Figure 18. Breaking new land with disk plows, Greely County, Kansas, 1925 (courtesy of Kansas State Historical Society).

By 1928, when Hugh Bennett and W. R. Chapline published the first national soil erosion a.s.sessment, topsoil loss amounted to five billion tons a year-several times faster than soil loss in the nineteenth century and ten times faster than soil formed. Nationwide, virtually all the topsoil had already eroded from enough farmland to cover South Carolina. Six years later Bennett and Chapline's report seemed understated. Even in drought and Depression, the number of tractors working Oklahoma farms increased from 1929 to 1936. New disk plows, with rows of concave plates set out along a beam, thoroughly diced the upper layers of soil, leaving a pulverized layer that could easily blow away in dry conditions.

The first major windstorm of 1933 swept through South Dakota on November i1. Some farms lost all their topsoil in a single day. The next morning the sky remained dark until noon-one part air to three parts dust. No one knew this was just a preview.

On May 9, 1934, fields from Montana and Wyoming were ripped up by high winds. Blowing across the Dakotas, the wind kept picking up dirt until a third of a billion tons of topsoil was heading east at up to a hundred miles an hour. In Chicago four pounds of dust dropped out of the sky for each person in the city. The next day Buffalo, in eastern upstate New York, fell dark at noon. By dawn on May ii dust was settling on New York City, Boston, and Washington. The huge brown cloud could be seen far out in the Atlantic Ocean.

Resilient when under permanent vegetation, grazed (and manured) by millions of buffalo, the prairie fell apart when plowed up and dried out by prolonged drought. Without the gra.s.ses and their roots that stabilized the soil, winds that decades before blew harmlessly across the range ripped the countryside open like a sand-charged hurricane. Shifting drifts of dirt covered a vast region where high winds carried off desiccated soil exposed beneath the parched stubble of wilted crops. High winds stirred up enough dust to choke people, shred crops, kill livestock, and shroud distant New York City in an eerie veil.

The National Resources Board reported that by the end of 1934, dust storms had destroyed an area larger than the state of Virginia. Another hundred million acres were severely degraded.

In the spring of 1935 strong winds again tore through the parched fields of Kansas, Texas, Colorado, Oklahoma, and Nebraska. With the fields freshly plowed, there was no vegetation to hold the dry loess in place. The finest and most fertile soil formed dark blizzards rising ten thousand feet to blot out the midday sun. Coa.r.s.er sand blew around near the ground, gnawing through fence posts. Streetlights stayed on all day. High winds piled up Sahara-like dunes, blocking trains and paralyzing the plains.

On April 2, 1935, Hugh Bennett testified before the Senate Public Lands Committee about the need for a national soil conservation program. Bennett knew that a great dust storm from the plains was descending on Washington. With help from field agents who called to report the progress of the dirt cloud, he timed his testimony so that the sky went dark as he presented it. Duly impressed, Congress appointed Bennett head of a new Soil Conservation Service.

The agency faced a daunting challenge. Within a few decades of settlement, barren desert had replaced the short-gra.s.s prairie. President Franklin D. Roosevelt ended the era of land settlement in November 1934 by closing the remaining public lands to homesteading. American agricultural expansion was officially over. Displaced Dust Bowl farmers had to find work in someone else's fields.

More than three million people left the plains in the 193os. Not all of them were fleeing dust, but about three-quarters of a million displaced farmers headed west. The grandchildren of the original sooners became environmental refugees, unwelcome until they reached California's new, labor-hungry fields at the edge of the continent.

Figure i9. Dust storm approaching Stratford, Texas, April 18, 1935 (NOAA, George E. Marsh Alb.u.m; available at www.photolib.noaa.gov/historic/c&gs/ theb1365.htm).

The problem of soil erosion was not restricted to the Dust Bowl. In 1935 the Department of Agriculture estimated the amount of ruined and abandoned farmland at up to fifty million acres. Two to three times that much land was losing an inch of topsoil every four to twenty years. Two hundred thousand acres of abandoned Iowa farmland was eroded beyond redemption. The next year the new Soil Conservation Service reported that more than three-quarters of Missouri had lost at least a quarter of its original topsoil, more than twenty billion tons of dirt since the state was first cultivated. Only four of their original sixteen inches of topsoil remained on some fields. The U.S. Bureau of Agricultural Engineering reported that it was common for southeastern farms to lose more than six inches of soil in a single generation. In the aftermath of the Dust Bowl, which cost over a billion dollars in federal relief, the federal government began to see soil conservation as an issue of national survival.

State and federal commissions traced the severity of the 193os dust storms to a tremendously increased acreage under cultivation, much of which was marginal land. The Kansas State Board of Agriculture, for example, blamed the disaster on poor farming practices. "Soil has been cultivated when extremely dry, and no effort has been made, in most cases, to return organic matter to the soil.... When cultivated in a dry condition such a soil became loose and dusty. There are individual farmers throughout the region who have followed good methods of soil management and have found it possible to prevent soil blowing on their farms, except where soil blown from adjoining farms encroached upon their fields.' 17 The report of the Great Plains Committee convened in 1936 by the House of Representatives had identified economic forces as a major cause of the disaster.

Figure 20. Buried machinery in barn lot, Dallas, South Dakota, May 13, 1936 (USDA image No:oodi097I CD8151-97r; available at www.usda.gov/oc/photo/ oodi097I.htm).

The [First] World War and the following inflation pushed the price of wheat to new high levels and caused a remarkable extension of the area planted to this crop. When the price collapsed during the post-war period Great Plains farmers continued to plant large wheat acreages in a desperate endeavor to get money with which to pay debt charges, taxes, and other unavoidable expenses. They had no choice in the matter. Without money they could not remain solvent or continue to farm. Yet to get money they were obliged to extend farming practices which were collectively ruinous.8 Walter Lowdermilk, by then the a.s.sociate chief of the Soil Conservation Service, suggested using the erosion rate on undisturbed lands as the geologic norm of erosion to provide the benchmark for gauging humaninduced erosion. His concern seemed justified when the Soil Conservation Service compiled county-level soil erosion maps into a national map. The results were alarming. More than three-quarters of the original topsoil had been stripped off almost two hundred million acres of land, about a tenth of the area surveyed. From one- to three-fourths of the topsoil was gone from two-thirds of a billion acres, more than a third of the area surveyed. At least a quarter of the soil was missing from almost a billion acres of land. America was losing its dirt.

In a speech given before the annual meeting of the National Education a.s.sociation in July 1940, Hugh Bennett would describe the dust storm of May six years earlier as a turning point in public awareness. "I suspect that when people along the seaboard of the eastern United States began to taste fresh soil from the plains 2,000 miles away, many of them realized for the first time that somewhere something had gone wrong with the land."9 On April 27, 1935, Congress had declared soil erosion a national menace and established the Soil Conservation Service to consolidate federal actions under a single agency. A year later in his opening address to a conference convened by order of President Roosevelt, the agency's newly appointed chief Hugh Bennett compared the rapid loss of soil from U.S. farmlands to the slow pace of soil formation.

Citing federal studies, Bennett showed just how fast America was disappearing. The erosion research station at Tyler, Texas found that the region's best farming practices increased soil loss by almost two hundred times the soil replacement rate. Poor management practices increased erosion by eight hundred times. The research station at Bethany, Missouri showed that soil loss from typical corn lands was three hundred times that of comparable land under alfalfa.

Research also showed that after erosion of the loose topsoil, more rainfall ran off over the surface instead of sinking into the ground. This produced more runoff, which then removed even more soil, producing yet more runoff. It did not take long to lose topsoil once the process started.

Bennett calculated that it took more than five thousand years for rainfall to remove six inches of topsoil from native gra.s.sland in Ohio. This made sense; it was close to the rate at which he thought soil formedabout an inch every thousand years. In contrast, fields lost some six inches of topsoil in little more than three decades of continuous cultivation. The erosion research station at Guthrie, Oklahoma found that the fine sandy loam covering the plains eroded more than ten thousand times faster under cotton cultivation than native gra.s.s. Cotton farming could strip off the region's typical seven inches of topsoil in less than fifty years. The same topsoil under grazed gra.s.s would last more than a quarter million years. The message was clear, Bennett advised not plowing hilly and highly erodible land.

Echoing Bennett's warnings in 1953, his a.s.sociate chief described how almost three-quarters of U.S. farmland was losing soil faster than it formed. In particular, Lowdermilk stressed that the United States was following ancient civilizations down the road to ruin. He argued that seven thousand years of history cautioned against plowing hillslopes.

Here in a nutsh.e.l.l, so to speak, we have the underlying hazard of civilization. By clearing and cultivating sloping lands-for most of our lands are more or less sloping-we expose soils to accelerated erosion by water or by wind.... In doing this we enter upon a regime of selfdestructive agriculture.... As our population increases, farm production will go down from depletion of soil resources unless measures of soil conservation are put into effect throughout the land.10 Lowdermilk did not see this as a remote threat some centuries in the future. He viewed the wars of the twentieth century as a battle for control of land.

After the Second World War, conversion of military a.s.sembly lines to civilian uses dramatically increased tractor production, completing the mechanization of American farms and paving the way for high-output industrial farming in developed countries. Several million tractors were working American fields by the 195os-ten times as many as in the 192os. The number of U.S. farmers plummeted as farm acreage increased and more people moved to the swelling cities. The few farmers left on their land grew cash crops to pay off the loans on their new labor-saving equipment. Mechanization, like slave labor in the South, required doing the same thing everywhere instead of adapting agricultural methods to the land.

Figure 2i. Plowing a steep hillside circa 1935 (National Archives, photo RG-o83- G-36711).

Droughts in the Great Plains occur about every twenty years. In the wet 1940s, doubling the acreage under cultivation increased wheat production fourfold-enough to support record exports to Europe during the war. In 1956 drought again caused near failure of the wheat crop. The 1950s drought lasted almost as long as the 193os drought and was as severe as the 189os drought (though this time soil conservation programs were widely credited with preventing another Dust Bowl). Small farms were going bankrupt while large farms better able to withstand the periodic dry spells bought more and larger machinery.

The U.S. government had begun farm subsidies in 1933. Within a year, most Great Plains farmers were partic.i.p.ating in programs aimed at soil conservation, crop diversification, stabilizing farm income, and creating flexible farm credit. As much as anything, this last element, which allowed farmers to carry more debt, changed American farming. Within a decade, farm debt more than doubled while farm income rose by just a third. Despite a continual rise in government subsidies, more than four out of every ten American farms disappeared between 1933 and 1968. Corporate factory farms better able to finance increasingly expensive farm machinery and agrochemicals began to dominate American agriculture by the end of the 196os.

Although different in detail from Rome and the South, the economics of large corporate farms similarly discounted concern about soil erosion.

Corporations are, by nature, temporary land owners.... A tenant on corporate land has no a.s.surance whatsoever of staying on the farm more than a year.... A high proportion of corporate land tends to cause instability in land tenure and to foster erosion, unless the majority of corporations can be induced to adopt definite soil conservation programs on their land. Heavy mortgage indebtedness exerts a specific financial pressure upon the soil by forcing the farmer to squeeze out of his soil whatever he can to meet his financial obligations.11 The growth of mechanized industrial agriculture promoted rapid soil loss as farmers spent their natural capital to service loans for machinery and fertilizers.

Records at Woburn Experimental Farm, established about twenty-five miles north of London in 1876 by England's Royal Agricultural Society, inadvertently doc.u.mented the effects of changing agricultural practices on soil erosion. The first half century of crop yield experiments recorded little erosion. After the Second World War, the introduction of herbicides and heavy farm machinery changed that.

The first report of soil erosion problems came after a storm on May 21, 1950, when intense rainfall carved four-inch-deep, three-foot-wide gullies into bare fields, burying sugar beet plots beneath piles of dirt and unearthed potatoes. Serious erosion in the 196os sharply reduced the organic nitrogen content on experimental plots. By the 198os the farm served to validate soil erosion models, as more than a dozen erosional events occurred each year, especially on the farm's steepest slopes. Yet the detailed diaries kept by farm staff from 1882 to 1947 had focused on the subtleties of crop performance, cultivation techniques, soil pH, and crop damage from varmints, with no mention of erosion before the introduc tion of heavy machinery and agrochemicals. Adopting twentieth-century agricultural methods greatly accelerated soil erosion.

One of the most persistent agricultural myths is that larger mechanized farms are more efficient and profitable than smaller traditional farms. But larger farms spend more per unit of production because they buy expensive equipment, fertilizer, and pesticides. Unlike industrial enterprises in which economies of scale characterize manufacturing, smaller farms can be more efficient-even before accounting for health, environmental, and social costs. A 1989 National Research Council study flatly contradicted the bigger is more efficient myth of American agriculture. "Well-managed alternative farming systems nearly always use less synthetic chemical pesticides, fertilizers, and antibiotics per unit of production than conventional farms. Reduced use of these inputs lowers production costs and lessens agriculture's potential for adverse environmental and health effects without decreasing-and in some cases increasing-per acre crop yields."12 Small farms also can produce more food from the same amount of land. A 1992 U.S. agricultural census report found that small farms grow two to ten times as much per acre as do large farms. When compared to farms greater than six thousand acres in size, farms smaller than twenty-seven acres were more than ten times as productive; some tiny farms-less than four acres-were more than a hundred times as productive. The World Bank now encourages small farms to increase agricultural productivity in developing nations, where most landholders own less than ten acres.

A key difference between small farms and large industrial farming operations is that large farms typically practice monoculture, even though they may grow different crops in different fields. Single-crop fields are ideal for heavy machinery and intensive chemical use. Although monocultures generally produce the greatest yields per acre for a single crop, diversified polycultures produce more food per acre based on the total output from several crops.

Despite the overall efficiency of small farms, the trend is toward larger, more industrialized farms. In the 1930s seven million Americans farmed. Today fewer than two million farmers remain on their land. As recently as the early 199os, the United States had lost more than twenty-five thousand family farms a year. On average, more than two hundred American farms have gone under every day for the past fifty years. In the second half of the twentieth century, the average farm size more than doubled, from under one hundred to almost two hundred hectares. Less than 20 percent of U.S. farms now produce almost 9o percent of the food grown in America.

As crop yields increased two- to threefold from 1950 to the 199os, the cost of machinery, fertilizer, and pesticide rose from about half to over three-quarters of farm income. Two types of farms survived: those that opted out of industrialization and those that grew by working larger areas for a smaller net return per acre. By the 198os the largest farms, dubbed superfarms by the USDA, accounted for close to half of all farm income.

If small-scale agriculture is so efficient, why are America's small farms going under? The high capital costs of mechanization can be an economic disaster for a small operation. A farm must be large to profitably use technology-intensive methods instead of labor-intensive methods. Sold on the idea that modernizing meant mechanizing, small farms sank into debt once overleveraged; large companies then bought up their land. This process may not help small farms stay in family hands, but it pumps a lot of cash into companies that produce farm equipment and supplies-and advise farmers how to use their products.

The economic and social trends that drove mechanization turned farming into an industry and accelerated soil loss. New equipment made more intensive cultivation of land easier, to a deeper depth and more often. Just as in ancient Rome, the ground lay bare and disturbed for much of the year. As farms mechanized, soil conservation practices such as terracing, hedgerows, and trees planted for windbreaks became obstacles to maneuvering heavy machinery. Contour plowing practices were modified to accommodate large machines that could not follow tight turns on sloping land. Soil was now a commodity-the cheapest of many inputs to agricultural manufacturing.

Substantial progress in raising both public and governmental awareness slowed but did not stop soil loss. Some areas have fared worse than others. Across the heart of the Midwest, islands of native prairie rising up to six feet above neighboring plowed fields testify to soil loss of about half an inch per year since settlement. Iowa lost half its topsoil in the last century and a half. Fortunate by comparison, the Palouse region of eastern Washington lost only a third to half of its rich topsoil in the past century.

The first settlers arrived in the Palouse in the summer of 1869. They grew grain on the valley bottoms and raised cattle and hogs to sell to miners in nearby Idaho. The region's deep loess soil could produce more but there was no way to get crops to market. Completion of the railroads in the 188os opened the land to distant markets, new equipment, and more farmers. By the 189os most of the Palouse was under cultivation.

Soil erosion rapidly became a major problem once the loess was cleared and plowed. In the early i9oos Washington State Agricultural College's William Spillman toured the region lecturing on the threat of soil erosion from the common practice of leaving plowed fields bare each summer. Few heeded the young professor's warning that each year's annoying rills would eventually add up to a serious problem.

In the 1930s tractors began replacing horse-drawn plows in the Palouse and elsewhere, allowing a single operator to farm much larger acreages. Eager to capitalize on the greater labor efficiency, landowners changed the traditional arrangement for sharecropping on leased land. Instead of keeping two-thirds of what they grew, tenants were now allowed to keep just over half. So tenants worked the land that much harder, reducing outlays for luxuries like erosion control. Farmers were now working more land, but not necessarily making more money.

By 1950 a USDA survey reported that all of the original topsoil was missing from io percent of Palouse farmland. Between 25 and 75 percent of the topsoil was missing from an additional 8o percent of the land. Just io percent of the region retained more than 75 percent of its original soil. Annual surveys of soil loss from 1939 to i96o showed an average loss of half an inch a decade. On slopes steeper than about fifteen degrees, soil loss averaged an inch every five years.

A cistern installed in 1911 on a farm near Thornton dramatically ill.u.s.trates the effect of plowing sloping fields. Originally projecting about a foot and a half above the adjacent hilltop, by 1942 it stuck out nearly four feet above the surrounding field. By 1959 the same cistern stood six feet above the field. Four and a half feet of soil had been plowed off the slope in less than fifty years-about an inch a year. Some eastern Idaho soils that were more than a foot thick in the early twentieth century were barely deep enough to plow by the 196os when just half a foot of soil remained above bedrock.

From 1939 to 1979 the total erosion on Palouse cropland averaged more than nine tons per acre per year; and reached more than one hundred tons per acre per year on steep slopes. Erosion rates on unplowed rangeland and forested land averaged less than one ton per acre per year. Plowing the loess increased erosion rates by a factor of ten to a hundred, with most of the loss caused by erosion from runoff across newly plowed ground. Simple soil conservation measures could halve erosion without reducing farm income. But doing so requires fundamental changes in farming practices.

Figure 22. Bare, rilled field in the Palouse region of eastern Washington in the 1970s (USDA 1979, 6).

In 1979 the Soil Conservation Service reported that three decades of plowing had lowered fields as much as three feet below unplowed gra.s.sland. Berms of soil four to ten feet high stood at the downhill end of plowed fields. Experiments conducted with a typical sixteen-inch moldboard plow pulled along contours showed that plowing typically pushed soil more than a foot downhill. The process that stripped Greek hillsides in the Bronze Age was being repeated in the Palouse.

Simply plowing the land pushed soil downhill far faster than natural processes ever managed. Even so, this process is almost as hard to notice, occurring imperceptibly with each pa.s.s of a plow. Continued for generations, till-based agriculture will strip soil right off the land as it did in ancient Europe and the Middle East. With current agricultural technology, though, we can do it a lot faster.

Wind erosion contributes to the problem. A core from the bed of eastern Washington's Fourth of July Lake records that dust fall into the lake increased fourfold with the introduction of modern agriculture to the region. There are few reliable measurements of wind erosion under natural conditions, but under the right conditions it can be extreme. Before soil conservation measures were adopted, wind stripped up to four inches of soil a year from some Kansas fields in the Dust Bowl era. Dust blowing off of bare, dry field is still a problem in eastern Washington. In September 1999 dust blowing off of agricultural fields blinded drivers and triggered fatal traffic accidents on Interstate 84 near Pendleton, Oregon.

Plowing exposes bare, disrupted soil to dramatic erosion when storms ravage ground not yet shielded by vegetation. In the American Midwest, over half the erosion from land planted in corn occurs in May and June before crops grow large enough to cover the ground.

Crop yields fall once the topsoil is gone and farmers plow down into subsoil with lower organic matter, nutrient content, and water-retention capacity. Losing six inches of topsoil from Georgia and west Tennessee soils reduced crop yields by almost half. Severely eroded areas of Kentucky, Illinois, Indiana, and Michigan already produce a quarter less corn than they once did. Just a foot or two of erosion can dramatically reduce soil productivity-sometimes to the point of losing all agricultural potential. Less than 50 percent of U.S. croplands have slopes gentler than 2 percent and therefore little threat of accelerated erosion. The steepest 33 percent of U.S. cropland is projected to fall out of production over the next century. Since 1985 the Gra.s.sland Reserve Program has been paying farmers to restore and preserve gra.s.slands in areas vulnerable to soil erosion.

Soil erosion is not only a problem of capitalist agriculture. The rich black earth of the Russian steppe eroded rapidly once the native vegetation was cleared. Although deep gullies surrounded Russian settlements as early as the sixteenth century, the fragile nature of these soils did not slow efforts to industrialize Soviet agriculture in the twentieth century. The first fiveyear plan, produced in 1929, included a blunt call to convert the steppe to factory farms. "Our steppe will truly become ours only when we come with columns of tractors and ploughs to break the thousand-year-old virgin soil."13 Contrary to plan, dust storms blossomed after plows broke up the gra.s.sland.

The Soviet's virgin land program of the 195os and 196os brought a hundred million acres of marginal farmland into production. Against the advice of prominent scientists aware of the American Dust Bowl, Premier Nikita Khrushchev ordered state collectives to plow forty million acres of virgin land from 1954 to 1965. Food production was not keeping pace with postwar consumer demand.

With the ground left bare during fallow periods, severe erosion reduced crop yields within a few years on much of the newly cleared land. At the program's peak, Soviet agriculture lost more than three million acres a year-not a good way to fulfill a five-year plan. Severe erosion damaged almost half the newly plowed land during the next dry spell in the i96os, creating a little-publicized Soviet dust bowl that helped drive Khrushchev from office.

Before 1986 Soviet censors hid the extent of environmental problems. Foremost among those problems was the Aral Sea disaster. In 1950 the Soviet government initiated a major effort to achieve "cotton independence" by turning the region into monocultural plantations. The Soviets greatly increased crop yields through improved cultivation techniques, aggressive fertilizer and pesticide use, and by expanding irrigation and mechanized agriculture. From 1960 to 199o thousands of miles of new ca.n.a.ls and more than six hundred dams diverted rivers from the Aral Sea. Not surprisingly, the sea began to shrink.

As the Aral Sea dried out, so did the surrounding land. By 1993, decades of continuous water diversion lowered the water level almost fifty-five feet, creating a new desert on the exposed seabed. Major dust storms in the 199os dropped a hundred million tons of Aral salt and silt on Russian farms a thousand miles away. The collapse of both the fishing industry and agriculture triggered a ma.s.s exodus.

A post-glasnost regional a.s.sessment revealed that desertification affected two-thirds of and lands in Kazakhstan, Uzbekistan, and Turkmenistan. Proposals to address this growing threat went nowhere before the Soviet Union disintegrated. Independence only increased the desire to pursue cash crops for export, moving the fight against soil erosion to the bottom of the political agenda. Despite the clear long-term threat, more immediate concerns prevailed.

A similar situation developed in the small southern Russian Kalmyk Republic tucked between the River Volga and the Caspian Sea. Between the Second World War and the 199os, aggressive plowing of rangelands desertified most of the republic. Almost a tenth of the country turned into barren wasteland.

Kalmykia's native gra.s.slands were ideal for livestock. As early as the twelfth century, Kalmyks brought cattle to the region where horses were said to graze without bending their heads. Traditional land use centered around horse breeding and sheep or cattle grazing. Accused of collaborat ing with the Germans, Kalmyks were exiled en ma.s.se to Siberia in 1943. By the time they returned fifteen years later, the Soviets were busy creating Europe's first desert.

Throughout the cold war, Soviet policies favored plowing up Kalmyk pastures to increase cereal and melon production. The number of sheep almost doubled on the remaining gra.s.sland. Forage crop yields declined by half between the i96os and the 199os. Each year the desert consumed another 50,000 hectares of bare fields and overgrazed pasture. By the 1970s more than a third of the republic was partially desertified.

Plowing the native gra.s.slands of this semiarid region led to problems reminiscent of the Dust Bowl. Developed on sand deposited on the bed of the once more extensive Caspian Sea, Kalmykia's rich soils were held together by the roots of the lush native gra.s.s. Within decades of plowing, more than a third of a million hectares of gra.s.sland were transformed into moving sand seas. In 1969, after extensive agricultural development, a major dust storm blew soil to Poland. Fifteen years later another dust storm sent Kalmyk dirt all the way to France. The republic's president p.r.o.nounced a state of ecological emergency on August I, 1993-the first such proclamation from a national government in regard to soil erosion.

The superpowers of the late twentieth century were not alone in losing soil faster than nature makes it. Erosion outpaces soil production by ten to twenty times in Europe. By the mid-i98os roughly half of Australia's agricultural soils were degraded by erosion. Soil erosion from steep slopes in the Philippines and Jamaica can reach four hundred tons per hectare per year-the equivalent of carting off almost an inch and a half of soil a year. Half of Turkey is affected by serious topsoil erosion. Once done, the damage lasts for generations.

In the 1970s sub-Saharan Africa experienced its own dust bowl. Until the twentieth century, West African farmers used a shifting pattern of cultivation that left fields fallow for long periods. Grazing was light as animal herders moved long distances across the landscape each year. In the twentieth century, the combination of a rising population and encroachment of agricultural fields on traditional pastures intensified land use by both farmers and pastoralists. Extensive land clearing and degradation led to extreme soil loss that created a flood of environmental refugees.

The African Sahel lies in the semiarid zone between equatorial forests and the Sahara. On average the region receives six to twenty inches of rainfall annually. But the rain varies widely year to year. In a good year, it rains more than one hundred days in northern Senegal; in bad years it rains fewer than fifty. Studies of ancient lake levels show that long droughts occurred repeatedly over the past several thousand years. Tree ring studies from the Atlas Mountains just north of the Sahel reveal at least six droughts lasting twenty to fifty years between AD 1100 and 1850. The next run of dry years proved catastrophic after almost a half million square miles of West African forest were cleared in under a century.

The 1973 West African famine killed more than a hundred thousand people and left seven million dependent on donated food. Triggered by drought, the roots of the crisis lay in the changing relationship of the people to the land. Extensive removal of ground-protecting vegetation triggered severe soil erosion and a humanitarian disaster during the next run of drier than average years.

The nomads and sedentary farmers of the Sahel traditionally practiced a symbiotic arrangement in which the nomads' cattle would graze on crop stubble, manuring farmers' fields after the harvest. When the rains came the herds would head north following the growth of new gra.s.s. Continuing north until the gra.s.s was no longer greener ahead of them, the nomads would turn back south, their cattle grazing on the gra.s.s that grew up behind them after they pa.s.sed through on their way north. They would arrive back in the south in time to graze and manure the farmers' harvested fields. In addition, Sahelian farmers grew a variety of crops and let land lie fallow for decades between episodes of cultivation. The division of the Sahel into separate states disrupted this arrangement.

Rapid expansion of French colonial authority across the Sahel in the late nineteenth century altered the social conventions that had prevented overgrazing and kept fields manured. Colonial authorities set up merchants in new administrative centers to stimulate material wants. Poll and animal taxes compelled subsistence farmers and nomads alike to produce goods for French markets. Held to new political boundaries, nomadic tribesmen who had moved their herds across the landscape for centuries increased their livestock density to pay taxes. Farmers moved north into marginal lands to plant crops for export to Europe. Pastoralists expanded south into areas where lack of reliable water and insecurity had previously limited the number of cattle and sheep. Large concentrations of animals around new wells destroyed pastures and left the soil vulnerable to erosive runoff and high winds during violent summer storms.

The Sahel became more evenly and continuously used for increasingly intensive grazing and farming. Between 1930 and 1970, the number of grazing animals doubled; the human population tripled. New French plantations to grow cotton and peanuts as cash crops pushed subsistence farmers onto smaller areas of marginal land. Fallow periods were reduced or eliminated and crop yields began to fall. Ground exposed beneath parched crops dried out and blew away with the wind.

Then in 1972, no rain fell and no gra.s.s grew. Livestock mortality was high where continuous overgrazing left little gra.s.s from the previous year. The few fruit trees that survived bore little fruit. Millions of refugees flooded into huge shantytowns. Between a hundred thousand and a quarter of a million people died of hunger. While the drought was the immediate cause of this disaster, colonial-era cultural and economic changes led to exploitation of the Sahel and allowed the population to grow beyond what the land could sustain during dry spells. It didn't help that crops grown on large plantations were still exported during the famine.

Destruction of perennial plant cover from overgrazing causes desertification by exposing the soil surface to erosion from wind and rain. Erosion rates of half an inch to three-quarters of an inch per year have been reported in semiarid regions once the native perennials are gone. The process is generally irreversible; plants cannot survive the dry season without the water-holding topsoil. Once the soil is gone, the ability to support people disappears too.

During the famine, an image taken by a NASA satellite provided stark confirmation of the human hand in creating the crisis. A mysterious green pentagon in the center of the drought-ravaged zone turned out to be a quarter-million-acre ranch separated from the surrounding desert by a simple barbed wire fence. The ranch, established in the same year that the drought began, was divided into five sectors, with the cattle allowed to graze one sector each year. Limiting the intensity of grazing prevented the problems that brought starvation to the surrounding countryside.

Desertification began in both the Sahel and North Africa during the 195os and r96os, despite above average North African rainfall during these years. Large state-owned ranches established during the i96os showed no evidence of desertification if stocked at the estimated long-term capacity of the rangelands. Although drought reinforces the effects of land degradation, climate variability is not the root cause. Droughts naturally recur in semiarid regions. Drought-adapted ecosystems and societies weathered them in the past. Traditional African pastoralists practiced de facto population control through social structures and rules developed over centuries of alternating scarcity during droughts and abundance during wetter times.

Soil erosion rates from West African fields range from about threequarters of an inch per century on savanna cropland to an extreme of more than ten inches a year on bare-plowed fields in steep formerly forested areas. Some estimates put average erosion rates for Sahel cropland at about an inch per year. In many parts of West Africa the topsoil is only six to eight inches thick. Cultivation after forest clearing quickly strips it off. Maize and cowpea yields in southwestern Nigeria dropped by 30 to 9o percent with the loss of less than five inches of topsoil. As the Nigerian population increased, subsistence farmers moved to steeper land that could not support sustained cultivation. Ca.s.sava plantations on land with slopes steeper than eight degrees lose soil more than seventy times faster than fields sloping less than a degree. Soil erosion rates on Nigerian hillslopes planted in ca.s.sava reach over an inch a year, far beyond any conceivable replacement rate.

Social conventions hindered soil conservation. Subsistence farmers were reluctant to invest in erosion control because they moved their fields every few years. Erosion problems were most severe in areas where communal land ownership discouraged individual efforts to conserve soil. In many West African countries, tractor-hiring schemes are heavily subsidized, so farmers get their fields plowed regardless of the steepness, soil type, or cropping system. Soil erosion rates in sub-Saharan Africa increased twentyfold in the past thirty years. The rapid soil erosion typical of West African agriculture means that only a few years of cultivation are needed to ruin the soil. This, in turn, fuels the drive to clear more land.

In the late 1970s University of Washington professor Tom Dunne and two of his graduate students-one of whom was my graduate advisorcompared recent and long-term erosion rates on the gentle slopes of semiarid grazing lands in Kenya by using the height of dirt pedestals where vegetation of known (or reasonably estimated) age still held the soil on denuded slopes as well as the amount of incision into land surfaces of known geologic ages. Remnant mounds of soil standing eight inches above the general ground surface around the base of fifteen- to thirty-year-old dwarf shrubs indicated modern erosion rates on the order of a quarter to half an inch per year.

Dunne's team determined that the average rate of erosion since the time of the dinosaurs averaged about an inch every three thousand years. The average erosion rate over the past several million years was about an inch every nine hundred years, a little higher than their estimated rate of soil formation of no more than about an inch every two thousand five hundred years. Present rates of erosion, however, ranged from about an inch every decade to half an inch a year. Based on the discrepancy between rates of soil formation and modern erosion rates, they estimated that it would take between two and ten centuries to strip Kenya's gentle slopes to bare rock.

Soil erosion can destroy the vitality of the land-but land can be healed too. Some subsistence farmers in Nigeria made a few simple changes and transformed their fields-at no cost. Tethering their sheep and feeding them crop stubble instead of letting them wander freely allowed collecting manure to fertilize the next crop. Planting cowpeas as part of a crop rotation also helped enhance soil fertility. Low earth-and-stone walls built around the fields kept soil from leaving in heavy rainfall. Crop yields doubled or even tripled without chemical fertilizer. What was required was labor-exactly what subsistence farmers can afford to give. Labor-intensive techniques that restore soil fertility turned the liability of a dense population into an a.s.set.

Ethiopia provides another example of how human societies more often bring soil erosion with them. Medieval deforestation of the northern portion of the kingdom triggered such extensive erosion in Tigre and Eritrea that the hillsides could no longer support grazing animals. By about AD 1000 the economic impact of soil degradation forced the kingdom to relocate its capital to better land in the south. There the process was repeated, as extensive soil erosion followed widespread deforestation. The region remains impoverished, unable to feed itself when the weather doesn't cooperate.

Drought-triggered crop failure brought starvation to almost ten million people in Ethiopia in the mid-i98os. Hundreds of thousands died despite the largest global famine relief effort in history. Long before the twentieth century, farming had expanded from the best agricultural lands onto erosion-p.r.o.ne slopes. Since the 193os, deforestation left Ethiopia with just 3 percent of its original forest cover and increased the silt concentration in the Blue Nile fivefold. The average rate of cropland soil loss in the western highlands would erode the native topsoil in little more than a century. In addition to direct losses to erosion, soil fertility has been projected to fall by as much as i percent annually owing to persistently intensive cultivation by desperate farmers.

Ethiopia's environmental refugee crisis shows how, over the long run, soil security is national security. Recognition of w.a.n.gari Maathai with the 2004 n.o.bel Peace Prize for her work on environmental restoration in Ethiopia's countryside shows that environmental refugees, who now out number political refugees, are an emerging global concern. People may endure temporary droughts, but desertification forces emigration once the land can no longer sustain either grazing or farming.

Desertification is not just happening in Africa. More than a tenth of Earth's land area is desertifying-about a third of the planet's dry lands. Studies over the past fifty years report a pace of desertification in regions with between 5 and 20 inches of annual rainfall that, if continued, would desertify most of the entire semiarid zone in this century. A decade ago, at the 1996 World Food Summit in Rome, global protection and sustainable management of soil were emphasized as critical for the security of future generations.

Before the Second World War, western Europe was the world's only grain-importing region. Latin American grain exports were nearly double those from North America in the late 193os. Exports from the Soviet Union's virgin lands were comparable to those from North America's Great Plains. Self-sufficient before the Second World War, Asia, Latin America, eastern Europe, and Africa all now import grain. By the early i98os over a hundred countries relied on North American grain. Today North America, Australia, and New Zealand are the world's only major grain exporters.

Famine returned to the global scene after decades of unprecedented prosperity in the postwar years when highly variable rainfall, coupled with increasingly severe land degradation, led to regional crop failures. In the mid-196os, the United States shipped 20 percent of its wheat crop to India to prevent famine from two consecutive crop failures. When Indian crops failed again in 1972, more than eight hundred thousand Indians starved to death. This time there was no American bailout; increased Soviet imports had tied up available wheat supplies. In addition, the 1972 Russian grain purchase encouraged U.S. farmers to plow up marginal land, undermining decades of soil conservation efforts. Today the impact of regional crop failures on global grain prices reflects the close balance between world food supplies and demand. The ongoing availability of surplus North American crops is an issue of global security.

Worldwide, over two billion acres of virgin land have been plowed and brought into agricultural use since i86o. Until the last decades of the twentieth century, clearing new land compensated for loss of agricultural land. In the i98os the total amount of land under cultivation began declining for the first time since farming reached the land between the Tigris and Euphrates. In the developed world, the rate at which new (and generally marginal) land was brought under cultivation fell below the rate at which land was being exhausted. Although we use a little more than a tenth of Earth's land surface to grow crops, and another quarter of the world's surface for grazing, there is little unused land suitable for either. About the only places left that could be used for agriculture are the tropical forests where thin, highly erodible soils could only briefly support farming.

Because we are already farming about as much of the planet as can be done sustainably, the potential for global warming to affect agricultural systems is alarming. The direct effects of rising temperatures are worrisome enough. A recent study published in the Proceedings of the NationalAcademy of Sciences reported that an average daily increase in the growing season's minimum temperature of just i0C results in a io percent reduction in rice yields; similar projections hold for wheat and barley. Beyond the immediate effects on crop yields, global warming scenarios that project anywhere from a i0C to a 5C temperature rise over the next century carry a far greater risk.

The world's three great regions of loess soils-the American Midwest, northern Europe, and northern China-produce most of the world's grain. The astounding productivity of modern agriculture depends on the climate of these extensive areas of ideal agricultural soils remaining favorable to crop production. The Canadian and American prairie is already marginal as agricultural land in its western extent. Yet global warming is predicted to increase the severity of droughts here in North America's heartland enough to make that of the Dust Bowl era seem relatively mild. Given the projected doubling of humanity in this century, it is far from certain that the world's population will be able to feed itself.

Other places are predicted to become wetter as global warming leads to a more vigorous hydrological cycle. More frequent high-intensity rainfall events are predicted to substantially increase rainfall erosivity in New England, the mid-Atlantic states, and the Southeast. Models of soil erosion predict from 20 percent to almost 300 percent increases-depending upon how farmers respond to changing rainfall patterns.

Global warming and accelerated erosion are not the only problems facing agricultural land. Growing up in California's Santa Clara Valley, I watched the orchards and fields between Palo Alto and San Jose turn into Silicon Valley. One of the more interesting things I learned from my first job as a foundation inspector was that preparing a building site means carting the topsoil off to a landfill. Sometimes the fine topsoil was sold as fill for use in other projects. Completely paved, Silicon Valley won't feed anyone again for the foreseeable future.

Enough American farms disappeared beneath concrete to cover Nebraska in the three decades from 1945 to 1975. Each year between 1967 and 1977, urbanization converted almost a million acres of U.S. farmland to nonagricultural uses. In the 197os and 198os over a hundred acres of U.S. cropland was converted to nonagricultural uses every hour. Urban expansion gobbled up several percent of the best European farmland in the 196os. Already, urbanization has paved over 15 percent of Britain's agricultural land. The growth of urban areas continues to consume farmland needed to feed cities.

During the cold war, the Department of Agriculture developed tolerance values for soil loss to evaluate the potential for different soils to sustain long-term agricultural production. These values were based on both technical and social inputs-what was considered economically and technically feasible in the 1950s. Soil conservation planning based on this approach typically defines acceptable rates of soil erosion as 5 to 13 tons per hectare per year (2-1o tons per acre per year), equivalent to a loss of an inch of soil in 25 to 125 years (0.2 to i mm per year). However, agronomists generally argue that maintaining soil productivity requires keeping erosion under i ton per hectare per year-loss of less than one inch in 250 years2 to 1o times lower than USDA soil loss tolerance values.

Until recent decades, little hard data were available on rates of soil production. So it was hard to know how big a deal to make out of this problem. Farms were losing soil faster than desired, but it was easy to lose sight of the big picture when farmers were struggling to deal with overproduction and food was cheap. Recent studies using a variety of methods, however, all point to soil production rates much lower than USDA soil loss tolerance values. A review of soil production rates from watersheds around the world found rates of from less than o.1 to 1.9 tons per hectare per year, indicating that the time required to make an inch of soil varies from about 16o years in heather-covered Scotland to more than 4,000 years under deciduous forest in Maryland. Likewise, a global geochemical ma.s.s balance based on budgets of the seven major elements in the earth's crust, soils, and waters pegs the average global rate of soil production at between an inch in 240 years to an inch in 820 years (equivalent to an erosion rate of 0.37 to 1.29 tons per hectare per year). For the loess soils of the Great Plains a soil replacement rate of an inch every 500 years is more realistic than the USDA's acceptable soil loss rates. Hence, the currently "acceptable" rates of soil loss are unsustainable over the long run, as they allow soil erosion to proceed four to twenty-five times faster than soil production.

In 1958 the Department of Agriculture found that almost two-thirds of the country's agricultural land was eroding at what it considered a destructive rate-faster than its soil loss tolerance values. A similar survey a decade later found no progress; two-thirds of the country's farmland still lost soil far faster than was acceptable. Despite soil conservation practices promoted after the Dust Bowl, almost two hundred million acres of American farmland were marginalized or lost to crop production by the 1970s. After two centuries of independence, erosion had stripped away a third of the nation's topsoil. At this pace, we would run out of topsoil in less time than has pa.s.sed since Columbus reached the New World.

By the 1970s many soil conservation plans worked out over the previous decades were abandoned as government policies shifted to support more aggressive cultivation. U.S. farm policy under secretary of agriculture Earl Butz encouraged plowing fence-row-to-fence-row to grow crops to sell to the Russians. Cash crops replaced gra.s.s and legumes in crop rotations as bigger tractors increasingly turned soil conservation measures like contour plowing and terracing into annoying nuisances.

In the late 1970s some in Congress viewed with alarm the fact that soil erosion continued to undermine American agriculture despite forty years of effort. The 1977 Soil and Water Resources Conservation Act required the USDA to conduct an intensive appraisal of the nation's soil. Four years in the making, the 1981 report concluded that American soil still eroded at alarming rate more than four decades after the Dust Bowl. In the 1970s the nation lost four billion tons of soil each year-a billion tons a year more than in the 193os. A freight train filled with all that dirt would stretch around the world twenty four times. At that rate it would take only a century to lose the country's remaining topsoil.

Realistically, political support for spending money to save soil was hard to reconcile with official encouragement to aggressively grow as large a crop as possible to sell overseas. Adjusted for inflation, government support for agricultural conservation programs fell by more than half in the 1970s. No amount of data was going to change congressional perception that the real problem was low prices because of overproduction. Why spend taxpayers' money to save soil when grain bins were bursting?

Part of the problem was that after decades of substantial expenditures on soil conservation programs, there was little solid information on their effectiveness at reducing erosion from America's farms. One of the few welldoc.u.mented examples of such studies found a substantial reduction in soil erosion from c.o.o.n Creek, Wisconsin, for the period from 1936 to 1975. Designated the nation's first conservation demonstration area in 1933, the c.o.o.n Creek basin was severely eroded. Fields plowed in a regular pattern on even steep slopes lacked cover crops, were inadequately manured, and had poor crop rotations. Pastures were overgrazed and eroding. Guided over four decades by the Soil Conservation Service, farmers adopted contour plowing, included cover crops in crop rotations, increased manure applications, and plowed crop residues back into the soil. By 1975 widespread adoption of improved farming practices had reduced hillslope erosion in the basin to just a quarter of what it had been in 1934.

Recent USDA estimates show soil erosion from U.S. cropland as dropping from about three billion tons in 1982 to just under two billion tons in 2001, substantial progress to be sure-but still far ahead of soil production. In the late 199os Indiana farms still lost a ton of soil to harvest a ton of grain. Even though we know that the soil conservation efforts of ancient civilizations proved too little, too late in case after case, we remain on track to repeat their stories. Only this time we're doing it on a global scale.

Across the planet, moderate to extreme soil erosion has degraded 1.2 billion hectares of agricultural land since 1945-an area the size of China and India combined. One estimate places the amount of agricultural land used and abandoned in the past fifty years as equal to the amount farmed today. The United Nations estimates that 38 percent of global cropland has been seriously degraded since the Second World War. Each year farms around the world lose 75 billion metric tons of soil. A 1995 review of the global effects of soil erosion reported the loss of twelve million hectares of arable land each year to soil erosion and land degradation. This would mean that the annual loss of arable land is almost i percent of the total available. Clearly this is not sustainable.

Globally, average cropland erosion of ten to a hundred tons per hectare per year removes soil about ten to a hundred times faster than it forms. So far in the agricultural era, nearly a third of the world's potentially farmable land has been lost to erosion, most of it in the past forty years. In the late 198os a Dutch-led a.s.sessment of global soil erosion found that almost 2 billion hectares of former agricultural lands could no longer support crops. That much land could feed billions of people. We are running out of dirt we cannot afford to lose.

In the mid-199os David Pimentel's research group at Cornell University estimated the economic costs of soil erosion and the potential economic benefits of soil conservation measures. They considered on-site costs for replacing water-holding capacity lost to soil erosion and for using fertilizers to replace lost soil nutrients. They also estimated off-site costs for increased flood damage, lost reservoir capacity, and dredging of silt-choked rivers to maintain navigation. They estimated that undoing damage caused by soil erosion would cost the United States $44 billion a year, and about $40o billion a year worldwide, more than $70 per person on the planethigher than the annual income for most people.

Pimentel's group estimated that it would take an annual investment of about $6 billion to bring erosion rates on U.S. cropland into line with soil production. An additional $z billion a year would do so on U.S. pasturelands. Each dollar invested in soil conservation would save society more than five dollars.

In the short term, though, it can be cheaper for farmers to disregard soil conservation; the cost of reducing soil erosion can be several times the immediate economic benefit of doing so. Farmers with high debt and/or a narrow profit margin can be forced to choose between conserving soil and going bankrupt or working the land until it becomes economically futile. Economic and political incentives encourage practices that destroy soil productivity over the long run, yet preserving the agricultural foundation of civilization requires protecting land from accelerated soil erosion and conversion to other uses.

Many soil conservation measures are proven technologies. Measures adopted to curb soil erosion after the Dust Bowl were not new ideascontour plowing and cover cropping were known more than a century before. Crop rotations, mulching, and the use of cover crops are ancient ideas. So is terracing, which can reduce erosion by 9o percent, enough to offset the typical increase in erosion rates from cultivation.