Analysis of spatiotemporal variations in b-values before the 6.8-magnitude earthquake in Luding, Sichuan, China, on September

Using the earthquake catalog provided by the Sichuan Earthquake Network Center, spatial and temporal b-value scans were calculated for large-and small-scale regions based on assessing the completeness of the earthquake catalog and aftershock removal. The results show that (1) b-values in the large-scale region ranged from 0.689 to 1.169, with a mean value of 0.928, while the b-values in the small-scale region ranged from 0.694 to 1.223, with a mean value of 0.925. The b-values in the study area were below the mean value before the medium and strong earthquake occurrence, and all exhibited the anomalous feature of a sudden drop-low peak rise. (2) The sliding rate of the northwest section of the Xianshui River Fracture Zone was higher than that of the southeast section; therefore, a large amount of stress was accumulated in the mill-west section of the southeast section, leading to a 6.8-magnitude earthquake in Luding. Before the earthquake, the study area was a low b-value area. The b-value decreased within a short period after the earthquake, dividing the area into concave and convex bodies. This area still has a future risk of moderate to strong earthquakes. (3) The error in the b-values for most of the earthquakes in the large-and small-scale regions is between 0.05 and 0.15, and only individual grid points have larger b-value errors (>0.2), indicating high con�dence in the information. In addition, when conducting a b-value study, choosing a suitable study area is important to avoid missing the b-value anomaly area.


Introduction
The b-value is an important parameter in the empirical magnitude-frequency formula, , which was derived from the study of global seismicity patterns (Gutenberg and Richter 1944).In the formula, N denotes the frequency of earthquakes above magnitude M, a denotes the level of seismicity in the region, and b denotes the proportional relationship between earthquakes of different magnitudes in the region.Experimental rock results have shown that the b-value is inversely proportional to the stress magnitude, and that low b-value regions have high stress accumulations (Scholz 1968;Amitrano 2003;Schorlemmer et al. 2005).In particular, there is a clear trend of decreasing b-values before a rock rupture or an earthquake tremor.
Many seismologists have studied the physical signi cance of the b-value in depth, and many case studies have corroborated the phenomenon of lower b-values often occurring in the source and adjacent areas before earthquakes.For example, Nanjo and Yoshida (2021) studied the variation in the b-value in and around the source area of magnitude 6.9 and 6.8 earthquakes off the coast of Miyagi Prefecture, Japan.The authors found that the earthquakes occurred near a region with very small bvalues, even after the earthquake.Zeng et al. (2020) studied the changes in the b-values before an Ms-6.0 earthquake in Changning, China, and found that low b-value anomalies (≤ 0.85) were present in the epicenter area and adjacent areas before the Changning earthquake, with a decrease in the b-values near the epicenter ve months before the earthquake.Jiang and Feng (2021) studied the characteristics of the pre-earthquake b-value anomaly in Jiuzhaigou, China, for a 7.0-magnitude earthquake and found that this region had signi cantly low b-value anomalies before the earthquake.Furthermore, Xie et al. (2022) studied the pre-seismic b-value variation of a 6.0-magnitude earthquake in Luxian, China, and found that anomalous low b-value features occurred in and around the source area.In summary, a long-term b-value decrease in a certain area re ects the high internal stress of the medium.The possibility of a signi cant rupture and a strong earthquake increases in these cases, so special attention should be paid to these areas.Spatiotemporal b-value scan results can provide information about the locations of future strong earthquakes and can be used to analyze the likelihood of strong earthquakes.They can reveal and infer the relative levels of stress accumulation in active ruptures at different stages, outline possible concave and convex bodies or closed fracture segments, and further determine the substantial earthquake hazard posed by active ruptures.The b-value has gradually become an effective tool for judging the level of regional stress, inferring the locations of future strong earthquakes, and determining potential earthquake source areas.
On September 5, 2022, at 12:52, a 6.8-magnitude earthquake occurred in Luding County, Ganzi Prefecture, Sichuan Province, China, at a depth of 16 km.The epicenter was located near the Moxi Fracture in the southeastern section of the Xianshui River Fracture Zone (29.59°N, 102.08°E), and the earthquake was a mainshock-aftershock type earthquake with a maximum intensity of IX.This earthquake caused 97 deaths, and 20 people were missing.The direct economic losses amounted to 154.80 billion yuan.Located on the eastern edge of the Qinghai-Tibet Plateau, the Xianshui River Fracture Zone is a northwesttrending arc-shaped left-lateral strike-slip fracture zone.This zone is also the boundary of the slip activity between the Bayankara Block and the Sichuan-Yunnan Block, starting in the north near the east valley of Ganzi and extending in the northwest-southeast direction, with a dip angle of roughly 55-80°.The zone passes through Luhuo, Daofu, Qianning, and Kangding and intersects with the Longmenshan Fault Zone and the Anning River Fault near asbestos, having a total length of approximately 400 km.
The Xianshui River Fracture Zone is divided into two major sections, northwest and southeast, with the Huiyuan Temple La Division Basin as the boundary.The northwest section has a sliding rate of less than 8.4 mm/yr, while the southeast section has a sliding rate of 4.0-5.2mm/yr (Liang, 2019).The northwest section can be subdivided into the Luhuo section (sliding rate of 9.13 mm/a), the Daofu section (sliding rate of 8.57 mm/a), and the Qianning section (sliding rate of 7.67 mm/a), with a relatively homogeneous geometry.The southeast section can be subdivided into the Kangding section (sliding rate of 6.14 mm/a) and the Moxi section (sliding rate of 4.41 mm/a) (Li et al., 2019).The structure of the Kangding section is relatively complex, and it is mainly composed of three fractures, including the Seraha Fracture (sliding rate of 1.2 mm/yr), Foldotang Fracture (sliding rate of 1.3-3.4mm/yr), and Yala River Fracture (sliding rate of 0.7-1.0mm/yr), which are nearly parallel to each other.
The seismic activity within the Xianshui River Rift Zone is strong, with more than 50 earthquakes with magnitudes of 5.0 or higher occurring since 1700, including eight earthquakes with magnitudes of 7.0 or higher (Fig. 1).The surface rupture caused by the earthquakes covers almost the entire section of the rupture.The most recent strong earthquake with magnitudes of 5.8-6.3occurred on November 22, 2014, more than a year before the strong earthquake in Luding.The Sichuan-Yunnan rhombic massif and the Bayankara massif exhibit the characteristic of strong earthquake activity with a high frequency.There have been a number of earthquakes in the Sichuan-Yunnan rhombic massif and in the boundary or interior of the Bayankara massif: on May 21, 2021, there were 5.8-6.4-magnitudeearthquakes in Yangbi County, Dali, Yunnan; on May 22, 2021, there was a 7.4-magnitude earthquake in Maduo County, Qinghai; on January 2, 2022, there was a 5.5-magnitude earthquake in Ninglang County, Lijiang, Yunnan; on June 1, 2022, there was a 6.1-magnitude earthquake in Lushan County, Yaan, Sichuan; and on June 10, 2022, there was a 6-magnitude earthquake in Markang, Sichuan.
In view of this, we calculated the spatial and temporal b-value scans of the earthquake catalog before the Luding 6.8magnitude earthquake in a large-scale region (100-103.1°E,29-32°N) and a small-scale region (29-30.5°N,101.5-103°E).Based on error analysis, we also analyzed the spatial and temporal variation characteristics of the pre-earthquake b-values and the in uence of different scales on the spatial and temporal variation characteristics of the pre-earthquake b-value.In addition, the spatiotemporal image characteristics of the b-value were interpreted in relation to the distribution of moderate to strong earthquakes and active ruptures.The low b-value zones were further delineated to determine the potential seismic hazard zones.
2 Information and methods

2.1Data sources and preprocessing
The study area is the source area of the 6.8-magnitude Luding earthquake and the adjacent areas (100-103.1°E,29-32°N), and the seismic catalogs used in this study were all obtained from the Sichuan Earthquake Network Center between January 1, 2020, and September 17, 2010.A total of 47,719 earthquakes with an M ≥ 0.1 were collected in the catalog, including four earthquakes with magnitudes of 6.0 or higher; 10 earthquakes with magnitudes of 5.0-5.9;71 earthquakes with magnitudes of 4.0-4.9;554 earthquakes with magnitudes of 3.0-3.9;3,915 earthquakes with magnitudes of 2.0-2.9;22,261 earthquakes with magnitudes of 1.0-1.9;and 20,904 earthquakes with magnitudes of 1.0 or less.The empirical relationship between the national M S and M L was obtained from the statistical regression of Wang et al. (2010) as M L = M S .Therefore, in this paper, M S and M L are not converted into each other, and the earthquake magnitude scale is directly expressed as M uniformly.
It is important to remove the foreshocks and aftershocks and to analyze the minimum magnitude of completeness (M c ) of the seismic catalog to obtain a reasonable and reliable seismic catalog for further study without the in uence of data selection.In this study, we rst selected the empirical relationship based on the time and distance windows of the seismic events proposed by Uhrhammer (1986) to remove the foreshocks and aftershocks and obtain the independent mainshock catalog.After deletion, a total of 20,749 earthquakes with magnitudes of M ≥ 0.1 were collected, including three earthquakes with magnitudes of 6.0 or higher, two earthquakes with magnitudes of 5.0-5.9, 31 earthquakes with magnitudes of 4.0-4.9,225 earthquakes with magnitudes of 3.0-3.9,2,206 earthquakes with magnitudes of 2.0-2.9,11,439 earthquakes with magnitudes of 1.0-1.9, and 6,843 earthquakes with magnitudes of less than 1.0.Second, the integrity magnitude range method (Ogata and Katsura 1993) was used to calculate the minimum integrity magnitude M c in the study area, obtained from Fig. 3 as being 1.4.

2.2Method of calculating b-value
Common methods used to calculate the b-value include the maximum likelihood (MLE) method (Aki 1965;Utsu 1965;Shi and Bolt 1982) and the least squares method (LSM) (Okal and Kirby 1995;Main 2000;Zöller et al. 2002).When the least squares method is used to calculate the b-value, the cumulative frequency is generally used to reduce the effect of the earthquake binning, emphasizing the role of earthquakes containing a larger amount of rich information and providing a small weighting for earthquakes with lower magnitudes.The maximum likelihood method for calculating b-values was proposed by Aki et al. (1965).In this method, the study area is rst gridded, after which the b-value of each grid is calculated, thereby reducing the effect of a few earthquakes on the overall b-value results for the area.This method for calculating its b-value is expressed as follows: 1 The standard deviation of the b-values is calculated as follows: 2 where M c is the minimum integrity magnitude and ΔM is the magnitude binning.
In this paper, the magnitude interval is taken as 0.1, is the average magnitude.e is the natural constant, and n is the sample size used to calculate the b-value.Using Equations ( 1) and ( 2), the maximum likelihood method averages the magnitudes of all earthquakes with the same weight, equivalent to weighting the information of a larger number of small earthquakes.Since this paper primarily analyzes the regional seismic activity using the small-earthquake catalog, the maximum likelihood method was chosen to calculate the seismic activity parameters, i.e., the a-value and b-value.
3 Calculation of b-values in a large-scale region

Characteristics of the temporal distribution of the b-values
The calculated b-values for the study area from 2010 to 2022 range from 0.689 to 1.169, with a mean value of 0.928.The solid blue line in Fig. 4(a) shows the location of the mean value.The analysis of the earthquake catalog in the study area during the period when the b-value was greater than the average and the graph of the magnitude variation over time in the study area (Fig. 4(b)) show that the earthquakes in the study area were mainly small earthquakes, with only a few moderate to strong earthquakes.Most of the moderate to strong earthquakes occurred during the period when the b-value was less than the average.Most of the small earthquakes occurred during the period when the b-value was above the average.This observation is consistent with the conclusion by Tormann et al. (2015) that the global average b-value is about 1, the b-value is less than 1 in areas with many moderate to strong earthquakes, and the b-value is greater than 1 in areas with many small earthquakes.
Figure 4(a) shows that the b-values in the study area uctuated widely, and the b-values suddenly decreased and increased several times.The periods with signi cant b-value changes in the region were selected and analyzed in conjunction with the moderate to strong earthquakes.Several moderate to strong earthquakes (indicated by the solid red line in Fig. 4(a)) were preceded by an abrupt decrease in the b-values, which dropped to a minimum and then gradually increased to a maximum peak.The earthquakes occurred during the period from the minimum to the maximum peak in the b-value, including the earthquake investigated in this study.However, the magnitude of the drop in the b-values before each of the moderate to strong earthquakes was different.For the earthquakes with serial numbers 1, 3, 4, and 7, small decreases in the b-values occurred three months before these earthquakes.For those with serial numbers 2, 5, 6, and 8, large decreases in the b-values occurred three months before the earthquakes.The relationship between the decrease in the b-values and the magnitude of the moderate to strong earthquakes is still unclear and requires further study.At this stage, we can only note whether the preearthquake b-value decreases during a speci c period to determine whether an earthquake has occurred.The decrease in the b-values was not evident one to two months before the earthquake with serial number 4, but it was apparent starting one year before the earthquake.Therefore, when analyzing the b-value change before an earthquake, choosing the right time to study is very important.In addition, for the earthquake with serial number 7, the b-value decreased insigni cantly more than one month before the earthquake, but six months before the earthquake, the b-value decreased signi cantly.
In this paper, a spatial interval of 0.05°×0.05° is used for gridding.Taking each grid cell's node as the center of a circle with a radius of 50 km, earthquakes within these circular statistical units with an M ≥ M c were selected.The b-value spatial scan was performed using the MLE method.Grids with fewer than 50 earthquakes were not included in the calculation to ensure su cient samples and are shown as blank areas in Fig. 5.In Fig. 5, the different colors represent different b-values, and the spatial scan results show that the b-values of this region are between 0.6 and 1.5.In this section, the spatial and temporal variations in the b-values in the source and neighboring areas and the Xianshui River Rift Zone before this earthquake are analyzed in terms of both temporal and spatial evolution.The Kangding earthquake on November 22, 2014, which occurred in the Xianshui River Rift Zone, as well as the Lushan earthquake on April 20, 2013, and the Lushan earthquake in the Longmenshan Rift Zone on June 1, 2022, were also selected for analysis of the spatial and temporal variations in the b-values in the source and neighboring areas before the earthquakes.
The spatial variations shown in Figs.5a-f re ect the differences in the stress accumulation levels of the different ruptures and, thus, the spatial variability of the seismic hazard.The results show slightly higher b-values in the southeast section of the Xianshui River Rift Zone than in the northwest section.As can be seen from the temporal evolution shown in Figs.5a-f, the bvalues differ in the different periods.Before the Luding earthquake, the b-values of the source area and adjacent areas gradually decreased from 1.15 in period I (Fig. 5a) to 0.9 in period IV (Fig. 5d).The b-values in the source and adjacent areas further decreased shortly after the earthquake, and their values dropped to 0.8 (Fig. 5f).
Similarly, the b-value of the Xianshuihe Fault Zone gradually decreased and is now at a lower b-value, indicating that this fault zone is prone to accumulating stress and experiencing moderate to strong earthquakes.In this paper, we conclude that the sliding rate of the northwest section of the Xianshui River Fracture Zone is higher than that of the southeast section, such that the northwest section slides toward the southeast section at a higher rate.The extrusion pressure caused by the northwest section accumulates at a decreasing rate interval, resulting in a higher stress accumulating in the southeast section.In addition, the sliding rate of the Kangding section in the southeast section is 6.14 mm/a, and the sliding rate of the mill-west section is 4.41 mm/a.The mill-west section accumulates a higher stress and has a higher possibility of rupture, which makes it more prone to earthquakes.Hence, the 6.8-magnitude Luding earthquake occurred in the mill-west section of the southeast section of the Xianshui River Fracture Zone.
The spatial distribution of the b-values in Fig. 5 is block-like or band-like, which is correlates well with the distribution of the regional fractures, especially the Xianshui River Fracture Zone and Longmenshan Fracture Zone.Studies have shown that concave and convex bodies or closed fracture segments with high stress accumulation in active fracture zones are prone to strong earthquakes (Aki 1984; Wiemer and Wyss 1997; Wyss et al. 2000).A concave-convex body can be interpreted as the area of stress concentration and strength on the fault surface before the earthquake.This area is the starting point of a rupture, or else the most severe damage occurs at this location during an earthquake (Xu et al. 2022).The closed fracture section is where strain energy or stress concentration is easily accumulated.Therefore, according to the spatial variations in the b-value with a block or band distribution, the region can be delineated as concave and convex bodies or as a closed fracture section.The concave and convex bodies, or bumps, with serial numbers 1, 2, and 3 can be clearly outlined from Figs. 5a-f.The concave and convex bodies' b-value size and spatial location change with time.
The No. 1 bump is located in the Longmenshan Fault Zone, where the M6.1 earthquake on June 1, 2022, and the M7.1 earthquake on April 20, 2013, occurred.The pre-earthquake b-values dropped to less than 0.7 for the M6.1 earthquake and less than 0.9 for the M7.1 earthquake.This is also consistent with the above conclusion that it is unclear how the magnitude of the b-value drop relates to the magnitude of moderately strong earthquakes, but the pre-earthquake b-values were both in the low b-value region.
The No. 2 bump is located in the source area and adjacent areas of the Luding earthquake.In addition, the southeast section of the Xianshui River Fracture Zone is prone to accumulate stress and can be classi ed as a closed fracture section.The No. 3 bump is located to the left of the M6.3 Kangding earthquake that occurred on November 22, 2014, and a small spatial change in the bump's location has occurred.As can be seen from Figs. 5a-b and 5c-d, the b-value of the bump changed abruptly, re ecting the abrupt change in stress, indicating that the concave-bump body activity was intense and should be taken seriously.The b-value decreased before the M6.3 earthquake, dropping to less than 0.7.
In summary, the pre-earthquake b-values decreased to less than 1 for the above three strong earthquakes.The above three concave and convex bodies suggest the locations of future strong earthquakes and require continuous attention.However, from the overall perspective (Fig. 5e), the b-values of the southeast section of the Xianshui River Fracture Zone, the source area, and the neighboring areas are around 1, and their anomalous features are not very obvious.Therefore, selecting data for the appropriate year for the b-value spatial scan is important.Usually, we divide the data into periods based on the monitoring capability of the seismic network, the occurrence of large earthquakes in that period, and whether the spatial and temporal variations in the b-value are apparent.

Spatial scan error estimation of b-values
Factors such as the starting magnitude of the sample, sampling interval, sample size, minimum integrity magnitude, magnitude binning, and the spatial and temporal extents of the sampling all directly affect the maximum likelihood error of the b-value.In order to check whether the b-value calculation results are reliable, Eq. ( 2) is used to calculate the b-value tting error in this paper.Generally, the more sample sizes involved in the calculation within each grid area, the smaller the b-value error is.Therefore, the sample size used to t the b-value of each grid point requires no less than 50 samples (Fig. 6).The results show that most earthquakes have b-value errors in the range of 0.05-0.1,and only individual grid points have larger bvalue errors (> 0.2), suggesting high con dence in the information.
4 Calculation of b-values in a small-scale region

Characteristics of the temporal distribution of the b-values
In this paper, we further analyze the b-values in a small-scale area near the source (29-30.5°N,101.5-103°E) to obtain a more obvious relationship between the b-value changes and this earthquake and to verify the conclusions drawn from the largescale area.Figure 7(a) shows that from 2010 to 2022, the b-values in the study area ranged from 0.694 to 1.223, with a mean value of 0.925.The solid blue line in Fig. 7(a) shows the location of the mean value, which is not signi cantly different from the large-scale regional mean b-value of 0.928 (Fig. 4a).In conjunction with Fig. 7b, it can be seen that all of the moderately strong earthquakes with M ≥ 4.0 in this region occurred at low b-values (less than the average b-value).The b-value also dropped back to a lower value before the Luding earthquake; before the moderately strong earthquakes, the b-values tended to be below the average value.This is consistent with the conclusion obtained for the large-scale region, but the phenomenon observed in the small-scale region is more pronounced.In addition, the b-values in the study area uctuate widely, and the bvalues have suddenly decreased and increased many times.Before the occurrence of several moderate to strong earthquakes, the b-values suddenly decreased; after the earthquakes, the b-values slowly rebounded.Combined with the temporal distribution characteristics of the large-scale b-values, it can be determined that the b-value changes can re ect the changes in the stress state.The abnormal characteristics can be used as a sign of the occurrence of moderate to strong earthquakes in the region.

Spatial distribution characteristics of b-values
The results of the b-value time scan for the large-and small-scale regions show a correspondence between the increase or decrease in b-values and the occurrence of moderate to strong earthquakes, which can be used as a reference for earthquake prediction in the future.However, the structure of the study area is complex, so the time scan features only re ect the possibility of earthquake occurrence on the time scale of the study area, and the location of earthquake occurrence cannot be discerned.Therefore, a spatial scan of the b-values should be performed for the study area, and an explanation should be given for the areas in the study area where the b-values are in an abnormal state.In this paper, the study area is divided into six time periods when conducting the b-value spatial scan, as in the large-scale region.Since the region's area is small, the scan radius is reduced to 40 km when the gridding is conducted, and the other parameters are the same as for the large-scale region.
The calculation results are shown in Fig. 8.The small-scale region can better re ect the spatial distribution of the b-values and further verify the conclusions drawn from the results for the large-scale region.In this region, the b-values ranged from 0.7 to 1.6, as seen from the temporal evolution shown in Figs.8a-f.The b-values in the source area and neighboring regions gradually decreased before this earthquake, from 1.2 in period I (Fig. 8a) to 0.85 in period IV (Fig. 8d).The b-values in the source area and neighboring regions further decreased to 0.75 within a short time after the earthquake (Fig. 8f).These results are not much different from the change in the b-values in the large-scale region, i.e., a zone of signi cantly low b-values.
Similarly, the b-value of the Xianshui River Fracture Zone has gradually decreased, and it has been at a lower b-value, which needs our attention.In addition, the abrupt change in the b-values in this region is evident from Figs. 8a-b, indicating that the region is highly active.Spatially, the b-value of the southeast section of the Xianshui River Fault Zone is higher than that of the northwest section, and the concave and convex bodies with serial numbers 1, 2, and 3 can be clearly outlined.Compared with the concave and convex bodies outlined in the large-scale area, the anomalous characteristics of the concave and convex bodies in the small-scale area are more pronounced, and it can be seen that the low b-value area migrates farther with time.
The low b-value area is usually accompanied by the occurrence of moderate to strong earthquakes, while the high b-value areas are scattered throughout the study area.Figures 8d and 8f show the results of the spatial scan of the b-values two years before and after this earthquake.Before this earthquake, the source area and its nearby areas had low b-values, and the bvalue magnitude decreased further in a short time after the earthquake, making this a signi cant low-anomaly area.

Spatial scan error estimation of b-values
This study adopted Eq. ( 2) to calculate the b-value tting error for the small-scale region.Since the area of the region is small, the sample size used in the tting of the b-value of each grid point is required to be at least 40 (Fig. 9).The results show that the error of the b-values for most of the earthquakes is between 0.05 and 0.15.The error of the b-values calculated for individual grid points is signi cant (> 0.2), this corresponds to the large-scale regional ndings with high con dence in the information.
5 Relationship between the b-value and medium-intensity earthquakes and ruptures The determination of the low b-value zone and the medium-strong earthquake risk zone depends to a large extent on the knowledge of the active rupture in the region.Therefore, a comprehensive understanding of the development degree of the active ruptures in the region and their activity and tectonic system is a necessary prerequisite for studying the b-value hazard zones and analyzing medium to strong earthquake hazards.The study region is located at the Tibetan Plateau's eastern edge, a strong earthquake activity area.Fracture tectonics are developed in the region.The Holocene active fractures mainly include the Xianshui River Fracture Zone, Longmen Mountain Fracture Zone, Anning River Fracture, Daliang Mountain Fracture, Yulongxi Fracture, Ganzi-Litang Fracture, and Anning River Fracture, which were intensely active during the Quaternary, and especially since the Late Quaternary.These fractures are seismic faults for strong and large earthquakes.Table 1 presents a list of the features of the main active fractures in the region.Figure 10 shows the spatial distribution of the major ruptures, earthquakes with magnitudes of 3 or higher, and b-values within the region.The data further illustrate the relationship between the b-value and moderate to strong earthquakes and ruptures.
As can be seen from Fig. 10, there are apparent differences in the spatial distributions of the medium and strong earthquakes and the b-values along each rupture zone.Most of the medium and strong earthquakes in the region occurred in the rupture and its adjacent areas.In particular, the most earthquakes occurred in the Xianshui River Rupture Zone, Longmenshan Rupture Zone, and Anning River Rupture, re ecting that these three rupture zones are the most active, strongest, and most dangerous rupture zones in Sichuan Province.The b-value of the Xianshui River Rift Zone is less than 1, and this is an area with many moderate to strong earthquakes.The moderate to strong earthquakes in the Xianshui River Rift Zone are mainly concentrated in the southeast section and the moderate to strong earthquakes in the southeast section are mainly concentrated in the millwest section.This is also consistent with the conclusion that the sliding rate of the northwest section of the Xianshui River Rift Zone is higher than that of the southeast section and the mill-west section.
The Longmenshan Fracture Zone is composed of the Maowen-Wenchuan Fracture, the Beichuan-Yingxiu Fracture, and the Jiangyou-Guanxian Fracture.It is a boundary fault between the Sichuan-Qinghai Block of the Qinghai-Tibet Plateau and the Sichuan Basin in Southern China extending along the Longmen Mountains, with a total length of about 500 km and a width of 40-50 km.The area affected by this fault can be divided into a concave-convex body, and the b-value of the concave-convex body is less than 0.8, making this an area prone to moderate to strong earthquakes.The An Ninghe Fracture Zone only spreads to the northern end of its northern section, and its characteristics are not as obvious as those of the Xianshuihe Fracture Zone and the Longmenshan Fracture Zone, the b-values of which are less than 1.However, it is also an area prone to moderate to strong earthquakes.The three ruptures mentioned above need to be focused on, and they are likely to produce moderate to strong earthquakes in the future and are strong earthquake danger zones.The Daliangshan Fault, Yulongxi Fault, Ganzi-Litang Fault, Anning River Fault, and Daliangshan Fault are prone to small earthquakes, are not prone to moderate to strong earthquakes, and their b-values are approximately 1.This paper analyzes and discusses the earthquake catalogs before and after the 6.8-magnitude Luding earthquake on two regional scales, large and small, based on the earthquake catalogs provided by the Sichuan Earthquake Network Center for 2010/01/01-2022/09/17.The following conclusions can be drawn from this study.
(1) The characteristics of the temporal distribution of the b-values show that the b-values in the large-scale region ranged from 0.689 to 1.169, with a mean value of 0.928, whereas the b-values in the small-scale region ranged from 0.694 to 1.223, with a mean value of 0.925.The b-values were below the average value before the occurrence of strong earthquakes in the study area, and all of them experienced a sudden decrease, a decrease to a minimum, and then a gradual increase to a maximum value.
(2) The spatial distribution characteristics of the b-values show that the b-values of the southeast section of the Xianshui River Fracture Zone are slightly higher than those of the northwest section, and they further outline the concave and convex bodies or closed fracture sections with serial numbers 1, 2, and 3 in the region according to the spatial variation of the bvalues with block-like or band-like distribution characteristic.The No. 1 bump is located in the Longmenshan Fault Zone, the No. 2 bump is located in the source area and adjacent areas of this earthquake, and the No. 3 bump is located to the left of the 2014/11/22 M6.3 Kangding earthquake.The b-value size and spatial location of the bump change over time and may be the location of future strong earthquakes, so it requires continuous attention.
(3) The spatial low b-value anomaly can be used to determine the potential risk area for future moderate to strong earthquakes.The Xianshui River Fracture Zone, Longmen Mountain Fracture Zone, and Anning River Fracture have many small and moderate to strong earthquakes with b-values of less than 1.They are currently in a high stress locking state (with concave and convex bodies or a locking fracture section), have a high probability of producing moderate to strong earthquakes in the future, and require continuous attention.The Daliangshan Fault, Yulongxi Fault, Ganzi-Litang Fault, Anning River Fault, and Daliangshan Fault are in the high b-value region of the study area, and they currently have relatively low-stress accumulation levels, with minor earthquake activity dominating in the future.
(4) The error of the b-values for most of the earthquakes in the large-and small-scale regions is in the range of 0.05-0.15,and only the errors of the b-values of individual grid points are larger (> 0.2).This information is highly reliable.When the b-values are calculated via spatial and temporal scanning, the analysis of a large-scale area is bene cial to identifying more b-value anomalous areas, but it masks the anomalies in local areas.The analysis of a small-scale area can better represent the bvalue variation characteristics of the study area and can more clearly outline the anomalous characteristics of the concave and convex bodies.Therefore, when performing b-value calculations, an appropriate study area should be selected to avoid missing b-value anomalies.

Figure 2 Distribution
Figure 2

Figure 3 Minimum
Figure 3

Table 1
List of main active fractures in the region