Two different types of human activity -- the creation of reservoirs and the pumping of fluids in deep wells -- have been shown to induce a change in seismicity rates. The process behind each is probably the same, and it involves the effects of water pressure on rocks at depth.
Civil engineers have considered a link between reservoirs and seismicity for over 100 years, but the first real evidence that the phenomenon of increased seismicity relating to water storage might be a genuine effect came in 1936, after the creation of Lake Mead, held back by Hoover Dam. Since then, numerous other cases of such increases in seismicity rates following the creation of a new reservoir have been reported.
Because earthquakes don't seem to be induced by reservoirs built in seismically inactive areas, it appears that reservoirs don't create earthquakes, they simply "encourage" the slipping of stressed faults in an area. Supporting this is the fact that the seismicity rate near a new reservoir will typically peak within the first decade after filling, and then lower and stabilize at a rate more closely resembling pre-reservoir levels. But how does this "encouragement" of fault slip happen?
Some of the water from the reservoir seeps into the rocks below, helped by the pressure of the water column above. As this water permeates into the local faults, it may begin to act as a kind of lubricant, both by weakening the rocks in a fault zone, and also by exerting pressure on the walls of a fault, which could act to separate the cracks ever so slightly, reducing the force needed to grind the walls of the fault past each other. When this begins to happen in faulted rocks under stress, they will become temporarily more active, and then, once the excess stress has been released, the fault slip will decrease in frequency and reach a new equilibrium.
Support for this idea came in 1962 near Denver, Colorado, when an army project designed to dispose of waste water began pumping the liquid into a deep borehole, at high pressure. Over the next year, the seismicity rate in the area near this borehole began to skyrocket! Residents in the area worried about the correlation between this disposal method and the increase in seismicity, and the project was stopped. In 1969, the U.S. Geological Survey (USGS) conducted an experiment to investigate the possible correlations between injecting water into deep wells and changes in seismicity rate. They found that, indeed, the seismicity rate of an area could be increased by changing the water content and pressure in faulted rocks at depth.
It has been suggested that a network of boreholes and pumping stations could be set up along an active fault zone to effectively control the slip along that fault, and help it release the increasing stress in small, non-damaging earthquakes instead of in larger, more destructive ones. Toying with the forces of nature like this obviously presents a risk (suppose you triggered a very large earthquake instead?), but the more fundamental problem with this idea is that it takes a lot of small earthquakes to equal the energy release of a larger rupture. How many? Let's calculate:
The largest non-destructive earthquake you would probably want to induce would be about magnitude 4.5. To release the equivalent amount of strain energy released by a single M 5.5 earthquake, you would need about 31 quakes of M 4.5. To equal the energy of a M 6.5 earthquake (about the size of the 1994 Northridge earthquake) would require 1000 M 4.5 tremors. And to match the energy of a M 7.5 earthquake, the kind produced by the rupture of the Mojave segment (only) of the San Andreas fault, it would take over 31,000 earthquakes of M 4.5! Assuming that this sort of rupture normally occurs every 140 years, we would have to induce over 200 M 4.5 earthquakes every year (4 each week) along its entire length of 150 kilometers just to keep this one section of the San Andreas fault zone "safe" -- free from the threat of a major, destructive rupture! While not impossible, the "control" of faults would certainly be quite an undertaking, both in terms of engineering and public policy.
Back to the reality of the present, there is a growing awareness outside of scientific circles of the link between reservoirs and seismicity rate increases. In southern California, seismologists have been keeping close watch on the seismicity rate in the area near the developing Eastside Reservoir project, in order to have a well-known baseline against which future rates can be compared.
A good source for more detailed information on this issue is Chapter 9 of Earthquakes ("Stimulation of Earthquakes by Water") by Bruce A. Bolt. In this chapter, the author provides several different examples of changes in seismicity apparently induced by newly filled reservoirs.Return to the main text