Water from Heaven

Water circulates incessantly between heaven and Earth, or rather between ocean, air, and land. From tropical sea surfaces warmed by abundant sunshine, water evaporates, and the wind carries the water in the gaseous state toward the equator. The warm humid air rises, and as it rises it cools, and once the temperature falls below the dew point, the water vapor condenses. Where the winds converge and strong convection lifts air upward, towers of cumulonimbus clouds rise to altitudes of 15 km (49,000 ft.) or higher, marking what is often called the “meteorological equator.” The water comes back down in the form of rain, and the air that rose, now dry, must also descend. Much of the descent occurs locally in the same storm system, in strong downdrafts that pilots try to avoid. Updrafts and downdrafts inside tropical convective clouds can have speeds as high as 10 meters (33 ft.) per second. Overall, however, upward motion dominates in the zone of maximum heating. In this zone shifting north and south of the equator, more or less following the Sun, more air is moving up than down.

Although air rises in the equatorial zone with an average upward motion of only a centimeter (less than half an inch) per second, it must be replaced. Replaced it is by the convergence of the trade winds coming from the tropical latitudes north and south of the equator. Passing over the sunny warm tropical seas, these low-level winds carry abundant moisture and with it latent heat of evaporation. Although they sometimes use the term “meteorological equator,” meteorologists prefer to call this the “intertropical convergence zone” or ITCZ. With the moisture-laden trade winds converging on the ITCZ, the only direction the air can take is up, and in another sense the upward motion sucks in the converging winds. The explanation may appear to involve circular reasoning, but atmospheric circulation is what this is about. The first cause is the geographical distribution of sunshine, but beyond that it is futile to try to distinguish causes from effects in this system of interrelated processes.

Following a mass of moist air as it rises above the equator, pressure decreases, for atmospheric pressure corresponds precisely to the weight of the air above. Temperature also decreases with altitude, generally between 5 and 8°C per kilometer (roughly 3 to 4°F per thousand feet), and so the rising air eventually reaches the dew point level at which its moisture can no longer remain entirely in the gaseous state. The water vapor condenses, forming the liquid water droplets or ice crystals that constitute clouds. The condensation releases latent heat, warming the air and further stimulating its ascension, giving rise to the towering cumulonimbus clouds of the ITCZ that climb as high as 18 km (60,000 ft.). Having gone up part or much of this way, most of the condensed moisture comes back down as rain. The remaining dry air, having reached the uppermost layers of the troposphere, must leave the equatorial zone to make way for the rising air masses that follow. At the tropopause, the upper limit of the troposphere, the air flow diverges: some goes north, some south, the flow generally being stronger toward the winter hemisphere. Finally, the air having risen over the equatorial zone descends in the neighborhood of the Tropics, near 23½° north and south latitude, in a slow motion called atmospheric subsidence. As the dry air descends, its temperature rises; its already low relative humidity becomes even lower, practically excluding any chance of condensation or precipitation. Zones of atmospheric subsidence are zones of aridity, and most of the world’s great deserts lie there, both north (the Saharan and Arabian Deserts, near the tropic of Cancer) and south (the deserts of Namibia and Australia, and the Atacama Desert of northern Chile, all near the tropic of Capricorn).

With air piling up along the Tropics, surface atmospheric pressures are slightly higher than elsewhere, and from this band of atmospheric subsidence and high surface pressure, the air spills out in divergent motions in the lowest layers. This accounts for the trade winds over the oceans, and the dry harmattan wind of the Sahara, completing the loop. On a north-south cross-section of the atmosphere along a meridian, one can sketch the atmospheric motion as two cells circulating north and south of the equator (fig. 7.1), called Hadley cells in honor of the English scientist George Hadley, who first imagined this arrangement to explain the trade winds.¹ To sum up: warm moist air rises in the equatorial zone, losing its moisture in abundant rainfall; the dried-out air at upper atmospheric levels moves away from the equator and descends in the high-pressure zones near the tropics of Cancer and Capricorn; once at low levels, much of that air then returns to the equatorial zone in the trade winds, picking up moisture from the sunny warm seas along the way. Actually, this Hadley circulation is stronger by far in the winter hemisphere around the winter solstice. Nevertheless, both cells appear in the circulation of the atmosphere averaged over all longitudes.

FIGURE 7.1 The general circulation of the atmosphere. On the left, a very schematic representation of near-surface winds averaged over the year, with a cross-section showing how mean meridional (north-south) motions vary with height in the troposphere. Typical positions of the polar jet streams are marked ⊗. On the right, typical cloud cover, also in cross-section. Vertical scale in the cross-section is highly exaggerated; it corresponds to about 10 km near the poles and 18 km near the equator, whereas the Earth’s radius is 6,370 km. The “polar front” corresponds to the mean trajectory of midlatitude cyclonic disturbances (the average storm track).

On average, air moves toward the equator near the surface, away from it at upper levels, but not just in a north-south or south-north direction. Near the surface, the air tends to follow the spin of the solid Earth, 40,000 km in twenty-four hours, over 1,000 miles per hour from west to east. As the air masses having reached the tropopause high above the equator move away either north or south, they follow the curve of the Earth and so approach the Earth’s axis of rotation. In such a case, the law of conservation of angular momentum requires that they speed up in their motion eastward around the Earth’s axis, just as a the spinning of a figure skater speeds up when he brings his arms closer to his body. On the other hand, near the surface where the air masses moving toward the equator necessarily move further away from the Earth’s axis, they must slow down in their eastward motion, and relative to the solid Earth they are deviated westward. Thus north of the equator, the southward drift of air gives trade winds blowing from the northeast, while to the south, northward drifting air gives the trades from the southeast. The Coriolis² force that deviates all motions to the right in the Northern Hemisphere, to the left in the Southern, must be included when writing the equations of motion of the atmosphere (or artillery shells or anything else) on the spinning Earth.

The Hadley cell does not stretch all the way to the poles, and trade winds don’t blow over all the seas. The Earth turns, and with it everything else, oceans as well as atmosphere. The strong solar heating in the latitude band around the equator gives rise to the Hadley cell overturning: air rises from the surface to the tropopause, then moves either north or south before descending and drifting back toward the equator. The Coriolis force due to the Earth’s rotation transforms the north-south atmospheric drift into the northeasterly and southeasterly (westward blowing) trade winds near the surface. However, neither the overturning of the Hadley cell nor the trade winds can extend all the way to the poles. Generally, sunshine is stronger near the equator than near the poles, but at any location the Sun spends as much time to the west as to the east. East-west winds must cancel out over the whole planet. Trade winds blow systematically from the east, exerting a frictional force that tends to slow down the solid and liquid Earth’s rotation. Somewhere on the globe, winds must blow in the other direction. Indeed they do, tending to come from the west in the zone of middle latitudes, these westerlies compensating the easterlies of the tropical zone, in a sort of global-scale gyration. In reality, things aren’t so simple. Especially in the middle latitudes, weather systems take the form of moving smaller-scale eddies, with strong rotating motion around traveling low-pressure centers. New Englanders keep a lookout for the development of what are known as “nor’easters.” The winds in the most dangerous sector of the rotating storm tend to be those loaded with water from the Atlantic and blowing from the northeast before the center of the storm itself arrives from the south or southwest; and in the winter, the water may come in the form of heavy snow. Other aspects of atmospheric dynamics involve concentrating and channeling a considerable amount of momentum into narrow fast-moving currents of air—the jet streams—and as a result it can take an hour or an hour and a half less to fly a Boeing or Airbus east from New York to Paris than it does to fly back. Although the jet streams flow in a sense above the weather, they play an important role in the birth and development of atmospheric “waves” and the resulting traveling low-pressure weather systems.

The trade winds, pushing surface waters of the tropical seas to the west, tend to slow down the Earth’s spin. Counteracting this, the strong prevailing westerlies and “wild west winds” of the middle latitudes drive ocean surface waters eastward and accelerate the planet’s rotation. Sailing south toward Antarctica means confronting the “roaring forties,” the “furious fifties,” and the “screaming sixties” where no land masses can interrupt the westerly air flow and the eastward drift of the Southern Ocean. The global-scale pattern of prevailing winds, easterlies in the Tropics, westerlies at middle latitudes, drives the pattern of surface currents, circulating in gyres both in the Atlantic and the Pacific (fig. 7.2).

In the Tropics, both south and north of the equator, surface ocean waters, driven by the trade winds, drift for the most part to the west. However, unlike the Southern Ocean south of Cape Horn where no land masses block the eastward drift, the tropical seas are limited by continents and divided into the Atlantic, Pacific, and Indian oceans. In the Atlantic, waters of the equatorial current warm up as they cross the Atlantic (running up against the Americas, where they exert a force slowing down the Earth’s rotation). The water must go somewhere, either north or south. A northern branch circumvents Cuba and makes a tour of the Gulf of Mexico, exiting to form the Gulf Stream off the southeast coast of the United States. Moving north, the current comes under the combined influence of the Coriolis force and the prevailing westerly winds; and starting at Cape Hatteras and the North Carolina coast, it turns gradually away from America and finally crosses the Atlantic in the direction of Europe. During the entire crossing, Gulf Stream water transfers warmth and humidity to the air above, and they are picked up by the west winds. As a result, Ireland and Brittany are blessed with mild moist climate, quite a contrast from the icy rigors of Labrador and northern Quebec at the same latitude on the other side of the Atlantic. The great Prussian and European naturalist and explorer Alexander von Humboldt noted that with the discovery of the New World, the observation of this difference between the eastern and western shores of the Atlantic constituted an important new contribution to the understanding of climate, until then focused on north-south contrasts.³

FIGURE 7.2 The main ocean surface currents. (1) Gulf Stream; (2) Kuro-shiyo; (3) North Pacific equatorial current; (4) equatorial countercurrent; (5) south equatorial current; (6) West Wind drift; (7) Peru current; (8) Labrador current; (9) Greenland current; (10) North Atlantic drift; (11) Brazil current; (12) North Atlantic equatorial current.

The Gulf Stream of the North Atlantic carries back east the waters that were driven to the west in the Tropics, and similar gyres dominate ocean circulation in the North Pacific with the Kuro-Shiyo, in the South Pacific, and in the South Atlantic oceans. Under the trade winds, the water heated to 30°C (86°F) reaches a western shore (an east coast) where it leaves the tropical zone going either north or south. It then moves to higher latitudes and crosses the ocean from west to east, finally bringing warmth and humidity to the west coast of another continent. There, part of the flow continues on to subpolar latitudes, while part turns back in the direction of the Tropics, becoming a cold west coast current. In these gigantic gyres flow millions of cubic meters of water per second (in fact, south of Nova Scotia, 150 million cubic meters per second, 750 times the Amazon River’s flow). Near its start off Florida, Gulf Stream waters stand out clearly from neighboring ocean waters, in color as well as in temperature, flowing at several kilometers per hour (a few knots). At that speed, a bottle tossed in the water can cross the ocean in a few months. In contrast, the water hardly moves at all in certain areas within the gyre, in particular in the Sargasso Sea, where all sorts of debris accumulate.

Both the color and the temperature of seawater can be monitored from space, provided that no clouds block the view. The color of reflected sunlight gives information on biological activity—more precisely, on chlorophyll of living algae and on suspended sediments. It can be measured by radiometers with appropriate filters, a difficult measurement requiring that the weak reflection be distinguished from sunlight scattered by air molecules (the blue of the sky). Measurement of sea surface temperature (SST for specialists) depends on radiometers sensitive in “window” wavelengths of the infrared for which the atmosphere is transparent, and it also requires cloud-free skies. With such observations from space, scientists can keep track of phenomena such as the formation of cold-water and warm-water rings, eddies that spin off on each side of the Gulf Stream. Also, locating warm-water and cold-water fronts helps fishermen to go after certain species such as tuna that have particular temperature preferences.

Observation from space has revolutionized our knowledge of the world’s oceans, but it cannot provide all the information needed to understand how the ocean works. Oceanographic measurement campaigns and in situ observations from instruments on buoys remain essential, and other types of observations can be imagined. No joke: a bottle can carry a message, and in the course of organized measurement campaigns of the last few decades, oceanographers have tossed well over 100,000 bottles into the sea, recovering 2 to 3 percent of them. And there have been a number of unplanned but welcome experiments as well. On May 27, 1990, the Korean containership Hansa Carrier, bound for the United States, met up with a huge wave and lost part of its cargo of Nike Air sports shoes. The containers washed away opened, releasing more than 60,000 shoes in the sea; but because of their air soles the shoes didn’t sink, and after several months afloat they began to reach the shores of Canada and the northwest United States (and in fairly good shape, but as single shoes rather than as matched pairs). To take advantage of this unexpected bounty (some insurance company’s loss), beachcombers organized exchanges so as to match left and right shoes. Once oceanographers heard of it, they set about collecting information on the dates and places where the shoes washed ashore. It was a bounty indeed: releasing 60,000 “tracers” at a single time and place would have been far too expensive an experiment, but here commerce and nature cooperated without asking for money from the oceanographers or their funding agencies. Another such unplanned experiment took place two years later, this time a cargo of plastic bathtub toys (floating ducks, frogs, and turtles) usually meant for baby’s bath. These acts of God unwelcome to shippers and their insurers came as a maritime manna for oceanographers, providing enormous data on the details of the fine structure of the Kuro-Shiyo and other currents of the North Pacific. Who knows? Maybe some of the toys will succeed in crossing the Arctic and finally reach the North Atlantic. Some bottles released in the ocean at Nome, Alaska, in 1900 actually made it through the Bering Strait and reached the Irish and Norwegian coasts some ten years later. If you come across an interesting object on the beach, tell your favorite oceanographer about it. In addition to the major ocean gyres, all sorts of smaller-scale eddies, currents, and (mostly equatorial) countercurrents complicate the picture, not to mention the complexities of the third vertical dimension. We’ll dive into that further on.

Sailors early learned to count on the reliability of the trade winds, but tropical seas do not stretch round the Earth. Interrupting the oceanic expanses, land masses complicate the simple picture drawn so far. In the course of the summer, land heats up more quickly than the oceans, this for a variety of reasons. First of all, the warming produced by absorption of solar radiation cannot penetrate far into the ground; it only affects a layer limited to several feet in depth. With a correspondingly small effective heat capacity, the land surface temperature varies strongly between night and day, even more between winter and summer. By comparison, tides and the winds agitate the liquid ocean, mixing the relatively thick surface layers. Only in extremely calm seas does the surface “skin” temperature rise significantly during the day to fall at night. Everywhere else, solar radiation must in effect warm a layer of water mixed down to at least 10 meters (33 ft.). In addition, especially where winds are strong, evaporation from the sea surface carries away latent heat, leaving less energy for warming the water. As a result, sea surface temperatures rarely rise above 30°C (86°F). At night, and during the winter, sea surface cooling proceeds slowly because, as the uppermost layer cools, the denser water sinks and is replaced by deeper not-so-cold and dense water. The resulting mixing of the surface layer can extend down as deep as 50 fathoms (300 ft. or nearly 100 m) and so involves enormous masses of water. As a result, variations of sea surface temperature are usually moderate and relatively slow.

On land, by contrast, nothing stirs up the soil: in the topmost few inches, temperature can climb strongly during the day, especially for dry land with no loss of latent heat by evaporation. As a result, summer daytime temperatures reach extremely high values, often more than 50°C (122°F) in the deserts of central Asia, the Sahara, Mexico, and the southwest United States. The high temperatures lead to the development of centers of lower atmospheric pressure, and the low-pressure centers attract cooler air masses from the neighboring oceans, where temperatures do not rise so much and atmospheric pressures remain higher. Over the Indian Ocean, the development of stronger and stronger contrast in surface temperature and pressure triggers the phenomenon called the monsoon, a name given by Arab navigators of the Middle Ages because the reversal in wind direction over the Arabian Sea actually reverses the flow in the Somali current east of Africa. The trade winds blowing from the southeast in the southern part of the Indian Ocean, drawn north of the equator by the low pressures, veer to the right because of the Coriolis force, and end up as southwesterly winds. The zone of towering clouds associated with rising warm humid air masses is driven by the monsoon winds further and further north of the equator, in the direction of the Indian subcontinent.

Reaching India, the thick water-laden clouds burst over the hills of the Western Ghats lining the coast near Bombay, but enough moisture remains in the air flow for monsoon rains to water the interior as far as Delhi and beyond. Once it reaches the foothills of the Himalayas, the moist air must rise and cool; and this produces a true deluge, with sometimes as much as 20 meters (780 in.) of rain falling at Cherrapunji in Assam. The great summer monsoon, the southwest monsoon, is a lifegiver, essential for food production for most of India. If the monsoon is good, everyone is happy and smiling, except perhaps for the European tourist who complains because it’s raining in August. But late or weak monsoons are a catastrophe. Famine now can be avoided by assistance and commerce, transferring grain from areas with surpluses (mostly North America, Europe, and Australia) to those where crops have failed, but many problems remain. Hoarders and speculators may benefit more than those who need the food, import of “dumped” surpluses discourages local farmers, and grain is of use only if it can be transported without spoilage to the consumer.

Note that the Indians often speak of the “winter monsoon,” referring to the winds that blow from the northeast in November and December, bringing rain to the southeast coast of India, in particular to the coast in the states of Madras and Pondicherry. These regions don’t get much water from the summer monsoon, the southwesterly air flow losing most of its moisture as it crosses the highlands of the subcontinent. But is “monsoon” the right word? In fact, these are the northeasterly trade winds, part of the “normal” air flow of the Tropics, reinforced however by strong outflows of cold air from the Siberian winter anticyclone. Many specialists prefer to reserve the term monsoon to refer to the flow of warm moist air from the ocean, across the equator, toward the land, and this also occurs in West Africa, where the arrival of moisture in the low-level monsoon is crucial for survival in the Sahel. A somewhat similar situation occurs in North America, even though it is quite far from the equator. Warm moist air masses from the Gulf of Mexico bring rain to areas in the interior of the continent far from the Atlantic and separated from the Pacific by the high barriers of the coast ranges, the Sierra Nevada, and the Rockies. And on much smaller scales, contrasts between land and sea—specifically, stronger faster daytime warming and nighttime cooling of the land, and higher thermal inertia of the water masses—drive the daytime sea breeze and nighttime land breeze familiar to visitors to the seashore. Fair-weather cumulus clouds form as warming air rises over the land during the day, and dissipate at night when the warming ceases and air begins to subside. The opposite occurs over the sea. Similar alternation of cloudy and clear skies also occurs between lakes of large area and their surroundings. In Africa, Meteosat (infrared) images show Lake Victoria covered by clouds at night, clearing up in the day as clouds form over the surrounding land.

The Pacific Ocean, with eleven time zones separating the shores of Chile and China, covers more than a third of the Earth’s surface, spanning more than 7,000 miles at the equator. More than 7,000 miles west of the coast of Ecuador, the equatorial Pacific bathes the Indonesian archipelago (often called the “maritime continent”), and further west, the equatorial zone traverses the Indian Ocean. From the shores of East Africa to the coast of Ecuador and Peru, this enormous oceanic domain and, above it, the atmosphere partly blocked by the Andes Mountains, pulsate with El Niño and the Southern Oscillation. Why these two names? In fact, until the 1950s, oceanographers knew El Niño in eastern Pacific waters, and meteorologists knew the Southern Oscillation of atmospheric pressure over the southern equatorial and tropical Pacific, as separate interesting phenomena. Only starting around 1960 did they recognize that the two phenomena were in fact connected parts of a single wider-scale pulsation, recognizing more recently the distinctive “personality” of La Niña. Integration of oceanography and meteorology continues as observational tools are improved and coverage extended to the entire globe, as vast increases in computer power make possible the modeling of coupled atmospheric and oceanic processes.

Fishermen along the coasts of Peru and Ecuador have always known about El Niño. Every December, as they prepare to celebrate Christmas, the birthday of El Niño (the little boy, the infant Jesus), the trade winds tend to weaken, and warm waters take the place of missing upwelling cold water from the depths of the Pacific. Fish are harder to find. It’s a good time to rest, repaint the boats, and enjoy the festivities of Christmas and the New Year. But sometimes the vacation drags on: the trade winds don’t come back, coastal waters become warmer and warmer, fish rarer and rarer. For fishermen, such an El Niño warm event constitutes a calamity, and repercussions extend to prices in the world grain markets because of increasing dependence on fishmeal as an alternative to grain for livestock feed. El Niño also decimates the bird populations that feast on fish, and that reduces their droppings of guano on the rocky islets where they roost. In the nineteenth century, when these dejections constituted an important and valuable source of fertilizer, each El Niño event produced a steep increase in market price. With warm waters invading the coastal zone, rainfall over nearby usually arid land increases enormously. El Niño may be a disaster for the fishermen, but it brings a year of abundance (año de abondancia) to farmers, insofar as they can keep flooding under control. Written records of these dramatic events have been kept since the arrival of the Spanish conquistadors in the sixteenth century. Analysis of layers of snow and ice accumulation in the glaciers of the Andes shows that such strong fluctuations of precipitation have been going on for several thousand years at least.

It is fairly easy to understand how the warm waters of El Niño lead to reduction of fish stocks (at least of the commercially interesting species, anchovies in particular). In the early 1800s, the great explorer Alexander von Humboldt described the cold current flowing north along the Pacific coast of South America. Sometimes called the Humboldt current, but officially known as the Peru current ever since oceanographers decided no longer to use the names of persons, this is part of the great gyration of South Pacific ocean currents. Near the equator, surface waters are usually driven west by the easterly and southeasterly trade winds. Off the coast of South America, the westward drift has a pumping effect, producing upwelling of cold water from the ocean bottom to replace the warmer surface water. Thanks to this upwelling of bottom water rich in nutrients, marine life thrives: the nutrients fertilize the growth of phytoplankton or algae, the zooplankton (microscopic sea animals) graze in these marine meadows and are eaten in turn by fish, with birds and humans feeding on the fish. However, when the trade winds weaken, the upwelling dies down. But why should the trade winds weaken?

Between 1900 and 1930, many centuries after Peruvian fishermen first discovered El Niño, an English meteorologist, Sir Gilbert Walker, director of the Indian Meteorological Department, was trying to find a way to predict failures of the Indian monsoon. After an exhaustive study of measurements made at weather stations all around the Pacific basin, he established the existence of coherent variations of atmospheric properties over all of this vast domain, a phenomenon that he named the Southern Oscillation.⁴ Near the equator and in the Tropics, whenever surface atmospheric pressure in the western Pacific (for example at Darwin, northern Australia) falls below its average value, higher-than-average pressures prevail over the eastern Pacific (Tahiti, or Easter Island); and vice versa. In the western equatorial Pacific, low pressures go with intense convective activity, strong ascending motion of warm moist air, towering cumulonimbus clouds, and high precipitation. Higher pressures correspond to much less of this, indeed to drought. A sort of seesaw of atmospheric pressure variations spans the vast domain of the equatorial Pacific, with effects extending still further east and west along the tropical belt, and also north and south over all the Pacific. Today we consider the phenomenon to be of global scale.

Let’s not get ahead of ourselves! At the time, meteorologists were far from realizing all the implications of the Southern Oscillation. Even today, they are far from being able to produce reliable forecasts of the fluctuations of the Indian monsoon, and that was Walker’s original objective. The east-west surface-pressure seesaw goes along with a shift of the area of strong convective activity from Indonesia in the west toward the central and eastern Pacific, and back again, interpreted in terms of what is now called the “Walker circulation.” In the so-called “normal” situation, low atmospheric pressure, strong convective activity, and abundant rainfall are observed over Indonesia, while at the same time to the east atmospheric pressure is higher, cloud cover reduced, and hardly any rain falls at all. This goes together with strong trade winds blowing to the west near the surface, with return flow to the east at upper levels of the atmosphere, somewhat like a Hadley cell oriented east-west rather than north-south. This Walker circulation cell involves strong uplift of warm humid air and accompanying precipitation over the Indonesian archipelago, and subsidence of dry air over the central and eastern equatorial Pacific. The subsidence of dry air with virtually no rain is perfectly consistent with the extremely arid conditions that usually prevail over the few islands to be found in this part of the Pacific, as well as over the coast of Peru. In the other phase of the Southern Oscillation, with higher pressure to the west, flow in the Walker cell is reversed: subsidence and aridity over Indonesia, ascending air and anomalously high rainfall in the east.

It was only in 1957 that Jacob Bjerknes, a Norwegian meteorologist settled in California, recognized that El Niño on the one hand, and the Southern Oscillation on the other, were in fact two aspects of a single phenomenon involving strongly linked oceanic and atmospheric processes (fig. 7. 3), today often called ENSO (El Niño/Southern Oscillation) by specialists. With pressures high to the east, the trade winds are stronger, vigorously driving the surface ocean waters to the west, thus activating the pumping and upwelling of cold bottom waters in the Peru current. Together with reduced evaporation from colder waters the subsidence of dry air practically rules out rainfall over coastal Peru and the eastern and central Pacific. Drifting to the west under the strong trade winds, warmer surface waters pile up in the western Pacific, and sea level there is several tens of centimeters (a few feet) higher than to the east.

FIGURE 7.3 The Southern Oscillation: El Niño (bottom) and La Niña (top). The two panels show schematic east-west cross-sections of the Pacific Ocean and the atmosphere above it, along the equator, with vertical scales highly exaggerated. The thickness of the warm water layer (shaded) is only a small fraction of the depth of the ocean, and the tilt of the sea’s surface, though real, is very small. Surface atmospheric pressure is above the average in the areas of subsidence and drought, below average in areas of convection and rain.

Sea-level measurements were once only possible along coasts, using tide gauges. Some such gauges have long been installed in Pacific island harbors, but there are precious few islands in the central and eastern Pacific. The advent of satellite radar altimeters has completely changed the situation, and beginning with the Franco-American Topex-Poseidon oceanography satellite launched in 1992, data on sea level are available over all of the oceans with an accuracy of a few centimeters. In the 1960s it had already been noted that sea-level changes between east and west tended to follow the Southern Oscillation’s seesaw of atmospheric pressure, and the satellite data provide much more precise and comprehensive coverage. In the “normal” situation, with high pressure to the east, water warmer than 30°C (86°F) piles up in the west, blown there by the trade winds strengthened by the “attraction” of the lower pressures over Indonesia, the maritime continent. But at some point, this pileup can no longer continue. Triggered perhaps by a burst of westerly winds, a sort of giant wave of warm water moves off to the east, inexorably invading the eastern equatorial Pacific all the way to Peru. El Niño is on his way, and his extended stay in Peruvian coastal waters always corresponds to the phase of the Southern Oscillation with pressures higher in the west and lower in the east. There, the trade winds are weakened by the repulsive effect of the stronger pressures to the west, and weak westerly surface winds take their place. The layer of warm water invading the eastern equatorial Pacific prevents cold bottom water from reaching the surface while at the same time providing humidity to the local atmosphere by enhanced evaporation. The humidity added to the atmosphere condenses when the warm air rises, releasing latent heat, stimulating more rising motion and further lowering the surface pressure; and that reduction of surface pressure in the east further weakens the trade winds. The atmospheric and oceanic processes reinforce one another, and El Niño extends his stay. What is the cause, which is the effect? What triggers the El Niño event, and how and when will it come to an end? In fact, the same set of oceanic and atmospheric processes can operate in the other direction, restoring strong easterly trade winds, the warm water moving back west, and the resumption of the upwelling of cold bottom water along the coast.

Up to the end of the 1980s, the tendency was to consider the El Niño events as a sort of anomalous state, returning on average twice a decade, to be distinguished from the “normal” situation. However, the El Niño or ENSO warm event is part of an oscillation, the Southern Oscillation, with El Niño as an extreme phase of that oscillation. What about the other extreme phase? In 1988, for example, a strong east-west difference of pressure was observed, with unusually high values to the east just off Peru, and temperatures at the surface of the eastern Pacific significantly lower than “normal.” And it seems now well established that the shift of the atmospheric jet stream that accompanied this extreme ENSO phase (with cold water off Peru) played a major role in the extremely hot and dry conditions that prevailed in summer 1988 over much of the United States, which some had hastily attributed to “global warming.” After some hesitation regarding the name to give to cold events of this type, the scientific community seems to have settled on the name of La Niña (the little girl).

As we shall see, the ENSO (or better still, ENLNSO?) phenomenon has strong impacts on regional ecology, the regional and world economy, and society. Are reliable forecasts of the extreme phases possible? There certainly is a need, and there has been significant progress in the research and operational weather service community. Following the very strong El Niño of 1982–83, major efforts were made to strengthen monitoring of the Pacific. A comprehensive network of buoys was set up and has since been operating, with some buoys drifting, some fixed, some making measurements at depth as well as at the sea’s surface, all along the equatorial band of the Pacific, transmitting their data by way of communications satellites. At the same time, new tide gauges were put into operation at many additional sites, and new satellites have provided more and more data. An important first step in the research was to identify some advance signals of particularly strong El Niño or La Niña phases.⁵ After all, the warm water mass takes several months to cross the Pacific, before the warm or cold event or “anomaly” reaches its peak. Does the “advance signal” tell us something about the cause of these phenomena? Is there in fact a cause to be identified? After all, in the ENSO phenomenon, one observes oscillations around an average state called “normal.” On a year-to-year graph of the Southern Oscillation Index (the difference between atmospheric pressure in the east and in the west of the Pacific, e.g., between Darwin and Tahiti), the curve goes above and below the average value, sometimes more, sometimes less, sometimes hesitating, but it is virtually never steady. The combined ocean-atmosphere system of the Pacific basin is hardly ever in its normal state, if by that its average state is meant.

Let me propose another way to look at this oscillation. Let’s go to the park. In the Parc Montsouris near my home in Paris, I occasionally see and hear Spanish-speaking children at play, because it’s just across from the international student campus. In New York’s Central Park or Riverside Park, near where I used to live, you can easily find Spanish-speaking children, a little boy and a little girl, El Niño and La Niña, together on a seesaw. You don’t know how long they’ve been there. At one moment, El Niño is up, then La Niña. Sometimes they teeter a bit, nearly balanced in a horizontal position, but they never stay that way long; that may be the average state of the seesaw, but it’s not normal in the sense that nearly half of the time one side is up or going up, the other down, and vice versa. Occasionally, La Niña will give herself a strong kick to get up high quickly; or maybe it’s El Niño who gave the strong kick, or who shifted position closer or further away from the central axis. Sometimes, El Niño seems stuck longer on high. How can you explain all these ups and downs? You don’t know who started things off, but between seesaw mechanics and child psychology, the kids keep going up and down. Of course, if one of them is much heavier than the other, a parent is needed. But just as for the seesaw in Central Park, for an overall explanation of the ENSO (ENLNSO) phenomenon, rather than look for an external cause, research tries to understand how the many interactions between atmospheric pressure, wind, sea level, atmospheric temperature and humidity, and water temperatures at different depths combine to keep the oscillation going. During the last ice age, 18,000 years ago, the ENSO seesaw didn’t work quite the way it works today, but for the last 10,000 years El Niño and La Niña have been at it, up and down, a bit more, a bit less, a bit faster or more slowly.

This section started off with the El Niño warm events and the associated decimation of anchovy populations off the coast of Peru, but this is only a part of the enormous oceanic and atmospheric fluctuations affecting the entire Pacific basin and beyond. Reversal of the east-west atmospheric motions in the Walker cell necessarily affects circulation between the equator and the Tropics in the Hadley cell, with additional consequences beyond the Tropics, over the whole Pacific at least. In particular, the changes in the distribution of atmospheric pressure lead to the formation of more typhoons traveling to islands such as Tahiti and Hawaii normally not visited by such storms. The shift of the jet stream sends one winter storm after another in the direction of California, and with a little bad luck, luxury sea-view homes, built on slopes that are typically threatened at least once a decade, end up on the beach or in the sea below. Maybe the insurers will pay, but the rates will go up.

In Indonesia, extreme El Niño events bring drought, as in 1982–83 and 1997–98, and any fire, in particular those set to clear forests for palm oil plantations, can easily go out of control, burning vast areas and sending stifling smoke to pollute cities as much as a thousand miles away. In the eastern Pacific, suddenly abundant rains (ten times the average!) fall on usually desert islands, radically transforming the landscape as luxuriant vegetation quickly takes advantage of this water from heaven. It becomes a veritable deluge in Peru (3 m, i.e., 120 in. of rainfall, in six months!), Ecuador, and Colombia. Some farmers know how to profit from this (planting rice instead of cotton, for example), but the high humidity and flooding raise serious safety and health risks.⁶ And further off? Walker was aiming at forecasts of Indian monsoon failures, and statistically, it seems that the monsoon is indeed weaker during the El Niño years. In any case, the effects of the changes in atmospheric circulation and ocean temperatures extend across the Indian Ocean to the East African coast, and depending on the season, drought afflicts the south and east of Africa from Ethiopia to Mozambique.

And still further away? The atmosphere knows no borders, so that everything is related, but that doesn’t mean that all the “teleconnections” that can be imagined really exist or are significant if they do. In many cases, no chain of cause and effect can be established, although statistics suggest a link. But the world is round, and it would be astonishing if an upheaval on such a vast scale as ENSO were to stop at the Andes. In the tropical zone, when El Niño rains drench the western slope of the Andes, drought reigns to the east, in the Amazon basin of Brazil. The strong perturbation of the Hadley cell over the Pacific and Indian Oceans also has effects over the Atlantic, although it is by no means the only factor governing variations there. Does a strong El Niño constitute a global disaster? It’s certainly a gigantic upheaval over and in the Pacific. Billions and billions of tons of warm water go on the move, the locations where water evaporates (and where now fresh water falls as rain) shift by thousands of miles, winds change direction, and all this happens in a matter of months. Elsewhere on the globe, less dramatic but nonetheless real and important changes are observed. Nevertheless, the catastrophe, in the mathematical meaning of a radical change of state, is not necessarily a worldwide calamity. No doubt El Niño hurts Peruvian fishermen, as it does those improvident Californians whose perilously perched houses slip into the sea. But for prepared Peruvian farmers, El Niño rains bring abundance. For the United States as a whole, the balance sheet must be worked out. The associated distribution of atmospheric pressure and the route taken by the jet stream tend to keep hurricanes away from the U.S. East Coast and the Caribbean, otherwise strongly under threat. Accountants tell us that the gains (or rather, the absence of losses) by residents of Florida and their insurance companies more than make up for losses in California, an El Niño benefit seldom mentioned by the media. Moreover, the El Niño weather teleconnections generally bring mild winters to the northeast United States. Cold-related deaths as well as heating fuel costs are reduced. During the exceptionally mild 1997–98 El Niño winter, in place of unneeded heating oil that could not be sold, refiners switched to producing and selling gasoline at bargain prices, encouraging still more gas-guzzling.

Despite the headlines, El Niño does not constitute a climate catastrophe everywhere in the world. What about La Niña, the opposite phase of the oscillation? Not much in the news before 1988, La Niña also has mixed effects, both boon and bane. It’s fine for fishing off Peru and for fire-quenching rainfall in Indonesia, but drought afflicts farmers both in much of the United States and west of the Andes from Colombia to Peru. Moreover, while El Niño tends to keep hurricanes away from the U.S. East Coast, the Caribbean, and the Gulf of Mexico, La Niña brings more active hurricane seasons. Hurricane Mitch, the cause of many thousands of deaths in Honduras and Nicaragua, came in October 1998 after the end of the 1997–98 El Niño and a rapid switch to La Niña conditions.⁷ The 1999 hurricane season was extremely active, but fortunately no hurricane hit such a vulnerable target. The media have helped to make the public aware of the worldwide impacts of El Niño (especially in 1997–98), but often they give the impression that the apocalypse has come. Far be it from me to play down the damage and pain brought by violent weather events—tornadoes, typhoons, hurricanes, or wild winter storms—but we have to recognize that the climate of the Earth works through ceaselessly changing weather, winds, and waves. Like it or not, we have to live with the weather, and ignorance of its violent side is no excuse. We know the statistics fairly well, and reliable forecasts of some though not all of the most dangerous events are increasingly available. For the media, however, the saying “No news is good news” seems often to have been transformed into “Good news—or no bad news—is no news.”

Especially in the atmosphere, the “waves” made by the Southern Oscillation spread out over the globe, but in Europe, on the opposite side of the world from the South Pacific, their effects can be detected only by sophisticated statistical analysis of changing weather patterns. In Europe, and also in northeastern North America, much more noticeable effects arise from what has long been known as the North Atlantic Oscillation (NAO). As with the Southern Oscillation, meteorologists observe a seesaw of atmospheric pressures and associated wind and water variations, but in this case between north and south rather than east and west. It shouldn’t come as a surprise. Compared with the vast east-west spread of the Pacific, the Atlantic is squashed, especially between the Brazilian nordeste and West Africa, where only 1,800 miles separate the two shores. Furthermore, North Atlantic waters mix quite freely with water and ice from the Arctic, whereas the Bering Strait and the Aleutian Islands form a bottleneck between Arctic and North Pacific. In the subtropical North Atlantic, an area of high atmospheric pressure often known as the Azores anticyclone separates the zone of easterly trade winds in the Tropics, to the south, from the zone of prevailing westerly winds at northern midlatitudes. Sometimes the anticyclone spreads out and shifts in the direction of Europe, providing clear skies there while blocking weather systems arriving from the west. European vacation-planners look forward to the expansion of the Azores anticyclone, many farmers fear it. As for fishermen, they know perfectly well that they have to manage with the Iceland Low and its bad weather where Arctic air meets mild and moist midlatitude air.

Like everything else in the atmosphere, the Iceland Low and Azores High are never fixed, either in location or in intensity. Here too there is a sort of seesaw of atmospheric pressure, between atmospheric pressures over Greenland and Iceland in the north, and over the Azores Islands in the south, changing the strength of winds and the tracks of Atlantic storms. When the Iceland Low deepens and shifts north toward Greenland, when at the same time pressures over the Azores get still higher, the increased pressure difference strengthens the west winds over the midlatitude Atlantic, and also the trade winds blowing westward off the Sahara Desert over the tropical Atlantic (fig. 7.4). Additional desert dust reaches Caribbean islands such as Barbados and even Florida, and surface waters off the southeastern United States tend to be warmer than average. The stronger west winds further north pick up additional warmth and moisture from the Gulf Stream before reaching Europe. Exactly where rain falls and where skies are clear can vary, but generally this positive NAO phase involves mild and wet winters in western Europe.⁸ However, if the Azores High spreads over France, conditions can be dry and cold in the winter, dry then becoming hot during the long sunny days in the summer. By contrast, in the negative phase of the NAO, the Iceland Low is pushed south and weakened by the development of high pressures over Greenland, and also pressures over the Azores are not quite so high. Then west winds over the Atlantic weaken, but with blocking by the Azores High weakened, they can bring more winter rain to southern Europe, the Mediterranean, and North Africa. Also in the negative NAO phase, extremely cold dry air from the Arctic can reach northern Europe, giving quite cool summers and bitter cold winters, sometimes as far south as France.

From one year to the next, the winds will vary, blowing warmer or colder as they exchange more or less heat and evaporate more or less moisture while passing over warmer or cooler Atlantic Ocean surface waters. Thanks to the Gulf Stream, the waters that wash the shores of Ireland and Brittany are far less chilling than those along the shores of Labrador and Newfoundland, at much the same latitude to the west. Thanks to the warmth that northeastern Atlantic waters give up to the atmosphere and that west winds carry to Europe, winter temperatures in Dublin, London, Paris, Brussels, and Geneva are far milder than those that are the common lot of citizens of Minneapolis, Montreal, or Moscow. But by the same token, western European city-dwellers cannot count on using the subway to get to cross-country skiing or open-air natural ice-skating, one of the pleasures associated with cold winters. North of Iceland, and sometimes also to the south, floating ice forms each winter and melts the following summer, but on occasion icebergs wander off to the south without immediately disappearing. Since the sinking of the Titanic, a systematic effort is made to keep track of such drifting hazards, and in recent years satellite radar has made it possible to monitor icebergs from space even when the sky is completely overcast. The limit of the floating ice pack advances and recedes, not only with the seasons but also from year to year. Indeed, careful study of the archives of maritime meteorology has revealed that in addition to the year-to-year fluctuations, there are strong variations from one decade to another. In the Labrador Sea, maximum ice cover occurred in 1957, 1972, and 1984; and each time, the average North Atlantic sea surface temperature went through a maximum a year or two later. What can explain these variations, with warmer water following closely the maximum amount of ice? Oceanographers have some ideas but no certainties. Further complicating the issue, the North Atlantic Oscillation appears to be related to a north-south pressure seesaw between the North Pacific and northwestern North America (Alaska, northwestern Canada).⁹ But anything happening over the Pacific is necessarily affected by the Southern Oscillation and, likewise, also has an effect on it. The Earth is round, and the atmosphere knows no borders.

FIGURE 7.4 The North Atlantic Oscillation. In the positive phase, the stronger high-pressure center near the Azores and the northward shift of the low-pressure center from Iceland to Greenland strengthen both westerly winds at midlatitudes and the northeast trade winds in the Tropics. These accompany shifts in wet and dry areas, in warm and cold ocean currents, and in many other phenomena.

These north-south variations, especially those that operate more slowly than the Southern Oscillation with its switch from El Niño to La Niña in a few years, complicate the problem of determining to what extent we are already seeing the global warming trend that must result from the reinforcement of the greenhouse effect. For one thing, weather satellites have existed only since 1960, and true global coverage was not possible before then. With only four decades of such global observations, and with strong NAO changes from one decade to the next (or even periods of twenty to thirty years), it is risky to say that the true long-term trend has been unambiguously extracted from the data. The importance of the Atlantic Ocean should not be underestimated, despite its being much smaller than the Pacific. For one thing, the Atlantic actually does export water to the continents, the Americas as well as Europe and Africa, whereas overall, the Pacific imports water, with slightly more rainfall than evaporation. Even more noteworthy, the Atlantic Ocean transports heat northward not just in the Northern Hemisphere starting in the Tropics, but all the way from the Southern Ocean, across the equator, up to Greenland and Scandinavia. To learn how this happens, we have to dive into the deep ocean.