Active faults of the northern Tien Shan: tectonophysical zoning …€¦ · Introduction Detecting...

www.elsevier.com/locate/rgg

Active faults of the northern Tien Shan: tectonophysical zoningof seismic risk

Yu.L. Rebetsky a,*, S.I. Kuzikov b

a United Schmidt Institute of the Physics of the Earth, Russian Academy of Sciences, ul. Bol’shaya Gruzinskaya 10, Moscow, 123995, Russiab Science Station of the Russian Academy of Sciences, Bishkek, 720049, Kyrgyzstan

Received 13 January 2015; accepted 28 May 2015

Abstract

This study continues the work by Mikhail Gzovsky on geological (tectonophysical) criteria for seismic risk. It is suggested to performseismic-risk zoning according to parameters of normal and shear stresses on fault planes converted from results of tectonophysical stressreconstructions. The approach requires the knowledge of both dip and strike of the respective fault segments. Slip geometry is estimated fromstress tensor, assuming that it is directed along shear stress. The suggested approach is applied to faults in the northern Tien Shan, and thecurrent stress parameters are reconstructed using source mechanisms of catalogued earthquakes recorded by the KNET seismological networkof the RAS Science Station in Bishkek. Stress modeling is performed by the method of cataclastic analysis providing constraints on stressellipsoids, as well as on relations between the spherical and deviatoric components of the stress tensor. Plotted on the Mohr diagram, the faultstress points allow estimating whether the respective fault segments are close to the critical state (brittle failure). The suggested seismic-riskzoning of faults in the northern Tien Shan reveals up to 25 km long hazardous fault segments.© 2016, V.S. Sobolev IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reserved.

Keywords: current stress; seismic risk; tectonophysical zoning; active fault; hazardous fault; Coulomb stress

Introduction

Detecting faults prone to generation of earthquakes is thekey point in seismic risk assessment. Discrimination of faultsegments in terms of potential hazardous slip or creep isimportant for safe and environment-friendly operation oftransportation and pipeline systems. Currently seismotectonicsand GPS measurements, as well as methods based on earth-quake records, are most commonly used for these purposes.Hazardous faults are identified from signatures of present andpast activity on the surface (rock falls, landslides, etc.) or intrenched faults. Seismotectonic analysis allows qualitativeassessment of hazard in different fault segments while inten-sity of paleoearthquakes from trenching data provides quanti-tative support.

Land-base and remote GPS measurements likewise canreveal activity of faults but mostly within short observationsperiods restricted to ongoing faulting. Instrumental records ofearthquakes highlight brittle failure processes deep in thecrust, but seismic analysis applies to historic events which

often have magnitudes much smaller than the maximumpossible values. The latter are found from recurrence plotsupdated using information on historic and prehistoric events.

Tectonophysical zoning of active faults is an alternativeapproach, which was suggested by Gzovsky (1975) fifty orsixty years ago but has been almost forgotten. Gzovsky usedthe term geological criteria of seismicity referring to signaturesof seismic hazard showing the most probable highest magni-tude of shocks and their expected recurrence. In the present-day terminology, they are rather tectonophysical criteria, aswe call them in the consideration below.

Tectonic and seismic activity occurs in areas of ruggedterrain and structurally differentiated crust, with high-gradientslip rates, and also in fields of young volcanism (Gzovsky,1975). The geological (tectonophysical) criteria of seismicitymafe basis for contouring areas of highest shear stress andhigh-gradient vertical motions within different periods: mil-lions of years for neotectonic methods, thousands and tens ofthousand years for geomorphology, and decades for GPSmeasurements. The results were used to predict stress andstrain trends and to assess the energy of pending events andduration of critical stress periods (Gzovsky, 1975).

Russian Geology and Geophysics 57 (2016) 967–983

* Corresponding author.E-mail address: [email protected] (Yu.L. Rebetsky)

Available online at www.sciencedirect.com

ScienceDirect

ed.

+1068-7971/$ - see front matter D 201 IGM, Siberian Branch of the RAS. Published by Elsevier B.V. All rights reservV.S. S bolevo

http://dx.doi.org/10.1016/j.rgg.201 .0 .006,

5 5 4

By tectonic stress Gzovsky meant primarily the stressmeasured directly in rocks with implications for principalstress directions (the method of conjugate par faults), whileestimation of total crustal stress was considered as a problemresolvable in the future.

We develop the approach by Gzovsky proceeding fromtoday’s advance in rock mechanic studies of critical preseismicstress.

Gzovsky’s approaches to seismic zoning

In his synthesis of seismic risk data available at that time,Gzovsky (1975) noted that geological (tectonophysical) crite-ria of seismicity sometimes contradicted real observations.Specifically, they would indicate high activity in the Alpswhere actually the seismicity was low, but low activity in theHungarian depression which was shocked by several earth-quakes of shaking intensity 8. To explain the paradox,Gzovsky hypothesized deceleration of mountain growth in theQuaternary with decreasing deviatoric stress and proposed tostudy ongoing crustal movements using GPS.

Some other paradoxical examples are high seismicity in theGarm valley in spite of tectonic and geomorphic signatures ofstability (Gzovsky, 1975) or low seismicity in the northwesternGreater Caucasus (zones of probable shaking intensity 6),though some parts of the shore between Sukhumi andNovorossiysk rather should belong to the intensity 8 zoneaccording to geomorphology and tectonics. Gzovsky notedthat the seismic process changed in time and crust movementshad slower or faster rates, which tectonic and geomorphologi-cal data, averaged over long periods, failed to resolve.

Gamburtsev (1955) also mentioned that tectonic move-ments varied with time and earthquakes had greater intensityduring reactivation of seismic sutures in the crust which wouldappear to be stable from parameters averaged over hundredsor thousands of years.

The earthquake magnitude was hypothesized (Gamburtsevand Belousov, 1960) to depend on the size of the high-stresszone and the magnitude of largest events to influence thefrequency of earthquakes, as the duration of the pre-faultingmaximum shear stress was related with the stress level.Therefore, the size of the zone of high potential energy ofelastic strain was related to the length of seismogenic faults.Gzovsky (1975) noted in this respect that triaxial compressionimpeded earthquakes while faults could be healed.

He interpreted nucleation of large earthquakes in largefaults as prolonged formation of small faults showing up assmall earthquakes; in their turn, large earthquakes wereunderstood as rapid motions along large faults as a result ofpulse-like growth of originally isolated small faults (Gzovsky,1975). The potential energy of elastic strain in the zone ofhigh maximum shear stress drops dramatically as it becomescrosscut by a growing fault. Gzovsky (1975) considered aseismic source as an area where the potential elastic strainchanged as a result of fault reactivation rather than as atectonic suture (Dobrovolsky, 1991), an idea which agreed

with the views by Benioff (1951), Bullen (1955) and Savaren-sky (1954). Gzovsky also mentioned the importance offaulting mechanisms in their relation to crustal stress.

The stress responsible for the ongoing seismic process candiffer from that in which the respective faults were originallyforming. Therefore, it is crucial to compare stress in activefaults with the highest level and with stress estimates basedon the amount of slip accumulated for hundreds of thousandor millions of years in the past.

Natural stress and strain: state of the art

Decades ago, when Gzovsky (1975) formulated the objec-tives of tectonophysics in seismic risk assessment, data ofnatural stress were collected by measurements of strain in crustand directly in mines during hydraulic fracture; by geomor-phological and GPS measurements; and by methods of fieldtectonophysics and seismology. Geomorphological and GPSestimates gave rates of terrain changes and required theknowledge of relations between stress and gradients of motion.The measurements of joints in tectonophysics and fault planesolutions in seismology could show only principal directionsbut not the level of stress.

The situation has changed over the recent decades, both inmethods and in amount of collected data. Remote sensing hasallowed great progress in GPS measurements, with horizontalcrust movements measured ten times more precisely than thevertical ones. Recent estimates of horizontal movements inCentral Asian intracontinental orogens (Kuzikov and Muk-hamediev, 2010; Sankov et al., 2011; Timofeev et al., 2009,2013; Zubovich et al., 2007) were interpreted mostly in termsof regional evolution models, but they should be convertedinto gradients along some directions or rates of longitudinalstrain. Strain can be also estimated by triangulation methods,which are quite well developed (Sankov et al., 2011; Tychkovet al., 2008). In continental active tectonic areas, lateral strainhas annual rates of n × 10–8–n × 10–9 yr–1; higher rates areobtained at shorter base lines and smaller scales of averaging(Karmaleeva, 2012; Kuzmin, 2004): three or four orders ofmagnitude higher at tens to hundreds of kilometer lines, whilestrain is localized in vicinities of large faults.

Stress patterns have been largely documented for the upper3 km of crust (Brady and Bzown, 2004; Brown and Hoek,1978; Herget, 1973; Potvin et al., 2007; Zubkov et al., 2010).Much evidence was obtained previously on horizontal com-pression shortening exceeding the standard lateral pressure(Dinnik, 1926) and vertical (overburden) pressure. Currentlyit is known that (i) vertical stress varies geographically butapproaches the overburden pressure; (ii) most of measure-ments show horizontal compression exceeding the verticalstress in areas of long-lasting uplift; (iii) most of measurementsshow stress about the standard of Dinnik (1926) in areas oflong-lasting subsidence and nearly vertical maximum com-pression; (iv) horizontal compression axis is especially vari-able near the surface (from 0.3 to many times as high as theoverburden pressure) but the variations decrease with depth

968 Yu.L. Rebetsky and S.I. Kuzikov / Russian Geology and Geophysics 57 (2016) 967–983

(0.8 to 1.1 times the overburden pressure at a depth about3 km); (v) principal stress directions are strongly variable atsome localities, and the principal direction of compression ishard to estimate at large horizontal stress.

The methods for studying natural stress in mines have notchanged much while tectonophysical studies have advancedconsiderably since the 1960s when conjugate shears was theonly method (Gzovsky, 1975). The new method of cataclas-tic analysis (MCA) of discontinue displacements (Rebetsky,2001) allows estimating stress tensor components, strainincrement, and mechanic properties (including strength andyield) on the basis of plasticity theory. The MCA procedureconsists of four stages. It begins with independent parallelcalculations of principal stress directions and the Lodé-Nadaícoefficients for stress tensors and seismotectonic strain incre-ment (Rebetsky, 2001) proceeding from basic energy postu-lates of mechanics. The second stage aims at estimatingthe relative effective isotropic pressure and the magnitude ofhighest shear stress that remained unknown at the first stage(Rebetsky, 2003). Modeling stems from experimental resultsof Byerlee (1968) showing that partly healed old joints ofdifferent directions can reactivate as shear stress on theirsurfaces exceeds the surface friction. At the third MCA stage,calculations apply to effective cohesion averaged on the scalecorresponding to that of the reconstructed stress (Rebetsky,2009a,c), using stress release in the sources of largest localearthquakes as additional data. Finally, tectonic and pore fluidpressure values are estimated which are related with effectivepressure known from the previous stage, assuming that verticalstress approaches the overburden pressure (Sibson, 1974); anupdated version of the latter hypothesis is based on the verticalmomentum conservation in the thick-plate approximation(Rebetsky, 2007a).

Thus, the MCA approach can provide constraints on thetotal stress tensor and fluid pressure, and thus give a new

perspective of detecting active faults and grouping them interms of seismic risk.

Active faults in the northern Tien Shan

MCA data on natural stress in the northern Tien Shan(Fig. 1) obtained by a joint team of the United Institute of thePhysics of the Earth, laboratory of tectonophysics (Moscow)and the Science Station (Bishkek) will be used further fortectonophysical zoning of active faults. We found and digit-ized 102 active faults (Fig. 2) with constant dips, each withmultiple straight segments, marked by the coordinates ofstarting and final points. The length of these segments(0.01–0.02 deg) are mainly a few kilometers or rarely10–15 km. The mapped faults show strike variations, on bothshort (a few kilometers) and long (tens of km) base lines,according to varying strikes of basins and ranges and otherterrain elements. The accuracy of strike estimates can beimproved by averaging over 10 km.

Note that digitized azimuths of fault segments are necessarybut insufficient for tectonophysical analysis in which theinferred 3D stresses are projected the onto a surface (faultplane): both strike and dip of faults have to be known,especially, in the case of dipping fault planes. In this respect,constant dips were assigned to fault segments of differentlengths on the basis of additional geomorpholgical analysis ofgeological maps and field measurements (Fig. 3).

The mapping faced several problems. One is the lack offault maps of a due scale to be used for reference, while theavailable composite geological and tectonic maps (1:500,000and smaller) are obsolete and often disagree with moredetailed maps. Thus, special care was taken to draw correctlythe continuous lines of main faults. More difficulty came fromdiscontinuous exposure of faults impeding field measurements.Poorly exposed and closely spaced parallel faults could be

Fig. 1. Topography and large tectonic units of Central Asia. Numerals in circles stand for names of intermontane basins: Tajik (1), Fergana (2), Issyk Kul (3), Ili (4),Jonggar (5), Alakul (6), Zaisan (7, 8), Turfan (9). Box frames study area in northern Tien Shan.

Yu.L. Rebetsky and S.I. Kuzikov / Russian Geology and Geophysics 57 (2016) 967–983 969

mistaken for a single structure or, vice versa, one long faultcould be misinterpreted as multiple faults.

Another problem was that the study has focused on activefaults, which are much fewer than inactive ones in the existingmaps; furthermore, different maps assume different durationsof activity required to consider a fault as active.

Some 1:50,000 geological survey maps and most ofsmaller-scale maps lack indications of dip directions andaverage angles, but show only directions and geometry of slip(thrust, normal, reverse or strike slip) or more often nothingbut fault lines.

One more thing, not critical but annoying, is the lack of aregional catalog of fault names which would be helpful to pickdata from different texts and figures.

Note that the maps of Figs. 2 and 3 were compiled withreference to other maps and text publications. They are,namely, more than sixty sheets of geological and tectonic mapsof Kyrgyzstan and southern Kazakhstan of different scales;the map of Late Pleistocene–Holocene active faults of thePamir–Tien Shan region (Makarov, 2005); the tectonic mapof Kazakhstan and surroundings (Bespalov, 1975) and severalother 1:500,000 maps (Chedia, 1988; Igemberdiev, 1982;Tursungaziev and Petrov, 2008; Zhukov, 1988); eight sheetsof the 1:200,000 geological map of the USSR, northern TienShan series, compiled by Grishchenko and Turbina in 1964;Okhotnikov and Novikov in 1962; Chabdarov et al. in 1967;Burtman et al. in 1960; Pomazkov and Burov in 1958;Zakharov and Zakharova in 1958; Pomazkov in 1962; andLasovsky and Mozolev in 1961, all joined in the collection ofBespalov et al. (1960–1973); some later versions of unpub-

lished 1:200,000 maps K-43-IX and K-43-XV compiled byZhukov (1975); a later version of the 1:200,000 geologicalmap published as an open file (Mikolaichuk and Apayarov,2004).

Mountains flanking the easternmost Chu basin weremapped using the 1:100,000 unpublished structural map byEfremov et al. compiled in 1989, along with analysis of1:50,000 unpublished geological and hydrological maps (fewpieces from the archives of the Geological Survey of Kyr-gyzstan) by Zakharov et al.; Levchenko et al.; Bondar et al.;Semenov et al.; Galanin et al.; Khristov et al.; Rubtsov et al.;Morozov et al.; Bondar; and Chernyavskaya et al. compiledin the 1960s through 2001.

Faults were considered active if they crossed Cenozoicrocks (or structures). In the case of controversy in faultlocations among different authors, more trust was given tolater references, larger-scale maps, and personal observations.

The fault planes with ambiguously estimated dips werecharacterized by average values in the case of minor variance;if the difference exceeded 30°, the fault was divided intosegments, and the missing field-measured dips were calculatedapproximately according to horizontal topography elementscrosscut by the fault. In a few cases, unknown steeply dippingsides of faults were inferred from geology and tectonics andestimated by analogy with the surrounding areas. We tried toreduce subjectivity-related biases, which are inevitable in theseconditions.

The fault dips were corrected for possible increase withdepth and the results were used together with data on stressin the upper 10 km of seismogenic crust. The accuracy of dip

Fig. 2. Active faults of northern Tien Shan. Numerals in circles stand for names of faults: Aksu (1), Chonkurchak (2), Baitik (3), Alamedin (4), Issyk-Ata (5), Shamsi(6), Karakunuz (7), Zaili (8), North Kemin (9), South Kemin (10), Toguz-Bulak (North Issyk Kul) (11), Aramin (12), Kyzyloi (13), Karakul (14), Ulunbulak(Uzunbulak-Oikain zone) (15), Karakol (16), Ortok (17), Akchop (18), South Kochkor (19), Karakurdjur (20). Pentagon marks epicentral area of Belovodskearthquakes of 1865 and 1986.


estimates, though not very high (10°–20°), was sufficient forestimating reliably the stress parameters. The effect of dipaccuracy was assessed by additional stress calculations inwhich dips were allowed to be 10° higher of lower than inFig. 3 (see below).

The map of Fig. 3 provides a better idea of faulting in thenorthern Tien Shan than the map of A. Mikolaichuk publishedby Kalmetieva et al. (2009) and than that of Delvaux et al.(2001). Most of N–W faults dip south- or northward (Fig. 3)and some N–S faults dip to the east or west at 20° to 80°(relative to the horizon), most often at 50°–70°.

Most of faults in Fig. 3 dip beneath ranges (the Suusamyrbasin, southern border of the Chu basin, etc.), except for thosein the southern slope of the Kendyktas Range and the northernslope of the Kastek Range. In the central Kyrgyz Range andnear the eastern end of the Chu basin, there are parallel faultsdipping to the north and to the south, i.e., faults of differentslip geometries may be spatially close, which is important forestimating principal stress directions.

In this study we determine fault dips using field geologicaldata and topographic maps, from changes in curvature of faultprojections on the surface. Dip angles are taken from geologi-cal reports and field data.

Crustal stress in the northern Tien Shan and theoretical background for tectonic zoning

Seismological data. The stress pattern of the northern TienShan was reconstructed using the catalog of earthquakemechanisms recorded by the local KNET network run by theRAN Science Station (Bishkek), with reference to our previousresults (Rebetsky and Sycheva, 2008; Rebetsky et al., 2012).The catalog includes more than one thousand M = 1.16 to 5.40events (Sycheva et al., 2003, 2005; Sycheva and Kuzikov,

2012) that occurred from 1994 through 2012 (Fig. 4). Forstress modeling we used the mechanisms of M = 1.5–3.5earthquakes, which corresponded to the regional scale of stressaveraging on 0.05° × 0.05° grids. The grid spacing was about5–6 km in longitude and latitude, according to the 10–15 kmlinear averaging scale inferred from the density and magni-tudes of the catalogued events. The modeling was applied tooverlapping 10-km thick horizontal layers with their middleat the depths 5, 10, 15, and 20 km, but the consideration belowis confined to the upper 10 km for which stress parameterswere obtained in 286 domains.

Principal stress directions (maximum deviatoric extensionand maximum compression, Fig. 5a, b) were estimated at thefirst stage of MCA stress reconstruction (Rebetsky, 1997,2001, 2007a) as three Euler (triple orthogonal vectors) anglesor as dip directions and angles, along with the Lodé-Nadaícoefficient µσ, to an accuracy of 5°–10° and 0.1–0.2, respec-tively, in stable areas but at a worse accuracy in areas ofvariable earthquake mechanisms.

The patterns obtained at stage I of MCA modeling (Fig. 5)were discussed earlier (Rebetsky and Sycheva, 2008; Rebetskyet al., 2012). Note that stage I yields four out of six stresstensor components, which represent deviatoric stress accurateto the values normalized to maximum shear stress. However,even these results provide additional information on faults iftheir dip angles are known, at least approximately (besides thestrike). With three Euler angles and the dip and strike of faults,one can estimate the directions of shear stress applied to thefault plane. Only the magnitude of this stress depends on themaximum shear stress (which remains unknown after the firststage of modeling) while its directions presumably coincidewith the relative displacement of the fault walls (Bott, 1959;Wallace, 1951). The latter hypothesis is used below to inferthe slip geometry of faults.

Fig. 3. Map of faults in northern Tien Shan (compiled by S.I. Kuzikov, see text). Dip angles, in degrees (published for the first time), and directions are shown onhanging walls. Histogram in top left corner shows dip distribution of fault segments.


Of course, the estimation is possible only for the faults thatrun across or near the crust domains characterized by stressdata. Depth-dependent variations of fault coordinates weretaken into account by extending the faults to a depth of 5 kmalong the known dip and using the next grid node (no fartherthan 15 km or three times the characteristic grid spacing).

Reduced stress. The deviatoric stress components normal-ized by the unknown cohesion strength of rocks τf can beestimated from earthquake mechanisms and are called reducedstress (Rebetsky, 2005, 2007b). Figure 6 shows the lateral

distribution of reduced effective confining pressure (p∗ =p − pfl is the pressure with regard to the fluid pressure pfl,according to (Terzaghi, 1943)); such patterns are also availablefor reduced values of maximum shear stress and principalstresses.

The cohesion τf of rocks averaged over a few tens ofkilometers is constant within a zone of regional geodynamicsetting. Thus comparison becomes possible for reducedstresses obtained at different grid nodes, as well as on planesof different fault segments, and can discriminate zones of highand low stress (Fig. 6).

The state of faults depends on stress in their vicinities, andthis dependence shows up in fault morphology. The shearstress τn that controls convergence of fault walls is higherwhen the fault plane is closer to the plane of maximum shearstress. On the other hand, this shear stress should over-come friction that depends on effective (Terzaghi, 1943) stress

σn∗ normal to the fault, i.e., reactivation of faults depends on

the relation between normal and shear stresses on the faultplane.

Fault strength. The Coulomb stress (τC = τn + kf σn∗ at

σn∗ < 0 and kf = 0.6) includes shear and effective normal

stresses, and its critical value reflects the rock strength (yield).The proximity of stress to the critical value can be judgedfrom the Mohr diagram (Fig. 7). The large and small Mohr

circles refer to the effective principal stress components σ1∗,

σ2∗, σ3

∗ while points between them (light gray shade) representeffective normal and shear stresses at randomly oriented planesin the crust. The zone between the lines of brittle strength(yield) and minimum friction (dark gray shade) correspondsto the state of faults prone to activity (slip) as the Coulombstress τC at the given fault segment reaches the surface

cohesion τfi (0 ≤ τf

i ≤ τf).The Coulomb stresses of faults can be positive or negative.

The former refer to stress points within the field of brittlefailure in the Mohr diagram, between the lines of yield andminimum friction (Fig. 7). The closer the point to the yieldline (C), the greater the hazard (white points) and the higherthe shear stress released in slip. The points near the line ofminimum dry friction are within the field of the least probableactivity (open circles). Negative values show that the pointlies within the elastic field, i.e., to the right of the minimumfriction line (triangles).

The fault segments were divided into three groups accord-

ing to reduced Coulomb stresses τC / τf (τC = τn + kf σnn∗ ):

(i) shear stress τn at the fault plane is below the minimumfriction (τC ≤ 0); (ii) shear stress is above the minimum frictionbut the Coulomb stress is not very high (0 < τC / τf ≤ 0.8); (iii)the Coulomb stress is high 0.8 < τC / τf ≤ 1. In the considera-tion below, the fault segments of the three groups are called,respectively, inactive, active, and hazardous.

Division of active faults in the northern Tien Shan

Slip geometry of active faults can be inferred from relativemagnitudes of shear stresses along the dip and strike of thefootwall fault plane (Fig. 8). Namely, stress projection on thefault dip exceeding that on the strike corresponds to the normal

Fig. 4. M = 1.1–5.4 earthquakes in northern Tien Shan, stations of KNET seismological network (pentagons), and large faults.


Fig. 5. Stress pattern of northern Tien Shan (upper crust, 0 to 10 km) obtained by MCA, stage I. Horizontal projections of minimum (a) and maximum (b) principalstresses and Lodé-Nadaí coefficients (c). Insets in top left corners show rose diagrams of principal stresses (a, b) and Lodé-Nadaí coefficients (c). See text forexplanation.


or reverse slip geometry, with right- or left-lateral strike-slipcomponents; thus, the concept of reverse slip includes low-an-gle thrusts and steeply dipping faults. The setting with greatershear stress along the strike than along the dip corresponds toright- or left-lateral strike slip, with or without reverse ornormal slip components. The stress-based division applies onlyto the fault segments for which stress data are available fromareas within 15 km around; otherwise, faults are drawn as thinlines (Figs. 9–11).

The relative magnitudes of shear stresses along the dipand strike of fault segments used for their classificationare measured by the ϕ angle: 60°–120° and –60°…–120°for reverse and normal slip, respectively; 120°–150° and–120°…–150° for reverse and normal slip, respectively, witha right-lateral component; 30°–60° and –30° to –60° for reverseand normal slip, respectively, with a left-lateral component;and the ranges –30° to 0° and 0 to 30° as well as 150° to 180°and –150° to –180° for left- and right-lateral strike slip,respectively (Fig. 8).

Thus the fault segments form four groups: (i) mostlyreverse slip with a minor right- or left-lateral strike-slipcomponent (Fig. 9a); (ii) mostly normal slip with a minorright- or left-lateral strike-slip component (Fig. 9b); (iii)mostly right-lateral strike slip with a minor reverse or normalcomponent (Fig. 9c); (iv) mostly left-lateral strike slip with aminor reverse or dip-slip component (Fig. 9d). Reverse slip isobviously present in zones subject to shortening (Fig. 5a, b),while dip-slip and strike slip geometries are typical ofextension and shear settings, respectively.

Faults over most of the study area, especially in its centraland northeastern parts, show reverse slip with a minor right-or left-lateral component (Fig. 9a) and strike mainly in theENE–WSW directions. The assumption of 10° steeper dips inreverse faults makes them generally less representative. Thesituation improves if the faults are reidentified as thoseincluding segments with a right- or left-lateral strike slipprevalent over reverse slip, with 10° shallower dips, respec-tively.

Fig. 6. Reduced maximum shear stress τ∗ / τf (a) and effective pressure p∗ / τf (b). Histograms in top left corners show representativeness of stress data. See text forexplanation.


The faults with prevalent normal slip are much fewer; theyare located mainly in the east and west of the study area(Fig. 9b) and vary in strike from ENE–WSW to ESE–WNW.All faults in the central part of the northern Tien Shan havereverse slip geometry. Segments with reverse slip are muchlonger than those of dip-slip. Unlike the case of reverse-slipfaults, normal faults become more representative when as-sumed to have 10° steeper dips but the assumption of lowerdip angles does not make them less representative.

Right-lateral strike-slip faults with minor reverse anddip-slip components are commensurate in length with preva-lently reverse-slip faults (Fig. 9a). They strike mainly in theWNW–ESE direction and are rather evenly distributed overthe study area (Fig. 9c), except for zones along 74°05′ and75°10′ E. The reverse slip component is present in a greaterpart of right-lateral strike-slip segments and in almost allleft-lateral segments. The latter are located mainly in the eastand in the west of the area (Fig. 9d), are most often shorterthan 15 km, and strike in the NE–SW or ENE–WSWdirections.

Faults with prevalent strike slip become slightly morerepresentative at the assumption of 10° steeper dips. Thereidentified fault segments are short and connect those ofprevalent reverse slip. Note that our division may differ fromsurface geological observations because of errors in fault dipestimates. Some faults turned out to be more distant from deepearthquakes than it was expected according to the assigneddips (Fig. 3, 4), this being evidence of dip change with depth.

The estimation of fault dips from geological data is themost problematic point in the suggested approach. The dipsshown in Fig. 3 correspond to the presumed behavior of faultsin depth, as the stress patterns were analyzed to depths within10 km. Thus, the identified slip geometry may slightly differ

from that observed on the surface, though the differencecannot be as dramatic as leading to a wrong choice betweendip-slip and reverse slip or between right and left strike slip.

Comparison of inferred slip geometry with availablegeological data. It is important to compare the inferred slipgeometries (Fig. 9) for faults in the upper 10 km of crust withpublished surface geological data. Our results agree with thegeometry of slip according to the mechanism of the M = 8.3Kemin earthquake of 1911 which occurred in the Chon-Keminzone (northern Issyk Kul region), in the northeastern part ofthe study area. The earthquake mechanism was interpretedpreviously as reverse slip with a minor left-lateral strike-slipcomponent (Bogdanovich et al., 1914; Kuchai, 1969). TheChon-Kemin zone distinguished by Shults (1948) and Maka-rov (1977) consists of ENE faults that dip southward (Bach-manov et al., 2008) and have a reverse-slip geometry with aleft-lateral strike-slip component. In the west, the zone joinsWNW–ESE faults with their geometry interpreted as right-lat-eral strike slip (Abdarakhmatov et al., 2001a), which isconsistent with our predictions of right-lateral strike slip withor without a reverse component.

Our interpretation of the South Kochkor fault geometry(Fig. 9a) is in line with signatures of reverse and right-lateralslip in prehistoric earthquakes as ~25 km long ruptures inbedrock found in the southern Kochkor basin along thenorthern slope of the Terskei Range (Korjenkov, 1999).

The results for the Shamsi-Tyundyuk fault dipping to thesouth agree with previous geological estimates: a reverse slipwith a right-lateral strike-slip component (Makarov, 2005).

The main reverse fault in the Uzunbulak–Oikain zone(Figs. 3 and 9), which begins at the southwestern terminationof the Kochkor basin, dips to the north according to Delvauxet al. (2001) and Kalmetieva et al. (2009). Our modeling alsosupports the SE dip inferred by Bachmanov et al. (2008) forits largest pinnate fault, though we suggest a higher angle: 60°(Fig. 3) instead of 30°.

There are, however, some cases of disagreement, either inthe inferred slip geometry or in dip angles. For example, WSWfaults in the Kastek Range have a left-lateral strike slipcomponent according to Delvaux et al. (2001) but reverse,

Fig. 8. Slopes ϕ (counted relative to strike; positive along dip) that controldirection of slip along fault plane (β is dip angle) correlated with slip geometry.See text for explanation.

Fig. 7. Mohr diagram for analysis of stress in fault segments. σj∗, j = 1, 2, 3 are

effective principal stresses, with regard to fluid pressure, τf is cohesion. Shear

stress (τn) is along vertical axis and effective normal stress (σnn∗ ) is along

horizontal axis (negative to the right). Domain of large Mohr circle above smallMohr circles corresponds to possible stresses. Solid line oblique to large Mohrcircle corresponds to maximum brittle strength (yield) of rocks (τf) and heavydash line below corresponds to minimum strength (friction); zone between themcorresponds to critical state and encloses points of positive Coulomb stress; thin

dash line shows intermediate strength τfi at point i. Pentagon (C) marks maxi-

mum Coulomb stress. Slope of friction and yield lines is related to frictioncoefficient as kf = tan φ. Domain of elastic behavior right and left of minimumfriction includes points of negative Coulomb stresses (triangles). See text forother explanations.


Fig. 9. Slip geometry of faults according to current stress in upper 10 km of crust (Fig. 5). a: reverse slip, b: normal slip, c: right-lateral strike slip, d: left-lateral strikeslip. Mixed vertical and horizontal components of motion are shown as well: reverse and dip slip with strike-slip components. Thin lines are faults for which no stressdata is available.


normal, and right-lateral strike slip components in our results;the change from reverse to normal slip may be due to quitea high dip angle of 60° and nonverticality of the principalstress.

Another example concerns faults of the Karakol zone,which extend from the northwestern end of the Kochkor basinalong the Karamoinok Range in the north and change theirstrike from NW to W–E near the Suusamyr basin. Ourinterpretation of right-lateral strike slip with reverse andnormal components (Fig. 6) disagrees both with left-lateralstrike slip suggested by Bachmanov et al. (2008) and Kal-metieva et al. (2009) and north-dipping reverse slip accordingto Delvaux et al. (2001). A large fault running in the north ofthe zone across the southern Kyrgyz Range appears as a W–Eleft-lateral strike slip in (Kalmetieva et al. (2009). We suggeststrike slip with a dip-slip component for a small segment ofan oblique fault in the Karakol zone (Fig. 9b) dipping at ahigh angle of 70°. This is a case of a vertical shear setting(Yunga, 1990), when the maximum and minimum principalstresses are at ~45° to the vertical and minor variations of thisangle can lead to either reverse or dip slip.

Like the Karakul fault, the faults in the northern border ofthe Suusamyr basin are interpreted as having a right-lateralstrike slip component by V. Trifonov (Makarov, 2005) but asleft-lateral strike slip faults by Kalmetieva et al. (2009). Ourresults agree rather with the former interpretation.

In a few cases, our estimates of fault dips differ fromgeological data. This is, specifically, a higher dip angle of 50°(Fig. 3), corresponding to deeper crust, we infer for theIssyk-Ata fault which was reported (Korzhenkov et al., 2012)to thrust northward at 190° and to dip at 19° near the surface.

For the South Kochkor fault, we suggest a dip angle smallerthan that derived from seismotectonic and seismic data(Omuraliev et al., 2009): 40° (Fig. 3) against 60° and 45°above and below the 17.5 km depth, respectively.

Thus, the results of our modeling for the current deforma-tion in the northern Tien Shan are generally consistent withpublished data of neotectonic motions observed on the surface.

Stress at fault segments. Reduced shear stress |τn| / τf and

effective normal stress |σn∗| / τf values (Fig. 10a, b) were

calculated at stage II of the MCA procedure with regard topore fluid pressure. Note that extension is considered positivein MCA, as in the classical mechanics. Normal and shearstresses often vary along faults according to their strike anddip variations; stresses may also change from one grid nodeto another.

High shear stress (Fig. 10a), along with normal compres-sion (Fig. 10b), appears over the longest segments of W–Efaults at the junction of the Kyrgyz and Karamoinok Ranges.An NW fault in the central Kyrgyz Range experiences lowcompression normal to the fault plane, which is expected toproduce local rupture and springs on the surface. Locally,extension is inferred as well for this segment, but it should betaken with caution, as the stress is estimated at depths to10 km, which is the accuracy limit of the method.

Inasmuch as normal compression results from friction thatimpedes shearing, the shear-to-normal stress ratio can beinterpreted in terms of seismic risk. This applies to faultsegments of high shear stress at the junction of the Kyrgyzand Karamoinok Ranges and the central Kyrgyz Range(Fig. 10a), as well as to some faults where both shear andnormal compression stresses are low. Additional calculationswith the assumption of 10° shallower dips show that the faults

become more representative, with higher |σn∗| / τf ratios, almost

fully at the account of faults north of the Djumgal Range andthe Sandyk Mountains.

Pore fluid pressure, which contributes largely to thecritical stress, can be estimated by the MCA algorithm usingregional average cohesion τf, which is either calculated fromreleased stress (Rebetsky, 2009a,b,c) or assumed a priori, for

Fig. 9 (continued).


Fig. 10. Zoning of faults according to stress. a, Reduced shear stress |τn| / τf ; b, reduced normal stress |σn∗| / τf (compression is negative); c, shear-to-normal stress ratio

τn / |σn∗|. Thin lines are faults for which no stress data is available.


instance, as 6 MPa obtained for the crust of the Altai–Sayanarea (Rebetsky et al., 2013). With this assumption and a rockdensity of ρ = 2.7 g/cm3, the pore fluid pressure for the faultsof the area is from 0.37 to 0.99 of the overburden pressure(Fig. 11).

The pore fluid pressure reaches its limit (0.95 < pfl / plt

< 1) in quite many faults but is rather low and close to thehydrostatic pressure (pfl / plt ≈ 0.4) in some faults. Increase inpore pressure can trigger slip even in faults (active or inactive)that are not hazardous at present. They are, namely, a largegroup of fault segments within an N–NW zone along 74°20′ E,where the pore pressure is 0.4–0.7 of overburden pressure(Fig. 11). The zone is located at the western termination of alarge cluster of stress values in the central Northern Tien Shan;stress estimates west of this zone are much fewer, and thestress pattern is generally less stable. There are two more localzones of 0.6–0.8 pore/overburden pressure ratios, slightlyhigher than in the zone along 74°20′ E, where the faults canbecome hazardous upon a 10% pore pressure increase: onenorthwest of Lake Issyk Kul and the other in the southernslope of the Kastek Range.

According to our experience (Rebetsky, 2007a, 2009b;Rebetsky and Marinin, 2006a,b; Rebetsky and Tatevossian,2013), shear stress is medium or even low in areas where largeearthquakes are being nucleated. The reason is that effectivenormal stress in these segments is low and the released energyis spent more on rupture growth than on resisting friction(Rebetsky, 2007b; Rice, 1980). In this respect, the segmentsunder low normal compression (Fig. 10b) that fall within thezone of brittle failure (high Coulomb stress) are hazardous aswell.

Identification of hazardous faults from Coulomb stress.The patterns of normal and shear stress, along with pore fluidpressure data, have implications for division of faults accord-ing to seismic risk. In order to check the accuracy of thedivision based on Coulomb stress, we performed calculationsassuming different fault dips: those as in Fig. 3 and 10° steeper

or shallower (Fig. 12), and obtained rather long (15–25 km)hazardous segments in all cases.

The longest 25 km fault segment, with stress about thecritical value (0.8 < τC / τf < 1), along the southern slope of theDjumgal Range and Sandyk Mountains (Fig. 12b; dip anglesare same as in Fig. 3) can generate M = 6.5–7.0 earthquakes.Shorter hazardous fault segments (10–15 km) at the junctionof the Kyrgyz and Karamoinok Ranges (northward fault dips),in the central Kyrgyz Range, and in the southern slope of theKastek Range, can generate M = 5.5–6.5 events, about thelargest possible magnitude for the area.

There are also several shorter segments of critical state nearthe Bishkek RAS Science Station, the longest one (12 km) atthe junction of the Chu basin and the northern slope of theKyrgyz Range (Fig. 12b), within the GPS network of theScience Station. Still shorter (within 5 km) hazardous seg-ments fall within the GPS network along N–S profiles east ofthe Science Station. Monitoring of crust deformation alongthese profiles would mitigate the risks. Nearly critical stresshas been inferred also for two 15 km long fault segments inthe eastern end of the Chu basin along the southern slope ofthe Kastek Range.

In the northern Issyk Kul area, hazardous fault segmentsare no longer than 5 km. The currently active Toguz-Bulakfault (Fig. 12b) appeas nonhazardous according to the Cou-lomb stress estimated for a dip of 70° (Fig. 3) but hazard ispredicted for its western segment for a smaller dip of 60°(Fig. 12a). A similar situation occurs along the southern slopeof the Kastek Range. Our modeling shows that the Toguz-Bu-lak fault is the most active in its western segment, where itchanges its strike from W–E to NW, but it becomes progres-sively less active in the eastern direction. The hazardoussegment of the fault (Fig. 12a) corresponds to its ~15 km longwestern flank. This result agrees with seismotectonic data(Korjenkov et al., 2011) from the Lake Issyk Kul areaindicating that the fault has been inactive in the Quaternary,and amount of motion along it decreases eastward.

Fig. 11. Zoning of faults according to pore fluid pressure pfl /plt.


Currently there are no hazardous faults in the area of theBelovodsk events (Fig. 12): two large earthquakes of M = 6.4and M = 6.9 in 1865 and 1885, respectively, and an M = 6.9event of 1838 near the eastern end of the Chu basin, on the

northern slope of the Kungei Range (Kalmetieva et al., 2009).However, the stress is critical in a 10 km long NE segment10–20 km southwest of the zone (Fig. 12b). The assumptionof 10° higher or lower dips does not change the predicted

Fig. 12. Zoning of faults according to reduced Coulomb stress τC / τf, for different fault plane dips. a, 10° less than in Fig. 3; b, same as in Fig. 3; c, more than in Fig. 3.


number of hazardous faults in this zone. Note that there iscontroversy about the causative fault of the Belovodsk earth-quakes, attributed previously to the Issyk-Ata fault (Chedia etal., 2000; Korjenkov and Nikonov, 2011) and then to theChonkurchak fault (Korzhenkov et al., 2012), by the sameauthors. Therefore, relating historic and prehistoric earth-quakes to specific faults is problematic.

We also predict high activity of the Issyk-Ata fault(Fig. 12a) with the assumption of a dip 10° lower than inFig. 3. The fault generated the Balasagun earthquake of 1475,of shaking intensity 8–9 (Korzhenkov et al., 2012), with itsepicenter in the Shamsi River catchment, in the eastern Chuvalley, at the northern slope of the Kyrgyz Range. The ShamsiTyundyuk border fault, inactive in the Holocene (Korzhenkovet al., 2012), was an unlikely candidate for being the cause ofthe earthquake. The activity of the Issyk-Ata and Ulunbulakfaults confirmed by our modeling (Fig. 11a) was also inferredby Abdarakhmatov et al. (2001b).

Although no long hazardous fault segments have beenrevealed in the northwestern Issyk Kul area shocked by theM = 6.9 earthquake of 1838 (Kalmetieva et al., 2009), thereis a tangle of faults comprising short (3–7 km) hazardoussegments (Fig. 12). Minor changes in stress or pore fluidpressure, which is currently low (Fig. 11), may trigger motionon these segments with variable dips.

Our predictions of hazard from a complex source withseveral fault planes in the northwestern Issyk Kul area are inline with the seismotectonic evidence (Korzhenkov et al.,2011) of at least four ~M =7 medieval earthquakes (in the 5th,latest 7th, 9th, and 14th centuries), with slip on two differentplanes in the event of the latest 7th century.

Our results indicate activity of the Kyzyloi fault, whichgenerated the M = 7.3 Suusamyr event of 1992 near thewestern border of the study area; the activity is higher in thesouth than in the north.

Thus, only 20–30 % of crustal faults appear to be activein the present stress field and only about 20 % of theirsegments are hazardous and prone to large-scale brittle failurein earthquakes. The fault pattern forms over tens or hundredsof million years, and only a small portion of these faultsaccommodates dissipated elastic energy accumulated duringmountain growth (Dobretsov et al., 2013).

Conclusions

In his studies of geological (tectonophysical) criteria ofseismic hazard, Gzovsky proceeded from stress and strain ofupper crust expressed in velocity gradients of tectonic motions,tectonic topography, and volcanism, which are commonly usedin assessment of seismotectonic activity (Reisner et al., 1993;Rogozhin, 2000; Ulomov, 2008), as well as stress and straintrends.

We develop the ideas of Gzovsky (Rebetsky, 2007b,c,2011) and suggest a pioneering approach for detection ofearthquake nucleation zones proceeding from stresses esti-mated on the basis of earthquake mechanism data. It consists

in relating fault planes with slip corresponding to the currentstress in rocks and in using the Coulomb rather than shearstress for estimating seismic hazard.

The approach stems from the modern views of the relationbetween stress and brittle failure which differ in many aspectsfrom those of the 1950–1960s. Now it has become clear thatshear strength of rocks depends on maximum shear stress andcohesion, as well as on friction. That is a reason whyhazardous seismic effects observed in nature may differ fromthe theoretical predictions by the method of Gzovsky.

Another possible reason for the discrepancy between thetheory and the practice is that the seismic process is unevenin time. The tectonic and geomorphic features (vertical sliprates) used in estimates of seismic risk were based on gradientsof crustal movements averaged over long times, which makesit impossible to judge whether the current activity corre-sponded to a period of slower or faster movements. Therefore,stress inferred from earthquake mechanisms represents theongoing phase of the geological process. The problem ofaveraged slip rates is resolved in our tectonophysical approachto seismic risk division of faults.

Note in conclusion that principal stress directions are usedfor analysis of seismotectonic and geodynamic settings, aswell as for explaining the mechanism of faulting (Rebetsky,2007b,c; Rebetsky and Alekseev, 2014; Rebetsky and Polets,2014; Rebetsky et al., 2012, 2013). According to the modernviews of the earthquake physics, reduced stresses have regularpatterns in nucleation zones of large earthquakes (Rebetsky,2007, 2009b; Rebetsky and Marinin, 2006a,b; Rebetsky andTatevossian, 2013). Thus, crustal stress has implications forthree basic problems in geosciences: faulting mechanisms,earthquake source physics, and seismic hazard.

The study was supported by grants 12-05-00234-a, 12-05-00550-a, and 13-05-00892a from the Russian Foundation forBasic Research and was carried out as part of Program 6 ofthe Geosciences Department of the Russian Academy ofSciences.

References

Abdrakhmatov, K.E., Thompson, S., Weldon, R., Delvaux, D., Klerkx, J.,2001a. Active faults of the Tien Shan. Nauka i Novye Tekhnologii, No. s2,22–28.

Abdarakhmatov, K.E., Weldon, R., Thompson, S.I., Burbank, D., Rubin, Ch.,Miller, M., Molnar, P., 2001b. Onset, style, and current rate of shorteningin the central Tien Shan (Kyrgyzstan). Geologiya and Geofizika (RussianGeology and Geophysics) 42 (10), 1585–1609 (1502–1526).

Bachmanov, D.M., Trifonov, V.G., Mikolaichuk, A.V., Vishnyakov, F.A.,Zarshchikov, A.A., 2008. The Minkush-Kekemeren zone of neotectonictranspression in the Central Tien Shan. Geotektonika, No. 3, 30–50.

Benioff, H., 1951. Earthquakes and rock creep. BSSA 41, 31–62.Bespalov, V.F. (Ed.), 1975. 1:1,500,000 Tectonic Map of Kazakhstan and

adjacent areas of USSR republics: 1971 [in Russian]. Aerogeologiya,Leningrad.

Bespalov, V.F., Grigoriev, A.V., Knauf, V.I. (Eds.), 1960–1973. 1:200,000Geological Map of the USSR, Series Northern Tien Shan [in Russian].Sheets K-43-IX, X, XI, XIV, XV, XVI, XVII, XXII. VSEGEI, Leningrad.

Bogdanovich, K.I., Kark, I.M., Korolkov, B.Ya., Mushketov, D.I., 1914. TheEarthquake of 22 December 1910 (4 January 1911) in the Northern Tien


Shan [in Russian]. Transactions, Geological Committee, New Series, Issue89, St. Petersburg.

Bott, M.H.P., 1959. The mechanics of oblique slip faulting. Geol. Mag. 96,109–117.

Brady, B., Bzown, E., 2004. Rock Mechanics for Underground Mining, thirdedition. Kluwer Academic Publishers.

Brown, W., Hoek, E., 1978. Technical Note. Trends in relationships betweenmeasured in-situ stress and depth. Int. J. Rock Mech. Min. Sci. 15, 211–215.

Bullen, K.E., 1955. On the size of the strained region prior to an extremeearthquake. BSSA 45, 43–43.

Byerlee, J.D., 1968. Brittle-ductile transition in rocks. J. Geophys. Res. 73(B14), 4741–4750.

Chedia, O.K. (Ed.), 1988. 1:500,000 Map of Neotectonics and NaturalResources of Kyrgyzstan [in Russian]. GUGK SSSR, Frunze.

Chedia, O.K., Abdarakhmatov, K.E., Lemzin, I.N., Mihel, G., Mikhalev, V.,2000. Issyk-Ata, Northern Tian-Shan fault in Holocene. J. EarthquakePred. Res. 8, 379–386.

Delvaux, D., Abdrakhmatov, K.E., Lemzin, I.N., Strom, A.L., 2001. Land-slides and surface breaks of the Ms = 8.2 Kemin earthquake of 1991,Kyrgyzstan. Geologiya i Geofizika (Russian Geology and Geophysics) 42(10), 1667–1677 (1583–1592).

Dinnik, A.N., 1926. Overburden pressure and round mine support design.Inzhenernyi Rabotnik, No. 3, 1–12.

Dobretsov, N.L., Polyansky, O.P., Reverdatto, V.V., Babichev, A.V., 2013.Dynamics of the Arctic and adjacent petroleum basins: a record of plumeand rifting activity. Russian Geology and Geophysics (Geologiya iGeofizika) 54 (8), 888–902 (1145–1161).

Dobrovolsky, I.P., 1991. The Theory of Earthquake Nucleation [in Russian].IFZ AN SSSR, Moscow.

Gamburtsev, G.A., 1955. Earthquake prediction: State of the art and prospects.Bull. Council on Seismology AS USSR, No. 1, pp. 7–14.

Gamburtsev, G.A., Belousov, V.V., 1960. Earthquake prediction, in: Gam-burtsev, G.A., Selected Works [in Russian], Izd. AN SSSR, Moscow,pp. 436–443.

Gzovsky, M.V., 1975. Fundamentals of Tectonophysics [in Russian]. Nauka,Moscow.

Herget, G., 1973. Variation of rock stresses with depth at Canadian iron mine.Int. J. Rock Mech. Min. Sci. 10, 37–51.

Igemberdiev, S.A., 1982 (Ed.). 1:500,000 Geological Map of the Kyrgyz SSR:1978 [in Russian]. VSEGEI, Leningrad.

Kalmetieva, Z.A., Mikolaichuk, A.V., Moldobekov, B.D., Meleshko, A.V.,Zhantaev, M.M., Zubovich, A.V., 2009. Earthquakes of Kyrgyzstan. AnAtlas [in Russian]. CAIIZ, Bishkek.

Karmaleeva, R.M., 2012. Strain monitoring at active tectonic sites, in: CurrentGeodynamics of Central Asia and Natural Hazard: Quantitative Results[in Russian], Proc. All-Russian Workshop and Training Seminar. Irkutsk,23–29 September 2012, Book 1, IZK, Irkutsk, pp. 86–89.

Korjenkov, A.M., 1999. Morphological structures and seismicity of theKochkor basin and bordering mountains (northern Tien Shan). Izv. NANKR, No. 3–4, 33–39.

Korjenkov, A.M., Nikonov, A.A., 2011. The Belovodsk earthquake of 1885,the first largest event in the northern Tien Shan (125th anniversary): Acomprehensive study, in: Geodynamics and Environment of ContinentalOrogens. Abstracts, 5th Int. Symp., Bishkek, 19–24 June, 2011. Book 2,NAC KR, Bishkek, pp. 45–52.

Korjenkov, A.M., Abdieva, S.V., Vakhrameeva, P.S., Djumabaeva, A.B.,Mamyrov, E.M., Morozova, E.A., Orlova, L.A., Fortuna, A.B., 2011.Strong historical earthquakes in the northwestern Issyk Kul’ basin(northern Tien Shan). Russian Geology and Geophysics (Geologiya iGeofizika) 52 (9), 1007–1015 (1276–1286).

Korzhenkov, A.M., Kolchenko, V.A., Rott, F.G., Abdieva, S.V., 2012. A largemedieval earthquake in the Chu basin, Kyrgyzstan. Geotektonika, No. 4,62–74.

Kuchai, V.K., 1969. Deformation in the epicentral area of the Kebinearthquake: Results of repeated surveys. Geologiya i Geofizika, No. 8,101–108.

Kuzikov, S.I., Mukhamediev, Sh.A., 2010. The field of current GPS velocitiesin the crust of Central Asia. Izv. RAN, Fizika Zemli, No. 7, 33–51.

Kuzmin, Yu.O., 2004. Modern geodynamics of fault zones. Izv. RAN, FizikaZemli, No. 10, 95–112.

Makarov, V.I., 1977. Neotectonic Structure of the Central Tien Shan [inRussian]. Nauka, Moscow.

Makarov, V.I. (Ed.), 2005. Modern Geodynamics of Intracontinental MountainGrowth (Central Asia) [in Russian]. Nauchnyi Mir, Moscow.

Mikolaichuk, A.V., Apayarov, F.Kh., 2004. Geological Map (Kyrgyz Rangeand Chu Basin), Kyrgyzstan. http:.www.kyrgyzstan.ethz.ch/fileadmin/download/geological_map.pdf.

Omuraliev, A., Omuraliev, M., Dzhumbaeva, A., 2009. Local 3D velocityheterogeneity and seismicity of the South Kochkor fault and the Kochkorearthquake of 2006 in the Tien Shan. Izv. NAN KR, No. 2, 32–46.

Potvin, Yu., Hadjigeorgiou, J., Sracey, D., 2007. Introduction, in: Challengeson Deep and High Stress Mining. Australian Centre for Geomechanics,The University of Western Australia, pp. 1–13.

Rebetsky, Yu.L., 1997. Reconstructions of tectonic stress and seismotectonicstrain: methods and current stress field of South-East Asia and Oceania.Dokl. Akad. Nauk 354 (1), 101–104.

Rebetsky, Yu.L., 2001. Principles of stress monitoring and the method ofcataclastic analysis of shears. Bull. MOIP, Ser. Geol. 76 (4), 28–35.

Rebetsky, Yu.L., 2003. Development of the method of cataclastic analysis ofshear fractures for tectonic stress estimation. Dokl. Earth Sci. 388 (1),72–76.

Rebetsky, Yu.L., 2005. Estimation of relative stress: Second stage of stressmodeling from fault slip data. Geofizicheskii Zhurnal 27 (1), 39–54.

Rebetsky, Yu.L., 2007a. Tectonic Stress and Rock Strength [in Russian].Akademkniga, Moscow.

Rebetsky, Yu.L., 2007b. Stressed state corresponding to the formation oflarge-scale brittle failure of rocks. Dokl. Earth Sci. 417 (8), 1216–1220.

Rebetsky, Yu.L., 2007c. New data on natural stress in zones of earthquakenucleation and an earthquake source model. Geofizicheskii Zhurnal 29(6), 92–110.

Rebetsky, Yu.L., 2009a. Stages II and IV of stress reconstructions by themethod of cataclastic analysis of shears. Geofizicheskii Zhurnal 31 (2),93–106.

Rebetsky, Yu.L., 2009b. Crustal stress in the Kuriles and Kamchatka beforethe Simushir earthquakes. Tihookeanskaya Geologiya 28 (5), 70–84.

Rebetsky, Yu.L., 2009c. Estimation of stress values in the method ofcataclastic analysis of shear fractures. Dokl. Earth Sci. 428 (7), 1202–1207.

Rebetsky, Yu.L., 2011. Seismic risk assessment by tectonophysical analysisof seismological data, in: Problems of Seismotectonics [in Russian], Proc.XVII Workshop with International Contributions. 20–22 September 2011,Moscow. IFZ RAN, Moscow, pp. 444–451.

Rebetsky, Yu.L., Alekseev, R.S., 2014. Current stress field in Central andSoutheastern Asia. Geodinamika i Tektonofizika 5 (1), 257–290, http://dx.doi.org/10.5800/GT-2014-5-1-0127.

Rebetsky, Yu.L., Marinin, A.V., 2006a. Stressed state of the Earth’s crust inthe western region of the Sunda subduction zone before the Sumatra–An-daman Earthquake on December 26, 2004. Dokl. Earth Sci. 407 (2),321–325.

Rebetsky, Yu.L., Marinin, A.V., 2006b. Preseismic stress field before theSumatra-Andaman earthquake of 26.12.2004: a model of metastable stateof rocks. Russian Geology and Geophysics (Geologiya i Geofizika) 47(11), 1173–1185 (1192–1206).

Rebetsky, Yu.L., Polets, A.Yu., 2014. Lithospheric stress in Japan before theTohoku earthquake of 11.03.2011. Geodinamika i Tektonofizika 5 (2),469–506, http://dx.doi.org/10.5800/GT-2014-5-2-0137.

Rebetsky, Yu.L., Sycheva, N.A., 2008. The present stress field of the CentralTien Shan, in: Correlation of Surface and Deep Structures, Proc. 14th Int.Conf., 27–31 October 2008, Petrozavodsk. KSC, Petrozavodsk, Book 2,pp. 146–150.

Rebetsky, Yu.L., Tatevossian, R.E., 2013. Rupture propagation in strongearthquake sources and tectonic stress field. Bull. Soc. Geol. Fr. 184,335–346.

Rebetsky, Yu.L., Sycheva, N.A., Kuchay, O.A., Tatevossian, R.E., 2012.Development of inversion methods on fault slip data. Stress state inorogens of the central Asia. Tectonophysics 581, 114–131, doi: 10.1016/j.tecto.2012.09.027.


Rebetsky, Yu.L., Kuchai, O.A., Marinin, A.V., 2013. Stress state anddeformation of the Earth crust in the Altai–Sayan mountain region.Russian Geology and Geophysics (Geologiya i Geofizika) 54 (2), 206–222(271–291).

Reisner, G.I., Ioganson, L.I., Reisner, M.G., Baranov, Yu.E., 1993. Types ofCrust and Current Geological Processes [in Russian]. IFZ, Moscow.

Rice, J.R., 1980. The mechanics of earthquake rupture, in: Dziewonski, A.M.,Boschi, E. (Eds.), Physics of the Earth’s Interior, Proc. Int. School ofPhysics “Enrico Fermi”; course 78, Soc. Italiana di Fisica, Bologna,pp. 555–649.

Rogozhin, E.A., 2000. Structure of source areas of large earthquakes in Russiain the latest 20th century, in: Natural Hazard in Russia [in Russian].Book 2. Seismic Hazard. Section 1.2, Kruk, Moscow, pp. 19–37.

Sankov, V.A., Parfeevets, A.V., Lukhnev, A.V., Miroshnichenko, A.I.,Ashurkov, S.V., 2011. Late Cenozoic geodynamics and mechanic couplingof crustal and upper mantle strain in the Mongolia–Siberia mobile zone.Geotektonika, No. 5, 1–18.

Savarensky, E.F., 1954. Seismicity of the USSR: results and prospects [inRussian], in: Trans. Geofiz. Inst. AN SSSR, Vol. 152, No. 25, pp. 5–24.

Shults, S.S., 1948. Neotectonics and Terrain of the Tien Shan [in Russian].Geografgiz, Moscow.

Sibson, R.H., 1974. Frictional constraints on thrust, wrench and normal faults.Nature 249, 542–544.

Sycheva, N.A., Kuzikov, S.I., 2012. Velocity models of the Tien Shanlithosphere and the method of double difference for earthquake locationfrom data of the KNET seismological network. Geofizicheskie Issle-dovaniya 13 (2), 5–22.

Sycheva, N.A., Aladiev, A.V., Mukhamadieva, V.A., Yunga, S.L., 2003.Earthquake mechanisms from data of the KNET seismological network,in: Geodynamics and Environment of Mountainous Region [in Russian].NS RAN, Bishkek, pp. 241–253.

Sycheva, N.A., Yunga, S.L., Bogomolov, L.M., Mukhamadieva, V.A., 2005.Seismotectonic crustal strain in the Northern Tien Shan (earthquake

mechanisms from data of the KNET seismological network). Izv. RAN,Fizika Zemli, No. 11, 62–78.

Terzaghi, K., 1943. Theoretical Soil Mechanics. Wiley, New York.Timofeev, V.Yu., Ardyukov, D.G., Boiko, E.V., 2009. Current crust move-

ments in Gorny Altai. Fizicheskaya Mehhanika 12 (1), 45–54.Timofeev, V.Yu., Kalish, E.N., Stus, Yu.F., Ardyukov, D.G., Arnautov, G.P.,

Smirnov, M.G., Timofeev, A.V., Nosov, D.A., Sizikov, I.S., Boiko, E.V.,Gribanova, E.I., 2013. Gravity variations and modern geodynamics of thesouthwestern Baikal region. Geodinamika i Tektonofizika 4 (2), 119–134.

Tursungaziev, B.T., Petrov, O.V., 2008. 1:500,000 Geological Map of theKyrgyzstan: 2008 [in Russian]. VSEGEI, St. Petersburg.

Tychkov, S.A., Kuchai, O.A., Bushenkova, N.A., Bragin, V.D., Kalme-tieva, Z.A., 2008. Current crustal deformation in the northern Tien Shan:GPS and seismological data. Russian Geology and Geophysics (Geologiyai Geofizika) 49 (4), 280–290 (367–381).

Ulomov, V.I., 2008. Detection of nucleating earthquakes and long-termprediction of large earthquakes in the northern Caucasus, in: Laverov, N.P.(Ed.), Global Change. Natural and Man-Caused Hazard. Book 1. SeismicProcesses and Catastrophes [in Russian], IFZ, Moscow, pp. 127–146.

Wallace, R.E., 1951. Geometry of shearing stress and relation to faulting.J. Geol. 59, 18–130.

Yunga, S.L., 1990. Studies of Seismotectonic Deformations: Methods andResults [in Russian]. Nauka, Moscow.

Zhukov, Yu.V. (Ed.), 1988. 1:500,000 Tectonic Map of Kyrgyzstan: 1988.VSEGEI, Leningrad.

Zubkov, A.V., Zoteev, O.V., Smirnov, O.Yu., Lipin, Ya.I., Khudyakov, S.V.,Krinitsyn, R.V., Selin, K.V., Ershov, A.A., Valiullov, L.R., 2010.Time-dependent patterns of crustal stress and strain in the Urals. Litosfera,No. 1, 84–93.

Zubovich, A.V., Makarov, V.I., Kuzikov, S.I., Mosienko, O.I., Shchelo-chkov, G.G., 2007. Intracontinental mountain growth in Central Asia:evidence from GPS measurements. Geotektonika, No. 1, 16–29.

Editorial responsibility: V.S. Seleznev


Active faults of the northern Tien Shan: tectonophysical zoning …€¦ · Introduction Detecting...

Documents

Transcript of Active faults of the northern Tien Shan: tectonophysical zoning …€¦ · Introduction Detecting...