Search

The Minle-Yongchang fault is an active reverse fault-related fold structural belt developed in the Hexi Corridor Basin on the northeastern edge of the Tibetan Plateau. An earthquake of Ms6.1 occurred near the Minle-Yongchang fault zone in 2003. The deformation pattern of the Minle-Yongchang fault and its relationship with this strong earthquake, however, are still not well known. In this study, we used the methods of HYPOINVERSE absolute location and double-difference location with waveform cross-correlation technology to relocate the 2003 Minle earthquake sequence. In total, 383 earthquakes are precisely relocated. Based on the results of precise seismic relocation, using the method of determining fault planes by small earthquakes, the seismogenic fault is found to be a low-angle thrust with a strike of 311°, a dip of NE, and a dip angle of 14°. It does not rupture the surface, extends to 19–20 km depth, and is hidden beneath the Yonggu Anticline. We also employed the cut-and-paste (CAP) method with a broadband waveform to determine the focal mechanism of the mainshock in 2003: the strike is 311°; the dip is 34°; and the rake is 90°. The fault plane parameters obtained in these two ways are roughly consistent. We also used a digital elevation model (DEM) derived from the SPOT 6 stereo image pair and high-precision differential Global Positioning System (GPS) to measure the displacement of terraces. Topographic profiles along the terraces across the Minle-Yongchang fault show that high alluvial terrain exhibits fold deformation. The vertical offsets of the T2 and T3 terraces along the Tongziba River are approximately 2.3 and 22 m, respectively. Optically stimulated luminescence (OSL) dating indicates that the ages of T2 and T3 are 11.3 and 106 ka, respectively. We calculated an average uplift rate of 0.21 ± 0.05 mm/a by dividing the vertical offset by age. According to the spatial distribution of the relocated earthquake sequence and terrace deformation in the study area, the Ms6.1 Minle earthquake in 2003 was caused by the latest activity of a blind reverse fault-related fold in the Hexi Corridor Basin.

Long records of paleoearthquakes are essential for understanding earthquake recurrence behavior of active faults and for evaluating regional seismic hazard. However, paleoseismic data on the Altyn Tagh fault (ATF), one of the longest strike-slip fault in Asia, are scarce. We document a long paleoseismic record along the Xorkoli section of the central ATF. Eight and probably nine earthquakes are identified based on event evidence in the form of open fissures, folds, unconformities, and upward fault terminations, with modeled mean (95% confidence) ages of A.D. 1598 (1491–1741) yr (event A), A.D. 797 (676–926) yr (B), B.C. 668 (732–589) yr (C), B.C. 956 (1206–715) yr (D), B.C. 1301 (1369–1235) yr (E), B.C. 2105 (2233–1987) yr (F, probable), B.C. 2663 (2731–2601) yr (G), B.C. 2818 (2878–2742) yr (H), B.C. 3396 (3522–3205) yr (I). The mean recurrence interval is 620 ± 410 yr with a coefficient of variation (COV) of 0.67, indicating that earthquake recurrence is weakly periodic, with individual intervals ranging from as short as 150 yr to as long as 1460 yr. A global compilation of 35 strike-slip paleoseismic sites yields a similar average COV of 0.69. Synthesis of paleoseismic sites from the central ATF indicates that not all earthquakes ruptured to the eastern end of the Xorkoli section, within the Aksay restraining bend. Given that the 420-yr elapsed time since the most recent event, well within a COV of mean interval, a large surface-rupturing earthquake could occur at any time along the central ATF.

Stepovers in strike-slip fault systems play important roles in controlling the propagation of earthquake ruptures, depending partially on the competition between the stepover width and the coseismic slip gradient approaching the step. The 1920 Mw 7.9 Haiyuan earthquake is the most recent major earthquake that has occurred along the Haiyuan fault. The earthquake rupture broke through multiple stepovers along the fault and finally ended at the 4 km-wide Jingtai pull-apart basin (releasing stepover) at the western end. To understand the process of this termination of the Haiyuan earthquake rupture, we conducted detailed mapping of the surface rupture geometry, coseismic slip measurement and slip gradient calculation in the vicinity of the endpoint based on the 0.2 m-resolution SfM-derived DEM along the 20 km section east of the Jingtai pull-apart basin. Combining coseismic slip measurements from this study and published slip data from fieldwork in the 1980s, we calculated slip gradients of 84–160 cm/km as the rupture approached the Jingtai releasing stepover. These values are high for the 4 km-wide Jingtai releasing stepover compared to those from a dataset of worldwide historical earthquakes compiled by Elliott et al. (2009). The high values imply that the rupture could have a relatively high likelihood of breaching through the Jingtai releasing stepover. Alternatively, the slip gradient may be overestimated. Detailed mapping and field investigation, however, show that the rupture may extend ~1.86 km farther west of the end location indicated in previous studies. We acquired the new slip gradient with lower value of 50–82 cm/km, which is still considerably high compared to dataset of Elliott et al. (2009). We speculate that the slip gradient could be larger variation along the stepover boundary to stop fault slip in different geologic setting. Another factor control slip termination is related to fault properties. Creep along the southern boundary fault of the Jingtai stepover, a velocity-strengthening region, may also have played a role in stopping the rupture at this location. Our observations further indicate that the increasing LOS velocity of InSAR data within the Jingtai stepover is probably related to fault behavior rather than a nontectonic signal of subsidence.

Historical archives of earthquake occurrence provide a millennial view of fault system behavior with precision beyond the capability of radiometric dating techniques. Understanding the long historical record of earthquake activity in China presents a great opportunity to develop such an archive. However, the complex distributed nature of faulting here leads to uncertainty and requires matching historic events to the paleoseismic record from fault excavations. We present paleoseismic evidence for the most recent rupture of the north Danghe Nanshan thrust (NDNT) and correlate the age of this event to nearby historical archives. We use high‐resolution topography generated from airborne drone photography to measure an average coseismic fault throw of 0.8±0.2 m. Three trenches excavated across these small scarps show shortening of ∼1.3 m, consistent with event magnitude of M w 7.0±0.5 that probably ruptured the entire ∼45‐km‐long northern strand of the eastern NDNT system. Geochronology data and the historical record from Dunhuang and Anxi together indicate that this event likely occurred at A.D. 1289, during the Yuan Dynasty. Electronic Supplement: Description of the test procedure of optically stimulated luminescence (OSL) samples, and figure showing natural OSL decay curve, growth curves, and equivalent dose (De) distributions.

Strike-slip rupture propagation falters where changes in fault strike increase Coulomb failure stress. Numerical models of this phenomenon offer predictions of rupture extent based on bend geometry, but have not been verified with field data. To test model predictions of rupture barriers, we examine rupture extent along a section of the sinistral Altyn Tagh Fault punctuated by three major double bends. We measure 3–8 m offsets and map >95 km of continuous scarps that define the most recent surface rupture. We document the eastern terminus of this rupture within the Aksay bend, where an undeformed Pleistocene alluvial fan we mapped and dated overlaps the fault. We conclude, based on this geomorphologic evidence, that multiple Holocene ruptures have stopped in the Aksay bend. Our field data validate model predictions of rupture termination at a >18° restraining bend and support use of geometric parameters to define expected earthquake sizes in seismic hazard models.

Restraining double-bends along strike-slip faults inhibit or permit throughgoing ruptures depending on bend angle, length, and prior rupture history. Modeling predicts that for mature strike-slip faults in a regional stress regime characterized by simple shear, a restraining bend of >18° and >4 km length impedes propagating rupture. Indeed, natural evidence shows that the most recent rupture(s) of the Xorkoli section (90°–93°E) of the eastern Altyn Tagh fault (ATF) ended at large restraining bends. However, when multiple seismic cycles are considered in numerical dynamic rupture modeling, heterogeneous residual stresses enable some ruptures to propagate further, modulating whether the bends persistently serve as barriers. These models remain to be tested using observations of the cumulative effects of multiple earthquake ruptures. Here we investigate whether a large restraining double-bend on the ATF serves consistently as a barrier to rupture by measuring long-term slip rates around the terminus of its most recent surface rupture at the Aksay bend. Our results show a W—E decline in slip as the SATF enters the bend, as would be predicted from repeated rupture terminations there. Prior work demonstrated no Holocene slip on the central, most misoriented portion of the bend, while 19–79 m offsets suggest that multiple ruptures have occurred on the west side of the bend during the Holocene. Thus we conclude the gradient in the SATF's slip rate results from the repeated termination of earthquake ruptures there. However, a finite slip rate east of the bend represents the transmission of some slip, suggesting that a small fraction of ruptures may fully traverse or jump the double-bend. This agreement between natural observations of slip accumulation and multi-cycle models of fault rupture enables us to translate observed slip rates into insight about the dynamic rupture process of individual earthquakes as they encounter geometric complexities along faults.

Studies of Late Quaternary geological slip-rates and the GPS decadal slip-rates along the major strike-slip faults in the northern margin of the Tibetan Plateau show that slip-rates remain relatively constant along the middle sections of each fault and decrease rapidly toward the ends. Our recent studies suggest that the slip rates are low (

International audience; Paleoseismology provides fundamental data for generalizing earthquake recurrence behavior, by revealing past surface-rupturing events. Determining the size of paleoseismic events is notoriously more challenging than their timing. Paleoearthquakes exposed in trenches are vaguely defined as large enough to break to the surface and often assumed to be similar in size. Here we show an example where the paleoseismic record includes events of both moderate and large magnitudes. At the Salt Lake site on the active left-lateral Haiyuan Fault, northern Tibetan Plateau, a high-resolution stratigraphic sequence recorded three and possibly four events since A.D. 1500, constrained by accelerator mass spectrometry 14 C dating. Historical accounts of earthquake damage in the study region suggest that several earthquakes exposed in the trenches markedly differ in magnitude. With the exception of the most recent M~8 earthquake that occurred in A.D. 1920, two earlier events, which occurred in A.D. 1760 (or 1709) and 1638, respectively, are considerably smaller, with magnitude M < 7 and more likely M~6 or less. Thus, this section of the Haiyuan Fault that broke during moderate-magnitude events failed again after a short interval during a large M w 7.8–8.3 earthquake, as part of a larger multisegment rupture. Our study shows that moderate-magnitude events can be preserved in the stratigraphy and exposed by paleoseismic trenching under ideal conditions, for instance, if sedimentation is fast enough and there is no hiatus in deposition. Eventually, the data presented add to the growing body of paleoseismic records containing events of different magnitudes with a large variability in rupture length and coseismic slip.