The distinction between earthquakes and underground explosions. Earthquake control

Earthquakes are associated with large fractures or faults in the Earth's outer surface. One of the most widely accepted explanations of shallow earthquakes is Reid's (1911) elastic-rebound theory, which attributes earthquakes to the accumulation of strain energy in tectonic regions and the sudden release of this energy by faulting when the fracture strength is exceeded.

In great earthquakes the offset across the fault may be as much as 15 m. Once the frictional bond is broken the elastic strain energy, which has accumulated slowly for tens or hundreds of years, is suddenly released and the resulting rupture will travel at a speed of about 3.5 km per second, continuing in some cases for as much as 1000 km. The effect of underground nuclear explosions is quite different, consisting of outwardly radiating P waves.

An important application of seismology is the detection of such explosions, and their discrimination from natural earthquakes. An earthquake, being associated with a mechanical couple (a pair of forces acting in parallel but opposite directions), should produce alternate phases of compression and dilation in four quadrants related to the fault plane.

Seismographs in two quadrants will therefore record a compression as the first motion detected, and those in the other two will record an initial dilatation. Almost all earthquakes give this quadrant distribution of compressions and dilatations from which the direction of the fault plane can be discovered. An explosion, on the other hand, in which energy is radiating outwards from a single point, should cause compression everywhere, so that all seismograph stations record an initial compression.

Consequently it was initially hoped that seismology would be able to distinguish between nuclear explosions and earthquakes, but uncertainty arises when there are not stations close to an earthquake's epicentre. Since there are many parts of the world, such as ocean bottoms and isolated land masses, in which it is impossible to place stations, earthquakes occasionally give initial compressions at all available stations.

Earthquake control. The concept of fracture followed by frictional sliding across a fault face is plausible only for shallow shocks. At depths below a few kilometres the overburden pressure exceeds the shear strength of the rock, and dry frictional sliding is impossible. If a pore fluid is present under pressure the effective normal stress across the fault is reduced. This has the effect of lubricating the fault, allowing sliding to occur at greater depths. In 1966 a series of small earthquakes in Denver, Colorado, was attributed by some seismologists to the injection of fluid into a waste disposal well.

This interpretation was questioned on the grounds that the area was seismically active in any case and that earthquakes continued to occur after injection ceased. The correlation between the amount of water injected and the number of earthquakes was. however, quite strong, and this chance discover}' was followed up in 1969 by a planned field experiment by the United States Geological Survey at the Rangely Oil Field in western Colorado. A large number of oil wells were available at the site, and water could be regularly injected into the wells or pumped out of them.

The results of the experiments showed an excellent correlation between the quantity of fluid injected and local earthquake activity. When the pore fluid pressure reached a critical level earthquake activity increased. When the pressure dropped as a result of water withdrawal seismic activity decreased. Again it should be pointed out that the wells at Rangely penetrated pre-existing faults and that the crust in the region was already under some tectonic strain, as indicated by the occurrence of small local earthquakes over the previous years.

The Denver and Rangely investigations suggest the possibility of earthquake control. One suggestion is to pump water through deep boreholes into faults in a region where natural earthquakes might be particularly hazardous. An outbreak of small earthquakes might thus be induced, reducing the amount of strain energy stored in the crust in the vicinity and reducing the probability of a large earthquake.

Tampering with the forces of nature in this way is, however, dangerous. If control was attempted along a major active fault the consequences might be especially damaging. On the other hand ‘de-straining’ crustal rocks by water injection at the site of a large structure, such as a dam. might be feasible.