EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Analysis of surface rupture complexity sheds light on coseismic slip during the last earthquakes along the Bulnay-Tsetserleg fault zone (Mongolia)

Yacine Benjelloun1, Yann Klinger1, Solène Antoine1, Ganbold Baatarsuren2, Laurent Bollinger3, Yungbeom Cheon4, Jin-Hyuck Choi4, and Ganzorig Davaasuren1,2
Yacine Benjelloun et al.
  • 1Institut de Physique du Globe de Paris, Université de Paris, Paris, France (
  • 2Institute of Astronomy and Geophysics, Ulaanbaatar, Mongolia
  • 3CEA, DAM, Arpajon, France
  • 4Korea Institute of Geosciences and Mineral Resources, Daejeon, South Korea

In 1905, two M ~ 8 continental strike-slip earthquakes occurred along the Bulnay fault system, in the northwestern part of Mongolia. After a first earthquake that ruptured the Tsetserleg oblique fault strand, the second event ruptured the main Bulnay fault 14 days later. With a total rupture of 676 km, these two earthquakes constitute the largest continental strike-slip earthquake sequence ever documented. Hence, the Mongolian earthquake ruptures offer a unique opportunity to document large-magnitude earthquake continental ruptures. Indeed, due to dry climatic conditions, limited erosion and anthropization, the surface ruptures have been preserved almost unaltered. This allows for accurate documentation of the rupture trace and coseismic slip distribution along the Bulnay fault, based on field observation and satellite imagery.

Along the Tsetserleg rupture, the available coseismic offset measurement data coming from high-resolution satellite imagery show a significant variability, ranging between 1.5 and 4 m for the horizontal component. It is presently difficult to assess the most representative value for the 1905 slip, which in turn impacts the magnitude estimation for this event. Another factor to take into account is the possibility of a vertical slip component, which is only poorly constrained.

In order to have a better estimate of the 3D coseismic slip, drone images were acquired on selected sites along the Bulnay 1905 rupture, near the junction with Tsetserleg fault, and along the Tsetserleg rupture. We favored sites showing structural complexities and significant surface fracture development (succession of cracks and ridges, stepovers, branching zones…).

High-resolution DEMs and orthophotomosaics were produced using the MicMac software. The geometrical characteristics of the complexities and their fracture network were then measured in order to compute the volumetric changes associated to the 1905 earthquake. These data were finally converted to 3D surface slip estimates. On certain sites, we also discussed the presence of features inherited from previous ruptures, overprinted by the 1905 earthquake.

How to cite: Benjelloun, Y., Klinger, Y., Antoine, S., Baatarsuren, G., Bollinger, L., Cheon, Y., Choi, J.-H., and Davaasuren, G.: Analysis of surface rupture complexity sheds light on coseismic slip during the last earthquakes along the Bulnay-Tsetserleg fault zone (Mongolia), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5325,, 2020

Display materials

Display file

Comments on the display material

AC: Author Comment | CC: Community Comment | Report abuse

Display material version 2 – uploaded on 06 May 2020, no comments
Version description: Minor modifications: (i) homogenization of x-axes on slide 7; (ii) fault design clarification on slides 8 and 9.[...]
Display material version 1 – uploaded on 03 May 2020
  • CC1: Comment on EGU2020-5325, Christoph Grützner, 05 May 2020

    Thanks for the very detailed presentation and the amazing drone & field photos!
    I have a question regarding the scaling relationships that lead to the discrepancy between magnitude and observed surface deformation. Did you use Wells & Coppersmith? Have you checked if other proposed relationships would fit better? As far as I know, there are problems with several huge historical quakes in Central Asia. Assuming your field observation captured the whole surface deformation, do you think that scaling relationships for such quakes simply only make sense with huge error bars, or that there would be a way to come up with better scaling relationships, or that there is significant off-fault deformation that can't be captured by these formulas? (In the latter case we should always underestimate the magnitude, shouldn't we?)


    • AC1: Reply to CC1, Yacine Benjelloun, 06 May 2020

      Thank you for your feedback !

      The observed discrepancy comes from the study of Choi et al. (2018, who mapped the 1905 rupture. The magnitude was computed using Mw = 2/3 log M0 – 10.7 (Hanks & Kanamori, 1979). For the calculation of M0, they used the average slip deduced from geomorphic offset measurements, the rupture length measured from surface rupture extent, and tested different depths.

      It should be noted that this calculation does not lead to a similar discrepancy for the Bulnay earthquake (provided an unusual rupture depth of around 50 km). So, I am not sure the relationship used is the main source of “misfit” in the present case. Another thing to keep in mind is that the seismological estimate of the moment by Schlupp & Cisternas (2007, relies on very limited data coming from a narrow range of azimuth. Not sure how this translates in terms of uncertainty…

      Then about the scaling relationships, I went through a similar issue last year and had a look at alternative scaling relationships to Wells & Coppersmith. In the end, I was not convinced to find one which would really perform better. Ideally, one should use a relationship adapted to the area and geodynamic context of interest. But I wonder if we have enough well constrained events in the slowly deforming Central Asia to build a meaningful scaling relationship. Meanwhile, sticking to Wells & Coppersmith may be the least bad option, as long as people use it correctly and do not forget to report the large uncertainties.

      Finally, the Bulnay rupture is a good case study for the significance of off-fault deformation and distributed damage: it was observed that the coseismic offsets were in average lower in the western part of the rupture, and this was correlated with a much higher amount of distributed deformation compared with the eastern part. This also happens in some parts of the Tsetserleg rupture but maybe we could quantify this effect better? To come back to the scaling relationships, I agree with your remark: if these relationships systematically missed a significant part of the deformation, we should likely see it in general, not just for Central Asia.

      I hope this answers your questions.

      • CC2: Reply to AC1, Christoph Grützner, 06 May 2020

        Thanks a lot for the very detailed reply, makes sense.
        All the best