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.

With the development and breakthrough of a series of techniques such as the fault surface morphology measurement, the geochemical element determination and Quaternary dating methods, it becomes possible to study paleo-earthquake using information recorded by the bedrock fault surface. At present, more and more scholars domestic and overseas have carried out a large number of paleo-earthquake studies on bedrock fault surfaces in different professional perspectives and have achieved a series of innovative results. This paper systematically introduces the development history, the current situation and the basic principles and applications of paleo-earthquake study on bedrock fault surface. Moreover, after the thorough discussion of the existing problems in paleo-earthquake research of bedrock fault surface, some suggestions for optimizing the current work are proposed. Finally, on the basis of comparison of the characteristics, advantages and disadvantages of various research methods, the prospects and development trends of the bedrock fault paleo-earthquake study are predicted. Lots of weaknesses and limitations in the current study are pointed out in this paper:Firstly, for the method of faullt surface morphology measurement, different morphological expression parameters exist nowadays, however, their advantages and disadvantages are unknown. Secondly, the TCNs method still has a large uncertainty in the age determination of the paleo-earthquake, and the mature cosmogenic nuclides dating methods is too few to meet the dating requirements of different lithologic fault surfaces. Besides, a reliable relationship between relative dating parameters such as morphologicl and physicochemical characteristics and the absolute dating method such as TCNs are not closely established to build a reliable chronology framework. The last but not the least, the lack of mechanical research on the physical and chemical biological processes that the bedrock fault surface experienced before and after the faulting and exposure, and insufficient multi-method comprehensive comparison are also the obstacles for the paleo-earthquake study on bedrock fault surface. It is suggested that in the future study of paleo-earthquakes on bedrock fault surfaces, more attention should be paid to the following aspects:Firstly, strengthen the evaluation of the reliability, applicability and accuracy of the parameters of each morphological model in time and improve the mathematical model of current dating techniques, optimize the mechanism of cosmogenic nuclide production, and introduce new high-precision dating technology timely; Secondly, strive to establish a reliable age framework between relative dating index(X)and absolute dating age(T)regionally; In addition, the morphological structure and mineral compositions of bedrock fault surface are analyzed proactively on the microscopic scale, and the mechanical study is conducted on a series of physical, chemical and biological processes that the fault surface experienced before and after the exposure. At last, comprehensive and comparative research need to be conducted by the multi-disciplinary and multi-method approaches. In conclusion, the paleo-earthquake study on the bedrock fault surface is going through the processes from the qualitative description to the quantitative expression, from the single-disciplinary method to the multi-disciplinary integration, from the exploration of a certain technical index to the comprehensive application of multi-source data technology. The combination of relative dating indicators(X)and absolute dating(T), and putting more emphasis on the mechanical study on the microscopic scale are the development trends of paleo-earthquake study on the bedrock fault surface. The close combination of the paleo-earthquake study of the bedrock fault surface with the traditional method of trenching conducted in the Quaternary sediment region is considered to help more effectively reconstruct a more complete paleo-earthquake sequence and the faulting history on the active fault zone, thus a more reasonable evaluation of the regional seismic hazard can be obtained.

Fault-related tectonic geomorphologic features are integrated expressions of multiple strong seismological events and long-term surface processes, including crucial information about strong earthquake behavior of a fault. It's of great significance to identify the strong seismic activity information from faulted landscapes, which include the date and sequence of the seismic activities, displacements, active fault features, for studying the seismic rupture process, predicting the future seismic recurrence behavior and evaluating the seismic hazard of the fault.
However, due to the restriction of measuring techniques and the subsequent poor quality of the acquired data, it has been difficult to accurately extract such information from complex tectonic landforms to study active faults for a long time. Recently, "small Unmanned Aerial Vehicle(sUAV)" photogrammetric technique based on "Structure from Motion(SfM)" provides a cost-efficient and convenient access to high-resolution and high-accuracy "digital elevation models(DEMs)" of tectonic landforms.
This paper selects the Tangjiapo area at the Haiyuan Fault to conduct data collection, in which the structural and geomorphic features are well preserved. Using a small quadrotor unmanned aerial vehicle(Inpire 2), we collect 1598 aerial photographs with a coverage area of 0.72km². For calibrating the accuracy of the aerial data, we set 10 ground control points and use differential-GPS to obtain the spatial coordinates of these control points. We use model software Agisoft PhotoScan to process these digital pictures, obtaining high-resolution and high-accuracy DEM data with the geographic information, in which data resolution is 2.6cm/pix and the average density of point cloud is 89.3 point/m². The data with these accuracy and resolution can fully show the real geomorphic features of the landform and meet the requirements for extracting specific structural geomorphic information on the surface.
Through the detailed interpretation of the tectonic landforms, we identify a series of structures associated with the strike-slip fault and divide the alluvial fan into four stages, named s₁, s₂, s₃, and s₄, respectively.Wherein, the s₁ is the latest phase of the alluvial fan, which is in the extension direction of the Haiyuan Fault and there isn't any surface fracture, indicating that the s₁ was formed after the M8.5 Haiyuan earthquake in 1920. The rupture zone on the s₂ fan is composed of varied kinds of faulting geomorphologic landforms, such as a series of en echelon tension-shear fractures trending 270°~285°, fault scarps and seismic ridges caused by the left-lateral motion of the seismic fault. In addition, a number of field ridges on the s₂ fan were faulted by the 1920 Haiyuan M8.5 earthquake, recording the co-seismic displacements of the latest earthquake event. Relatively speaking, the surface rupture structure of the s₃ fan is simple, mainly manifested as linear fault scarp with a trend of 270°~285°, which may indicate that multiple earthquakes have connected the different secondary fractures. And a small part of s₄ fan is distributed in the southwest of the study area without fault crossing.
Furthermore, we measured the horizontal displacements of river channels and vertical offsets of fault scarps. The faulted ridge on the s₂ fan and faulted gully on the s₃ fan provide good linear markers for obtaining the fault left-lateral dislocation. We used the graphical dislocation measurement software LaDiCaoz developed based on Matlab to restore the gully position before the earthquake by comparing the gully morphology on both sides of the fault, and then determined the horizontal offset of s₂, which is(4.3±0.4)m and that of s₃ is(8.6±0.6)m. In addition, based on the DEM data, we extracted the fault scarp densely along the fault strike, and obtained the vertical offset of s₂, which is(4.3±0.4)m and that of s₃ is(1.79±0.16)m.
Moreover, we detect slope breaks in the fault scarp morphology. For compound fault scarps generated by multiple surface rupture earthquakes, there are multiple inflection points on the slope of the topographic section, and each inflection point represents a surface rupture event. Therefore, the slope break point on the scarp becomes an important symbol of multiple rupture of the fault. The statistical result shows that the slope breaks number of s₂ is 1 and that of s₃ is 2. Based on the analysis of horizontal displacements of river channels and vertical offsets of fault scarps as well as its slope breaks, two surface rupturing events can be confirmed along the Tangjiapo area of the Haiyuan Fault. Among them, the horizontal and vertical displacements of the older event are(4.3±0.95)m and(0.85±0.22)m, respectively, while that of the latest event are(4.3±0.4)m and(0.95±0.14)m, which are the coseismic horizontal and vertical offsets of the 1920 Haiyuan earthquake.
These recognitions have improved our cognitive level of the fine structure of seismic surface rupture and ability to recognize paleoearthquake events. Therefore, the high-resolution topographic data obtained from the SfM photogrammetry method can be used for interpretation of fine structure and quantitative analysis of microgeomorphology. With the development of research on tectonic geomorphology and active tectonics toward refinement and quantification, this method will be of higher use value and practical significance.

The quantitative analysis of morphologic characteristics of bedrock fault surface is a useful approach to study faulting history and identify paleo-earthquake. It is an effective complement to trenching technique, specially to identifying paleo-earthquakes in a bedrock area where the trenching technique cannot be applied. This paper focuses on the Luoyunshan piedmont fault, which is an active normal fault extending along the eastern boundary of the Shanxi Graben, China. There are a lot of fault scarps along the fault zone, which supply plentiful samples to be selected to our research, that is, to study faulting history and identify paleo-earthquakes in bedrock area by the quantitative analysis of morphologic characteristics of fault surfaces. In this paper, we calculate the 2D fractal dimension of two bedrock fault surfaces on the Luoyunshan piedmont fault in the Shanxi Graben, China using the isotropic empirical variance function, which is a popular method in fractal geometry. Results indicate that the fractal dimension varies systematically with height above the base of the fault surface exposures, indicating segmentation of the fault surface morphology. The 2D fractal dimension on a fault surface shows a ‘stair-like’ vertical segmentation, which is consistent with the weathering band and suggests that those fault surfaces are outcropped due to periodic faulting earthquakes. However, compared to the results of gneiss obtained by the former researchers, the characteristic fractal value of limestone shows an opposite evolution trend. 1)The paleo-earthquake study of the bedrock fault surface can be used as a supplementary method to study the activity history of faults in specific geomorphological regions. It can be used to fill the gaps in the exploration of the paleo-earthquake method in the bedrock area, and then broaden the study of active faults in space and scope. The quantitative analysis of bedrock fault surface morphology is an effective method to study faulting history and identify paleo-earthquake. The quantitative feature analysis method of the bedrock fault surface is a cost-effective method for the study of paleo-earthquakes in the bedrock fault surface. The number of weathered bands and band height can be identified by the segment number and segment height of the characteristic fractal dimension, and then the paleoearthquake events and the co-seismic displacement can be determined; 2)The exposure of the fault surface of the Luoyunshan bedrock is affected and controlled by both fault activity and erosion. A strong fault activity(ruptured earthquake)forms a segment of fault surface which is equivalent to the vertical co-seismic displacement of the earthquake. After the segment is cropped out, it suffers from the same effect of weathering and erosion, and thus this segment has approximately the same fractal dimension. Multiple severe fault activities(ruptured earthquake)form multiple fault surface topography. The long-term erosion under weak hydrodynamic conditions at the base of the fault cliff between two adjacent fault activities(intermittent period)will form a gradual slow-connect region where the fractal dimension gradually changes with the height of the fault surface. Based on the segmentation of quantitative morphology of the two fault surfaces on the Luoyunshan piedmont fault, we identified four earthquake events. Two sets of co-seismic displacement of about 3m and 1m on the fault are obtained; 3)The relationship between the fault surface morphology parameters and the time is described as follows:The fractal dimension of the limestone area decreases with the increase of the exposure time, which reflects the gradual smoothing characteristics after exposed. The phenomenon is opposite to the evolution of the geological features of gneiss faults acquired by the predecessors on the Huoshan piedmont fault; 4)Lithology plays an important role in morphology evolution of fault surface and the two opposite evolution trends of the characteristic fractal value on limestone and gneiss show that the weathering mechanism of limestone is different from that of the gneiss.

The geological structure exposed by paleoearthquake trenches is the key material to the right cognition of fault activity and paleoearthquake. However, paleoearthquake trenching inevitably destroys active tectonic geomorphic evidence and trench exposures are usually difficult to reserve. The conventional process of recording the delicate geological information, manually constructing photomosaics by image-editing software, is time-consuming and produces undesirable artificial distortions. Herein, we explored the process of constructing trench orthophotomosaics and the 3D image model using the Image-based Modeling technology and applied it to the Liutiaohe trench across the Tianqiaogou-Huangyangchuan Fault, Gansu Province. Based on the 3D image modeling and orthophotomosaic, we firstly constructed the control points and scale bars on cleaned trench walls and collected photos of all sections of the trench with a digital camera in the field, and then reconstructed the 3D model of the trench through the Agisoft PhotoScan, an efficient image-based modeling software, and finally yielded the 3D image model of the trench and othophotomasaics of the trench exposures. The results show that the automated workflow can produce seamless, sub-millimeter-level high-resolution photomosaics more quickly, with precision in the centimeter range, and the 3D image model is of great help to identify strata and geological structures in trenches with much lower capital and labor costs and low expertise levels compared with LiDAR, meanwhile, the 3D archive benefits the share and communication and even allows future reinterpreting the site using new insights.

As an important technology to paleoseismologic research, trenching has been used to identify paleo-earthquakes recorded in strata, combined with dating technology. However, there have been some bigger uncertainties and limitations. For instance, subtle strata in loess sediment cannot be interpreted only by naked-eye, which seriously affects identifying paleo-earthquake horizon and time. Therefore, how to improve the accuracy and reduce the uncertainty of paleo-earthquake identification is the important problem we are currently facing. Dongyugou loess section, located in the northeastern corner of Linfen Basin, Shanxi Province, cuts across the Huoshan piedmont fault. The section exposes not only the well-developed loess sequence, but also several obvious faulting events. Thus, this loess section is a better site to make a high resolution study to improve the accuracy and reduce the uncertainty of paleo-earthquake identification. Based on the high-resolution grain size and magnetic susceptibility analysis, and associated with visual interpretation by naked-eye, we made a high-resolution stratification of Dongyugou loess section, including high-resolution thickness of each stratum and its upper and bottom boundaries. Based on the high-resolution stratification and their comparison between two fault walls, we identified three earthquake events, which occurred after formation of u5-7, u4 and u2, corresponding to their stratification depth of 7.1m, 4.7m and 2.9m in hanging wall. Based on results of OSL dating and average sedimentation rate of hanging wall, we estimated that the three events occurred around 45.8ka(between (48.1±1.5)～(43.2±2.5)ka), 32.8ka(between (35.0±2.4)～(30.6±1.3)ka) and 23.3ka(between (26.4±0.8)～(20.9±0.7)ka). According to the thickness difference of three loess-paleosol sedimentary cycles between two fault walls, we calculated the coseismic vertical displacements of the three events as 0.5m, 0.4 and 1.3m, respectively. Compared with other segments of the Huoshan piedmont fault zone, we found the southernmost segment is the weakest, with longer recurrence interval of about 11ka and lower vertical slip rate of 0.048mm/a. The high-accuracy grain size and magnetic susceptibility analysis offers an effective method for reducing the uncertainties of the paleo-earthquake research in loess area.