Wind’s Intricate Patterns - Windswept: The Story of Wind and Weather - Marq de Villiers

Windswept: The Story of Wind and Weather - Marq de Villiers (2006)

Chapter 4. Wind's Intricate Patterns

Ivan's story: By the afternoon of September 3 the system had moved almost 300 miles farther into the tropical Atlantic, and was located 863 miles south southwest of the Cape Verde Islands. Around it and above it the air was still, but the system itself was still spinning slowly, moving just south of west at about 30 miles an hour. Strong winds stretched outward about 60 miles from the center. Within a few hours the winds picked up and passed the 39-mile-an-hour threshold that turned the system officially from a tropical depression into a tropical storm. A bulletin from the National Hurricane Center in Miami assigned the system the name Ivan, according to custom and its place in the annual cycle, and found it was strengthening steadily. By five P.M. on the 3rd, Bulletin 3 found that the storm's central pressure had dropped to 1,000 millibars, and maximum sustained winds were in the region of 30 miles an hour, with higher gusts. This was still well shy of the 74 miles an hour of a hurricane, but the bulletin was cautious: Strengthening was expected, and it could turn westward onto a new track.

In the small print of the bulletin, issued under the signature of Forecaster Beven, there were hints of what Miami thought would happen. "Thought," because they didn't know; winds are predictable in their larger patterns and behaviors, but horribly intricate in their local behavior. Aircraft data had found flight level winds of g3 knots near the center.1 Beven expected the system to turn slightly north and reach hurricane status within two days. Further ahead than 36 hours, he wrote, the forecast was of low confidence.

He expected Ivan would head for the Gulf of Mexico, but the system could still go anywhere, do anything.

That bulletin was the first time Ivan impinged on my consciousness. I logged on to the National Hurricane Center Web site and squinted along the predicted track. It looked likely that Ivan could go to Puerto Rico and then Cuba, and after that southern Florida and the Gulf. It didn't appear dangerous for the northeast corner of America. But that "low confidence" niggled. The northern edge of the track forecast would take the storm north of the Bahamas, where it could curve northward, as most storms do under the influence of the Coriolis force, and head for Bermuda. That could be Trouble.

At least for me. I confess I was rather less concerned at this point about Florida than I was about the storm showing up in the northern Atlantic and bashing away at my beach. I figured Florida was more able to cope with Trouble than I was. NIMBY, Not in My Back Yard, yes. Sorry.

I

Sometimes you really can see the wind. Directly, not just by watching trees sway or ripples skitter across open water. Last year we had a two-day blizzard that dropped almost a yard of snow in gusty 50-mile-an-hour winds. At its height, you could stare out the window and watch the white wind dipping and swirling through the trees, heaping the snow up in great sculptured drifts, darting around the corner of the house. We had often wondered why the greatest damage in wind storms seemed to be done to a row of rhododendrons on the lee side of the house, and suddenly we could see the reason for ourselves—the wind dashed itself at the corner of the building and split, some of it going straight up and tugging at the eaves but the rest swirling around the corner, where it was squeezed between the house and a cut in the bank, suddenly accelerating just before it passed the rhodos, tearing at their leaves. I had already "seen" the wind in the violent dust storms of my boyhood in South Africa, where the prevailing color was not white but a brutal violet, and later in Saharan sandstorms, where the wind blows a gritty ochre and dun, and of course tornadoes are lethally visible. But these are crude views, made visible only by wind's awful force. And so when I read a lyrical passage in Sebastian Smith's sailing memoir, Southern Winds, I found myself nodding in agreement: "Sometimes, staring into the sky, I tried to imagine what it would be like to see the wind. As if there might be special glasses to understand [the wind's] secrets—the paths used to reach seemingly impossible places, the mystery of opposing airs and the drama of the katabatic wind. In simulations these can be created but to ordinary men they remain as inscrutable as any god … Imagine standing on deck and being able to watch jet streams swoop through the summer skies. Or the slow turning wheel of low-pressure systems. Or the avalanche of a squall."2 Just so. How wonderful to be able to watch the wind's intricate dance across the planetary surface and high into the troposphere, to see how great air masses move and collide and meld and wrestle, how little zephyrs tickle the fine hairs on the forearm, how the breezes tease their way through trees and over rocks and up hillsides. Sometimes airplanes hit an air pocket—which is really vertical-shear wind plunging violently downward through pressure differentials in the surrounding air. Wouldn't pilots, and their passengers, love to see that wind before they hit it?

We can understand wind—we understand it now almost to the molecular level—but it would be much more agreeable, and infinitely more useful, to be able to see it.

Find starts, of course, as with so much else on our planet, with the sun.

Deep inside the seething cauldron of the sun, thousands of miles beneath the corona, or what passes for its surface, is a constant cascading series of hydrogen fusion reactions. Every nanosecond millions of hydrogen atoms crash together, and for every four that destroy themselves in this furious suicide, one helium atom is created. Since four hydrogen atoms weigh slightly less than one helium atom, a fraction of the mass is lost each time, and this shortfall—pace Einstein—is released as pure energy.3 So much energy, indeed, that the temperature of the sun is maintained at a fairly steady 15 million degrees Celsius. Some of that energy is radiated into space. A tiny fraction, about two billionths, reaches the earth.4 This might not seem a lot, but the sun is so massive relative to Earth that the solar radiation reaching us is around 175 trillion kilowatt hours of energy every hour.5 Somewhere between 1 and 2 percent of this energy input is converted into wind, about fifty to one hundred times more than all the energy converted into biomass by all living things on Earth.

These trillions of watts of solar energy strike the earth directly at the equator and more obliquely closer to the poles. This was far from obvious in ancient days, when the earth was a flat disk and the sun directly overhead, but to us the mechanism is obvious—equator at high flame, midlatitudes at a simmer, poles just barely affected, a straightforward consequence of Earth's spherical shape, a simple pattern complicated only by Earth's rotation on its rather tilted axis and its annual rotation about the sun. Pretty obviously, the air is therefore hotter at the equator and cooler at the poles, and it is in these differentials that all winds, and all weather, and therefore climate, are derived.

The differentials are set in motion by one simple governing principle: entropy. Entropy, disorder—or "mixed-upness," as Richard Dawkins called it—is the substance of the second law of thermodynamics, which is that entropy, or disorder, always increases in a closed system. In this definition, order means that different parts of a system have different characteristics (heat, pressure, odor); disorder means no part is different from any other. In nature, if something is hot and something is cool, reordering will occur. That is, if a zone of high pressure is near a zone of low pressure, nature will try to equalize the two zones through movement of air from the high to the low, and winds result. Nature is striving for balance; climate—derived from the Greek word klima, meaning degree of latitude—is the earth's way of seeking to balance its energy intake.

This is just as well. If this didn't happen—if entropy wasn't at work, if there was no balancing—there would be no wind, no weather, and no life on earth as we know it. The poles would go into a much deeper freeze, the equatorial zones would overheat, and what little organic life remained would be huddled at the interstices.

And so, in some ways winds, the movement of the air relative to ' the surface of the earth, are simplicity itself: Hot air one place, cooler air another, and there you have it—wind. Pressure differentials— wind. Adjacent climate zones—wind. Altitude differentials— wind. Planetary rotation—deflected wind. The physics is not complicated: Wind is air moving from high pressure to low pressure, in a straight line, deflected by the rotation of the earth (the Coriolis force).

Because winds begin with the sun, the key to understanding global wind patterns is to start where solar radiation is most intense, at the equator. The air warmed by the radiation rises quickly, causing a quasi-vacuum of low pressure that draws air toward the equator from semi tropical latitudes. The winds so created head directly for the equator but because of the planet's rotation are turned by the same Coriolis force that twists the ocean's currents. This "turn" pushes the winds "right" in the northern hemisphere and "left" in the southern hemisphere, until they parallel the equator. These are the most reliable winds on earth, the so-called trade winds; they are the winds that made transoceanic travel possible in the days of sail. Eventually these steady trade winds, because they are paralleling the equator, are also warmed, and they then follow the same pattern— they rise, are cooled, drift toward the poles, and sink again, at about 30 degrees of latitude, more or less the southern Mediterranean and northern California.

Some of this newly cooled air moves at high altitudes back toward the equator, completing what is known as the Hadley cell, named after George Hadley, an English lawyer in the eighteenth century. But some of it moves toward the midlatitudes, and on its way the Coriolis force "turns" the wind right in the northern hemisphere and left in the southern, causing the prevailing midlatitude westerlies, the winds that the canny New Englanders learned to exploit when their trade with Europe began to expand. But the westerlies only account for a fraction of the air mass in motion. The rest, which forms a second overturning cycle of air, carries surface air poleward and upper-level air back toward the Hadley cells. These are Ferrel cells, which generally have a motion opposite to planetary rotation; they exist largely to balance the Hadley cells and their equivalents at the poles.6 Ferrel cells and Hadley cells meet at the horse latitudes.

This intricately three-dimensional pattern of winds is the planet's general circulation. It explains why most of the air movement in the lower atmosphere is vertical rather than horizontal. It is also directly responsible for the latitudinal movement of air masses, and therefore of weather and the long drifts of weather called climate.

II

This business of the Coriolis force, named after French physicist and mathematician Gustave Gaspard Coriolis, who first described it in 1835,7 merits a small digression, because it is not quite as simple as it sounds. The physics that govern what scientists call a rotating frame of reference are quite complex, but the effects of the earth's rotation on natural phenomena are simple enough to see. East-west motions create a Coriolis force, or Coriolis effect, which is directed radially inward (for easterly motion) or outward (for westerly motion) from the axis of rotation. Motion relative to the earth's, to the west or the east, will produce an acceleration (because of the force) to the north or the south.

Hadley and Ferrel cells, showing a simplified version of the major vertical air movements that help balance planetary heat distribution.

North or south motions also give rise to a Coriolis force because the motions are toward (or away) from the axis of rotation. Vertical motions also give rise to a horizontal Coriolis force, but it is negligible and usually ignored.

For our purposes, the result is what is important: Whether from east-west or north-south motions, there is always a deflection to the right in the northern hemisphere, and to the left in the southern hemisphere.

The easiest commonsensical way to see the effect is to imagine firing a rocket or an artillery shell while you are standing at the equator. Aim the rocket to hit a target a thousand miles away. If you fire it toward the north pole, something odd seems to happen. Even if your calculations were spot on, and there was no wind, the rocket wouldn't land due north of where you are. Instead, it will appear to have drifted off course, to the east. Surely something must have diverted it? Was Newton wrong?

Gustave Coriolis

The answer lies in the rotation of the earth. All points on Earth rotate 360 degrees in twenty-four hours, a planetary day, but obviously some points must be rotating much faster than others. The fastest speeds are at the equator, perhaps a thousand miles an hour; close to the poles the rotational velocity is negligible, a mere dozen or so miles an hour.

So what has happened to your rocket is this: At the moment you fire it in a northward direction, you and the launcher that fired it, and therefore the rocket itself, are traveling eastward at a high rate of speed—the speed the equator is traveling. Your rocket travels in a straight line, but throughout its flight it keeps moving eastward at a constant rate, the speed of the equator. When it nears its destination, however, the ground below it is not moving eastward nearly as fast as the rocket itself. And so the ground "seems" to have moved westward, and the missile "seems" to have drifted eastward. That is, an object traveling away from the equator will eventually be heading east faster than the ground below it, and may seem to have been driven east by some unknown force. Objects traveling toward the equator, similarly, will seem to be have been driven west. Which means, if you can turn your head to squint in the right direction, that in the northern hemisphere objects will turn to the right, and in the southern hemisphere to the left.

South Pole

The Coriolis force, showing how movement is deflected to the right in the northern hemisphere and to the left in the southern hemisphere.

Newton's law is conserved, after all. It is just that we are in a rotating frame of reference.

Some of the effects are critical ones, and not just for long-range ballistic missiles. Over the course of a four-hour flight, a jetliner traveling north or south must compensate for the Coriolis effect, or a pilot from Toronto will find himself depositing his passengers in New Orleans instead of Atlanta.

On a very small scale, say that of the average bathtub or kitchen sink, the angle of deflection is so tiny as to be unnoticeable, estimated at i/3600th of a degree, so the way water flows out of a bathtub depends on the angle it was poured in, the configuration of the tub itself, abrasions on the surface, the temperature differentials in the water, and dozens of other minute effects. It is theoretically possible, in a carefully controlled laboratory experiment, to show the effect at work; most of us, to see it at all, would need a tub the size of one of the Great Lakes.8

I once saw a "demonstration" of the Coriolis force by a cheerful Maasai at the Equator near the town of Eldoret, in Kenya. I had come to Eldoret after a rather harrowing all-night, breakneck drive in an aging Renault, piloted by a former game warden; we shared the car with a massive python, which the game warden called Brenda, that creaked and rustled around in the backseat and kept me awake. I was therefore feeling somewhat fragile when the car stopped at the "Line," where we joined a busload of tourists who were having their pictures taken with one leg in each hemisphere, and who were being subjected to the aforementioned Coriolis demo. Two buckets had been filled with water and placed a yard on either side of the line. Then the demonstrator, with a flourish, pulled the plug on each in turn, and lo and behold, the water exited by spinning counter clockwise in the northern hemisphere bucket, and clockwise in the southern. This produced the requisite "ooohs" and "aaahs" from the busload, as well as a satisfactory yield in tips, but it was all a fraud, set up by the direction in which the con artist had poured the water into the buckets, and by a minute difference in the way he pulled the plugs. The game warden, whose name was Ibrahim, gave him a disgusted snort for his pains, and so did I, but he wasn't offended; he was making a good living from human gullibility and the occasional skeptic wasn't going to deter him.

The Coriolis effect influences massive movements of air just as much as it does artillery shells or aircraft, and is therefore important for an understanding of global winds. Just like an artillery shell, freely moving air (winds) will deflect to the right in the northern hemisphere, and to the left in the southern. So air moving toward a low-pressure system will deflect to its right; but because the forces which got the air moving toward the north in the first place are still in play, the result will be a vortex of air, spinning counterclockwise— air will try to turn to the right, the low pressure pocket will try to draw the air into itself, and the result is that the air is held into a circle that actually turns to the left. Without the Coriolis effect, air rushing into a point could still form a vortex, but the direction of rotation would be random.

With the Coriolis effect in play, the randomness disappears, and northern hemisphere cyclones, including hurricanes, always revolve in the same direction.

Bear in mind that the Coriolis force is not the only factor determining large scale winds. One force is generated by the pressure gradient, of course, but there are three others: the curvature of the wind (wind that turns speeds up or slows down depending on whether it is turning clockwise or counterclockwise [the opposite in the southern hemisphere]); changing pressure differences (called the isallobaric effect), which can dramatically boost or inhibit the wind, a major consideration in east coast "weather bombs"; and the stability of the air in the lowest part of the atmosphere (stable air, either warm or cold, is reluctant to overturn, whereas unstable air is highly turbulent, and its overturning and mixing brings down the established winds from higher elevations to where we live).

III

The general circulation model—the Hadley and Ferrel cells and the generally vertical overturnings of huge air masses—is important for understanding global winds, but since most of us see winds as horizontal, not vertical, it's more useful too look at planetary winds in another way: at how solar radiation and the Coriolis force together conjure into being a consistent pattern of latitudinal "belts" paralleling the equator.

Starting from the equator and going north, these belts are the doldrums, the trade winds, the horse latitudes, the prevailing south-westerlies of midlatitudes, and the northeasterlies of high polar latitudes. In the southern hemisphere, the same belts exist but the wind directions differ.

The doldrums straddle the equator and girdle the earth, a belt of low pressure, static and ever-present, a windless region known to all sailors. The doldrums are more properly known as the intertropical convergence zone (ICZ), or thermal equator, which generally occupies about 5 degrees each side of the equator, though it migrates somewhat north or south with the seasonal position of the sun. (It also shifts farther to the south over land masses such as South America, Africa, and Australia, and farther to the north over open water, such as the Pacific or Atlantic.9 It can occasionally reach beyond the 30th parallel.) Confusingly, it is sometimes called the equatorial convergence zone or the intertropical front.

The trade winds are next, bounded on the doldrums side by a zone of sharply rising winds, creating towering cumulonimbus thunderclouds and torrential rains. The trade winds flow from the next band out, the subtropical high-pressure zone called the horse latitudes, toward the low-pressure zone of the doldrums, and are "turned" westward by the Coriolis force. They were named, obviously enough, for their useful ability to push sailing vessels quickly and economically across oceans; they blow steadily at about 12 miles an hour between the ICZ and the 30th degree of latitude. In the northern hemisphere, the trade winds are northeasterlies; in the southern hemisphere, they are southeasterlies.

The horse latitudes are bands of calmer air, zones of stillness that have caused many a mariner to rue his profession. Maritime lore has several explanations for the name, all of them more or less implausible; the most popular is that ships became becalmed for so long in this zone that sailors and colonists were forced to toss their horses overboard, no longer having enough water to keep them alive. The horse latitudes are sometimes called the calms of Cancer in the northern hemisphere and the calms of Capricorn in the southern hemisphere. On land, most of the world's great deserts lie in this region.

Outside the horse latitudes are the prevailing westerlies (southwesterly in the northern hemisphere, northwesterly in the southern hemisphere) that cover most of Europe, North America, China, and similar latitudes south of the equator. They're as consistent—and as useful—as the trade winds. In the southern hemisphere they're closer to the equator and are more vigorous than their counterparts in the north, and are known as the roaring forties. In the northern hemisphere, this is where most of the North American and European weather is generated.

In high latitudes the flow reverses, and easterlies are the dominant wind patterns of both Arctic and Antarctic regions.

The most turbulent, changeable and perversely complicated weather on the planet occurs in the midlatitudes, where the warm equatorial air and the cooler polar air intersect in an apparently patternless turbulence. This is the downside of living in temperate regions, otherwise the most benign climate on Earth. The intersections of these winds—frontal zones—cause violent storms, tornadoes, thunder cells, and gales, as well as mild breezes and balmy temperatures. Mark Twain was moved to comment on the weather in northeastern America: "The people of New England are by nature patient and forbearing," he said in a speech to the New England Society in 1876, "but there are some things which they will not stand. Every year they kill a lot of poets for writing about 'Beautiful Spring.' These are generally casual visitors, who bring their notions of spring from somewhere else, and cannot, of course, know how the natives feel…"

The prevailing winds, showing the three major bands of steady planetary winds: the trade winds, the midlatitude westerlies, and the subpolar easterlies.

North Atlantic weather is complicated further by more or less permanent zones of low pressure around the 60 degree mark, the latitude of Greenland, and zones of high pressure in the horse latitudes, around the arid zones of the Middle East and the American southwestern deserts. In the Atlantic, weather forecasters call this the Bermuda high, and its presence, strength or weakness, is crucial for forecasting Atlantic weather; it can help "steer" hurricanes and other low-pressure systems. This high migrates east and west with varying central pressure. In summer and fall it is located near Bermuda; in the winter and early spring it is primarily centered near the Azores Islands, and then—surprise!—it is called the Azores high. The actual flow depends on a number of incalculables, including seasonality and jet-stream positioning. The circulating air tends to skirt the highs and get funneled into the lows, a tendency that is further complicated by topography (air flows more smoothly over water than over land, and less smoothly over mountains than plains), and even further complicated by the smoothly flowing stratospheric winds, that sometimes have the benign effect of shearing the tops off massive cyclones before they can coalesce into hurricanes.

Tropical cyclones such as hurricanes or typhoons usually decay as they reach cooler higher latitudes. But sometimes they decay in dangerous ways, interacting with the unstable baroclinic environment in higher latitudes and turning into what are called extratropical cyclones, a process known as extratropical transition, or ET for short. ET events, as they are called, can reenergize waning hurricanes and carry massive amounts of moisture, with the risk of flooding on land and storm surges along the shores. By some measures, almost half of the tropical cyclones that form in the Atlantic undergo some form of ET. New Zealand gets a few of these in the southern hemisphere, but the global champion in ET events, the one place where they are more likely to occur than any other, is northern New England and Atlantic Canada. One of the tasks of the Canadian Hurricane Centre has been to better understand ET; in the year 2000, Chris Fogarty, a meteorologist who operates out of the center, and others flew into Hurricane Michael while it was undergoing transition, and reported on the results in the journal of the American Meteorological Society. At the very least, they concluded, there is an urgent need to develop numerical and conceptual models that will enable weather forecasters to better anticipate changes and to improve warnings.10

A "baroclinic" storm enveloped us in the winter of 2004. On the morning of February 17, 2004, the forecasters told us to expect snow. By midday, the forecast had changed to "expect heavy snow," and by evening a blizzard warning was in effect. They had spotted a low-pressure system that had organized southeast of Cape Hat-teras, in North Carolina; it had taken on the classic cyclonic shape as cold air from eastern North America clashed with the warm waters of the Gulf Stream, and was heading northeast at almost 24 miles an hour. By the morning of the eighteenth, there was a broad cloud shield and a high cirrus deck spreading out over the region, the storm already beginning to stretch out to the northeast in a comma shape. A strikingly sharp cold front extended southward as far as the Bahamas. The barometer was dropping fast, and in just over twenty-four hours had gone from 1,000 millibars to 959.

The forecast did its job. Everyone in the storm's path knew it was coming. Snowplow crews gassed up, power company workers checked their equipment and cranked up the AT Vs. We stocked up on drinking water in case the power went off, and inventoried our food supplies; we made sure there was enough firewood for the fireplace and checked the kerosene supply for the lamps—we were reasonably well prepared.

We woke up to a howling wind and heavy snow. It snowed all day. It blew all day. It snowed all night. It blew all night. Shutters banged, trees cracked, the entire landscape seemed to groan in pain. Below the irritant of the banging shutters there was a constant basso roar, as the wind-born waves hurled themselves at our rocky beach. I wasn't worried they'd come over the rampart—only the rare hurricane causes that—but it still sounded like an express train going by at high speed. I seemed to feel the bedrock shaking, though I knew it was my imagination. Out on Sable Island, which was nothing but sand, the giant waves literally did make the whole island quiver, but here …

On top of the roar of the waves and the clattering shutter was the sound of the wind gusting in the spruces, varying in pitch from a howl to a whine. On the morning of the nineteenth, it was still snowing, the wind still blowing, visibility only a few yards. The storm had slowed down. It was not quite stalled, but it was now centered over Sable Island, a hundred miles to our east, and was drifting in a leisurely way to the northeast. We had maybe three feet of snow on the ground, swirled around and piled high in drifts. I couldn't see the cars, except for their antennas. The front door had snow piled against it well over the door handle. The back door, in a wind shadow, was completely clear, but halfway up the path the snow was over my head. Our deck was buried in a snow cover maybe four feet thick. The picnic table had disappeared entirely.

The media in the next few days took to calling the storm White Juan, in memory of Hurricane Juan, which had passed through only five months earlier. Winds during the storm were steady at 48 miles an hour, with gusts up to 75 miles an hour in exposed areas, decently into the hurricane range.11

An intense storm, but not that atypical in winter. Perhaps more snow than usual, but every year in these latitudes there are two or three storms pretty much like it.

That's really interesting about these hurricane-force winter gales is not their animus but the mundane and mechanically simple nature of their origins. The western edge of the Gulf Stream is where these storms are made. As the current snakes its way past Cape Hatteras and turns back out to sea after a close swipe at land, it mingles briefly with the cold tongues of the southbound Labrador Current. In winter, when a dome of high pressure from the Arctic drifts southeast, bringing with it vast eddies of freezing air, it may come to the edge of the Gulf Stream and stall. If a core of warm air is encountered off the coast, just east of the Gulf Stream, and if the jet stream is flowing toward the northeast, the two air masses collide, the cold air fighting to move east, the warm air prodded by the jet stream. A pocket of turbulence develops in the crook between them. Wind flows east, then is bent quickly to the north. Unable to resist the centrifugal force, it begins to move full circle, creating a system of low pressure that deepens violently.12

Because the winds are flowing counterclockwise around these storms, the winds come out of the northeast as the storms move offshore, and the fishermen just call them nor'easters. So commonplace are these violent winter storms that the load lines painted on modern ocean vessels denoting the depths to which they may safely be loaded always have as the lowest line, the lightest loading, a line marked "WNA"—Winter North Atlantic.13 Storms strong enough to severely damage large vessels happen on average about once or twice a year, particularly in the wintertime.14 Such storms happen in all the midlatitudes of the northern hemisphere. Insurance companies have tracked the damage they do; the worst in the last one hundred years was a 1953 storm that made its way to Europe, causing massive Atlantic storm surges, killing almost two thousand people. (For some of the worst winter storms, see Appendix 8.)

Exactly why some mixes of cold and warm air deepen explosively and others remain benign is still unknown. It all depends, says Peter Bowyer of the Canadian Hurricane Centre, on how fast the cold dry air and the warm moist air masses are forced together. What happens after they deepen is much better understood. Bowyer tells me, "Our computer models are quite adept at handling these storms. The physics of winter storms is better understood than that of hurricanes, so our models do quite well, and we can predict many days ahead of time when a storm is likely to form over, say, the Carolinas, and how long it will likely take to reach our region."

High overhead in both hemispheres are the stratospheric gales, and the jet stream of popular weather forecasting. This jet stream, which is essentially a standing wave of high pressure pushed eastward by the Coriolis force at the boundary between cool polar air and warm tropical air, flows at speeds up to 240 miles an hour, and sometimes more, at elevations around 30,000 to 35,000 feet, a little over five miles. Jet streams are generally the reason why long-distance flights are faster going eastward, because airline bosses like pilots to ride the jet stream to conserve fuel. Not that this is always easy; jet streams do meander, and substantial vertical wind shear, which can cause clear-air turbulence, can often be encountered at their edges; passengers who'd like a quick crossing don't necessarily sign on to a roller coaster and, as cabin crews can attest, are easily irritated by bumpy rides. Jet streams were first discovered during the Second World War, which was when high-altitude transatlantic flying first became commonplace. They were dubbed jet streams because they seemed to flow in narrow ribbons at high speeds—and jet aircraft had just been invented. More than one jet stream exists—the midlatitudes jet stream, the polar jet stream, and the polar vortex—and they are found above all oceans and all continents. The first to exploit them were the Japanese military, which in 1944 and 1945 launched experimental bombs suspended from balloons into the Pacific midlevel jet stream; some of these weapons traveled five thousand miles in three days; one reached the coast of Oregon, exploding near a Sunday school picnic, killing five children and the minister's wife.15 Another made it over the Rockies and reached the thinly populated Canadian prairie province of Saskatchewan, a most unexpected assault.

A jet stream, too, can be deflected by pressure systems and is generally closely watched by meteorologists in winter—a jet stream represents an area the scientists call a zone of baroclinic instability, and a deflected jet stream in midlatitudes can mean the difference between mild temperatures and severe ones.16 Jet streams that dive southward into the United States typically mean intense cold over much of the continent; when they retreat into midlevel Canada, the weather will be unusually mild. In addition, a jet stream pattern that swoops down toward the southern United States in late summer can cause a flow pattern that can steer hurricanes my way; I have learned to keep a wary eye on its position. The jet stream boundary is the locus of squalls, cyclones, and storms.17

Jet streams can work as a steering mechanism for short-term winter weather too. In January 2005 warm, moist air from the southern Pacific flowed up the jet stream all the way to the Aleutians, then dumped massive quantities of rain on British Columbia, where they called it the "Pineapple Express." The same system followed the jet stream across the Rockies, where it became a clipper— Montana clipper in the United States, Alberta clipper in Canada. Snow and snow squalls duly followed. But that wasn't all: The system followed the jet stream into the southern Great Plains, then back up to the Ohio valley and out to sea somewhere around Virginia. After which, yes, it rode the jet stream back up past New England and Maritime Canada, bringing snow squalls and high winds. Thank you, Hawaii.

At high altitudes winds are organized into a sequence of high-pressure ridges and low-pressure troughs, with a wavelike motion. The largest of these wave patterns, the so-called standing waves, have three or four ridges and a corresponding number of troughs in a broad band. Shorter waves, several hundred miles in length, are called traveling waves. These form the upper parts of near-surface cyclones and anticyclones, and guide their movements and development. These high altitude winds are global in scope and are called geostrophic winds; they are driven by temperature and consequently pressure differentials, and are not influenced by the surface of the earth. These winds were unknown before weather balloons and are now generally measured by aircraft.

At low altitudes winds are essentially circular, organized into cyclones and anticyclones (low-pressure and high-pressure areas). As we have seen, cyclones rotate counterclockwise around lows in the northern hemisphere and clockwise in the southern hemisphere. Anticyclones revolve in the opposite direction. Extratropical cyclones are generally benign, little more than eddies in the overall system, but they are nonetheless significant: They are an essential part of the transference of excess heat received in tropical latitudes from the sun to polar regions;18 without them the poles would be much colder and the tropics much hotter. In tropical latitudes the cyclones are smaller, usually not much more than 300 miles or so across, but they can carry winds of terrifying violence. When the sustained winds at their centers reach 74 miles an hour, they are called hurricanes or typhoons.

All in all, it seems like an orderly system, and it is. But on closer analysis it gets more complicated—much more complicated, for in other ways winds make up the most intricately beautiful and complex of the great engines that sustain life on the planet. Winds are steered and diverted and distorted by continents, mountains, forests, deserts, oceans, and large lakes, even cities, that muddy their flow and retard their passage. They also change winds' intensity. Land heats faster than water, and so localized pressure differentials are caused at every coast. Deserts, for their part, radiate heat faster than grasslands, and grasslands faster than forests, and each retains different degrees of moisture. All these factors complicate wind patterns.

Out to sea the patterns are simpler, and much more direct. If the world were perfectly flat—all ocean, perhaps—and didn't rotate, air would flow smoothly in perfectly predictable directions. Aerodynamicists call smoothly flowing air "laminar flow," as opposed to "turbulent flow"—we'd call it "streamlined." It is possible only where little interferes with the movement of air, and in nature this condition is very uncommon, so in practice pretty well any movement of air greater than 2 or 3 miles an hour is predominantly turbulent. It is one of the main tasks of car and aircraft designers to change turbulent flow into laminar flow to reduce resistance.

Large air masses don't mix easily. If they come together at speed, which is common enough, they set in motion complicated, swirling eddies of turbulent air. These eddies are what we call "weather."19

IV

Winds and weather are complicated in three other ways that make their understanding, and their prediction, much harder.

The first of these is the large-scale and relatively long-term cyclical fluctuations in the movement of air masses.

The second is the almost universal tendency of air to coalesce into vortexes. Hurricanes and tornadoes are the most obvious of these, perhaps because of their destructive potential, but there are many others, both useful and apparently useless.

The third is the purely local or regional winds affected by mi-croscale topographic and geographic features that are not plottable on larger weather maps, but which nevertheless have significant consequences. Examples are the notorious harmattan of the Sahara, the mistral of the Mediterranean, and, closer to my home, Ies Suites of Cape Breton Island, whose astonishing accelerating effects on moving air I have experienced myself.

Take climate cycles. Perhaps the best known of these is El Nino, referred to by scientists as ENSO (for El Nino Southern Oscillation). Typical El Nino events, as they are called, last somewhere between six to eighteen months; their most obvious feature is a massive upwelling of warmer water in the eastern tropical Pacific, whose main consequences lie in the tropics but whose climatic fingerprints cause widespread droughts as far away as southern Africa and generally warmer winters in places like northeastern America. Its companion phenomenon is La Nina, which is its opposite: unusually cold Pacific temperatures. La Ninas occur after many, but not all, El Ninos; their net effect is colder-than-usual northeastern American winters and warmer southwestern temperatures. As with dry years in the Sahel, there tend to be fewer hurricanes in El Nino years; the best guess is that one doesn't cause the other, but that some still-unknown factor causes both phenomena.

It's known that in the current global climate, El Nino years are warmer and La Nina years are cooler. It's also known that in 1976 the equatorial Pacific, potentially driven by anthropogenic warming, switched from a weak La Nina state to one in which El Nino occurs with greater frequency and intensity.

Does it therefore follow that more persistent El Ninos would amplify global warming? Or that more global warming would result in more El Ninos?

Perhaps, but not necessarily. A new study has found that in the early- to mid-Pliocene (5 to 2.7 million years ago), a steamy era that was the last time that global temperatures were warmer than they are today (atmospheric temperatures probably 10 degrees higher), the Pacific system was under an extended La Nina—like state, rather than the predicted El Niiio one. These results were unexpected, and remain to be explained.20

El Ninos were first acknowledged by fishermen from Chile, and because the phenomenon generally occurred around Christmas and brought them mostly beneficial results (more fish in the up welling water), they gave it the name El Nino, which means Christ child in Spanish. La Nina, for her part, was originally referred to as el Viejo ("the old guy"), but was given its present name by the American media.

El Ninos were first plotted by a British meteorologist, Gilbert Walker, in the 1920s, from as far away as India. Walker was trying to get a grip on what caused the often sizeable fluctuations in the strength of the Indian monsoons, and discovered that strong monsoons often correlated with severe droughts in Australia, Indonesia, and southern Africa. He also noted, without interpreting them, correlations between stable periods of high pressure in the eastern Pacific and periods of low pressure in the Asian Pacific. In his journals he called this "the southern oscillation." And there it rested, until the 1950s, when climatologists finally connected his hypothesis with what the Chilean fishermen had already observed. El Nino affects everything from large-scale climatic trends to microscale events, like wildflower blooms in the southern California deserts.

It is still impossible to predict when an El Nino will happen, a simple fact that makes climate-change skeptics raise their eyebrows— if you can't predict a simple recurring cycle a year or two in the future, how can you possibly predict climate change over hundreds and even thousands of years?

El Nino is not the only "oscillation" to affect winds and weather. There are at least a dozen others, and researchers seem to be discovering more every year. I spent a few months talking to atmospheric scientists and plowing through research papers in an attempt to understand their somewhat dizzying interconnections; at one point the wall in my office was slathered with charts labeled with impenetrable acronyms (AO, NAO, PDO, MJO, QBO, and others) of dubious utility, and eventually I tore them all down. Even to scientists, the impact of most of these cycles is only hazily understood.

Still, some things are clear. For example, the Arctic Oscillation (the "AO" in the list above) directly affects weather in the northeastern quadrant of North America and in western Europe, and tantalizing research has indicated some connections between the AO and tropical cyclone formation in the hurricane season. (There is also a subset of this cycle, called the North Atlantic Oscillation, but no one yet knows what it does.)

As its name implies, the AO circles the Arctic and extends high into the stratosphere. Its timescale is shorter than El Nino's—only a few months, or even weeks—and it cycles through a negative or cold phase, which brings high pressure to Arctic regions, along with lower-than-normal pressure over midlatitudes, and a positive or warm phase, whose effects are the opposite.

Perversely, warm AOs result in extracold weather in America and Europe. By contrast, when the AO high-level circulation is cool, it inhibits cold surface air dipping southward, warming up cities from Moscow to Vancouver, Calgary to Boston, London to Warsaw.21 Similar oscillations exist in the southern hemisphere. Some studies suggest that the Antarctic Oscillation, the southern equivalent to the AO, was corrupted by the recent hole in the ozone layer, which resulted in extraordinarily cold winds, which in turn may explain why the southern polar regions were warming more slowly than northern ones before the ozone hole repaired itself. (Warming in Antarctica is once again accelerating.)

Two other cycles with implications for wind and weather are the Pacific Decadal Oscillation and the Quasi-Biennial Oscillation.

Climatologists rather like the PDO because it is a way of showing the general public that so-called normal climatic conditions can change, sometimes radically, over a period less than a human lifespan. The cycle often causes wild swings in Pacific marine species like salmon, and in local weather patterns.

It's possible the PDO is merely an El Niiio writ large, with much longer cycles; there have been only two full PDO cycles in the last one hundred years—"cool" ones from 1890 to 1924, and again from 1947 to 1976, and "warm" ones from 1925 to 1946, and again from 1977 to the mid-1990s. It's also possible that the oscillation has two cycles, and not one—one from fifteen to twenty-five years, the other from five to seventy years. Its causes are unclear; the closest scientists can come is to suggest that it arises from some air-ocean interactions, which at least suggests some lines of inquiry.

Perhaps the most curious cycle of interest to wind scientists is the Quasi-Biennial Oscillation, in which the lower stratospheric winds of the tropics abruptly change direction, about every twenty-eight months. It is an enigmatic phenomenon, even by global oscillation standards. Why should winds that are easterly one month abruptly become westerly a few weeks later? And why should the easterly phase be about twice as strong as the westerly phase?

The QBO was only discovered in the 1950s, because it is only detectable at higher (stratospheric) altitudes. In the 1970s it was found that the periodic abrupt switches were caused by atmospheric waves starting in the tropical troposphere that travel upward into the stratosphere, where they are dissipated by cooling. The nature of these waves remains mysterious. The current culprit of choice is gravity waves, but what causes those and in this periodicity is opaque.

But they are important for millions of people, because hurricane activity is more common when these stratospheric winds are westerly, a pattern that is true also for cyclones in the Pacific. In its easterly phases, the PDO tends to knock hurricanes off their balance before they can really get going, while westerlies seem to act as catalysts. Why this should be so is another of the many unknowns.22

Reinforcing these oscillations of air, and probably partly caused by them, is a similar periodicity in global ocean movement that scientists call the thermohaline circulation, the steady movement of the world's oceans—with the Gulf Stream, the world's most powerful, and fastest, ocean current, as one of the prime engines. So important is the Gulf Stream that the whole phenomenon of thermohaline circulation is sometimes referred to as the North American conveyor belt. This conveyor belt has complex and still little-understood effects on winds and storms.

The conveyor belt is formed by water from the Florida Current, which circulates through the Gulf of Mexico and the Straits of Florida, and the North Equatorial Current, which flows westward along the equator. The resultant Gulf Stream parallels the coast of North America along a boundary separating the warm and more saline waters of the Sargasso Sea to the east from the colder, slightly fresher continental slope waters to the north and west. It more or less bounces off Cape Cod and is bent eastward, in the general direction of Ireland.

The Gulf Stream then feeds into the North Atlantic Current, which splits in northern and southern directions along the west coast of Ireland. The southward flow turns into the Canary Current, named for the Canary Islands off the coast of southwestern Morocco in North Africa, and thence bends westward parallel to the equator, winding up once more off Florida. Water flowing north along the west coast of Britain becomes the Norwegian Current, as it moves along the coast of Norway.

At least on the surface. Deeper down, the patterns are more complex.

When Gulf Stream water enters northern latitudes, it cools and sinks, becoming saltier and denser in the process (the haline in thermohaline). This happens in curious, slowly revolving "pipes" that take water from the surface to the seabed, mostly in the Labrador Sea and the Greenland Sea. At this low level, the water moves south and circulates around Antarctica; thence north again to the Indian, Pacific, and finally the Atlantic basins. The Smithsonian Institution estimates that it can take a thousand years for water from the North Atlantic to find its way into the North Pacific.

It is pretty obvious that changes in this massive circulatory device would have profound impacts, not just on wind but on climate generally. It is one of the main worries about global warming that increasing Arctic ice melt might alter or, worse, stop, the Gulf Stream, at least for a period. This indeed seems to be happening: The known vertical "pipes" have been reduced in number in the last few years from about a dozen to two, in part because the water is too warm to sink. The computer models all show that global warming would have a perverse short-term cooling effect on some northern places; instead of warming, Maritime Canada and northern New England, Ireland and the British Isles would go into a temporary deep freeze. I've sometimes contemplated this notion of my little house turning into Iceland, but it doesn't do to dwell too much on the possibility, because there are other, more immediately worrying things to be concerned about. For instance, a good deal of evidence suggests that changes in the velocity and direction of the conveyor belt might be a prime cause of the peculiar fact that hurricanes seem to wax and wane on a more or less thirty-year cycle. Is it just coincidence that the conveyor belt slowed down in the 1960s, cooling the North Atlantic slightly, in a period of fewer annual hurricanes? Or that the conveyor belt seems to have picked up speed starting in the 1990s, the decade when hurricanes started to increase in frequency again? Bob Sheets, former director of the National Hurricane Center in Miami, has asserted that the meteorological evidence suggests that the coming quarter century will produce more, and more intense, storms, and that "the thousands of people who moved to their dream homes during the hurricane low … could be in for some unpleasant time."23 It's also possible that it might be the other way around: The frequency of hurricanes may, by contrast, affect the thermohaline circulatory system.

For us on the American northeast coast, therefore, this expected deep freeze would be a mixed blessing. More ice, but fewer hurricanes. If the Gulf Stream, on the other hand, moved farther north, we'd be warmer, but we'd be hit harder and more often by severe storms. When Hurricane Juan hit Nova Scotia in 2003, the ocean water temperatures were two or three degrees above normal. I asked Chris Fogarty, who has made a particular study of the relationship between surface water temperatures and hurricanes, what would happen if this increase persisted. "More Juans," he said laconically24

Evidence is also accumulating that indicates the ocean's currents and hurricanes act together, in a feedback loop. The direction and speed of the conveyor belt can affect hurricane frequency, but the size and frequency of the storms can also push the Gulf Stream faster and farther. It was once thought that the prevailing winds alone, acting with our old friend the Coriolis force, caused the ocean's currents; it now seems that hurricanes have their part to play too. Kerry Emanuel, an atmospheric researcher at MIT, said in 2004 that the theory he had learned earlier in his career of how hurricanes began and sustained themselves was wrong. "It was always felt they were freak storms that didn't have much to do with climate in general … but we've come to different conclusions. Really, they are very integral to the climate system." His theory is that because hurricanes churn up the top 600 feet or so of the oceans, they lead ultimately to the circulation of all the world's oceans, which directly affects winds and storms, which in turn affect climate.25

This is the cheerful side of typhoons and hurricanes: Typhoons (from the Chinese ta, "great," and feng, "wind")26 and hurricanes (from Hunraken, the Mayan storm god)27 may be disruptive to a species that likes to build big-windowed homes by the sea, but within the complicated and coupled hydrographic-atmospheric system, they are nonetheless important agents of planetary self-government, redistributing air, moisture, and heat both vertically and latitudinally, scrubbing the air of accumulated pollutants, and accelerating the movement of the great ocean currents that keep our planet stable. A mature hurricane can export more than three and a half billion tons of air every hour, contributing greatly to redistribution of the troposphere, and can transport a billion tons of water over several degrees of latitude. They help make Earth work.

I should have thought of this when I was tracking Ivan. I mean to keep it in mind in the hurricane seasons to come. I just don't want a hurricane to drop any of that billion tons of water on me.

V

So much for climate cycles, the first of the three complicating factors in weather analysis. The second of these is the apparently irresistible need of flowing air (and liquids too) to form into vortexes.

By definition, a vortex is a rotation around a common center, often with a slow radial inflow or outflow superposed on the circular flow. As the air converges on the center, it begins to rotate faster and faster. "The process is surprisingly similar to an ice-skater's spin that accelerates as the skater's outstretched arms are drawn closer to the body"28

It's easiest to see in water—a whirlpool is a liquid vortex. Watch any stream that contains obstructions—a large rock, say—and you'll see a vortex, usually just upstream. I once watched a large log caught in just such an eddy, in a fast flowing river in British Columbia; it circled lazily in the eddy for more than three hours before its tip caught in the surrounding flow and it was dragged clear and went hurtling downstream (well, I had nothing better to do that day).

Vortexes occur throughout the universe, not just on our rotating little planet. Spiral galaxies are themselves versions of a vortex, caused by a combination of velocity and gravity. In our solar system, Jupiter's giant red spot, which is thought to be a long-lived vortex, is visible even to a small telescope. On Earth, cyclones are vortexes, and thus hurricanes and typhoons as well as tornadoes. Fire-caused whirlwinds are a hazard to firefighters. Dresden and Hamburg were both largely destroyed by fire vortexes. You can see vortexes in every campfire, in every little brook, in the water exiting a bathtub. Vortex motions, usually caused on Earth by pressure differentials in the atmosphere, are critical to a surprising number of human technologies, some economically important, others amusing but trivial. For example, airplane flight depends largely on vortex motion. Birds use tiny tornadoes to turn sharply in the air. The singing sound of high-tension wires is caused by vortexes. So is the ubiquitous slice of an amateur's golf ball, and, as we shall see when we get to Bernoulli's principle, even the annoying habit of shower curtains to billow inward and stick clammily to the person showering. A smoke ring is an example of a ring vortex. Some volcanoes blow smoke rings; so do steam locomotives. In nature, vortexes can have lifetimes ranging from a few seconds to several days.

In water, vortexes are called whirlpools, in which the flow is downward, and kolks, in which it is upward. Perhaps the most famous vortexes in history were Homer's Charybdis, off the Ca-labrian coast, and the Maelstrom, off Norway.

In air, the most common vortexes are whirlwinds and dust devils, which occur almost everywhere in the atmosphere, almost always accompanied by some degree of wind shear, or rapid interchange of air between layers; whole academic careers have been built on these boundary-layer studies. A typical whirlwind is the Australian cockeyed bob, which picks up leaves, light twigs, and dust as it goes; the vortex winds we called die duiwel gee om, "the devil cares," near where I was born, were strong enough to carry tumbleweeds, some of them as large as hippopotamuses. Winds that build from the bottom up, like these whirlwinds, are sometimes called willy-willys; their more deadly cousins, tornadoes, are made from the top down, and of course are far more ferocious.

Tornado comes from the Spanish word for thunderstorm, tronada, which in turn comes from the Latin for turn, tornare, which is what vortexes do. And turn tornadoes do, spinning tightly with awesome speed. This is the most violent of all winds. Just how violent is still unknown, because tornadoes routinely destroy even the most robust of measuring devices, even supposing one could be placed in a storm's unpredictable path. But estimates have placed vortex winds at somewhere around 290 miles an hour, much faster than a Category 5 hurricane like Camille, and it is possible that occasional tornado winds might even exceed 480 miles an hour. The highest wind ever measured within a tornado was near Red Rock, Oklahoma, in April 1991, when a twister was clocked at 286 miles an hour.

The story is similar for barometric pressure. Standard barometers can't cope with the rapid changes in pressure caused by tornadoes, but pressure drops of 100 millibars are not uncommon, and drops of 200 not unknown. Because such drops occur in mere seconds, the normal pressure inside a building simply doesn't have time to adjust before the roof is blown off and walls are blown outward. The energy within a single tornado is not much less than the 20-kiloton bomb dropped on Hiroshima.29

I've had three near-encounters with tornadoes, and I've seen the results of others. The first was just a few years after I had nearly been blown out to sea in a Cape southeaster. My family had moved to Johannesburg, a notorious locus of thunderstorms and massive hailstorms. One day we were in the family's aging sedan on our way from somewhere to somewhere else when the sky suddenly turned black, and then a violent yellow, and a deep rumble scraped across the city and across our nerves. My father brought the car to a stop and we saw the twisting funnel of a tornado touch down, perhaps a mile away, and then it was gone. Afterward he took me to see where it had been, a furrow of destruction two hundred yards wide carved through the edge of the city. Astonishingly, in the center of the path, a house still stood. Its roof was gone entirely, even the rafters, but nothing indoors was disturbed. Nearby was the remnant of a tree, only the first six feet still upright. The wind had driven a wooden clothespin deep into the wood so hard I couldn't pull it out. The family that lived in the house had taken refuge, as per the conventional wisdom, in the bathtub, and were unharmed. This being the South Africa of a certain period, there were servants quarters out back; these had been destroyed, and the woman who lived there had vanished and was presumed dead. The "savage and baleful Zephyrus," god of the winds of the sea, could not, surely, have been any more capricious than this.

The second encounter was in Arizona. I was at a conference of magazine editors, meeting in one of those gloomy, subterranean hotel conference rooms, when the lights abruptly went off. I remember it mostly because a publisher was whining about the ingratitude of her editor, and the rest of the assembled editors clearly felt the lights going out meant god was on their side, but then they came on again, and the discussion, such as it was, resumed. Outside … no more than five hundred yards from the hotel, a tornado had torn through. We were stunned at the proximity and the extent of the damage. The storm had ripped a path through town, leaving behind a ghastly jumble of mangled cars and demolished billboards, road signs, and small buildings. A pole with a traffic light still attached was poking out through the ruined windshield of a pickup truck. By some miracle, no one had been killed.

The third was in Ontario, which gets few tornadoes. At that time we owned some woodlands in the Ontario deciduous forest belt, and one day, when we were away in the city, a tornado tore through the woods not far from our cabin, so we saw it only by its results. Its path was peculiar, not uncommon with tornadoes. The whole thing seemed to have begun and ended on our small property. It tore through the forest for no more than a few hundred yards, but while it was there, it demolished a straight line of maples and beeches. It didn't just knock them over, as a hurricane would have. It tore them right out of the ground, roots and all, and pushed them into an untidy heap. No more than a few feet off its path, the trees were untouched. Even the leaves were still on their branches.

Tornadoes can happen anywhere, but the United States has the dubious honor of being far and away in first place, both in frequency and violence, with Australia an unenthusias-tic runner-up. Other common-enough spots are the Ganges basin of Bangladesh and the Yangtze River valley of China; and I knew for myself that they happened on the great plains of South Africa, and occasionally in central Canada. Tornadoes, formerly just called whirlwinds, or occasionally, typhoons, were not unknown in Europe— Aristotle described a tornado in the handbook he called Meteorologica, and a tornado ruined parts of Rome in 1749. But the champ is America, where the tornado belt, or Tornado Alley, runs in a swath across the Great Plains from north Texas through Oklahoma, Kansas, and Iowa to southern Minnesota. Somewhere between six hundred and one thousand tornadoes touch down in the United States every year. May is generally the worst month. In its results, America's worst tornado was in March 1925, when a twister roared at 60 miles an hour through a series of small mining towns from eastern Missouri to western Indiana, killing 695 people. But in sheer perverse capriciousness, the unfortunate loser has got to be a small town in Kansas called Codell, which was hit by tornadoes three years in a row on exactly the same date, May 20, in 1916, 1917, and 1918. The tornado-free May 20, 1919, must have been rather a big day in town.

Squall lines, no matter how severe, seldom generate tornadoes, nor do normal thunderstorms—neither a squall nor a thunderstorm is a vortex, and to conceive a tornado the mother storm must show at least the beginning of a cyclone effect, a true vortex. The deadliest tornadoes are the creatures of mammoth and long-lived storms called supercells, whose winds are already rotating (they are themselves vortexes, albeit slow-moving ones) and may carry updrafts and downbursts exceeding hurricane strength. Some of these supercells can be 30 miles wide and 60,000 feet tall. Some of the other ingredients necessary for birthing tornadoes are warm, humid air near the ground, cold air at higher altitudes, and shearing winds. As with hurricanes, it is the humid air rising rapidly into colder air above that precipitates ice or rain, which in turn releases enormous latent energy, which then refuels the storm. Supercells almost always carry massive amounts of moisture, which often comes down as hail—many observers have reported what they call "hail roar" during a thunderstorm, the sound of billions of hailstones clattering together on the way to the ground.

Tornadoes can form very quickly, and are very hard to predict. Warm air rising rapidly into colder air above is a necessary precondition, but if the warm air rises steadily and smoothly, tornadoes are actually unlikely. A much more likely result would be another series of rather weak thunderstorms. But if a shallow layer of just-warm-enough air hovers above the surface—warm enough to prevent the ground-level air from rising—the potential is much greater for serious damage. Because if that cap is somehow moved or damaged, say by an incoming cold front, the pent-up warm air on the ground can burst through very rapidly. Then, watch out. Tornadoes can appear in less than an hour.

Fortunately, only one in a thousand thunderstorms becomes a supercell, and only one in about ten supercells causes tornadoes. The exact mechanism for tornado formation is obscure. They are more likely when surface winds blow in a direction other than high-altitude winds and the stronger the winds and the greater the height of the storm, the more intense the results. But as with hurricane beginnings, the actual tipping point is not understood. What is understood is that the Great Plains are the perfect kitchen for cooking tornadoes. This is because the eastern half of the continent is overlain in summer by warm moist air coming in from the Gulf of Mexico, and the western states, where the prevailing winds are westerlies, are very dry—they are in the rain shadow of the Sierras and the Rockies. Thus, Tornado Alley.

In the peak season, hundreds of tornado chasers (known, bizarrely, as "the chase community") spread out across Tornado Alley, usually in Kansas but also anywhere from Texas to South Dakota. Guessing tornado touchdowns is a sophisticated, if hazardous, game, and the Internet is full of more or less fanciful boasts from people who claim predictive powers that range from implausible to deranged. In high season tour buses packed with gawkers who want to experience nature's ferocity for themselves barrel their way down rural highways, hoping to get lucky. Some of these come closer and get luckier than they would have liked, and the occasional bus lurches out from under a supercell with windows shattered by flying debris or side panels dented by furious hailstones.

Some of these chasers are like the ham radio operators who bombard the National Hurricane Center with their track predictions, "useful fools," as they are often described by the professionals. With tornadoes, because of their elusiveness and short duration, a curious symbiosis has developed between the chasers and professionals from places like Oklahoma's National Severe Storms Laboratory, who themselves fan out across Tornado Alley hoping to plant instrument packages directly into a twister's path. In practice, both "communities" keep in touch via cell phones and radio; they are occasionally plugged into emergency services and police bands when tornados are thought to be imminent. Tornado watches, which are little more than a guess at probabilities, are released to the public and the media several hours before tornadoes are expected, but warnings of actual twisters are released in a much shorter time frame. A network of Doppler radar units covers much of Tornado Alley, but the rupture in the cap that can produce tornadoes can be too small for the radar to easily see. As a consequence, sightings from the public, and from the tornado chasers, are taken seriously. With luck, warnings can come up to fifty minutes before the tornado strikes, but they can be issued as little as a dozen minutes before zero-point, perilously little time to take shelter, if, indeed, any shelter is to be found. Tornadoes seldom last more than an hour.

The first sight of a tornado is its funnel shape. It always seems to be striking downward at the earth, but this is an illusion. Tornadoes do form from the top down, but they aren't visible until they pick up debris from the ground—what you actually see is a grotesque mixture of earth, shrubs, fragments of trees, window glass, household effects, flying barns, bits of houses, whole cows, even car bodies and sheets of plywood and metal roofing. Their forward speed is usually around 30 or 36 miles an hour, but they can be nearly static or move well over 60 miles an hour. Their paths are usually narrow, no more than several hundred yards and sometimes less. Length varies widely from not very much to dozens of miles. The 1925 Missouri twister was huge, 9 miles wide and more than 180 miles long. A series of tornadoes crossed Grand Island, Nebraska, in June 1980 at a stately 4.8 miles an hour; the path of the final one included two complete 360-degree circles—the damn thing just would not go away. Five people were killed. On April 22, 2004, a tornado roared through Utica, Illinois, killing eight residents who had taken shelter in the local tavern. "The sky turned purple and then the air screamed," said one of the survivors, Mary Paulak. "It sounded like a woman shrieking with rage."30 The storm had struck too quickly for residents to react. There had been fifty-one reports of tornadoes the previous day, and everyone was nervous, but this one came without any warning whatever.

As usual with tornadoes, there were anomalies in its 200-yard-wide path of destruction. One house had its back walls sheared right off, but a small cluster of cheerful orange tulips at the front porch was untouched.31 Anomalies, curiosities, quirks—these are the nature of tornadoes. Almost every twister leaves behind some curious fact—a children's doll driven feet-first into the trunk of a tree, but otherwise undamaged; the whole roof of a house, still intact with its gables and gutters, five hundred yards from the house it once adorned; a car containing two children hurled into the air and back to the ground, the children miraculously unhurt; a house demolished, the one next door, a mere three feet away, untouched; blades of straw embedded in fence posts; a schoolhouse with eighty-five pupils inside demolished and the children carried one hundred and fifty yards, unharmed but seriously frightened; five railway coaches, each weighing seventy tons, moved thirty yards …

A 1985 tornado in Barrie, Ontario, sheared a house in half, peeling it open like a doll's house. After it had passed, an ironing board was still standing in an upstairs bedroom, the iron still on it, as though ready for use. Of course I had seen for myself in Johannesburg how by cruel fate a twister had unfairly mirrored the apartheid system by demolishing the quarters of a black servant, leaving the main house and its white residents more or less intact, and how a little wooden clothespin, made of soft wood and a short length of twisted wire, had been driven hard into a tree. I'd also seen the aftereffects of a tornado that had passed through an oasis near Timbuktu in Mali in 1999, leaving the poorly constructed mud-built houses undamaged, but tearing out by the roots all the date palms, the reason for the community's existence. Leaving the houses alone was doubly unfair—they were of no further use. Less than a week after it had happened, the entire oasis was deserted. No repair crews were working away, no builders, no gardeners or planters, no herdsmen or householders. The oasis was empty, abandoned. The houses had been stripped, the camels moved off. Everyone had left.

Some of these anomalies are caused by smaller vortexes that spin around the edges of the larger one; videotapes of large tornadoes often show three or more smaller vortexes curling around the main funnel. Sometimes the strongest winds are generated in these associate twisters; wind tunnels have created small tornado-generation chambers that show how they work. In other cases, the quirks are caused by changes in the core circulation of the main body of rotating air—as it waxes or wanes, the ground-level effects can change from a few yards to tens of yards in seconds. Sometimes the funnel leaves the ground altogether, only to touch down again a hundred yards or so farther on—sparing one or two houses in a long row, with, always, apparently demonic unfairness.32

Less-violent wind vortexes than tornadoes are often just called landspouts. They're common at the higher elevations of Colorado and Kansas and in the Caucasus of eastern Russia, where the height of the land makes it difficult for strong tornadoes to form. Witnesses who have seen these elusive and fleeting apparitions describe them as curiously beautiful, almost luminous and translucent, perhaps because the low level of available moisture is not enough to fill them completely. They're called landspouts because they rather resemble their aqueous cousins called, for rather obvious reasons, waterspouts.

Water vortexes—waterspouts—are true tornadoes though their debris fields, and therefore their visibility, are rather different. Waterspouts have made their way into a good deal of fantastical literature—whole ships are said to be sucked up, and one early novel even had a stable community of spout-dwellers living comfortably at the spout's apex. Alas, even the more prosaic legends are mostly untrue. Waterspouts may pose some danger to small fishing boats, but none whatever to larger ships; nor do they suck up massive quantities of water, although they may lift water a few yards. They're really only visible because they contain clouds formed by condensation.

The most famous real, as opposed to fictional, waterspout appeared off Massachusetts in 1896. It was witnessed by thousands of holiday makers on a variety of beaches, for it appeared three times, lasting some thirty-five minutes. Estimates cobbled together from numerous excited eyewitnesses put it at over 3,000 feet high and maybe 250 feet at base.

It was widely believed that interrupting this column of air could be dangerous. "The violence of the wind retains the column in the air, and when that long spout of water comes to be cut by the masts or yards of the ship entering into it, when one cannot avoid the same, or the motion of the wind comes to be interrupted by rarifying the neighboring air with cannon or musket shot, the water being then no longer supported falls in prodigious quantities [upon the vessel]."33 But no, contrary to legend, firing a cannonball into a waterspout will have no effect whatever, except to wet the cannon-ball on its way through. 34

If you could interrupt a vortex, you would, indeed, destabilize it and cause it to fail; this is the theory behind controlling hurricanes. But the energy to do so is almost as great as that carried by the vortex itself, and the notion rather lacks for practicality.

VI

The third complicating factor in weather analysis is the microclimate winds, local systems that are geography and topography dependent, that ride on the back of global wind systems but that have a profound effect on local climate and weather. Engineers have to pay attention to local winds. The "local wind climate" can affect how buildings and bridges need to be designed.

But local winds have affected more than that. Winds affect not just myth and mood; they have also, in a very direct way, affected history, changing it for better or worse. If global winds have affected human history, in the sense of delineating those regions where cultures might best flourish, local winds and storms have affected it also, but much more abruptly—the "what if" school of history is full of weather-related stories. For example, a Saharan sandstorm foiled the Persian invasion of Egypt in the fourth century B.C. The Khanate assault on Japan was called off when a typhoon sank half the fleet in 1275. In 1529 heavy rains and high winds fatally delayed the progress of the huge Ottoman army under Suleiman the Magnificent, which would otherwise have captured Vienna and dethroned the Habsburgs centuries before their time. Sixty years later the Spanish Armada went down to defeat because the winds conspired with the British to blow in the wrong direction. When a gale blew the Armada back into port and one of his advisers suggested it was an omen from the Almighty, Philip II responded with what historian Geoffrey Parker called "naked spiritual blackmail": "If this were an unjust war," Philip declared, "one could indeed take this storm as a sign from our Lord to cease offending Him. But being as just as it is, one cannot believe that he will disband [the armada], but rather grant it more favor."35

A French fleet sent to sack Boston in 1746 was destroyed by savage gales, once off the Bay of Biscay, a second time near Halifax; the hapless commander, dispirited by the debacle, fell on his sword in his cabin, and retired from history. Many a time the fate of the American Revolution turned on wind. The 1775 "Independence Hurricane" drowned about 4,000 British sailors just south of Newfoundland, which in turn affected the British presence in New England in the months that followed. In 1776 George Washington called off what might have been a disastrous attempt to take Boston from Imperial troops because a "hurrycane" (in the word of one of his junior officers) intervened. The same year a violent Atlantic nor'easter allowed him to escape from New York. And in the following year he defeated Cornwallis in a crucial battle whose fortunes turned when a north wind froze muddy roads along the Delaware and enabled the Americans to reposition their artillery. Two violent storms sealed the fate of the British troops trapped at Yorktown in 1781. And more recently too: In the Second World War several campaigns were won or lost when the harmattan, the hot wind of the desert, knocked out communications. Not to mention the D-day invasion itself, which hinged for an awful moment on a break in an Atlantic gale. And the country of Bangladesh was birthed in a typhoon: In 1970 a tropical cyclone hit Bangladesh, then still called East Pakistan, killing more than 300,000 people; the central government's mishandling of the crisis that followed was a prime reason for the eastern province to break away and form its own state.

Like the global wind systems, local winds are a complex and shifting amalgam of factors, including the presence of bodies of water nearby, the accelerating effect of narrowing valleys, and the presence of mountains. These may be small, a particular valley in a prevailing wind, or very large, the Sahara Desert, for example, which is as huge as the continental United States. The most common local winds are sea and land breezes, because land and sea heat up at different rates, and mountain and valley winds (katabatic and anabatic winds)—the Foehn, the Chinook, the Santa Ana, and many others—because valleys and areas of rock heat up faster on one side or the other, depending on the sun's position, which results in thermal lift on the sunlit side and a cool downdraft on the other. These winds can be powerful. "Local" winds can also be quite large: The monsoons of Southeast Asia are really an overscale form of the sea and land breeze.

In a way, hurricanes and tornadoes, indeed all storms, are local winds too, although they are traveling winds with no fixed address.

In the Mediterranean, locally unique winds have been mapped for millennia. The fine-grained permutations are almost endless. Corsica, for example, has been known to record a Force 6 or 7 gale on the west coast, calm on the east, and quite a different gale in the Strait of Bonifacio, places only a dozen miles apart.36 This has much to do with the presence of many islands, which distort and redirect winds. "The heat of an island creates new winds, mountains arrest some, accelerate others and all of these must meet, mingle and die or dance. The pairing of two islands means the re-diverting of already diverted winds."37

Wind anomalies exist, in short, everywhere. Within a few hundred miles of where I live are three of the most interesting. All three are due in some ways to the accelerating effect of local landscape, or what the forecasters, without apparent irony, call channeling. This is nothing more mysterious than the garden hose effect—if you pinch a hose, the water emerging through the pinch will come out faster than that going in. The same thing happens with winds. A small-scale example are the winds that are killing my rhododendrons. A somewhat more dramatic example is along the Presidential Range of mountains in New England, where Mount Washington for years held the record of the fastest wind ever recorded. The mechanism is simple: The westerlies race across Vermont, slide into the Connecticut River Valley, and then roar up the west slopes of the Presidential Range, compressed by the valleys into less and less space. At the summit of Mount Washington these winds exceed hurricane strength four days out of ten; in 1934 a weatherman recorded a wind speed of 231 miles an hour, faster than even the most terrifying hurricane. For decades this survived as the record speed for wind, until an observatory in Guam recorded a mountain summit wind of 236 miles an hour in a typhoon in 1997.38 A wind-tunnel terrain study of Hawaii and Guam by NASA showed that wind funneled up a diminishing valley can eject peak velocities of about 250 percent of the flow coming in from the ocean.

The Canadian province of Newfoundland, a few hundred miles to my northeast, is one of the windiest places on the planet—it is surely destined, in the post-fossil-fuel era, to become to wind power what Saudi Arabia is to oil. E. Annie Proulx, in her novel The Shipping News, caught the unique flavors of the Newfoundland psyche when she described the wind scouring the bleak landscape:

By midnight the wind was straight out of the west and he heard the moan leap to bellowing, a terrible wind out of the catalog of winds. A wind related to the Blue Norther, the frigid Blaast and the Landlash. A cousin to the Bull's-eye squall that started in a small cloud with a ruddy center, mother-in-law to the Vinds-gnyr of the Norse sagas, the three-day Nor'easters of maritime New England. An uncle wind to the Alaskan Williwaw and Ireland's wild Doinion. Stepsister to the Koshava that assaults the Yugoslavian plains with Russian snow, the Steppenwind, and the violent Buran from the great open steppes of central Asia, the Crivetz, the frigid Viugas and Purgas of Siberia, and from the north of Russia the ferocious Myatel. A blood brother of the prairie Blizzard, the Canadian arctic screamer known simply as Northwind, and the Pittarak smoking down off Greenland's ice fields. This nameless wind scraping the Rock with an edge like steel.39

For most of its five hundred years of settlement, Newfoundland was populated only around the edges, in little villages called out-ports reachable only by boat. It wasn't until the coming of the railway in the late 1800s that people actually traveled over land. That had its drawbacks: Shortly after the Newfoundland Railway was completed, Passenger Train No. I was blown over during a winter storm and its mail car caught on fire.

The story is told by Mont Lingard, who wrote a number of books about the narrow-gauge "Newfie Bullet" trains, including one called Next Stop: Wreckhouse. Lingard, who had traveled the route many times himself, quoted the appalled engineer: "When we went down across Bennett's Siding [on the island's west coast, facing Labrador], we struck it. Just like running into a concrete wall. WHOMP! Mike had one of those big army parkas on. When he come out of the station, the wind picked him up and took him about twenty feet up in the air and slapped him down on the ground. He said, 'I nearly blew away' The caboose and five cars went overboard. The caboose just cleared the bridge, except for that she would have went out in the water and we all drowned."

It happened again, a few decades later. As Lingard wrote: "Here was one of those big trailers, she was on the full length of the 40-foot flat car. She lifted up just the same as you lifted her up with a crane and moved her out about 15 feet clear of the track and brought her down. Never turned over or nothing, just lifted her up the same as you did it with a crane. It was unbelievable. They had tied her down and everything but that didn't matter. A good squall of wind took her and broke all that like nothing."40

The company that built the railways petitioned the government for money to build a new line elsewhere, but was refused. Since the company couldn't afford it, and could not afford to continue losing the occasional train to the nearby ditch, as was happening, it reached an ingenious solution: It hired a man called Lauchie McDougall as a human wind gauge.

Bruce Whiffen, who works for Environment Canada in its Newfoundland office, has written an engaging account of the Lauchie story. The McDougall family had settled in the area in the 1870s. Lauchie was born in 1896 and lived his entire life at what was already called, for fairly obvious reasons, Wreckhouse. He developed an intuitive and almost uncanny knack for figuring out the signs that indicated an approaching storm, especially the deadly south-easterlies. The railway company signed him on, paying him the then-substantial fee of twenty dollars a month, and over the years, Lauchie was credited with delaying hundreds of trains because of treacherous conditions. At least once, a train conductor spurned his advice and twenty-two cars blew over into the ditch. Lauchie died in 1965. You can still see a bronze plaque at the Marine Atlantic Terminal in Port aux Basques that says, in part: "This plaque is dedicated to the memory of Lauchie McDougall (1896-1965). McDougall had extraordinary skills in determining wind velocities … through this area. Often called the human wind gauge, McDougall provided this service to the railway for over 30 years."

These appalling Wreckhouse winds were themselves caused by the garden hose effect. When an intense storm approaches Newfoundland from the south, gusty winds converge along the south coast and extend inland several miles from shore. When they encounter the Table Mountains a few miles north of Port aux Basques, the winds channel through the valleys and gulches between the mountains. The wind that then passes over and through the Table Mountains is far stronger than the original wind, sometimes devastatingly so.

The railroad no longer exists, but the Trans-Canada Highway was pushed through the same area, and in stormy weather truckers and car drivers are forced to delay their passage. Sometimes the more stubborn among them refuse; and the local police are resigned to pulling the shaken motorists from the ditch afterward.

The other local wind effect that I know well is even closer, no more than a four-hour drive away, on the western side of Cape Breton, the northerly portion of the province of Nova Scotia. Cape Breton is a lovely place, but its weather can be awful (it is a standing joke around here that a forecast will say "windy today, with gales in Cape Breton"). Les Suites, an Acadian corruption of the French words sud-est, or south-east, are also channeled winds, but here the effect is more complicated. Les Suites are so-called mountain waves, which happen when stable air flows over mountains or hills and combines with other effects such as drainage winds. In Cape Breton, southeasterly winds often blow ahead of an approaching warm front. Above the surface and ahead of the front, the air is very stable and causes a frontal inversion (warm air over cold air), which pinches the approaching air stream between the inversion and the surface. This has the by-now-familiar accelerating effect, creating very strong surface winds along the coastal plains on the lee side of the hills. If there are strong gale-force southeasterlies inland and to the east everything collides, and hurricane force gusts to 100 knots or more will be caused on the west coast and ten to fifteen miles out to sea. In March 1993 an intense storm starting in the Gulf of Mexico moved up the east coast to Maritime Canada, causing Les Suites winds measuring 126 knots, 150 miles an hour, in the little town of Grand Etang, lifting the roof off Cheticamp hospital.41

VII

As meteorologists, and then mechanical engineers and the aircraft industry, understood more and more about the turbulent and occasionally violent nature of air and wind, and as they came more and more to see the importance of vortexes, a focus of these new wind engineers became the physics of boundary layers, defined as the distance from a surface at which a velocity of wind reaches 99 percent of the unobstructed "free stream." This might all seem too esoteric for real-world applications, but boundary layer studies, combined with studies of turbulent flow, have generated a much greater understanding of how buildings and structures like bridges withstand the force of wind. It was the belated realization that wind's "energy content," in the scientific jargon, was affected by turbulence at the boundary layers that finally pushed wind engineering from something theoretical to something that would have a real impact on how buildings were designed.

The pioneers in this endeavor were Jack Cermak, a professor emeritus at Colorado State University in Fort Collins, Colorado, and Alan Davenport, who founded the Boundary Layer laboratory and wind tunnel at the University of Western Ontario, in London. Others include, most notably, Richard Kind, at Carleton University in Ottawa, whose research includes snow and sand movement in wind, wind-damage control mechanisms for roofers, and wind studies of retractable stadium roofs. But Cermak and Davenport are by far the best known. Cermak had built a massive wind tunnel in Colorado in the 1960s, big enough to model full-scale atmospheric boundary layers. In 1964 Davenport and his associate Les Robertson showed up at Cermak's lab to see if they could borrow his wind tunnel to check out designs for a new building project in Manhattan. They needed the best models they could get, because the project was difficult and expensive. They wanted to test everything— wind loads, pressures, and potential flexibility, everything they could. This new building would call for innovative structural concepts, and would be an incredible 1,368 feet tall. It was to be called the World Trade Center.42

The British-born Davenport has done the critical wind studies for many of the world's most complex engineering projects, among them the world's longest bridges and tallest buildings—not just the World Trade Center but also the Sears Tower in Chicago, the CN Tower in Toronto, the proposed 1,900-yard span crossing the Messina Straits in Italy, the Normandy bridge in France, the Storebaelt bridge in Denmark, and the Tsing Ma Bridge in Hong Kong. The lab has re-created a miniature city and simulated wind conditions for Hong Kong, where turbulent typhoon winds put every building, including the Hong Kong Bank and the Bank of China, to the test. Davenport's resume seems endless; he helped write the Caribbean Uniform Building Code, implemented in the 1980s, which is intended to render low-rise buildings more resistant to hurricanes. He is a founder of the insurance-industry-funded Institute for Catastrophic Loss Reduction, and the chairman of the steering committee of Project Storm Shelter, a facility sponsored by the International Association of Wind Engineering to improve the wind resistance of housing. And he is developing a series of high performance, low cost, prefabricated housing for use in postdisaster situations and other applications.43

It was a study of the proposed CN Tower in Toronto that cemented Davenport's reputation beyond his professional colleagues. The tower, a needle like structure that dominates Toronto's skyline, was for years the world's tallest free-standing structure, a banal boast pretty much calculated to leave out guyed towers, of which there are several that are taller, and to finesse the fact that the tower isn't really a building but a communications tower with a restaurant on top. But since it was to be built in the middle of a city, and it would be by far the tallest concrete structure ever built, the architect's design was put to the test in Davenport's lab. It failed, and the original design was abandoned for a new-and-improved version. Davenport had shown that it would not stand up to the elements. "There were problems," Davenport says delicately, "with the amount of sway in public areas."44

Among the most important case-studies of local winds are those that involve long-span bridges. These are very complicated structures, susceptible to wind in many ways. They are liable to swaying, and to oscillations; the cables are liable to quiver dangerously, like massive violin strings, in high winds; and all components undergo stress and therefore fatigue. Oscillation is the enemy of bridges, and engineers must install what they call vortex-dampers to head it off. Out-of-control oscillations have destroyed a number of such structures over the years, including the 1836 collapse of the Brighton Pier in England, the 1879 collapse of the Tay Bridge in Scotland, the 1940 collapse of Tacoma Narrows Bridge in Seattle, and the 1986 Amarube Tekkyo rail bridge in Japan. Perhaps the most notorious of these was the Tacoma Narrows Bridge, which was captured on film—and which can be viewed on any of dozens of Web sites run by disaster junkies. The most interesting result of the collapse was to cause a new note of sobriety to enter the world of bridge design. Since the nineteenth century, bridge engineers had been besotted with new techniques and new designs, and suspension bridge designers competed with each other to achieve a maximum of structural grace and slenderness; artistic merit more or less overshadowed conservative engineering—a fact of life now endlessly drilled into the skulls of first-year engineering students. With its very shallow trusses and slender towers, the Tacoma Narrows Bridge was the high point in bridge artistry Alas, it liked the winds far too much. It was shaped not unlike an aircraft wing—except that bridges are not supposed to generate lift. Only a few weeks after opening, the bridge had already developed the nickname Galloping Gertie for its tendency to heave in even moderate winds. A gusty windstorm in November 1940 was enough to do it in altogether. For an alarming few minutes, it twisted violently in the wind—at maximum twist one sidewalk was twenty-eight feet higher than the other—before a six hundred-foot section broke off entirely and plunged into Puget Sound.

All bridges are now subjected to full wind-tunnel tests before any concrete is poured or steel fabricated. Some parts of this testing are straightforward—the aerodynamics of the bridge cross-section and its towers are easy enough to model. Where it gets complicated is the introduction into the analysis of the "local wind climate," an incredibly complex study of historic meteorological records, historic wind directions and speeds, and the local topography—are there any accelerating topographical features around that would make normal winds into gales? And what about hurricanes? Not every area is hurricane-prone, but all areas might get hurricanes on rare occasions. How to model for those?

The problem with these local wind climate studies is that there is so much data, far too much to make accurate calculation possible: historic storm intensity data, storm track data, pressure differentials and the like, data on midspan and quarter-point pressures, defections and deflections, cable tensions, and many other variables, some of them with short-term periods and little apparent predictability. Even massive number-crunching computers are not up to the task; and even if they were, it would take far too long to input the data. And so engineers use a curious statistical sleight of hand with a whimsical name to provide answers about the sensitivity of the design to winds of various strengths and directions. They call it a Monte Carlo simulation.

As you might expect, the name is derived from the casino at Monte Carlo. Like scientists trying to model the real world on computers, gamblers too are faced with large sets of apparently random numbers. Each gambler, notoriously, has his or her own method of assessing the odds. In the casino as in the humdrum world beyond its walls, the numbers may be random, but very large sets of runs will provide statistical patterns that are more or less valid—as chaos theory would predict. To take the most widely known example, it is impossible to predict whether a coin toss will come up heads or tails, but a very large number of such tosses will always yield a fifty-fifty ratio of heads to tails. The Monte Carlo simulation, then, is simply a use of random numbers and probability statistics to investigate problems. It makes possible the examination of problems otherwise too complex for computation. For example, solving equations that describe the interactions between two atoms is fairly simple; solving the same equations for hundreds of thousands of atoms is impossible. Monte Carlo allows the sampling of such large systems, and the wind climate around bridges is a real-world example. (MC methods are used everywhere in science, in disciplines as diverse as economics and nuclear physics.)45

An intriguing test case for local wind studies was something not nearly as, well, dire, as the winds that destroyed trains in Newfoundland or the bridge near Tacoma. It was a wind that affected, and still affects, only the pocketbooks of a handful of very wealthy men. These are the winds around the eleventh green and twelfth tee of Augusta National Golf Course, the home of the Masters tournament in Augusta, Georgia. Golfers have ruefully called this Amen Corner, in recognition that only prayer seems to help the ball go where it is directed. Notoriously gusty winds from the left can be in the face of a golfer teeing off at the twelfth, but at the twelfth green, the target of the current shot, the flag shows wind coming from the right. How, then, to judge the shot? If you aim the ball to match the wind in your face, halfway through its flight it will abruptly be seized by a contrary wind and blown deep into the rough. If you compensate for that, you might still lose. The gustiness of the site means that the compensating wind late in the shot may not be there, dragging the ball into the rough on the other side. Tournaments have been won and lost at Amen Corner, which means substantial money is at stake.

Amen Corner is at the bottom of a slender valley between two hills. Tall trees surround the twelfth tee and green, in contrast with the relatively exposed eleventh green. For years, golfers have blamed those trees for causing the wildly eccentric wind directions at the three locations of prime interest: the twelfth tee and green and eleventh green.

In 2002 the Augusta National placed a call to Alan Davenport. His Boundary Layer wind tunnel researchers had never done a golf course before. It was hard to resist, so they took the commission.

The first trick was to construct a model of Amen Corner they could use in the lab. From topological maps, photographs, and sketches, they put together a 1:200 model made of high-density foam. Fairways were made of drywall compound; more than six hundred trees were constructed with sponge branches and wire-and-foam trunks; Rae's Creek was acrylic over a foam base, with sil-icone to replicate wind-induced wavelets. Tiny people were added for scale. Stage two was to model the surface wind speeds and directions. Data were available from a nearby airport dating back to 1949; from the raw data they extrapolated seasonal and annual frequency histograms corrected to the standard meteorological height of thirty-three feet. From this, wind speed and direction probability distributions were plotted for the full year and the month of April, when the Masters is played.

The next step was to plot the trajectory of a ball. They chose an eight iron for the twelfth tee. Data from golf ball manufacturer Maxfli showed that such a shot typically lasts just over 5.2 seconds, and its arc is known. The trajectory was shown by a thin copper wire coated with titanium-tetrachloride; a current fired through the wire created the necessary smoke.

Then various wind directions were simulated, together with the necessary gustiness, and the results plotted on digital video through a technique called stereoscopic particle imaging velocimetry. Because of the gustiness, which in reality is greater for the first few hundred feet of the ball's flight near the ground, no two visualizations were ever quite the same.

Still, the videos clearly illustrated the conflicting information anxious golfers must deal with and confirmed that the anecdotal evidence of wind behavior at Amen Corner is, in fact, largely true. The wind changed significantly along the shot trajectory. Near the twelfth tee, the wind is either directly in the golfer's face or slightly in the direction of the eleventh fairway. But near the peak of the trajectory the wind is moving more closely in the direction of the thirteenth fairway. At the twelfth green is a swirling flow with low wind speeds. If struck in the direction the golfer senses must be right, the ball will travel true for about a third of its path, about fifty yards. Then suddenly it would be hit by a strong crosswind from the left, carrying it depressingly out of plane into the thickets to the right of the fairway. For the last third of its flight these strong cross-winds diminish—they are still there, but weaker. Not one of these directions mirrored the prevailing winds of the day. As a consequence, each shot a golfer makes from the tee will be slightly different because of the natural gustiness. And that very gustiness made accurate prediction almost impossible.

The laboratory had provided detailed evidence that local topography influences wind patterns, and had shown how. A nice byproduct was to provide architects, landscape designers, and farmers with their windbreaks and snow fences with more evidence that careful planting can mitigate wind damage and protect both buildings and crops.

The Augusta National, for its part, could solve the problem of Amen Corner by removing the trees, as they had suspected. But they had no intention of doing so. To watch the world's most skilled golfers turn occasionally into the rankest hackers was much more fun than getting out the chainsaws for a litde silviculture.46

This was a classic case of "seeing" the wind through instrumentation and devising ways to deal with it. In just such ways, engineers can predict worst cases and best cases and use probability theory to devise protections against bad outcomes and uses for the good ones. Predicting when they'll happen, and at what intensities, however, turned out to be far more difficult. Variables can be subtle and hard to see; effects can be dramatic.

Weather forecasts are erratic for very good reasons.