Description of the Mexico City Aquifer and Its Exploitation
The unique geology of the Basin of Mexico has historically provided abundant water resources for the inhabitants in spite of the scarcity of free flowing surface water. This chapter briefly describes the physical characteristics and hydrogeology of the basin, focusing on the southern portion of the basin where human presence has been a major factor since the Aztec capitol of Tenochtitlán. The history of exploitation of the Mexico City Aquifer and the associated problems of subsidence are briefly examined, and the quantity of water available in ground water storage is reviewed.
PHYSICAL CHARACTERISTICS AND HYDROGEOLOGY
The Basin of Mexico is located in the central part of the Trans-Mexican Volcanic Belt and has an approximate area of 9,000 square kilometers. The basin, at an altitude of 2,200 meters above sea level, is the highest valley in the region, and is surrounded by mountains that reach elevations of over 5,000 meters above sea level. Average annual temperature is 15 degrees centigrade (or about 60 degrees fahrenheit). Most of the 700 millimeters of annual rainfall is concentrated in a few severe storms from June through September with little or no precipitation the remainder of the year.
The basin is a naturally closed depression that was artificially opened in the late 1700s to control urban flooding. Sources of ground water recharge in the basin are largely derived from infiltrated precipitation and snow melt in the
surrounding mountains and foothills, which move as subsurface flow toward the lower elevations. In its natural state, the basin contained a series of lakes, ranging from fresh water lakes at the upper end to those at the lower end where salt is concentrated due to evaporation. The ground water flow produced numerous springs in the foothills and upwellings in the valley (Figure 3–1).
The geology of the southern portion of the basin, south of Sierra Guadalupe, is the best-investigated portion of the Basin of Mexico. This area, which contains Mexico City, is often referred to as the Valley of Mexico, because it is partially divided by several low mountains from the remainder of the basin. Likewise, the aquifer system in this region is often referred to as the Mexico City Aquifer. Details of the subsurface geology of this area, shown schematically in Figure 3–2, have been described by Mooser (1990) and Mooser and Molina (1993). The information is based on data from a series of four deep test holes and reflection seismic profiles performed by Petróleos Mexicanos (or Pemex, the government-owned petroleum company) following the earthquake of September 19, 1985.
Superficial lacustrine clay deposits (i.e., the layer of clay at the former and existing lake bottoms) cover 23 percent of the lower elevations of the Basin of Mexico. The deposits are present in two formations divided by what are referred to as “hard layers” (Capa Duras) composed primarily of silt and sand. The Capa Duras occur between 10 and 40 meters deep and are only a few meters in thickness. The lacustrine clay layers, which reach to a depth of 100 meters, are considered to be an “aquitard,” because they are considerably less permeable than the Capa Duras or the underlying alluvial sediments. The Capa Duras produced the first artesian wells when ground water was exploited early in the 19th century.
Alluvial fill occurs below the lacustrine clays and ranges in thickness from 100 to 500 meters. This material is interstratified with Pleistocene and recent basalt deposits that together comprise the upper portion of the principal water supply aquifer (Units 2, 3, and 4 of Figure 3–2). A lower unit of the principal aquifer is composed of stratified volcanic deposits, 100 to 600 meters thick, and reaches to a depth of 500 to approximately 1000 meters (Unit 6 of Figure 3–2). This lower unit is bounded by a Pliocene lacustrine clay deposit.
Three major hydrologic zones are defined for the Basin of Mexico—the lacustrine zone described above, the piedmont or transition zone, and the mountain zone. The distribution of these three zones can be inferred from the elevation map in Figure 3–3. The lacustrine zone corresponds to the lowest elevations. The piedmont region generally occurs between the historic lake bed and the steep mountains. Here, the lacustrine clays become interbedded with silts and sands, or, in the areas closer to the base of the mountains, the piedmont is composed largely of fractured basalt from volcanic flows. The basalt formation is highly permeable with good storage capacity, and is considered to be a component of the principal aquifer. It is exposed near the upper portion of the
piedmont, and extends below the alluvial deposits of the valley floor. The piedmont, also known as the transition zone, is important for natural recharge of the aquifer. The mountains ringing the Basin of Mexico are volcanic in origin. The Sierra Nevada range is to the east and the Sierra de las Cruces is to the west. The Sierra Chichinautzin to the south forms the most recent chain. Its eruption, approximately 600,000 years ago, blocked what was primarily a southerly drainage and effectively closed the basin. The Sierra Chichinautzin is the most natural important recharge zone for the Mexico City Aquifer due to the high permeability of its basalt rock. The large Xochimilco springs on the basin floor are a discharge point for the underground flow, and some of the most productive wells are located there.
The conceptual model for the Mexico City Aquifer identifies two deeper permeable units—an intermediate and a deep aquifer. These are poorly characterized but are considered to be independent of the principal aquifer. The intermediate aquifer is composed of Miocene volcanic deposits (Units 9, 10, and 10a of Figure 3–2). The underlying Cretaceous limestone formation (Units 11a and 11b of Figure 3–2) may also be an aquifer. Where it is exposed outside the southern portion of the basin, it is currently exploited for ground water.
Historically, the principal aquifer and the shallow aquifer (or Capa Duras) were subject to artesian pressure so that all wells on the valley floor flowed at the surface without pumping. The natural hydraulic gradients caused water to move upward and through the overlying clay aquitards (as was shown in Figure 3–1). In the past century, the proliferation of wells has changed the natural hydrologic conditions. Now, the gradients and flow in the upper layers of the deposits are generally downward toward the heavily pumped zones.
WATER LEVEL DECLINES IN THE AQUIFER AND LAND SUBSIDENCE
Beginning in the fourteenth century, the Aztec city of Tenochtitlán made use of an elaborate system of aqueducts to carry spring water from higher elevations of the southern portion of the Basin of Mexico down to the city situated on land reclaimed from the saline Lake Texcoco. After defeating the
Aztecs in 1520, the Spaniards rebuilt the aqueducts and continued to use spring water until the mid 1850s. The discovery in 1846 of potable ground water under artesian pressure motivated a well-drilling furor (Orozco and Berra, 1864). Over the next century, a combination of increasing ground water extraction and artificial diversions to drain the valley resulted in the drying of many natural springs, shrinking of lakes, and a loss of pressure and subsequent consolidation of the lacustrian clay formation on which the city is built. Consequent land subsidence has been a serious problem in the MCMA since the early 1900s. By 1953, it had been demonstrated that subsidence was linked to ground water extraction, and many wells inside the urban area were shut down.
The first signs of ground water level declines were the drying up of natural springs in the 1930s, coinciding with intensive exploitation of the main aquifer through deep wells (100–200 meter depths). Although ground water levels have
been measured for decades, measurements were performed for particular projects and thus did not give a good indication of regional draw down. In 1983, systematic monitoring of the water levels in the aquifer began (Lesser-Illades et al., 1990). Since that time, the average annual declines in ground water levels range from 0.1 to 1.5 meters per year in the different zones of the MCMA. Water level measurements during the period from 1986 to 1992 show a net lowering of 6 to 10 meters in the heavily pumped zones of this region.
When the shallow aquifer was pumped extensively in the mid to late 1800s, land subsidence was already in progress. By 1895, subsidence had reached an
average of 5 centimeters per year. From 1948 to 1953, increased pumping rates, combined with deeper extraction wells, resulted in subsidence rates of up to 46 centimeters per year in some areas. According to the Mexico Valley Water Authority (Gerencia de Aguas del Valle de México), the net subsidence over the last 100 years has lowered the central, urbanized area of the MCMA by an average of 7.5 meters. The result has been extensive damage to the city’s infrastructure, including building foundations and the sewer system.
The site of Mexico City on the valley floor has always been subject to flooding. One of the most serious problems resulting from subsidence is the lowering of the elevation of the MCMA relative to Texcoco Lake—the natural low point of the southern portion of the basin. In 1900, the bottom of the lake was 3 meters lower than the median level of the city center. By 1974, the lake bottom was 2 meters higher than the city. These changes have aggravated the flooding problem and are reflected in the evolution of the complex drainage system to control flooding (see Figure 3–4). In the early 1900s, drainage of the city was gravity-conducted through the Grand Drainage Canal and out the Tequisquiac Tunnel at the north end of the valley. By 1950, the sinking of the city was such that dikes had to be built to confine the stormwater flow, and pumping was required to lift the drainage water under the city to the level of the drainage canal. The relative rise of the lake level continued to threaten the MCMA with flooding, leading to work on the deep drainage system and excavations to deepen Texcoco Lake.
In 1953, because of the severe subsidence in the city’s center, many wells were closed and new wells were constructed in the southern regions of Chalco, Tláhuac, and Xochimilco. Current pumping rates of 12.2 cubic meters per second (cms) in this region have likewise resulted in subsidence and lowering of the water levels. Several lakes have formed in depressions created by lowered ground levels in the pumping area. As pumping continues, these lakes are expanding. Figure 3–5 shows comparative subsidence in the city center area of Mexico City and the Chalco plain from about 1935 to the present (Ortega et al., 1993).
In 1925, Roberto Gayol reported to the Mexican Society of Engineers and Architects that surveys showed Mexico City was sinking, and that the cause was likely subsurface drainge related to the recently completed construction of the Grand Drainage Canal and the Tunnel of Tequisquiac. The linkage between subsidence and exploitation of the aquifer has been closely examined since that time. Nabor Carrillo was the first to develop a mathematical model linking subsidence to the hydrology (Carrillo, 1948). Observation wells were installed, and institutional research programs began with the Comisión Hidrológica del Valle de México and its successors, the Comision de Aguas de Valle de México, and the Gerencia de Aguas de Valle de México (SARH, 1953–1969). An early and comprehensive review of subsidence was performed by Hiriat and Marsal (1969). Modern ground water models were developed for the
semiconfined, multiple aquifers in the southern portion of the Basin of Mexico, and applied to the question of subsidence in the MCMA (Herrera and Figueroa, 1969; Herrera, 1970), and in related studies (Bredehoeft and Pinder, 1970). The Federal District is currently using more recent developments of these models (Herera, et al., 1989; Herrera et al., 1994) in conjunction with a network of 320 observation wells for determination of water level and flow direction. Every two years, more than 1,400 surveys are performed to gauge changes in subsidence.
WATER BALANCE FOR THE AQUIFER
It is common practice to refer to a water balance in order to determine the quantities of water available for use, and a ground water balance is often at-
tempted where appropriate. At a gross level, estimating a water balance for surface waters is fairly straightforward because the main input to a surface water catchment—precipitation—can be measured. Estimates are less precise for a ground water system because all data for the computations (properties of the media, subsurface geology, and definition of the flow systems) have inherent errors that make all calculations questionable. Finally, most ground water systems respond to stresses much more slowly than surface water systems so that water balances are not often used except for very long-term considerations. A further complication is that the water balance for the aquifer may be quite different than the balance for the entire ground water system; much of the water that enters the ground water system may not reach the main aquifer under consideration.
By far, the best way to determine the water balance for an aquifer is to use long term records of pumping and ground water levels. Declining water levels demonstrate that more water is leaving the system than entering, and indicate a state of overdraft. Most unexploited aquifers are in a state of quasi-equilibrium. Seasonal or cyclic fluctuations are expected, but in the absence of major climatic shifts, the long-term water levels remain stable under natural conditions.
Field measurements have shown that the water table of the aquifer supplying Mexico City has been declining by approximately 1 meter per year (Herrera, et al., 1994), and overdrafting of the aquifer has been occurring at least since the early 1900s. The question of how long this rate of exploitation can continue has been a subject of debate.
The best estimates of the quantity of ground water in storage have come from investigations of the southern portion of the Basin of Mexico (generally from Sierra Guadalupe, south) where the majority of geological studies have been made. In estimating the volume of ground water in storage, it is important to consider the contributions from the upper clay layer, and the fact that the clay layer (the aquitard) is not acting as a confining layer for approximately 30 percent of its extent where water levels have sunk below this limiting layer. On the basis of both field measurements and modeling in this region, the total saturated volume of the aquifer in the southern portion of the basin has been estimated to be 1,189.3 billion cubic meters. The annual draft in this region is estimated at 27.9 cms1. This rate of extraction equates to a ground water loss of either 3.45 or 5.59 billion cubic meters per year. The difference depends on whether or not the calculations consider the water held in the aquitard as contributing to the volume of water in the main aquifer. At these rates of depletion, the calculated volume of storage is 212 to 344 times the annual draft (see Herrera et al., 1994, and AIC-ANIAC, 1995, for further details of the calculation).
Ground water extraction rates for the entire Basin of Mexico, for supplying the MCMA, are estimated to be 43 cubic meters per second (see Chapter 4).
While this kind of water balance is commonly used to estimate changes in ground water volumes, it does not provide a basis for developing long-term rates of extraction. In the MCMA, subsidence is the penalty for overdrafting. The damage to the drainage system and other public works has been previously noted. In addition, and as described more fully in Chapter 4, the aquifer is vulnerable to the contamination that accompanies the consolidation, dewatering, and fracturing of the clay layers of the aquitard. A simplistic water balance approach does not account for other realities as well. The actual volume available in the main aquifer would likely be less than estimated because of probable decreasing porosity with increasing depth. Also, there are practical, economic limits to the depth of pumping. Finally, the deep well tests by PEMEX (the government-operated oil company) in the late 1980’s have indicated a likelihood of geologically-induced water quality problems with increasing aquifer depth.
More precise information on the sustainability of continued extraction from the aquifer would require specific studies including field observations and the use of computational models (see AIC-ANIAC, 1995, for additional details on water balance, hydrogeology, and exploitation of the aquifer).