Fifth Straight Year of Accurate Hurricane Predictions from William Gray and Colorado State Forecast Team

Note to Editors: The complete 2003 season verification report, as well as related research, detailed descriptions of forecast factors, previous forecasts and explanations of individual storms, are available on the Web at or

For the fifth consecutive year, William Gray and the Colorado State University hurricane forecast team were on target with their seasonal hurricane forecast, providing skillful predictions of named storms, named storm days, hurricanes, hurricane days and intense hurricanes in 2003 as highlighted in a summary report of 2003 Atlantic tropical cyclone activity released today. The report, available in its entirety on the Web at, also shows the team performed well with their newer August, September and October-only forecasts.

Following a suppressed 2002 hurricane season, Gray and his colleagues predicted above-average activity for 2003. Last December, the team’s long-range seasonal outlook for this year called for 12 named storms, eight hurricanes and three intense hurricanes. In their May 30 update, issued two days before the official start of hurricane season, the team adjusted their prediction of named storms to 14, leaving the forecast for hurricanes and intense hurricanes at eight and three respectively. As of Nov. 20, observed totals for the season were 14 named storms, seven hurricanes and three intense hurricanes.  

"We consider 2003 a successful forecast year with regard to most of our forecast categories," said Gray. "As the last five years indicate, we are making progress in better understanding and consequently improving seasonal prediction skill. With more research, this understanding will likely continue to improve, and we hope these forecasts will continue to be of assistance to coastal populations, emergency managers, insurance providers and others concerned about Atlantic basin hurricane activity."

Today’s report summarizes tropical cyclone activity in the Atlantic basin during the 2003 hurricane season and compares the team’s seasonal and monthly forecasts to what actually occurred. The Atlantic basin, including the Atlantic Ocean, Caribbean Sea and Gulf of Mexico, experiences more year-to-year hurricane variability than any other global cyclone basin.

Gray and his colleagues issued a long-term seasonal forecast in early December 2002, with updates on April 4, May 30, August 6, Sept. 3 and Oct. 2 of this year.

"We increased our later seasonal forecast updates a bit to take into account changing hurricane signals that indicated a slight increase in the potential number of named storms, named storm days and intense hurricane days," said Gray. "Overall, however, we stayed the course with most of our early seasonal predictions."

A new aspect of the Colorado State forecast team’s climate research is the development of tropical storm activity predictions for the individual months of August, September and October, which historically account for 91 percent of the total storm activity in the Atlantic basin. August-only forecasts have now been made for four seasons and September-only forecasts for two seasons. This is the first year the team issued an October-only forecast.

"There often are monthly periods within active and inactive hurricane seasons that do not conform to the overall season," said atmospheric research scientist and team member Philip Klotzbach. "To this end, we have developed new schemes to forecast individual monthly activity for August, September and October that have proven to be highly accurate in hindcast analysis of the past 50 years."

The team’s August-only forecast was nearly perfect and is credited to former project member, Eric Blake, now at the NOAA National Hurricane Center, who developed the scheme while a graduate student at Colorado State. It called for three named storms, one hurricane and one intense hurricane; three named storms, two hurricanes and one intense hurricane were observed.

The September-only forecast called for four named storms, two hurricanes and one intense hurricane; four named storms, three hurricanes and one intense hurricane were observed. The longevity of these storms was underestimated, however. The October-only forecast predicted three named storms, two hurricanes and no intense hurricanes; October observed three named storms, no hurricanes and one intense hurricane (because Kate became a hurricane in September but did not reach intense hurricane status until October).

"We are encouraged by our new individual monthly forecasts of August-only, September-only and October-only activity," said Gray.

According to the seasonal summary report, a major reconfiguration of the distribution of Atlantic thermohaline circulation began in 1995. This increase warmed North Atlantic sea surface temperatures and lowered surface pressure which enhanced Atlantic basin hurricane activity, a trend well associated with the large increase in major hurricane activity throughout seven of the last nine years. Despite the El-Nino-linked seasonal reductions in 1997 and 2002, the last nine years constitute the most active consecutive years on record for Atlantic basin tropical cyclone activity.

"These and other trends toward increased hurricane activity give strong support to the theory that we have entered a new era of enhanced major hurricane activity," said Klotzbach. "We expect this era to span the next two or three decades."

Today’s seasonal summary report also addresses the downturn in United States major hurricane landfall despite the large increase in Atlantic basin major hurricane frequency. The nine years of 1995-2003 have seen 32 major hurricanes to form in the Atlantic.

Throughout the last century, an average of about one in every three major Atlantic basin hurricanes made landfall. Based on this century-long average, nine or 10 intense hurricanes should have made landfall in the United States within the last nine years, but only three crossed the shoreline. The East Coast has been especially lucky to have experienced only one major hurricane landfall (Fran, 1996) since 1995. During the active hurricane period of 1944-1961, there were 16 major landfall events along the East Coast.

Gray explains that good fortune has been manifest during recent years in large part due to a persistent upper-air trough located along the U.S. East Coast that caused a large portion of otherwise northwest-moving major hurricanes to curve to the north before they reached the coastline. However, he warns that this good luck cannot be presumed to continue.

"The United States has been extremely lucky over the past nine years, but climatology will eventually right itself and we must expect a great increase in landfalling major hurricanes," said Gray. "With such large coastal population growth in recent decades, it is inevitable that we will see hurricane-spawned destruction in coming decades on a scale many times greater than what we have seen in the past."

William Gray and his Colorado State University team have provided seasonal Atlantic basin hurricane forecasts for the past 20 years. Until Gray began developing his forecast model in the early 1980s, there were no objective methods for predicting whether forthcoming hurricane seasons were likely to be active, inactive or near average. Ongoing research by Gray and his colleagues has since indicated that there are a number of atmospheric and oceanic precursor circulation signals that can be used for skillful Atlantic basin seasonal hurricane activity as early as December of the prior year. The team’s new individual monthly forecasts issued zero to two months in advance are also found to have substantial hindcast skill.  

Gray and his team are already working on their first seasonal forecast of the 2004 Atlantic basin hurricane season. This forecast report will be issued on Dec. 5, 2003 and will be available on the Web at In addition to Gray and Klotzbach, other team members include William Thorson, Jason Connor and other ex-project members presently working with NOAA.