A hurricane, which is a Caribbean Indian word for "evil spirit and big wind", is a large rotating system of oceanic tropical origin with sustained surface winds of at least 74 mph somewhere in the storm. Due to the earth’s rotation, these storms spin counterclockwise in the Northern Hemisphere, and clockwise in the Southern Hemisphere; both types of hemispheric spins are referred to as cyclonic rotation. These storms occur worldwide, and are called different names in different locations (Holland 1993). In the Northwest Pacific across 180 degrees E, they are called typhoons. In the Philippines, they are called chubasco. Around Australia, they are called severe tropical cyclones, while India calls them severe cyclonic storms. In deference to United States readers, storms in this 74-mph or faster category will be called hurricanes in this book.

A hurricane does not form instantaneously, but reaches this status in an incremental process. Initially such a tropical system begins as a tropical disturbance when a mass of organized, oceanic thunderstorms persists for 24 hours (NHC 1998). Sometimes partial rotation is observed, but this is not required for a system to be designated a tropical disturbance. The tropical disturbance becomes classified as a tropical depression when a closed circulation is first observed and sustained winds are less than 39 mph everywhere in the storm. When these sustained winds increase to 39 mph somewhere in the storm, it is then classified as a tropical storm and given a name.

It is important to note that these categories are defined by sustained winds, not instantaneous winds. Sustained winds are the average speed over a period of time at roughly 30 feet above the ground. In the Atlantic and Northern Pacific Oceans, this averaging is performed over a 1-minute period, and in other ocean basins over a 10-minute period (Holland 1993). The actual wind will be faster or slower than the sustained wind at any instantaneous period of time. Therefore, a hurricane with maximum sustained winds of 90 mph will actually contain gusts of 100 mph or more. Also, one should note that these categories are defined by maximum winds somewhere in the storm, almost always near the center, and that winds may be slower in other parts of the storm. For example, maximum sustained winds of 90 mph may only be concentrated in the northeast section near the hurricane center, with the southwest quadrant containing weaker winds.

By international agreement, the most general term for all large cyclonically rotating thunderstorm complexes over tropical oceans is tropical cyclones. The category tropical cyclones includes depressions, tropical storms, and hurricanes in addition to other large, tropical, cyclonically rotating thunderstorm complexes which contain distinctly different temperature and organization characteristics. Examples include monsoon depressions which bring moderate winds and heavy rain to India, and subtropical cyclones which produce heavy rains and moderate winds in the Atlantic and Pacific Oceans. This book focuses on depressions, tropical storms, and hurricanes, and the term "tropical cyclone" will be used when generalizations regarding all stages of these storms are implied. The details of the development process from tropical disturbance to hurricane will be discussed in a moment, but the seasonal location of these storms will be examined first.

Many United States residents perceive the North Atlantic Ocean basin as a proliferate producer of hurricanes due to the publicity these storms generate. In reality, the North Atlantic is generally a marginal basin in terms of hurricane activity. Every tropical ocean except the South Atlantic and Southeast Pacific contains hurricanes, several of which produce more hurricanes annually than the North Atlantic. For example, the most active ocean basin in the world --- the Northwest Pacific --- averages 17 hurricanes per year. The second most active is the eastern North Pacific, which averages 10 hurricanes. In contrast, the North Atlantic mean annual number of hurricanes is 6. Table 1 summarizes each basin’s average number of hurricanes and total storms (Landsea 1998a).

Hurricanes are generally a summer phenomenon, but the length of the hurricane season varies in each basin, as does the peak of activity (Table 2). For example, the Atlantic hurricane season officially starts on June 1 and ends November 30, but most tropical storms and hurricanes form between August 15 to October 15. The Atlantic hurricane season "peaks" around September 10. In contrast, hurricanes occur year-round in the western North Pacific, with a longer active regime lasting from July to November peaking around September 1. In general, this pattern exists for other basins (with a late summer peak), although Southern Hemisphere hurricanes are 6 months out of phase. The exception is the North Indian basin, where there are two peaks in May and November (the reason for this will be discussed later).

Hurricane formation occurs in two distinct phases. The first phase is called the genesis stage, and includes tropical disturbances and tropical depressions. The second phase includes the tropical storms and hurricanes, and is called the intensification stage. These phases are separated because many disturbances and depressions never reach tropical storm intensity, and eventually dissipate.

Tropical disturbances form in regions where surface air "piles up," known as convergence. Where convergence occurs, the atmosphere will respond by causing air to rise or sink. Since air cannot penetrate a solid surface such as water or land, it must ascend to balance this accumulation of air. As air rises, it will saturate and form the base of a cloud. Once the air is saturated, ascent is enhanced because in the warm, moist tropics the atmosphere is in a state of instability. In an unstable atmosphere, saturated air forced upwards by convergence is less dense than surrounding unsaturated air and accelerates upwards, forming towering puffy clouds. The concepts of cloud formation and atmospheric instability are beyond the scope of this book, and the reader is referred to other meteorology books on the subject (Ahrens 1994; Nese and Grenci 1998; Danielson, Levin, and Abrams 1998).

Several conditions must simultaneously exist for a tropical disturbance to reach the tropical storm stage. First, the disturbance must be in a trough, defined as an elongated area of low pressure. (Pressure is the "weight" of the air above a given area of the earth’s surface; its standard unit of measurement is the millibar, or mb). Furthermore, these troughs must contain a weak, partial cyclonic rotation; however, in general all troughs at least 10 degrees from the equator will obtain a partial cyclonic spin due to the earth’s rotation. Troughs 10 degrees or further from the equator fall in three general categories: monsoon troughs, frontal troughs, and surface troughs.

Monsoon troughs occur in regions where air or water temperature actually increases away from the equator. In these regions, the Intertropical Convergence Zone (which is a region where Southern Hemisphere and Northern Hemisphere air converges near the equator) is displaced 10-20 degrees poleward, resulting in westerly winds on the equatorward side and easterly winds on the poleward side. A frontal trough is the remnant of a front (defined as the boundary between two air masses of different temperature and/or moisture characteristics) which has lost its contrasting temperature characteristic and entered the tropics. A surface trough encompasses all other kinds of troughs, such as those associated with a region of moisture contrast or broad thunderstorm complexes. Some surface troughs are even triggered by weather features 40,000 ft aloft.

The vast majority of genesis cases are associated with the monsoon trough. Many of the disturbances undergo the transition to tropical depression and then to tropical storm in the monsoon trough itself. However, a few experience this transition as tropical waves that have broken off from the monsoon trough and traveled westward a considerable distance before strengthening. In particular, most Atlantic hurricanes actually originate from tropical waves (more frequently called easterly waves in the United States) which broke off from the African monsoon trough and have propagated into the Atlantic! In the Atlantic about 55-75 tropical waves are observed, but only 10-25% of these develop into a tropical depression or beyond. Clearly additional factors are required for genesis to occur.

The second condition required for genesis is a water temperature of at least 80F. Heat transferred from the ocean to the air generates and sustains the thunderstorms in the disturbance through the instability mechanism discussed earlier. The third genesis condition is weak vertical wind shear, defined as the difference between wind speed and direction at 40,000 feet and the surface. In other words, for genesis to occur the wind must be roughly the same speed and blowing from the same direction at all height levels in the atmosphere.

Under these conditions, the embryo of a hurricane is born. The ascending air in the disturbance stimulates low-level inflow toward the center. This inflow slowly increases the cyclonic circulation of the disturbance, similar to the mechanism through which an ice skater spins faster as the skater pulls the arms inward (a process called conservation of angular momentum). When a closed circulation is observed, the disturbance is upgraded to a tropical depression. As long as the depression remains over warm water in a low vertical wind shear environment, the system will likely develop. Weak wind shear is a crucial factor because it allows vertical orientation of the thunderstorms and maintains the low-level inflow. Should the disturbance move into an environment where winds change dramatically with height, the thunderstorms are "torn apart" in different directions, the structure of the system is disrupted, and inflow weakens. Should adverse wind shear persist, or the system moves over colder water or move over land, the disturbance or depression will weaken and eventually dissipate.

If optimum conditions persist, the sustained cyclonic winds will continue to slowly increase. Typically, the genesis timeframe of disturbance and depression lasts for several days or longer. However, well-organized, rotating disturbance over warm water in weak shear can evolve much quicker. When the cyclonic sustained winds increase to 39 mph somewhere in the disturbance, the depression is upgraded to a tropical storm. At this point, the intensification stage begins.

For tropical storm intensification to a hurricane, the same conditions that allowed its initial development (warm water and weak wind shear) must continue. Should this favorable environment persist, the rate of development increases compared to the genesis stage. This is because as the wind increases, more heat and moisture is transferred from the ocean to the air. The column of air begins to warm, which lowers surface pressure. Air will flow from higher pressure to lower pressure, trying to redistribute the atmosphere’s weight, resulting in faster winds. In addition, the faster cyclonic winds also enhance convergence. Both factors increase thunderstorm production and low-level inflow. A feedback mechanism now occurs in which faster cyclonic winds breeds more potent thunderstorms, dropping surface central surface pressure more and creating stronger inflow, which breeds faster cyclonic winds, etc. Under favorable environmental conditions, a tropical storm can "spin-up" rather quickly, with winds increasing an additional 50 mph or more in a day. When sustained winds reach 74 mph somewhere in the storm, it is classified as a hurricane.

Water temperature is unquestionably linked to these storms’ development. Hurricanes rarely form over water colder than 80F (although exceptions do occur for reasons still not understood). They also weaken dramatically if a mature system moves over water colder than 80F, or if they make landfall, since their heat and moisture source has been removed. For a hurricane to maintain thunderstorms through atmospheric instability, warm, moist surface air is required near the low-pressure center. This warmth is provided by sensible heat transfer from warm ocean water, because otherwise air flowing toward lower pressure would expand and cool. (To convince yourself of this, let air out of a tire and feel how cool it is. When one lets air out of a high-pressure tire, the air expands as it enters the lower pressure environment and cools. This cooling occurs because the motion of gas molecules slows down as the air expands, and temperature is essentially a measurement of molecular motion). In other words, sensible heat flux from the warm water compensates for expansional cooling due to lower pressure, maintaining warm surface air near the storm center.

When a hurricane moves over cold water, expansional cooling dominates which stabilizes the atmosphere, the thunderstorms disintegrate, and the hurricane weakens. When a hurricane moves over land, the weakening occurs even faster because not only has the surface heat flux been lost, but so has the moisture source for cloud formation. Since land has more friction than water, this also weakens landfalling hurricanes, but this is a minimal influence compared to the loss of sensible heat and moisture flux.

While warm water is significant, it is nearly as important that the warm water be at least 200 feet deep. This is because hurricanes generate huge oceanic waves which mix the water to great depths. If the warm water only covers a thin film at the top, the hurricane will bring colder water to the surface and cut off the storm’s warm water energy supply, thereby weakening the system. In fact, sometimes a hurricane will kill itself when it become stationary for an extended period of time. Should a hurricane stop moving for several days, it can mix the ocean so much that all the warm water is replaced by cold water, and the hurricane dissipates. An example is Hurricane Roxanne in 1995 when it became stationary in the Bay of Campeche.

The warmer the water, the greater are the chances for genesis, the faster is the rate of development, and the stronger these storms can become. Under conditions of prolonged weak wind shear and water temperature greater than 85F, sustained winds may reach almost 200 mph. Table 3 shows the maximum potential intensity a tropical storm or hurricane can achieve for a given water temperature. Fortunately, few hurricanes over warm water reach their potential because some inhibiting factor (such as wind shear, landfall, or movement over colder water) occurs.

Table 3. Maximum potential intensity (as measured by sustained winds in mph) of a mature hurricane for a given water temperature (in F). Please note that this table is not valid for the genesis stage since hurricanes do not form over water colder than 80F. However, it is valid for a mature hurricane moving over colder water. Adapted from DeMaria and Kaplan (1994).

As shown earlier, hurricanes are most numerous and strongest in late summer for most ocean basins. This is because the three favorable conditions --- warm water, weak wind shear, and cyclonic disturbances --- are optimum in late summer. In particular, water’s temperature peaks in late summer. This seems paradoxical since the longest day is in June. However, the days are still longer than nights until fall, therefore the water is still accumulating heat into late summer. The monsoon troughs are most active in late summer as well, and large-scale circulation patterns favor weak wind shear in late summer. The exception is the North Indian oceans, where the season peaks in spring and fall (Table 2). This is because strong wind shear occurs over India during the summer, and because the Indian monsoon moves inland during the summer.

When a tropical depression is upgraded to a tropical storm, it is assigned a name. Before this practice was started, tropical storms and hurricanes were identified in many confounding ways. Some legendary storms have been inconsistently named for the holiday they occurred on (Labor Day Hurricane of 1935), the nearest saint’s day (Hurricane Santa Anna of 1825), the area of landfall (Galveston Hurricane of 1900), or even for a ship (Racer’s storm in 1837). However, most storms before the 1940s never received any kind of designation.

During forecast operations, the situation was just as perplexing. At first forecasters used cumbersome latitude-longitude identifications. This naming convention was too long, and was confusing when more than one storm was present in the same ocean basin. Right before World War II, this procedure was changed to a letter designation (i.e., A-1943).

The naming of hurricanes was amusingly initiated by the Australian forecaster Clement Wragge, who occasionally named severe storms after politicians he was displeased with. Wragge is apparently also the first person to give hurricanes female names. However, the naming convention began in earnest when western North Pacific World War II forecasters began informally naming tropical storms after their girlfriends and wives. This practice became entrenched in the system, and beginning 1945 Northwest Pacific storms officially were given female names.

Atlantic storms were officially given names starting in 1950 that were radio code words (i.e., Able, Baker, Charlie). In 1953 the US Weather Bureau switched to a female list of names. The Northeast Pacific began this convention in 1959 for the Hawaii region, and for the rest of the basin in 1960. Southwest India storms were first named during the 1960-1961 season. Australian storms were first named in 1964. Storms are not named in the North Indian oceans. Male names were added to the lists in the following years: Northwest Pacific in 1979, Northeast Pacific in 1978, Atlantic in 1979, Australia in the 1974-1975 season, and Southwest India in the 1974-1975 season (not sure about the last).

Today, six lists of names are used. Table 4 shows an example for the Atlantic Ocean. Each list is used alphabetically for a particular year, then is recycled six years later, with the exception of the Northwest Pacific. The latter ocean is active year round, and when the last name is reached on a list, the next list is used. Should an Atlantic storm be very destructive, or should it have a noteworthy impact on human lives or the economy, its name is retired (Table 5.) and replaced by a new name beginning with the same letter (LePore 1996). The letters "Q," "U," "X," "Y," and "Z" are not used by NHC, leaving 21 names on the list. It has never happened in the historical database, but should more than 21 Atlantic storms occur in a year, the next storms would be given Greek letter names (Alpha, Beta, etc.).

Other oceans contain a similar list of names. All names are determined by the World Meteorological Organization (WMO), and reflect the regional culture. For example, east Pacific storms tend to have Hispanic names.

The structure of a hurricane is certainly one of the most fascinating, awesome, and bizarre features in meteorology. Distinct cloud patterns exist for each stage of a hurricane’s life cycle. These patterns are so unique that a meteorologist can estimate the intensity of a depression, tropical storm, or hurricane solely based on cloud organization and cloud height using a methodology known as the Dvorak technique (Dvorak 1975). During genesis, typically a mass of clouds with a weak rotation is first observed, known as "stage 1 of genesis". Usually these clouds will temporarily dissipate, leaving a residual circulation, although in the case of a tropical wave sometimes a cloud pattern resembling an upside down "V" is observed. When the clouds return, and the circulation increases, "stage 2 of genesis" commences which coincides with the depression stage.

Clouds start forming a curved pattern, and when the winds reach tropical storm strength, intricate patterns emerge. In the center where convergence is strongest, a 100-mile wide region of clouds grows to 50,000 feet in height, surrounded by less tall but still potent thunderstorm bands out to 500 miles from the center. These curved thunderstorm bands, known as spiral bands, are clustered about 20 miles apart, with light to moderate rain in between. In the periphery of a tropical storm, wind speed will fluctuate, with fastest sustained winds of 30-40 mph in the spiral bands. As one approaches the center of a tropical storm, winds will consistently increase, with the strongest winds close to the center.

As the winds in the center increase to hurricane strength (74 mph), a ring of thunderstorms form around the center, known as the eyewall. In this eyewall are the fiercest winds and the most dangerous part of a hurricane. In the center itself, a clear region devoid of clouds form, known as the eye. In the eye, winds become weak, even calm! This transition from hurricane force winds to calm is rather sudden (often within minutes), and is truly one of the most bizarre features in the atmosphere.

The eye size can vary from 10 miles wide to 50 miles wide (there have even been cases of 100-mile wide eyes in the western North Pacific Ocean). Typically an eye starts at about 25 miles wide during the transition from tropical storm to hurricane. As the hurricane intensifies, usually the eye contracts. However, sometimes a second eyewall forms outside the original eyewall, known as the concentric eyewall cycle (Willoughby et al. 1982). This outer eyewall "cuts-off" off the inflow to the inner eyewall, causing the inner one to weaken and dissipate. As the eye expands, temporary weakening occurs. The outer eyewall will begin to contract inwards to replace the inner eyewall, and about 12 to 24 hours later, intensification resumes.

The cause of eye formation is still not understood, but there is some consensus among meteorologists that near 74 mph the strong rotation impedes inflow to the center, causing air to instead ascend about 10-20 miles from the center. Consider what happens as one drives at fast speeds in a vehicle. As the driver makes a sharp turn, an "invisible" force makes the driver lean outward. This outward-directed force, called the centrifugal force, occurs because the driver’s momentum wants to remain in a straight line, and since the car is turning, there is a tugging sensation outwards. The sharper the curvature, and/or the faster the rotation, the stronger is the centrifugal force.

As rotation increases in a developing hurricane, air is subjected to these outward accelerations which counteracts inflowing air. The centrifugal force eventually dominates near hurricane strength, causing inflowing air to not reach the center and ascend about 15 miles from the center. This strong rotation also creates a vacuum of air at the center, causing some of the air flowing out the top of the eyewall to turn inward and sink to balance this loss of mass. This subsidence suppresses cloud formation, creating a pocket of generally clear air in the center (although low-level short clouds or high-level overcast skies may still exist). People experiencing an eye passage at night often see stars. Trapped birds are sometimes seen circling in the eye, and ships trapped in a hurricane report hundreds of exhausted birds resting on their decks. The landfall of Hurricane Gloria (1995) on southern New England was accompanied by thousands of birds in the eye.

The sudden change of violent winds to a calm state is a dangerous situation for people ignorant about a hurricane’s structure. Those experiencing the calm of an eye may think the hurricane has passed, when in fact the storm is only half over with dangerous eyewall winds returning from the opposite direction within 20 minutes or less. Marjory Stoneman Douglas, the famous Florida Everglades activist, writes in the well-known book "Everglades: River of Glass" of the eye passage experience of survivors in the 1926 Miami hurricane:

In fact, many killed in the Labor Day Hurricane had wandered outside during the passage of the eye. In addition to the return of fierce winds, sometimes the highest storm surge (to be discussed later) occurs on the backside of the eye.

Outside the eyewall region, the weaker spiral bands accompanying the hurricane typically affect a large area. On average, the width of a hurricane’s cloud shield is about 500 miles, but it may vary tremendously. For example, the cloud shields of Hurricane Gilbert (1988) and Hurricane Allen (1980) covered an impressive one-third of the Gulf of Mexico. Some are giants such as Supertyphoon Tip (1979) which was twice the size of Gilbert or Allen! In contrast, the cloud shield of hurricanes may be less than 100 miles wide (typically called midgets by meteorologists) such as the Labor Day Hurricane of 1935 and Tropical Cyclone Tracy which hit Australia in 1974. However, storm size does not correlate with storm intensity. Both midgets and giant hurricanes can have sustained eyewall winds in excess of 100 mph. In fact, the Labor Day Hurricane and Tropical Cyclone Tracy were extremely devastating.

Coastal communities devastated by strong hurricanes usually take years to recover. Many forces of nature contribute to the destruction. Obviously, hurricane winds are a source of structural damage. As winds increase, pressure against objects increases at a disproportionate rate. For example, a 25-mph wind causes about 1.6 pounds of pressure per square inch. In 75-mph winds, that force becomes 450 pounds, and in 125-mph winds it becomes 1250 pounds. When the wind exceeds design specifications, structural failure occurs. Debris is also propelled by the winds, compounding the damage. For example, a four by eight sheet of plywood starts to become pushed by 50 pounds of wind pressure.

Isolated pockets of enhanced winds occur in hurricanes, too. As hurricanes make landfall, interactions with the thunderstorms form columns of rapidly rotating air in contact with the ground, known as tornadoes. Also accompanying the thunderstorms are areas where heavy rainfall accelerates air to the ground, known as downbursts, and spreads out at speeds greater than 100 mph. In addition, another phenomena was documented in Hurricane Hugo (1989) and Hurricane Andrew (1992) called mesoscale vortices (Willoughby and Black 1996; Fujita 1993). These are whirling vortices 150 to 500 feet wide which form at the boundary of the eyewall and eye where there is a tremendous change in wind speed. The updrafts in the eyewall stretch the vortices vertically, making them spin faster with winds up to 200 mph. Damage by these three wind phenomena occurs in narrow swaths inland, although sometimes it is difficult to discern which wind event caused a particular swath’s destruction during the post-storm analysis.

Floods produced by the rainfall can also be quite destructive, and is currently the leading cause of hurricane-related fatalities in the U.S. Hurricanes average 6-12 inches at landfall regions, but this amount varies tremendously and is not necessarily proportional to intensity. In fact, weaker tropical storms often produce greater amounts of rain, such as a US-record 42 inches of rain in 24 hours by Tropical Storm Claudette (1979) near Houston, Texas, causing $400 million (in 1979 dollars) making it the costliest Atlantic tropical system which was not a hurricane. Rainfall accumulation generally increases for slow moving storms. Hurricane Danny (1997) sat over Mobile Bay for almost one day before moving inland, dumping at least 37 inches of rain in coastal Alabama in 36 hours, of which 26 inches fell in 7 hours!

Heavy rainfall is not just confined to the coast. The remnants of hurricanes can bring heavy rain far inland, and is particularly dangerous in hills and mountains where acute concentrations of rain turn tranquil streams into raging rivers in the matter of minutes. In addition, mountains "lift" air in hurricanes, increasing cloud formation and rainfall. Rainfall rates of 1 to 2 feet per day are not uncommon in mountainous regions when hurricanes pass through. In fact, the highest hurricane rainfall amounts have occurred in the mountains of La Reunion Island (see Chapter 4 for details).

Some examples regarding rainfall damage and fatalities follow. Hurricane Camille (1969), which made landfall in Mississippi, dumped 30 inches of rain in 6 hours in the Blue Ridge Mountains, triggering flash floods and mudslides that killed 114 people in Virginia and 2 in West Virginia. One of the most widespread floods in United States history was caused by Hurricane Agnes (1972), which caused 188 deaths and $2.1 billion (in 1972 dollars) in property damage along most of the Eastern United States. Mudslides are also often a problem in hilly terrain, burying homes and destroying property, particularly in underdeveloped mountainous countries where the warning system, hurricane preparedness, and infrastructure are insufficient. Many of these countries have high terrain near the coast where mud slides and floods can occur with tragic results. Eastern Hemisphere countries sustain the worse fatalities since more storms occur there. A few of many examples include the Philippines (Typhoon Kelly (1981) with 140 killed; Tropical Storm Thelma (1991) with 3000 killed), China (Typhoons Peggy (1986) with 170 killed; Herb (1996) with 779 killed; and Nina (1975) with at least 10,000 killed), and Korea (Typhoon Thelma (1987) with 368 killed or missing). However, Western Hemisphere countries also suffer sizable casualties due to flooding and mudslides, such as in the Caribbean nations (i.e., Hurricanes Fifi (1974) with 8000-10000 killed; Flora (1963) with 8000 killed; and Gordon (1994) with 1145 killed), Mexico (i.e., Hurricanes Gilbert (1988) with 202 killed; and Pauline (1997) with 230 killed), and Central America (i.e., Hurricane Mitch (1998)).

As discussed in the preface, the recent Hurricane Mitch (1998) produced an estimated 75 inches of rainfall in the mountainous regions of Central America, resulting in floods and mud slides which destroyed the entire infrastructure of Honduras and devastated parts of Nicaragua, Guatemala, Belize, and El Salvador. Whole villages and their inhabitants were swept away in the torrents of flood water and deep mud that came rushing down the mountainsides, destroying hundreds of thousands of homes, and killing an estimated 11,000 to 18,000 people at press time. Mitch is the most deadly hurricane to strike the Western Hemisphere in the last 200 years, and the President of Honduras, Carlos Flores Facusse, has claimed the storm destroyed 50 years of progress. Typically in all the examples listed above, the number of homeless is in the 10,000 to over 100,000 range.

Although all these elements (wind, rain, floods, and mudslides) are obviously dangerous, historically most people have been killed in the storm surge, defined as an abnormal rise of the sea along the shore. The storm surge, which can reach heights of 20 feet or more, is caused by the winds pushing water toward the coast. As the transported water reaches shallow coastlines, bottom friction slows their motion, causing water to pile up. Ocean waters begin to rise gradually, then quite quickly as the storm makes landfall (it does not occur as a tidal wave as depicted in at least one Hollywood movie!).

Some factors that determine a storm surge’s height include storm intensity, storm size, storm speed, and the angle at which the hurricane makes landfall. The storm surge increases with storm intensity, size, and speed. The storm surge is also greater when landfall is perpendicular to the coastline, since some of the surge will be deflected offshore when storms land at a sideways angle. The shape of the coastal estuary is another important component, since coastal points and channels tend to enhance the surge in certain regions. One more element is the proximity of shallow water near the coast. Low-lying regions adjacent to shallow seas (such as the Gulf of Mexico in the southern U.S. and the Bay of Bengal bordering Bangladesh and India) are particularly vulnerable to the storm surge, since this allows more water to pile up before inundating the coast. Another minor contribution to the storm surge is the low pressure of a hurricane which allows water to expand a little (known as the inverse barometer effect). For every 10-mb pressure drop, water expands 3.9 inches. The storm surge is always highest on the side of the eye corresponding to onshore winds, which is usually the right side of the point of landfall in the Northern Hemisphere, called the right front quadrant. Winds are also fastest in the right front quadrant because storm motion (which averages about 10 mph but varies substantially) is added to the hurricane’s winds.

The total elevated water includes two additional components --- the astronomical tide and ocean waves. The astronomical tide results from gravitational interactions between the earth, moon, and sun, generally producing two high and two low oceanic tides per day. Should the storm surge coincide with the high astronomical tide, additional feet will be added to the water level, especially when the sun and moon are aligned since this produces the highest oceanic tides (known as syzygy). The total water elevation due to the storm surge, astronomical tides, and wave setup is known as the storm tide. Therefore, the storm surge is officially defined as the difference between the actual water level under the hurricane’s influence and the level due to astronomical tides and wave setup. In practice, water level observations during post hurricane surveys are always storm tides, and it is difficult to distinguish between storm surge and storm tide water elevations. Therefore, the two terms are used interchangeably.

Water is very powerful, weighing some 1700 pounds per cubic yard, and therefore most inundated structures pounded by waves and the storm surge will be demolished. Ocean currents set up by the surge, combined with the waves, can severely erode beaches, islands and highways. Most people caught in a storm surge will be killed by injuries sustained during structural collapse, or by drowning. Death tolls for unevacuated coastal regions can be terrible. The worst natural disaster in United States history occurred in 1900 when a hurricane-related 8-15 feet storm tide inundated the island city of Galveston, Texas and claimed over 6000 lives. In 1893, nearly 2000 were killed in Louisiana and 1000 in South Carolina by two separate hurricanes.

Hurricane Camille (1969), with sustained winds of at least 180 mph, produced a storm tide of 23 feet in Pass Christian, Mississippi. As the storm tide penetrated far inland, Camille killed 137 people in Mississippi and 9 in Louisiana, including 20 of 23 people who ignored evacuation warnings and stayed for a "hurricane party." A survivor of the hurricane party, who clung to floating debris, traveled ten miles or more in the storm tide. In Louisiana, Camille produced a storm tide which pushed water over both levees near the mouth of the Mississippi River, "removing almost all traces of civilization" as one U.S. Department of Commerce (1969) report states. Floods from the storm tide penetrated 8 miles inland in the Waveland-Bay St. Louis region, and in river estuaries the flood extended 20 to 30 miles upstream (Corps of Engineer 1970). When combined with the 116 Virginia flood deaths mentioned earlier, Camille killed a total of 262 people (Corps of Engineers 1970). A total of $1.5 billion in property damage (in 1969 dollars) was incurred, with total devastation on the immediate coastline and severe damage further inland.

However, these statistics pale compared to the lives taken in coastal India and surrounding countries by storm tides. The most vulnerable area is coastal Bangladesh, a huge river delta fertile enough to support large numbers of farmers and fishermen (about 1500 per square mile). Many are essentially nomads, staking claims on frequently changing temporary islands. The geography of this area favors large storm surges because the narrow inlet of the bay funnels large amounts of water inland, and because shallow water extends 60 miles offshore. Storm surge warnings are issued more than a day in advance, but communicating this information is difficult due to the rural and nomadic nature of the population. Furthermore, many choose to ignore the warnings, or find it too difficult to evacuate since no transportation infrastructure exists (Rosenfeld 1997). The results have been the worst storm surge fatalities in history.

Six hurricanes have hit Bangladesh since 1960, killing at least 10,000 each time (Rosenfeld 1997). The most tragic incident occurred on November 12, 1970, when a hurricane with 125-mph winds caused a 20-foot storm tide that killed 300,000 people. This storm triggered a revolution against Pakistan which brought Bangladesh independence. Unfortunately, this tragedy was still repeated in 1991 when 139,000 people perished in another Bangladesh hurricane. In general, this area has a history of hurricane-induced fatalities: 300,000 people were killed near Calcutta, India in 1737; 20,000 were killed in Coringa, India in 1881; and 200,000 were killed in Chittagong, Bangladesh in 1876. Unfortunately, evacuation procedures and public response has not changed much since 1991, so the chance of similar tragedies remains high. China, Thailand, and the Philippines have also seen losses in the tens to hundreds in recent years due to storm surges.

The combination of heavy rainfall and a storm surge can also be particularly devastating, especially along coastal streams. Under these conditions, residents will first experience the surge, which typically last for one day, followed by flooding from run-off that persists for weeks. For example, coastal homes in Pasagoula, Mississippi were devastated by slow-moving Hurricane Georges (1998) because the storm surge first inundated the region, followed by flooding from the Pasagoula River.

To quantify the expected levels of damage for a given hurricane intensity, Herbert Saffir and Robert Simpson devised the Saffir-Simpson scale (Simpson 1974). This scale, only valid for the Atlantic Ocean, classifies hurricanes into five categories according to central pressure, maximum sustained winds, storm surge, and expected damage (Table 6). Although all categories are dangerous, categories 3, 4, and 5 are considered major hurricanes, with the potential for widespread devastation and loss of life. Whereas only 21% of U.S. landfalling tropical systems are major hurricanes, they historically account for 83% of the damage (Pielke and Landsea 1998). On average, the Atlantic has two major hurricanes per year (Table 1). Fortunately, Category 5 hurricanes are infrequent in the Atlantic Ocean, and seldom sustain themselves at such intensities for very long before weakening to a lower category. Only two Category 5 hurricanes have made landfall in the United States this century (Hurricane Camille in 1969, and the Florida Keys’ Labor Day Hurricane of 1935). Major hurricanes are much more common in the eastern North Pacific and in the western North Pacific. In fact, in the western North Pacific Ocean major hurricanes (or typhoons, as they are called there) are so common that they’ve developed a new definition for storms with sustained winds greater than 149 mph --- supertyphoons. Some Australians give major hurricanes the colorful name "cock-eyed Bob."

While hurricanes wrought much misery, they can also be beneficial. Hurricanes often provide much needed rain to drought-stricken coastlines. Their ocean interactions can flush bays of pollutants, restoring the ecosystem’s vitality. After the record rainfall from Hurricane Claudette (1979) in Texas, fish were being caught in the northern industrialized reaches of Galveston Bay that had vanished for several years. Finally, in cruel Darwinism fashion, weak sea life and plants will perish, leaving only the strong to survive and reproduce.

In this vein, sometimes hurricanes "correct" man’s mistakes. For example, in the early 1900s non-native foliage such as Australian pine trees had been planted on the tip of Key Biscayne, Florida (now the Bill Baggs State Park). These non-native plants had few natural enemies in their new environment, and quickly dominated plant life, resulting in a loss of natural habitat. However, these Australian non-natives lacked the ability to withstand hurricane-force winds, and Hurricane Andrew destroyed them all in 1992. Park officials seized the opportunity to replant the park with native foliage.

Hurricanes are one of the most difficult phenomena to forecast in meteorology. A forecaster must understand all facets of meteorology for hurricane prediction since these storms encompass all weather processes, from individual thunderstorms to rainband physics to air-sea coupling to interaction of the hurricane itself with the surrounding atmosphere. These large and small-scale weather features also interact with each other in complicated ways that even a computer can only crudely simulate and predict these processes. Even worse, since hurricanes occur over the data-sparse ocean, forecasters have few observations to see the current state of a hurricane or to input into the computers. Without knowing what the storm is doing now, how can one anticipate its future? Forecasting hurricanes is indeed a challenge.

Government hurricane forecast centers exist worldwide (Table 7.) The center charged with Atlantic and Northeast Pacific hurricane forecasts is the National Hurricane Center (NHC) located on the campus of Florida International University in Miami. All the hurricane centers utilize similar forecast procedures, so this book will discuss NHC operations.

For evacuation and emergency preparedness purposes, forecasters are most concerned with predicting where the storm will go and, should the hurricane threaten land, where the eye will make landfall. Of nearly equal importance is forecasting hurricane intensity and the width of the tropical and hurricane force winds around the hurricane. Forecast statements are also issued for expected rainfall and storm surge. A tornado watch, which means conditions are favorable for tornado development, is always issued preceding hurricane landfall by the nearest National Weather Service office. When a tornado has been observed by trained people called "weather spotters" or inferred by an instrument called Doppler radar (which can detect areas of rapid rotation in a thunderstorm), a tornado warning is issued by the National Weather Service.

Hurricanes have a reputation for being unpredictable at times. Hurricanes can suddenly turn, speed up, slow down, stall, or loop. They can also reform their center of rotation when thunderstorms are not uniformly distributed around the center, making the storm suddenly "jump" from one spot to the next (an example is Hurricane Earl in 1998). Track forecasts have improved considerably the past 20 years, but errors still occur. Intensity forecasts, however, currently exhibit little skill. The process of anticipating whether a hurricane will strengthen, weaken, or not change intensity ---- and predicting how quickly any intensity change may occur ---- is still an unsolved forecast problem. Such uncertainties are fraught with potential disasters, as described in the "near-miss" for Hurricane Opal (1995) below.

Hurricane Opal formed in the southern extent of the Gulf of Mexico off the Yucatan peninsula. Initial computer models forecasted it would accelerate toward Florida. However, it only slowly drifted northwards. Therefore, NHC forecasters became skeptical as each subsequent model run projected a fast northward motion that did not materialize. Opal was also a Category 1 hurricane, and not expected to intensify past Category 2. Based on Opal’s slow motion and the fact that it was a minimal hurricane, at 5PM on October 3, 1995, NHC decided a hurricane warning was not necessary until the following morning for the Florida panhandle.

Late that night, Opal intensified explosively to a Category 4 with winds of 150 mph, just shy of Category 5 status. (Later analysis attributed this rapid development to movement over an isolated pool of very warm water. Interactions with environmental 40,000 foot winds may have also promoted the intensification.) Even worse, Opal finally began accelerating towards Florida. Since most people were sleeping, it was difficult to alert the public about this new, unanticipated danger. Many people near Pensacola Beach awoke the next morning with a near-Category 5 hurricane just offshore. The last minute evacuation procedures clogged the roads and Interstate 10, creating traffic jams which moved less than 10 mph. It was NHC’s worst nightmare coming true.

Fortunately, Opal weakened before landfall from a Category 4 to Category 3 hurricane. While still dangerous and destructive, this was a bit of a reprieve. Furthermore, while many residents were unable to use the roads and had to stay home during the storm, most residents on the immediate coast did manage to evacuate. Opal caused $3 billion in damage and killed 20 people in Florida, Alabama, Georgia, and North Carolina, but none of these were from the storm surge. The people of Florida had dodged a bullet.

Opal demonstrated that large track forecast errors still occur, and when combined with unexpected intensification, many congested coastal regions are today toying with human catastrophe since seashore development has proceeded without seemingly any considerations for speedy evacuation. These issues will be discussed later, but first a background on storm motion and its prediction is necessary.

To understand hurricane motion, it is helpful to use the analogy of a wide river with a small eddy rotating in it. To a first approximation, the river transports the eddy downstream. However, the eddy will not necessarily move straight because the speed of the current varies horizontally; for instance, it may be faster in the center and slow down towards the river banks. As a result, the eddy may wiggle off a little to the left or right as it moves downstream, and depending on the situation, may speed up or slow down. Furthermore, this eddy’s rotation may alter the current in its vicinity, which in turn will alter the motion of the eddy.

Likewise, one may think of a hurricane as a vortex imbedded in a river of air. The orientation and strength of large-scale pressure patterns basically dictate the hurricane’s motion, except that the steering depends both on the horizontal and vertical wind distribution. There is a tendency for stronger hurricanes to be steered by winds higher aloft, thus requiring the forecaster to identify the best "steering height" for a particular storm. To make a track forecast, one must first predict the large-scale flow within which the hurricane is imbedded, and this can be difficult --- how does one separate the hurricane winds from the environmental winds, and how does one know what height best represents the steering flow?

Forecast errors also occur when the steering current is weak and ill-defined, thus causing the hurricane to stall or slowly drift. However, the largest errors are associated with rapid changes in the strength and orientation of the steering current. Under these situations, a hurricane may accelerate its motion and/or turn. For example, sometimes forecasters are faced with the dilemma of a hurricane moving toward the United States East Coast, not knowing for sure whether it will hit land before being turned back to the right by westerly steering currents poleward of the hurricane.

The hurricane can internally change its course as well. Just as in the river example, the hurricane can interact with surrounding pressure fields, thus altering its own steering current. Research has also shown that the earth’s rotation theoretically affects the motion by inducing a weak poleward and westward drift of 2-3 mph (known as the beta effect). On rare occasions when two hurricanes get within 850 miles of each other --- especially in the Pacific and Southern Hemisphere oceans --- they may begin to move towards each other and rotate about a common midpoint between them (known as the Fujiwhara effect). Even when a hurricane moves relatively straight, a detailed analysis shows small oscillations about the mean path, called trochoidal motion.

Until the late 1970s, NHC forecasters relied on meteorological intuition, statistical schemes, and the storm’s past trend to make track predictions. They would also look at past tracks from other years to see which way similar storms moved. These techniques are still used today.

However, forecasters began to realize such difficult predictions required the use of computer models. Computer models ingest current weather observations, and approximate solutions to complicated equations for future atmospheric values such as wind, temperature, and moisture. NHC uses a suite of models which differ in their mathematical assumptions and complexities in describing atmospheric processes. Some of these differences exist because certain atmospheric features have been removed to make the computer program run faster. Other differences exist because meteorologists are still uncertain about how to formulate certain weather features on a computer, such as cloud processes. The most complex models must be run on the fastest computers in the world, known as supercomputers. Model predictions can vary because of these mathematical differences. Model forecasts also contain small errors because the mathematical solutions contain a small amount of uncertainty which accumulates with time; therefore, all forecast errors grow with time. Models will also produce incorrect forecasts if they are initialized with bad data, which is a frequent problem over the data-sparse ocean (as they say "garbage in, garbage out.")

Nevertheless, computer models have revolutionized all aspects of meteorology, including hurricane track forecasts, and they are improving each year. Recently, NHC started using a model developed at the Geophysical Fluid Dynamics Laboratory that is generally 20 percent better than its predecessors. Unfortunately, computer guidance for intensity forecasts is still unreliable. Forecasters use statistical schemes and intuition to make their intensity predictions.

Hurricane forecasters also monitor the latest observations from satellites, ships, radar, buoys, oilrigs, and other sources to assess any track or intensity changes. When a hurricane is far offshore in data-void regions, the intensity is estimated from satellite cloud patterns using the Dvorak technique. While generally robust, the Dvorak technique can produce wrong intensity values, as in the case of Typhoon Omar (1992). The Dvorak technique estimated Omar to be a Category 1 storm, but was discovered to actually be a Category 2 storm hours before hitting Guam, surprising the island’s inhabitants and perhaps amplifying damage due to lack of preparation.

Therefore, the most important data platform is reconnaissance planes that fly into the hurricane’s eye and take critical meteorological measurements. The precise information required for obtaining hurricane structure, location, and intensity cannot be obtained in any other way, and is crucially important for forecasts and evacuation procedures. Forecasters credit a reconnaissance plane for discovering the Hurricane Camille (1969) had strengthened from sustained winds of 115 mph to 150 mph on August 16, 1969 at 5PM CDT, 30 hours before landfall. Based on this new information, additional evacuations were prompted which may have saved up to 10,000 lives by one estimate. A reconnaissance plane also measured Hurricane Opal’s (1992) unexpected rapid intensification. Reconnaissance flights are expensive, and only occur in the Atlantic Ocean since only the United States government has the financial capability to fund them. At one time, Congress had considered halting reconnaissance, but once their forecast and evacuation impacts were made clear by scientists and evacuation managers. Not only was funding preserved, a line item was added to the U.S. budget explicitly setting aside funds for Atlantic reconnaissance. At one time, Air Force reconnaissance flights also occurred in the western Pacific Ocean, but budget restrictions forced those flight missions to end in 1987. Occasional surprises such as Typhoon Omar are the result.

Continuous reconnaissance flights begin once a tropical system moves close enough to land. The 53rd Weather Reconnaissance Squadron at Keesler Air Force Base near Biloxi, Mississippi --- also known as the "Hurricane Hunters" --- takes most of these measurements on WC-130s. Two additional P-3 Orion planes with more sophisticated instruments (including a radar) are also available by the NOAA Aircraft Operations Center at MacDill Air Force Base in Tampa Bay, Florida. The P-3s are deployed on less routine flights to perform analysis of hurricane structure in conjunction with scientists at the Miami NOAA Hurricane Research Division (HRD). The P-3s also transmit data to NHC.

Obviously, reconnaissance flights carry an element of risk. In September 1955, a Navy plane and its crew of nine plus two Canadian newsmen were lost in the Caribbean Sea while flying in Hurricane Janet. Three Air Force aircraft have been lost flying in typhoons in the Pacific. Fortunately, no planes have been lost in the last 25 years, but scary moments still happen. In one extreme circumstance during Hurricane Hugo (1989), one of the P-3 planes encountered a severe mesoscale vortex, started sinking, and regained altitude moments before it would have crashed into the ocean. Most flights are less eventful, but do contain some "bumps" between thunderstorm updrafts and downdrafts, and occasional roller coaster-like drops (or ascents) of 3000 feet in less than a minute due to strong updrafts and downdrafts. Since hurricane winds are strongest just above the surface (about 1000 feet) and weaken with height, reconnaissance flights are performed at 5000 or 10,000 feet. During the hurricane penetration, information about the horizontal wind and temperature structure is transmitted to NHC. Once the plane enters the eye, they deploy a tube of instruments (called a dropsonde) which parachutes downward from flight level to the sea, sending valuable intensity measurements back to NHC.

In summary, forecasters at the National Hurricane Center critically evaluate each computer prediction for reasonableness and consistency with other forecasts, observed data, statistical projections, the historical record, and extrapolated positions. They also monitor the latest observations for last-minute adjustments. Based on all of this information, every six hours the forecasters produce a new forecast of the storm’s projected location and intensity for the next 72 hours. When hurricane conditions on the coast are possible within 36 hours, a hurricane watch is issued. When hurricane conditions are likely within 24 hours, a hurricane warning is issued. These watches and warnings refer to the arrival of hurricane force winds of 74 mph, not eye landfall which generally occurs a few hours later. Since 24-hour forecast track errors currently average about 100 miles, the warning is issued for a rather large coastal area (about 350 miles), and is quantified by the percent chance the hurricane’s center will pass within 65 miles of a particular location.

Since Hurricane Camille killed 256 people in 1969, U. S. hurricane-related fatalities have dropped dramatically as storm surveillance, evacuation procedures, public awareness, and forecasts have improved. However, new concerns have emerged as coastal population growth and property development has exploded during this period. In fact, currently U.S. population increases are largest in coastal communities. The population has grown on average by 3-4 percent a year in hurricane-prone regions, and in some states like Florida the growth is greater. As a consequence, property damage costs due to hurricanes have skyrocketed in the 1990s (Table 8).

This trend is disturbing in other ways. First of all, during this period of larger property costs, the number of major hurricanes (Category 3 or better) has actually decreased since 1970 (Landsea et al. 1996). This decrease corresponds with an active period of intense hurricanes during the 1940s to 1960s which suddenly changed to a quiet period of intense hurricanes during the 1970s to early 1990s. (Such shifts in weather activity are known as multidecadal changes). This change is likely due to a cooling in the Atlantic Ocean water temperature during this period, and may also be related to a drought which started in 1970 in Africa where many tropical waves form before propagating out into the Atlantic.

Second, the number of major hurricane making landfall on the East Coast of the U.S has dramatically decreased since 1965. During the period 1944 to 1964, 17 major hurricanes hit the East Coast (Landsea 1998a). In contrast, only 5 major hurricane made landfall on the East Coast between 1965 to 1995 (with a period between 1965 and 1983 where no East Coast landfall occurred!). It is during this period much population growth and development has occurred on the East Coast, apparently ignoring or not knowing what happened during 1944-1964.

Some evidence suggests that major hurricane activity occurs in 40-60 year cycles, where 20-30 years of active Atlantic hurricane seasons will be followed by 20-30 years of relatively quite hurricane seasons (Gray et al. 1996). In fact, there are signs that the downward trend in hurricane activity may be ending. 1995 was an extraordinarily active year with 19 tropical storms and hurricanes, 11 hurricanes, and 5 intense hurricanes. The following years 1996 and 1998 were also above normal. In fact, the four-year period of 1995-1998 had a total of 33 hurricanes --- an all-time record in the Atlantic Ocean. Global weather patterns are also emerging which are favorable for more active hurricane seasons in the future, such as a return of warm Atlantic water temperatures (Gray et al. 1998). When the more active hurricane phase returns (if it hasn’t already), property damage costs could easily exceed $50 billion in one year, especially should a major city be hit! Irregardless of trends in Atlantic hurricane season activity, each year the U.S. has at least a 1 in 6 chance of experiencing hurricane-related damage of at least $10 billion [in 1995 dollars] (Pielke and Landsea 1998).

Even more disturbing is the sharp population increase. Most new coastal residents have never experienced a hurricane, increasing the chances some may ignore evacuation orders or not take proper precautions. Also, some coastal regions are so congested now it is becoming more difficult to perform a timely evacuation. Track forecasts have improved by about 1 percent per year, but the 3-4 percent population growth could overwhelm these better predictions, resulting in higher casualties again. Even worse, should a hurricane unexpectedly change course or accelerate its motion, there may not be enough lead time for orderly evacuation of intensively developed coastal regions, trapping many residents at landfall.

Even before a hurricane makes landfall, millions of dollars are lost in evacuation costs. At least $300,000 of business losses and hurricane preparation costs is incurred per day for every mile of coastline evacuated. Some experts claim the cost is even larger, perhaps $1 million per mile.

Therefore, a top priority has become accelerating the improvements in forecasting accuracy. Research has shown that more detailed reconnaissance observations of hurricanes may achieve this goal. As a result, HRD has improved the dropsonde using Global Positioning System technology so that hurricane observations may be obtained with unprecedented accuracy. In addition, the new dropsondes have the ability to take observations below 1500 feet --- something the older dropsondes could not do. A new, faster plane has also been added to the fleet --- the Gulfstream IV. Since the Gulfstream is a jet, it can reach altitudes of 45,000 feet, unlike older propeller-driven planes belonging to the Hurricane Hunters and to the Aircraft Operation Center which can only reach 25,000 feet at best.

Starting in 1998, the objective is to release 46 dropsondes from 40,000 feet in 20 flights around hurricanes. This will relay a better picture of the atmospheric conditions that surround and steer a hurricane. Other Air Force and HRD planes will deploy drifting buoys in the path of a hurricane so as to better understanding storm conditions near the ocean surface. Research has shown that this additional data can reduce computer model forecast track errors up to 30%.

If this estimated track improvement is correct, it may allow NHC to shrink the warning zone by 50 to 80 miles which would save millions in evacuation costs, especially if a major city is excluded from the warning area. It would also reduce traffic congestion, speeding up evacuation from the most threatened region. Furthermore, if homeowners take advantage of more accurate forecasts by seriously protecting their homes, it is estimated another 10-15 percent in costs would be saved --- and any lives saved by the improved forecasts would be priceless.

Other exciting technological developments are about to occur. The Hurricane Hunters will be replacing their WC-130H aircrafts with WC-130J planes in 1999; the latter are faster, more fuel-efficient (resulting in longer flights), and most importantly, can travel at 37,000 feet or more. This will further augment operational measurements critical to hurricane forecasting. In addition, starting in 1998 NASA scientists, NOAA scientists, and universities are collaborating to collect data at all levels of hurricanes using multiple aircrafts, remote sensing technologies, and portable surface measuring platforms (see Chapter 2 for details about the). The datasets produced by this and future missions will be unprecedented in their comprehensiveness, providing researchers with new information towards understanding hurricanes.

By the 21^st century, another new type of plane may be taking oceanic observations. Scientists in Australia and Canada are experimenting with small, pilotless aircrafts typically called "drones," or sometimes "aerosondes." These drones theoretically can be programmed to fly a fixed path for over 24 hours, continuously ingesting weather observations that would improve weather forecasts. The first successful cross-Atlantic flight by an unmanned aerosonde was accomplished in August 22, 1998. Their unique aerodynamic structure and relative weightlessness also allows them to theoretically withstand hurricane winds and strong updrafts, although this has yet to be convincingly demonstrated.

In the early 1980s, Dr. William Gray at Colorado State University asked the question, "Why do some Atlantic hurricane seasons have many storms and why are other seasons inactive?" He further wondered if one could predict, months in advance, the number of tropical storms and hurricanes for the upcoming Atlantic tropical season. Based on years of research, in 1984 Gray began to publicly predict how active the Atlantic hurricane season would be before it starts. His predictions, at times, have been remarkably accurate, and have scientifically proven to be skillful compared to guessing. His forecasts, which are well publicized in the media, are issued every December, April, June, and August. The forecast is also available on the web at http://tropical.atmos.colostate.edu.

Gray and his students have discovered several surprising global signals that affect Atlantic hurricane activity (Gray et al. 1998). The reasons for some of these associations remains unclear, and are still being researched. These include:

1. The El Niño-Southern Oscillation (ENSO), also sometimes just called El Niño. El Niño occurs when warm water in the equatorial western Pacific Ocean shifts east, and occurs about every 3-7 years. What has the Pacific Ocean got to do with Atlantic hurricanes, you ask? This change in the Pacific Ocean alters global wind patterns and modifies many weather elements, including reducing hurricane activity. In El Niño years, wind shear is enhanced in the Atlantic Ocean which is destructive to tropical storm formation. An El Niño officially ends when the warm eastern Pacific waters cool to average temperatures; however, usually the cooling continues, which is called a La Niña. La Niña years are associated with weaker than average Atlantic wind shear and enhanced Atlantic hurricane activity. The occurrence of an El Niño or La Niña is one of the most important modulators of Atlantic hurricane activity.
2. African rainfall. There is a strong correlation between rainy years in the western Sahel of Africa and major (Category 3, 4, or 5) hurricane activity. Likewise, during drought years in Africa major hurricane activity is reduced. Most major hurricanes come from tropical waves which originate over Africa. It has been hypothesized that, during rainy years, tropical waves are stronger and more conducive to genesis as they move westward off the African coast. However, Goldenberg and Shapiro (1996) show that years with strong (weak) wind shear are associated with droughts (rain) in Africa. It is also likely related to changes in water temperature in the Atlantic Ocean, where years with warm (cold) water are correlated with wet (dry) years in Africa and more (less) strong hurricanes (C. Landsea, personal communication 1998).
3. Pressure and temperature difference between western African coast and the Sahel region during the previous February-May period. When these differences are conducive to onshore flow that transports moisture over Africa, hurricane activity is often enhanced; when it is conducive to offshore flow, hurricane activity is often suppressed.
4. Caribbean Sea Surface Pressure. When pressure is lower (higher) than normal in the Caribbean Sea, Atlantic hurricane activity is enhanced (suppressed). High pressure indicates sinking air, which suppresses thunderstorm formation in easterly waves and other tropical disturbances. Knaff (1997) also showed that higher pressure is also correlated with strong wind shear.
5. Quasi-Biennial Oscillation (QBO). The QBO is an oscillation of equatorial winds between 13 and 15 miles aloft. These winds change direction between westerly and easterly every 12-16 months. Westerly winds are associated with more hurricanes than easterly, especially when the wind is westerly at both 13 and 15 miles aloft. It has been postulated that a westerly QBO reduces wind shear, but the reason for this hurricane association remains unclear.
6. Caribbean wind shear. During seasons when shear is greater than normal in the Caribbean Sea, hurricane activity is suppressed, and vice versa for seasons with weak wind shear.
7. Atlantic Ocean water temperature. Obviously, warmer than normal sea surface temperature is favorable for active hurricane years, and vice versa for colder than normal water.
8. Strength of Azores High Pressure System. The Azores High is a permanent feature located between 20 and 30 deg W in the Atlantic Ocean. When pressure is higher than normal in the Atlantic, tropical northeast winds tend to be stronger than normal. This, in turn, upwells cold water to the surface off the northwest African coast, thereby reducing the chance of tropical genesis as tropical waves propagate off Africa. There also is some long-term feedback, such that a strong Azores high in the fall results in high pressure in the Caribbean Sea the following year (Gray et al. 1998; Knaff 1998). Therefore, when the Azores high is stronger (weaker) than normal in the central Atlantic, hurricane activity is enhanced (suppressed).

In general, when more of these predictors are favorable for hurricane activity than unfavorable, Gray predicts an above average hurricane season, and when most are unfavorable, a quite hurricane season is predicted. When the same number of predictors have positive and negative influences, an "average hurricane season" is predicted (about 9 named storms, 6 hurricanes, and 2 intense hurricanes; see Table 1). However, caution must be advised here since some predictors are more important than others. For instance, despite the fact that most factors were favorable for an active hurricane season in 1997, a record El-Niño that year dominated the weather resulting in a quiet hurricane season. Also, one should be careful attributing Atlantic hurricane activity to one feature in any year, because many of these features are inter-related; for example, an El Niño tends to be associated with strong Caribbean wind shear.

Gray quantitatively predicts several parameters using statistical techniques and intuition, such as the number of named storms (which includes tropical storms and hurricanes), the number of hurricanes, the number of days with tropical storms or hurricanes, etc. He also forecasts the number of major hurricanes, which includes Category 3, 4, or 5 hurricanes, since these are the storms which cause about 83% of total hurricane damage (Pielke and Landsea 1998). Table 9 summarizes Gray’s forecast of named storms versus the observed number.

Note from Table 9 that during 11 of 15 years, Dr. Gray correctly predicted whether the observed number of storms would be above or below the average of 9 named storm. However, the years that were incorrect (1989, 1993, 1997, and 1998) are instructive because they led to new insight on how to improve the forecasts. For example, after analyzing the 1989 underforecast of storm activity, the importance of African rain was discovered since 1989 was the only non-drought year that decade. Starting in 1990, African rainfall was included in Gray’s seasonal forecasts. Also, since the predictions are based on statistical probabilities, the forecast will fail in some years (just as local National Weather Service statistical forecasts of rain and temperature will occasionally fail). Gray has to also make a correct prediction of parameters such as African rain and El Niño for an accurate seasonal hurricane forecast. If he forecasts a basic predictor incorrectly (such as not predicting the record El Niño of 1997), the seasonal forecast will probably be wrong.

Starting 1998, Gray will begin issuing forecasts for hurricane landfall probability (Gray et al. 1998). It will be interesting to observe how this new extended range forecast methodology will evolve.

All hurricane seasons should be taken seriously by coastal residents, even if the seasonal forecast is "below normal." Many devastating hurricanes have occurred in otherwise inactive seasons, such as Hurricane Andrew (1992), Hurricane Alicia (1983), and Hurricane Allen (1980).

Some environmental groups claim hurricanes are becoming more numerous and stronger (Leggett 1994) due to a reputed phenomena known as "global warming." Media outlets also have made such assertions (Newsweek 1996). First of all, the topic of global warming is itself very political and controversial. Although a majority of scientists believe global warming is occurring, many scientists still discredit these allegations with solid arguments. This book will not the global warming controversy in any detail since many other books are devoted to the subject, including an ABC-CLIO book (Newton 1993).

However, this author will discuss two issues concerning any links to possible global warming and hurricanes. The first question is: have hurricanes increased in numbers and/or intensity in the past two decades due to possible global warming? The second question is: will hurricane activity increase in the future due to possible global warming?

In short, the answer to the first question is: no. The answer to the second question is: maybe, although reductions in hurricane activity are also possible. Before addressing these questions in detail, a review of the greenhouse effect and global warming is required.

A balance between incoming solar radiation from the sun and outgoing radiation from the earth primarily regulates the atmosphere’s temperature. This is a complicated process requiring some explanation. When an object becomes warmer than –273.15C (the temperature where all molecular motion ceases, also called absolute zero on the Kelvin temperature scale, and equivalent to –459.67F), it begins to emit electromagnetic radiation. Therefore, as the sun warms the earth environment, the earth begins to emit its own radiation. However, radiation is dependent on temperature. The hotter an object is, the shorter the wavelength of peak emission. The much hotter sun emits most of its radiation as visible light. The cooler earth emits its energy at a much longer wavelength which human eyes cannot see called the infrared spectrum.

It turns out that molecules in the atmosphere behave differently with regard to the passage of infrared radiation and visible radiation through them. It’s a complex process, but the end result is that some air molecules allow visible radiation to pass through them, but obstruct infrared radiation. With this background, the heat balance of the earth and its atmosphere can now be discussed.

As solar radiation flows from the sun toward the earth, it passes freely through the atmosphere and warms the earth. However, as the earth tries to emit this heat back to space in the infrared spectrum, some of it becomes "trapped" by the atmosphere, thus warming the air. A portion of this energy is radiated back to the earth, which warms the surface. The earth, in turn, reradiates this infrared energy upwards, where it is again absorbed and warms the lower atmosphere some more. In this way, the atmosphere acts as an insulating layer, keeping part of the infrared radiation from escaping rapidly to space. Consequently, the earth’s surface and lower atmosphere are much warmer than they would be without this selective absorption of infrared radiation. This process is often called the greenhouse effect since it is analogous to a glass building which allows visible light inside but prevents some infrared radiation from leaving, thus keeping the plants inside warm (even in winter). It’s important to realize that the atmosphere’s greenhouse effect is a natural process, and without it the earth would be a much colder, unlivable planet with an average surface temperature of 0F.

There is a distinction between the atmosphere and a greenhouse --- in a greenhouse the glass allows visible light to pass through but restricts the passage of infrared, while in the atmosphere molecules are differentiating between visible and infrared radiation. Also, a greenhouse warms quickly because glass is a physical barrier to air movement, while in the atmosphere mixing occurs freely between warm air at the ground and cooler air aloft, thus slowing down warming. Nevertheless, the phrase "greenhouse effect" is used throughout the media and meteorology, so the nomenclature will be used in this book as well.

Certain molecules are more effective at trapping infrared radiation than others. The most important greenhouse gas is water vapor because it strongly absorbs a portion of outgoing infrared radiation and is a plentiful gas. The other greenhouse gas is carbon dioxide (), which is equally as absorptive but far less plentiful than water vapor, and therefore plays a much smaller role in the greenhouse effect. is a natural component of the atmosphere, produced mainly by the decay of vegetation.

However, is also produced by the burning of fossil fuels such as coal, oil, gasoline, and natural gas. Observations show that has increased by more than 10% since 1958, coinciding with increases in fossil fuel emissions from automobiles, factories, and other power sources. Global warming theory states that the greenhouse effect will be enhanced as increases because more outgoing infrared radiation will be absorbed. There is also the possibility that increasing the temperature will also increase water vapor concentrations (the major greenhouse gas) due to increased evaporation rates, further enhance warming prospects. Some scientists contend global warming will increase the earth’s average surface temperature by 1.5 to 4.5C. However, other scientists contend feedback processes involving clouds and the ocean may cancel that potential warming. Some scientists, based on decades of observations, also say global warming is already occurring. However, other scientists using different observation techniques show no true warming has happened yet; others attribute any perceived warming to natural climate variability. The honest truth is no reputable scientist is 100% confident that global warming has occurred, or that it will ever occur. But let’s play devil’s advocate for the moment and discuss how these scenarios could change hurricane activity.

The global average of 86 tropical storms has probably not changed in the last few decades (Landsea 1998a). However, some ocean basins do experience 10-20 year cycles where the total number of storms changes. For example, since 1980 the total number of tropical storms has increased in the Northwest Pacific, but this increase was preceded by a nearly identical decrease from about 1960 to 1980. A downward trend in tropical storm number since the mid-1980s for the Australian region has been observed, but this is probably an artificial decrease due to a change in tropical cyclone wind designation by that country in the mid-1980s.

As discussed earlier, the number of Atlantic tropical storms has substantial year-to-year variability. However, no significant trend in tropical storm number has been observed since 1944 (Landsea et al. 1996). In contrast, the number of major hurricanes (Category 3 or better) has shown a significant downward trend during this period, corresponding with an active period of intense hurricanes during the 1940s to 1960s which suddenly changed to a quiet period of intense hurricanes during the 1970s to 1990s. (Such shifts in weather activity are known as multidecadal changes). This change is likely due to a cooling in the Atlantic Ocean water temperature during this period, and may also be related to a drought which started in 1970 in Africa where many tropical waves form before propagating out into the Atlantic.

The early 1990s were a particular inactive period in the Atlantic. No hurricanes were observed over the Caribbean Sea during the years 1990-1994 - the longest period of lack of hurricanes in the area since 1899. 1991-1994 is the quietest four-year period on record since 1944 (in terms of frequency of: total storms - 7.5 per year; hurricanes - 3.8 per year; and major hurricanes - 1.0 per year). However, one major hurricane making landfall during this period was Andrew (1992), which caused records amount of damage in Miami.

Because Andrew, a Category 4 hurricane, hit a metropolitan area, it attracted much attention, including claims that global warming must have caused such a powerful storm because of the unprecedented damage (Leggett 1994; Newsweek 1996; U.S. Senate Bipartisan Task Force on Funding Disaster Relief 1995; Dlugolecki 1996). However, it is normal to have 2 major hurricanes each year (Table 1). In fact, it is unusual to have only 1 major hurricane per year during a 4-year period! Andrew just happened to hit a highly developed coastal region.

Therefore, claims of global warming increasing hurricane activity or intensity in the last few decades are incorrect. In fact, during the last 50 years the overall trend has been a decrease in major hurricanes during this period of increased fossil fuel emissions. If global warming is already occurring (and no one knows this for sure yet), thus far it definitely has not increased hurricane activity.

The moral is that a reader should view most sensational science reports with some skepticism. Stories about global warming (and most sensational science stories in general) need not necessarily be dismissed as incorrect, but they should be carefully scrutinized. Questions one should ask when preparing a science report, or trying to learn about a particular science topic, are:

1. Who is the source of the report? For example, is it in a scientific magazine written by an author with a science degree from college, or in a newspaper/news/TV report by an author with no background in the sciences? Just because an article is in popular media outlet doesn’t mean its shoddy, but it does depend on the quality of the journalist, his interview procedures, and scientific aptitude. Some writers work hard to produce an accurate, informative article. On the other hand, many scientists can describe instances where a journalist inaccurately reported their information, or had a quote taken out of context.
2. Was the report peer-reviewed? The most reputable reports are in journals which undergo a peer-review process. The peer-review process involves rigorous commentary by several experts in the field who either accept or reject the article based on its scientific merit. Usually several rewrites are needed. Examples of general science journals available at bookstores and newsstands which at least undergo some minimal peer review are Scientific American, Nature, and Science. Examples of peer-reviewed journals in meteorology are Monthly Weather Review, Weather and Forecasting, Journal of Climate, and the Bulletin of the American Meteorological Society. The latter are all published by the American Meteorological Society (see Chapter 5.
3. Is it a well-balanced report presenting all the issues, or is it strongly biased towards one point of view? Often a journalist, or a magazine funded by a particular special interest group or political party, will strongly skew a report towards their point of view. These reports may have legitimate information, but one should investigate other sources before jumping to conclusions. To assess whether a report is biased, simply look at other reports by the same reporter or magazine to see if similar topics are reported in the same context. Even peer-reviewed articles can contain biases.

In summary, one should be skeptical and inquisitive about most controversial or sensational scientific information they encounter, and should always obtain their information from more than one source.

Research has shown that global warming could increase tropical sea surface temperatures and tropical rainfall. Based on this research, some have suggested hurricanes may increase in frequency, area of occurrence, and intensity (Ryan et al. 1992; Emanuel 1987). However, any changes in hurricane activity will also be associated with large-scale changes in the tropical atmosphere (Landsea 1998a; Henderson-Sellers et al. 1998). For example, the instability threshold necessary for thunderstorm maintenance could change, since temperature may increase throughout the atmosphere, not just at the surface.

In addition, any changes in global wind patterns will profoundly affect hurricane activity. For instance, if global warming increases wind shear, one would see a significant decrease in hurricane activity. Likewise, a reduction in wind shear would dramatically increase hurricane activity. If monsoon activity increases, then tropical cyclogenesis will increase, and vice versa for weakened monsoon troughs. Another wildcard is how global warming would change El Niño activity --- if more El Niños occur, hurricane activity would be reduced, and vice versa for fewer El Niños. In summary, the combined changes in water temperature, wind shear, monsoon activity, atmospheric instability, and El Niño will dictate how hurricane activity changes. Therefore, it is difficult to assess how potential global warming could alter hurricane activity. Besides, there is considerable motivation for society to prepare better for hurricanes independent of global arming concerns!

Because of their destructive and life-threatening nature, experimental attempts have been made to weaken hurricanes. The main hypothesis involved converting liquid cloud water to ice just outside the eyewall. Water gives off enormous quantities of stored heat (also called latent heat) when it changes phase from liquid to ice. Many clouds exist in a supercooled state, which means the liquid droplets’ temperature is below freezing (32F), but lack a "triggering" mechanism to turn to ice. For this conversion to occur, supercooled liquid water needs to attach to a floating aerosol with a molecular structure similar to ice, known as ice nuclei. However, ice nuclei are often sparse in the atmosphere, and many supercooled droplets are never converted to ice. If somehow one could introduce artificial ice nuclei (such as silver iodide) into a supercooled cloud, the water would be converted to ice, thus releasing lots of heat into the air and causing the cloud to grow. This premise, known as cloud seeding, has been attempted to increase rainfall in drought stricken-regions (with unproven results) and snowpack in the mountains (with successful results).

To theoretically weaken a hurricane, the seeding process is more complicated. At first scientists thought they could seed the eyewall and perturb the hurricane’s wind outwards, thus weakening the hurricane (Anthes 1982; Willoughby et al. 1985). By the mid-1960s scientists realized this theory was flawed. A revised theory evolved in which one seeds the clouds just outside the eyewall to stimulate cloud growth away from the eyewall. The new outer eyewall would grow, depriving inflow into the older, inner eyewall. The result is a weakening inner eyewall, resulting in less subsidence in the eye and a rise in central pressure. If the pressure increases, inflowing air is unable to penetrate to as small a radius, and most of the new ascent occurs at the new outer eyewall. Eventually the new eyewall would replace the old eyewall, but at a larger distance from the center. Just as ice skaters slow their rotation when their arms are spread out, a larger eyewall radius would cause a reduction in wind speed.

Cloud seeding was first tested when several U.S. government agencies collaborated in a pioneering weather modification effort known as Project Cirrus (Willoughby et al. 1985). Among other notable firsts was the first cloud seeding of a hurricane. On October 13, 1947 a plane dropped silver iodide into a hurricane moving to the northeast. Observers on the plane noted changes in the visual appearance of the cloud, but could not demonstrate any changes in structure or intensity. However, shortly afterwards the hurricane reversed course to the west, making landfall on the coasts of Georgia and South Carolina. It is extremely unlikely the seeding altered the course, since hurricanes are mostly guided by constantly shifting large-scale atmospheric currents. However, the political and legal implications taught scientists to be more careful about where they conducted their hurricane seeding experiments. Future attempts at hurricane modification occurred only in hurricanes which were: 1) far from all land masses; and 2) unlikely to make landfall within 24 hours.

The next experiment occurred on September 16, 1961 when a naval aircraft dropped eight canisters of silver iodide into Hurricane Esther. Esther, which has been intensifying, stopped strengthening and the eyewall’s distance from the center increased. The next day, a second seeding attempt was made, but the canisters missed the eyewall and the hurricane’s intensity did not change. At the time, the experiment was considered successful, although we now know the eyewall-seeding hypothesis was flawed and that Esther’s changes must have occurred naturally. The encouraging results from Esther led to the formal establishment of a hurricane modification program known as Project STORMFURY in 1962.

STORMFURY was a collaboration between National Oceanic and Atmospheric Adminstration (NOAA) and the U.S. Navy, and directed by the National Hurricane Research Project (now called the Hurricane Research Division). STORMFURY started out well when Hurricane Beulah’s eyewall was seeded on August 24, 1963. The eyewall disintegrated, followed by formation of a new eyewall at a larger distance form the eye. The maximum winds decreased by 20% and moved further away from the center. STORMFURY seemed to have a promising beginning.

Four more years were to pass before the next modification experiment. The years 1964-1968 were generally inactive hurricane seasons, and the hurricanes which did occur were either too close to land or out of flight range. Scientists also realized that the eyewall-seeding hypothesis was incorrect, resulting in the revised hypothesis of seeding outside the eyewall. The modified hypothesis was tested on Hurricane Debbie on August 18 and August 20, 1969 when more than a thousand seedings of silver iodide were made each day. The eyewall shifted outwards each day, and the winds decreased by 31% and 15%, respectively.

Ironically, this was essentially the end of Project STORMFURY (Willoughby et al. 1985). The 1970 hurricane season yielded no suitable candidates for seeding. In 1971, the only eligible storm was Hurricane Ginger, a late-season, diffuse system. It was seeded twice, but was a poor candidate because it lacked a small, well-defined eye; the seeding had no effect on Ginger. In 1972, all the storms were too weak, too close to land, or out of flight range. In fact, in general the 1970s were a period of below average hurricane activity, especially for intense hurricanes. Other difficulties also plagued STORMFURY. The Navy ended its support to pursue goals more closely related to national defense. Several of the aircrafts had become too old for reliable use anymore. Permissions for seeding from Caribbean countries and Mexico became more difficult to obtain, and the State Department increased the time restriction for seeding before landfall (Posey 1994). Eventually, only a narrow zone north of Puerto Rico in which hurricanes were at least 36 hours from landfall was allowed for seeding. Unfortunately, no storms ever passed through the small permissible trapezoid of ocean north of Puerto Rico between 1973 and 1979. STORMFURY scientists attempted to move the project into the Pacific where storms are more numerous, but Japan and Australia blocked that move. Finally, weather modification experiments in general were falling into public disfavor (Pielke and Pielke 1997). In 1983, Project STORMFURY was terminated.

Also, from the very beginning of STORMFURY, and throughout the project’s lifetime, several scientists expressed concerns about the seeding hypothesis, as well as the interpretation of STORMFURY’s experiments. These doubts became substantiated by additional reconnaissance observations in the 1970s and 1980s, and by increased knowledge of hurricane structure and evolution during this period. These findings, which refute the basic premise of project STORMFURY, are summarized below (Willoughby et al. 1985):

1. Observational evidence shows that hurricanes actually contain too little supercooled water and too much natural ice for seeding to be effective.
2. Hurricanes go through a natural (but temporary) weakening process in which a new eyewall forms outside the original eyewall. The outer eyewall "chokes off" inflow to the inner eyewall, causing it to dissipate. The outer eyewall then propagates inwards, replacing the original eyewall. This process, known as the concentric eyewall cycle, lasts 12 to 36 hours, and is associated with temporary weakening (Willoughby et al. 1982). Therefore, much of the observed weakening in Esther, Beulah, and Debbie may be the result of natural internal evolution rather than a consequence of seeding. Hurricane intensity is also controlled by external influences, such as water temperature and wind shear. For example, some evidence exists that Debbie moved into a strong wind shear environment conducive to weakening (P. Black, personal communication 1994). Therefore, the expected results of seeding are indistinguishable from naturally occurring intensity changes.

Project STORMFURY should not be viewed as a failure, however. Even though only a few seeding experiments were performed, many reconnaissance flights were conducted into hurricanes to understand their formation, structure, and evolution. The legacy of Project STORMFURY is a wealth of observations which has augmented our understanding of hurricanes, and have improved hurricane forecasts. For more information about Project STORMFURY, the reader is referred to papers by Posey (1994) and Willoughby et al. (1985).

The notion of hurricane modification still remains though (Posey 1994). Occasionally, someone will ask, "Why don’t we just nuke a hurricane and destroy it?" Radioactivity aside, such a question demonstrates an extreme underestimate of hurricane power. For example, Hurricane Andrew (1992) generated the equivalent energy of a ten-megaton bomb continuously during its existence --- not in a split second as in a bomb explosion! It is doubtful a hurricane would even be affected by a nuclear blast, other than spewing radioactive fallout throughout the region. Other ideas involve removing a hurricane’s energy source. Some have proposed covering the ocean ahead of a hurricane with an impermeable chemical film which impedes evaporation from the sea, thus weakening the hurricane as it moves into that region (Posey 1994). This is unrealistic, though, because it is unlikely any surface film could withstand the 30 to 50-foot ocean waves in a hurricane, and covering hundreds of square miles of ocean with a substance is a formidable task. Others have promoted stimulating cloud growth in the outer core of a hurricane by increasing the surface temperature with carbon black, which possibly would restrict inflow to the hurricane’s inner core (Gray et al. 1993); however, such a scheme would suffer the same obstacles as Project STORMFURY. Besides, both schemes pose environmental problems. The wisest course of action is to reduce societal exposure to hurricane impacts, not through modification attempts. Pielke and Pielke (1997) offer several recommendations to improve hurricane preparedness policies.

At this point, a reader may ask, "Meteorology sounds like a fascinating field. How do I obtain employment in this profession?" Meteorology certainly is an exciting field. Meteorologists work with state-of-the-art computers and technology in a non-stagnant arena, because the weather changes every day for forecasters and new unexplored research opportunities are always emerging. Employment opportunities include weather forecasting for the National Weather Service and other government agencies, as well as broadcast television/radio and private companies. Scientists perform research on a variety of unresolved issues such as severe thunderstorms, tornadoes, hurricanes, global warming, climate change, and air pollution. Scientists generally work at universities and governmental agencies. Meteorologists with a particular specialty may also be consultants for companies.

Obtaining a Bachelor of Science degree in meteorology requires dedication and hard work. A strong background in science, math, and computers is a must. High school students should take as many science, computer programming, and math classes as possible, including Calculus and Physics. Good grades (As and Bs) are essential in all technology-driven fields, including meteorology. Children should be exposed to computers as early as possible. In addition to a bachelor’s degree, certain aspects of meteorology require graduate school education. Most researchers must have at least a Master of Science degree, and a Ph.D. is preferred. Some advanced forecast applications also require advanced degrees. For example, forecasters at the National Hurricane Center need a Master’s or a Ph.D.

Table 10 lists universities in the United States that offer meteorology degrees or degrees closely related to meteorology (i.e., environmental science, theoretical atmospheric physics, broadcast meteorology, etc.). This list is adapted from the "1998 Curricula in the Atmospheric, Oceanic, Hydrologic, and Related Sciences" published biannually by the American Meteorological Society. This publication may be ordered at a price of $40 (including shipping and handling) from:

American Meteorological Society

45 Beacon Street

Boston, MA 02108-3693

ISBN 1-878220-28-4

(617) 227-2425

http://www.ametsoc.org/AMS

It is important one contacts several schools to decide what is best for the student, because different institutions have widely varying specialties. For example, a student interested in weather forecasting would want to attend a program with a strong weather analysis and forecasting program. Another student interested in global warming may want to attend a school which has a program in climate change or environmental sciences. Some programs offer a "concentration" or "minor" in meteorology, with the actual Bachelor’s degree in another field. A student interested in basic research of weather phenomenon (i.e., severe thunderstorms, tornadoes, hurricanes, etc.) would want to attend a school with a broad theoretical background of the atmosphere and also has professors who specialize in the specific weather phenomenon.

It again needs to be emphasized that research is conducted at the graduate school level. A student must first earn a Bachelor’s degree in a science field, then apply to a graduate school and be accepted. Competition is stiff for acceptance into graduate school, and good grades as an undergraduate are essential. It is extremely helpful to get the Bachelor’s degree in meteorology, but it is not a requirement. Many graduate schools accept applicants from students with a Bachelor’s degree in math, physics, computer science, or engineering.

Ahrens, C. D., 1994. Meteorology Today. West Publishing Co. 591 pp.

American Meteorological Society, 1998. Curricula in the Atmospheric, Oceanic, and Related Sciences. [Available from the American Meteorological Society, 45 Beacon Street, Boston, MA 02108-3693; (617) 227-2425; http://www.ametsoc.org/AMS]

Anthes, R., 1982. Tropical Cyclones: Their Evolution, Structure, and Effects, American Meteorological Monographs, Volume 19, American Meteorological Society, Boston, MA, 208 pp.

Corps of Engineers, 1970. Hurricane Camille --- 14-22 August 1969. U.S. Army Engineer District, Mobile, Alabama, 80 pp.

1998 Curricula in the Atmospheric, Oceanic, Hydrologic, and Related Sciences ---- Colleges and Universities in the United States and Canada, American Meteorological Society, Boston, MA, 567 pp.

Danielson, E. W., J. Levin, and E. Abrams, 1998. Meteorology. WCB/McGraw-Hill Co., 462 pp.

DeMaria, M., and J. Kaplan, 1994: Sea surface temperature and the maximum intensity of Atlantic tropical cyclones. J. Climate, 7, 1324-1334.

Dlugolecki, A. F., Ed. 1996: Financial Services. Climate Change 1995: Impacts, Adaptations, and Mitigation of Climate Change: Scientific-Technical Analyses. Cambridge University Press, 539-560.

Dvorak, V. F., 1975: Tropical cyclone intensity analysis and forecasting from satellite imagery. Mon. Wea. Rev., 103, 420-430.

Emanuel, K. A., 1987: The dependence of hurricane intensity on climate. Nature, 326, 483-485.

Fujita, T. T., 1993: Wind fields of Andrew, Omar, and Iniki, 1992. Preprint from the 20^th Conference on Hurricanes and Tropical Meteorology, American Meteorological Society, 46-49.

Goldenberg, S., and L. Shapiro, 1996: Physical mechanisms for the association of El Niño and West African rainfall with Atlantic major hurricane activity. J. Climate, 9, 1169-1187.

Gray, W. M. and W. M. Frank, 1993: Hypothesis for hurricane intensity reduction from carbon black seeding. Preprint from the 20^th Conference on Hurricanes and Tropical Meteorology, American Meteorological Society, 305-308.

Gray, W. M., J. D. Sheaffer, and C. W. Landsea, 1996: Climate trends associated with multi-decadal variability of intense Atlantic hurricane activity. Chapter 2 in "Hurricanes, Climatic Change and Socioeconomical Impacts: A Current Perspective", H. F. Diaz and R. S. Pulwarty, Eds., Westview Press, 49 pp.

Gray, W. M., C. W. Landsea, P. W. Mielke Jr, and K. J. Berry, 1998: Extended range forecast of Atlantic seasonal hurricane activity and U.S. landfall strike probability for 1999. Department of Atmospheric Science Paper, Colorado State University (Also available at http://tropical.atmos.colostate.edu).

Henderson-Sellers, A., H. Zhang, G. Berz, K. Emanuel, W. Gray, C. Landsea, G. Holland, J. Lighthill, S-L. Shieh, P. Webster, and K. McGuffie, 1998. Tropical Cyclones and Global Climate Change: A Post-IPCC Assessment. Bull. Am. Meteor. Soc., 79, 19-38.

Holland, G. J., 1993. In: Global Guide to Tropical Cyclone Forecasting. World Meteorological Organization Technical Document, WMO/TD No. 560, Tropical Cyclone Programme, Report No. TCP-31, Geneva, Switzerland. Also available on the web at http://www.bom.gov.au/bmrc (look for additional links under the mesoscale section to find this book).

Knaff, J., 1997. Implications of summertime sea level pressure anomalies in the tropical Atlantic region. Mon. Wea. Rev., 10, 789-804.

Landsea, C., 1998a: Climate Variability of Tropical Cyclones: Past, Present, and Future. Submitted to WMO and Global Response to Storms. Edited by R. A. Pielke Sr., and R. A. Pielke Jr.

Landsea, C., 1998b: Frequently Asked Questions: Hurricanes, Typhoons, and Tropical Cyclones (available at http://www.aoml.noaa.gov/hrd/tcfaq/).

Landsea, C. W., N. Nicholls, W. M. Gray, and L. A. Avila, 1996: Downward trends in the frequency of intense Atlantic hurricanes during the past five decades. Geo. Res. Letters, 23, 1697-1700.

Leggett, J. Ed., 1994: The Climate Time Bomb, Greenpeace International, Amsterdam.

LePore, F., 1996: Interview with Earth and Sky.

Nese, J. M., and L. M. Grenci, 1998. A World of Weather: Fundamentals of Meteorology. Kendall/Hunt Publishing Co., 539 pp.

Newsweek, 1996: The HOT ZONE: Hurricanes, floods, and blizzards: Blame global warming. Shown on the January 21 magazine cover.

Newton, D. E., 1993. Global Warming, ABC-CLIO, 183 pp.

NHC, 1998: National Hurricane Center homepage, http://www.nhc.noaa.gov

Pielke Jr., R. A., and R. A. Pielke Sr., 1997. Hurricanes. Their Nature and Impacts on Society. John Wiley & Sons. 279 pp.

Pielke Jr., R. A., and C. W. Landsea, 1998. Normalized Atlantic hurricane damage: 1925-95. Wea. Forecasting, 13, 621-631.

Posey, C., 1994: Hurricanes --- Reaping the whirlwind. Omni, 16, 34-47.

Rosenfield, J., 1997: Storm Surge! Hurricanes' Most Powerful and Deadly Force. Weatherwise, 50, 18-24.

Ryan, B. F., I. G. Watterson, and J. L. Evans, 1992: Tropical cyclone frequencies inferred from Gray's yearly genesis parameter: Validation of GCM tropical climates. Geo. Res. Letters, 24, 1255-1258.

Simpson, R. H., 1974: The hurricane disaster potential scale. Weatherwise, 27, 169, 186.

U.S. Department of Commerce, 1969. Hurricane Camille, August 14-22, 1969 (Preliminary Report). Environmental Science Services Administration. Weather Bureau. 58 pp.

U.S. Senate Bipartisan Task Force on Funding Disaster Relief, 1995: Federal Disaster Assistance. 104-4.

Willoughby, H. E., and R. A. Black, 1996: Hurricane Andrew in Florida: Dynamics of a disaster. Bull. Am. Meteor. Soc., 77, 543-549.

Willoughby, H. E., D. P. Jorgensen, P. G. Black, and S. L. Rosenthal, 1985: Project STORMFURY: A scientific chronicle 1962-1983. Bull. Am. Meteor. Soc., 66, 505-514.

Willoughby, H. E., J. A. Clos, and M. G. Shoreibah, 1982: Concentric eyewalls, secondary wind maxima, and the evolution of the hurricane vortex. J. Atmos. Sci., 41, 1169-1186.