Detailed crustal structure of big earthquake resource regions is of good significance for knowledge the earthquake generation mechanism. Numerous big earthquakes have occurred in the NE Tibetan Plateau, consisting of the 1920 Haiyuan M8.5 and also 1927 Gulang M8 earthquakes. In this paper, we obtained a high-resolution three-dimensional crustal velocity model around the source regions of these two large earthquakes using an enhanced double-difference seismic tomography method. High-velocity anomalies encompassing the seismogenic faults are observed to extend to depth of 15 km, suggesting the asperity (high-velocity area) plays critical role in the preparation process of large earthquakes. Asperities are solid in mechanical strength and could accumulate tectonic stress much more easily in lengthy frictional locking periods, huge earthquakes are as such prone to generate in these areas. If the near relationship between the aperity and also high-velocity bodies is valid for most of the large earthquakes, it deserve to be offered to guess potential huge earthquakes and also estimate the seismogenic capacity of faults in irradiate of framework studies.

You are watching: 1927 china earthquake magnitude length and damage

Earthquakes take place when the stored energy in the earth lithosphere is unexpectedly released. Huge earthquakes typically cause good hazards top top natural setting and/or humans. There has actually been an evident surge of good earthquakes through magnitudes ≥ 8.0 during the past decade with good diversity at miscellaneous aspects1. The diversity that earthquake resource is generally linked with the geometrical complexities that the fault systems, which are attributed to the heterogeneity of the dynamic rupture process2,3,4,5. Moreover, earthquake procedures are significantly impacted by the heterogeneity of mechanically properties of the fault zone, frequently represented conceptually together asperities6,7,8. Faults the are totally or partially locked by strong asperities breed good seismic hazard since the collected stress ~ above the asperities are prone come be exit through large earthquakes, contrasting sharply with the faults i beg your pardon are identified by climb deformation9,10,11. Seismological studies indicate the incident of solid earthquakes is carefully related to the abnormal distribution of crustal velocity12,13,14,15. Therefore, examining the well velocity structure approximately the source regions could shed light top top the relationship between the velocity functions and large earthquakes and also furtherly the seismogenic system of huge earthquakes.

Various studies have been carried out to discover the crustal framework of the large earthquake source regions and also link the observations to system of earthquake generation16,17,18. Besides, there have actually been a growing number of observations and numerical simulation studies on asperity7,19,20,21,22. However, so much there are few studies top top the high-resolution regional crustal structure bordering the resource regions of large earthquakes. To investigate the effects of structure heterogeneity on huge earthquake generation, we explored the body-wave crustal velocity structure in the NE Tibetan Plateau where is high on seismicity (Fig. 1) and is therefore suitable natural laboratory to examine the seismogenic device of large earthquakes. As one of the most active areas in continental China, numerous large earthquakes arisen there, including the 1920 Haiyuan M8.5 earthquake and also 1927 Gulang M8 earthquake. These 2 earthquakes are very destructive and also have resulted in a heavy loss top top life and also property. Though nearly 100 years have actually passed since the Haiyuan earthquake, there has not been sufficient research on this two huge events because of the lack of data. Fortunately, in current decades, plentiful seismic data space increasingly available in the NE Tibetan Plateau, permitting us to attain its high-resolution velocity structure and explore the seismogenic device of the associated huge earthquakes.


Tectonic background of the NE Tibetan Plateau and its surrounding areas. The white circles signify the areas of earthquakes larger than size 6.0 and also their sizes are proportional to magnitude. The two pink circles stand for the 1920 Haiyuan M8.5 earthquake (on the right) and also 1927 Gulang M8 earthquake (on the left), respectively. The 2 thick blue lines display the rough surface rupture area of this two big earthquakes23,24, while their focal mechanisms are shown by the two surrounding beach balls respectively23,25. The black color thin lines highlight the main active faults26, among which the significant ones room abbreviated together followings: LSF, Longshoushan fault; HSF, Huangcheng-Shuangta fault; TJF, Tianjingshan fault; HYF, Haiyuan fault; LPF, Liupanshan fault. The maps are produced using share Mapping tools (GMT)27 (v.4.2.1,

In this work, we enhanced the double-difference seismic tomography an approach and used it to acquire a high-resolution three-dimensional (3-D) ns wave and S wave velocities (Vp and also Vs) model roughly the focal regions of the 1920 Haiyuan M8.5 and 1927 Gulang M8 earthquakes. Through assessing their velocity features, us investigated the relationship between huge earthquakes and velocity structure. Furthermore, we explored the seismogenic device of huge earthquakes in irradiate of the asperity (high-V patches) which offers a overview for additional seismic risk assessment and associated catastrophe reduction. The comprehensive geometrical shape of fault aircraft is usually hard to depict, we therefore mainly discuss the seismogenic system in terms of the structure features of the neighboring rocks.

Sufficient observed seismic data and additional picking of later phases are critical for high-resolution tomography. In this work, us collected huge amount of travel time data in the study region from 2 sources: the background data indigenous Gansu, Ningxia, inner Mongolia and Qinghai districts (1985–2008) and uniform data from national Earthquake Data center (2009–2018) ( We lugged out the adhering to criteria top top data selection: (1) all of the seismic hypocenters and also seismic train station are spread in the variety of 32°–41°N latitude and 97°–109°E longitude. (2) Each occasion was recorded at least by 5 stations. (3) The epicentral ranges for Pn and Sn tide are bigger than 2°29. (4) The take trip time residuals are much less than 5.0 s. We finally acquired 325,829 Pg wave, 27,903 Pn wave, 311,646 Sg wave and also 8,783 Sn tide from 34,997 local earthquakes taped by 124 permanent and temporary seismic station (Figure S1). The selected take trip times the Pg, Pn, Sg and also Sn wave exhibit universally direct relationship through the epicentral distance (Fig. 2).


The distribution of take trip times matches epicentral distance. The four different seismic phases are presented by various colors, with Sg (cyan), Sn (magenta), Pg (green) and also Pn (red) from peak to bottom.

Resolution tests

Checkerboard check is a routine procedure for assessing the resolution and also reliability of tomographic images. Us assigned additionally positive and negative velocity perturbations of 5% to the surrounding inversion grids that the early velocity model. The artificial travel times to be calculated utilizing the same ray distribution as the real data. Consequently we applied our boosted tomoDD method to acquire the recovered images and then compared with the initial checkerboard model to inspect the recovery degree. Number S3a and S3b reflects the recovered checkerboard fads for Vp and also Vs tomographies about the resource regions that Haiyuan and also Gulang earthquakes through the lateral network interval of 0.2°, respectively. ~ above the whole, the checkerboard patterns were fine recovered with resolution reaching to 0.2° above 15 km. Artificial tests were also carried out to assess the reliability of the velocity anomalies, and these tests demonstrate that the main functions of our tomographic results are well recovered at many of the locations (Figure S4 and S5).

The velocity structure approximately the 1920 Haiyuan earthquake

The 1920 Haiyuan M8.5 earthquake is just one of the biggest seismic events in continental China through an approximated hypocenter of longitude 105.3°E, latitude 36.6°N and also depth 11 km30. It occurred on the Haiyuan fault, i m sorry is a sinistral strike-slip error zone with a length span of ~ 1000 km and also it connect the Qilian orogen come the west and also the Qinling orogen to the east31. The fault rupture the this occasion is said to be about 220 km long32. Intraplate earthquakes in basic occur in ~ the brittle upper crust above 15 km. Figure 3a,b and also 4c,d present the inverted velocity perturbations of P and S tide at different layers relative to the median velocity at each depth, respectively. The fads of Vp and also Vs images are generally similar to each other, return the size and also amplitude of anomalies room slightly various somewhere which could be attributed come the relatively lower resolution of Vs maps. In ~ the depth that 5 km, Haiyuan earthquake source region is conquered by southeastward elongated high-V anomalies surrounding by low-V anomalies, although the low-V anomalies in the Vs map room not together clear as that of Vp framework at the southwest of the source region. The high-V anomalies in Haiyuan earthquake source region extend downwards come the depth of 10 km, and also are even visible at 15 km at Vp results however shrink in size and amplitude.


Tomographic Vp picture (a) and Vs image (b) approximately the source regions the the 1920 Haiyuan and 1927 Gulang earthquakes, respectively. The magenta lines in number (a) stand for the location of the 4 vertical tomographic profiles called AA′, BB′, CC′ and also DD′ in Fig. 4.


Vertical P and also S tide tomographic images with relocated earthquakes. The thick red present on the file of AA′ and also CC′ display the rough surface rupture zones of the two large earthquakes, and also the white lines represent the major faults.

The tomographic velocity structure approximately the Haiyuan earthquake is equivalent to the outcomes from a much more accurate deep seismic sounding study along longitude ~ 105.5°, which observed high-V anomalies over 25 km crossing Haiyuan fault30. The magnetotelluric study about the Haiyuan earthquake zone also demonstrated high resistivity in the focal area33, i m sorry is consistent with ours results.

The velocity structure approximately the 1927 Gulang earthquake

The M8 Gulang earthquake occurred on 23 may 1927 is another huge event in the Haiyuan-Qilian fault belt after ~ the 1920 Haiyuan earthquake. The epicenter is around located in ~ 37. 6°N and also 102.6°E34. The surface-rupture region is but much shorter, only 23 km long along the thrust35. Indigenous the high-resolution 3-D models obtained in this paper (Figs. 3a,b and also 4a,b), Gulang earthquake source an ar has comparable velocity trends as Haiyuan earthquake and also is conquered by high-V anomalies between 5 and also 15 km. Particularly on the Vs map, the high-V features are much more obvious over 10 km. The focal area is surrounding by low-V anomalies, such as the Hexi corridor and southern area to the Gulang earthquake.

The far-reaching high-V features which room noticeable above 15 km in ~ the focal distance area that the Gulang earthquake have likewise been it was observed by a collection of vault seismic tomography studies36,37, yet our results present the anomlies in a much higher resolution due to the improved technique and plentiful data. The magnetotelluric sounding outcomes show similar patterns through high-resistivity features above the focal distance area38. The heaviness study additionally find high-density anomalies in the Gulang earthquake source an ar with high-V structures39.

The dominant finding in this document is the the focal areas of the 1920 Haiyuan and also 1927 Gulang earthquakes space both defined by 3-D high-V anomalies above 15 km, i m sorry has likewise been revealed by previous deep seismic sounding and also magnetotelluric profiles30,33,38. Besides, the seismogenic faults exactly reduced through the high-V bodies (Fig. 4). The high-velocity structures approximately these two large earthquakes room located approximately the Precambrian basement and/or Neoproterozoic to early Paleozoic ophiolite sequences40,41.

Similar high-V anomalies around large earthquakes have actually been it was observed by many huge earthquakes, such together the 1966 Parkfield M6 earthquake14,16, 1995 southerly Hyogo M7.3 earthquake18, 2004 M6.8 Niigata–Chuetsu earthquake42, 2008 Mw7.9 Wenchuan earthquake43, 2010 Ms7.1 Yushu earthquake44,45, 2011 Tohoku-oki Mw 9.0 earthquake46, 2013 Ms7.0 Lushan earthquake47,48 and 2015 Mw7.8 Gorkha earthquake15. Therefore, that is probably a global phenomenon the the high-V body in the resource regions of huge earthquakes represent the asperities in error planes, which space essential facets for huge earthquake generation.

The asperity model8,49 has actually been widely embraced to describe the seismogenic device of main shock of huge earthquakes. The term asperity was originally characterized as "unevenness the surface, roughness or ruggedness". Smooth surfaces, even those refined to a mirror, room not important smooth top top a microscopic scale. Surface asperities exist across multiple scales, and seismic asperity top top a fault airplane is at a macroscopic scale. Simply speaking, the asperity is a locked area before an earthquake (Fig. 5a), which has stronger localized mechanical coupling in the source an ar and is able to offer greater than mean resistance come rupture8. The asperity will certainly block error from slipping or the fault slip will destroy the asperity in the error plane. For this reason the stamin of asperity plays an essential role in the initiation of error slip. The more powerful the asperity is, the more challenging the error sliding is. Tension incessantly accumulates on the asperity in the interseismic duration due to tectonic loading (Fig. 5b) till fault fail (Fig. 5c), and the preferably slip normally occurs in ~ the asperity zone8,50. Generally, high-V bodies have actually high density, tiny porosity and also then high strength51. Therefore, high-V bodies stand for asperities and are the basic cause for the locking of error planes, i m sorry finally add to the generation of huge earthquakes.

Figure 5


The cartoon illustrates the mechanism of preparation and also generation of big earthquakes.

Full size image

If the close relationship in between the aperity and also high-V body is valid for many of the huge earthquakes, it have the right to be used to suspect potential huge earthquakes and estimate the seismogenic ability of faults in irradiate of structure studies. In this work, there room clear high-V anomalies above 15 km cross the Tianjingshan error (Fig. 3), i m sorry suggests strong potential for large earthquakes. Us then searched the background earthquakes indigenous Chinese Earthquake Catalog and also found the 1709 Zhongwei M7.5 earthquake (latitude 37. 4°N, longitude 105.3°E) just arisen there, which validates our conclusion. Therefore, if comparable tomographic framework researches are carried out on all active large faults, it will be a an excellent contribution come earthquake prevention and also disaster reduction.

The double-difference seismic tomography an approach (tomoDD)52 is emerged from the double-difference hypocentral location technique (hypoDD)53 and has the benefit to attain the hypocentral locations and velocity framework simultaneously. TomoDD is an ideal for local-scale tomography43,54. There are at the very least two worries need to be handle when used to big study locations for guaranteeing a trusted result. First, tomoDD just sets the Moho discontinuity as a gradational rather a sharp interface, which inevitably biases the velocity structure around the Moho discontinuity. This difficulty is especially severe in our study area wherein the Moho depth comparison can be as large as 30 km between the special NE Tibetan Plateau and also the Ordos and Alxa blocks55,56. Second, tomoDD usually only supplies first-arrival phases in the inversion, yet there are plenty of later-arrival phases such together Pg and also Sg waves in huge epicentral street over 200 km. These phases are vital for enhancing the coverage and crossing the ray courses in the middle-to-lower crust and thus could much better constrain the velocity framework there.

To improve the tomographic resolution, we boosted the tomoDD method mainly in two aspects. First, comprehensive Moho depths beneath the research area are included as a prior details in the inversion, which aid to acquire high-accuracy travel times the first-arrival Pn and also Sn phases. The Moho results are mainly acquired from the thick stations of ChinArray55. For the areas out the coverage that ChinArray stations, the Moho outcomes are collected from that et al.57. Second, us modified the ray tracing of tomoDD method for allowing to calculate the travel times the later-arrival Pg and also Sg phases.

We collection up a 3-D velocity design with horizontal spacing that 0.2° and also ~ 5 km in depth follow to the data density and also checkerboard tests. Referring to the comprehensive 2-D velocity outcomes from active-source seismic profiling in the examine region58, we built the early velocity version with the Moho depth taken right into account (Figure S2).

The LSQR algorithm59 is supplied to fix the huge and sparse system of monitoring equations which attach the observed travel times to the unknown hypocentral and 3-D velocity parameters in tomoDD method. In bespeak to minimize the instability during the inversion and also balance model smoothness versus data fitting, smoothing and damping regularizations are adopted in the inversion. ~ testing various values that smoothing and also damping parameters because that finding the optimal trade-off suggest between the RMS the travel-time residual and the share of the 3-D velocity model, the values of 10 and 600 were selected for the smoothing and damping parameters, respectively.


Lay, T. & Kanamori, H. In Earthquake prediction (eds David W. Simpson & Paul G. Richards) 579–592 (1981).

Download references

We thank the center of the China Earthquake Networks ( for giving the seismic data in this study. This occupational was sustained by the National an essential R&D program of China (2017YFC1500303), the strategic Priority Research program of Chinese Academy of scientific researches (XDA20070302) and also the National natural Science structure of China (U2039203, 41674090, 41490610).


Key laboratory of continent Collision and Plateau Uplift, institute of Tibetan Plateau Research, Chinese Academy of sciences (CAS), Beijing, 100101, China

Quan Sun, Shunping Pei, Yanbing Liu, Xiaotian Xue, Jiawei Li, Lei Li & Hong Zuo

University that Chinese Academy that Sciences, Beijing, 100049, China

Quan Sun, Shunping Pei, Yanbing Liu, Xiaotian Xue, Jiawei Li, Lei Li & Hong Zuo

CAS facility for Excellence in Tibetan Plateau earth Sciences, Chinese Academy of scientific researches (CAS), Beijing, 100101, China

Shunping Pei

Department of Earth and Environmental Sciences, Lehigh University, Bethlehem, PA, 18015, USA

Zhongxiong Cui

Department of ocean Science and also Engineering, southerly University the Science and also Technology, Shenzhen, 518055, China

Yongshun man Chen


J.L., L.L. And also H.Z. Are responsible because that collecting the seismic data. Y.L. And also X.X. Room responsible for processing the data. Q.S. And also S.P. Enhanced the technique and conductd the inversion. Q.S., S.P., Z.C. And Y.J.C. Contributed to interpretation and also writing.

Corresponding author

Correspondence come Shunping Pei.

Ethics declarations

Competing interests

The authors explain no competing interests.

Additional information

Publisher"s note

Springer continues to be neutral through regard come jurisdictional claims in published maps and institutional affiliations.

See more: Iverson Movie Ranch: Where Was Daniel Boone Tv Series Filmed

Supplementary information

Supplementary information.

Rights and permissions

Reprints and also Permissions

Cite this article

Sun, Q., Pei, S., Cui, Z. Et al. Structure-controlled asperities the the 1920 Haiyuan M8.5 and 1927 Gulang M8 earthquakes, NE Tibet, China, revealed by high-resolution seismic tomography. Sci Rep 11, 5090 (2021). Https://

Download citation

Received: 01 July 2020

Accepted: 09 February 2021

Published: 03 march 2021

Share this article

Anyone you share the following connect with will be able to read this content: