Assessment of Strong Earthquake Risk in Maqin–Maqu Segment of the Eastern Kunlun Fault, Northeast Tibet Plateau

: The East Kunlun Fault Zone, as a highly seismically active fault, has witnessed five earth-quakes with magnitudes exceeding M7.0 to the west of Animaqing Mountain since 1900. Conversely, the historical records for the Maqin–Maqu segment in the east of Animaqing Mountain show no M7.0 or above earthquakes, designating it as a distinctive seismic gap within this fault zone. We analyzed the tectonic background and structural features of the Maqin–Maqu segment within the East Kunlun Fault Zone to evaluate its potential seismic capacity. Utilizing a new established probability recurrence model, we calculated the seismic hazard for both segments over the next 100 years. The results indicate that the probability of M7.0 or above earthquake occurring in the Maqu segment in the next 100 years is 11.47%, classified as a moderate probability event. The joint probability of at least one M7.0 or above strong earthquake occurring in the entire Maqin–Maqu segment in the next 100 years is 16.14%, also classified as a moderate probability event, while the probability for the Maqin segment alone is 5.36%, classified as a low probability event. Considering the uncertainty of the probability model, a qualitative hazard classification for each segment was further conducted. The comprehensive evaluation suggests a low risk of a major earthquake occurring in the Maqin segment in the next 100 years, while the Maqu segment is assessed to have a higher risk.


Introduction
The Eastern Kunlun Fault constitutes a significant fault zone characterized by primarily sinistral strike-slip movement.It originated along the Eastern Kunlun Variscan paleotectonic suture within the Tibetan Plateau due to the collision between the Indian and Asian plates.This fault spans approximately 2000 km, following a general west-northwest trend, extending from the west of Bukedaban Peak, situated north of Hoh Xil Lake in Qinghai province, to the Maqu region in Sichuan province [1] (Figure 1).Situated at the northern margin of an upwelling area in the middle of the Tibetan Plateau, the primary portion of this fault zone lies approximately along the 64-km depth contour of the Moho surface, extending in an east-west direction.An isobathic line from the southeast of Maduo-Maqu to the Mohorovicic discontinuity shifts southward, coinciding with a gradual reduction in crustal thickness along this trajectory [2].Recent seismic refraction data reveal that the crustal thickness of the BaiYan Har Mountain Group, positioned in the southern sector of the fault zone, measures 64 km.This thickness notably decreases to 60 km as one progresses northward, reaching the southern rim of the Qaidam Basin and the Touraine area.This creates a region characterized by variable crustal thickness in the west-northwest direction.The Moho topography and crustal flow mechanism within this fault zone have experienced discernible shifts, delineated by the Aemye Ma-chhen Mountains in an east-west orientation [3,4].As a result, this paper employs the Aemye Ma-chhen Mountains as a reference to partition the Eastern Kunlun Fault into eastern and Since 1900, five earthquakes with magnitudes exceeding 7.0 have occurred in the western part of the Aemye Ma-chhen Mountains in the Eastern Kunlun fault, encompassing the Ms 7.9 Mani earthquake in 1997, Ms 8.1 Kunlun Mountains earthquake in 2001, Ms 7.0 Xiugou earthquake in 1902, Ms 7.1 Akake Lake earthquake in 1963, and Ms 7.5 Tuosuohu earthquake in 1937 (Figure 1).As depicted in Figure 1, there is a notable absence of documented large earthquakes (Ms ≥ 7) in the historical records of the eastern segment of the Eastern Kunlun Fault, marking it as a distinctive seismic gap within this fault zone.Moreover, the Wenchuan earthquake of 2008 heightened the accumulation of Coulomb stress within the eastern segment of the Eastern Kunlun Fault zone [5], consequently elevating the latent risk of earthquakes in this section.Therefore, the Maqin-Maqu segment, constituting the eastern part of the Eastern Kunlun Fault zone, emerges as the seismic gap with the highest population density along the entire fault zone.As such, there is an exigent need to scrutinize the earthquake risk and forecast potential large-scale seismic events.
To predict earthquakes in the medium and long term, a quantitative approach involves expressing the likelihood of earthquake occurrence as a probability [6,7].Assuming that the time for earthquake occurrence approximately follows a specific recurrence model [8,9], we can employ this model to estimate the probability of recurrent major earthquakes within a defined time frame [10][11][12].Many research findings indicate that the model may exhibit various distribution types, including normal, logarithmic normal, and Weibull distributions.In the northeastern Tibetan Plateau, occurrences of severe earthquakes are relatively frequent, and research on ancient earthquakes along the fault zone is notably Since 1900, five earthquakes with magnitudes exceeding 7.0 have occurred in the western part of the Aemye Ma-chhen Mountains in the Eastern Kunlun fault, encompassing the Ms 7.9 Mani earthquake in 1997, Ms 8.1 Kunlun Mountains earthquake in 2001, Ms 7.0 Xiugou earthquake in 1902, Ms 7.1 Akake Lake earthquake in 1963, and Ms 7.5 Tuosuohu earthquake in 1937 (Figure 1).As depicted in Figure 1, there is a notable absence of documented large earthquakes (Ms ≥ 7) in the historical records of the eastern segment of the Eastern Kunlun Fault, marking it as a distinctive seismic gap within this fault zone.Moreover, the Wenchuan earthquake of 2008 heightened the accumulation of Coulomb stress within the eastern segment of the Eastern Kunlun Fault zone [5], consequently elevating the latent risk of earthquakes in this section.Therefore, the Maqin-Maqu segment, constituting the eastern part of the Eastern Kunlun Fault zone, emerges as the seismic gap with the highest population density along the entire fault zone.As such, there is an exigent need to scrutinize the earthquake risk and forecast potential large-scale seismic events.
To predict earthquakes in the medium and long term, a quantitative approach involves expressing the likelihood of earthquake occurrence as a probability [6,7].Assuming that the time for earthquake occurrence approximately follows a specific recurrence model [8,9], we can employ this model to estimate the probability of recurrent major earthquakes within a defined time frame [10][11][12].Many research findings indicate that the model may exhibit various distribution types, including normal, logarithmic normal, and Weibull distributions.In the northeastern Tibetan Plateau, occurrences of severe earthquakes are relatively frequent, and research on ancient earthquakes along the fault zone is notably extensive.In this study, through the collection and analysis of data regarding ancient earthquakes, we ascertain sample parameters and derive the probability density function governing earthquake recurrence in this region.This model is subsequently applied in assessing the seismic probability for the Maqin-Maqu segment of the Eastern Kunlun fault zone.

Regional Tectonic Setting
The Maqin-Maqu segment, situated at the confluence of Gansu, Sichuan, and Qinghai provinces, constitutes the eastern boundary fault for the Qaidam and Bayan Har blocks.It spans from the northern base of the Aemye Ma-chhen Mountains in the west, passing the north sides of Dongqinggou, Dawutan, Kendingna, Xigongzhou, and Tangdi.Progressing through Keshengtuoluo where it traverses the Yellow River, it extends into the Ruoergai Basin via Maqu.Eventually, it emerges in the swampland north of Luocha, intersecting with the Tacang fault.The total length approximates 310 km.Its overall trend aligns with 295 • northwest, with a prevailing southwest inclination [2].
By delineating tectonic indicators along the segmented boundary of the strike-slip fault [6], it can be divided, from west to east, into the Maqin segment and Maqu segment (Figure 2A).The calculated second strain rate invariant indicates that the strain rate is mainly distributed along the Maqin-Maqu segment of the East Kunlun Fault Zone.In the vicinity of the intersection of Animaqing Mountain and the West Gong Co-seismic Fault, there is a high strain rate zone with maximum values of approximately 65 nstrain•a −1 and 70 nstrain•a −1 , respectively.East of Maqu, the strain rate is diffusely distributed between the Gongga Shan-Dieshan Fault and the East Kunlun Fault Zone, ranging from 30 to 40 nstrain•a −1 (Figure 2B) [13].Boundary markers between these segments and the western segment include the uplift of Animaqing Mountain (exhibiting double bending as a marker for segmentation of seismic ruptures) [14,15] (Figure 3A-C), the region of fault intersection and decomposition (triple junction where the Maqin fault intersects with the Awancang fault near the Xike River as a marker for segmentation of seismic ruptures) [16] (Figure 2A-C), and the pull apart basin of the left rank (namely, the Maqu fault aligns with the left rank of the Tacang fault through Ruoergai Basin (Figure 3D,E).Considering the Holocene slip rates for these three segments (Figure 2C), it is evident that each has maintained its distinctive level of activity throughout this period, with the slip rate gradually decreasing from west to east.This trend aligns consistently with the broader Eastern Kunlun fault zone, where activity diminishes in the same west-to-east fashion.

Database
This study entails the collection of data pertaining to ancient earthquakes across different segments of 22 active fault zones in the northeastern Tibetan Plateau (refer to Appendix A Research data Table A1).Data selection adheres to three primary criteria: (1) Collection, sorting, and comparative analysis of data to scrutinize ancient earthquake events exposed by various trench profiles within the same fault segment, while excluding repetitive events.(2) Exclusion of ancient earthquake events that transpired an excessive length of time ago.(3) Computation of the recurrence interval (T) for each fault segment to establish the average recurrence interval of ancient earthquakes (T ave ).Subsequently, the ancient earthquake data (T/T ave ) for each segment are normalized to mitigate data dispersion.Finally, a total of 119 data points (T/T ave ) are available for statistical analysis based on the aforementioned principles.

Methodology and Model
Employing SPSS 26.0 statistical software, we construct a frequency distribution histogram illustrating the recurrence probability (T/T ave ) of ancient earthquakes in the northeastern Tibetan Plateau (Figure 4). Figure 4 reveals that while variations in recurrence probabilities among different ancient earthquakes are substantial, they notably diminish after normalization processing, resulting in a distribution skewed towards values near 1.0.The concentration of T/T ave is confined within a range proximate to 1.0.Furthermore, we posit that this frequency distribution may conform to potential fits with a normal distribution, logarithmic normal distribution, and Weibull distribution.Subsequent sections utilize the Q-Q plots to evaluate potential distribution forms Figures 5-7.

Database
This study entails the collection of data pertaining to ancient earthquakes across different segments of 22 active fault zones in the northeastern Tibetan Plateau (refer to Appendix A Research data Table A1).Data selection adheres to three primary criteria: (1) Collection, sorting, and comparative analysis of data to scrutinize ancient earthquake events exposed by various trench profiles within the same fault segment, while excluding repetitive events.(2) Exclusion of ancient earthquake events that transpired an excessive length of time ago.(3) Computation of the recurrence interval (T) for each fault segment to establish the average recurrence interval of ancient earthquakes (Tave).Subsequently, the ancient earthquake data (T/Tave) for each segment are normalized to mitigate data dispersion.Finally, a total of 119 data points (T/Tave) are available for statistical analysis based on the aforementioned principles.

Methodology and Model
Employing SPSS 26.0 statistical software, we construct a frequency distribution histogram illustrating the recurrence probability (T/Tave) of ancient earthquakes in the northeastern Tibetan Plateau (Figure 4). Figure 4 reveals that while variations in recurrence probabilities among different ancient earthquakes are substantial, they notably diminish after normalization processing, resulting in a distribution skewed towards values near 1.0.The concentration of T/Tave is confined within a range proximate to 1.0.Furthermore, we posit that this frequency distribution may conform to potential fits with a normal distribution, logarithmic normal distribution, and Weibull distribution.Subsequent sections utilize the Q-Q plots to evaluate potential distribution forms Figures 5-7.In statistics, a Q-Q plot (quantile-quantile plot) serves as a probability plot, offering a graphical method for comparing two probability distributions by plotting their quantiles against each other.Each point (x, y) on the plot corresponds to a quantile of the second distribution (y-coordinate) plotted against the same quantile of the first distribution (x-coordinate).This results in a parametric curve where the parameter is the index of the quantile interval [20].
The primary purpose of a Q-Q plot is to compare the shapes of distributions, providing a visual representation of how properties such as location, scale, and skewness are either similar or different in the two distributions.Q-Q plots can be applied to compare collections of data or theoretical distributions.Utilizing Q-Q plots for comparing two data samples can be seen as a non-parametric approach to assessing their underlying distributions.While Q-Q plots are generally more diagnostic than comparing sample histograms, they are less widely known.Commonly, Q-Q plots are employed to compare a dataset to a theoretical model, offering a graphical assessment of goodness of fit rather than reducing to a numerical summary statistic.Additionally, Q-Q plots are used to compare two theoretical distributions to each other.As Q-Q plots compare distributions, there is no requirement for the values to be observed as pairs, as in a scatter plot, or for the numbers of values in the two groups being compared to be equal [21].In statistics, a Q-Q plot (quantile-quantile plot) serves as a probability plot, offering a graphical method for comparing two probability distributions by plo ing their quantiles against each other.Each point (x, y) on the plot corresponds to a quantile of the second distribution (y-coordinate) plo ed against the same quantile of the first distribution (xcoordinate).This results in a parametric curve where the parameter is the index of the quantile interval [20].
The primary purpose of a Q-Q plot is to compare the shapes of distributions, providing a visual representation of how properties such as location, scale, and skewness are either similar or different in the two distributions.Q-Q plots can be applied to compare collections of data or theoretical distributions.Utilizing Q-Q plots for comparing two data samples can be seen as a non-parametric approach to assessing their underlying distributions.While Q-Q plots are generally more diagnostic than comparing sample histograms, they are less widely known.Commonly, Q-Q plots are employed to compare a dataset to a theoretical model, offering a graphical assessment of goodness of fit rather than reducing to a numerical summary statistic.Additionally, Q-Q plots are used to compare two theoretical distributions to each other.As Q-Q plots compare distributions, there is no requirement for the values to be observed as pairs, as in a sca er plot, or for the numbers of values in the two groups being compared to be equal [21].
Based on the aforementioned analysis, in conjunction with Figures 5-7, it is evident that the normal distribution yields the most favorable fit, followed by the Weibull distribution, with the logarithmic normal distribution performing the least satisfactorily.Consequently, this study opts for the normal distribution as the probability density function.T/Tave is expected to adhere to the normal distribution: T/Tave~N(µ, σ 2 ), µ represents the expected value of the normal distribution; it can be interpreted as a position parameter.σ represents the standard deviation; it can be interpreted as the amplitude of the distribution.
Based on the statistical analysis results based on Figure 4, µ = 0.9538, σ = 0.3879, T/Tave is expected to adhere to the normal distribution N(0.9538, 0.3879 2 ), and the fi ing result successfully passes the χ 2 test within the 95% confidence interval.This implies that recurrent events in the same location possess an approximate 95% confidence interval falling within [µ−2σ, µ + 2σ].Subsequently, the probability density function for T/Tave conform- Based on the aforementioned analysis, in conjunction with Figures 5-7, it is evident that the normal distribution yields the most favorable fit, followed by the Weibull distribution, with the logarithmic normal distribution performing the least satisfactorily.Consequently, this study opts for the normal distribution as the probability density function.T/T ave is expected to adhere to the normal distribution: T/T ave ~N(µ, σ 2 ), µ represents the expected value of the normal distribution; it can be interpreted as a position parameter.σ represents the standard deviation; it can be interpreted as the amplitude of the distribution.
Based on the statistical analysis results based on Figure 4, µ = 0.9538, σ = 0.3879, T/T ave is expected to adhere to the normal distribution N(0.9538, 0.3879 2 ), and the fitting result successfully passes the χ 2 test within the 95% confidence interval.This implies that recurrent events in the same location possess an approximate 95% confidence interval falling within [µ − 2σ, µ + 2σ].Subsequently, the probability density function for T/T ave conforming to the normal distribution can be derived as follows: The probability distribution function for the northeastern margin of the Tibetan Plateau can be derived as follows: In Equation ( 2), Φ represents the standard normal distribution function, R denotes the average recurrence interval of earthquakes at a given location, and t signifies the elapsed time (the duration between the last earthquake and the present).
Employing the aforementioned probability distribution function, the conditional probability (P) of earthquake recurrence can be expressed by the following equation: In Equation (3), t represents the elapsed time (the duration between the last earthquake and the present), R denotes the average recurrence interval of earthquakes, while ∆t signifies the time increment used to project the probability of future earthquake occurrences.To compute the conditional probability of earthquake recurrence, only the three fundamental parameters-t, R, and ∆t-are required.
Regarding the conditional probability model for earthquake recurrence, Zhang (1996) [12] conducted a comprehensive analysis by collating recurrence interval data from 46 ancient earthquakes within the continental interior.Employing the logarithmic normal distribution, Zhang fitted the cumulative frequency and established a universal distribution relationship applicable to all active faults within dynamic continents [12]: Applying the maximum likelihood method, we obtain estimates for the mean (µ = 0.1206) and standard deviation (σ = 0.5054).The resulting probability distribution function is as follows: In the subsequent sections, we employ the previously mentioned two distribution models to examine the earthquake risk within the Maqin-Maqu segment of the Eastern Kunlun fault zone.Subsequently, we conduct a comparative analysis and discussion of the results, thereby mutually validating their feasibility.

Estimation of Largest Earthquake Magnitudes
In 1994, American scholars D. L. Wells et al. [22] established an empirical relationship between the surface rupture length and the moment magnitude of strike-slip earthquakes based on a large number of samples, which is known briefly as the WC empirical relationship: Mw = 5.16 + 1.12LgL, where Mw represents the moment magnitude, L represents fault length, and the residual standard deviation is 0.28 [23].When applying the WC empirical relationship, we encounter the problem of conversion between moment magnitude Mw and surface wave magnitude Ms. D. L. Wells suggested in the 1994 work that when Ms is in the range of 5.7-8.0,there is no systematic difference between them, namely, Mw ≈ Ms.However, there is a difference between the magnitude determination method and station data adopted in China and the United States; namely, Ms and Mw are not the same in the continent of China.In 2009, Ran [24] derived a transformation relationship for these two earthquake magnitudes according to seismic data in mainland China during 1973-2008: Ms = 1.412 + 0.845Mw, where Ms represents surface wave magnitude, Mw represents moment magnitude, and the residual standard deviation is 0.11 [23].In this paper, we use the aforementioned equation of statistical relationship and combine it with the length of the fault to calculate the maximum earthquake magnitude that could occur in each segment (Table 1).Based on the above analysis, regardless of Mw or Ms, the earthquake may be induced by the fault belonging to the strong earthquake (Ms ≥ 7.0).For the evaluation of seismic risk, we must comprehensively understand the average recurrence interval of ancient earthquakes in this area and the elapsed time since the last earthquake.Quantitative studies on ancient earthquakes in the east segment of the Eastern Kunlun fault zone began in the 1990s and achieved a succession of research findings.However, the difference between results is relatively large.In this paper, we compare the research results of previous studies and obtain relatively reasonable recurrence intervals for ancient earthquakes and elapsed times since the earthquakes, which provide comparatively accurate source materials for calculating the recurrence probability for each segment.

Maqin Segment
The tectonic geomorphology preserved in the segment at the surface, such as earthquake swells and fault scarps, reflects that the segment had multiple active episodes during the Holocene Epoch.Previous research utilized methods such as excavating trenches and using natural outcrops to identify ancient earthquake events along this segment during different stages (Figure 8).Various researchers [17,[24][25][26] determined different periods and times of ancient earthquakes, as shown in Figure 8. Li (2009) established that the most recent age of the buried ancient surface is 358-430 Cal a BP from the footwall (Tc3) of a secondary bench fault profile east of Gequ River, the sampling location of which is an ancient plant layer [18].However, no records of a relevant earthquake were found after referring to historical documents covering nearly 400 years.Therefore, this age still requires further investigation.It is found from Figure 8 that when the age of an ancient earthquake is 977-1090 Cal a BP, the conclusions of researchers are essentially consistent [17,24] and agree with the time of an earthquake recorded in one passage of the 'Epic of King Gesar' historical literature.It is speculated that this age is relatively reliable, and it can be used as the latest elapsed time for the occurrence of an earthquake in this segment.Finally, by comparing and analyzing the main research results, ancient earthquake events of the Maqin segment can be mainly divided into six events: the first is 977-1090 Cal a BP; the second is 2000 ± 300 Cal a BP; the third is 3342-3454 Cal a BP; the fourth is 6600 ± 700 Cal a BP; the fifth is 7971-8050 Cal a BP; and the sixth is 10,000 ± 200 Cal aBP.The calculated average recurrence period of ancient earthquakes is approximately 1765 a.
King Gesar' historical literature.It is speculated that this age is relatively reliable, and it can be used as the latest elapsed time for the occurrence of an earthquake in this segment.Finally, by comparing and analyzing the main research results, ancient earthquake events of the Maqin segment can be mainly divided into six events: the first is 977-1090 Cal a BP; the second is 2000 ± 300 Cal a BP; the third is 3342-3454 Cal a BP; the fourth is 6600 ± 700 Cal a BP; the fifth is 7971-8050 Cal a BP; and the sixth is 10,000 ± 200 Cal aBP.The calculated average recurrence period of ancient earthquakes is approximately 1765 a.

Maqu Segment
The seismic activity along the Maqu fault is clearly weaker than that along the Maqin fault.Ancient earthquake data of this period were also analyzed intensively by predecessors (Figure 9).There are two trenches near the Xike River and four near Maqu County [2,25,26].In this paper, we compare the research results of previous studies and derive the elapsed time of earthquakes and the average recurrence interval of earthquakes.By comparing the three profiles of ancient earthquakes near Maqu County, He (2006) used the

Maqu Segment
The seismic activity along the Maqu fault is clearly weaker than that along the Maqin fault.Ancient earthquake data of this period were also analyzed intensively by predecessors (Figure 9).There are two trenches near the Xike River and four near Maqu County [2,25,26].In this paper, we compare the research results of previous studies and derive the elapsed time of earthquakes and the average recurrence interval of earthquakes.By comparing the three profiles of ancient earthquakes near Maqu County, He (2006) used the successive restriction method proposed by Zhang (1996) for ancient earthquake events to obtain comparatively accurate events.Four events occurred during the Holocene Epoch, considering that the first and second times of ancient earthquakes are relatively reliable [12,25].However, because the number of prospecting trenches is small, the record of ancient earthquakes is still incomplete.Lin (2008) recognized five ancient earthquake events by lamination and sampling of three natural outcrops and two trench sections in the Maqu segment [26].The location of trench sections among them matches the He (2006) research result, while the obtained age result is also similar.In this paper, by comparing the research results of previous studies, we further combine the sampling positions of outcrops and trench profiles and dating methods to describe seven ancient earthquake events that are relatively reliable: the first is 1210 ± 40 Cal a BP; the second is 1730 ± 50 to 2530 ± 40 Cal a BP; the third is 3736 ± 57 Cal a BP; the fourth is 4850 ± 40 Cal a BP; the fifth is 6100-6700 Cal a BP; the sixth is 8590 ± 70 Cal a BP; and the seventh is 9000-10,000 Cal a BP.We can determine that the past event of the most recent ancient earthquake is 1210 ± 40 Cal a BP, and the average recurrence time period of ancient earthquakes is approximately 1465 a. [12,25].However, because the number of prospecting trenches is small, the record of ancient earthquakes is still incomplete.Lin (2008) recognized five ancient earthquake events by lamination and sampling of three natural outcrops and two trench sections in the Maqu segment [26].The location of trench sections among them matches the He (2006) research result, while the obtained age result is also similar.In this paper, by comparing the research results of previous studies, we further combine the sampling positions of outcrops and trench profiles and dating methods to describe seven ancient earthquake events that are relatively reliable: the first is 1210 ± 40 Cal a BP; the second is 1730 ± 50 to 2530 ± 40 Cal a BP; the third is 3736 ± 57 Cal a BP; the fourth is 4850 ± 40 Cal a BP; the fifth is 6100-6700 Cal a BP; the sixth is 8590 ± 70 Cal a BP; and the seventh is 9000-10,000 Cal a BP.We can determine that the past event of the most recent ancient earthquake is 1210 ± 40 Cal a BP, and the average recurrence time period of ancient earthquakes is approximately 1465 a.

Qualitative Classification of Risk Degrees in Maqin-Maqu Segments
Based on the fact that the average recurrence intervals of earthquakes calculated in this paper are long, we conduct a qualitative classification on the degree of earthquake risk in the Maqin-Maqu segment of the Eastern Kunlun fault zone.We usually use E = t/R to describe the time urgency of earthquake occurrence on the faults, E represents the quantitative parameter of describing the time urgency of earthquake occurrence, t represents

Qualitative Classification of Risk Degrees in Maqin-Maqu Segments
Based on the fact that the average recurrence intervals of earthquakes calculated in this paper are long, we conduct a qualitative classification on the degree of earthquake risk in the Maqin-Maqu segment of the Eastern Kunlun fault zone.We usually use E = t/R to describe the time urgency of earthquake occurrence on the faults, E represents the quantitative parameter of describing the time urgency of earthquake occurrence, t represents the elapsed time since the last earthquake, R represents the average recurrence interval, while, to some extent, we ignore the contribution of R to the time urgency of earthquake occurrence.Even the value of E is the same.In two segments with different R values, the degrees of earthquake risk are different.Generally, the smaller the value of R, the higher is the degree of risk.Therefore, we can use two parameters to divide the degree of risk: (1) it is class C if E < 0.5 and R > 2500 years, which indicates that there is no risk in this section within one thousand years; (2) it is class B if 0.5 < E < 0.8 and R < 2500 years, which indicates that there is a relatively small risk in this segment during the coming several hundred years; and (3) it is class A for 0.8 < E < 1 and R < 2500 years, which indicates that the risk is relatively large in this segment during the coming several hundred years.E takes 0.5, 0.8, and 1.0 as the critical values, which are determined according to the distribution characteristics of earthquake recurrence T/T ave , and R is 2500 years because the recurrence intervals along the two segments are smaller than 2500 years.As the prediction time increases, the increase of risk is also relatively rapid [28].We divide the segments according to this principle as follows: the Maqin segments belong to class B, and the Maqu segment belongs to class A. According to the probability distribution function (2) fitted in this paper and the probability distribution function (5) suitable for all active faults in mainland China, we combine the elapsed time and average recurrence interval of the last earthquake for Maqin-Maqu segments in the Eastern Kunlun active fault zone and calculate the conditional probability of earthquake recurrence in the future 20 years, 50 years, and 100 years after 2023 for different segments of this fault zone (Table 2).
In 1984, Wallace et al. [29] analyzed the probability of seismic risk and proposed that recurrence probabilities in the range of 0-10% are considered low; recurrence probabilities in the range of 11-49% are considered intermediate probability events; and recurrence probabilities in the range of 50-100% are considered high probability events.Given the recurrence probabilities calculated in Table 2, the recurrence probability of the Maqu segment in the next 100 years indicates an intermediate probability event, and the Maqin indicates low probability events.
Finally, we find that a comparison result between the probability density function fitted in this study and the conditional probability values calculated using commonly used generic functions reveals that the generic recurrence probability function becomes less sensitive to the dependent variable P (conditional probability) with an increase in the independent variable t/R compared to the fitted function proposed in this paper.According to the aforementioned recurrence probability of different fault segments in Table 2 (i.e., the fitting function used in this paper), we calculate the joint probability for the occurrence of at least one large earthquake in the Maqin-Maqu segment of the Eastern Kunlun active fault zone in the future: P1, P2 are the probabilities for the occurrence of an earthquake in the Maqin segment and Maqu segment of the Eastern Kunlun fault.
Table 3 shows that the probability for the occurrence of a strong earthquake in the next 20 years in the east segment of the Eastern Kunlun fault zone is low, and the probabilities for the occurrence of an earthquake in the next 50 and 100 years are intermediate.The probabilities in Table 3 likely represent only the minimum probability values of this active fault zone.Because the data for ancient earthquakes are likely somewhat inadequate, such as the lack of certain active events, they overestimate the recurrence interval and the value of elapsed time.

Conclusions
We conducted an analysis of the tectonic background and structural characteristics pertaining to the Maqin-Maqu segment within the East Kunlun Fault Zone.The segment was divided into two seismogenic structures, and an evaluation was performed to determine the potential maximum magnitude of earthquakes that these structures could trigger.Additionally, a recurrence probability model for the region was established by collecting and organizing paleoseismic data from the northeastern part of the northeastern Tibetan Plateau.Subsequently, the model was employed to assess the seismic hazard associated with the Maqin-Maqu segment of the East Kunlun Fault Zone over the next century.The acquired insights and prevailing issues can be summarized as follows: (1) The calculated recurrence probability values for strong earthquakes in the Maqin-Maqu segment of the Eastern Kunlun fault zone reveal distinct patterns.Over the next 100 years, the Maqu segment presents an intermediate probability of 11.47%, whereas the Maqin segments exhibit values below 10%, signifying events of low probability.Although current research results indicate that the Maqin segment exhibits stronger tectonic activity compared to the Maqu segment, such as a decreasing trend in the sliding rate of the East Kunlun Fault from west to east, the seismic hazard in the next century is closely associated with the time since the last earthquake.The Maqu segment has an elapsed time of around 1200 years, signifying a seismic gap on the East Kunlun Fault throughout the last millennium.Furthermore, it has experienced Coulomb stress loading since the occurrence of the 2008 Wenchuan earthquake.Consequently, there is a critical need to concentrate on monitoring and mitigating the seismic hazard in the Maqu segment.
(2) The sensitivity of conditional probability P, calculated from the strong earthquake recurrence probability model to the prediction time period ∆t, is intricately linked with the length of the recurrence cycle R.In general, as R increases, the probability value P experiences a gradual rise with the expanding prediction time interval ∆t.However, when R is exceedingly large and ∆t is very small, the prediction outcome may be relatively imprecise, underscoring a limitation of the recurrence model in earthquake prediction.The recurrence intervals for different segments in the east segment of the Eastern Kunlun fault zone all hover around one thousand years.While the calculated conditional probabilities are generally modest, they still offer valuable insights into the seismic activity levels of these three segments for several centuries into the future.
(3) We utilized quantitative data on ancient earthquakes pertaining to 22 fault zones in the northeast Tibetan Plateau and subjected it to a normalization process to construct a recurrence probability model for strong earthquakes.A comparative analysis of probability values calculated using the model proposed in our paper and the prevailing universal seismic model reveals that, when the independent variable of the universal model (t/R) approximates 1, the amplitude for the growth of the calculated recurrence probability value (P) is less sensitive compared to the fitting model in this paper.Consequently, for fault zones with a substantial body of studies providing data on ancient earthquakes, it is arguably more accurate to employ the model developed in this paper for assessing the future risk of a large earthquake, particularly in segments characterized by small average recurrence intervals and extended elapsed times.Furthermore, the current universal recurrence model is likely more applicable to faults with a relatively limited number of studies available on ancient earthquakes and segments with relatively long recurrence intervals.

Figure 1 .
Figure 1.Topographic map of the northern Tibetan Plateau region.Major active faults are adapted from Deng et al., 2007 [1].The upper-left figure shows the location of Northern Tibet.The Blue dashed box indicates the location of Figure 2. The Eastern Kunlun Fault zone is represented by bold colorful lines, with historic earthquake surface ruptures marked by red or yellow lines.Purple lines represent the faults for which paleoseismology records are used in this study.Detailed earthquake timing for these faults is listed in Table A1.Fault name abbreviations: BYF, North Bayinguole River fault; EHF, Eastern Helanshan piedmont fault; ELF, Eastern Luoshan piedmont fault; EYF, Eastern Yumushan piedmont fault; FHF, Fodongmiao-Hongyazi Fault; HSF, Huangcheng-Shuangta fault; THF, Tianqiaogou-Huangyangchuan fault; TLSF, Tuolaishan fault; YF, Yumushan fault; ZTF, Zhongwei-Tongxin fault.

Figure 1 .
Figure 1.Topographic map of the northern Tibetan Plateau region.Major active faults are adapted from Deng et al., 2007 [1].The upper-left figure shows the location of Northern Tibet.The Blue dashed box indicates the location of Figure 2. The Eastern Kunlun Fault zone is represented by bold colorful lines, with historic earthquake surface ruptures marked by red or yellow lines.Purple lines represent the faults for which paleoseismology records are used in this study.Detailed earthquake timing for these faults is listed in Table A1.Fault name abbreviations: BYF, North Bayinguole River fault; EHF, Eastern Helanshan piedmont fault; ELF, Eastern Luoshan piedmont fault; EYF, Eastern Yumushan piedmont fault; FHF, Fodongmiao-Hongyazi Fault; HSF, Huangcheng-Shuangta fault; THF, Tianqiaogou-Huangyangchuan fault; TLSF, Tuolaishan fault; YF, Yumushan fault; ZTF, Zhongwei-Tongxin fault.

Figure 2 .
Figure 2. (A) Fault trace distribution of the Maqin-Maqu segments.The fault has been divided into different segments based on geometrical complexity, such as restraining bends, triple junctions, and stepover (marked by pink arrows and dashed circles).(B)The calculated second strain rate invariant indicates that the strain rate is mainly distributed along Maqin-Maqu of the East Kunlun Fault Zone.Data were compiled from Wen Y M et al. (2023) [13].(C) Compilation of slip rates showing the partitioning of slip rates at the Maqin-Maqu-Awancang Triple Junction [17-19].

Figure 2 .
Figure 2. (A) Fault trace distribution of the Maqin-Maqu segments.The fault has been divided into different segments based on geometrical complexity, such as restraining bends, triple junctions, and stepover (marked by pink arrows and dashed circles).(B)The calculated second strain rate invariant indicates that the strain rate is mainly distributed along Maqin-Maqu of the East Kunlun Fault Zone.Data were compiled from Wen Y M et al. (2023) [13].(C) Compilation of slip rates showing the partitioning of slip rates at the Maqin-Maqu-Awancang Triple Junction [17-19].

Figure 3 .
Figure 3. (A) Hillshaded DEM and (B) Oblique Google Earth image showing the geometry and shape of the Anemaqen Transpressive Range and restraining double bend.(C) Topographic profile extracted from SRTM 30 m DEM showing the relief of the range.The profile location is shown in Figure 3A.(D) A simplified geological map illustrating the fault traces of the Maqu Segment and Tazang Fault.The buried Maqu fault segment was determined based on unpublished geophysical

Figure 3 .
Figure 3. (A) Hillshaded DEM and (B) Oblique Google Earth image showing the geometry and shape of the Anemaqen Transpressive Range and restraining double bend.(C) Topographic profile extracted from SRTM 30 m DEM showing the relief of the range.The profile location is shown in Figure 3A.(D) A simplified geological map illustrating the fault traces of the Maqu Segment and Tazang Fault.The buried Maqu fault segment was determined based on unpublished geophysical profiles.(E) A Google Earth image showing the Tazang Fault obvious sinistral strike-slip displaced mountain ridges, gullies, and Holocene alluvial fans [19].

Figure 4 .
Figure 4. Histogram of normalized data of ancient earthquake recurrence intervals in the northeastern Tibetan Plateau T/Tave and the normal distribution fitting curve.T represents recurrence interval of ancient earthquake, Tave represents average recurrence interval, T/Tave represents normalized data of ancient earthquake recurrence intervals.

Figure 4 .
Figure 4. Histogram of normalized data of ancient earthquake recurrence intervals in the northeastern Tibetan Plateau T/Tave and the normal distribution fi ing curve.T represents recurrence interval of ancient earthquake, Tave represents average recurrence interval, T/Tave represents normalized data of ancient earthquake recurrence intervals.

Figure 5 .
Figure 5. Diagram of normal Q-Q plot and reverse trend normal Q-Q plot.A normal Q-Q plot comparing randomly generated, independent standard normal data on the vertical axis to a standard normal population on the horizontal axis.The linearity of the points suggests that the data are normally distributed.

Figure 6 .
Figure 6.Diagram of logarithmic normal Q-Q plot and reverse trend logarithmic normal Q-Q plot.A Q-Q plot of a sample of data versus a logarithmic distribution.The deciles of the distributions are shown in the plot.Many outliers are evident at the high end of the range.

Figure 5 .
Figure 5. Diagram of normal Q-Q plot and reverse trend normal Q-Q plot.A normal Q-Q plot comparing randomly generated, independent standard normal data on the vertical axis to a standard normal population on the horizontal axis.The linearity of the points suggests that the data are normally distributed.

Figure 4 .
Figure 4. Histogram of normalized data of ancient earthquake recurrence intervals in the northeastern Tibetan Plateau T/Tave and the normal distribution fi ing curve.T represents recurrence interval of ancient earthquake, Tave represents average recurrence interval, T/Tave represents normalized data of ancient earthquake recurrence intervals.

Figure 5 .
Figure 5. Diagram of normal Q-Q plot and reverse trend normal Q-Q plot.A normal Q-Q plot comparing randomly generated, independent standard normal data on the vertical axis to a standard normal population on the horizontal axis.The linearity of the points suggests that the data are normally distributed.

Figure 6 .
Figure 6.Diagram of logarithmic normal Q-Q plot and reverse trend logarithmic normal Q-Q plot.A Q-Q plot of a sample of data versus a logarithmic distribution.The deciles of the distributions are shown in the plot.Many outliers are evident at the high end of the range.

Figure 6 .
Figure 6.Diagram of logarithmic normal Q-Q plot and reverse trend logarithmic normal Q-Q plot.A Q-Q plot of a sample of data versus a logarithmic distribution.The deciles of the distributions are shown in the plot.Many outliers are evident at the high end of the range.

21 Figure 7 .
Figure 7. Diagram of Weibull distribution Q-Q plot and reverse trend Weibull distribution Q-Q plot.A Q-Q plot of a sample of data versus a Weibull distribution.The deciles of the distributions are shown in the plot.many outliers are evident at the high end of the range.

Figure 7 .
Figure 7. Diagram of Weibull distribution Q-Q plot and reverse trend Weibull distribution Q-Q plot.A Q-Q plot of a sample of data versus a Weibull distribution.The deciles of the distributions are shown in the plot.many outliers are evident at the high end of the range.

Figure 8 .
Figure 8. (A) Distribution map of the Maqin segment and adjacent segments of the Eastern Kunlun fault.(B) The occurrence ages of paleoearthquakes since 12 ka are compiled [25-27].

Figure 8 .
Figure 8. (A) Distribution map of the Maqin segment and adjacent segments of the Eastern Kunlun fault.(B) The occurrence ages of paleoearthquakes since 12 ka are compiled [25-27].

Figure 9 .
Figure 9. (A) Distribution map of the Maqu segment and adjacent segments of the Eastern Kunlun fault.(B) The occurrence ages of paleoearthquakes since 12 ka are compiled [25-27].

Figure 9 .
Figure 9. (A) Distribution map of the Maqu segment and adjacent segments of the Eastern Kunlun fault.(B) The occurrence ages of paleoearthquakes since 12 ka are compiled [25-27].

4. 5 .
Joint Probability of the Occurrence of Strong Earthquake with Magnitude Ms ≥ 7 in the Next 100 Years

Table 1 .
Estimate of the magnitude of Maqin and Maqu Segment for single or joint rupture.
4.2.Estimation of Elapsed Time since the Last Earthquake (t) and Average Recurrence Interval (R)

Table 2 .
Conditional probability of strong earthquake recurrence in the future 20 years, 50 years, and 100 years for Maqin-Maqu segment of the Eastern Kunlun active fault zone (calculated from 2023).

Table 3 .
Joint probability for the occurrence of at least one large earthquake in Maqin-Tazang segment of the Eastern Kunlun fault zone during the future 20 years, 50 years, and 100 years (calculated from 2023).

Table A1 .
Data of ancient earthquakes for 22 active fault zones in the northeastern Tibetan Plateau.