Significant resistivity changes in the fault zone associated with the 1999 Chi-Chi earthquake, west-central Taiwan

The purpose of this paper is to demonstrate the possible influence of an earthquake on the resistivity distribution in a fault zone. We collected resistivity image profiles across a proposed fault trace prior to the 7.3 magnitude 1999 shallow earthquake in the Chi-Chi area of Taiwan. Significant positive resistivity anomalies were observed in the hanging wall after the earthquake. However, there were negligible resistivity changes in the footwall. From an examination of geoelectric phenomena and surface rupture in the fault zone, it is believed that geoelectric anomalies are associated with abrupt displacement along the active Chelungpu fault. This result indicates a potential for resistivity methods to provide a basis for the monitoring of an active fault.     

Co-seismic displacements, folding and shortening structures along the Chelungpu surface rupture zone occurred during the 1999 Chi-Chi (Taiwan) earthquake

A nearly 100-km-long surface rupture zone, called Chelungpu surface rupture zone, occurred mostly along the pre-existing Chelungpu fault on the northwestern side of Taiwan, accompanying the 1999 Chi-Chi Ms 7.6 earthquake. The Chelungpu surface rupture zone can be divided into four segments based on the characteristics of co-seismic displacements, geometry of the surface ruptures and geological structures. These segments generally show a right-step en echelon form and strike NE–SW to N–S, and dip to the east with angles ranging from 50 to 85°. The co-seismic flexural-slip folding structures commonly occurred in or near the surface rupture zone from a few meters to a few hundreds of meters in width, which have an orientation in fold axes parallel or oblique to the surface rupture zone. The displacements measured in the southern three segments are approximately 1.0–3.0 m horizontally and 2.0–4.0 m vertically. The largest displacements were measured in the northern segment, 11.1 m horizontally and 7.5 m vertically, respectively. The amount of co-seismic horizontal shortening caused by flexural-slip folding and reverse faulting in the surface rupture zone is generally less than 3 m. It is evident that the co-seismic displacements of the surface rupture zone are a quantitative surface indicator of the faulting process in the earthquake source fault. The relations between the geometry and geomorphology of the surface rupture zone, dips of the co-seismic faulting planes and the striations on the main fault planes generated during the co-seismic displacement, show that the Chelungpu surface rupture zone is a reverse fault zone with a large left-lateral component.     

Tomographic image of crustal structures across the Chelungpu fault: Is the seismogenic layer structure- or depth-dependent?

In 1999, a large earthquake (M_w = 7.6) occurred along the Chelungpu fault in the fold-and-thrust belt of western Taiwan. To shed more light on the subsurface structures and the seismogenic layers, three-dimensional velocity structures were inverted by using the travel times of both P- and S-waves from 2391 aftershocks recorded by the Central Weather Bureau during the 15 months that followed. From tomography, a typical image of the large-scale thrusting structures in the upper crust across the Chelungpu fault was obtained. In general, high velocities beneath the Western Foothills and Central Ranges are separated from low velocities beneath the Coastal Plain by an east-dipping boundary that is roughly consistent with the Chelungpu fault on the surface. The contrast in velocity on either side of the Chelungpu fault is indicative of about a 7- to 9-km vertical offset in the upper crust. The relocated hypocenter for the Chi-Chi earthquake shifts by 2.2 km toward the northwest, and its focal depth decreases by 0.7 km. A plot of focal depths versus rock velocities where the aftershocks occurred shows earthquakes are more inclined to occur in rock with a velocity of around 5.6 km/s. This strongly suggests the seismogenic layer in the fold-and-thrust belt of Taiwan is more structure-dependent than depth-dependent.     

Why does the co-seismic slip of the 1999 Chi-Chi (Taiwan) earthquake increase progressively northwestward on the plane of rupture?

The Chi-Chi 1999 earthquake ruptured the out-of-sequence Chelungpu Thrust Fault (CTF) in the fold-and-thrust belt in Western Central Taiwan. An important feature of this rupture is that the calculated slip increases approximately linearly in the SE–NW convergence plate direction from very little at its deeper edge to a maximum near the surface. We propose here a new explanation for this co-seismic slip distribution based on the study of both stress and displacement over the long-term as well as over a seismic cycle. Over the last 0.5 My, the convergence rate in the mountain front belt is accommodated by the frontal Changhua Fault (Ch.F), the CTF and the Shuangtung Fault (Sh.F). Based on previously published balanced cross sections, we estimate that the long-term slip of the Ch.F and of the CTF accommodate 5–30% and 30–55% of the convergence rate, respectively. This long-term partitioning of the convergence rate and the modeling of inter-seismic and post-seismic displacements suggest that the peculiar linear co-seismic slip distribution is accounted for by a combination of the effect of the obliquity of the CTF to the direction of inter-seismic loading, and of increasing aseismic creep on the deeper part of the Ch.F and CTF. Many previous interpretations of this slip distribution have been done including the effects of material properties, lubrication, site effect, fault geometry and dynamic waves. The importance of these processes with respect to the effects proposed here is still unknown. Taking into account the dip angle of the CTF, asperity dynamic models have been proposed to explain the general features of co-seismic slip distribution. In particular, recent works show the importance of heterogeneous spatial distribution of stress prior to the Chi-Chi earthquake. Our analysis of seismicity shows that previous large historic earthquakes cannot explain the amplitude of this heterogeneity. Based on our approach, we rather think that the high stress in the northern part of the CTF proposed by Oglesby and Day [Oglesby, D.D., Day, S.M., 2001. Fault geometry and the dynamics of the 1999 Chi-Chi (Taiwan) earthquake. Bull. Seismol. Soc. Am. 91, 1099–1111] reflects the latitudinal variation of inter-seismic coupling due to the obliquity of the CTF.     

Analysis of the Chiufengershan landslide triggered by the 1999 Chi-Chi earthquake in Taiwan

A disastrous earthquake rocked Taiwan on September 21, 1999, with magnitude M_L=7.3 and an epicenter near the small town of Chi-Chi in central Taiwan. The Chi-Chi earthquake triggered landslide on the dip slope at the Chiufengershan. In this study, a review of the topography and geology of this area was followed by field investigations. Laboratory testing was applied to understand the geomaterial composing the slope. Then, based on a series of limit equilibrium analyses, the failure mechanism of this landslide and the risk of the residual slope were studied.

According to the stability analyses, the pre-quake slope is quite stable, with factor of safety of 1.77 (dry) to 1.35 (full groundwater level); explaining why there is no written record of a landslide here for the past 100 years. In contrast, a back analysis shows that the Chi-Chi earthquake-induced dynamic loading is far more than the dip slope can sustain, due in part to the short distance to the epicenter. A Monte Carlo type probability analysis suggests that the residual slope is more dangerous than the pre-quake slope and needs more attention.     

Fluid infiltration associated with seismic faulting: Examining chemical and mineralogical compositions of fault rocks from the active Chelungpu fault

In order to understand the fault zone architecture and mechanisms that caused the Chi-Chi earthquake, the Chelungpu drilling project was conducted during April 2000 through a collaborative project between Japan and Taiwan. In this study, chemical and mineralogical variations within the overall Chelungpu fault zone, including variations between less damaged host rocks, damaged zones, and fault cores caused by the Chi-Chi earthquake were examined. Slopes of TiO₂ immobile isocons were consistently > 1 for analyses comparing host rocks with rocks from damaged zones or with gouges from fault cores, indicating that volume loss occurred in damaged zones and the fault cores. These results strongly imply that pervasive fluid infiltration occurred within the fault zone. Volume loss within the damaged zone and fault core is interpreted to result from a two-stage process involving: (i) coseismic mechanical wearing and/or dissolution in the fault core, and (ii) fluid infiltration within the fault zone during postseismic and interseismic periods along cracks caused by seismic failure. Semi-quantitative XRD analysis indicates that the kaolinite content consistently increases from the less damaged host rocks to the damaged zone and gouges in each fault core. Mineralogic changes indicate that pervasive acidic fluid infiltration occurred within the fault zones and reacted with the feldspars or muscovite to form kaolinite. Enrichment of kaolinite and illite found in the fault zones of southern drilling site could play some role on the slipping behavior of the southern part of the Chelungpu fault. Greater volume loss in the fault core may have resulted from moderate permeability, combined with the very fine grain nature of pulverized material in the fault core, which enhanced chemical reactions including transformation of feldspars and muscovite to clay minerals. The study results indicate that pervasive fluid infiltration occurred and changed the mineralogical and chemical architecture of fault zones caused by the cyclic earthquakes.     

Geological causes and geomorphological precursors of the Tsaoling landslide triggered by the 1999 Chi-Chi earthquake, Taiwan

The Tsaoling landslide, one of the largest landslide areas in Taiwan, has been affected by catastrophic events triggered by rain or earthquakes six times since 1862. These landslides, including that caused by the 1999 earthquake, have essentially not been reactivated old slides, but were sequential new ones that developed upslope, retrogressively. The landslide area is underlain by Pliocene sandstone and shale to form a dip slope with a bedding plane, dipping uniformly at 14°. The slip surface of the 1999 landslide was smooth and planar, parallel to the bedding plane with a slightly stepped profile; it formed within thinly alternated beds of fine sandstone and shale with ripple lamination or in a shale bed. The shale is weathered by slaking and probably by sulfuric acid, which is inferred to be one of the major causes of the intermittent retrogressive development of the landslides. The weathering was likely accelerated by the removal of overlying beds during earlier landslides in 1941 and 1942. The top margin of the 1999 landslide, in plan view, coincided with a V-shaped scarplet, which can be clearly recognized on aerial photographs taken before the landslide. This geomorphological feature indicates that this landslide had already moved slightly before its 1999 occurrence, providing precursory evidences.     

Seismic behavior in central Taiwan: Response to stress evolution following the 1999 Mw 7.6 Chi-Chi earthquake

Following the 1999 M_w 7.6 Chi-Chi earthquake, a large amount of seismicity occurred in the Nantou region of central Taiwan. Among the seismic activities, eight M_w ⩾ 5.8 earthquakes took place following the Chi-Chi earthquake, whereas only four earthquakes with comparable magnitudes took place from 1900 to 1998. Since the seismicity rate during the Chi-Chi postseismic period has never returned to the background level, such seismicity activation cannot simply be attributed to modified Omori’s Law decay. In this work, we attempted to associate seismic activities with stress evolution. Based on our work, it appears that the spatial distribution of the consequent seismicity can be associated with increasing coseismic stress. On the contrary, the stress changes imparted by the afterslip; lower crust–upper mantle viscoelastic relaxation; and sequent events resulted in a stress drop in most of the study region. Understanding seismogenic mechanisms in terms of stress evolution would be beneficial to seismic hazard mitigation.     

Detailed surface co-seismic displacement of the 1999 Chi-Chi earthquake in western Taiwan and implication of fault geometry in the shallow subsurface

The northern segment of the Chelungpu Fault shows an unusually large co-seismic displacement from the event of the Mw 7.6 Chi-Chi earthquake in western Taiwan. Part of the northern segment near the Fengyuan City provides an excellent opportunity for characterizing active thrust-related structures due to a dense geodetic-benchmark network. We reproduced co-seismic deformation patterns of a small segment of this Chelungpu Fault using 924 geodetic benchmarks. According to the estimated displacement vectors, we identified secondary deformations, such as local rigid-block rotation and significant shortening within the hanging wall. The data set also allows us to determine accurately a 3D model of the thrust fault geometry in the shallow subsurface by assuming simple relations between the fault slip, and the horizontal and vertical displacements at the surface. The predicted thrust geometry is in good agreement with borehole data derived from two drilling sites close to the study area. The successful prediction supports our assumptions of rigid displacement and control of displacement in the hanging wall by the fault geometry being useful first approximations.     

Ultrashallow seismic experiment on a trenched section of the Chelunpu fault zone, Taiwan

Ultrashallow P-wave seismic reflection experiments were conducted at a model test site and in a trenched shallow fault zone along the Chelunpu fault line. The field layout was designed to have the shallowest undistorted reflection from about 1 m depth with 0.5 m vertical resolution. The smallest group interval tested in this study was 0.05 m with a 0.25 ms sample interval, which can avoid spatial aliasing of ground roll if the target is very shallow and the velocities are low. Data processing was designed to be simple but consistent. As the ultrashallow reflections may be contaminated with high-amplitude coherent noise in many aspects, first break muting and surgical muting were performed on each file as detailed as possible, and f–k filtering was applied mainly for the purpose of attenuating the aliasing energy and back-scattered noise. Data acquired in this study show that the low P-wave velocities (< 200 m/s) and high dominant frequencies (120–200 Hz) of near-surface layers may have a potential vertical resolution of 0.4 m or even better.Comparing the test profile with the ground-penetrating radar (GPR) control profile of the same test site and correlating the results obtained from the study site with those of the geologic cross-section of the trench, this experiment demonstrates the possibility of using seismic methods in investigating shallow structures at depths of less than a few meters with vertical resolution comparable to the GPR technique.     

Imaging the ramp–décollement geometry of the Chelungpu fault using coseismic GPS displacements from the 1999 Chi-Chi, Taiwan earthquake

We use coseismic GPS data from the 1999 Chi-Chi, Taiwan earthquake to estimate the subsurface shape of the Chelungpu fault that ruptured during the earthquake. Studies prior to the earthquake suggest a ramp–décollement geometry for the Chelungpu fault, yet many finite source inversions using GPS and seismic data assume slip occurred on the down-dip extension of the Chelungpu ramp, rather than on a sub-horizontal décollement. We test whether slip occurred on the décollement or the down-dip extension of the ramp using well-established methods of inverting GPS data for geometry and slip on faults represented as elastic dislocations. We find that a significant portion of the coseismic slip did indeed occur on a sub-horizontal décollement located at 8 km depth. The slip on the décollement contributes 21% of the total modeled moment release. We estimate the fault geometry assuming several different models for the distribution of elastic properties in the earth: homogeneous, layered, and layered with lateral material contrast across the fault. It is shown, however, that heterogeneity has little influence on our estimated fault geometry. We also investigate several competing interpretations of deformation within the E/W trending rupture zone at the northern end of the 1999 ground ruptures. We demonstrate that the GPS data require a 22- to 35-km-long lateral ramp at the northern end, contradicting other investigations that propose deformation is concentrated within 10 km of the Chelungpu fault. Lastly, we propose a simple tectonic model for the development of the lateral ramp.     

Electrical structures of the source area of the 1999 Chi-Chi, Taiwan, earthquake: Spatial correlation between crustal conductors and aftershocks

Wide-band magnetotelluric (MT) data were collected on an east–west profile, approximately perpendicular to the local strike of the Chelungpu thrust, through the hypocentral area of the Chi-Chi earthquake for imaging the seismogenic structure. MT data were then inverted for two-dimensional resistivity models plus best-fitting static shift parameters using a nonlinear conjugate gradient algorithm that minimizes the sum of the normalized data misfits and the smoothness of the model. As shown in the inverted 2D resistivity models presented in this paper, an electrical conductor beside the hypocenter of the Chi-Chi earthquake indicates that deep-crustal fluids may participate in the rupture process of the Chi-Chi earthquake. A striking spatial correlation between the crustal conductor and occurrence of aftershocks beneath the Chelungpu fault suggests a postseismic pore pressure adjustment ongoing after the mainshock. Additionally, the hypocenter exhibits an electrical resistive zone, consistent very well with a predicted compact zone from a crustal deformation and transient fluid flow modeling.     

龙门山汶川-茂县断裂带的岩石磁学特征及其地震作用的指示意义

汶川-茂县断裂带是龙门山后山断裂,是松潘-甘孜褶皱带与龙门山断裂带之间的边界断层,然而,在2008年M_W 7.9级汶川大地震中并没有发生破裂。同时,汶川-茂县断裂带在汶川地震之前是否发生过历史大地震仍缺少确切的证据。这不仅制约着汶川地震发生机制的认识,而且还影响对龙门山形成演化过程的理解。因此,确定汶川-茂县断裂带的断裂作用环境对于认识龙门山断裂带的地震发生机制至关重要。断裂岩的岩石磁学可以有效地揭示断裂带的物理和化学环境。本文以汶川-茂县断裂带北部地表露头的断裂岩为研究对象,通过岩石磁学研究,并结合显微结构观察和地球化学分析,探讨汶川-茂县断裂带的断裂作用环境。断层泥和断层角砾岩最大磁化率值分别约为围岩的30倍和15倍,具有高磁化率值特征。断层泥的主要载磁矿物为磁铁矿、磁黄铁矿、针铁矿;断层角砾岩的主要载磁矿物为磁黄铁矿、针铁矿。断层泥的高磁化率值异常的主要原因是围岩中的含铁矿物在地震摩擦热和流体作用下新生成磁铁矿和磁黄铁矿。断层角砾岩的高磁化率值异常是围岩在含有大量硫元素的低温热液流体作用下生成了磁黄铁矿。大量针铁矿指示了震后期还原性的低温热液流体作用。断裂岩的高磁化率值异常和新生磁铁矿指示了汶川-茂县断裂带曾经发生了摩擦热温度>500℃的大地震活动,发震和震后期均为还原性的硫元素含量较高的低温热液流体环境。

     

四川汶川5.12大地震同震滑动断层泥的发现及意义

2008年汶川8.0级地震沿龙门山断裂带内的映秀—北川断裂和灌县—安县断裂产生了近300 km的同震地表破裂带。震后地质科学考察发现地表变形以逆冲为主,并伴有右旋走滑。地震地表破裂带大多沿古生代碳质泥岩、页岩和三叠系煤系地层内的滑动面出露地表,这些软弱地层为地震破裂带冲到地表提供了超低摩擦滑动带。我们发现在同震垂直和水平位错达6m左右的地表破裂带,地震的同震滑动发生在厚度约0.5~2cm 的狭窄滑动带内,以发育新鲜的灰色断层泥为特征,这些断层泥是地震断层快速滑动过程中岩石—流体相互作用的结果。     

四川汶川5.12大地震同震滑动断层泥的发现及构造意义

2008年汶川8．0级地震沿龙门山断裂带内的映秀-北川断裂和灌县-安县断裂产生了近300km的同震地表破裂带。震后地质科学考察发现地表变形以逆冲为主,并伴有右旋走滑。地震地表破裂带大多沿古生代碳质泥岩、页岩和三叠系煤系地层内的滑动面出露地表,这些软弱地层为地震破裂带冲到地表提供了超低摩擦滑动带。我们发现在同震垂直和水平位错达6m左右的地袁破裂带,地震的同震滑动发生在厚度约O．5—2cm的狭窄滑动带内,以发育新鲜的灰色断层泥为特征,这些断层泥是地震断层快速滑动过程中岩石-流体相互作用的结果。     

Fracture-zone conditions on a recently active fault: insights from mineralogical and geochemical analyses of the Hirabayashi NIED drill core on the Nojima fault, southwest Japan, which ruptured in the 1995 Kobe earthquake

An 1800-m-deep borehole into the Nojima fault zone was drilled at Nojima-Hirabayashi, Japan, after the 1995 Hyogo-ken Nanbu (Kobe) earthquake. Three possible fracture zones were detected at depths of about 1140, 1300, and 1800 m. To assess these fracture zones in this recently active fault, we analyzed the distributions of fault rocks, minerals, and chemical elements in these zones. The central fault plane in the shallowest fracture zone was identified by foliated blue-gray gouge at a depth of 1140 m. The degree of fracturing was evidently greater in the hanging wall than in the footwall. Minerals detected in this zone were quartz, orthoclase, plagioclase, and biotite, as in the parent rock (granodiorite), and also kaolinite, smectite, laumontite, stilbite, calcite, ankerite, and siderite, which are related to hydrothermal alteration. Biotite was absent in both the hanging wall and footwall across the central fault plane, but it was absent over a greater distance from the central fault plane in the hanging wall than in the footwall. Major element compositions across this zone suggested that hydrothermal alteration minerals such as kaolinite and smectite occurred across the central fault plane for a greater distance in the hanging wall than in the footwall. Similarly, H₂O+ and CO₂ had higher concentrations in the hanging wall than in the footwall. This asymmetrical distribution pattern is probably due to the greater degree of wall–rock fracturing and associated alteration in the hanging wall. We attributed the characteristics of this zone to fault activity and fluid–rock interactions. We analyzed the other fracture zones along this fault in the same way. In the fracture zone at about 1300 m depth, we detected the same kinds of hydrothermal alteration minerals as in the shallower zone, but they were in fewer samples. We detected relatively little H₂O+ and CO₂, and little evidence for movement of the major chemical elements, indicating little past fluid–rock interaction. In the fracture zone at about 1800 m depth, H₂O+ and CO₂ were very enriched throughout the interval, as in the fracture zone at about 1140 m depth. However, smectite was absent and chlorite was present, indicating the occurrence of chloritization, which requires a temperature of more than 200 °C. Only smectite can form under the present conditions in these fracture zones. The chloritization probably occurred in the past when the fracture zone was deeper than it is now. These observations suggest that among the three fracture zones, that at about 1140 m depth was the most activated at the time of the 1995 Hyogo-ken Nanbu (Kobe) earthquake.     

A fault zone cause of large amplification and damage in Avcılar (west of Istanbul) during 1999 Izmit earthquake

Twenty seven buildings collapsed and 2076 buildings were heavily damaged during 17 August 1999 Izmit Earthquake in Avc?lar (west of Istanbul), which is nearly 100 km from the epicenter of the earthquake. The geology of Avc?lar consists of Upper Miocene poorly bounded conglomerate and sandstone (Çukurçe?me formation), claystone with sandstone and limestone interbedding (Güngören member) and limestone with sandstone and claystone interbedding (Bak?rköy formation). Lithological and geotechnical parameters of these formations in Avc?lar are not different from non damaged parts of the western Istanbul such as Zeytinburnu, Bak?rköy and Beylikdüzü, but these formations were cut by several faults in the damaged area. Collapsed and damaged buildings are located on this fault zone. Thus, cause of large amplification and damage in Avc?lar might be related with this fault zone because the fault zone behaves as a waveguide trapping seismic energy.     

长诏断裂带垂直运动与地震关系初探

利用跨断层短水准资料分析了长诏断裂带断层活动特征,结果表明:长诏断裂带总体活动性质基本相同,分时段活动性质有所不同。断层各段多数以压性逆断为主,不同于历史地质考察结果。此外,长诏断裂带各断裂都具有分段活动特点。     

Geometry of the fault zone of the 2003 Ms = 7.5 Chuya earthquake and associated stress fields, Gorny Altai

The co-seismic deformations produced during the September 27, 2003 Chuya earthquake (Ms = 7.5) that affected the Gorny Altai, Russia, are described and discussed along a 30 km long segment. The co-seismic deformations have manifested themselves both in unconsolidated sediments as R- and R′-shears, extension fractures and contraction structures, and in bedrock as the reactivation of preexisting schistosity zones and individual fractures, as well as development of new ruptures and coarse crushing zones. It has been established that the pattern of earthquake ruptures represents a typical fault zone trending NW–SE with a width reaching 4–5 km and a dextral strike–slip kinematics. The initial stress field that produced the whole structural pattern of co-seismic deformations during the Chuya earthquake, is associated with a transcurrent regime with a NNW–SSE, almost N–S, trending of compressional stress axis (σ₁), and a ENE–WSW, almost E–W, trending of tensional stress axis (σ₃). The state of stress in the newly-formed fault zone is relatively uniform. The local stress variations are expressed in insignificant deviation of σ₁ from N–S to NW–SE or NE–SW, in short-term fluctuations of relative stress values in keeping their spatial orientations, or in a local increase of the plunge angle of the σ₁. The geometry of the fault zone associated with the Chuya earthquake has been compared with the mechanical model of fracturing in large continental fault zones with dextral strike–slip kinematics. It is apparent that the observed fracture pattern corresponds to the late disjunctive stage of faulting when the master fault is not fully developed but its segments are already clearly defined. It has been shown that fracturing in widely different rocks follows the common laws of the deformation of solid bodies, even close to the Earth surface, and with high rates of movements.     

Temporal variation in the geometry of a strike–slip fault zone: Examples from the Dead Sea Transform

The location of the active fault strands along the Dead Sea Transform fault zone (DST) changed through time. In the western margins of Dead Sea basin, the early activity began a few kilometers west of the preset shores and moved toward the center of the basin in four stages. Similar centerward migration of faulting is apparent in the Hula Valley north of the Sea of Galilee as well as in the Negev and the Sinai Peninsula. In the Arava Valley, seismic surveys reveal a series of buried inactive basins whereas the current active strand is on their eastern margins. In the central Arava the centerward migration of activity was followed by outward migration with Pleistocene faulting along NNE-trending faults nearly 50 km west of the center. Largely the faulting along the DST, which began in the early–middle Miocene over a wide zone of up to 50 km, became localized by the end of the Miocene. The subsidence of fault-controlled basins, which were active in the early stage, stopped at the end of the Miocene. Later during the Plio-Pleistocene new faults were formed in the Negev west of the main transform. They indicate that another cycle has begun with the widening of the fault zone. It is suggested that the localization of faulting goes on as long as there is no change in the stress field. The stresses change because the geometry of the plates must change as they move, and consequently the localization stage ends. The fault zone is rearranged, becomes wide, and a new localization stage begins as slip accumulates. It is hypothesized that alternating periods of widening and narrowing correlate to changes of the plate boundaries, manifest in different Euler poles.