Follow "Surveying and mapping academic information "
2020 Stanley earthquake co-seismic deformation and fault sliding distribution inversion
Jiang Xin 1,2,3,Wang Shijie 1,2,3,Li Wei 1,2,3,Yang Zhenkang 4
(1. School of Surveying and Mapping and Geographic Information, Lanzhou Jiaotong University, Lanzhou 730070;
2. National Joint Engineering Research Center for the Application of Geographic National Condition Monitoring Technology, Lanzhou 730070;
3. Gansu Provincial Geographic National Condition Monitoring Engineering Laboratory, Lanzhou, Lanzhou 730070;
4. Qinghai Province Geographical Space and Natural Resources Big Data Center, Xining 810000)
Abstract Need : Research on the co-seismic deformation field and fault motion characteristics caused by the Stanley earthquake on March 31, 2020, this article obtained the co-seismic deformation information of the Stanley earthquake based on D-InSAR technology. Based on the two-step inversion method, the distributed sliding distribution of seismic faults is obtained. This method obtains the geometric parameters of the Stanley earthquake fault through nonlinear inversion, and obtains the fine sliding distribution on the fault surface through linear inversion. The co-seismic deformation results show that the uplifting deformation of the Stanley earthquake is 24.2 cm, the settlement is 12.5 cm, the downlifting deformation of the uplifting deformation of the downlifting rail is 22.3 cm, and the settlement is 14.1 cm. The sliding distribution results show that the fault rupture did not extend to the surface, and the fault sliding is mainly left-handed strike-slip, mainly concentrated 10~12 km underground, with a maximum sliding volume of about 0.9 m, and appears 12 km underground.
Introduction
At 7:52 on March 31, 2020, a magnitude 6.6 earthquake occurred in Stanley, USA. The China Earthquake Network determined that the epicenter was at 44.46º N, 115.13º W, and the focal depth was 10 km (http://www.cenc.ac.cn/). The epicenter is 30 km from Stanley City and 120 km from Boise City. As of 24:00 on April 15, a total of 392 aftershocks were recorded, including 5 earthquakes of magnitude 4.0 to 4.9 and 133 earthquakes of magnitude 3.0 to 3.9. The largest aftershock was the magnitude 4.8 earthquake that occurred at 12:27 on April 1. According to the United States Geological Survey (USGS) survey on earthquake intensity in the earthquake zone, the maximum earthquake intensity caused by this strong earthquake in the earthquake zone is VIII.
2020Stanley earthquake occurred at the northern border of sawtooth fault. After the Stanley earthquake, literature [1] found that the serrated fault boundary extending northwards was squeezed by the remains of the Challis fault, which made the seismic activity in the area more active in the next few years. Literature [2] inverted the fault sliding distribution of this earthquake based on the lifting rail InSAR deformation field, mainly concentrated in depth 10~15 km underground, and found that there is a complex after-slip phenomenon, but there is no good correlation between the line of sight (LOS) simulated deformation and the observed deformation. The horizontal deformation of the Stanley earthquake was calculated using GPS data and the source position was inverted. This paper studies the co-seismic deformation caused by the Stanley earthquake using the lifting orbit SAR image of the Sentinel-1A satellite (https://www.asf.alaska.edu/) of the ESA Sentinel-1A satellite. Based on the elastic semi-space dislocation theory, the calculated LOS deformation was used to establish the co-seismic sliding distribution model of the Stanley earthquake. The triggered static coulomb failure stress (CFS) was evaluated. Combined with the spatial distribution of aftershocks, the impact of regional seismic activities was analyzed, and the potential earthquake hazards and their significance in the study area were explored to make up for the shortcomings of geological tectonic research in this area.
Geological structure
Idaho, USA (Idaho), is located at the geopolitical intersection of the basin expansion boundary and mountain range squeeze. The basin and mountains extend in the east and west directions since the Miocene, forming a normal fault boundary. The hidden faults in the basin generally develop along the north-south direction [3]. The fault in the northwest direction of at the junction of central Idaho basin and mountain ranges induces the Ms7.3 magnitude earthquake in Hebgen Lake in 1959, resulting in obvious surface rupture and forming fault cliffs. The 1983 Borah Peak Ms6.9 earthquake may lead to the formation of a normal fault east of the epicenter area [4].The Stanley earthquake occurred at the boundary between the interaction between the basin and the mountains. Some scholars believe that the extension and extrusion of the basin and the mountains are caused by the extension and extrusion of the basin and the mountains, and it shows that the existing older silent faults are reactivated under the action of different structures and expand along the mountains to form a new normal fault. Under the action of extrusion of different directional forces, it becomes a strike-slip fault [4]. Most of the faults in the trans-Challis fault zones develop into regular faults, controlling the topographic structure of the region. The
Stanley earthquake occurred on the northern tectonic belt of the Snake River plain. The Snake River plain migrated eastward and northeastward at a certain rate, forming a new tectonic belt in the Yellowstone area. It is an area with frequent modern tectonic movements and seismic activities. Existing research has found that the normal faults present on the tectonic belt develop locally into strike-slip fault [5], with a large number of fault branches and steep ridges distributed. Most surface ruptures are limited to a certain section of the fault, making the fault movement mostly dominant with positive tilt sliding, and at the same time, they also have a little left-hand component [6]. The earthquake is located about 16 km NA of the sawtooth fault. The sawtooth fault was active 12,000 years ago. It is a normal fault with an inclination length of about 60 km along the northeast. Due to mountain ranges and basin extension, the sawtooth fault deflected in the northwest near 44º N, which affects the occurrence of folds and hidden faults on the tectonic belt [7]. Under the influence of fractures distributed along the east-west direction, the tectonic activity of the western section of the sawtooth fault has a strike-slip component, while the faults distributed along the north-south direction control the expansion of the plain boundary. Figure 1 shows the aftershocks and fault distribution of the seismic region of the Stanley earthquake mapped using data from the United States Geological Survey (USGS). It can be seen that the aftershocks are distributed in the north-south direction near the epicenter, parallel to the direction of the expansion of the plain boundary. In the past 30 a, strong seismic activity in the epicenter area was sparse, and the existing active structures showed that there were no other obvious faults in the epicenter area, so there was no clear explanation for the seismic structure of this earthquake. The Stanley earthquake was the largest earthquake in Idaho since the 1983 Borah Peak Earthquake Ms 6.9 magnitude earthquake and was considered a left-hand strike-slip event.
InSAR Co-seismic deformation observation
The Sentinel-1A launched on April 3, 2014 short return period, convenient data acquisition, and has been widely used to obtain surface deformation caused by earthquakes, landslides and human activities. Sentinel-1A C-band data (5.6 cm), with a width of up to 250 km, and a single image can complete the full coverage of the deformation area caused by the Stanley earthquake. Precision orbit data and wide-range (IW) orientation high-resolution data (Table 1) are from the European Space Agency (ESA). NASA uses the National Aeronautics and Space Administration (NASA) to release the 90-m resolution shuttle radar topography mission (SRTM) DEM to eliminate the impact of terrain ups and downs, and interference phase noise is reduced through adaptive filtering. The phase unwrap is performed using the minimum cost flow [8] (minimum cost flow, MCF) algorithm. The atmospheric phase [9] related to the terrain is fitted based on the correlation between the terrain and the atmospheric phase. The partial error is weakened based on the above data processing method and the differential interference displacement is obtained. Finally, the InSAR co-seismic deformation field under geographic coordinates is obtained by geocoding.
The InSAR image used in this study covered the entire deformation field caused by the Stanley earthquake. Except for the snow and vegetation covered areas, other areas have good coherence. As shown in Figure 2(a), the LOS direction deformation observed by the ascending track shows that the uplifting volume of the north deformation center reaches 24.2 cm, the deformation distribution is dense and the range is large, and the settlement volume of the south deformation center reaches 12.5 cm, and the distribution range is small and sparse. As shown in Figure 2 (b), the LOS direction deformation observed by the descending orbit shows that the maximum uplift is 22.3 cm, and an irregular settlement distribution of 14.1 cm is observed in the epicenter area, which may be related to the concentrated deformation of aftershocks. These characteristics suggest that earthquakes in areas near the epicenter may rupture deeper at fault in depths. Away from the center of the seismic source, the LOS directional deformation variable gradually decreases, which is consistent with the physical mechanism of seismic activity [10].
fault sliding distribution
To gain an in-depth understanding of the seismic source mechanism of this earthquake and further explore fault motion, after obtaining InSAR co-seismic deformation, in order to improve the efficiency and accuracy of inversion, the quadtree method [11] (quadtree method) is used to downsample the LOS to the deformation field, that is, set a small threshold for the near-field areas with large deformation gradients, and the sampling is relatively dense. A large threshold is set for the far-field areas with small deformation gradients, and the sampling interval is relatively sparse, which preserves the spatially related characteristics of the deformation field to the maximum. After downsampling, there are 264 and 360 sampling points for the lifting orbit for inversion. Based on the empirical error of the one-dimensional covariance function, the relative weight ratios of the sampled rail data and descending rail data are set to 1:0.86[11], and the two-step inversion method [12] is used to invert the source parameters of this earthquake to obtain the sliding distribution of the fault.
.1 fault geometric parameters
is based on Okada elastic semi-space dislocation theory [13] ignores the influence of geological stratification, assuming that the sliding amount on the fault is a constant, combining the nonlinear relationship between fault parameters and surface deformation, as shown in Equation (1).
where: dInSAR represents the observation value of the co-seismic deformation field; m represents the geometric parameters of the fault (including epicenter position, source depth, fault length, width, direction angle, inclination angle, sliding angle, etc.); G is the Green function; ε represents the observation error. The simplex algorithm [14-15] is used to loop iterate in a multivariate function to search for the optimal solution of the earthquake fault parameters (Table 2). The Monte Carlo algorithm [16] (Monte Carlo bootstrap Simulation Technique) is used to estimate the weight and uncertainty of the parameters. In Figure 3, 100 simulation models under the noise interference of the statistical characteristics of the one-dimensional covariance function are used to estimate the standard deviation of their distribution.
Table 2 2020 Stanley MW6.6 earthquake source parameters
Note: Results: Fault parameters obtained by inversion; SD: Standard deviation; GCMT: Global Centroid Moment Tenso.
.2 finite fault model
determines the fine sliding distribution on the fault surface through linear inversion, extending the fault length and width to 50 km along the direction and 30 km along the downward tilt direction respectively. Assuming that the seismic fault is a smooth rectangular plane, the fault plane is evenly divided into 600 single-segment 2 km × 2 km single-segment shaping deformation observations, and the least squares algorithm of fault boundary constraint inverts the sliding amount of each sub-fault, and adds the Laplace smooth constraint [17] to avoid the divergence of the sliding distribution solution. The equation to be solved is shown in Equation (2).
where: dInSAR is the observation value of surface deformation; m2 is the fault geometric parameter; G is the Green function between the fault parameter and the surface deformation value; κ2 represents the smoothing factor; h represents the Laplace smoothing operator; ε represents the observation error. The resulting sliding distribution is shown in Figure 4. The best fault model obtained by
InSAR observation inversion shows that the rupture process of this earthquake is mainly dominated by left-hand strike-slip motion, and the fault rupture did not reach the surface. There is a significant slip area along the fault direction 40 km and downward 28 km. At 12 km underground, the maximum sliding volume of the fault can reach 0.9 m, with an average sliding angle of -38º. The released seismic moment is 8.72×1018 N·m, and the moment magnitude MW6.56. The direction and aftershock direction lines of the seismic fault determined by the InSAR deformation field are similar to the northward direction of the sawtooth fault. It is judged that the seismic fault of this earthquake may be a hidden fault in the northeast direction of the sawtooth fault. The inclination and direction angle obtained by inversion differ from the results given by USGS and GCMT, which are close to the node parameters given by literature [2]. Analysis believes that near-field geodescending data is more sensitive to fault geometry constraints than far-field seismic wave data. The inversion results in this paper are attached to InSAR co-seismic deformation field constraints, and the differences may be due to the different types and methods of inversion data.According to the optimal fault model, based on the Okada elastic semi-space dislocation theory, the forward fitting deformation field is simulated and calculated respectively. Overall, the residuals of the simulated LOS displacement are mostly less than 3 cm. Statistically, it can be concluded that the standard deviations corresponding to the lifting and lowering rails are 1.39 cm and 2.18 cm respectively, indicating that this fault model can better explain the surface deformation caused by this earthquake, and the residuals in the near-field area of the deformation may be related to the atmospheric phase, concentrated deformation aftershocks, and the impact of gravity on faults [18-19].
Figure 5 Fault sliding distribution inversion fitting
4 Static Coulomb stress
Cohomicolar stress
Cohomicolar sliding on faults may affect the state of adjacent faults and surrounding areas and the seismic activity frequency [20-22] through stress transfer. The stress state of the fault and surrounding areas is a key indicator for future seismic probability assessment. Increased Coulomb stress will promote fault rupture activity, while reduced Coulomb stress will inhibit fault rupture [16]. Although the Coulomb stress after the earthquake is smaller than the tectonic stress required by the earthquake, a large number of domestic and foreign research results show that the low Coulomb stress of [23-25], 0.01 Mpa is enough to trigger an earthquake.
CFS[26] is a quantitative description of stress states that can be used to describe the potential destructiveness of rock formations, which helps further explore the mechanisms of related earthquakes and can also provide new starting points for earthquake prediction. In order to evaluate the effect of Stanley earthquake on seismic activity in the surrounding area, this paper calculates the effect of this earthquake co-seismic dislocation on the Coulomb stress distribution in the surrounding area at three different depths based on the sliding distribution results, and analyzes the impact of the main earthquake on the occurrence of aftershocks. The effective friction coefficient and shear modulus [27-28] are set to 0.4 and 3.32×1010N/m2, respectively. The calculation results show that (Figure 6) the Stanley earthquake produced a significant CFS reduction in the rupture area of the seismic fault. In the area where CFS is significantly increased, no major earthquakes have occurred in history. The disturbing stress caused by the 2020 Stanley earthquake may make it instable under further tectonic loading, causing other seismic events. Most aftershocks occur within 5~10 km, triggered by Stanley's main shock. The area where CFS increases has fewer aftershock distribution, and there are sparse sections within the range of 10~15 km. The sparse aftershock area is located at a place where the fault slides are large. Analysis believes that before the earthquake, CFS was effectively released when the main shock rupture, so there were fewer aftershocks, and in areas with smaller fault sliding volume, the distributed aftershocks may be CFS-triggered [29-30]. The spatial distribution of main shock and aftershocks was analyzed, and the dense distribution of aftershocks along the northeast at 44.3º N was observed. Combined with the analysis of the fine positioning aftershock model of literature [1], serrated fractures may develop into slippery fractures. Overall, this area is a earthquake-prone area and geological structures are still active.
junction buoy
2020 Stanley earthquake produced maximum deformation along the Sentinel-1A satellite rise and descending orbit LOS directions, respectively, to 24.2 cm uplift and 14.1 cm settlement. The coseismic deformation field based on InSAR observations simulates the coseismic sliding distribution of the Stanley earthquake in the United States in 2020. The inverted fault model shows that the fault motion is dominated by left-hand strike-slip motion and there is a small amount of tilt-slip component. The maximum sliding volume of the fault is 0.9 m, and the total seismic moment released is 8.72×1018N·m, corresponding to the moment magnitude Mw 6.56. Fault direction angle of fitting the optimal sliding distribution is 156°, inclination angle of 58°, and sliding angle -38°. It shows that the Stanley earthquake fault sliding occurs on a fault surface with a relatively gentle inclination angle, which is significantly different from the steep tilt sliding characteristics of sawtooth fault [31-32]. As shown in Figure 4(a) and Figure 4(b), the shallow part of the fault surface (depth 0~3 km) and the deep part (depth 25 km) have almost no sliding. Having studied [30] shows that the friction properties on the fault surface are usually uneven, that is, the sliding of the friction stability zone at the upper edge of the fault and the sliding of the friction-reducing zone alternately interact from the upper part of the fault to the depth of the earth's crust. Therefore, it is speculated that the friction characteristics of the area around the seismic zone of the Stanley earthquake in 2020 (depth 0~3 km, depth 25 km) are enhanced. This also shows that due to the enhanced friction characteristics of the shallow part of the fault, the earthquake fault movement did not cause surface rupture.The squeeze of the basin and the mountains leads to the northward flow of material in the central Snake River Plain, which is very likely to induce earthquakes and is blocked by deep underground rock foundations. The aftershock distribution of Stanley earthquake is parallel to the northern expansion boundary of the plain. Since aftershocks often occur on the same fault as the main shock, aftershocks can provide information on fault structure (such as position, inclination, depth). Combined with the spatial distribution characteristics of the InSAR co-seismic deformation field in Figure 2, the downward inclination distribution along the sliding surface and the fault sliding distribution model, it is determined that the seismic fault of the Stanley earthquake is a hidden fault developing northeast along the sawtooth fault, and this area may develop into a tilt fault at 44.3º N. Other more accurate results rely more on geological surveys and other multivariate data fusions. The CFS calculation results show that the Coulomb stress along the sawtooth fracture increased by at least 0.1 Mpa, and the tectonic movement in the epicenter area is active, and the earthquake risk in the area needs to be paid attention to.
Author Profile : Jiang Xin (1996—), male, from Huining, Gansu, master's degree, with a main research direction in geodesics. E-mail: [email protected]
