by Viktor MORGUNOV, Dr. Sc. (Phys. & Math.), Joint Institute of Earth Physics named after O. Schmidt, RAS

When can we expect the next earthquake to strike? Questions of this kind are not mere curiosity for the residents of tremor-prone regions, to say nothing of expert seismologists. The problem of earthquake prognostication has been a matter of priority for geophysicists for more than a century. Its practical importance continues to grow with the growth of megalopolises and what are called ecologically hazardous technologies.

HOPES AND REALITY

According to time scale, experts distinguish three types of quake forecasts: long-term (tens and hundreds of years), medium (several months) and short-term (days and hours).

In ancient times people learned to identify the signs, or harbingers, of an approaching "doom", such as changing water levels in wells, new springs gushing up, and others running out of water, strange behavior of domestic animals and night-time glow of the sky... Such signs and phenomena cannot be simply shrugged off as legends or freaks of imagination and this is especially so in the light of modern instrumental data. And the list of the usual "omens" today includes computer failures, radio and TV interferences, and/or upsets in power grids such as those caused by major magnetic storms. This being so, we still cannot fully rely upon such omens and forebodings, which means that in our modern day and age earthquakes mostly remain as unpredictable as centuries ago.

On the more optimistic side, however, experts have been able to cope with the task of what they call seismic zoning * , which simply means long-term prognostication. Detailed maps with "diagrams of probability" of tremors of different strength are an indispensable "foundation" of all major building projects now. But having this "safety warranty" alone is not enough for the prevention of often tragic aftermaths of tremors. This underlines the importance of short-term prognostication.

Now, let's take a look into a likely scenario. 24 hours before an expected tremor all schools are closed, factories and public transport are brought to a halt, city dwellers are moved to safety, gas and oil pipelines as well as electricity networks and other potentially hazardous enterprises are "deactivated". Then strikes the quake, as expected, and although the material damage can be appreciable, there are practically no casualties among the population. And the economic losses from fires and environmental pollution are minimal (one can recall at this point that the main losses, such as those from the Tokyo disaster of September 1, 1923, were caused not by crushing buildings but by a giant fire; and in San Francisco in 1906 most of the damage was also caused by fires).

And now let's get back to the present. Earthquakes, often accompanied by tsunami, are surpassed in the number of casualties only by flood disasters. And the situation is aggravated by increased density of the population and concentration of industries in quake-prone regions which, as a rule, boast attractive climate. And something like the Spitak tragedy (1988) in Armenia can add up to the likelihood of the Chernobyl (1986) disaster in Ukraine in different spots of the planet because tectonic shifts happen even in some relatively geologically stable areas (platforms) of the earth crust. This being so, experts have every reason to say that there are no "tremor-proof regions, but only slow-slipping faults.

And while we can rely on satellite data and networks of ground weather stations for weather forecasts, there are no matching conveniences for earthquake prognostication. And this is not just a matter of expenses.

At the present time Japan holds the pride of place in the world by the volume of i nvestments and intensity of prognostication research. But still and all there is no denying the fact practically that none of the forecasts made over the past 30 years within the framework of the appropriate programs has come true.

Bearing this in mind, one could come to the seemingly unavoidable conclusion about the impossibility of solving this problem. This would be in line with the concept which has gained popularity over the past decade and according to which the earth crust is in what experts call the state of self-organized criticality (SOC). This has no characteristic dimensions and, consequently, denies any reliable assessments of the place, time and magnitude of coming seismic events. Views of this kind are shared by prominent scientists in the United States, Western Europe and Russia. And whether SOC is a reflection of a physical process, or a mere generalization of the insufficient elucidation of the problem, is something that can only be established by a physically substantiated experiment.

* See: V. Ulomov, "Seismic Menace in Russia", Science in Russia, No. 6, 2001. - Ed .

Pages. 5

Burst of acoustic emission registered by the Bakuriani seismic station 16 hours before the main quake and in the active phase of the after-shock activity of the focus of the Spitak earthquake of December 7, 1988. The arrow marks the moment of the event.

FORECAST COMES TRUE!

The above outline of our "cloudless" future looks fantastic even for the 21st century. But the paradox consists in the fact that this has been a quotation from a report about one short-term forecast of a major quake.

On February 4, 1975 the authorities in the North Korean town of Khaichen declared a state of emergency for 48 hours. All factories were closed and hundreds of thousands of residents were evacuated to a field camp. And one can only guess at the responsibility assumed by the local authorities who issued the order. And at 19 h 26 min of the same day the city was struck by an earthquake of strength of 7.4 points. Practically all residential districts were turned into rubble, same as hospitals and schools. Also damaged were factories, dams and other engineering facilities. But the human toll was minimal.

That was the world's first really tangible achievement of seismologists ever recorded. The following year, however, Mother Nature took revenge with the devastating (M=8.0) Tan Shan quake which took a toll of 700,000 to 1 mln lives. And before the end of the 20th century there were several more unexpected calamities of this kind in the Northern hemisphere.

So, what can we really expect: earthquakes can be forecast in theory, but in practice such timely predictions is sheer luck? Let's take a closer look at the aforesaid examples.

The success of the Khaichen prognostication was determined by two factors: a succession of portents, or precursors, and the unhesitating actions by the authorities who heeded scientists' recommendations. But with the Tan Shan disaster the picture was quite different. There, too, there were timely signs of the approaching doom, but they were not so clear as in the first example although the ultimate blow was 7 times stronger. And experts did issue timely warnings-which failed to convince the authorities. It could be that this time around the scientists were not persistent enough, or had some doubts themselves.

Summing it up, the success of a forecast depends on three key factors, or components: scientific methods and the degree of their acceptance, or recognition, the availability of a special seismic prognostication service, and the existence of adequate legal norms. Incidentally, in Khaichen the inadequate observation network was "made up for" by residents' reports about unusual natural phenomena, such as of numbers of snakes escaping from their holes at their unusual time of winter hibernation. These signs and warnings were enough for the authorities to issue their unprecedented orders of evacuation of the people and factories.

Looking back at the Tan Shan quake, experts discovered a truly extraordinary situation which had been registered by both civil and military services: for three days before the disaster radio communications had been practically paralyzed by strong static interferences within a radius of 250 km from the epicenter of the subsequent quake. But that warning was somehow ignored.

One of the main causes of the tragedies in Spitak, on the Russian Island of Shikotan (1994) and in

Pages. 6

Clear anomaly of electromagnetic emission intensity as registered by the Inubo Observatory (near Tokyo) 4.5 hours before the Shikotan earthquake of October 4, 1994.

Neftegorsk on the Sakhalin (1995) was the absence of well-equipped stations. But tragedies of that kind also happen in technically advanced countries: the devastating earthquakes in Los Angeles (1989) and in Kobe, Japan (1995) with the death-toll of more than 6 thousand and 27,000 injured. In all of them the residents, and scientists, were taken by surprise. Some signs of the approaching doom were registered there, but only several observation stations were functioning in these areas and in what specialists call passive regime. Until recently in the United States, and also in the European Geophysical Union, the prevailing opinion was that prognostication studies were unpromising. Most specialists even regarded as "bad taste" producing publications on that subject. That attitude was produced mostly by a stream of superficial reports and hypotheses based on haphazard factual evidence. It was only recently that things began to change for the better.

CONCEPT OF SELF-ORGANIZED CRITICALITY

The skeptical attitude to all sorts of portents and precursors observed in the US Geological Service stems from the lack of precise measurements registering deformation of the earth crust and its inclinations at geophysical testing sites. But is that really the case?

Before the quake at Lander (California) on June 22, 1992 of magnitude of 7.3 points, experts measured the level of water in wells. In three of them before the event it dropped from 0.15 to 0.3 m and in the fourth one, located only 100 to 250 m away it dropped down by 5 m. The difference of more than 10 times at a distance of several hundred meters only! This fact could offer tangible proof of the complex and mosaic geological structure and of the field of stresses as supporting the opinion of the Japanese geophysicist Keyo Mogi that "signs of the impending peril" have a tendency to manifest themselves in some "chosen" points of tectonic structures. According to some specialists such contradictory results cannot be regarded as factual material.

A few more examples. The 1989 quake near Los Angeles struck as a surprise. Although, according to a Reuter's report, two days before a local short-wave fan published a warning in the local press about a probable disaster. But the warning remained unheeded.

As people were going to work in Spitak early in the morning on December 7, 1988 they watched with surprise rats escaping in numbers from a local storage depot (there were similar reports from Japan before the Kobe disaster). And at the same time instruments, developed at the Institute of Earth Physics of the USSR Academy and mounted at the Bakuriani seismic station located 80 km from Spitak, registered clear signals of the approaching quake. For a period of 16 hours before the first strike recorders of the station started "tracing" the upcoming tragedy. This being so, why did the station director failed to "ring the bell"? To begin with, there were week-long breaks in the local telephone and teletype networks. Second, to raise an alarm he obviously had to obtain the formally required permissions which were, unfortunately, quite a problem in the Soviet Union in those years. There were no formally accepted warning procedures for the population

Pages. 7

Anomaly of electromagnetic emission intensity at 81 kHz as registered by the Sugaida observatory (Japan) one hour before two successive strikes in 1980.

then, and the situation has not improved in any significant way since.

ELECTROMAGNETIC PRECURSORS

The seismic shock that hit the Shikotan Island on October 4, 1994 was so strong (M=8.1) that this biggest "member" of the Kuriles sunk by 0.6 m. A geological catastrophe of this magnitude, even from the angle of common sense, could not have happened all by surprise. It is true, though, that geophysical measurements had been stopped some time before for lack of funds. But specialists at the Inubo radiophysical observatory (near Tokyo) had been conducting measurements of electromagnetic ground emissions for a whole year using the methods of the Russian Institute of Earth Physics named after Schmidt. The Japanese registered an anomalous signal four and a half hours before the event and regarded it as a clear warning.

These Japanese observations continued studies of electromagnetic precursors of earthquakes started in 1980 under a joint Soviet-Japanese project. And the very first measurements on Japanese territory registered clear signals in the electromagnetic field before two successive strikes. Later on this method proved its effectiveness also in Greece, France, Italy, Bulgaria and other countries.

As for Shikotan, reports on radio communications and interrogations of members of ground and marine crews revealed that short-wave anomalies had been observed across the epicenter region 7 to 10 days before the earthquake. Directly before the event one could clearly see glow of the lower atmosphere, and according to a navigator from a steamer located some hundred of kilometers away "it looked like aurora polaris emanating from the Earth". Naturally enough, observations of this kind- important as they may be-will remain a net of casual facts if there is no proper scientific interpretation thereof.

EFFECT OF ROCK CREEP

It takes tens and even hundreds of years for an earthquake to "mature". The foci of future tectonic shoves, or shifts, are produced by stresses in the earth crust and upper mantle. The geological timescale determines the progress or rate, of the rock creep.

These processes have been well studied for building materials in strength-of- structures calculations. As experts point out, material begins to "flow", or creep under prolonged stress in practically any solid body, including rock, which is clearly visible on geological structures where layers look like manifold folds. Rock creep of this kind has three stages: established, unestablished and what is known as flowing avalanche. During the first the rate of deformation is reduced with time and the material is consolidated. The second-which is the longest-takes place in the earth crust at an almost constant rate over tens and hundreds of years. The third process is of a short duration and this one is linked with the appearance of short-time (days-hours) precursors. During that time the probability of authentic forecasts is the greatest. Incidentally, a similar process can be observed in a lab: under a long and permanent strain a sample develops a "neck". But if the same force continues to be applied to a diminishing cross-section, the strain grows and the rate of destruction is increased.

In rock layers this, what we call avalanche creep, progresses mainly due to the formation of cracks up and until the final rupture, or break.

The scale of deformations and their rate diminish with distance from the minimal "neck" cross-section while the characteristic "signs" of the future fault, or rupture can be preserved. With the approaching "hour" of the earthquake these defects become more apparent and can be identified as short-time precursors.

But it is not always that the stage of avalanche creep ends in the destruction of the medium. Redistribution of stress can slow down the process or stop it. This accounts for the "variety" of precursors, including "false alarms" (of which there are but few).

Laboratory experiments confirm the appearance of specific "signs" of destruction as manifested in changing electromagnetic and acoustic emissions and the electric field and currents. Observations in nature prove that these parameters can be best controlled during "rapid" deformations. The Spitak catastrophe, for one, had been proceeded by an outburst of sonic signals in the range of 1 - 10 kHz. But because of strong damping in rock they could not reach out beyond 80 km from the hypocenter and up to the seismic station in Bakuriani. And that

Pages. 8

Diagram of creep curves of a deformed solid body under permanent load P: 1-relative deformation; 2-rate of relative deformation; S-sample cross-section at "neck"; t-time. I-unsettled stage, II-established stage, III-accelerated, avalanche creep.

Model of earthquake emergence after the precursor anomaly. Solid line- deformation. Dotted line-rate of deformation. Arrow indicates the moment of quake when deformation reaches the limit of rock strength. Crossed area-anomaly of electromagnetic emission intensity. On the vertical-number of pulses per hour; on the horizontal-time (in days).

Pages. 9

means that the sources of this emission (fractures) were in a direct proximity of the sensor which registered it. And that means that on the eve of the Spitak tragedy intensive rock erosion was in progress in the area of the local seismic station.

It should be noted that the phenomenon of acoustic emission provides the basis for the methods of defectoscopy and safety methods in mines. But the registration of sonic precursors of earthquakes calls for some specific conditions of instrument installation and is hampered by all sorts of natural and technogenic "noises".

More effective, therefore, is a method of identification of electromagnetic precursors. Suffice it to recall the strange radio interferences and noises accompanying the Tan Shan, Los Angeles and Shikotan earthquakes.

Radio interferences which accompanied them can be explained by electromagnetic emissions generated at the stage of non-elastic deformation of upper rock layers; the same accounts for the strange behavior of animals. There are known laboratory experiments with animals in electromagnetic fields proving their definite response to its alterations.

The first to study such phenomena was Soviet physicist Dr. Alexander Stepanov. In 1933 he discovered an effect, later named after him, showing that deformation- electric processes are observed not only in quartz crystals (piezo-effect), but in any solid body submitted to plastic deformations. Later on these studies were continued in this country and the results obtained proved that this effect not only exists, but can be used for practical purposes in flaw-detection of structural materials and for studies of major deformations in rock (monitoring of tectonic shifts, short-term earthquake prognostication, mapping and prognostication of landslides and avalanches).

The problems of localization of sources of seismo-electromagnetic emissions and mechanisms of their generation in nature call for more comprehensive studies. These processes are more intense in the focal zones, but their registration on the surface is obstructed by the rapid damping of emissions during their propagation through conducting rock. Such radio emissions are of homogeneous intensity and does not clearly depend on the depth, earthquake strength and distance from the epicenter. And that means that its sources should be looked after near the earth surface.

Pages. 10

SHORT-TERM FORECASTS ARE POSSIBLE

On April 5, 1997 Chinese experts issued a warning about the possibility of a major earthquake. On the night before the D-day authorities in the Sinjan autonomous region evacuated more than 150 thous. residents. The two strikes that followed (M=6.4 and 6.3) ruined 2,000 homes and damaged another 1,500. And there were human casualties.

The two successful Chinese earthquake warnings - on February 4, 1975 and on April 3, 1997-over a period of 20 years can be regarded as lucky coincidences. But, as they say-there is light at the end of the tunnel. A method has been developed which makes it possible to forecast with great certainly seismic threats one or two days, or at least several hours, in advance. Years of studies conducted by our Institute experts in different areas of seismic activity of this and some other countries have proved the practical value of this method as formulated in the ALARM-SEISMO complex which has won diplomas and gold medals at international shows in Brussels (1999), Moscow (2000) and Geneva (2001). It is ment for short-term prognostication of quakes and can also be used for monitoring of landslides and avalanches and tectonic shifts. After its introduction in the Northern Caucasus a decade ago the KAVKAZGEOL-SYEMKA Agency of the RF Ministry of Natural Resources has been able to predict a number of local earthquakes. Using one or several such complexes it is possible to identify periods of increased seismic danger. This helps solve the problems of public warnings, prevention of hazardous situations at atomic power stations and other ecologically unsafe industrial units. But with all of its advantages, this method does not guarantee absolute accuracy of short-term predictions. Comprehensive solution of this problem calls for considerable expenses, the establishment of a network of observation ground stations, the use of space technologies and for having new centers of data collection and processing on a real time-scale.

The well-known "regularities" of outbreaks of seismic activity manifesting themselves in cascades of strong tremors occurring at short intervals of time in regions located far apart or on different continents, accord with our notions of the lithosphere as one common structure. One recent example was a series of destructive quakes which occurred in 1999 on an area from Taiwan to Turkey and Greece (the impact was the most catastrophic in Turkey for more than a century with a death toll of over 18 thous, 30 - 35 thous people under ruins, 23 thous wounded and 500 thous people left homeless; the material damage was estimated at 25 - 40 bin dollars). Some experts, such as Dr. Abxei Nikonov of our own Institute, are confident that the catastrophe could have been anticipated if more attention was given to an analysis of the tragic past of that region. * That being so, there is really no alternative to the establishment of a common worldwide sendee of earthquake prognostication. This geophysical problem is something quite worthy of the new century!

* See: A. Nikonov, "Catastrophe in Turkey: A Surprise?, Science in Russia, No. 1, 2000. - Ed.

Permanent link to this publication: