A Qualquer Custo (Hell or High Water) 2016 | Planet With | Player 1: StreamTo Add: Depuis 1 jours

Long-term prediction for seismic hazard for radioactive waste disposal

Long-term prediction for seismic hazard for radioactive waste disposal

Available online at www.sciencedirect.com ScienceDirect Russian Geology and Geophysics 56 (2015) 1074–1082 www.elsevier.com/locate/rgg Long-term pre...

610KB Sizes 2 Downloads 15 Views

Available online at www.sciencedirect.com

ScienceDirect Russian Geology and Geophysics 56 (2015) 1074–1082 www.elsevier.com/locate/rgg

Long-term prediction for seismic hazard for radioactive waste disposal B.T. Kochkin *, V.A. Petrov Institute of Geology of Ore Deposits, Petrography, Mineralogy and Geochemistry, Russian Academy of Sciences, Staromonetnyi per. 35, Moscow, 119017, Russia Received 16 December 2013; accepted 8 December 2014

Abstract We consider possible approaches to the long-term prediction for seismic hazard in relation to the practical need for the safety of geological disposal of long-lived radioactive waste. The required period of prediction significantly exceeds the one reflected in the set of maps of General Seismic Zoning of the territory of the Russian Federation (GSZ-97). The first geological repository in Russia is planned to be set up in the Nizhnii Kan granite massif in the Krasnoyarsk Krai. This region is an intraplate territory with a relatively high seismic activity. We summarize the analysis of the known empirical generalizations and theoretical principles underlying the seismic hazard prediction. Real seismic events constantly violate forward-looking statements even for relatively short periods of time. These and other arguments suggest that the hypothesis of stationarity of the seismic regime, which is the basis of long-term prediction today, has limited and uncertain applicability in time. Intraplate earthquake prediction is especially uncertain because of the uncertainty in the factor responsible for generating tectonic stresses in these regions. The short horizon of the prediction, based on statistical methods, can be attributed to the nonlinearity of seismic geodynamic processes. Fundamental laws of tectonic processes should be used as the scientific basis for long-term predictions for seismic hazard at the sites chosen for geological disposal of long-lived radioactive waste. These processes can be reflected in models for the migration of the seismically active boundaries of lithospheric plates and the occurrence of seismic activity in intraplate regions. © 2015, V.S. Sobolev IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reserved. Keywords: seismic hazard; long-term prediction; radioactive waste; geological disposal; safety

Introduction Earthquake prediction and the development of seismic hazard maps is a problem that has not yet been solved. According to the period of time for which an earthquake prediction should be made, it is common to distinguish between long-term, medium-term, and short-term predictions (Sobolev, 1993). Each of them is based on its own specific theoretical propositions and formalized rules of processing observation data. According to experts’ opinion, long-term predictions are more reliable, medium-term predictions are less reliable, and short-term predictions are even less reliable (Mogi, 1985; Sobolev, 1993). The longer the prediction period, the more it can be considered as a seismic hazard prediction for the given region as a whole, rather than a prediction for a specific event. Here an analogy with climate and weather forecasts is appropriate. Climate is more reliably predicted for longer

* Corresponding author. E-mail address: [email protected] (B.T. Kochkin)

periods, although with a loss of specificity of the time and place of occurrence of separate events. Seismic hazard maps are generated for the purposes of long-term prediction. Seismic zoning maps of the former Soviet Union compiled previously were in one way or another inadequate to the real environmental conditions. Considerable progress in detailing and refining the seismic hazard of the country is observed in the last GSZ-97 set of maps. However, it also involves a number of problems related to the reliability of predictions for long periods of time (Morozov et al., 2001; Seismotectonics..., 2009). Obviously, existing methods of earthquake prediction are not fully suitable for long periods of time, although the need for such predictions exists and is associated, in particular, with the regulatory requirements for safety analysis of radioactive waste (RW) repositories containing long-lived radionuclides. In particular, the Russian regulations, which generally follow the internationally accepted guidelines, require ensuring reliable operation of geological repositories and protecting the population over the period of the potential danger of isolated radionuclides (Radioactive..., 2004). For high-level waste and spent nuclear fuel (HLW and SNF) and other types of RW

1068-7971/$ - see front matter D 2015, V.S. So bolev IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.rgg.201 + 5.06.008

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

containing long-lived radionuclides, the period of potential hazard is millions of years. A prediction for especially dangerous objects, such as nuclear power plants and radioactive repositories, with a reasonable probability of not exceeding a given earthquake intensity is calculated for a period of 10,000 years and is based on (GSZ-97D map). However, it is recommended that the stability of the geological environment that will ensure the reliability of the disposal system over the period of the potential hazard of long-lived RW should be evaluated for the whole of this period (Considering..., 2009; Kochkin, 2013; Methods..., 2012). Hence the need for longterm predictions of seismic hazard for periods of hundreds of thousands and millions of years and there is a question about the tools for predicting seismotectonic activity for such long periods. These predictions should be considered to be superlong-term even in geological terms, and considering tools for their development is the main purpose of the paper.

1075

Dangerous consequences of seismic fault activation with magnitudes maximal for the south of the Siberian region (M > 8) are manifested mainly on the surface and near it as secondary faulting, slope processes, and other phenomena at distances up to tens of kilometers or more. At a distance from a seismogenic source, the frequency of dangerous geological processes decreases exponentially (Lunina et al., 2014). It is known that at a depth of hundreds of meters in rocks, the dangerous consequences of earthquakes (seismic intensity) are much weaker. This somewhat reduces the potential danger of high seismic activity for geological repositories. Under the current regulations, sites in regions with possible seismic events of more than 9 points or with signs of active fault are considered unsuitable. Regions in which the seismicity is characterized by a maximum calculated earthquake intensity over 7 points are unfavorable (Radioactive..., 2004). Thus, seismogeodynamic processes at the site of the future repository in Krasnoyarsk Kari is to be studied in detail to assess its long-term safety.

Safety of geological disposal of long-lived RW The geological disposal of RW is a technology designed to ensure the reliable protection of the population and the biosphere from radioactive contamination (Falck and Nilsson, 2009). It is planned to dispose waste in special underground facilities (repositories) fitted with a multibarrier system of protection of the environment from radionuclides. First, this is a matrix comprising radionuclides. It is packaged in a metal can, which provides physical disposal of radionuclides from groundwater in the initial stages of isolation. Clay fillings of the chambers and tunnels are used for absorption slowing of the leakage of radionuclides due to the inevitable corrosion of the cans. The last barrier in this system is the geological environment. It is to minimize the dispersion of radionuclides for as long as they are a hazard. Over millions of years, different events and processes will occur in the repository and its surrounding, and the rate and duration of these processes can vary within wide limits. In particular, the dangerous processes that can affect the stability of the isolating geological environment and deform the engineering barriers include the geodynamic processes occurring not only along the active margins of tectonic plates but also in intraplate blocks (Stein, 2007). Obviously, it is best to place geological repositories in regions characterized by low seismicity and having no active faults. However, if such a region is turned out to be seismically hazardous, as is the case with the site for the Russian repository in Krasnoyarsk Krai (Lobanov et al., 2011), the location of active faults should be clearly established, and their seismotectonic activity should be predicted for the entire potentially dangerous period. The seismic hazard of the future disposal site in Krasnoyarsk Krai is estimated to be 8 points (GSZ-97D map) or according to the refined zonation of probable earthquake centers (PEC) for the territory of the Krasnoyarsk agglomeration developed in accordance with the methodology for constructing medium-scale GSZ maps is 7 points (Sibgatulin et al., 2004).

Empirical foundations and theory of long-term prediction for seismic hazard The seismicity of a region us usually assessed based on three main criteria: seismic activity, earthquake recurrence, and the maximum possible earthquake magnitude (Mmax) (Bune and Gorshkov, 1980; Ulomov and Shumilina, 1999). Seismic activity is a stochastic variable. It is unstable both in time and in space. Using the results of analysis of seismic activity in time and space, Fedotov (1968) introduced the concept of so-called “seismic gaps.” He has shown that the regions of the sources of catastrophic earthquakes that occurred during the observation period occupy a significant part of a particular seismic zone but do not cover it completely and do not overlap one another. He assumed that the regions where strong earthquakes had not been observed for a long time are possible places of future major earthquakes. Sooner or later, strong earthquakes are believed to occur repeatedly in the same places. Fedotov proposed the term seismic cycle to refer to the course of the seismic regime at the same point of a seismogenic fault in the time interval between two earthquakes of maximum intensity. Thus, in the Kuril-Kamchatka zone, the average return period of catastrophic earthquakes M > 73/4 is approximately 140 years. In less active regions, it is many hundreds and thousands of years. Instrumental monitoring for any region is performed only in the last 100 years or less, so that there are no data about any full seismic cycle. Even in the Kuril-Kamchatka zone, continuous instrumental monitoring of the seismic activity of individual sections during complete cycles will probably end only in about 2044 (1904 + 140 years). Fedotov believed that the seismic cycle is a common feature of seismic processes. Its typical characteristic is the long period of stabilization of the regime, which lasts for 3/4 of the cycle or more. During this period, the seismic activity fluctuates around

1076

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

Fig. 1. Stress–time (τ) and displacement–time (υ) graphs for faults localized in the seismogenic zone of the upper crust, according to (Scholz, 1990). Stages of the seismic cycle: α, interseismic or seismic calm; β, preseismic (foreshocks), γ, coseismic (mainshock); δ, postseismic (aftershocks); ∆τ, stress relief.

a constant level and is characterized by only relatively weak earthquakes. Fedotov’s concepts of on the seismic cycle and its stagewise nature have been widely recognized and used in the development of long-term earthquake predictions. Thus, according to Scholz (1990), each seismic cycle consists of four stages: preseismic, coseismic, postseismic, and interseismic (or “seismic calm”) (Fig. 1). In the preseismic stage, there is a rapid accumulation of stresses associated with the nonlinear deformation of rocks and seismic pulses (foreshocks) preceding the earthquake. In the coseismic stage, the deformation of rocks develops in an avalanche manner immediately after the relief of the accumulated stress (earthquake). In the postseismic stage, rocks deform nonlinearly for some time, which is accompanied by a number of tremors (aftershocks), which may occur along the seismically active fault at a considerable distance from the earthquake source. In the interseismic stage of seismic calm, the deformation of rock massifs is described as linear. In most earthquake-prone regions, strong earthquakes recur very rarely. The return periods of earthquakes on active faults outside seismically active zones are even longer. Special studies have shown that the return periods of strong earthquakes on the ancient platforms of continental territories are 10–100 thousand years and more (Crone et al., 1997). In the same region, earthquakes of lower intensity than the established maximum seismic activity recur at short intervals. Thus, in a seismic zone with an intensity of, e.g., 9 points and a recurrence of 9-point earthquakes of, on average, 1200 year, 8 or 7-point earthquakes repeat on average in 230 and 75 years, respectively (Medvedev, 1968a). The recurrence of earthquakes has a logarithmic dependence of the number of earthquakes on their magnitude (M). For large territories and observation times, the slope of the corresponding graph was considered constant, which has until recently been used to calculate Mmax in a particular local area (Bune and Gorshkov, 1980). It turned out that these graphs are not rectilinear. Starting at M > 6.5 for all regions, graphs of the annual average rate of event flow indicate a higher

recurrence of such earthquakes than it should have been from the traditional linear extrapolation of the graphs. The real frequency of major earthquakes is three or more times higher than previously thought (Ulomov and Shumilina, 1999). Until now the magnitude Mmax for one or another local area is the most controversial seismicity parameter, whose determination directly from observations made in a particular area is limited by the extreme rarity of the events close to the maximum possible. To evaluate Mmax, one has to take into account indirect data related to geological and geophysical conditions in the areas of earthquakes. Indirect factors are taken into account through expert judgment or through certain formal calculations. Nevertheless, it is believed that adopting the hypothesis of stationarity of the seismic regime, we can expect that each point of the Earth is characterized by a finite value Mmax, which cannot be infinitely large. Erroneous determination of Mmax in any area may affect the determination of the length of the full seismic cycle. The number of places for which there are instrumental data for a more or less long period is limited, and historical documents cover no more than a few hundred years, or, in the best case, the first thousand years (Bulletin..., 1913; Kondorskaya and Shebalin, 1977). Therefore, topographical, geomorphological, and geological techniques are important tools for the study of past seismic history. Recently, the significance of such fieldwork has become obvious, methods of identifying active faults have been improved, and data on strong earthquakes of the past have been used to estimate the time of earthquake recurrence in a particular area. This and other empirical evidence of the stable period of recurrence of seismic events are used to predict earthquakes and generate seismic hazard maps. Supplementing the tool database with information on historical earthquakes is crucial for solving problematic issues of seismicity of the Siberian region (Nikonov and Fleifel, 2014). The scientific complexity of the problem of predicting seismic hazard is that it belongs to the category of predictions based on incomplete information and fuzzy methodological propositions. The observation base is too limited, and theoretical concepts do not allow one to reliably estimate the characteristics of the seismic regime for large time intervals. In the middle of the last century, there was a replacement of the then-existing paradigm of “seismic actualism” to a new one which involves fist identifying not only real but also potential source zones and only then assessing seismic tremors of the Earth’s surface. The new paradigm became the basis of all subsequent general seismic zoning maps, including the modern GSZ-97 (Ulomov and Shumilina, 1999). The GSZ-97 set of maps for the Russian Federation was developed to make predictions for the construction of facilities of different degrees of criticality and different service lives. This was done using a probabilistic-deterministic approach based on the statistically assessed return period of large earthquakes and the expected life of various types of structures. In the development of these maps, seismicity was treated as the result of deformation of the Earth’s crust and the entire lithosphere taking into account the fractal features of their

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

layered block structure, the interpretation of seismic sources as extended (rather than point) sources of earthquakes, the idea of the nonlinear development of seismogeodynamic processes, etc. (Ulomov and Shumilina, 1999). The last GSZ-97 set of maps includes four seismic hazard maps with indication of the expected maximum seismic intensity for return periods of tremors of 50, 500, 5000, and 10,000 years.1 The last of these, the GSZ-97D map, was compiled later than the others and corresponds to a return period of, on average, 10,000 years for a seismic effect on the ground at average soil sites. For this map, the probability of this intensity being exceeded within 50 years is 0.5%, and within one year, 10–4%. This map is designed for reconnaissance assessment of seismic hazard in the vicinity of extremely critical facilities, such as nuclear power plants, radioactive disposals, and other highly hazardous facilities of the nuclear fuel cycle. The paradigm change, the development of theoretical ideas about the nature of seismicity, and improvements in predicting methods have not prevented the occurrence of new seismic events of unexpected scale in one region or another. For example, after the preparation of the GSZ-78 map, devastating 8–9 and even 9–10 point earthquakes occurred almost every year for a decade in regions of the former USSR for which the hazard on this map was underestimated by at least 2 or 3 points. Among them were earthquakes having catastrophic consequences, such as the Spitak earthquake (1988) in Armenia and the Neftegorsk earthquake (1995) in Sakhalin (Ulomov and Shumilina, 1999). Recent events, such as the Kaliningrad earthquake (2004) have shown that the seismicity of some platform regions must be raised to the level of 6 and 7 points, and they cannot be considered aseismic as is adopted on the GSZ-97 map (Seismotectonics... 2009). The de facto recognition of this trend, nevertheless, usually does not cast doubt on the validity of the hypothesis of the stability of the seismic regime, and ad hoc hypotheses can be invoked to explain the curvature of the recurrence graph, as is done in (Morozov et al., 2001). Indeed, it is quite reasonable that the development of the tectonic process and related deformations are of inherited origin, at least, in the current tectonic stage, which began at about 25 million years ago, and, therefore, stability of the tectonic stress field may be expected in a certain period (maybe of the same length) in the future. Thus, it is commonly believed that a statistical analysis of available data on the seismicity of a region indicates that earthquakes of the maximum intensity possible in this region have a certain return period. The latter is explained by the seismic cycle of accumulation and relief of tectonic stresses under the conditions of stationarity of the seismic regime. It is assumed that the presence of the seismic cycle is consistent (at least, inextricably linked) to the hypothesis of stationarity of the seismic regime in a particular region. Indeed, without this hypothesis, it is impossible to explain the virtuality of the essentially stochastic constant of the duration of the seismic 1

The GSZ-2012 set of maps is currently discussed (http://seismos-u.ifz.ru/ 2012.htm).

1077

cycle and, in general, the presence of periodicity in the occurrence of the largest seismic events. Nevertheless, the established curvature of graphs of earthquake recurrence in the region of extremely strong events, along with the lack of continuous instrumental records of seismicity in individual areas during the full cycle are the empirical basis for doubting the validity of the hypothesis of stationarity of the seismic regime in relation to the distant future. There is no empirical evidence of the stability of the return period during several cycles. The hypothesis of stationarity of the seismic regime in a given region, which is valid for relatively short periods, obviously, has a limited applicability in time. Stationarity of the seismic regime in the distant future is an extrapolation of geodynamic and historical data for the previous periods of time. The uncertainty is how far into the future this extrapolation is valid.

Nature of uncertainties in earthquake prediction The main approach to earthquake prediction is based on stochastic regularities derived from observation data. The examples above show that a new seismic event could easily degrade such predictions and their theoretical foundations. Unfortunately, many scientific uncertainties relating to the safety assessment of long-lived radioactive waste repositories are unavoidable (Kochkin, 2004). Thus, the law of large numbers, which underlies the practical application of the theory of probability to earthquake prediction, is useless for predicting events occurring in periods much greater than the human experience. The strongest earthquakes are just such rare events. The required information can be obtained in principle, but a decision on the prediction cannot wait for these data to be received actually (if ever). In such circumstances, the researcher does not have objective data to substantiate the adequacy of the observations, their correctness and the representativeness of the result for the region described by the model, as well as the correctness of all model assumptions. Moreover, as regards the assessment of the adequacy of the model (true/false) selected for the prediction, there can be no statistics. On the one hand it is obvious that in the latter situation, increasing the observation period (no matter how long it lasted) has no relation to increasing the sample size (number of tests) used in probability theory. On the other hand, a new observation that goes beyond the previously known parameters can radically change the researcher’s view of the suitability of the model (Apostolakis, 1990). The uncertainties related to the latter situation cannot be eliminated at all. The objective uncertainty of a long-term earthquake prediction based on stochastic models necessitates the use of deterministic models of the seismogeodynamic process, but this path also has its obstacles. It is believed that an earthquake is related to disruption of the continuity of the crust material during a relative displacement of individual sections along a more or less extended plane. Rupture occurs under the action of elastic stresses

1078

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

Fig. 2. Postulated (a) and real (b) behavior of rocks during metamorphism, according to (Passchier and Trouw, 2005). P, Pressure; T, temperature; M1 and M2, metamorphic cycles; D1–D3, deformation events.

accumulated during tectonic deformation of this material, and it relieves these stresses, partially or completely, in the rupture plane. Rupture arises in a limited area (hypocenter) and propagates from it at a limited speed. Thus, in earthquakes with M = 7.5, the average length of ruptures reaches 100 km, and in those with M = 8.5, it is 300 km (Bune and Gorshkov, 1980). The 24 May 2013 earthquake under the Sea of Okhotsk with M = 8.3 propagated at a speed of about 4 km/s with a rupture length of ~180 km (Ye et al., 2013). Structural settings that can provide the accumulation of elastic stresses for their subsequent relief are amenable to modeling (Barbot et al., 2012), but this is beyond the scope of this study. The currently adopted mechanism of earthquake occurrence explains the periodicity of these events in one place and, accordingly, the seismic cycle, since the conditions for the occurrence of ruptures are maintained due to the stability of the tectonic stress field in a much larger volume of the Earth’s crust than that occupied by the hypocenter. Nevertheless, it is obvious that, after each rupture, stress accumulation will proceed under conditions where some of the factors of the seismogeodynamic system have changed and the new cycle will have parameters different from those of the previous cycle. For example, the geometry of the seismogenic fault (source zone), the degree of fluid and gas saturation of rocks, their rheology, etc. will first change. This implies that even if the magnitude of the tectonic forces acting in the region and the general characteristics of the stress field remain unchanged, the development of the system can take a different path. The question is, how different is this path and how long these differences will not fundamentally affect the seismic regime and the constant of earthquake recurrence in the region. We can assume that it is these slight differences that are ultimately responsible for the curvature of recurrence graphs in the region of strong earthquakes, which reflect the physical nonlinearity in the system. The nonlinearity of seismogeodynamic processes is considered as an explanation of the basic impossibility of even short-term predictions of specific events (Koronovskii and Naimark, 2012).

The concept of nonlinear systems is based on the idea of I. Prigogine (1960) of the fundamental instability of the processes occurring in complex systems with energy and mass transfer. The general response of such a system is not equal to the sum of responses to individual effects. The concepts of nonlinearity of the processes occurring in rocks were formulated in the works of M.A. Sadovskii in the 1970–1980es (Nikolaev, 1998). He defined the notion of a geophysical environment, following the above-mentioned idea of I. Prigogine. Physical nonlinearity is a fundamental property of rocks. Real rock environments are hierarchically inhomogeneous. The physical properties of rocks change in a relatively short time intervals under the influence of endogenous geodynamic and exogenous processes, including lunar and solar tides and other phenomena of extraterrestrial nature and man-made factors. A fundamental revision of the concepts of the properties of rocks laid the foundations for nonlinear geophysics and, in particular, nonlinear seismology. According to the tenets of classical science, the prediction accuracy is limited only by purely mathematical difficulties. In the last 30–40 years, it has been shown that the trajectories of future states of systems are described not only by a point, as for a mechanical pendulum or a circle as for reversible systems, but also by quite whimsical curves called strange attractors. Formally, the systems they describe are deterministic, but their behavior can be predicted only for a limited time. Here we can cite an example from petrology which is illustrated by the figure (Fig. 2) borrowed from (Passchier and Trouw, 2005). Thus, the actual pressure and temperature distributions in the interior and hence the response of rocks during metamorphism is described by a more complex curve than appears. A single metamorphic cycle can include secondorder cycles. The nonequilibrium of irreversible systems makes possible the occurrence of unique events. At any given time—a bifurcation point—there may be a change in the spatial-temporal organization of the system (Prigogine, 1991). Uncertainty in the knowledge of the initial data or details of the dynamic process sooner or later grows to a huge uncertainty in the prediction, if the current state of the system is extrapolated to long time periods. It has become clear that in

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

many cases where a prediction for the development of a complex system is required, there is a horizon of predictability, beyond which complete uncertainty reigns (Kapitsa et al., 2003). This fully applies to predictions of seismic events. Nevertheless, there is a lot of order in strange attractors, which opens up prospects for a successful prediction. Chaos at the micro level can lead to order at the macro level. Along with the major factors acting in irreversible systems, the processes involving energy dissipation play a key role in the self-organization of these systems. These include heat conduction, friction, and viscosity. Some of these secondary factors affecting the energy dissipation at the earthquake source, for example, water or gas saturation of the seismogenic fault, has long attracted the attention of researchers. An example of the occurrence of order at the macro level in the chaos of a set of micro events is the phenomenon of the seismic cycle which arises under the influence of stable geodynamic forces acting within a certain seismic belt. However, because of the hierarchy of the geophysical environment, the occurrence of any event with some M can be interpreted in two ways: this event can be considered either extraordinary, i.e., a bifurcation point (completion of the old spatial-temporal organization of the system and the start of a new history) or it can be considered an ordinary manifestation of stress relief in a broader system. The uncertainty with Mmax remains and the real answer can be obtained only after a long period of time. Instead, this uncertainty gives grounds to identify cycles of different orders and duration in the flow of seismic events. Then, any event completes one of the subcycles. Quasi-periodicity in the onset of events of the same order may be due to factors that are additional to the proper geodynamic forces—tidal, solar, galactic, etc. cyclic processes. They have been the subject of numerous studies (Khain and Khalilov, 2009). For this article, it is important to note that the cyclicity of seismogeodynamic activity with different recurrence periods might be useful for long-term assessment of the seismic safety of RW geological repositories as a factor in lowering or raising the level of seismic activity in a given region relative to the current level. Of interest are cycles that are associated with a change in the parameters of the Earth’s rotation and are of long-period nature. The inclination of the Earth’s axis varies with a period of 40 thousand years. The precession completes a full circle in 20 thousand years, and the eccentricity cycle in 100 thousand years. However, the possibility of obtaining empirical data in the foreseeable future to evaluate seismic activity cycles of such duration is questionable. Thus, the nonlinearity of seismogeodynamic processes in the source zones can be treated as a theoretical limitation imposed on the hypothesis of stationarity of the seismic regime. Due to uncertainty in the input data, the instability of a system whose evolution is to be predicted produces an inevitable chaos in the description of the system and, as a consequence, a short prediction horizon. The need for a long-term prediction for seismic activity in a particular region necessitates the transition from models of specific seismic sources to general models that have a smaller scale but cover

1079

increasingly longer periods of geological time and larger volumes of geological space. Geotectonic position of a repository site and seismic safety assessment Plate tectonics has provided some clarity about the forces driving the mechanism of earthquakes (Mogi, 1985). According to this theory, due to the stable network of convective cells, the mantle material rises from the depths of the Earth through narrow channels along the mid-ocean ridges and in some areas within the continental crust. The ocean floor builds up on both sides of the ridges, increasing the area of the oceanic crustal plates, which move to collide with continental plates and slip under them. According to these views, the collisions of the plates during their incessant movement lead to the formation of extended regions under very high tectonic stresses along the plate boundaries, and sooner or later these stresses are released in the form of earthquakes. The distribution of earthquakes over the Earth’s surface follows a certain pattern directly related to the nature of this phenomenon. According to modern concepts, earthquake sources are located along extended and relatively narrow zones of active fault structures of the crust and lithosphere, in which seismogeodynamic interactions of the plates take place. The latter are divided into three main types: the rift zones of the expanding oceanic crust, the zones of subduction or subduction of the oceanic crust under the continental crust, and the zones of transform movements of lithospheric plates. The converging lithosphere structures are seismically most active. They are represented by the arc-shaped boundaries between the lithospheric plates located in the form of subduction zones dipping under the continents on the periphery of the oceans, as well as relics of ancient subduction zones on the continents themselves (Ulomov and Shumilina, 1999). A more complex issue, in the opinion of many researchers, is related to the explanation of the mechanism of intraplate seismicity, which appears occasionally, develops in different regions of the crust, and has a tendency to migrate from site to site (Gangopadhyay and Talwani, 2003; Kenner and Segall, 2000; Schulte and Mooney, 2005). As for the Russian Federation and the adjacent territories, it has always been assumed that, most often, earthquakes sources are confined to regions of intense and contrasting latest tectonic movements, which, according to modern concepts, are just the structures of seismogeodynamic interaction between the plates. In the continental crust, such movements and earthquakes are confined to the Alpine folding zones (the Carpathians, the Crimea, the Caucasus and the Pamir), as well as to the parts of platforms that have experienced the latest tectonic activity (Tien Shan and Pribaikalie). In the zones of convergent junction of the oceans and continents in the vicinity of Kamchatka and the Kuril Islands, shallow-focus earthquakes occur on the Pacific side, and their depth gradually increases toward the continent, outlining the deep planes of the Zavaritskii–Benioff subduction (Bune and Gorshkov, 1980; Medvedev, 1968b).

1080

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

Fig. 3. Lithospheric plates of southern Siberia (fragment of V.I. Ulomov’s map (http://seismorus.ru/hazards/russia). 1, boundaries of lithospheric plates; 2, relics of convergent seismically active structures (subduction zones); 3, contour of the map in Fig. 4.

In the preparation of seismic hazard maps for the territory of the USSR in previous years, the algorithm of identification of seismically active structures did not allow the domination of any one of the existing general theoretical tectonic paradigms. Only in the preparation of the GSZ-97 map was its structural framework based on a single paradigm, namely, the theory of lithospheric plate tectonics. According to this paradigm, the seismicity of Northern Eurasia, on which the Russian Federation is located, is due to the intense seismogeodynamic interaction of several major lithospheric plates: European, Asian, Arabian, Iranian, Indian, Chinese, Pacific, Okhotsk, and North American (Ulomov and Shumilina, 1999). In contrast, shields and plates on the ancient platforms of the continents are characterized by relative tectonic stability and a small number of earthquakes differ. This makes them the most attractive for the location of geological repositories long-lived radioactive waste. However, strong devastating earthquakes in them also occur. Estimated probabilities of extreme events (M ≥ 7) for the sites recommended for radioactive waste disposal in the shields of ancient platforms, though uncertain, are very low (Fenton et al., 2006). The time of the existence of high-seismicity zones along the boundaries of tectonic plates and, hence, the low-seismicity regions located between these zones within the plates is hundreds of millions of years (geotectonic megacycles), which far exceeds the reasonable period of safety assessment for long-lived radioactive waste repositories. The megacyclicity of geotectonic processes, in our opinion, provides a fundamental geological basis for the long-term safety assessment of geological repositories located outside seismic zones on ancient platforms and shields. A key point in the disposal of long-lived radioactive wastes on ancient platforms is to avoid locating disposal sites on potentially active faults, for which, as mentioned above, the return period can exceed tens of thousands of years.

Fig. 4. Schematic of seismic zoning elements of the south of Krasnoyarskii Krai (prediction for PEC), according to (Sigbatulin et al., 2004). 1, seismically active faults; 2, regions of moderate seismic activity; 3, regions of increased seismic activity; 4, PEC; 5, radioactive waste repository site.

Unfortunately, the assessment of seismic regions adjacent to intraplate zones of high seismicity involves serious problems. For example, as judged by the seismogeodynamic zoning map (Fig. 3), the repository in the vicinity of Krasnoyarsk is located at a distance of about 1000 km to the north of the border with the China plate within the Asian plate. The region belongs to the Alpine–Himalayan belt of high seismic activity and is confined to a region of 7-point (or even 8-point) seismic hazard. Furthermore, this disposal site near Krasnoyarsk is located near seismically active faults (Fig. 4) which inherit an ancient subduction zone (see Fig. 3). This position does not exclude the scenario of activation of this ancient boundary of lithospheric plates in the distant future, which immediately makes it necessary to evaluate this probability. According to the theory, lithosphere plates move relative to each other, resulting in a change in the position and seismic activity of the active boundaries. In the geological perspective, the position of a region in the system of seismogeodynamic interactions does not remain constant. This is also true for radioactive waste disposal sites. The application of the modern mechanisms of the convective regime in the mantle to the theory of plate tectonics makes it possible to predict the movements of the plates and, accordingly, the positions of seismic zones in the distant future. Thus, Mitchell et al. (2012), using one of the possible scenarios, has made a prediction for the movements of the modern continents and their merger into a single supercontinent in the region of the Arctic Ocean in approximately 200 million years. Modern modeling techniques allow predicting movements of lithospheric plates, using a variety of hypothetical scenarios. Such exercises can be the basis for the prediction

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

for movements of seismically active regions in the very distant future. Despite the lack of common views on the dynamics of lithospheric plates, it is theoretically possible to predict (deterministically!) the migration of a radioactive waste disposal area together with the plate on which it is located for superlong (in the ordinary sense) periods. The uncertainty in this prediction is due to the evolution of the seismic regime of the area as the plate migrates, or it is due to the probability of plate splitting into separate fragments. Although the spatial position of the boundaries of lithospheric plates can be predicted from geodynamic positions for very remote periods, given that these boundaries will always be zones of high seismic activity, there is little that can be said about the evolution of seismicity at some sites of these zones or the possibility of formation of new such zones. Nevertheless, it is the theory of lithospheric plates that provides the most general principles for super-long-term prediction for seismicity, including for underground radioactive waste disposal sites.

Conclusions The seismicity of the area in which it is planned to make a radioactive waste repository in Krasnoyarsk Krai is characterized by a maximum calculated earthquake intensity of 7–8 points, which according to existing regulations, is considered a negative factor. The seismic safety of the future repository filled with radioactive waste containing long-lived radionuclides, which remain potentially hazardous over millions of years, should be assessed for the entire potentially dangerous period. Universally recognized indicators of seismic safety adequate to this task are not available to date. Seismic hazard prediction based on stochastic models of the seismic regime in individual regions remains suitable (with certain reservations regarding the range coverage of the hypothesis of stationarity of this regime) for a long-term period of thousands of years. For the superlong term, seismic hazard prediction should be replaced by seismogeodynamic prediction. This prediction can be based on deterministic models of migration of lithospheric plates, whose boundaries will always be characterized by increased seismic activity. Seismic safety assessment of geological disposal of longlived radioactive waste in a particular area based on any model is an extrapolation of geodynamic and seismotectonic laws identified in the analysis of data characterizing the development of the area in previous periods. Despite the uncertainty as to how far into the future this extrapolation will be acceptable to a particular prediction methodology, fundamental geological laws are the only guarantee of its authenticity. This work was supported by a program of the Presidium of the Russian Academy of Sciences.

References Apostolakis, G., 1990. The concept of probability in safety assessment of technological systems. Science 250 (4986), 1359–1364.

1081

Barbot, S., Lapusta, N., Avouac, J-P., 2012. Under the hood of the earthquake machine: toward predictive modeling of the seismic cycle. Science 336 (6082), 707–710. Bulletin PTsSK, 1913. Printing of the Imperial Academy of Sciences, St. Petersburg. Bune, V.I., Gorshkov, G.P. (Eds.), 1980. Seismic Zoning of the USSR. Methodical Basis and Regional Description of the 1978 Map [in Russian]. Nauka, Moscow. Considering Timescales in the Post-Closure Safety of Geological Disposal of Radioactive Waste, 2009. OECD/NEA, Paris. Crone, A.J., Machette, M.N., Bowman, J.R., 1997. Episodic nature of earthquake activity in stable continental regions revealed by paleoseismicity studies of Australian and North American Quaternary faults. Aust. J. Earth Sci. 44, 203–214. Falck, W.E., Nilsson, K.-F., 2009. Geological disposal of radioactive waste: moving towards implementation. Eur. Comm., Luxembourg. Fedotov, S.A., 1968. On the seismic cycle, the possibility of quantitative seismic zoning, and long-term earthquake prediction, in: Seismic Zoning of the USSR [in Russian]. Nauka, Moscow, pp. 121–150. Fenton, C.H., Adams, J., Halchuk, S., 2006. Seismic hazards assessment for radioactive waste disposal sites in regions of low seismic activity. Geotech. Geol. Engineer. 24, 579–592. Gangopadhyay, A., Talwani, P., 2003. Symptomatic features of intraplate earthquakes. Seismol. Res. Lett. 74, 863–883. GSZ-97D Map. http://seismos-u.ifz.ru/maps/ocp-97d.rar. Kapitsa, S.P., Kurdyumov, S.P., Malinetskii, G.G., 2003. Synergetics and Prognoses for the Future. Editorial URSS, Moscow. Kenner, S.J., Segall, P., 2000. A mechanical model for intraplate earthquakes: Application to the New Madrid seismic zone. Science 289, 2329–2332, doi: 10.1126/science.289.5488.2329. Khain, V.E., Khalilov, E.N., 2009. Cyclicity of Geodynamic Processes: Its Possible Nature [in Russian]. Nauchnyi Mir, Moscow. Kondorskaya, N.V., Shebalin, N.V., 1977. New Catalog of Strong Earthquakes on the Territory of the USSR from Ancient Times to 1975 [in Russian]. Nauka, Moscow. Koronovskii, N.V., Naimark, A.A., 2012. Unpredictability of earthquakes as a fundamental consequence of the nonlinearity of geodynamic systems. Vestnik MGU. Ser. 4. Geology, No. 6, 3–12. Kochkin, B.T., 2004. Geological uncertainties in the safety assessment of waste disposal systems. Geoekologiya, No. 2, 142–153. Kochkin, B.T., 2013. Assessment of the long-term safety of geological repositories for highly radioactive waste. Geoekologiya, No. 4, 342–352. Lobanov, N.F., Beigul, V.P., Kamnev, E.N., Lopatin, P.V., Revenko, Yu.A., Shramko, I.V., 2011. Federal facility of underground disposal of long-lived radioactive waste in a mining-chemical plant, Bezopasnost’ Yadernykh Tekhnologii i Okryzhayushchei Sredy, No. 1, 10–23, http://www.atomicenergy.ru/articles/2011/06/15/23447. Lunina, O.V., Andreev, A.V., Gladkov, A.A., 2014. Geologic hazards associated with seismogenic faulting in southern Siberia and Mongolia: forms and location patterns. Russian Geology and Geophysics (Geologiya i Geofizika) 55 (8), 1028–1042 (1294–1313). Medvedev, S.V., 1968a. Seismic zoning of the USSR territory, in: Seismic Zoning of the USSR [in Russian]. Nauka, Moscow, pp. 3–10. Medvedev, S.V. (Ed.), 1968b. Seismic Zoning of the USSR [in Russian]. Nauka, Moscow. Methods for Safety Assessment of Geological Disposal Facilities for Radioactive Waste. 2012. OECD/NEA, Paris. Mitchell, R.N., Kilian, T.M., Evans, D.A.D., 2012. Supercontinent cycles and the calculation of absolute palaeolongitude in deep time. Nature 482, 208–212. Mogi, K., 1985. Earthquake Prediction. Academic Press, Tokyo [Russian translation: Mir, Moscow, 1988]. Morozov, V.N., Rodkin, M.V., Tatarinov, V.N., 2001. On the problem of geodynamic safety of facilities of the nuclear fuel cycle. Geoekologiya, No. 3, 227–238. Nikolaev, A.V., 1998. Problems of Nonlinear Seismology, in: Issues of Nonlinear Geology and Geodynamics (Proc. III Seminar on Nonlinear Geology and Geodynamics, November 1998). GEOS, Moscow, pp. 65–69.

1082

B.T. Kochkin and V.A. Petrov / Russian Geology and Geophysics 56 (2015) 1074–1082

Nikonov, A.A., Fleifel, L.D., 2014. Forgotten accounts of historical earthquakes in Siberia (17th–18th centuries) by J.G. Gmelin. Russian Geology and Geophysics (Geologiya i Geofizika) 55 (4), 522–529 (669–677). Passchier, C.W., Trouw, R.A.J., 2005. Microtectonics, second revised and enlarged edition. Springer-Verlag, Berlin–Heidelberg. Prigogine, I., 1960. An Introduction to the Thermodynamics of Irreversible Processes. IL, Moscow; Academic Press, New York. Prigogine, I., 1991. The philosophy of instability. Vorposy Filosofii, No. 6, 46–52. Radioactive Waste Disposal: Principles, Criteria and Basic Safety Requirements. NP-055-04, 2004. Zaruchevskaya, G.P., Sorokin, V.T., Zakharova, K.P., Masanov, O.L., Zubkov, Yu.N., Iryushkin, V.M., Kochkin, B.T., Rybal’chenko, A.I., Levin, A.G., Smetnik, A.A., Sharafutdinov, R.B. Rostekhnadzor, Moscow. Seismotectonics of Ancient Platforms in the Field of Quaternary Glaciation [in Russian], 2009. Kniga i Biznes, Moscow. Scholz, C.H., 1990. The Mechanics of Earthquakes and Faulting. Cambridge Univ. Press, Cambridge.

Schulte, S.M., Mooney, W.D., 2005. An updated global earthquake catalogue for stable continental regions: Reassessing the correlation with ancient rifts. Geophys. J. Int., 161, 707–721. Sibgatulin, V.G., Simonov, V.V., Peretokin, S.A., 2004. Seismic Hazard Assessment of the South of Central Siberia [in Russian]. KNIIGiMS, Krasnoyarsk. Sobolev, G.A., 1993. Fundamentals of Earthquake Prediction [in Russian]. Nauka, Moscow. Stein, S., 2007. Approaches to continental intraplate earthquake issues, in: Stein, S., Mazzotti, S. (Eds.), Continental Intraplate Earthquakes: Science, Hazard, and Policy Issues. Geol. Soc. Am. Spec. Paper 425, pp. 1–16. Ulomov, V.I., Shumilina, L.S., 1999. GSZ-97 Set of Maps of General Seismic Zoning of the Territory of the Russian Federation (1:8,000,000). An Explanatory Note and a List of Cities and Towns Located in EarthquakeProne Areas. UIPE RAS, Moscow, http://seismos-u.ifz.ru/pdf/explanatorynotes-ocp-97.pdf. Ye, L., Lay, T., Kanamori, H., Koper, K.D., 2013. Energy release of the 2013 Mw 8.3 Sea of Okhotsk earthquake and deep slab stress heterogeneity. Science 341 (6152), 380–1384.

Editorial responsibility: V.S. Seleznev