by Mark A. Thornton - September 2005

Guess Who's Coming to Dinner
On the afternoon of August 13, 2004, the residents of Florida’s Gulf Coast were awaiting the arrival of a guest. The fact that this thirteenth day of August was a Friday did not improve the overall mood of the region – whether you were prone to superstitious feelings or not. The unwelcome guest was Hurricane Charley, who much earlier in the day had struck Cuba as a Category 2 hurricane on the Saffir-Simpson scale, and in the process had caused three deaths and destroyed 8,300 homes.

Charley survived its encounter with Cuba’s western tip and reentered the Gulf of Mexico in the wee hours of the 13th with maximum sustained winds of 90 knots. Charley spent the morning of the 13th as a strong Category 2 hurricane, charting a path for Florida’s Gulf Coast. Only a few short hours before landfall, Charley rapidly intensified to Category 4, made a moderate and somewhat unexpected adjustment to a more easterly course, and set its sights on the Port Charlotte, Florida area. Media focus on Tampa as the area of landfall had caused many of Port Charlotte’s residents to be caught unprepared. According to the US Census Bureau, 950,000 people were living within 30 miles of the cyclone’s projected landfall, nearly a third of who were over the age of 65.

Charley made landfall at Cayo Costa, a barrier island protecting the entrance to Charlotte Harbor, at approximately 3:45PM delivering maximum sustained winds near 130 knots and torrential rain. The hurricane reached Punta Gorda and neighboring Port Charlotte about an hour later, resulting in the destruction of one in seven businesses and rendering an estimated 11,000 homes uninhabitable. Mobile homes were particularly hard hit.

By midnight on the 13th, Charley had exited the Florida peninsula near Daytona Beach retaining its hurricane status, although significantly weakened to Category 1. Charley’s trip across Florida would prompt federal disaster declarations in 25 of the state’s 67 counties, and insured losses would approach $7 billion (USD). Charley ultimately dissipated near Cape Cod on August 15, but not before wreaking havoc along most of the East Coast.

Montage of visible images and path from CIMSS

Hurricane Charley, August 9 to 15, 2004 from CIMMS

Charley is an interesting hurricane for a number of reasons, in particular its largely unexpected intensification from a Category 2 to a Category 4 hurricane in only three hours. Although significant by any measure, the compact nature of Charley’s destruction was surprising.

The element of surprise associated with its last minute course change also was a factor. I am particularly curious about the influence that the media and non-governmental meteorologists had on the general public’s awareness and understanding of the NHC forecast advisories and discussions. Katrina is a much more recent example where the NHC had a good handle on the situation but the general public either wasn’t paying attention or was confused by what was being presented. Perhaps I will pursue this topic as an addition to my e-portfolio.

Charley’s landfall also had a family connection. Several family members live in Englewood, a community less than twenty miles north of Port Charlotte. Since their homes are only about 10 feet above sea level, they wisely chose to evacuate to northern Florida. If landfall had been only a few miles further north and the expected storm surge had been realized, it is likely that their homes would have been flooded. Fortunately, they returned a couple of days later and found that no damage had occurred. The fact that they had a larger understanding of Charley’s potential, and were not distracted by all of the focus on Tampa is comforting.

Finally, you just have to admire Charley’s timing. Who can’t appreciate a hurricane that makes landfall on Friday the 13th?

Birth of a Hurricane
Like most tropical cyclones, Charley began as an unimpressive area of low pressure off the coast of Western Africa on August 4th. The system’s rapid westerly transit of the Atlantic was accompanied by gradual improvements in organization and the development of cyclonic circulation. Satellite imagery and surface observations to the northeast of Venezuela on the morning of August 9th indicated that the system had grown into a tropical depression. By late afternoon on the 11th, the system had progressed through tropical storm status and had become Hurricane Charley. By the time it reached Cuba late in the day on the 12th, Charley was a Category 2 hurricane.

Charley ultimately slammed ashore at Cayo Costa at 3:45PM on the 13th as a solid Category 4 hurricane, packing estimated sustained winds near 130 knots. The automated weather observation site (ASOS) at Punta Gorda observed a gust of 97 knots at 4:36PM, just before it stopped reporting wind speed due to failure of the anemometer—a common occurrence with ASOS sites in the path of landfalling hurricanes. Considering that the barometric pressure continued to fall, subsequent surface wind speeds are thought to have been much greater.



Doppler Radar Image - NWS Tampa, Florida

The NHC Tropical Cyclone Report noted that Charley made landfall with a very small, well-defined eye. This statement is consistent with the observations made by a reconnaissance aircraft at 3:57PM of a circular eye with a diameter of five nautical miles (nm). A review of the reconnaissance observations in the 24-hour period prior to landfall shows the diameter tightened from 18nm to 5nm. The small eye size resulted in Charley’s maximum winds being restricted to an area approximately 10 miles wide centered on the path of the eyewall, sparing the region from more widespread destruction. The radar image from 4:47PM (2047Z) shows the overall structure of the storm and small eye.

Although well within the three-day cone (Advisory 14) issued by the NHC at 5PM on Thursday, August 12, 2004, most residents of the Punta Gorda/Port Charlotte area were of the opinion that Charley was headed to Tampa Bay. In fact, a review of subsequent three-day cones issued at 5AM (Advisory 16) and 11AM (Advisory 17), indicates that the projected track of Charley changed very little during this interval.

Perhaps the track was a little more easterly, but  clearly Tampa Bay was at the center of the cone. The media’s constant focus on Tampa Bay and St. Petersburg as Charley’s target is principally blamed for the overall lack of preparation by Punta Gorda and Port Charlotte residents. It had been quite some time since the last hurricane threat, and many residents weren’t familiar with how to interpret NHC’s three and five-day cones. They incorrectly focused on the center lines of these forecasting products rather than the larger warning area that was intended.

Charley’s more easterly path was identified and published on the three-day cone issued at 2PM (Advisory 18). The cause of this shift was a cold front that was stronger than the NWS expected. By then, of course, Charley was less than two hours from making landfall and only two and one-half hours from Punta Gorda and Port Charlotte. On the afternoon following landfall, the Associated Press reported NHC spokesman Robbie Berg as saying “We’re kind of surprised that people were caught by surprise”.

According to the NHC, the overwhelming majority of hurricane fatalities are caused by storm surge and flooding. Although the NHC special advisory issued at 2PM on the 13th expected “storm surge flooding of 10 to 15 feet near and south of where the center crosses the Florida west coast”, actual storm surge was on the order of 6 to 7 feet. Charley’s fast moving character is credited with this positive outcome. The hurricane was moving too rapidly to produce wave heights of sufficient magnitude to cause the anticipated storm surge.

An Intensifying Hurricane
The most intriguing aspect of Hurricane Charley is the profound intensification during the mid-day hours of the 13th. Let’s go behind the scenes and examine the data that forecasters were reviewing on this date.

Charley reached the Dry Tortugas at approximately 8:00AM on the 13th as a Category 2 hurricane, with maximum winds of 95 knots and a central sea level barometric pressure of approximately 970mb. Over the next three hours, it changed very little with surface winds and barometric pressure remaining fairly steady. The hours between 11:00AM and 2:00PM, however, would see Charley transform into a powerful Category 4 hurricane. Surface winds during the period increased from 95 to 126 knots as the barometric pressure fell to 954mb. In the two hours preceding landfall, Charley’s barometric pressure would fall another 13mb to 941mb. It would maintain its Category 4 status until landfall.

The NHC only hinted at the possibility of Charley strengthening in Discussion #17 issued at 11:00AM on the 13th. By 2:00PM the evidence was clear and the NHC issued Special Advisory #18 announcing that, based upon surface wind observations, Charley was a Category 4 hurricane and had assumed a more easterly track.

Observations and Analysis
Forecasters at the National Hurricane Center (NHC) had many tools at their disposal to monitor and dissect the structure, intensity, and path of Charley during its Friday the 13th journey to Florida’s southwestern coast. These include vortex messages from reconnaissance aircraft, satellite imagery, scatterometry and wind analyses.

Vortex Messages
One such tool is aircraft reconnaissance, a resource that came into common use during the 1950s. These “flying weather stations” have trained meteorologists and a wealth of sophisticated instruments on board that allow for both direct and inferred determination of storm features and characteristics as the aircraft passes through the center of a tropical cyclone. The meteorological information, such as the location of the hurricane, various temperature measurements, wind speeds, sea level pressure and the character and orientation of the eye, that is collected during a hurricane-hunter flight is transmitted real-time in short bursts of data called a vortex message. Once decoded, meteorologists at the NHC are able to compare this on-the-spot data with other, more remote forms of observation to help them produce their forecasts, advisories and associated warnings.

It is instructional to compare the data from the decoded vortex messages from 8:19AM and 3:30PM on August 13th. During the 8:19AM vortex message, Charley was in the vicinity of the Dry Tortugas. At the time of the 3:30PM mission, Charley was just 15 minutes from making landfall. The table below highlights Charley’s transformation during the period. In particular, note the dramatic increase in flight level wind speeds and the steep decline in sea level pressure.

Source: Vortex messages from the National Hurricane Center

As mentioned earlier, the diameter of the eye shrank by 50%, an indicator of a strengthening storm. Other sure signs of a strengthening storm are the decreases in 700mb heights and the continued increase of temperature in the eye. Meteorologists frequently use differences in temperature from within and outside the storm to gauge a storm's strength, in this case an eight degree difference existed at the time of landfall.

Satellite Imagery
Satellites provide an excellent opportunity to collect data that is either impossible or inconvenient to obtain by any other means. Almost everyone is familiar with satellite images of weather systems because they are a staple of most nightly weather forecasts. Meteorologists use water vapor, infrared and visible water satellite images to help determine and predict the structure and future development of tropical cyclones.

Visible image south of Cuba on the 12th from MODIS

Enhanced infrared image near landfall on the 13th from CIMSS

The image on the left shows the overall structure of Charley and the clearly discernible eye just below the center of the cloud mass. This image was captured very close to the same time as the scatterometer image displayed below. The enhanced infrared image on the right shows Charley just before landfall. The color scale displays cloud top temperatures which in this case provided clues as to the level of severity of the thunderstorms comprising Charley's eyewall.

Some weather satellites also carry an instrument that uses high-frequency radar to measure surface wind speed and direction. This extremely useful device, referred to as a scatterometer, “reads” the differences in the texture of the ocean’s surface and equates rougher surfaces to higher wind speeds. Scatterometer products, such as QuikScat, are capable of identifying wind patterns that are consistent with the early stages of tropical cyclones and, therefore, provide forecasters with the ability to identify developing weather systems sooner than any other method allows.

Unfortunately, scatterometer technology is not perfect. Because the instruments are deployed on polar-orbiting satellites, “blind spots” exist on progressive passes over the Earth’s surface. If a tropical cyclone’s location should happen to fall within one of these blind spots, no analysis will be available. Ironically, on Friday the 13th, Hurricane Charley spent its only day as a Category 4 hurricane in a QuikScat blind spot!

For an example of scatterometry, please review the QuikScat image from 6:47AM (1047Z) on August 12, 2004. At the time, Charley was located south of Cuba and just to the northwest of Jamaica. A review of the wind barbs suggests a center of circulation at approximately 79.9 degrees W longitude and 19.2 degrees N latitude and maximum surface winds of 50 to 55 knots (the red wind barbs) in the storm’s northeast quadrant. In the Northern Hemisphere the surface winds flow counter-clockwise around a tropical cyclone. The wind barbs containing a grey dot at the base should be approached somewhat skeptically due to the dampening effect of heavy rain – a recognized issue with scatterometry technology. This may explain why the QuikScat wind speeds are significantly below that which you would expect for a Category 2 hurricane.

QuikScat image 6:47am on the 12th from Remote Sensing Systems

Wind Analysis
Wind Analyses from the Hurricane Research Division (HRD), are a valuable source of surface wind information. In addition to a contour-like representation of wind speeds, these images provide the overall structure of wind velocity distribution associated with a tropical cyclone. Each analysis also contains an arrow pointing in the direction of the maximum wind. These HRD Wind Analyses are produced by combining a variety of surface wind observations, which are conveniently listed at the top of each analysis, and then adjusting the output to a standard surface wind height of 10 meters.

HRD Wind Analysis - 4:30am on August 12, 2004

HRD Wind Analysis - 3:30pm on August 13, 2004

The analysis above and to the left was created a little over three hours before the accompanying QuikScat analysis. Notice that the wind speeds and their distribution correlate reasonably well with the QuikScat image and suggest that heavy rain was indeed suppressing the wind speed values on the QuikScat analysis. At the time, the maximum winds were 70 knots from the east. The 3:30PM (1930Z) surface wind analysis (above and to the right) provides an excellent overview of the speed and extent of southwesterly hurricane force winds in the minutes just before landfall. The reduction of surface wind speeds are clearly evident when the wind analyses from 3:30PM (1930Z) and 4:32PM (2032Z) are examined.

Rapid Intensification of Hurricanes
The forecasting products discussed above, although only a few of those available, assisted the National Hurricane Center in observing and predicting Charley’s rapid development, but what forces where responsible for the storm’s rapid intensification in the hours before landfall?

Recent research suggests that the existence of deep warm-water pools in the Gulf of Mexico may act as “fuel injectors” to intensify hurricanes. These warm pools are regions of warm water that extend to much greater depths that what is typically found in the Gulf, perhaps as much as 100 meters or more. As a hurricane passes over warm water, it draws energy from that water but eventually the resultant mixing that occurs diminishes the amount of energy that remains available to the storm. A deep warm pool of water, however, is thought to provide a richer source of energy that could contribute to the development of a much more powerful storm.

Another contributing factor to rapid hurricane development may be the interaction of the outer rain bands and the hurricane’s eye. Scientists from the University of Miami, the University of Washington and the National Center for Atmospheric Research have recently announced a study that will investigate this interaction and its influence on a storm’s intensity. The project will use data harvested from Hurricane Hunter flights during the 2005 hurricane season.