cyclones with Dr. William Gray's earliest prediction each year of the number of tropical cyclones with the number cyclones observed you find that WRC has come within +/-2 storms more or less ten times and Dr. Gray has come within +/-3 storms more or less five times. Dr. Gray adjusts his forecast up and down several times during the season. For example this year, Dr. Gray predicted 11 named storms in the Atlantic Basin in May and updated his forecasts to 8 named storms at the end of August. WRC never changes their forecasts of the number of storms. See the table and graphs below:

Weather Research Center versus Dr. William Grays
Tropical Cyclone Outlooks 1985-2001
1986 -5 2 Gray
1987 0 1 WRC
1988 -4 -1 Gray
1989 -2 -4 WRC
1990 -5 -3 >2
1991 0 0 TIE
1992 0 2 WRC
1993 -1 6 WRC
1994 0 2 WRC
1995 -9 -9 >2
1996 -2 -3 WRC
1997 0 4 WRC
1998 -5 -4 >2
1999 -2 2 Tie
2000 -6 -3 >2
2001 -6 -3 >2

In 2002, WRC was 6 storms below and Dr. Gray was 1 storm below. WRC forecast 6 named storms and Dr. Gray forecast 11 named storms. Positive numbers indicates the number of storms over-forecast each year. Negative numbers indicate the number of storms under forecast. If you use +/-2 then WRC was the closest 7 out of the 18 years with 3 years in a tie which makes a total of 10 years. Dr. William Gray was closest 3 years out of the 18 years with three ties making a total of 6 years. This is based on Dr. Gray's April or May forecasts since 1995 and his earliest forecasts for years prior to 1995. There were 5 years when both forecast methods had errors greater or less than 2 storms.

A peek at the 2003 forecasts shows that the OCSI is looking for storms along the east coast of the United States.

Louisiana to Alabama59%40%
West Florida50%70%
East Florida30%39%
Georgia to N. Carolina60%54%
East Coast of US60%32%
Gulf oil Blocks72%78%

The 2003 outlook calls for 7 named storms with 5 becoming hurricane.

In addition to its ongoing research, the Center also provides storm and hurricane information via the Internet through Storm Navigator. This service helps people navigate weather information on the Internet as well as providing detailed storm updates and related information. All of the Center's projections including past predictions can be found on the Internet, http://www.wxresearch.com/outlook.

Backgroun on Researchers: Jill F. Hasling and Dr. John C. Freeman are both Certified Consulting Meteorologist and Fellows of the American Meteorological Society. Both are researchers and founding directors of the Weather Research Center, which is a non-profit educational and research facility, based in Houston, Texas. Dr. Freeman has over 50 years of meteorological experience and Ms. Hasling has over 28 years experience. Both have been involved in tropical meteorology research and have been working together on this forecast method since 1985.