EARTHQUAKE PREDICTION: AN ALTERNATIVE OF EXPLORING CRUSTAL BEHAVIOR BY MANMADE SEISMICITY.
----- A draft released for discussions on Reservoir Induced Seismicity
Opinions and comments are appreciated and please address them to the author: fax: 403-250-6711 or E-mail: retom@cadvision.com
"Earthquake prediction in many ways still resembles a black art raised to the highest order of scientific method"
----- Allen W. Hatheway (Rolla, Missouri) in Engineering Geology (#43, 1996) by Lomnitz, 1994.
"The universe, so far as known to us, is so constituted, that whatever is true in any one case, is true in all cases of a certain description; the only difficulty is, to find what description."
----- John Stuart Mill, A System of Logic
RETROSPECT:
ATTENTION TO INDUCED SEISMICITY MIGHT PROVIDE SEISMOLOGISTS THE INTEGRATION OF KNOWING AND DOING.
Is the fate of earthquake prediction similar to the proof of Pierre de Fermat's [1601-65] last theorem? The theorem he presented was without proof, although he claimed to have discovered one [and lost on a side of an unclaimed envelope] was widely accepted and attempted proofs were recognized addition to the theory of numbers. Why generations of mathematicians struggled with? ... it was a peak to be conquered or a necessity to master another peak for the continuation of logical mathematical developments. Human approach requires that peaks must be conquered, even of no use other than "to conquer them."
Similar too, the "dream of earthquake prediction" is an unavoidable burden of the modern age seismologists, living in a world where everybody along seeing the weather watch often chooses taking his umbrella based on the acceptable inaccurate last meteorological forecast.
More than 350 years ago, Pierre de Fermat wrote down his apparent simple "theorem" (FLT) in the margins of a mathematical book, he was reading, adding that he discovered a remarkable proof for it but that there was no room to include the proof in the margins too. He died without ever presenting the step-by-step logic to prove this tantalizing claim, confounding the best of mathematicians ever since trying to do so. Like many a quiet English mathematician, Andrew Wiles, was drawn into the Fermat's puzzle and after years an extraordinary idea linked these irritating problems with the Taniyama-Shimura Conjuncture, named after a young Japanese who tragically committed suicide. The link meant that if Taniyama was true then so must be FLT. Mathematics hit the front pages of the world's press. The answer is people tried and they never found a way of counting, and this was why this is the key breakthrough that Andrew found a way to count not the original problem, but a modified one.
Just as a similarity a seismic behavior of rather different origin, man's activity, is proposed as an alternate way of investigating the seismic prediction issue and better understanding the earth crust bearing and response. The difference is dealing with a man-made phenomenon versus a natural one.
The triggering of induced seismicity may be an intermediate step of development of improved understanding of the earth's crust and then structuring better means for predicting seismic behavior of the crust.
Recent seismological reviews recognized the difficulties now considered as impossible, road blocks on prediction of natural earthquakes, which in press languages is highlithed as "Earthquakes Cannot Be Predicted"(1) or "Long odds on prediction"(2). The media reflection is the briefing of the specialized symposium: "Assessment of schemes for earthquake prediction" [ASEP,1996] organized by the Royal Astronomical Society and the Joint Association for Geophysics, held in November 1996, in the Geological Society, Burlington House, London.
Several types of engineering developments are suspected, with varying amounts of evidence, of triggering the occurrence on a limited localized scale and referred as induced-seismicity. The strength of induced seismic (IS) activity range from moderately sized, damaging, earthquakes with the range of failure of the order of kilometers, through to rock bursts, and down to micro-seismic emissions, which are mostly sensed by instrumentation and have range of failure on a scale of meters. Historically, the development of large reservoirs for power generation was the first type of engineering activity that significantly affected the earth's crust and seismicity is associated with such stress modifications. The largest event recorded was 6.5 on Richter scale and associated with the impoundment of very large reservoirs. The actual (3) incidence of Reservoir Induced Seismicity (RIS) is associated to about hundred twenty reservoirs of different sizes. Other induced seismic activities are related with mining ( MIS - 5.7 in potash), fluid injection and disposal ( Denver Arsenal 5.2), fluid withdrawal (Strachan, Alberta - 4.3) and geothermal (Goldsmith, TX, 3.5), enhanced oil recovery (3.0) and storage of oil and gas in the underground caverns, mostly associated to microseismic activity generated by the gas pressure variations.
Reservoir Induced Seismicity (RIS) used to be considered as a strange unintentional environmental phenomenon, which was generally detected without instruments. It was observed in relation to new impoundments and presented a problem to the professional and scientific community for many years. Today, RIS is recognized mostly as an environmental and dam safety-related concern, which is largely understood but for which there are limited means of prediction on a consistent basis. The first engineering operation to be aborted on environmental grounds, was associated with seismic activities at a series of deep wells for disposal of contaminated fluids at the Rocky Mountain Arsenal in Denver Colorado USA, in 1966. Not all seismic activity was proved to be connected with the fluid disposal, but awareness and social sensitivity brought the operation to an end because of environmental concern associated to increased social sensitivity rather than real threats. RIS is of particular concern in areas of low natural seismicity, where potential larger induced events might be higher than the natural activity.
In depth references on RIS are well known ( 4 ) and updated along scientific exchanges, conferences' ie: ISORIS'95 Beijing China (5), IUGG 1994 (6 ) , NATO's ARW, Moscow 1994 (7) and most recent ICOLD Congress 1997 (8).
RIS impoundments were classified initially in the first two major groups (9),and some other specifics refeered recently ( 8)indicate some other particularities:
But the "opportunity" to predict the man-made seismicity is limited and its running out in time, and along with the loss of a pre-eminent contingency to study the response of the crust under controlled load condition, generated by the impoundment processes. This might be a unique window of opportunity to observe the crust loading which is happening anyway, but may be dull monitoring data collection. The impoundment process involves the largest human being crustal interferences, with an extreme large amount, from insignificant to billions of cubic meters of water or their equivalent loads [billions of tonne].
Generally the 120 RIS cases (Most of them listed) are referred to the known number of very large reservoirs (VLR- depth larger than 100 meters) and associated with the construction of about 300 very large dams (VLD), but the seismic occurrence is predominant on smaller reservoirs than the storage with average sizes of 116 meter depth and volume of about 25 cubic kilometers. As an accepted "rule of thumb" for seismic risk, critical reservoirs will be equal or larger depth of 100 meters or a minimum volume of one cubic kilometer. Prediction of RIS or IS is an exception, without the benefit of any statistical support. A seismic risk evaluation procedure was developed by a comparation of reservoirs around the globe with only few applications, and is applied only incidentally. Risk mitigation might achieved by more complex undertaking like the application of Risk Prediction Model (RPM) endorsed at the discussions for Q.60 (13) at the 16th Congress in San Francisco and supported by four confirmed impoundments with RIS associated with hydro developments(12) in Quebec (Manic3/5 ,LG2-LG4) and more recently 1996 at Katze Dam(Lesotho).
The issue of RIS prediction has several aspects, and is a potential nuisance for any large dam development through the social concerns. At this stage the issues of RIS in practical terms continue to be still a neglected issue. Examples could be unaware residents around Katze reservoir who had their houses affected, when no mitigation of risk was done for a site predicted to be at risk [Water Power 1993, issue May, page 33(12)] along with others ie: Ertan in China.
"Truly, the nation was shocked to learn of the Mapeleng tremor while to the Mapeleng people the incident was a real nightmare." (14). In the surrounding area of Katze dam provision of temporary housing for seriously damaged and repairs compensate for the 54 dwellings affected. The rehabilitation program includes replacement for affected springs, a voluntary relocation program and assistance for increasing earthquake resistance of buildings and an income restoration program.
Most practitioners will report for new impoundments that monitoring procedures are in place. Monitoring is a good contrameasure for continued or increased seismicity but lack the prevention component make the cost of monitoring questionable or to be used as an experimental procedure by others, ie: seismologists with interest or desire.
Pessimism now surrounds the field of earthquake prediction was considered as excessive by some authors in the past(15). Of forty proposed precursors considered recently by the International Association of Seismology and Physics of the Earth's Interior subcommission on earthquake prediction, only five were judged as significant, and of those five, two were of hydrogeochemical in nature and most probably important to reservoir seismicity as well.
There has been little progress in over 25 years of research into earthquake prediction (15), where we use "prediction" in the classical sense of specifying the magnitude, time and place (how, when, and where). RIS prediction does not detail time sequences or location, mostly controlled by the location and type of faults. and "how" is based by broad understanding not on quantitative elements implied in a reliable prediction manner. Despite all this demonstrable unpredictability, there is one phenomenon that can be reliably monitored. A large earthquake cannot take place without there being sufficient deformation to yield the energy released by the earthquake. Although details of the time and place of the earthquake itself may be unpredictable, most recent studies confirm that the buildup of deformation before a large earthquake can be monitored. RIS studies include these elements on the known or chosen side of the equations. Monitoring the buildup of deformation,which often is called earthquake forecasting, are a much simpler and better understood phenomenon than earthquake prediction. Forecasting is used in the sense of recognizing the buildup of stress before earthquakes, analogous to the use of air pressure to forecast the weather. Monitoring the long-term buildup of stress before earthquakes depend on appropriate swarms of small earthquakes to provide the source signals within the shear-wave window of digital high-sampling-rate three-component recording stations. It is rare to have such swarm events within the shear-wave windows of seismic stations near to the location of impending earthquakes. Consequently, routine monitoring would require controlled-source experiments and monitoring seismic risk prone reservoirs might be a realistic alternatives.
Three types of earthquake prediction models, deterministic, specific and models that assess the magnitude of events are usually considered by the general public and most of geophysicists to be `real' earthquake predictions.
The author wish to recognize that some aspects of the paper might be controversial and sometimes too optimistic but researchers are able to identify if the goal is not biased as defined by Richter well-known comment (1964) on some earthquake prediction attempts: "Claims to predict usually come from cranks, publicity seekers, or people who pretend to foresee the future in general." This is still the consensus of the scientific community today. Richter comment is mostly reminded associated to the facts that often some seismologists neglect the potential contribution of induced seismicity as a tool for actively study the crust's response to human activities or to consider as an environmental issue.
The window of opportunity to use the RIS soon will come to be closed as fewer important dams and reservoirs will be constructed because of the physical limitation of available proper sites. Sustainabel development for the exploding population of the world include supplementary important future water supply demands which include an increase of 20 percent addition of water storages for the next 25 years (1500 Kmc by about 130 large reservoirs, to the existing 7500 Kmc., provided by about 300 large dams).
Using the past occurrence of RIS the statistical risk estimate suggests a probability of induced seismicity hazard at about fifty new impoundments. The natural limits of growth must be an inherent consideration to actively try to use this reservoirs and associated potential seismicity, before this unique window of opportunity to study the effects on the crust's behavior,will close, relatively soon.
Despite years of work by scientists around the world, no generally accepted prediction method has ever been developed. Many claims to have developed successful prediction methods have later proved groundless, leading to general skepticism by the scientific community. Charles F. Richter, said in 1964 that "Claims to predict usually come from cranks, publicity seekers, or people who pretend to foresee the future in general." This is still the consensus of the scientific community today.
The author's hope that involving more seismologists in studying IS, both the public as the seismology may be closer to achieve some realistic steps to achieve some elements of earthquake prediction in the future. Acknowledge must be made that more seismology involvement is hindered by practical social support, and luck of any funding for RIS and IS, is a common ground generated by critical level of funding for earthquake prediction activities. .
RETROSPECT:
ATTENTION TO INDUCED SEISMICITY MIGHT PROVIDE SEISMOLOGISTS THE INTEGRATION OF KNOWING AND DOING.