Figure 5.-Geologic map of the Kharga-Aswan region of Egypt. The colors of the Multispectral Infrared Radiometer spectra indicate rock types. Gravel and sandstone contain two different clay materials. The solid line shows the average spectrum, and the dashed line indicates variation.

channels, are keyed to the map using colors to indicate the rock type measured in each area. The solid dark line is the average of several spectra, and the color envelopes bounded by dashed lines express the variation. The spectrum located in the upper left corner is a single spectrum.

Mineralogical determinations are made by comparing the radiometer spectra acquired during the shuttle flight with 10-channel radiometer spectra made in the laboratory prior to the flight for a large number of rocks and minerals. Only the most conspicuous absorption features are evident in the radiometer spectra. Nevertheless, the shapes of the radiometer spectra can be used for identifying certain important minerals. The radiometer spectra, shown in dark blue and green in figure 5, have the shape that is characteristic of limestone, particularly the marked decrease in reflectance in the 2.35-micrometer channel that is

caused by intense absorption in calcite (fig. 4). The other radiometer spectra indicate the presence of two types of clay minerals, one shown as red and the other as orange. The spectra shown in red indicate the presence of kaolinite in the sandstone.

The radiometer spectra shown for the rock units in figure 5 characterize the particular units for great distances; the spectra change abruptly at lithologic contacts. Although the radiometer collected spectra along narrow tracks, these results suggest that many lithologic units could be differentiated over large areas on the basis of specified mineralogical differences in highspectral resolution images recorded from a satellite platform. Realization of this goal would permit mapping of mineralogical differences that require years of laboratory and field work using conventional methods. Such systems are already in the developmental and initial evaluation stages.

More than 86 historically active faults with an aggregate length of 150 miles have been identified within and adjacent to the Houston, Texas, metropolitan area (fig. 6). Although scarps of these faults grow gradually and without causing damaging earthquakes, historical fault offset has caused millions of dollars in damage to houses and other buildings, utilities, and highways that were built on or across the faults (fig. 7). The historical fault activity results from renewed movement along preexisting faults and appears to be caused principally by withdrawal of ground water for municipal, industrial, and agricultural uses in the Houston area. Approximately one-half of the area's water supply is obtained from local ground water. Monitoring by the U.S. Geological Survey of heights of fault scarps indicates that many of the scarps have recently stopped increasing in height. The area where faulting has ceased coincides with the area where ground-water pumping was cut back in the mid-1970's to slow the damage caused by land subsidence

along Galveston Bay and the Houston Ship Channel. Thus, it appears that efforts to halt land subsidence in the coastal area have provided the additional benefit of arresting damaging surface faulting.

The Houston area is the largest metropolitan area in the United States dependent chiefly on local ground water for its municipal and industrial water supply. Not surprisingly, water levels in the freshwaterbearing sediments beneath the area have declined significantly since the beginning of the 20th century. These declines, in turn, have caused the water-bearing sediments to compact and the land surface to subside or sink over an area of more than 4,700 square miles. Maximum historical subsidence in the region exceeds 9 feet. The subsidence is particularly acute near Galveston Bay and along the Houston Ship Channel, where more than 31 square miles of low-lying coastal land has been permanently inundated. In addition, as the land sinks, the area susceptible to tidal flooding by tropical storms and hurricanes is increased greatly.

-2.0

EXPLANATION

3.0

.0

Figure 7.-Photograph of

house damaged by a fault in northwest Houston. Fault has deformed garage

and has required patch in

road.

For example, C. W. Kreitler of the Texas Bureau of Economic Geology estimated that Hurricane Carla, which struck the area in 1961, would have flooded at least 21 percent more land, a total of 146 square miles, if it had occurred in 1976. Today, more than 300,000 people would be affected by a similar event in an area where, in 1900, the worst natural disaster in United States' history occurred when a hurricane made landfall near Galveston Bay and killed more than 6,000 people. Even without a major hurricane, a study by L. L. Jones of Texas A & M University revealed property losses and flood damage attributable to subsidence in the coastal area from 1969 to 1974 were $31.7 million per year.

To reduce or stop future subsidence, surface-water supplies have been augmented by tapping local rivers. In addition, the Harris-Galveston Coastal Subsidence District was organized and empowered to regulate ground-water pumping. As a result of these efforts, water levels in the water-bearing sediments beneath the coastal area have partially recovered from their former low levels. Despite a slowdown in pumping, however, water levels continue to decline in the western part of the subsidence area. Changes in water levels since 1977 in both areas are shown in figure 8.

The first scarp probably related to groundwater pumping was recognized in 1936, and the number of active faults slowly increased as pumping and subsidence continued. The Geological Survey began to map the faults in late 1974. Most of the 50 or so faults known prior to that time had been recognized as a result of damage to manmade structures. Many of the faults mapped since 1974, however, are in undeveloped or recently developed areas, where damage has

not yet occurred and could, in principle, be avoided or minimized. Today, more than 160 faults are known, ranging up to 10 miles in length and 3 feet in height. Historical scarp growth has been confirmed on at least 86 of them, and many others are suspected to be active. The fault scarps increase in height at rates that range from 0.2 to 1.1 inches per year, and the average rate of growth is 0.4 inch per year. Because the faults grow gradually and without earthquakes, they generally go unnoticed during construction unless special site investigations are conducted. Thus, houses, utilities, and roads built on a fault are subjected to cumulative structural damage as the slow steady movement continues.

In 1978, the U.S. Geological Survey began to monitor scarp growth of 12 selected faults in the subsidence area. Their measurements revealed that all seven of the monitored faults in the area of water-level recovery (fig. 8) have either completely stopped growing or slowed to rates of less than 0.05 inch per year, rates well below their average during historical time (fig. 9). By contrast, all five faults in the area where water levels have continued to drop have continued to grow at their former annual rates. Although this monitoring program covers only 14 percent of the recognized active faults in the area, the perfect areal correlation between stable or slow-moving faults and the recovery of ground-water levels strongly suggests that continuing efforts to halt subsidence will also halt the growth of historically active faults.

The five faults in the area of falling water levels show annual patterns of movement that further support the link between pumping and faulting. Movement on all five faults varies seasonally; rates of faulting increase as water levels drop rapidly during summer pumping and then decrease as water levels partially recover when pumping declines during the winter. As can be seen in figure 9, fault movement sometimes stops entirely during the winter months.

The response of the Houston faults to water-level changes is similar to the response of monitored surface faults in subsidence areas in Arizona and California. In all of these areas, faulting ceases during periods when pumping declines and water levels recover. Although the fault displacements can never be reversed, it appears that man can at least arrest the subsidence and faulting that he has initiated.