Overloading: the Leaning Tower of Pisa

Classic cases of subsidence by overloading have occurred where engineers have built structures without considering the load-bearing capacity of the underlying rocks. Generally these classic cases involve soft sediments and heavy structures.

Very heavy structures can be built safely on strong substrates‹the Pyramids are a good example. Many of the castles and cathedrals of Europe are built on hills, and most hills are high because they are made of hard material that can stand heavy loading. But there have been cases where engineers have ignored the strength of geological substrate to their cost.

In 1173 a new bell tower was built for the cathedral in Pisa, at the time one of the richest cities in northern Italy. The tower was in trouble before it was even begun. The engineers dug down only 2 m to lay the foundation. If they had investigated the underlying sediment (digging wells was a routine chore in medieval times), they would have found 2 m of compressible clay only half a meter below their inadequate foundation. As it was, the tower was already sinking and leaning to one side before they had built it three stories high, as the clay was compressed unevenly. Naturally, this is unstable: any further building simply increases the uneven loading, and tends to cause even more uneven subsidence.

Construction then stopped for close to a century, and so did the tilting. In 1274 construction began again, and once again, the builders found that the tower sank and leaned more as it was built higher (and heavier). Ten years later the tower was up to its seventh level, 60 m high, requiring only the bell chamber, but the tilting was so rapid that the work had to be stopped: the seventh level was leaning out by a meter at this time. Because the tower was now heavier, it continued to subside on one side, and the seventh level was 1.5 m out of alignment when the Pisans had forgotten the problem and added the bell chamber in 1350.

Six hundred years later the Leaning Tower of Pisa is one of the most famous buildings in the world, but the seventh level now leans more than 5 m out of true, that is, it angles out at close to 5°. In the same time the building has sunk 2 m into the ground, in effect providing itself with a deeper foundation, and probably slowing the tilt. The rate of tilting today is slow, but the building remains unstable, and there is now a significant danger of collapse. The tower was closed to visitors in 1990. In 1992 engineers began to load the structure with 800 tonnes of lead, to try to even out the stresses on it. The latest plan is to burrow under the tower and take out a slice of sediment from under the uphill side.

Fluid Withdrawal

Most of the case studies about þuid withdrawal are related to water, because water is often pumped from rather shallow, poorly-consolidated rocks and soils, but pumping out oil from shallow reservoirs can and does give similar effects.

Two oil fields have been particularly well studied: the Goose Creek oil field in Texas, and the Wilmington oil field in Long Beach Harbor, California. Each is close to sealevel, so that the subsidence became obvious very quickly. Subsidence in the Wilmington oil field caused a great deal of damage because it was in the center of a busy port and industrial area, and it can be used to demonstrate the problem most effectively. In the greater Houston area, subsidence was first noticed over the Goose Creek oil field, but by now the area is having far greater problems with groundwater pumping than it ever did with oil extraction.

Long Beach, California.
The Wilmington anticline is a broad gentle structure about 5 km wide but 18 km long, running through the harbor of Long Beach, California. Oil was discovered in the area in 1932, and extensive pumping began in the late 1930s. By 1942 there were 1000 producing wells in the Wilmington field.

Because the Wilmington field stands only a few feet above sea level, and is centered right in a major international harbor, subsidence was noticed by the summer of 1941.It was not clear at first whether the subsidence was related to oil extraction, or to the numerous water wells that supplied the city. Reports in 1947 and 1949 specifically related the subsidence to underground fluid extraction, however, with the oil extraction fingered as the most likely agent. Oil production was still increasing at that time, and continued to do so until 1951: and pumping stayed at high levels for years.

The harbor area was particularly affected. It stood originally 5­10 feet above sealevel. But the ground level subsided over a foot by 1941, and 4 feet by 1945. The Long Beach Harbor Board established benchmarks, and began to monotor them closely as portions of the harbor subsided toward sea level. The rate of subsidence rose to a foot a year by 1947, and over 2 feet a year by 1951, when oil production reached a maximum, and parts of the harbor area had sunk significantly below sea level.

By 1958, ground level had dropped by 27 feet in the center of the field, and about 25 square miles of ground had dropped 2 feet or more. In other words, a large area of heavily industrialized city had sunk appreciably, in places far below sea level, and in the process major industrial and port facilities were damaged, and the subsidence had required the construction of protective levees and breakwaters. The bill was at least $100 million.

Fortunately there was a partial solution to the subsidence that pleased everyone. The þuids that had been removed were partially replaced by pumping salty water into the oil-bearing rocks, beginning in 1953 and reaching effective levels by 1958. This was cheap in comparison with the prospect of further remedies for subsidence (the injection program cost $30 million). In fact, the underground flooding not only stopped subsidence and restored some of the þuid pressure, but it greatly stimulated oil production by driving out more petroleum. The subsidence slowed to a halt in 1962. What's more, water injection could hold the subsidence in check as pumping continued.

Oil pumping was actually expanded: the THUMS consortium of five giant oil companies (Texaco, Humble, Union, Mobil, and Shell) was allowed to expand its operations in 1965 to drill into the undersea extension of the Wilmington oilfield from artificially constructed islands. For the Long Beach, the equation included damage to one of the major industrial, commercial, and strategic ports of the United States, but was balanced by an oilfield that was the largest ever exploited in California. (When it produced its billionth barrel in 1964 the Wilmington oilfield became a "supergiant" field, the first in California, and the second in the United States.)

The City of Long Beach drew massive royalties from the oil production, and damage to the port was mitigated not from its own resources, but by the Federal taxpayer at large through the Army Corps of Engineers and the Navy. Water injection had turned the problem from critical to minor, and this partial solution for the subsidence allowed the continuation of both port and oilfield operations. By 1971 the Wilmington Oilfield was producing 22% of California's entire production, THUMS was pumping oil from 652 wells, and the City of Long Beach was drawing royalties at $4 million a month.

It is not clear what Long Beach would have done if there had been no easy solution to the subsidence problem. Contrast this with the situation in Venice.

Venice and the Po Delta.
The Po delta lies at the end of the river system that drains most of the southern face of the Alps, as well as a large section of northern Italy. Historically, many deltas are prone to natural subsidence, and the Po delta seems to subside at perhaps 2 mm a year.

Venice lies on an island in a lagoon just north of the Po Delta. But the Brenta and Piave rivers also deliver sediment to a continuous deltaic plain, and the whole system makes a geographical unit. The subsidence of the delta was counteracted to some extent by continuous delivery of silt by the rivers. In 1550, however, the Venetians diverted the rivers away from the lagoon, so they would no longer have to deal with the problems of siltation. The results were slow to have impact, but they were disastrous. The sinking of the city could have been counteracted by new building if it had not been accompanied by flooding. Venice is subject to episodes of tidal or storm flooding that invades much of the city, causing not only economic but cultural damage as its architecture and art are repeatedly flooded.

Natural gas (methane) was discovered under the Po delta after World War II, and used to supply new industry in the region. Two major new industrial centers were built on landfill on the edges of the Venetian lagoon, and at the same time the entrances to the lagoon were dredged to allow larger ships (especially tankers) in and out. Beginning in about 1950, the delta began to subside at 30 cm a year, and Venice experienced even greater floods than before.

Many factors were blamed for this environmental disaster (rising sea level, natural subsidence, peat shrinkage, and so on). Natural gas pumping was always suspected, and finally in 1960 it was stopped on an experimental basis in the area of most rapid subsidence. After 6 months, subsidence rates halved in that area, while they continued unchanged elsewhere.

The cause was now clear: the question was what to do about it? The cost of protecting the subsiding coast from flooding was running at about $30 million a year by 1960. One proposal was to close all methane wells in the Delta, thus sacrificing the natural gas industry, and forcing the industries that depended on it to seek other (much more expensive) feedstocks and fuel supplies.

Whatever was done about methane, Venice suffered its greatest flood to date on 4 November 1966, when a tide rose 2 m above normal in a storm, and caused damage estimated at $6 billion. Water was over a meter deep in the Piazza San Marco, St. Mark's Square. A major commission was set up to recommend action. It now appeared that groundwater extraction from the Delta sediments under Venice, and particularly under the industrial regions onshore was responsible for a good deal of the subsidence. In the 1970s, ground water extraction was limited, and a new aqueduct was built to supply water to the region from the Alps. The subsidence was only stopped, however; there was no recovery. Venice still remained desperately vulnerable to flood. Further recommendations bogged down in bureaucratic argument.

In the end, nothing more was done until 1982, when floating barriers to block high tides from the lagoon were suggested as a solution. The project was promised financing in 1984, but the recession of 1991 dried up Government funds before any floodgates were completed. This may have been a blessing in some ways, because the lagoon was already suffering horrendously from algal pollution. The pollution was caused in turn by the untreated sewage generated by (100,000) resident Venetians and 5 million transient tourists per year, and made worse because existing tidal currents are not very good at flushing the lagoon water out into the Adriatic Sea. It's clear that all the reasonable and fairly cheap fixes have been tried. Perhaps Venice will have to be allowed to sink.

Water and Oil in the Greater Houston area, Texas.
The Goose Creek field lies on the coast of Texas close to Houston. The field went into production in 1917, and by the end of 1918 the small peninsula that separated the field from San Jacinto Bay had subsided under water. Maximum subsidence was 5 feet by 1925, and overall the area affected was close to 4 square miles, corresponding in area and shape with the area that contained the producing oil wells. The correlation was dramatic, and the cause and effect were easy to see. The subsidence of the field caused faults to develop round its edges, and boundary faults 12-16 inches in offset developed along the north and south edges of the field as it sank.

The Goose Creek field is long ago exhausted, and it was only one of many oil fields in the region. Most of them extract oil from fields too deep to have caused surface subsidence. But the greater Houston region has come to rely heavily on ground water as the population has grown toward megalopolitan size. Ground water is used for public water supplies, for industrial use, and for rice irrigation, and groundwater pumping has come to be the greatest factor in continuing ground subsidence in the region. Groundwater pumping accelerated as early as the 1930s, and ground subsidence from water extraction was noticed in the 1940s.

Hurricane Carla reminded the area in 1961 of the dangers of a 5-m storm surge on low-lying coastal areas: Carla flooded 319 sq km, but no action was taken to slow groundwater pumping. By 1973 an area of 12,000 sq. km had subsided at least six inches, forming one of the largest subsidence bowls in the United States. Small faults were activated, sometimes with as much as a meter of vertical movement, and water and sewer lines and pavements cracked. The regional ground subsidence was small by Long Beach standards (0.3 m everywhere within 20 miles of Houston; 1.5 m throughout the area from Houston to Galveston Bay; 2 m in the Goose Creek oil field; and a maximum of 2.7 m along the Houston Ship Channel), but that is very significant in terms of damage and flood vulnerability in a large metropolitan area on a flat coastal plain very close to sea level and subject to hurricanes. Total damage from subsidence was estimated at $113 million, and tropical storm Delia in that year caused subsidence-related damage of $53 million.

By 1975 some parts of the suburb of Baytown were under water most of the time, and relocation costs approached $17 million. The Texas Bureau of Economic Geology warned that a Carla-sized hurricane in 1976 would flood at least 25 sq km more land because of the subsidence that had occurred since 1961. A levee built around subdivisions in the low-lying area did not stop the storm surge from hurricane Alicia in 1983, and over 100 homes were destroyed.

The only remedy is to stop further groundwater pumping, and to inject water back into the underground aquifers to try to get them to rebound at least a little. This will not restore the original land surface, but it will stop further subsidence. Given the thirst of modern megalopolitan areas, however, and the cost of importing surface water in this area of Texas, it is unlikely that this course of action will be taken.

After 1978, the rate of groundwater pumping slowed in some of the worst-affected areas, but was made up by pumping larger amounts from other areas. Subsidence rates followed the shift in groundwater pumping rates: in 1984, some parts of Greater Houston were still subsiding at 5 cm (2 inches) a year. Western Houston region shows increasing rates of subsidence, ranging up to 70 mm/yr from 1973 to 1987, only 7 miles west of downtown Houston. Such subsidence rates are critical: downtown Houston stands 15 m above sealevel, but the Johnson Space Center is only 4 m above sea level.

The San Joaquin Valley, California
In the San Joaquin Valley, California's most extensive problems have been building for years since the problem was noticed in 1935. Outrageous amounts of ground water have been pumped for irrigation in the semi-desert climate, at much greater rates than it is replaced by rainfall and percolation.

Groundwater use began early in the San Joaquin Valley, but increased dramatically with the advent of cheap power in the 1920s. By the 1950s the problem of subsidence became obvious. The Bureau of Reclamation constructed the Delta-Mendota Canal through the Los Banos-Kettleman City area, completing it in 1951. The canal bed immediately began to warp, and by 1964 there had been over 5 feet of differential settlement along part of its route. In 1956 the US Geological Survey was asked to study the whole problem of subsidence in the San Joaquin Valley. By this time groundwater pumping in the San Joaquin Valley alone accounted for 25% of all groundwater pumped for irrigation in the entire continental US. The water table had dropped 400 feet in some places, and at least 35% of the valley floor showed clear signs of ground subsidence, at rates that had reached 18 inches a year in some places.

Three areas in particular were having severe difficulties: the Los Banos-Kettleman Hills area, on the west side of the valley, where the ground had subsided as much as 23 feet; the Tulare-Wasco area (the western Tulare Basin, especially near Delano), where subsidence had reached 12 feet; and the Arvin-Maricopa area in the southern valley, with at least 6 feet of subsidence. By 1972 half the San Joaquin Valley had been identified as subsiding, and the maximum subsidence had reached 28 feet.

Subsidence and Faulting

Subsidence is often accompanied by large-scale ground cracking, and in some cases the cracking has movement across it, making it into incipient or actual faulting. The Baldwin Hills Reservoir.
The Baldwin Hills Reservoir was a 20-acre water-storage basin built in 1951 by the Los Angeles City Department of Water and Power on a low hilltop in the northwest Los Angeles Basin. It was an ideal location to provide good water pressure and supply storage for much of the city. A shallow basin was excavated in the hilltop, but the volume was increased by building low compacted earth dikes on three sides, while the lowest side was formed by a large earth dam 232 feet high and 650 feet long. On the afternoon of Saturday 14 December 1963, the earth dam failed, and 250 million gallons of water poured out, killing five people and destroying 277 homes. Evacuation teams had had three hours warning, and frantic evacuation efforts had just managed to clear the area below the dam in time to avoid an enormous loss of life.

Everyone agrees that the Baldwin Hills Reservoir was built right on a small active fault, now known as the Reservoir Fault, that is part of the same system as the much larger, active Inglewood Fault that passes only 150 m west of the Reservoir. Movement of up to 7 inches on the Reservoir Fault probably triggered the dam failure by breaking the asphalt floor of the reservoir, and allowing water to undermine the dam by washing soft sediment out of it.

However, the large Inglewood oilfield also lies under the Baldwin Hills. Although flood damage costs had initially been paid by the City's insurance companies, in 1966 the City of Los Angeles and the insurance companies jointly sued the companies operating the Inglewood oilfield, claiming that oilfield operations had led directly to the fault movement and therefore to the dam failure. The suits were settled out of court for nearly $4 million, so that the case never came to trial. How could oilfield operations have caused this disaster? And could it have been foreseen?

The Inglewood oilfield had discovered in 1924, and was developed rapidly. Pumping became more expensive as oil pressure dropped, and production declined. In 1954, three years after the Baldwin Hills Reservoir was completed, Standard Oil began to experiment by pumping salt water into the oilfield to help to drive more oil out in so-called "secondary recovery." The results were good, and Standard began full-scale brine pumping in 1957.

Meanwhile, ground subsidence had been taking place above the Inglewood oilfield. The Los Angeles Department of Water and Power had been concerned about it as they planned the Baldwin Hills Reservoir, but the data were not good enough to define the extent of subsidence until 1955, four years after the Reservoir had been built. By 1957, however, surface cracking and faulting had begun to show in surrounding streets, at precisely the time that Standard began its full-scale brine-injection program. Some areas of the Baldwin Hills had settled as much as 10 feet over 40 years, and the hilltop on which the reservoir was built subsided about 3 feet, with the southwest corner dropping more than the northeast corner. Eight more surface faults were activated in the local area by 1963, when finally the dam broke. The crucial crack in the reservoir þoor ran along a line that continued southward into the Inglewood oilfield.

Mexico City: a Case Study in Urban Geology

With perhaps 20 million people, Mexico City is one of the largest cities in the world, but it stands on a site that should never have been used for urban development. It became the center of a civilization because of its position in a well-watered and fertile basin in a dry and inhospitable region. That same geographical accident makes it vulnerable to major earthquake damage because it sits on an old lake bed in a thick alluvial basin. But man-made problems have added to the negative aspects of the Mexico City basin: most of them are related to the misuse of ground water from the sediments under the city.

The Valley of Mexico is a closed basin, so it does not drain. Evaporation is more than enough to keep pace with water supply, and under natural conditions the valley floor was covered with shallow natural lakes whose level fluctuated with the seasons and with the climatic cycles of years, decades, and centuries.

Streams from the surrounding hills laid down a great thickness of alluvial fill. The run-off from the basaltic volcanic rocks brought silt and clay rather than sand, so that the valley floor was quite flat, the sediments were soft and oozy, and the lakes that covered perhaps 15% of the valley floor were very shallow.

The Aztecs arrived in the valley about 1300 AD, after its edges had already been fully colonized. They made their living chiefly on their well-deserved reputation for savagery, and hired themselves out as mercenaries. Eventually they settled not on the shore, but on small swampy islands in Lake Texcoco that they modified into their capital city of Tenochtitlan.

The valley's inhabitants had already invented or learned the chinampa agricultural system. Mud is scooped from the valley floor and piled into long mounds or ridges (chinampas) separated by the trenches from which the mud was dug: the trenches form an interlocking network connected with the lake and fill up with water. Plants, including trees, are grown on the chinampas, which are stabilized by the tree roots while the crops are well lit and warmed directly by the sun. The plants are watered, humidified, and protected against spring and fall frosts because they grow close to the water level, which keeps the microclimate mild. The mud is full of nutrients, and by scooping new mud on to the chinampas after each spring flood, the crops receive fresh nutrients every year. Fish and waterfowl can be caught in the trenches and on the lake.

The Aztecs absorbed these techniques and characteristically turned them to military advantage. Tenochtitlan was, in effect, built on a giant chinampa in Lake Texcoco, and eventually neighboring islands such as Tlatelolco were annexed, given the same treatment, and amalgamated into Tenochtitlan by causeways built on lake mud. As the city grew in size, communal latrines were built over specially dedicated boats floating at the city edges, and specially dedicated boatmen then rowed their cargoes for sale to tanners and chinampa farmers. Canals, of course, provided the main means of transport within the city and between the city and the settlements on the lake shore. (This system was very effective, since pre-Columbian Americans had neither horses nor wheels.)

Naturally, the Aztecs came up against the same problems of urban planning as any other "developing" society. Their problems were accentuated, because they were constrained to live on these swampy islands by what they perceived as the requirements of military security.

As individual and civic buildings became larger, the Aztecs had to overcome the problems imposed by their unstable island home. They used piles driven into the mud for foundations on which they could build stone structures. Everyone but the aristocracy was limited to a one-story home. The largest buildings were confined by choice or necessity to the firmer sites of the original island cores of Tenochtitlan and Tlatelolco, and were constructed of light vesicular volcanic rock, not because it was the only rock available, but because it was light. Even so, heavy ceremonial buildings subsided. The "Templo Mayor" on Tenochtitlan, dedicated to Tlaloc and Huitzilopochtli, the major pyramid within Mexico City itself, is not very dense: it is primarily a mound of clay, faced with light vesicular basaltic rock. Yet even it is made of several structures. Each level is not parallel with the next: presumably the temple subsided askew, and as it became necessary to add more height the restorers took advantage of the opportunity to straighten it out.

As the cities grew, the Aztecs came to dominate the valley politically and militarily. Causeways connected the city to the lake shores, and new environmental problems arose. Fluctuations of the lake, which had not been more than a nuisance to chinampa farmers, affected daily life more. After a catastrophic flood, the Aztecs built dikes all the way across Lake Texcoco to hold back the floods which usually came from the north. Because the dikes interfered with water flow in the valley, an artificial water supply had to be built for Tenochtitlan in the form of a double aqueduct that ran from Chapultepec.

This civic system was in place by the time the Spanish arrived. Bernal Díaz wrote that the city looked enchanted, with towers, temples, and buildings rising from the waters. But its water supply was its weak point. Cortés was able to disrupt the aqueduct in his final assault on Tenochtitlan and Tlatelolco in 1521. As he advanced, he destroyed many of the buildings in the city, using the rubble to fill in many of the canals so that his cavalry could maneuver.

At this point the Spanish were faced with the question of holding down a hostile population hundreds of miles from the safety of their ships, and finding a secure base for further pillage, looting, and conquest. No 16th century European would normally dream of building a city in a swamp: contemporary Venice was a flourishing city only because it could use naval strength from a superbly sheltered harbor secure from invasion by numerically stronger armies. As Venetian traders and admirals chose sites for their colony-fortresses, they invariably chose rocky sites over swamps. The Venetian problems of subsiding buildings, contaminated water supply, civic pollution, and raging malaria were chronic even at the height of its power.

Tenochtitlan was as anomalous to the Spanish as Venice would have been: it had canals rather than streets; it had no easily accessible grazing for cattle, sheep, goats, or horses, and it provided no good ground for the style of building that the Spanish were used to live in.

Why, then, did Cortés choose an otherwise unsuitable site for his headquarters? Cortés was no visionary planner and builder: he wasn't even a particularly good general and soldier. He was a legalized bandit, a match for the Aztecs in cruelty and more than a match in ruthless determination. He was a superb judge of people and a skilled manipulator of their weaknesses. He put his finger and dagger point directly on the Aztec jugular. His motivations for rebuilding the new Mexico City on the site of Tenochtitlan were military and symbolic, and were immediately pressing. He did not want to leave the site of the Aztec capital in case they re-occupied it; he wanted the psychological impact of levelling the Aztec temples and replacing them with the imposing Spanish architecture of fortification, palace, and cathedral: it is no accident that the Palacio Nacional and the Cathedral stand on the leveled ruins of the great pyramid of Tenochtitlan. Furthermore, given the limitations of manpower imposed on Cortés, no other site in the valley was as defensible as Tenochtitlan, or as independent of the goodwill of some fickle ally.

The growth of the city involved draining the lake, and much of the Spanish construction was built on the old lake bed. The Spanish did build with some eye to the inferiority of the site. Like the Aztecs, their buildings were generally low. A 16th century writer ascribes this to an idealistic desire to keep ground level light and airy, and more pragmatically, because earthquakes would level anything higher.

The rebuilding of Mexico City with Spanish technology also had the result of deforesting the slopes of the valley: timber could now be transported easily on wheeled carts drawn by horse or mule, and over 25,000 trees a year were cut from the valley slopes and converted into piles driven into the lake bed as foundation materials. New buildings still subsided and tilted, and the floods still came (probably run-off was worsened by the deforestation). Three destructive floods in the 1550s made the Spanish realize that the site was untenable. Something had to be done: the choice lay between rebuilding somewhere else, or trying to solve the problems of the current site. They took the age-old step of trying to make the city inhabitable by a giant public works project: in other words, they set out to cure the symptom rather than the disease. Such decisions are, of course, commonplace in history: examples that are just as bad include Los Angeles and New Orleans.

The Spanish decision is less defensible than Cortés' decision thirty years previously. There was no longer any organized resistance to Spanish control, and therefore the military factors that had weighed so heavily with Cortés were no longer relevant. The Spanish administrators of the 1550s were not unsophisticated legalized bandits like Cortés: they included seasoned administrators and well-educated ecclesiastics. The decision was not made with a horizon of a few years, as Cortés' had been: the Spanish administration was looking at a horizon of centuries. Almost all the problems now faced by Mexico City today stem from the decision to maintain the new Hispanic capital on the site of Tenochtitlan instead of moving it to better ground nearby.

First, the Spanish tried the same method the Aztecs had tried nearly 200 years before. A dike was built to hold back the flood waters. But the city flooded again in 1579 and again in 1580. This time the Spanish, blaming the lake water for rising and flooding the city, made the incredible assumption that if they could drain the lakes the flood problem would go away. In such a case, they reasoned, the lake bed could be made into valuable farm land, and the city could extend out over the countryside.

Nothing much happened in the next few decades, which were generally drier. The canals of the city were gradually filled in as it was converted more and more into a European city. But a new flood in 1607 forced a new appraisal of the situation. This time the Spanish decided to build new dikes, and to begin to drain the lakes. An ambitious but abortive attempt to cut a tunnel right out of the basin to drain off water had failed by the time a devastating flood in 1627 left Mexico City under water for four years.

Once again, and for the last time, the Spanish had the opportunity to move their capital. This time, vested interests brought such pressure to bear that the decision was made to stay. The lakes were gradually drained, so that Mexico City was never again flooded to the same extent. However, the draining produced its own disastrous effects. The basis of the chinampa system, the renewal of fertility each year, was gone, and as the surface soils built up salts from evaporation, they became less and less fertile. The imported Spanish technology of cultivation by deep plough broke up the fine clay and silt at the surface, and as salinity rose and fertility dropped, the loose fine-grained soil was bare and barren for more of the year. Great dust storms began to blow about the valley, creating problems for health as well as agriculture.

Even 200 years after the conquest, Spanish engineers had not fully assimilated the lessons of building in the Valley that had been bequeathed to them by the Aztecs. In 1709 they built the Basilica of Our Lady of Guadelupe, an enormous structure that was begun on the old lake shore but extended out over old lake bed. The eastern towers and chapel now tilt precariously with cracked walls as the building has subsided unevenly over the decades.

Naturally, as normal (flood)water supplies to the city were intercepted, a shortage of surface drinking water arose, just as it had for the Aztecs. But the lake sediments formed a natural artesian basin, and shallow wells provided abundant water for the growing city. 140 artesian wells were þowing without pumping in 1854.

At first, shallow wells under the city were sufficient to supply the people. But as the city grew from 500,000 people in 1895 to 1 million in 1922 and 5 million by 1960, the artesian supply could not keep up with demand. By 1957 more than 3000 wells were needed to supply the city, and artesian pressure had been reduced so that water had to pumped from 20-30 m underground.

Of course, taking water from underground aquifers in soft sediments adds to any subsidence. The subsidence of large buildings became even more acute this century than it had previously, as the city's shallow aquifers have gradually been exhausted. Ground subsidence was reported in 1929: the heavy buildings of the Aztec and Spanish had always had a tendency to sink into the old lake bed, but the general subsidence added to the problem. Pumping of ground water for human and industrial use accelerated it in subsequent decades, and ground level was sinking about 25 cm a year in the 1950s. By 1959, much of the old city had sunk 13 feet, and some of it had sunk 24 feet. The Palacio de Bellas Artes sank 13 feet in 40 years after it was built. Parts of the city have subsided 25 feet! this century. In the late 1940s the rate was 31 inches a year.

Obviously, uniform subsidence would not be as much of a problem, but variable subsidence, with some areas and some buildings sinking more than others, is a serious problem not just for individual buildings but for underground city facilities such as gas, oil, and water pipes, and sewage systems. Ground subsidence in the basin altered the existing sewer lines so that they led in and under the city instead of out and away, so that all sewere and storm water must now be pumped out of the city. This problem had begun earlier: a giant drainage canal had been cut during the 17th century to drain lake water away from the city, and a larger one, the Gran Canal, in the 19th century. A tunnel was added this century, and yet another is being constructed now, to take waste water under the surrounding hills to drain away into a desert basin to the north.

Mexico City's setting in a closed basin has led to horrendous air pollution problems as the population has increased, but that is not the focus of this discussion. The final problem is relevant to earth science, however, and that is the threat of earthquake.

One of the great earthquake zones of the world lies off the west coast of Central America, where Pacific Ocean floor is being destroyed under the continental rocks of Mexico: in fact, the volcanoes that ring Mexico City are part of that process. Although the great earthquakes of Western Mexico are often several hundred kilometers away from the Valley of Mexico, the shape of the valley and its deep load of soft sediment combine to generate shaking of much greater amplitude and duration than one would normally expect. In other words, the intensity of an earthquake in Mexico City is amplified more than one would expect from the magnitude of that earthquake. This probably doesn't affect a largely rural population of chinampa farmers very much, but is critically dangerous for a dense urban population. Major earthquakes have shaken the city periodically, but the earthquakes of 1957 and 1985 gave two more in a series of reminders of that fact of nature.

Web pages