Hugh Willoughby, International Hurricane Research Center
Florida International University, Miami Florida

As the first cold fronts of autumn push southward, 2004 takes its place in the hurricane record books.  Three hurricanes — Charley, Frances, and Jeanne — struck Florida directly this year.  A fourth, Ivan, whose center crossed the coastline over Gulf Shores Mississippi, caused devastation in Florida’s western Panhandle.  The last time three hurricanes hit the state was 1960, and before that 1896.  Half a world away, Typhoon [i] Tokage, the deadliest Japanese tropical cyclone in a decade, bashed Honshu near Tokyo in late October.  A total of 10 tropical cyclones pounded Japan during 2004, tying the record number of landfalls on that nation in a single year.

Pine Island, Florida. Boat storage facility literally shredded apart by Hurricane Charley's winds.

Worldwide, we see the effects of climatic warming in sea-level rise, retreat of icecaps and glaciers, poleward migration of tropical species, and the stark alternation of deluge and drought.[ii]  Should we recognize the tropical cyclones of 2004 as heralds of more frequent, more intense, more costly, and more deadly cyclones on a warmer globe?  Or do they reflect natural cycles, perhaps combined with simple bad luck?

On the natural cycle side, meteorologists working at the National Oceanic and Atmospheric Administration (NOAA) and Colorado State University[iii] have shown that the number and intensities of hurricanes follow a 50-70 year cycle.  This cycle, called the Atlantic Multidecadal Oscil­lation (AMO), is controlled by gradual changes in the North Atlantic Ocean currents.  The “thermohaline circulation,” as it is called, moves warm water northward in all latitudes of the Atlantic Ocean, even south of the equator.  When seawater in high latitudes (near Greenland and Iceland) is warm and salty, the weight of the extra salt allows it to sink easily and the thermohaline circulation runs quickly.  Warm water moves northward freely to replace the foundering surface water.  When seawater in high latitudes is relatively fresh, it has to be colder in order to sink, and the circulation is more leisurely northward flow of not quite so warm water.  Rainfall and evaporation throughout the world’s oceans set the tempo of the AMO.

Bokeelia, Florida. Manufactured homes were no match for Hurricane Charley.

Through a complicated chain of cause and effect, the faster oceanic circulation causes the mid-latitude westerlies (winds blowing from the west) to stay north of the tropical Atlantic.  Then, deep tropical Trade Winds, which blow steadily from the east, produce conditions favorable for hurricane formation.  That's the phase of the AMO that we are in now.  When the thermohaline circulation is weaker, the west winds bend farther southward above the Trades at altitudes just below the stratosphere.  This situation causes increased vertical wind shear that suppresses hurricane activity.  That is the AMO phase we were in during the inactive 1970s through early 1990s. 

View from second-story window of house shown below in Pine Island

There was a previous active phase of the AMO from the late 1920s through 1969.  The current active AMO phase began in 1995.  Even counting 1997 and 2002, when El NiZo kept hurricane formation in check, the years since 1995 have been the most active on record in terms of number and intensity of hurricanes.  By these measures, 2004 has been typical of the non-El Nino years since 1995.

 The difference this year is the greater fraction of hurricanes that hit Florida.  During first nine years of the active AMO phase from 1995-2003, and also during the inactive phase from 1970-1994, very few of the hurricanes that formed made landfall anywhere in North America.  A simple return to the long-term average ratio of landfalls to hurricanes would be a big change from experience during the last third of a century. 

Cycle of frequency and intensity of tropical cyclones in
the North Atlantic

During the time when so few hurricanes hit North America, we as a society framed decisions about land use, construction standards, and other aspects of our lives around the shores of the Atlantic Ocean and Gulf of Mexico.  Built into these plans was the unstated assumption that hurricanes would continue to stay away from our shores as they had for the last third of a century.

Nonetheless, we can argue that the present active AMO phase may be a bit more active than the last one.  Despite the lack of US landfalls until this year, there have been more hurricanes, more intense hurricanes that can cause the worst destruction, and more days when hurricanes were present.  Is the apparent change from the 1925-1969 active phase the result of more comprehensive observations at the turn of the 21st century or is it real?

To address that question we need to understand the basics of topical cyclones. The pressure in the eye of a hurricane is low because the vortex center is filled with warm air.  Since warm air is less dense than cold air, the pressure, which is simply the weight of air above the surface, is lower near the center than it is outside.  Hurricane winds circulating around the center grow stronger with greater pressure difference.  Thus, the difference in temperature between inside and outside is what deter­mines the strength of the hurricane.

The sea supplies the heat required to keep the air column warm.  By late summer the sun has warmed the tropical ocean to the temperature of bath water (27-29o C).  Circulating storm winds promote evaporation from the sea as they spiral inward toward the hurricane center.  When the vapor is drawn into the clouds around the eye, it condenses to make rain.  Condensation turns the stored heat into a temperature increase, lowering the surface pressure, and strengthening the wind. 

For every hurricane there is a “Maximum Potential Intensity (MPI)” that depends upon the temperatures in the air around it and in the sea under it.  Most hurricanes, most of the time, are weaker than the MPI calculated from conditions before the storm— either because of storm-caused cooling of the sea or because of wind shear (the difference between the surface wind around the storm and the wind above 10 km altitude).  With the notable exceptions of Andrew and Hugo, the two costliest recent landfalls, and Charley in August 2004, hurricanes that have struck the U.S. since 1970 were weaker than their MPI and weakening as they came onshore, predominantly because of wind shear.

Pine Island, Forida. Two-story home near Hurricane Charley's landfall point with obvious destruction.

In global warming, we might suppose that the effect of a warmer surrounding atmosphere and a warmer ocean would result in the same temperature difference as before, so that MPI would be unchanged.  But the amount of vapor that air can hold increases rapidly with temperature.  A 3o C warming is equivalent to a 1% increase in absolute temperature, but causes a 20% increase in the energy that air can store in water vapor at 100% humidity.  This large increase in energy for a small increase in temperature in the tropics is one reason why global warming isn’t global at all, but concentrated in high latitudes generally during the cold season (See companion article by Mike MacCracken).

The New York Times recently highlighted studies[iv] by Tom Knutson and Bob Tuleya, well-respected climate and hurricane researchers at NOAA’s Geophysical Fluid Dynamics Laboratory.  This research provides evidence that the most intense hurricanes may indeed become more intense on a warmer globe.  Knutson and Tuleya applied ocean and air temperatures calculated for doubled CO2 by several Global Circulation Models as the storm environment in the computer model used by the National Weather Service to forecast hurricanes. 

Turner Civic Center, Arcadia, Florida. Located 25 miles inland, this building was nearly flattened.

They found that there was little reason to expect more hurricanes or hurricanes in different places, but the intensities of the few hurricanes that get close to their MPI increased as a result of the thermodynamics described above.  The amount of the increase was only 7-16 miles per hour, or about half a Saffir-Simpson[v] category.  In addition to the stronger winds, hurricanes on the warmed globe rained 28% more on average. CO2 doubling will not happen until late in the 21st Century. Nonetheless, examples of what to expect then are Ivan and Isabel of 2003.  Both were extreme category 5 hurricanes at maximum intensity, but weakened for days before U.S. landfall.  An even more threatening harbinger was Hurricane Jeanne’s torrential rainfall that drowned thousands of men, women, and children in Haiti. 

For now, the key questions are: Do Charley, Frances, Ivan, and Jeanne, which made US landfall as garden-variety hurricanes, presage more frequent landfalls in the context of a not-too-different climate or will the incredibly good luck of the late 20th century hold for the foreseeable future?  The consensus among hurricane researchers and forecasters is that the hurricane landfalls of 2004 resulted from the AMO, a natural cycle of hurricane activity, combined with a lapse in the incredibly good fortune of the previous 35 years.  The effect of global warming was at most second-order, and probably not present at all.

Notes: