Local stress environment
In Fig. 6, the spatial distribution of LURR anomalies at three different time periods before the Yutian earthquake was calculated to detect those regions with high stress accumulation. The figures are derived by evaluating LURR of Benioff strain of the Ms ≤ 4.0 earthquakes (data from the China Earthquake Networks Center catalog). The time window is 0.5 year and the spatial window 80 km radius which slide at a step of 0.25degree in both longitude and latitude. The internal friction coefficient used to evaluate CFS is 0.4 and the value of LURR greater than 1.0 is plotted at the center location by using the 2D bilinear interpolation of Matlab.
The results showed that there was no significant anomaly in Yutian region in the second half of 2019 (Fig. 6a), but the anomaly increased significantly in March 2020. It is clear that the spatial distribution of LURR anomaly expanded and the amplitude was also greatly enhanced, especially in the area around the future epicenter where the LURR values exceeded 4 (Fig. 6b). Since then, the LURR values in this region began to weaken before the Yutian earthquake (Fig. 6c).
In order to evidence that stress accumulation in Yutian earthquake leads to the change of LURR, we constructed a profile. We constructed a GPS profile along the direction of the Ashikule fault using GPS data from 1999 to 2019 to verify the motion properties of the extensional fault area at the southwest end of the Altun Tagh fault zone (Fig. 7a). Calculated data were obtained from the First Monitoring Center of the China Earthquake Administration. Velocity field was calculated using quasi-observational combination analysis26, GNSS (global navigation satellite system) at MIT (GAMIT)/Global Kalman filter (GLOBK)27,28,29 of the GPS velocity field associated with mainland China. Results show that the directions of the GPS velocity field within 200 km in the south and 300 km in the north of the profile are in NNE and NNW or N directions, respectively (Fig. 7b). The average velocities on both sides of the section are also significantly different. The NW–SE tensile state of the NE trending rupture fault is dominant, and the slip rates on north and south sides of the fault are approximately 9 and 16 mm/a, respectively, probably because of the northward subduction extrusion of the Southern Indian plate, the obstruction of the Tarim block in the north, and the extrusion of the western Kunlun thrust fault.
This stress state change in the crust calculated by LURR may also be recorded by the near geodetic data in the same time periods (Fig. 8). The vertical pendulum tilt observation of the Yutian Seismic station is located in a cave in Aqiang Township, Yutian County, Xinjiang Uygur Autonomous Region (Fig. 2) with geographic coordinates of 36.4°N and 81.9°E and an elevation of 1200 m. The site consists of clastic carboniferous metamorphic rock and marble. The observation site is located at the intersection of the eastern side of the West Kunlun fold belt and the Tarim Basin. The Pulu fault, which was active in early Quaternary in the NE direction, is distributed approximately 8 km to the north of the station. Approximately 40 km to the east of the station is the Altun Tagh fault zone, which is NE-trending and active in Holocene. Approximately 47 km to the south is the Karakax fault zone, which is NE-trending, and active in Holocene. The station was built in October 2009, within a room in the cave. The cave is 30 m long and it is powered by solar energy. Formal observation began in November 2016.
As shown in the daily mean curve (Fig. 8a), the EW component observed by the tilting of the Yutian vertical pendulum tilt was usually in an E-dip state from mid-September to mid-April of the following year. The time of the E-dip phase was from November 27th, 2019 to April 14th, 2020, and it is the same period as in previous years; however, the annual variation in this period is significantly reduced compared to those of same periods in previous years (Fig. 8c). There was no environmental interference in the site during the period of observation. Comparative statistics shows that the annual variation from 2019 to 2020 is 0.264″ (Table 2), which is significantly smaller than the annual variation during the same period in previous years, and hence, there is an annual variation anomaly. The Yutian Ms 6.4 earthquake occurred 73 days after the end of the anomaly (Fig. 8a).
Anomalous changes in Yutian vertical pendulum near the epicenter of the Yutian earthquake two months before the earthquake reflect the instability and rupture of the fault zone in this region. Previous studies showed that the Altun Tagh, Kunlun, and Karakax fault zones and other strike-slip faults were formed in the northern boundary of the Qinghai–Tibet Plateau due to the eastward movement of the Qinghai–Tibet Plateau block and the obstruction of the Tarim Basin30,31,32. The GPS velocity of the Altun Tagh fault gradually decreases from west to east33. The GPS velocity of the western segment is relatively small, its compression is the largest, and surface deformation mainly occurs in the south. However, in the north, due to the stable Tarim block, there are few surface shape variables. As shown in Fig. 7b, the south side of the Yutian earthquake’s seismogenic fault moves eastward at a speed of approximately 3 mm/a. Approximately two months before the earthquake, anomalous deformation observed in the northwest direction of the earthquake shows that Altun Tagh and Karakax faults show eastward movement, and the rate of movement of the former is higher than that of the latter. The instability and loosening of the fault junction had occurred two months before the earthquake because of the long-term pull action and continuous accumulation of stress at the fault junction, and the Karakax fault had temporarily escaped from the eastward extrusion of the Qinghai–Tibet Plateau. The westward movement occurred under the influence of the Tarim Basin, resulting in the tensile characteristics of the Ashikule fault. When the accumulation of tensile stress exceeded the critical rupture state, the Yutian Ms 6.4 earthquake occurred in 2020.
Correlation between Benioff strain and geoelectric LURR
LURR models can be used to detect the variations of crustal stress state in the source area before an earthquake occurs. LURR in this paper is evaluated based on the dynamic evolution law of rock constitutive relationship. All geophysical quantities that can reflect the medium instability process in a seismogenic area can be regarded as response quantities. LURR theory is based on the triggering mechanism of tidal stress. Since tidal stress is far lower than tectonic stress, it can only trigger but not create earthquakes. When the tectonic stress is low, small changes of tidal stress are difficult to trigger earthquakes, which mean the crustal medium is in a stable state, hence, LURR < 1. However, when the tectonic stress accumulates to a high level, any small increase of stress, such as tidal stress, may lead to the occurrence of small earthquakes, resulting in differences in the release of Benioff strain between loading and unloading stages, thus, LURR > 1. The Benioff strain response calculation method based on the LURR has a better continuity compared to geophysical observation data. Small changes in LURR volatility in the time curve can accurately identify the critical state before the occurrence of an earthquake, when the rock medium is in an elastic stage (steady state) and the expansion is full in the seismic zone.
LURR of geo-electric field is based on the premise that the expansion of rock volume caused by the cracks generated in critical state of earthquake can change geo-electric field near the focal area. In particular, new cracks were not generated in the solid medium during the loading and unloading process, the number of cracks remained unchanged (Fig. 9a), and the migration of fluid in the rock mass was not affected, and hence, the change in flow potential was very small (Fig. 9b). In addition, the change in the geo-electric field LURR time curve and the LURR value were close to 1.0. As stress accumulation reaches a high level, the rock enters the yield stage (instability state), and according to the Kaiser effect34, the number of cracks generated in the rock experiment at this stage was significantly higher than that in the unloading stage (Fig. 9c). At the loading stage, rock dilatation and the generation of microcracks lead to changes in fluid migration, resulting in filtration potential (Fig. 9d). Therefore, the generated local electric field will cause the LURR value to increase gradually and deviate from the stable value of 135,36.
Our calculation results display that the LURR has very significant prediction value. As shown in Fig. 5, the Benioff strain LURR anomaly near the epicenter began to rise rapidly around October 2019 and reach the peak in March 2020. At this time, the LURR anomaly of the Hotan geoelectric field, which is 290 km away from the epicenter, also changed simultaneously, but not significantly. After that, it began to rise rapidly around December 2019, reach the peak in 1 to 3 months before the mainshock. This phenomenon indicates that the change in electric field is related to fracture propagation and fluid infiltration during earthquake preparation18,19,37. Anomalous changes in the electric field often occur in the late seismogenic stage38, and hence, the peak value of electric field anomaly may be later than that of the LURR of the Benioff strain as the response quantity.
Visit our sponsors
Wise (formerly TransferWise) is the cheaper, easier way to send money abroad. It helps people move money quickly and easily between bank accounts in different countries. Convert 60+ currencies with ridiculously low fees - on average 7x cheaper than a bank. No hidden fees, no markup on the exchange rate, ever.