While paling in comparison to natural agents in the Earth System, human activities have induced earthquakes, and there is potential for compound catastrophes which have involved multi-billion dollar societal decisions. We'll wrap up the course by considering a few of the issues associated with triggered earthquakes and social dilemmas.
The construction of artificial dams can produce dramatic changes in water budget in a given region, with massive artificial reservoirs being created and sustained in regions where natural geological processes would have drained the landscape. Is there any dangerous potential from such activities? We have already considered the Vaiont Dam disaster in which the creation of high water levels destabilized the adjacent hillside, leading to landslide that caused a huge wave that broached the dam. In that case the water lubricated joint surfaces in the bedding planes of the rock formations, leading to failure. There have been many other cases in which actual earthquakes were triggered by the impounding of large reservoirs.
The first time this was noticed was in 1936, when earthquake activity in the previously quiescent region flooded by the Hoover Dam project were detected. Since that time at least 13 cases have been well documented with seismicity being clearly related to reservoir loading. Examples include:
It is always somewhat difficult to establish the causality between reservoir impounding and earthquake activity, and there may have been many more instances of earthquake triggering, but if this occurs in a region where there were frequent earthquakes prior to impounding of the reservoir, it may not be clear what is causing subsequent events. One way that the causal relationship can be established is by looking for systematic seasonal variations in small earthquake activity in relationship to natural cycles in lake level resulting from seasonal rainfall fluctuations. For example, tracking the monsoon cycles affecting the rainfall-derived input into Shivajisagar Lake, near Koyna, suggests that earthquakes tend to fluctuate significantly, with a weak tendency for higher levels some time after peak increases in water levels. The 1967 event was within a few months after the lake levels rose to their highest levels of the year, but those levels were not significantly higher than in previous years.
The observed fluctuations do not reveal a simple causality in most cases, and this leaves room for much debate over cause and effect. This is not a trivial issue, as their are very legitimate issues of liability that arise when a damaging earthquake occurs, possibly the result of a commercial or governmental modification of the regional hydrological budget. A large earthquake in India several years ago occurred in a region where prior seismicity was not recorded and where a large reservoir was impounded in the 1980s. There was massive destruction, with over 20,000 people killed. Was the water control the cause?
Closer to home, the same issue arose in 1975, associated with the Oroville earthquake in central California. This earthquake ruptured a fault dipping down into the ground a few kilometers south of Lake Oroville, created by a dam. The August 1975 mainshock had a magnitude of 5.7, and caused damage in the nearby town of Oroville. Unlike some earthquakes near reservoirs, the fault that ruptured was clearly remote from the lake, by at least a few kilometers. Could the earthquake have been a response to the lake formation? Could it have been anticipated? These issues are still debated, as the entire seismically active region of the western U.S. is speckled with large and small reservoirs.
What is happening? Is the added weight of the water in the lake causing an increase in crustal stresses that triggers earthquakes? This is probably not the case, as the associated stress perturbations can be calculated and appear to be implausibly small, relative to the stresses associated with other effects such as erosion of the terrain, topography, and ambient tectonic stresses. However, the role of water may be greatly enhanced if the water penetrates down into the ground and reduces the effective confining pressure on faults at depth. In this case, the reduction of the stresses normal (perpendicular to the fault) reduces the friction resisting sliding motion, and the fault may become unstable. It is believed that it is the underground role of water that plays the major part in inducing earthquakes, with changes in groundwater level producing variations in hydrostatic head which may affect friction on preexisting faults.
Much of our understanding of this process is actually the result of direct changes in groundwater content due to processes other than lake impounding (which is always a very complex process and suffers from lack of control behavior). In fact, the first time it was established that groundwater variations were directly controlling earthquakes was in 1962, and it came as a shock. This was the surprise occurrence of earthquakes near the Rocky Mountain Arsenal in Colorado (between Boulder and Golden). In April 1962 a swarm of earthquakes commenced in what had been a previously aseismic area, and this swarm persisted for over a years, with 700 recorded events with magnitudes from 0.7 to 4.3. It turns out that the Army was pumping contaminated waste water down a deep well, under high pressure, injecting the fluids into the ground. A clear relationship was established (albeit begrudgingly) between injection activity and earthquake frequency. This was clearly a major concern, not so much because of fear of triggering a very big earthquake (although that possibility could not be ruled out; in many places the crustal stresses may be quite high due to locked in deformation that can be released), but mainly because the hope was that the deep aquifer into which the waste was being injected would contain the waste in isolation from shallower aquifers used by regional agriculture and water wells. The occurrence of earthquakes is of course a direct indicator that faults are being activated at depth and that deformations are occurring which could potentially disrupt any one aquifer, allowing the contaminated fluids to leak out into other aquifers. In fact, decades later, there is extensive contamination of deep aquifers in the northwest Denver area, and there is massive litigation associated with the underground contamination and resulting health hazard.
Further testing of the influence of water injection and water extraction on earthquake activity was pursued for years, including experiments at the Rangely Oil Field in Colorado, and it became well established that fluid injection was a factor in enhanced earthquake activity, although the behavior is very complex. A minor earthquake in northern Ohio (magnitude 4.5) occurred in close proximity to a waste water injection facility, possibly a triggered event and possibly not. The main interest in that event was that the earthquake took place very close to a non-operating nuclear power plant, which had many structural failures. While there probably would not have been a major disaster even had the plant been running, it brought much attention to the concern about water injection triggering earthquakes that could in turn induce significant secondary hazards.
Other human activities have triggered significant earthquakes, including explosion of underground nuclear weapons. This is a common effect in weapons test sites, and is of concern for two reasons: 1) the faulting may disrupt the containment of the nuclear debris, allowing venting of radioactive material; 2) the earthquake signal may obscure the fact that a bomb was detonated, so that treaty monitoring is difficult. At the Nevada Test Site, one of the largest earthquakes to be triggered followed the explosion of the Benham device, which had a yield close to a megaton (1 million tons of TNT; the Hiroshima bomb was only 14,000 tons). The triggered faulting was complex, but surface breaks spanned about 10 miles in the surface of Pahute Mesa where the explosion was detonated and many small aftershocks occurred. The seismic radiation from the event is still identifiable as an explosion plus and earthquake, but it showed how difficult the problem can be. Two effects appear to be responsible for this so-called "tectonic release" which often accompanies large underground explosions. One is simply the driving stress from the explosion destabilizing the friction on a pre-existing fault and allowing it to fail. The second factor is that a large explosion like Benham produces a large cavity surrounded by a fractured medium in which any pre-existing tectonic stress is relaxed due to the loss of rock coherence. This relaxation emits seismic waves that can look like faulting even if it is really a volume of rock that is relaxing without a crack in it.
The recognition that underground explosions could be used to relax stress near the explosion has led to various ideas. In both the U.S. and former Soviet Union tests were conducted to serve 'Peaceful purposes', such as excavating large canals. An example at NTS was the 1962 shallow explosion called Sedan, which left a large crater in the surface as material collapsed into the cavity created by the explosion. The notion that one could perhaps judiciously set off explosions to regulate the stress on faults has received both scientific and popular consideration (did you see the Superman and James Bonds movies where crackpots were going to set off explosions along the San Andreas fault to produce huge earthquakes?). Well injections have also been considered. The idea would be to release strain accumulation early in the cycle of stick-slip motions of a fault, effectively having many small events occur rather than one large devastating one. Of course, this is not a safe bet, given that the crustal stress and fault distribution is so complex that one cannot predict safely what event will be triggered, and in fact small events may run away to become large events. Imagine the law suits then!
Another major societal issue associated with earthquakes and human activities has been citing of nuclear power plants. This is now a global issue, as many developing nations are looking to nuclear power as an energy source for the future, and earthquake hazards must be considered. The case history for California is rather telling. If we look at where actual and proposed nuclear power plants have been located in the state, there are 8 sites along the coast, which immediately bring to mind the approximate location of the plate boundary. The sites hug the coast because one of the primary needs of nuclear plants is ample water for the cooling systems, and the ocean is a ready source. Also the sites naturally tend to cluster near major metropolitan areas of San Francisco and Los Angeles because of the power needs of those communities. Of course the plate boundary also hugs the coast.
In the planning for many of these sites, the earthquake safety issue came to the fore, both in assessing the hazards posed by known active faults throughout the region, and in the potential for unrecognized faults to exist close by to a given plant. This is a very complex issue, as there are legitimate concerns about providing sufficient power for the burgeoning population of the state, but also concerns about the potential for an earthquake induced catastrophic failure of the plants. Many billions of dollars have gone into planning, building, operating and shutting down these plants, with only 3 surviving to provide power today. Geologists were brought in on both sides of the issue, some supporting the power company perspectives and some supporting the anti-nuclear power constituency. While the issue often has devolved to differences of professional opinion, perhaps biased by personal feelings about the nuclear power issue, one truism did develop out of the many site preparation and construction efforts. This is that just about anywhere that you dig into the crust in California you will find a fault in the rocks. The problem then becomes one of telling whether that fault has any potential to fail under present day tectonics motions, or whether it is an old scar that will not move in the foreseeable future.
One site of interest was the Bodega Head location, just north of Point Reyes and San Francisco. This area was considered for a nuclear power plant that was intended to service the large metropolitan area of San Francisco. The area is close to the San Andreas Fault, which ruptured in the 1906 earthquake, and that was a recognized fault of concern. However, the design for the plant was deemed sufficient to withstand shaking from a repeat of the 1906 earthquake with no catastrophic failure. This is not surprising, as the earthquake safety issue has driven remarkable design of nuclear power plants in this country, which has greatly contributed to the cost of the plant and the delay in breaking even on the up-front investment in the facility. Having fought many a battle to permit the plant to be developed at Bodega Head, excavation of the reactor site, which was to be largely contained underground, initiated. PG&E dug down into the ground and sure enough, to the dismay of their geological consultants, they found a fault right beneath the proposed reactor site. The geologists established that the fault had been active in the recent geological epoch, but they could not unambiguously establish the potential for future failure of the minor fault. While the engineering was sufficient to provide safety for shaking from a nearby San Andreas event, it taxes even the best engineering designs to withstand a possible fault dislocation directly below a building. Some isolation ideas, involving a space to accommodate fault movement around the reactor were put on the table, but ultimately, the uncertainty, projected increased costs, and renewed litigation brought the whole initiative to a halt. No plant was ever finished. Comparable stories exist for many of the other sites in California. Total financial losses in exploring the issue have been billions of dollars, passed on to you via PG&E bills.
Why has nuclear power plant siting received such intensive scrutiny and attention to possible geological catastrophe? There has never been a major geologically induced nuclear power plant failure to point to. It is the emotional intensity of the issue that has played a role in this prioritization, and similar concerns are playing out with respect to radioactive waste disposal issues. The concern about radioactivity is hard to argue with, but it is somewhat out of balance with respect to the relatively casual consideration of geological hazards associated with much more common and catastrophic phenomena. Part of the challenge for society is defining how to prioritize geological hazard issues given the infrequent nature of such events and there present unpredictability. There are always competing concerns associated with societal needs (power, water resources, disposal of wastes, etc.) and hazards raised by those needs, as we have been reviewing. The role for most of the students taking this class will be one of a lifetime of participating by voting, political action, education, and direct involvement in decision making about the balance of needs and hazards. Every election puts before us issues such as should we establish bonds to fund earthquake retrofitting of bridges to reduce freeway collapses for future earthquakes, or should be invest in prisons instead since the probability is that the earthquake won't happen before we have to vote again?
Let's close the discussion with an example of a human-induced environmental disaster with which we are all familiar. The event of interest was the March 24, 1989 rupture of the tanker Exxon Valdez, which ran aground on Bligh Reef in the inland passage. Ten million gallons of oil spilled, and the slick spread over 700 km long, contaminating 5300 km of shoreline in a previously beautiful natural environment. There was great devastation of water fowl and shallow marine life forms in the region, but what struck me most close to the heart was the travails of the Sea Otters along the coastline. The clean-up effort yielded 878 Sea Otter carcasses, with presumably many more lost in the sea. Some 357 sea otters were captured and efforts were made to save them. 123 died from the pollution and the trauma of capture. 37 were damaged and will live out their lives in aquaria. 197 were cleaned-up, deemed viable and released. Thus, only 18% of those that we know of were saved by human efforts, and the total labor involved in this one tiny part of the huge clean-up response was $18.3 million. That comes out to $80,000 per otter. Was this a good use of money, given all the competing concerns and human situations existing today? Even for those of us that love sea otters, the answer involves a dilemma or competing priorities akin to that which we face throughout our lives. Earth catastrophes, both natural and of human activity will accompany our collective journey through time, and by understanding and thinking deeply about the processes and issues involved, perhaps we can all find our way.
![[return icon]](../dinoheadt.jpg)
Return to the Earth Sciences 80A Lectures Page