by Oleg KONDRATYEV, Dr. Sc. (Phys.&Math.), All-Russia Scientific Research Institute of Geophysical Methods of Prospecting (VNII Geofizika) of the RF Ministry of Natural Resources

The last year of the 20th century brought to our planet a number of catastrophic underground tremors-in Turkey, Taiwan, and California. Those tremors have again brought to the fore a problem: will the humankind ever be able to predict such disastrous natural phenomena and prepare for them? Some Western scientists think that really effective prognostication of earthquakes is still impossible: although large research teams in many countries are engaged on this problem, no significant progress has been achieved for many years.

Articles in this rubric reflect the opinion of the author. - Ed.

page 45

IF THE PROBLEM IS SET CORRECTLY

I could agree with that assessment but with one reservation-one should adhere to the traditional ways of research. For a more clear formulation of the problem, it is necessary to study the phenomenon of catastrophic earthquakes as a whole. It can be divided into 3 main parts: prediction of the place and time of future events; prompt notification of those events that have already begun; prevention of earthquakes.

It may seem strange but the prospect of solution of the last two technically more labor- consuming stages is more understandable. To make the whole problem more clear, let's start from the third stage and proceed to the first on the basis of the common assumption: seismic tremors are the consequence of accumulation and discharge of colossal strains in the earth's crust conditioned by its tectonic movements.

Speaking of the prevention of earthquakes we have in mind efforts to change their "mode", i.e. sudden rupture of accumulated strains "to exchange" for split multistage discharge, and not efforts to stop or direct differently such movements. Thus, instead of a big catastrophic event, a series of minor tremors take place. Though we can pose the problem differently- figuratively speaking to provoke "premature confinement". Strong vibration of geological medium can serve as a trigger mechanism for the beginning of earthquakes. And for this the energy of an atomic explosion at safe depths (several hundreds of meters) is quite sufficient. In such cases the population can be evacuated in time from the danger zone and as far as it is possible big industrial sites can be secured.

Another problem is notification about an approaching dangerous event. But is this possible and really effective? The nature allows very little time from the beginning of the earthquake in its focus till the appearance of seismic waves on the surface as the velocity of their propagation amounts to several kilometers per second. Foci of strong tremors are usually located at the depth of 10-20 km, the radius of their destructive action from the epicenter does not exceed several tens of kilometers. So, from the beginning of an earthquake till its catastrophic consequences only a few seconds elapse. At first sight, this is a very short period of time, which does not allow for any measures. However, this is not exactly so. To prove this point of view, let's examine an analogous case. Up-to-date comfortable cars are equipped with air bags that save people during car accidents. The distance and time of propagation of a shock wave are too small. But automation helps prevent destructive consequences of the accident.

Besides, a strong earthquake does not start instantaneously, but it is preceded by a certain stage of development. And if its very beginning is detected, a few minutes and sometimes even hours are available for a timely notification.

What must be done to implement such a preventive system in cases of earthquakes? First of all, a transmitter, which detects the beginning of tremors, must be located as near to its focus as possible. In other words, it is necessary to place a "string" of seismographs into wells (made in each foci zone) at depths of 5-7 km. Connected to them are automatic systems which in case of a strong earthquake will disconnect electrical lines and stop dangerously explosive industrial processes, stop gas supply and cut off pipelines, and will also switch off reactors at atomic power plants, etc. A few seconds, let alone minutes, left for such measures are quite sufficient for performing the above-mentioned operations.

Such "alarm" systems for the most dangerous industrial sites, like atomic power plants, are already being worked out and are even used in some places. However, I must point out: they can be useful only if installed in all big cities situated in seismodangerous zones. No doubt, at first sight it may seem that we shall be dealing with very expensive equipment. But according to calculations, their cost is considerably lower in comparison with expenses incurred by cities and industrial sites as a result of natural calamities, let alone human losses.

It is obvious that both the notification and prevention of earthquakes are possible only given the exact definition of the foci and when specialists are continuously following the changes in the strained condition of rocks at that place. Only in this case it is possible (before the beginning of a quake) to achieve "fighting readiness" of safety automatic systems, evacuate people from big buildings,

page 46

temporarily close schools, and shift industrial enterprises to a special regime of work, etc.

Thus, the determination of the place and time of a dangerous seismic event plays a decisive role in the problem we are dealing with. Besides, the place of the epicenter (its projection on the earth's surface) must be known with an accuracy of 1 km and the time of the event- within 24 hours before its start. In any other case doing away with destructive consequences of strong earthquakes would be impossible.

At present by what we call long-and medium-term prognostication we understand the determination of the zone and force of probable seismic tremors with the indication of approximate time of their beginning. In fact, this is the task of what we call seismic zoning of the territory. For its solution probability assessments are widely used, while they in principle are inapplicable when determining the place and time of probable tremors. The latter task is called short-term prognostication. However, there are no exact requirements for it in relation to the space and time frameworks of predicted events without which it is impossible to attain the final

page 47

goal-to take timely measures for reducing the consequences of a catastrophe.

SEARCH FOR PRECURSORS LEAD NOWHERE

Earthquake prognostication first came to the fore in the USSR 50 years ago, after the 1948 Ashkhabad earthquake (*). The very first scientific program on this subject was compiled by one of the founders of prospecting geophysics, Director of the Geophysical Institute of the USSR Academy, Academician G. Gamburtsev (1903-1955). It dealt with the ways of development of physical bases and the formation of new (geophysical) methods of studies of the earth crust-in-depth seismic probing, correlative method of study of earthquakes, vertical seismic profiling in wells, azimuthal observations with the aim of polarizing analysis of seismic waves, etc. Unfortunately the aforesaid scientist died early, and shortly after the seismoprospecting studies (including the prognostication of earthquakes on their basis) at the Institute practically stopped.

The overwhelming tendency in the sphere of prognostication became the search for precursors of earthquakes according to the following scheme. An uninterrupted observation of series of changes in some parameter of geophysical field are carried out, curves of its changes in time are drawn and on the same time scale are indicated moments of comparatively big earthquakes which took place in this and any other region. It is considered that the forerunner of the latter can be this or other peculiarity on the chart of the corresponding parameter, repeated before several strong events. Along with geophysical fields, studies concern deformation of the earth's surface, emanation of inert gases, variations in the level of ground waters and other characteristics.

In the course of such search for precursors there regularly arose first promising and later quite unjustified methods. A number of pseudo-scientific speculations came to the foreground. So, at the meeting of the collegium of the USSR Ministry of Geology which took place soon after the 1988 Spitak disaster and was dedicated to the problems of prognosis, a report was made on the "successful discovery" of the precursors of the given earthquake according to the changes in the chemical composition of sea water... in the Crimea. Many such "discoveries" are characterized by the fact that the authors, out to substantiate the alleged validity of their assumptions, cite only facts corresponding to their theory, saying nothing or ignoring the contradictory ones. Moreover, usually absent are any kind of hypothesis or even a diagram of physical connection of a measured parameter with a predicted phenomenon.

It is, no doubt, impossible to deny that the search for quake precursors can by chance lead to the discovery of some new links between heterogeneous phenomena. However, we must remember that any of them is formed under the impact of a multitude of different factors and among them changes in the strained condition of the depths of the earth's crust may appear practically insignificant. So, for each of such indirect features it is first necessary to detect and exclude the impact of more significant factors, and then to substantiate the existence of some physical link of the given "forerunner" with the strained condition of the medium at the depths of several tens of kilometers.

Thus, we can conclude: search for earthquake precursors is fruitless. First, they are not universal for various seismodangerous zones with the heterogeneous geological structure. Second, tectonic processes themselves in different geological conditions proceed differently and will develop differently in indirect features. It is obvious: for different regions it is necessary to work out specific methods of earthquake prognostication. In general, examples of successful earthquake prognostication with the help of precursors are very rare and do not go beyond the probability limits of simple conjecture.

Let's discuss from this angle some prognostication methods substantiated on the basis of indirect features. Some years ago a hydrodynamic platform was set up in Northern Caucasia to predict earthquakes according to changes in the level of ground waters in a number of wells on a rather big territory. It was assumed that the changes in the strained condition of the geological medium in this seismodangerous zone embrace not only great depths but also the upper part of the earth's crust. However, the latter is at the same time but in a greater degree influenced by seasonal variations of atmospheric precipitation, technical sampling of water and other factors that are more significant than the strained condition of the medium. If the subsurface conditions are evident here, then depending on the character and place of seats of earthquakes we can expect stress-strain of rocks at the surface part of the cross-section and, consequently, also different behavior of the level of ground waters. As far as such forerunners as the emanation of inert gases from depths are concerned, it strongly depends on the porosity of rocks in different places and can react differently to the changes in the strained condition of the medium. The luminosity of upper layers of the atmosphere, which was observed from time to time before a strong earthquake, is not necessarily linked with them.

Along with the above-mentioned factors one more and rather significant shortcoming of prognostication methods according to precursors must be pointed out. For example, after the breakup of the USSR, the number of observations of seismodangerous regions in our country was considerably reduced and at

* See: A. Nikonov, "And There Was No Warning", Science in Russia, No. 6, 1998.-Ed.

page 48

present Russia has practically one prognostication platform near Caucasian Mineral Waters. According to the specialists working there, they have the experience of studying some characteristic factors arising before medium-force earthquakes in Daghestan and other neighboring regions. Let's assume that these factors are stable. Then under their declared "sensitivity" to underground tremors on rather big territory (for strong earthquakes it must be expanded), the significance of prognostication based on such factors consists in the cancellation of false predictions, and it no longer allows to sound an alarm and take prophylactic measures throughout the Caucasian region. The prognostication itself, though based on geophysical but still indirect factors, is hardly reliable. This proves, in particular, the following fact: according to observations on the prognostication platform in Turkey, local specialists as well as German, American and Japanese scientists have recently assessed the probability of big earthquakes in this region for the nearest 5 years as nearly 13 percent. But it was exactly there that in 1999 two strongest seismic quakes took place (*).

So, to my mind, we can conclude: solution of the problem of earthquake prognostication based on precursors can lead to a blind alley.

WAY TO SUCCESS-PROSPECTING GEOPHYSICS

As seismic tremors are directly linked with the accumulation and discharge of significant strains in the earth's crust, the only direct way of prognostication and control over this process is a search for ways and methods of remote measurement of strains. The latter necessarily lead to a definite transformation and change of physico-mechanical parameters of the medium in the seat zone, which form a geophysical response of the medium to its impact on it, first of all during excitation of artificial seismic and electromagnetic fields. Thus, we have every reason to conclude: the problem of earthquake prognostication can be solved only by methods of prospecting geophysics.

Such efforts, connected with the realization of Academician G. Gam-burtsev's ideas, were also made earlier. So, one of his students, V. Myachkin, Dr. Sc. (Phys.&Math.), in the 1970s carried out X-raying of rift zones on Kamchatka by the method of refracting waves. He hoped to fix changes of the velocity of their spread on accumulating

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

page 49

page 50

strains there. However, the observations turned out to be insufficiently informative for prognosis. The cause of failure consisted in the fact that when using the given method the seismic ray penetrates into the earth's crust very deeply only if the distance from the explosive source of vibrations is approximately 100 km. In this case the larger part of the path of waves along the ray goes beyond the X-rayed seat zone and the seismic characteristics of refracting waves are influenced by changes of the properties of the medium not only in the seat zone but also on the whole path of waves.

Another set of methods, worked out later at the Garmsky platform (Tajikistan) under the guidance of I. Nersesov, Dr. Sc. (Phys.&Math.), was not quite successful either-it used a change of the ratios of the propagation of linear and cross-sectional waves on the basis of recording the earthquakes by seismologi-cal stations. But in this case the seismic waves not always go through a concrete potential seat zone the observers are interested in.

To my mind, the solution of the problem lies in the application of one of the methods of prospecting geophysics-remote control of changes in the strained condition of the medium by means of reflected seismic waves. They can probe the seat zone proper and, what is more important, their dynamic (spectral) characteristics are mainly defined by physico- mechanical parameters of the medium in places of reflection, i.e. in points of a potential tremor. Today this method allows to study the earth's crust in all its depth-about 40-50 km. But seats of the strongest earthquakes are usually observed at depths of 10-20 km.

Let me give two examples to prove my point. The first one-a seismic cross-section obtained through special processing of seismograms in the area of Kola super-deep well. It gave a rather reliable picture of the structure of the earth's crust in upper 12 km. This is the location of the Pecheng series of stratified crystallic rocks which in the northern direction of the profile appear on the surface. This is what was reflected on the given seismic cross- section, which proves the authenticity of outlined depth reflections.

The second example is a result of processing of seismograms to a depth of 40 km on the 70 km long plot of the regional profile crossing the Urals. Evident here are also big zones of the earth's crust with increased and decreased reflective properties. According to the already mentioned physical factors, the seat zones of earthquakes will be reflected in the form of "bright spots"; their color changes depending on the strained condition of rocks. So, the problem consists in dividing the color scale in the parameters of the medium strain and marking a critical point of the beginning of catastrophic destruction of rocks on it.

Thus, the following techniques of prognostication observations are evident. First, regional profile studies of the earth's crust must be carried out in dangerous areas with the help of super-deep method of reflected waves to attract all existing geologo-geophysical and seismo-logical data and to discover potential seat zones (a number of repeated profile observations will be needed). Then, the monitoring of area probing must be carried out at potential seat zones with a frequency of observations that is necessary for tracking the process of strain accumulation. Near big cities or industrial sites (atomic, hydrostations, etc.), besides observations on the surface, the monitoring includes registration of seismic waves in deep wells in order to shift observations into the regime and pass a signal of notification about the beginning of earthquakes.

It is obvious that this method will require comprehensive studies, first of all those connected with the assessment of principal possibilities of remote tracking of the changing state of the medium in great depths and marking the above-mentioned scale of strain. The stumbling point in this matter is a choice of the right place and time for tests. It's no secret that measurements can be carried out at any place for years without any results as it is unknown whether the state of the medium there changes at all. However, the nature has shown us a way out: if strains are being accumulated for tens of years at unknown places then the reverse process-their discharge after the first destructive tremor-takes place at an exactly defined seat zone and in a shorter period of time. Nevertheless, there is enough time to organize and carry out necessary series of observations, if, of course, there are specially prepared geophysical expeditions studying abrupt reaction to such extraordinary situations. Having studied the peculiarities of changes of geophysical fields at various stages of discharge of the strained state of the medium, we can predict what will happen there in future and then select and apply the most effective complex of measurement methods and interpretation of the obtained data. It is also necessary to complement the seismological reconnaissance with other geophysical methods fit for determination of the make-up and strained state of the medium.

The proposed new paradigm of solving the problem of earthquake prognostication seem sufficiently grounded and promising. Unlike the concept of the search for forerunners, it, first, proves the following principle: complex problems have no simple solutions. And prognostication of the development of tectonic processes in the earth's crust is a very complicated problem. Second, it is suggested to solve physical problems through direct measurements and studies of the physical process itself and not some indirect manifestations.