Floods, Droughts, and Climate Change

Robert H. Webb; Michael Collier

Chapter 1: Entering the Atmosphere

One would think that the atmosphere—all 5,000 million million tonnes of it—would be more conspicuous. We spend our lives within this sea of air. It permeates every void and lodges itself between every hair. We breathe it, we walk through it, and yet the atmosphere remains a mystery, invisible to us. Are fish aware that they swim in water? The atmosphere is so intimate, yet it slips like mist through our fingers. Without the extremes of wind and rain, heat and cold, we would scarcely even acknowledge the atmosphere's existence.

If we are to understand why floods and droughts occur, we must begin with the air we breathe. Climate is the fundamental link between the two. Hippocrates wrote about the effects of climate on human health twenty-four hundred years ago. Aristotle expounded on patterns of weather fifty years later. The Greeks were aware of orderly zones of climate; indeed the word derives from klima, which refers to the gradational effects of a slope. Despite the Greeks' conjectures, subsequent observers could no more comprehend atmospheric dimensions than they could grasp dreams vaguely remembered upon waking.

It wasn't until people thought to measure temperature, pressure, and motion that air in some sense became tangible. Santorre Santorio invented the thermometer in 1612, and Evangelista Torricelli built a barometer in 1643. With these tools, Robert Boyle was able to demonstrate his law of gases in 1661, which explained the relationships between pressure, volume, and temperature. Edmund Halley drew the first map of trade winds seven years later. With these maps, George Hadley offered an elementary understanding of global atmospheric motion in 1735 that stands as the basis for much of global climatology today.

The study of the atmosphere has advanced dramatically over the last one hundred years. During the First World War, Norwegian meteorologists described embattled warm or cold bodies of air separated by "fronts." The jet streams were discovered during the Second World War by high-altitude pilots who were impeded while trying to make their way westward over the Pacific Ocean. In 1957 the International Geophysical Year focused scientists' attention on the global environment. Since then, a coherent system of oceanic and atmospheric measurement stations has gradually been developed. Tiros I was launched in 1960, the first weather-observation satellite to be put into orbit. Today the Tropical Ocean Global Atmosphere Project (TOGA) network of monitoring buoys collects information over large portions of the world's tropical oceans.

The National Oceanic and Atmospheric Administration (NOAA) and similar institutions worldwide have continuously gathered objective meteorologic data for decades: wind, temperature, precipitation, barometric pressure, and atmospheric moisture. Data gathered from the oceans, atmosphere, and space are the foundation for an understanding of the interconnectedness of our evolving atmosphere. All of these tools have sharply focused our vision of atmospheric processes. Indeed, only in the last third of the twentieth century did satellite imagery allow us to comprehensively see the awesome and complete fury of a hurricane.

From one day to the next, we're touched by the vagaries of weather: extremes of heat or cold, rain or drought. Storms break and rivers burst their banks; or skies remain relentlessly clear for months as we watch animals die and reservoirs dry up. Over the years, season after season, one learns what to expect from the weather. With time, the cumulative experiences of storm, wind, and heat coalesce into a sense of climate, an important component of our sense of place. Although this "sense of place" is invaluable to us as individuals, it doesn't offer scientists or society much of a foundation for the broader picture of a global climate, much less any insight into how climate might be changing or how it affects our lives. For that we need objective data from worldwide monitoring systems.

For more than a century, the U.S. Geological Survey (USGS) has assessed river flows throughout this country. USGS scientists have maintained river gauging stations since 1895, amassing long-term records of water flow and sediment transport. These records provide an indispensable backdrop against which one can evaluate "aberrant" conditions of flooding along a particular stretch of river or drought within a given river basin. Hydrologists and climatologists have long been aware of the role of regional climate in the prediction of floods or in understanding drought. With our growing sense of a variable climate, it's appropriate to reassess these concepts of flood and drought, not as isolated events, but as phenomena connected on a worldwide scale.

This book examines floods and droughts within the wider context of climate and climate change. It introduces the concepts of global weather, puts these processes into the longer-term framework of climate, and explores the idea that patterns of climate evolve through time. With this foundation, it becomes increasingly clear that floods and droughts, once considered isolated acts of God, are often related events driven by the same forces that shape the oceans and the entire atmosphere.

Chapter 2: Cutting Edges of Climate

Extremes of weather shape us as individuals and as a society. Storms and searing heat sharpen the edges of our culture; floods and droughts hone the edges of our landscape. The hurricane of September 1900 swept six thousand people from the streets of Galveston, Texas. In 1927 close to a million people—almost 1 percent of the U.S. population—left their homes to huddle atop levies from Illinois to Louisiana as the Mississippi River flooded a combined area equal to Massachusetts, Connecticut, New Hampshire, and Vermont. Six years later, winds of the Dust Bowl began to inhale 100 million acres of western Kansas, eastern Colorado, and the panhandles of Oklahoma and Texas; the summers of 1934 and 1936 were the hottest since the turn of the century. John Steinbeck forever focused this catastrophe in the American consciousness when he wrote The Grapes of Wrath in 1939:
And then the dispossessed were drawn west—from Kansas, Oklahoma, Texas, New Mexico; from Nevada and Arkansas, families, tribes, dusted out, tractored out. Car-loads, caravans, homeless and hungry; twenty thousand and fifty thousand and a hundred thousand and two hundred thousand. They streamed over the mountains, hungry and restless—restless as ants, scurrying to find work to do—to lift, to push, to pull, to pick, to cut—anything, any burden to bear, for food. The kids are hungry. We got no place to live. Like ants scurrying for work, for food, and most of all for land.

Floods and droughts are not aberrations. Floods are initially more conspicuous than droughts because they can occur over days or weeks instead of months or years. Droughts require a more persistent weather pattern before they are recognized. We commonly assume they are end members of the spectrum of possible meteorologic conditions for a given locale. Another emerging perspective, however, is that floods and droughts arise from conditions that are somehow different than the established norm. Climate may not turn out to be a smooth continuum of meteorologic possibilities after all, but rather the summation of multiple processes operating both regionally and globally on differing time scales.

Floods and droughts are neither random nor cyclic. They may seem that way when our noses are pressed against the windowpane of the present, but they aren't. To the Dakota farmer whose lips were cracked dry during the Dust Bowl, it seemed that rain would never sweep in over the horizon. To the Honduran peasant whose home and family have just been swallowed by the hurricane-swollen Río Choluteca, dry blue skies are the vanishing memory of a former life. Up close, it's all but impossible to see the true patterns of weather.

The extremes of flood and drought occur within the context of climate, a context that is both local and global. One must understand the geography and meteorologic response of a given watershed to understand its history of flooding. One should also look beyond basin boundaries to appreciate the coherent patterns that influence weather regionally. Almost fifty years ago, Jerome Namias suggested that local droughts can be the manifestation of anomalous patterns of atmospheric circulation arising from changing ocean-surface conditions half a world away. The same is true for floods.


Drought is more than a simple lack of rainfall. Drought is a persistent moisture deficiency below long-term average conditions that, on average, balance precipitation and evapotranspiration in a given area. Not all droughts are created equal; similar moisture deficits may have very different consequences depending on the time of year at which they occur, preexisting soil moisture content, and other climatic factors such as temperature, wind, and relative humidity. Drought can be defined in terms that go beyond the meteorologist's rainfall measurements. Hydrologic drought occurs when surface water supplies steadily diminish during a dry spell. If dry conditions continue, groundwater levels could begin to drop. Agricultural drought occurs when a moisture shortage lasts long enough and hits hard enough to negatively impact cultivated crops. Soil conditions, groundwater levels, and specific characteristics of plants also come into play in this functional definition of drought. Ecologic drought is detrimental to native plants that don't have the benefit of irrigation.

The dryland farmers of southeastern Alberta know about several kinds of drought, particularly agricultural drought. John Palliser had explored that country from horseback for the Canadian government during the dry years of 1857-59. He warned that southeastern Alberta and southwestern Saskatchewan would not reliably sustain human (i.e., agricultural) life. The land was fundamentally arid, receiving 400 or fewer millimeters (mm) of rain a year; some years there might as well be no rain at all. Fifty years later, Palliser's admonitions were swept away by the newly arrived Canadian Pacific Railway as it launched one of the most far-flung publicity campaigns ever undertaken.

In 1901, 600,000 hectares within the "Special Areas" (defined by a dryland ranch life that would all but collapse thirty-five years later) had seventy-five residents. Fifteen years of rain and railroad boosterism followed. By 1916, twenty-four thousand people were busting sod on their quarter-section homesteads. Throughout the larger Palliser Triangle of southeastern Alberta and southwestern Saskatchewan, 280,000 people were farming almost 3 million hectares before the First World War. About 100,000 hectares within the Special Areas would be broken by the plow by 1924.

Harry Gordon's parents left Calgary and joined the land rush to the prairies around Hanna, Alberta, in 1913; Harry was one month old. Like their neighbors, the Gordon family thrived during the wet years through 1915. A lucky farmer could expect 100 bushels to the hectare. Wheat prices soared as World War I raged. Then the rain ceased. Drought sucked the Gordon farm dry from 1917 through 1926. In nearby Medicine Hat, 1 hectare of land planted in wheat produced 17 bushels in 1917, a couple bushels in 1918, and none in 1919. Harry learned to do without; he avoided debt like the plague itself. He learned to always leave a year's worth of grass on the range for his cattle.

Times improved on the Alberta prairie during the mid-1920s, but drought conditions returned during the Dust Bowl years. Sixty-six thousand people gave up their plows from 1931 through 1936. The populations of towns within the Special Areas shrank by 40, 60, or even 80 percent. Canada's federal government created the Prairie Farm Rehabilitation Administration in 1935 to assist in the formation of irrigation districts. The Special Areas were officially designated in southeastern Alberta to assist dryland farmers who struggled to stay on their land. Harry Gordon hung on through good times and bad. He learned to live with drought.

Southeastern Alberta is drought prone, not because of its low average annual rainfall, but because its climate departs from average rainfall patterns frequently enough to hamstring people who try to live and farm there. South of Hanna, Henry Haugen raises cattle around Manyberries, near the Alberta/Montana border. His grandparents settled on this land in 1900; he was born just a kilometer up the road. Rainfall here averages 350 mm a year; in some years, it's less than 200 mm. Henry has run as many as four hundred cattle. Typically, each cow needs 200 hectares of grazing pasture. His well produces water from a depth of 100 meters (m) with an extremely high salt content: dissolved solids total more than 3,000 parts per million. The Haugens hung on through the 1930s by abandoning any expectation of return on their investment; income was plowed back into expenses year after year. That solution may not always be good enough for his three sons who continue to ranch with him. If water drops any lower in the well or if his sons' wives become any less content, one by one they will leave to live elsewhere.

Southwest of Alberta, in the Cascade Mountains of Oregon, Barry Lopez pondered the McKenzie River alongside his home as it dropped with drought one year:
I awoke one night and thought I heard rain—it was the dry needles of fir trees falling on the roof. . . . I have trained myself to listen to the river, not in the belief that I could understand what it said, but only from one day to the next know its fate. . . . It was in this way that I learned before anyone else of the coming drought. Day after day as the river fell by imperceptible increments its song changed, notes came that were unknown to me. . . . I would say very simple prayers in the evening, only an expression of camaraderie, stretching my fingers gently into the darkness toward the inchoate source of the river's strangulation. I did not beg. There was a power to dying, and it should be done with grace. I was only making a gesture on the shore, a speck in the steep, brutal dryness of the valley by a dying river.

It was drought, along with war and famine, that biblical people feared most. We have spent the last two thousand years trying to anticipate the rhythms of climate, trying to understand the machinations of rain. But these rhythms are elusive. Half a century ago, in a lovely book entitled Drought, Its Causes and Effects, I.R. Tannehill sighed as he observed, "In conclusion, the writer regrets that it is necessary to say that the problem of drought is not completely solved." We continue to search for an underlying cause for drought. Lorne Coles ranches near Hanna. He stands alongside a pasture fence with a handful of dry sand sieving though his fingers. Lorne and his dryland neighbors in southeastern Alberta are curious to know what we find.


The cadence of wet and dry years is difficult to comprehend, let alone predict. A tantalizing pattern links drought and its inverse, flooding, in many places around the world. Floods in one location and drought in another—so obviously different—often exist within similar global configurations of climate.

Flood and drought: Both are etched deeply into the collective human experience. Flooding on the Huang Ho River took the lives of 3.7 million Chinese in 1931. Drought in Africa's Sahel Desert claimed at least 150,000 lives during the second half of the twentieth century. Over time, drought has been more deadly than flooding. Drought depends on the persistence of dryness over months or years. Catastrophic floods can explode suddenly out of a single summer thunderstorm. Flooding, however, can also be caused by a months-long buildup of moisture, such as the fast melting of a heavy winter's accumulation of mountain snow or soil saturated by high seasonal rainfall. All floods, of course, are shaped by the basin through which they flow.

Typical springtime floods on the Mississippi River have peak discharges on the order of 30,000 cubic meters per second (m³/s). On March 21, 1997, the Mississippi River at Vicksburg, Mississippi, was flowing at twice that volume: an irresistible python of water 2.5 kilometers (km) wide and 40 m at its deepest, coursing past the delta countryside at 10 kilometers per hour (km/hr). Floods on the Amazon River are typically ten times as large as those on the Mississippi, but to find truly big floods one must travel surprising distances through time or space.

During the late Pleistocene, about 15,000 years ago, the breakup of ice dams on the now-vanished Lake Missoula triggered a series of floods that leveled everything in their path from Montana to the Pacific Ocean. Peak discharge of one of those floods—the Rathdrum Prairie event—has been calculated at 20 million m³/s, that is, a thousand Mississippi River floods flowing together.

Outwash channels on Mars—complete with streamlined uplands, longitudinal grooves, dry waterfalls, scour marks, and fan-shaped deposits of sediment—are 100 km wide and 2,000 km long. After guessing at a multitude of parameters, including channel slope and sediment load, scientists have estimated that the peak flow volume of these Martian floods was 300 million m³/s.

We need to look only as far as the earth to see the various mechanisms that trigger floods. Earthquakes and landslides that occur beneath an ocean floor can induce tsunamis. These are waves that flash unnoticed across an entire ocean to pile up against an unsuspecting coastline half a world away. Dams, either man-made or natural (such as ice jams or landslides), can burst. Hurricanes and cyclones whip up tidal surges that inundate coastal lowlands. Rivers can swell in response to precipitation. Tsunamis and dam failures have certainly wreaked havoc through the ages. In 1960, when Chile was rocked by the greatest earthquake ever recorded (magnitude 9.5), the associated tsunami crossed the Pacific in twenty-two hours, drowning 5,700 people in Hawaii and Japan. After the La Josefina landslide blocked the Paute Valley near Cuenca, Ecuador, in 1993, its catastrophic failure released a flood that approached 10,000 m³/s.

Of the four flood mechanisms listed above, the last two—tidal surges and especially river response—are most intimately tied to the whims of weather. It is on these two that we will focus. The Ganges-Brahmaputra-Meghna River system can swell dramatically when the Indian monsoon is especially torrential. In September 1987, the lives of 25 million people in Bangladesh were disrupted when monsoonal flooding inundated 2 million hectares of cropland. These floods were anticipated by days if not weeks; even so, 5 million homes were destroyed and 2,000 people died. But what the Bangladeshi truly fear is flooding that sweeps in, not from the mountains, but from the Indian Ocean. Tidal surges are domes of water 75 to 150 km across that are driven aground in front of an incoming storm. Surges from just two tropical cyclones killed 440,000 people in Bangladesh in November 1970 and April 1991.

Floods are most likely to kill when they occur without warning. Residents of Nelson, Nevada, awoke to blue skies on the morning of September 14, 1974. As the day progressed, convective clouds built over the mountains behind town. A single thunderstorm lingered over the upper end of the 59 km2 basin leading into El Dorado Canyon. Sluggish upper-level winds carried the cell slowly eastward toward the canyon's mouth, so that rain kept falling on the moving flood crest, amplifying floodwaters that had already begun to rush through the canyon. More than 2,000 m³/s swept down the usually dry wash, offering no more than minutes of warning to the people below. Nine drowned. The tiny community of Nelson vanished into Lake Mohave on the Colorado River.

In contrast, the Mississippi River's great flood of 1993 had its beginnings in the upper basin's gradual accumulation of above-average rainfall as early as July 1992. Warm waters in the Pacific Ocean had reinforced a ridge of high pressure over the western United States. The subtropical jet stream was enhanced, adding kinetic energy and moisture to the flow of air coming into the Midwest. Then during the summer, a ridge of high pressure became embedded above the East Coast. Clockwise rotation of wind around this ridge aimed a fire hose of moist air northbound from the Gulf of Mexico against the Midwest.

A river of atmospheric moisture was flowing from the Gulf of Mexico into the central and eastern United States. The westerly jet stream served as the match that ignited this fuel flowing in from the gulf. Embedded within this moist regional air mass were at least 175 cores of intense thunderstorm activity, each locally producing 150 mm or more of rain. Thunderstorms popped off like a string of firecrackers all summer long. Weekly average rainfall throughout the upper Mississippi and Missouri River basins from June through August 1993 was 28 billion m3. After forty weeks of this deluge, the precipitation total was twice the volume of Lake Erie.

Minor flooding lapped against levees at St. Louis for forty-four days in April and early May 1993. The Mississippi and Missouri Rivers rose, but it wasn't until June 20 at Hatfield, Wisconsin, that the first of a thousand levees broke. From that point on, the rivers rose inexorably, each day surpassing the previous day's estimate of flood height. Locks were flooded; the rivers were closed to barge traffic; grain prices soared. At St. Louis, the Mississippi River finally crested at 15.1 m on August 1, 1993, with a discharge of more than 28,000 m³/s. Record flooding was observed along 2,900 km of river within the upper basin, and major flooding occurred on an additional 2,000 km. By the end of August, 50,000 people had been displaced and almost 40,000 km2 of prime cropland were covered by floodwaters. Fifty-two people died. Losses totaled $18 billion.

Floods don't just happen because it rains. Floods happen because rain falls on saturated ground, because warm rain falls on an existing snowpack, because rain falls heavily throughout an entire basin, or because the basin has been changed (either naturally or otherwise) so as to retain or heighten floodwaters that would have otherwise rolled on through without making a mess. Floods are usually more localized than droughts—both in time and in space—because floods require these specific preexisting hydrologic and meteorologic conditions.

Spring melting of winter snow is always a time of high river flow. In many regions, river channels shape themselves to accommodate these annual events, building banks and gradually raising floodplain terraces in response to each year's high water. Meltwaters will usually be released from the snowbound mountains in an orderly progression as springtime temperatures begin to rise. Floods will occur if temperatures rise faster than expected or if rain falls on snow that is already near its melting point.

Destructive flooding doesn't always have to follow high precipitation. The mountains of western Washington and Oregon were buried beneath 30 m of snow in the winter of 1998--99, far more than usually falls. Communities along the Northwest Coast held their breath as spring arrived, waiting for the floods that seemed inevitable. But nothing out of the ordinary happened. Because the spring turned out to be relatively cool, the snowmelt occurred in an orderly fashion and rivers remained within their banks. Tallying precipitation does not necessarily equate with flood forecasting.

During the winter of 1997-98, California was inundated by a series of storms that broke records for precipitation throughout the state. Eureka received rain on fifty-two of the fifty-nine days of January and February. San Francisco registered 508 percent of normal precipitation during the month of February. The people of Rio Nido along the Russian River in the central Coast Ranges of California were blasted awake on February 7 by a debris flow that reduced homes to matchsticks along Upper Canyon Three. The community waited without defense, holding its breath 600 feet beneath a saturated hillside that threatened to erase 140 homes that were at risk if the hillside slipped. Pescadero Canyon was repeatedly blocked by mud slides that had taken lives at Loma Mar in the mountains above Palo Alto. Homes in Los Angeles were transformed into nightly news icons as foundations eroded and their walls gradually disappeared.

Floods are fickle, requiring very specific conditions beyond merely wet weather. A flood might hit one basin yet inexplicably ignore another basin nearby. Sometimes whole towns or villages are wiped out; sometimes the damages seem more like the luck of the draw. Whether widespread or local in scale, floods are set up by large-scale atmospheric processes that are in some ways the inverse of droughts. Whether the flood occurs or not when those conditions are present is another matter that befuddles flood forecasters.

Copyright © 2002. Arizona Board of Regents.

This sample is free for reading while you are visiting our web site but is not free for other use, distribution, or download anywhere other than your web browser's cache.

[Close Window]