Bedrock normal fault scarps, as classical topographic features and geomorphological markers along mountain range fronts, form in consolidated bedrock due to faulting in extensional settings. They generally preserve more complete records of paleo-earthquakes than fault scarps in unconsolidated sediments. With the development of technologies such as fault surface morphology measurement and terrestrial cosmogenic nuclide dating, bedrock fault planes have become a nice object for paleo-earthquake study in bedrock areas. The reconstruction of paleo-seismic history from a bedrock fault scarp in terms of the times, co-seismic slips and ages by a combination of quantitative morphological analysis, TCNs dating and other physical/chemical index has been proven feasible by several previous studies.

However, this success heavily relies on a suitable site selection along the bedrock fault scarp because erosional processes can exhume the bedrock fault surface, and the sedimentary processes can bury the bedrock fault surface. Namely, non-tectonic factors such as gully erosion, sediment burial, and anthropogenic activity make bedrock fault planes difficult to record and preserve paleo-seismic information.

Therefore, to successfully extract paleo-seismic information from the bedrock area, it is necessary to select suitable study points along the bedrock fault scarp in advance. Traditional survey and mapping methods are time-consuming and labor-intensive, and it is difficult to understand bedrock fault scarps. The resolution of satellite images cannot obtain the fine structure of bedrock fault scarps. Small unmanned aerial vehicle(sUAV), combined with Structure-from-Motion(SfM)photogrammetry has emerged over the last decade. It is used as an established workflow in acquiring topographic data by filling the spatial gap between traditional ground-based surveys and satellite remote sensing images. As a low-altitude photogrammetry technology, it can quickly obtain high-precision three-dimensional surface structures of bedrock fault scarps.

In this paper, taking the Majiayao bedrock fault scarp at the northern foot of Liulengshan in Shanxi Rift as an example, the high-precision and three-dimensional topographic data of the bedrock fault was obtained by using sUAV combined with SfM photogrammetry technology. The high-resolution and high-precision images of tectonic landforms can be obtained conveniently and efficiently by sUAV survey. The sUAV-obtained photos can be further processed by the SfM photogrammetry for generating a digital 3D structure of the bedrock fault scarp with true or shaded color.

The non-tectonic factors such as rock collapse, sediment burial, and gully erosion along the bedrock fault scarp are identified by interpreting the 3D model of the bedrock fault scarp. The profile shape characteristics of the erosion, burial and tectonic fault scarps are summarized through fine geomorphological interpretation and fault profile analysis. For the erosion profile, the hanging wall slope is down-concave, showing that the fault surface below the ground surface has been partially exposed. For the bury profile, the hanging wall slope shows an obvious concave-up shape, indicating that the lower part of the bedrock fault surface has been partially buried by the colluvium. For the tectonic profile, the hanging wall slope shows a smooth and stable slope, showing the exhumation of bedrock fault scarp is controlled purely by tectonics. Finally, the study sites suitable for paleo-earthquake study on bedrock fault surfaces were selected, showing the important role of sUAV aerial survey technology in the selection of paleo-earthquake study sites in bedrock areas.

This study illustrates that based on the high-precision three-dimensional surface structure of the bedrock fault plane from sUAV aerial survey, the existence of non-tectonic factors such as gully erosion, sedimentary burial and bedrock collapse can be clearly identified. These non-tectonic sites can be excluded when selecting suitable sites for paleo-earthquake study indoors. The shape analysis of bedrock fault scarp is also helpful to determine whether the bedrock fault surface is modified by surface process and suitable for paleo-seismic study. The sUAV aerial survey can play an important role in paleoseismic research in the bedrock area, which can accurately select the study points suitable for further paleo-seismic work in the bedrock area.

An M_S7.4 earthquake occurred in Madoi County, Guoluo Tibetan Autonomous Prefecture, Qinghai Province of China at 02:04 (Beijing Time) on May 22, 2021. A total of seven 800~3 000m trans-fault survey lines were targeted laid along different parts of the seismic surface rupture zone(the west, mid-west, mid-east, and the east), one month after the earthquake when the detailed field investigation of the coseismic displacement and the spread of the seismic surface rupture zone had been carried out. The soil gases were collected and the concentrations of Rn, H₂, Hg, and CO₂ were measured in situ.
The results show that the maximum value of Rn, H₂, Hg and CO₂ concentrations in different fracture sections of the surface rupture was 2.10~39.17kBq/m³(mean value: 14.15kBq/m³), 0.4×10^-6~720.4×10^-6(mean value: 24.93×10^-6), 4~169ng/m³(mean value: 30.72ng/m³)and 0.73%~4.04%(mean value: 0.59%), respectively. In general, the concentration of radon is low in the study area, which may be related to the thick overburden and the lithology dominated by sandstone. The concentration characteristics of hydrogen and mercury released from soil have good consistency, and the concentrations are higher at the east and west ends of the surface rupture zones but were lower in the middle of the rupture zone. This is consistent with the field investigation showing that the earthquake-induced surface rupture zone and deformation are more concentrated in the western section, while the eastern section has a large amount of seismic displacement.
The fault strikes at the east and west ends of the Madoi M_S7.4 earthquake surface rupture have deviated from the NW direction to a certain extent, and there also exits two branching faults and rupture complexities at the east end of the main fault of the Madoi earthquake. In the west end of the surface rupture, i.e., the south of Eling Lake, the fault strike turns to EW direction. We laid two survey lines(line 2 and line 3)at the west end of the rupture, the concentration of Rn, H₂ and Hg escaped from line 3 is the lowest one among all lines while the gas concentration of line 2 is significantly higher. In the vicinity of line 3, the field geological survey did not find the cracked and exposed surface rupture, and only a small number of liquefaction points were distributed near the Eling Lake. The soil gas concentrations and morphological characteristics were consistent with the field phenomena. At the east end of the rupture zone, the soil gas morphological characteristics of the south and north fault branches were inconsistent: the soil gas of the south branch showed a single-peak type which was more similar to that at the west end, but the gas concentration pattern of the north fault branch showed a multiple-peaks type. This phenomenon is consistent with the characteristic shown in the surface fracture mapping, that is, the deformation zone of the rupture where is wider.
To find out the source of soil gas and the possible influencing factors of soil gas concentrations in the study area, the carbon isotope and helium isotope of the collected gas samples were analyzed. The value of ³He/⁴He shows that the noble gas in the study area is mainly an atmospheric source, but the results of δ¹³C and CO₂/³He show that the soil gas along the surface rupture of the Madoi earthquake has the mixed characteristics of atmospheric components and crustal components, which to a certain extent reflects the cutting depth of main fault-Jiangcuo fault may be shallow, and it is speculated that the surface rupture caused by Madoi M_S7.4 earthquake may be confined to the shallow crust.

The strength properties of fault rocks at shearing rates spanning the transition from crystal-plastic flow to frictional slip play a central role in determining the distribution of crustal stress, strain, and seismicity in a tectonically active region. Since the end of the 20^th century, many experimental and modelling works have been conducted to elucidate the variation of the strength profile and mechanism of brittle-ductile transition(BDT)with temperature, pressure, and sliding rate. We review the substantial progress made in understanding the physical mechanisms involved in lithospheric deformation and refining constitutive equations that describe these processes. The main conclusions obtained from this study are as follows:

(1)The mechanical data and microstructure of friction and creep experiments indicated the transition from brittle to plastic deformation with the increasing crust depth, which not only controls the ultimate strength of the crustal profile but also limits the lower limit of the seismogenic zone. Moreover, based on the variation of rock characteristics, temperature, normal stress and sliding rate, the brittle-ductile transition zone distributes at different depths in the crust. The strength profile consisting of friction law and flow law is widely used to describe the strength and seismicity of the continental crust. However, this profile model is oversimplified in the BDT zone because this area involves a broad region of semi-brittle behavior in which cataclastic and ductile processes occur. At the same time, the model also lacks characterization of the transient dynamic properties of faults. Rate-and-state friction(RSF)law stipulates that the occurrence of slip instabilities(i.e. earthquake)can be linked with the velocity dependence of friction. Therefore, the RSF equations, when applied to the kilometer-scale of fault zones, models incorporation RSF equations can reproduce several important seismological observations, including earthquake nucleation and rupture, earthquake afterslip, and aftershock duration. However, these key microphysical processes of fault gouge evolution are unknown to this model.

(2)During numerical model-fitting experimental observations, the Friction-to-flow constitutive law merges crustal strength profiles of the lithosphere and rate dependency fault models used for earthquake modelling on a unified basis, which is better than controlling the boundary of BDT using the Mohr-Coulomb criterion, Von Mises criterion and Goetze’s criterion. The Friction-to-flow constitutive law can predict the steady-state and transient behavior of the fault, including the response of shear stress, sliding rate, normal stress, and temperature, in addition to simulating the transition of fault sliding stability from velocity-weakening to velocity-strengthening. It also solved seismic cycles of a fault across the lithosphere with the law using a 2-D spectral boundary integral equation method, revealing dynamic rupture extending into the aseismic zone and rich evolution of interseismic creep, including slow slip before earthquakes. However, these constitutive models do not base on microphysical behavior. Furthermore, at low to intermediate temperatures, the ductile rheology of most crystalline materials are different from those at high temperatures.

(3)A recent microphysical model, which treats fault rock deformation as controlled by competition between rate-sensitive(diffusional or crystal-plastic)deformation of individual grains and rate-insensitive sliding interactions between grains(granular flow), predicts both transitions well, called the CNS model. Unlike the numerical model, this model quantitatively reproduces a wide range of(transition)frictional behaviors using input parameters with direct physical meaning, which is closer to the natural strength of the fault. This mechanism-based model can reproduce RSF-like behavior in microstructurally verifiable processes and state variables. However, the major challenge in the CNS model lies in capturing the dynamics of micro- and nanostructure formation in sheared fault rock and considering the different processes of rock deformation mechanisms.

Since it is microphysically based, we believe the modelling approach can provide an improved framework for extrapolating friction data to natural conditions.

14 survey lines, with a total of 167 measuring points, were laid out in the northern section of the Red River Fault, the Longpan-Qiaohou Fault, the Heqing-Eryuan Fault, the eastern piedmont fault of Yulong Mountains, and the Lijiang-Jianchuan Fault in the northwest of Yunnan Province, China. Cross-fault soil gas radon, hydrogen, and carbon dioxide have been measured on the above-mentioned faults. The concentration intensity and distribution characteristics of soil gas in the study area were calculated and analyzed. The results show that:
(1)The concentrations and distribution patterns of soil gas radon and hydrogen vary greatly in different faults. The concentrations of radon vary from 6.18Bq/L to 168.32Bq/L, while that of hydrogen are between 7.72ppm to 429ppm, and carbon dioxide are from 0.73% to 4.04%.
(2)The average results of soil gas measurement show that the concentrations of radon are higher than 40kBq/m³ in the sampling sites of Yinjie, Niujie, Gantangzi, while the concentrations of radon in En’nu and Tiger Leaping Gorge measuring lines are smaller; The concentrations of hydrogen are higher than 60ppm in the sampling sites of Yangwang village, Houqing, Dawa, Yangcaoqing and Tiger Leaping Gorge, while the concentrations in Gantangzi and Niujie measuring lines are smaller.
(3)The spatial distribution characteristics of soil gas concentration in faults in northwest Yunnan are obvious, and the intensity of radon and hydrogen concentrations in different active fault zones vary greatly. The intensities of radon and hydrogen concentration are higher and have good consistency in Yinjie and Yangwang village measuring lines located in the northern section of the Red River Fault, the Houqing survey line located in Longpan-Qiaohou Fault, Dawa survey line in the Lijiang-Jianchuan Fault and Yangcaoqing in the south of Chenghai Fault. The soil gas concentration in such sample sites is high and the degassing ability is strong, indicating the different activity characteristics of different segments of the above faults to some extent.
Under the action of tectonic stress, the fault will slip and the rock properties and material structure of the fault will change, thus causing changes of underground material, gas transport channel and transport mode, which is characterized by the change of the concentration and distribution characteristics of escaped soil gas. Combined with the active characteristics of faults, slip rate and geomorphological features, the characteristics of concentration and spatial distribution of two soil gases(radon and hydrogen)are discussed, and the following conclusions are obtained.
(1)There is a large difference in the concentration of escaped soil gas from different faults in the study area, indicating that the content of soil gas is controlled by regional geochemical background values, and there are certain differences in the gas concentration of different sections of the same fault, indicating that the local concentration/flux change is greatly affected by the transport.
(2)The concentration of fault soil gas is related to fault activity, and for different faults, the higher the degree of fault activity is, the higher the concentration of fault gas will be. From the point of fault gas concentration characteristics, the concentrations in the survey lines in the northern section of the Red River Fault and Heqing-Eryuan Fault in the study area vary greatly, suggesting that the fault segmentation is obvious. Compared with other faults, the northern section of the Red River Fault has a higher concentration of soil gas, indicating that the fault is more active. However, there is no simple linear relationship between the soil gas concentration and the fault slip rate, and it may also depend on its material source, transport channel structure, etc.
(3)The concentration of soil hydrogen at the outcrop of faults(especially normal faults and strike-slip faults)is generally higher, which shows that hydrogen has better indicative significance in revealing the location of fault rupture, and the distribution pattern of soil gas radon concentration is a good indicator for analyzing the characteristics of fault movement.

An earthquake of M6½ occurred near Fushan County in the 9th year of Dading Period of the Jin Dynasty(in 1209), which caused a large number of casualties and property losses. Many experts and scholars speculated that the Fushan Fault might be its seismogenic structure, but no in-depth research has been conducted, which greatly hinders the development of earthquake prevention and disaster reduction in the region. The Fushan Fault is located on the east side of the Linfen fault basin in the Shanxi fault depression zone. It is a boundary fault between the Linfen fault basin and the uplift area of the Taihangshan block. Predecessors have done little research on the Fushan Fault. This paper carries out a quantitative study on the late Quaternary activity and displacement rate of the Fushan Fault. First, we carried out remote sensing interpretation, fault surface excavation, collection and testing of fault geomorphological samples in the area of Qianjiao village of Fushan Fault. It is determined that the Fushan Fault starts from Hanzhuang village, Beihan Town in the north, extends to the southwest through Yushipo village, Fenghuangling village, Baozishang village, Zhaojiapo village in Beiwang town, Nanwang village, Zhuge village, Qianjiao village, Guojiapo village, Qiaojiapo village in Tiantan town, Dongguopo village and Zhaishang village in Zhangzhuang town, Lijiatu village and Zhujiashan village in Xiangshuihe town, and terminates in Chejiazhuang village in Xiangshuihe town, with a total length of 24km. The formation age of geomorphological bodies was obtained. It is determined that the latest stratum dislocation event of the fault is later than 7ka, and the fault is a Holocene active fault and has the ability to generate earthquakes of magnitude 7 and above. A total of two phases of stratum dislocation events have occurred on the Fushan Fault since 17ka BP(Late Quaternary): The first-phase event E1 occurred between 17ka and 7ka BP, producing a displacement of 2.04m, the average displacement rate of the Fushan Fault is 0.20mm/a; the second-phase event E2 occurred since 7ka BP, producing a displacement of 3.93m, and the average displacement rate of the Fushan Fault is 0.56mm/a. The displacement rate of the fault has been increasing since the Late Pleistocene. The future seismic hazard of this fault is worthy of attention. This paper also uses land-based LiDAR scanning to obtain the topographic data of the fault plane on the Qiaojiapo village bedrock section of the Fushan Fault(4.5km away from the Qianjiao village section). The isotropic variogram method was used to calculate the fractal dimensions of the fault surface morphology, and the morphological weathering zone was divided, and two phases of ancient seismic events of the Fushan Fault since the Late Quaternary were determined, which are, from old to new, the first-phase event E1 which caused a co-seismic displacement of 3.18m, and the second-phase event E2 which caused a co-seismic displacement of 2.51m. Studies have shown that the bedrock fault plane fractal method is an effective method for studying ancient seismic events in the bedrock area, and its ancient seismic period division is consistent with that of the sedimentary coverage area. Finally, this paper discusses the seismogenic structure of the 1209 Fushan earthquake with magnitude of 6½, and believes that the seismogenic structure of the Fushan earthquake is most likely to be the Fushan Fault. However, due to the lack of a lower age limit and that the only upper limit age is far away from the historical earthquake time, it is necessary to conduct a more detailed investigation and research on the fault to determine whether there can be a revelation of ancient earthquake events with a younger age and comparable magnitude.
This study has greatly improved Fushan County’s risk prevention and control, and territorial planning capabilities.

Many of the large earthquakes in the continental crust nucleate at the bottom of the seismogenic zone in depths between 10 and 20km which is related to the broad so-called ‘brittle-to-plastic or brittle-to-ductile’ transition region. From the field studies and seismic data, we could know that the dominant deformation mechanism at the base of seismogenic zone is likely to be semi-brittle flow of fault rocks. The physical and chemical processes acting in the ‘brittle-to-plastic’ transition are of great interest for a better understanding of fault rheology, tectonic deformation of the continental lithosphere and the generation of strong earthquakes. So it’s of great significance to know more about this transition. Despite the importance of semi-brittle flow, only few experimental studies are relevant to semi-brittle flow in natural rocks. In order to study the semi-brittle deformation and rheological characteristics of granite, we performed a series of transient creep experiments on fine-grained granite collected from the representative rock of Pengguan Complex in Wenchuan earthquake fault area using a solid-medium triaxial deformation apparatus(a modified Griggs rig). The conditions of the experiments are under the temperatures of 190~490℃and the confining pressures of 250~750MPa with a strain rate of 5×10^-4s^-1. The temperature and pressure simulate the in-situ conditions of the Wenchuan earthquake fault zone at the corresponding depths of 10~30km. We observe the microstructures of the experimentally deformed samples under the scanning electron microscope(SEM). The mechanical data, microstructures and deformation mechanism analysis demonstrate that deformation of the samples with experimental conditions could be covered by three regimes: 1)Brittle fracture to semi-brittle flow regime. We could see the strain and stress curves of the samples characterizing with strain hardening behavior and without definite yield point under low temperatures and pressures, which correspond to the depths of 10~15km; 2)Brittle-ductile transition regime. The strain and stress curves of the samples tend to be in a steady state with definite yield point under temperature and pressure at the depths of 15~20km. The main deformation mechanism is cataclasis, and dynamic recrystallization and dislocation creep are activated; and 3)Ductile flow regime which is at depths of 20~30km. The strength of granite increases with depth and reaches to the ultimate at the depth of 15~20km, and then decreases with depth at 20~30km. Based on the analysis of strength of granite, microstructures and deformation mechanism, we conclude that the granitic samples deformed with the characteristics of transient creep, and the strength of Longmenshan fault zone reaches maximum at the depths of 15~20km where it is in the brittle-to-plastic regime. Based on the Mohr circle analysis, the rupture limit at depths of 15~20km is close to the limit of friction, and at the same time, this depth range is also consistent with the focal depth of Wenchuan earthquake. Therefore, it implicates that the deformation and strength of Pengguan complex granitic rocks should control the nucleation and generation of the Wenchuan earthquake.

Field studies and seismic data show that semi-brittle flow of fault rocks probably is the dominant deformation mechanism at the base of the seismogenic zone at the so-called frictional-plastic transition. As the bottom of seismogenic fault, the dynamic characteristics of the frictional-plastic transition zone and plastic zone are very important for the seismogenic fault during seismic cycles. Granite is the major composition of the crust in the brittle-plastic transition zone. Compared to calcite, quartz, plagioclase, pyroxene and olivine, the rheologic data of K-feldspar is scarce. Previous deformation studies of granite performed on a quartz-plagioclase aggregate revealed that the deformation strength of granite was similar with quartz. In the brittle-plastic transition zone, the deformation characteristics of granite are very complex, temperature of brittle-plastic transition of quartz is much lower than that of feldspar under both natural deformation condition and lab deformation condition. In the mylonite deformed under the middle crust deformation condition, quartz grains are elongated or fine-grained via dislocation creep, dynamic recrystallization and superplastic flow, plagioclase grains are fine-grained by bugling recrystallization, K-feldspar are fine-grained by micro-fractures. Recently, both field and experimental studies presented that the strength of K-feldspar is much higher than that of quartz and plagioclase. The same deformation mechanism of K-feldspar and plagioclase occurred under different temperature and pressure conditions, these conditions of K-feldspar are higher than plagioclase. The strength of granite is similar to feldspar while it contains a high content of K-feldspar. High shear strain experiment studies reveal that granite is deformed by local ductile shear zones in the brittle-plastic transition zone. In the ductile shear zone, K-feldspar is brittle fractured, plagioclase are bugling and sub-grain rotation re-crystallized, and quartz grains are plastic elongated. These local shear zones are altered to local slip-zones with strain increasing. Abundances of K-feldspar, plagioclase and mica are higher in the slip-zones than that in other portions of the samples (K-feldspar is the highest), and abundance of quartz is decreased. Amorphous material is easily formed by shear strain acting on brittle fine-grained K-feldspar and re-crystallized mica and plagioclase. Ductile shear zone is the major deformation mechanism of fault zones in the brittle-plastic transition zone. There is a model of a fault failed by bearing constant shear strain in the transition zone: local shear zones are formed along the fractured K-feldspar grains; plagioclase and quartz are fine-grained by recrystallization, K-feldspar is crushed into fine grains, these small grains and mica grains partially change to amorphous material, local slip-zones are generated by these small grains and the amorphous materials; then, the fault should be failed via two ways, 1)the local slip-zones contact to a throughout slip-zone in the center of the fault zone, the fault is failed along this slip-zone, and 2)the local slip-zones lead to bigger mineral grains that are in contact with each other, stress is concentrated between these big grains, the fault is failed by these big grains that are fractured. Thus, the real deformation character of the granite can’t be revealed by studies performing on a quartz-plagioclase aggregate. This paper reports the different deformation characters between K-feldspar, plagioclase and quartz under the same pressure and temperature condition based on previous studies. Then, we discuss a mode of instability of a fault zone in the brittle-plastic transition zone. It is still unclear that how many contents of weak mineral phase(or strong mineral phase)will control the strength of a three-mineral-phase granite. Rheological character of K-feldspar is very important for study of the deformation characteristic of the granitic rocks.

Paleo-seismic and fault activity are hard to distinguish in host rock areas compared with soft sedimentary segments of fault. However, fault frictional experiments could obtain the conditions of stable and unstable slide, as well as the microstructures of fault gouge, which offer some identification marks between stick-slip and creep of fault.
We summarized geological and rock mechanical distinction evidence between stick-slip and creep in host rock segments of fault, and analyzed the physical mechanisms which controlled the behavior of stick-slip and creep. The chemical composition of fault gouge is most important to control stick-slip and creep. Gouge composed by weak minerals, such as clay mineral, has velocity weakening behavior, which causes stable slide of fault. Gouge with rock-forming minerals, such as calcite, quartz, feldspar, pyroxene, has stick-slip behavior under condition of focal depth. To the gouge with same chemical composition, the deformation mechanism controls the frictional slip. It is essential condition to stick slip for brittle fracture companied by dilatation, but creep is controlled by compaction and cataclasis as well as ductile shear with foliation and small fold. However, under fluid conditions, pressure solution which healed the fractures and caused strength recovery of fault, is the original reason of unstable slide, and also resulted in locking of fault with high pore pressure in core of fault zone. Contrast with that, rock-forming minerals altered to phyllosilicates in the gouges by fluid flow through degenerative reaction and hydrolysis reaction, which produced low friction fault and transformations to creep. The creep process progressively developed several wide shear zones including of R, Y, T, P shear plane that comprise gouge zones embedded into wide damage zones, which caused small earthquake distributed along wide fault zones with focal mechanism covered by normal fault, strike-slip fault and reverse fault. However, the stick-slip produced mirror-like slide surfaces with very narrow gouges along R shear plane and Y shear plane, which caused small earthquake distributed along narrow fault zones with single kind of focal mechanism.

The transition from microscopic brittle deformation to microscopic plastic deformation is called brittle-plastic transition, which is considered as a key layer for determining the limit of lower continental crust seismicity. The depth and deformation mechanism of the brittle-plastic transition zone is controlled mainly by temperature. Besides, the strain rate and fluid pore pressure also affect the transition during the different deformation stages at the seismic cycle.
In this paper, microstructure observation of catalcastic samples collected from the Red River Fault was carried out using optical polarized microscopy and scanning electron microscopy. The morphology, microstructures of deformation characteristics, mineral composition, water-rock reaction, pressure solution, exsolution, crack healing in the samples were systematically observed. The mineral components quantitative analyses were examined using the EDS. Water-rock reaction and pressure solution were systematically observed under SEM. The fabric of the main minerals in the samples was measured using electron backscattered diffraction(EBSD). Based on these analyses, the deformation mode was setup for the brittle-plastic transition zone of the fault during the post-seismic relaxation period.
Both brittle deformation and plastic deformation were developed in the cataclastic samples. EBSD data shows that the c axial fabrics of quartz present low-temperature plastic deformation characteristics. The feldspar deformed as cataclastic rock, and the micro-fracture in feldspar was healed by static recrystallized quartz and calcite veins. The calcite vein underwent plastic deformation, which represents the post-seismic relaxation deformation.
Based on the analysis of deformation mechanism of cataclastic samples in brittle-plastic transition zone of the Red River Fault, and combined with previous studies, we concluded that the brittle fracture and fracture healing is the main deformation mode at brittle-plastic transition zone in the post-seismic relaxation. High stress and high strain rate at post-seismic relaxation lead to brittle fracture of high-strength minerals such as feldspar in rocks. Plastic deformation occurs in low-strength minerals such as quartz and mica. Under the fluid condition, micro-fractures were healed by quartz and calcite. The minerals such as quartz and calcite in the fracture transformed from static recrystallization to dynamic recrystallization with stress gradually accumulating. With fracture healing and stress accumulation, the fault strength gradually increases which could accumulate energy for the next earthquake.

A magnitude 7(3/4) earthquake happened in Linfen, Shanxi, on May 18, 1965(the 34th year of Qing Emperor Kangxi). In the Catalogue of Chinese Historical Strong Earthquakes, the epicenter of this earthquake is located at the northwest of Zhangli Village of Xiangfen County and Dongkang Village of Yaodu District, Linfen City(36.0°N, 111.5°E), and the epicentral intensity is Ⅹ. It was inferred by previous studies that Guojiazhuang Fault is the seismogenic structure of the earthquake. In this paper, in cooperation with the Archives of Linfen City and Earthquake Administration of Linfen, the author looked up in details the first-hand materials of the earthquake damage to the ancient town of Linfen and its surrounding areas, and based on this, drew the isoseismals of the earthquake. Through discussions with relevant experts, we consider that it would be more appropriate that the location of the macroscopic epicenter of this earthquake is in Donguan area of the ancient town of Linfen, the epicentral intensity is Ⅺ, and the major axis of the isoseismals is in NWW. Later, in the implementation of "Linfen city active fault detection and seismic risk evaluation", we found two earthquake fault outcrops near the macroscopic epicentral area of the 1695 Linfen earthquake. Shallow seismic exploration lines and drill rows perpendicular to the strike of the fault outcrops were arranged to implement the exploration. The results demonstrate that the right-lateral stepover composed of Guojiazhuang Fault and Liucun Fault, together with the Luoyunshan Fault(Longci segment), were involved in the 1695 Linfen earthquake, the intersection of the faults is the microscopic epicenter of the earthquake, and the above-mentioned three faults are the seismogenic structure of the earthquake. In addition, the seismic geological remains in this region(landslides, earthquake ground cracks, sand emitting channels, etc.) are mainly distributed on the hanging wall of the Guojiazhuang Fault, this proves from another perspective that the earthquake remains is the product of activity of Guojiazhuang Fault in 1695.

We performed deformation experiments using Carrara marble in dry and wet conditions under temperature of 400~700℃ and confining pressure 300MPa with two different strain rates. Water contents of deformed samples were measured using FTIR spectroscopy. The microstructure and deformation mechanisms of samples were observed under optical microscopy, scanning electron microscopy and energy spectroscopy analysis. The mechanical data show that samples display strain hardening at 400℃, and transition to steady creep at temperature from 500~700℃. The strength of marble reduced gradually with elevated temperatures or decreased strain rate. However, water effect to the strength of the marble is significantly weak. Microstructures observed show that the deformation is cataclastic flow in dry samples, fracture and pressure solution in wet samples at 400℃. Samples underwent brittle-plastic transition at 500℃. Dislocation glide is major deformation mechanism for dry samples at 600℃. Dislocation climb and dynamic recrystallization are major deformation mechanism for wet samples at 600℃ and for all wet samples and dry samples at 700℃. Lower strain rate and higher water content could promote the process of pressure solution and diffusion as well as dynamic recrystallization.

In this paper, rheological experiments are carried out on natural hornblende under high temperature and high pressure. We used polarizing microscope and scanning electron microscope to analyze the experimental samples' microstructure, investigating the mechanisms of hornblende under the condition of different temperature, pressure and strain rate. The experimental results reveal the following features of the stress-strain curves of deformed samples: As the temperature increases, the stress-strain curve of the samples changes gradually from strengthening to yielding and weakening, sample strength reduces significantly; with the increase of confining pressure, the sample strength increases; and with the decrease of strain rate, the sample strength reduces, and it significantly reduces in the samples with the compression direction heterotropic to foliation. Plenty of transgranular fractures as well as a small amount of cataclastic deformation occur in hornblendite at temperature of 500℃, and the deformed sample is dominated by brittle deformation. At temperature of 600℃, porphyroclast system consisting of residual plaques and cataclastic series grows in the samples, wavy extinction appears in part of hornblende crystals, the deformation is characterized mainly by cataclastic deformation with ductile deformation, locally. At temperature of 700℃, the deformation is mainly dominated by intragranular kink, and dehydration and fine-grained microcrystalline appear locally, containing microcracks. The deformation of the sample is in the brittle-plastic transition phase; At temperature of 800℃, almost no obvious brittle deformation is observed in the deformed samples, the samples are dominated by dynamic recrystallization, and dehydration appears. Therefore, at the temperature conditions of 500℃, 600℃, 700℃ and 800℃, the deformation of hornblende is characterized by brittle fracture, cataclastic flow, crystal kink, dynamic recrystallization and dehydration, which shows the deformation mechanism varying from brittle to brittle-ductile, and to ductile deformation.

Deformed fabric has been broadly observed from middle to lower crustal rocks. The deformed fabric of granitic rock not only affects the strength of rock, but also controls the later deformation, which transforms the former deformation. The experimental exploration on effect of pre-exsiting fabric on the rheological strength of rocks is thus a new research topic. In this paper, we re-analysed the semibrittle-plastic creep data of anisotropic rocks(mica schist-gneiss, synthetic layers and particulate quartz-anorthite (50:50) composites), combined with the rheological experimental results of granitic gneiss and mylonite by the authors under different fabric conditions and discussed the effect of pre-exsiting fabric on rheological strength of anisotropic rocks. The experimental results show that the angle between the compression direction and the original foliation of anisotropic rocks is an important factor, which controls the variation of strength. In the semi-brittle deformation regime, the flow strength of rocks with compression direction perpendicular (PER) to the foliation is basically similar with that of rocks with compression direction parallel (PAR) to the foliation. The fracture strength is minimum when the angle between the compression direction of experiment and the original foliation is 30 degrees. In the plastic deformation regime, the flow strength of rocks with PER is significantly stronger than that of rocks with PAR. The strength has a minimum value when the angle between the foliation and the orientiation of the maximum principal stress is 45 degrees. The degree of replacement of former deformed fabric by that of later deformation determines the strength of anisotropic rocks. Besides, the content, distridution and grain size of minerals have a significant effect on the strength of anisotropic rocks. Prediction by previous theories is consistent with experimental results of mica schist, but the rheological model of other types of anisotropic rock seems more complex than such theoretical models. Therefore, further investigation on the effect of pre-existing fabric on rheology of rock with comparison of the deformation between geological and experimental conditions is the most effective method to understand the rheology and deformation mechanism of anisotropic rocks in tectonic settings.

The seismogenic fault of Wenchuan earthquake is a high-angle reverse-slip fault. It is necessary for the sliding of such a high-angle reverse fault and the triggering of the Wenchuan earthquake on it to have special mechanical conditions at the deep part of fault. In this study, we investigated the deformation mechanism of cataclastic-mylonite rocks in ductile shear zones found in the Yingxiu-Beichuan Fault. The deformation temperature and the flow stress of brittle-plastic transition of fault were estimated by the deformation fabrics of quartz. The water contents and the distribution of major minerals in mylonite were measured using Fourier transform infrared spectroscopy(FTIR). The fluid inclusions were measured using Raman and microprobe. The rehological structures of brittle-plastic transition of the Longmenshan Fault zone under different fluid pressure and strain rate conditions were constructed to discuss the role of the high fluid pressure in the seismogenic and occurrence mechanics of Wenchuan earthquake.
The studies showed that inhomogeneous ductile deformation occurred in the brittle-plastic transition of the Yingxiu-Beichuan Fault. The complex deformation characters of quartz display different deformation temperatures in the ductile shear zone. The quartz in fine-grained mylonite was deformed by the grain boundary migration and recrystallization, implying the deformation temperature is from 500 to 700℃. The quartz in porphyroclastic mylonite was deformed by the subgrain rotation and recrystallization, implying the deformation temperature is from 400 to 500℃. The earlier quartz veins and healed cracks were deformed by the bulges and recrystallization, implying the deformation temperature is from 280 to 400℃. The later quartz veins which cut the earlier quartz veins were deformed by the cataclastics, indicating the deformation temperature is from 150 to 250℃. The deformation of quartz shows that the ductile shear zone experienced multi-phase brittle-ductile transitions. Based on the grain size of recrystallized quartz, the ductile flow stress of the fault is estimated to be 15～80MPa. The trace amount water in quartz and feldspar exists in the forms of hydroxyl in crystals, grain boundaries water and fluid inclusions water, and the water contents are higher with increasing strain of rocks, with a changing range from 0.01wt% to 0.15wt%. A lot of secondary fluid inclusions were found in the quartz in the brittle-plastic transition of fault, which were captured during crack healing. Based on measurement of the fluid inclusions, the capture temperature of the fluid inclusions is from 330 to 350℃, and fluid pressure is about 70～405MPa, the corresponding fluid pressure coefficient is estimated to be from 0.16 to 0.9, which stands for the characters of fluid inclusions captured during cracks healing process related with co-seismic and postseismic slip of fault.
Rheological structure was constructed based on the analysis data and flow law of wet quartz, as well as variation of fluid pressure and strain rate during periods of inter-seismic to earthquake nucleation, and after-slip to post seismic. Rheological structure shows that the strength of fault and depth of brittle-plastic transition change with strain rate and fluid pressure during inter seismic, earthquake nucleation, and after-slip period, and the depth of brittle-plastic transition is fit to the deformation mechanism of quartz, and the depth of transition of velocity weakening to strengthening of fault slip, as well as the focal depth of Wenchuan earthquake, which display that the Yingxiu-Beichuan Fault has the probability of weakening of sliding velocity and qualification of earthquake nucleation. However, the existing high fluid pressure in fault could be the most important factor for the high-angle reverse fault slip and triggering the Wenchuan earthquake.

The depth distribution of aftershocks of the 1996 Lijiang M_S7.0 earthquake is strongly time-dependent. Events occurring shortly after the main shock had deeper focal depths, and as the time going on, the focal depth of the aftershocks became shallower and shallower, i.e. the cutoff depth of seismicity became shallower and shallower with time. As we know, the lowermost events occur around the depth of brittle-plastic transition, and this depth depends on strain rate. The postseismic deformation model inferred from GPS data show that the major contribution of postseismic strain release comes from the lower layer of the crust. These results suggest that significant afterslip is related to viscous relaxation of lower layer. We estimated the lower bound of the strain rate according to Marone's et al.(1991)afterslip model and the slip data observed at the surface of Xianshuihe Fault. The results show that the strain rate is high after the main shock, and decreases gradually with time. We calculated the strength profile of middle crust based on flow law of wet quartz, estimated strain rate, temperature profile determined using the heat flow data at Lijiang, as well as crustal structure based on P wave velocity. By comparing the cutoff depth of seismicity and the brittle-plastic transition depth of the middle crust, we found the two depths are consistent to each other. We suggest the temporary existence of deeper small events after main shock and the depth distribution of aftershock is due to the changing brittle-plastic transition of the middle crust corresponding to strain rate variation from high to lower values after the main shock, and this kind of change is the manifestation of rheology of the middle crust.

It is crucial to research on creep tests of granulites in understanding rheology of lower crust,continental dynamics and seismogenic environment. However,based on analysis of creep data of mafic rocks,it was found that most of experimental studies of mafic rocks focused on creep of single-phase and two-phase aggregates,and creep data of mafic granulite is very limited. There are a lot of factors affecting rheology of mafic granulite. Besides the external causes of experimental conditions(temperature,pressure,strain rate),the internal causes,such as mineral components,grain size of samples,trace water,partial melt,mineral reactions,are very important influencing factors,which induced the results of creep tests on mafic rocks to be more complex. Therefore,there are a lot of uncertainties in discussing the rheological structures of lower crust according to creep data of single-phase and two-phase aggregates,which could not fulfill the requirement for studying lower crustal rheology. The calculation of the rheological strength of granulite constituted by multi-phase minerals based on end-member flow law parameters and empirical equation is only a kind of simple approximate,which could not replace experimental studies on rheology of granulite. The development trend of the experimental study on mafic rocks in the future will be to perform creep tests of multi-phase natural mafic granulite and obtain flow laws of samples which can be used to study the rheology of continental lower crust quantitatively. The scientific questions and technological problems during experimental studies of creep tests of mafic rocks are multi-phase mineral rheology,partial molten under high temperature and interactions of mineral reaction and rheology,and all of those need more detailed experimental studies under high temperature and pressure.

In this paper,we summarized deformation mechanisms of crustal rocks and the progress of the rheological experiments under high temperature and pressure,and discussed the quality of rheological experiment data. Although a great amount of creep test data have been published,the data of natural felsic rock,as well as data of quartz published in early time are hard to be reduplicated. However, the creep data of quartz and feldspar published in recent years are of very high quality. Some studies show that flow laws of two-phase rocks could be fitted using the empirical and theoretical models and rheological parameters of end member minerals. However,the flow laws for felsic rocks are hard to be determined because of complicated components and special rheological properties. As a result,when rheological parameters of felsic rocks are used to estimate the strength profile of the continental crust,the stress envelope and the depth of brittle-plastic transition zone are different even in the same temperature and strain rate condition. Therefore,we suggest more creep experiments should be done in order to obtain high quality rheological data of felsic rocks.
Based on recent progresses in rheological experiment,the influence factors on rheology of felsic rocks in laboratory conditions are discussed,especially,the covering water,mineral component,grain size and fabric of the rock. Little water in the rock would have a significant weakening role on the rheological strength. The effect of the rheological strength of melt is dependent on the observations and the distribution of the melt. If grain boundaries are wetted by melt films,the weakening effect of the melt will become more obvious. The effect of mineral component on rock rheology is reflected not only by stress exponents of samples,but also transition temperature from semi-brittle state to plastic state. The grain size mainly influences the deformation mechanism of rocks. In diffusion regime,the flow stress has negative linear correlation with grain size of fine grain samples,which is not only used to estimate as stress piezometer,but also suitable to determine the rheological strength in the ductile shear zone. However,in dislocation regime,the flow stress is independent of grain size,which is the foundation in estimating the rheological strength of the crust rocks in natural condition using experimental data. Finally,the effect of rock fabric and anisotropy,which is the common phenomenon in the crust,on rheology of experimental samples is very limited,which needs further experimental studies to get more data.

Constraints provided by field observation, laboratory experiments and seismic data have lead to a general consensus that the shallow crust deforms by brittle faulting, while the lower crust deforms by crystal plastic flow. These constraints provide the basis for the dual mechanism model for the rheology of the crust and lithosphere in which the strength of the upper brittle crust is limited by Byerlee's law, while the strength of the lower ductile crust is limited by power law creep. The maximum depth of microseismic activity is controlled by the broad zone of brittle-plastic transition that lies between the two extreme brittle and plastic layers. While the dual mechanism model is so simple that overestimates the strength of rocks near the brittle-plastic transition zone. Although many studies about the deformation mechanism of brittle-plastic transition zone have been made, a 'flow law' representation, which can describe the strength for the brittle-plastic transition, has not been formulated, and there has been little research about fluid effects; In addition, research on brittle-plastic transition usually focuses on temperature effects, while the research on the aspects of strain rate and fluid are relatively weak. Studies of deformation mechanisms of minerals in faults have indicated that brittle-plastic transition of some faults occurred in the same depth (temperature and pressure) and this phenomenon, which has been considered to be relevant to synseismic loading and postseismic creep in earthquake cycles and confirmed by distribution of focal depth, is due to the strain rate. The presence of high-pressure fluid in active fault at depth is proved by analysis of characteristics of fault fluids, and these fluids, which can evolve in pressure pertaining to fracturing and sealing processes, play a key role during the seismic cycle. The formation of high-pressure fluid (cracks sealing) has several mechanisms, but researches show pressure solution deposition is one of the main mechanisms which controls crack sealing kinetics around active faults. Studies on pressure solution under the action of water can supplement and correct the crustal strength profile defined by traditional relations describing brittle/frictional behavior (Byerlee's law) and dislocation creep. As a consequence, we believe it is necessary to further study the impact of strain rate and fluid pressure on the brittle-plastic transition through deformation samples both from field and high-pressure high-temperature experiments. Simultaneously, we may establish the equation for the pressure solution to approximately estimate the strength of brittle-plastic transition zone.

The co-seismic surface ruptures of the May 12,2008 Wenchuan earthquake in Bajiaomiao and Shengxigou were developed mainly at the outcrops of carbon mudstones of Xujiahe formation of Triassic system. The black color and textures of co-seismic gouge are similar to old gouge and bed rock. We excavated trenches along the surface ruptures and collected samples of wall rock, fault breccia, old fault gouge and co-seismic gouge. All samples were analyzed quantitatively by X-ray diffraction. The main rock-forming minerals and clay minerals of co-seismic gouge are similar to old gouge, but their content is different, which shows the co-seismic gouge was formed based on old gouge. The wall rock and fault breccias adjacent to co-seismic gouge are carbon mudstones. The mineral composition and texture of the fault zone are obviously simpler than that of the northern part of the surface ruptures of Yinxiu-Beichuan Fault. The major minerals of co-seismic gouge are quartz and clay minerals, containing a few amount of feldspar, without calcite; a small amount of dolomite was found in co-seismic gouge at Shenxigou, and the content of dolomite is much lower than that in bed rock and old gouge. The marked character of new gouge is abundant in clay minerals, and the content of clay minerals decreases in turn from co-seismic gouge to old fault gouge, fault breccia and wall rock. The main clay minerals are illite and illite/smectite (I/S) mixed layer, containing a few amount of chlorite; a few kaolinite was found in co-seismic gouge of Shenxigou, the bed rock and gouge of Bjiaomiao did not contain kaolinite. Mineral characters of co-seismic gouge are different from old gouge. The old gouge contains calcite and dolomite, and the co-seismic gouge contains a few amount of dolomite and without calcite; the old gouge does not contain illite. However, for co-seismic gouge, mineral characters are different between black and white gouges, the content of illite in black gouge is higher than that in white gouge. In this study, the main clay minerals are I/S mixed layer, illite and chlorite, which is similar with San Andreas Fault and Chelunpu Fault. However, kaolinite content is extremely low in this fault, only trace kaolinite was found in the co-seismic gouge of Shenxigou. The high content of I/S in co-seismic gouge shows that the rich K+fluid participated in the seismic fault slip. All of these characters show that minerals of co-seismic fault gouge in this study are somewhat different with that of San Andreas Fault and Chelunpu Fault.

Advanced Prediction is one of the most effective methods to assure the safety in tunneling.The histories and case experiences show that the TSP203(Tunnel Seismic Prediction203)combined with other techniques can meet the requirements of advanced prediction in tunneling and the prediction implementation costs.This paper firstly introduces the principle of the TSP203,and then compares the prediction result with the excavated conditions of the Dayaoshan railway tunnel.As an example,we introduce the prediction method,explain how to use TSP203 in complicated geological conditions.and analyze the main reason about predication errors according to several years'operation experiences in site as well as seismic theory.The method of TSP203 is a kind of mature technology,it is suitable for planar faults with large intersection angle to the axis of tunnel,but not fit for small intersection angle.However because the karsts terrain has the character of complexity,changeability and uncertainty,the predicting accuracy is a little lower in this kind of place.In addition,groundwater prediction is a technological problem in construction of tunnel that has not been solved both at home and abroad.TSP prediction accuracy is also related with operator's practice levels.Finally,some suggestions are given about how to improve data acquisition accuracy and reliability of explaining the reflected objects.Some important items in predication using TSP203 are summarized.

Pressure calibration for solid medium vessel under high pressure and high temperature is an important step before apparatus is employed,because precise pressure calibration directly determines the preciseness of measurement of experimental pressure.Sound and effective methods are premises of pressure calibration.Pressure calibration includes confining pressure calibration and axial load calibration.The best way of axial load calibration is to estimate axial friction by multi-cycle of piston-in and piston-out.There are two key points during the test:(1)ensuring the hit-point of piston and sample:The hit-point is determined by an intersection of two beelines,one is the linear fit to the part of load-displacement curve of piston contacting with soft metal,the other is the linear fit to the part of load-displacement curve of sample's elastic deformation;(2)confirming dynamic friction:The dynamic friction which is dependent with displacement is established by the linear fit to the part of load-displacement curve of piston contacting with soft metal.Then,the final axial calibration includes cutting the load-displacement curve before hit-point,and correcting the load-displacement with dynamic friction.The best method for calibrating confining pressure is mineral phase transition,such as quartz-coesite,albite-jadeite + quartz,fayalite + quartz-ferrosilite and farringtonite-Mg₃(PO₄)₂-Ⅱ,because those phase transitions are more stable and the transition equations are widely used in previous calibrating confining pressure.It is proposed that quartz-coesite and albite-jadeite + quartz be used as pressure standards for the piston-cylinder apparatus in the pressure-temperature range of 2.5～3.2GPa,500～1200℃ and 1.6～3.2GPa,600～1200℃.Tests were performed for calibrating the axial friction using 2GPa confining pressure vessel.The sample is gabbro.Two experiments have been performed:(1)a number of cycle-experiments with different piston-rate under 500MPa and 820℃,1000MPa and 900℃,1000MPa and 25℃;(2)under 500MPa and 1000℃,firstly cycle-experiments were conducted,and then piston rate reduced from 2×10^-4mm/s to 5×10^-5mm/s after rock sample's plastic deformation,and the rate dependence of creep observed.The result of the experiment shows that the factors which affect the dynamic friction are confining pressure,temperature and piston rate.Confining pressure is the main factor,dynamic friction increases with the increase of confining pressure.Temperature and piston rate only influence on intercept.Hence,axial calibration should be conducted under specific experimental condition.Finally,axial calibration of an experimental load-displacement curve was conducted.

Previous studies showed that the deformation of dry mafic lower crust is semi-brittle,which may locate in transition of semi-brittle slip to semi-brittle flow. Therefore,we perform tests on brittle-plastic transition of four kinds of dry and wet mafic rocks in order to comprehend the mechanical behavior of continental lower crust. The experimental confining pressure ranges between 450～500MPa,and strain rate is 1×10^-4s^-1. The results of experiments show that the samples of dry Jinan gabbro (sample C),dry Yanqin diabase (sample D) and wet Yanqin diabase (sample D) have experienced deformation regimes of faulting,cataclastic flow,semi-brittle flow and plastic flow under 300～900℃; and dry Panzhihua fine grained gabbro (sample A) and fine-to medium grained gabbro (sample B) have experienced deformation regimes of semi-brittle flow and plastic flow under 700～900℃. The temperature of brittle-ductile transition of dry gabbro is 100℃,higher than that of dry diabase. The temperature of brittle-plastic transition of all dry mafic samples is 700℃,but the microstructures of semi-brittle flow are variable. For example,grain-size reduction and preferred orientation of plagioclase and clinopyroxene occurred in deformed diabase,displaying the typical structures of protomylonite. However,this kind of microstructure did not appear in the deformed samples of gabbro. In contrast to the semi-brittle flow regime,the strength and microstructure of all dry mafic samples are basically the same in the plastic flow regime under higher temperature,when dislocation glide becomes the predominant deformation mechanism. The main effect of water on brittle-plastic transition of mafic rocks is embodied by both the change of strength and temperature of brittle-ductile transition and brittle-plastic transition. The strength of wet diabase is much lower than that of dry gabbro and diabase at the experimental condition. The temperature of both brittle-ductile transition and brittle-plastic transition of wet diabase is 100℃ lower than that of dry diabase,and 200℃ lower than that of dry gabbro. The temperature of transition from semi-brittle flow to plastic flow of wet samples is much lower than that of dry samples.

In summarizing the metamorphic temperature and pressure at hydrostatic condition of ultra high pressure rocks obtained from both field geology and experiments, we find that this problem is needed to further discuss, because the collision tectonic zone is not under hydrostatic conditions but under the action of differential stress. In this paper, we reassess and analyze the data of high temperature-high pressure experiment on quartz-coesite transition at differential stress condition made by Hirth and Tullis (1994). The results show that coesite is observed in both the semibrittle faulting and semibrittle flow regimes under temperature condition of 500～700℃ and pressure condition of 1.20～1.25GPa. Coesite is present mainly at the top and bottom of the tested samples adjacent to the pistons, as well as along fracture zones and along grain boundaries oriented perpendicular to σ₁ within the sample. The confining pressure (1.20～1.25GPa) required for quartz coesite transition in the presence of large differential stress is much lower than that (2.5～3GPa) at hydrostatic pressure condition. Obviously, the effect of differential stress is of great significance in the experiments. It is found that garnet in eclogite might be plastically deformed, indicating that differential stress do exist in collision tectonic zone, while the upper limit of the tectonic differential stress is constrained by the strength of rocks. Accordingly, differential stress is of great significance to ultra high pressure metamorphism. It is suggested, therefore, that systematic high temperature high pressure experiments are the essential and effective way to further investigate this problem.

Rheological parameters and deformation mechanisms of rocks are the basis for estimating crustal strength by using frictional constitutive equations and power law creep equations. In the past 30 years, substantial progress has been made in high-temperature and high-pressure experimental studies, which provide a large numbers of data on rheological parameters of crustal minerals and rocks, as well as new insights into the deformation mechanisms. In this paper, we summarize systematically the existing experimental data, and study crustal rheology of North China by using these data combined with focal depth distributions in this region. The results show that the upper crust as represented by granite and low-grade metamorphic rocks is deformed in brittle faulting and frictional sliding regime, the strength of which is controlled by friction on faults; the middle crust comprising felsic-gneiss, as well as the upper layer of lower crust composed of intermediate granulite behave in plastic flow regime; the lower layer of lower crust consists of dry mafic granulite, which behaves in brittle-plastic flow transition regime. The composition and rheological stratification of the crust in North China may cause decoupling of different crustal layers and provide mechanical conditions for strong earthquake generation. In addition, they may also serve as the bottom boundaries for blocks of different scales.

Seismic tomographic data show that crustal low-velocity layers and mantle uplift do exist in a vast area along Beijing-Tianjin-Tangshan-Zhangjiakou, where a lot of large earthquakes had occurred. However, such low-velocity layer and mantle uplift can not be found in Ordos region, where no strong earthquake had occurred. Geological study shows that the compositions of the crust are similar in these two areas. The main difference of the two areas is that rifting had occurred in Beijing-Tianjin-Tangshan-Zhangjiakou area at Cenozoic. We suggest, therefore, that the crustal low velocity layers in North China were controlled by the Cenozoic rifting and mantle uplifting, which caused the rise of temperature and hence giving rise to plastic deformation of lower part of the middle crust and lower crust. The lattice preferred orientation of minerals formed by plastic deformation can cause wave anisotropy, while subgrain substructure produced by dislocation creep may reduce the elastic wave velocity of rocks. The low velocity layers in lower part of the middle crust and lower crust, therefore, should be the result of plastic deformation of rocks. The low velocity layers in the upper part of the middle crust, however, might represent the low angle detachment or ductile shear zones formed during rifting, while low velocity might be resulted from anisotropy of rocks and possibly the effect of fluid. The crustal structure of Ordos region is similar to "ice-water" structure, and that of Beijing-Tianjin-Tangshan-Zhangjiakou area is like "sandwich". A large number of studies show that the crustal weak layers (lower velocity and plastic deformation layers) enhance the decoupling of stronger layers, and this may play an important role in the generation of large earthquake.