Abstract Constructing reliable earthquake catalogs in regions with sparse station coverage and strong noise remains challenging. We evaluate the performance of a deep‐learning–based detection and location workflow (LOC‐FLOW) using the 2017 Changdao offshore earthquake swarm in eastern China. Three catalogs are compared: an automatic deep‐learning catalog, a matched‐filter catalog, and a routine network catalog. For moderate‐magnitude events, the deep‐learning and matched‐filter catalogs show consistent temporal evolution and frequency–magnitude statistics, and both delineate a northwest–southeast–trending fault zone at 8–14 km depth. Differences are confined to small magnitudes. The deep‐learning catalog has a completeness magnitude of ∼1.1, similar to the routine catalog, whereas the matched‐filter catalog extends completeness to ∼0.8 and detects substantially more microearthquakes. Event correspondence analysis indicates that most deep‐learning detections are shared with the matched‐filter catalog, with missed events primarily occurring under low signal‐to‐noise and poor azimuthal coverage. These results demonstrate that deep‐learning workflows provide a stable backbone catalog under sparse‐network conditions, while matched filtering offers complementary sensitivity to weak events. Their joint application provides an effective strategy for constructing high‐resolution earthquake catalogs in coastal transition zones and other sparsely instrumented regions.

In this article, we collected the seismic phase arrival data (115 degrees-120 degrees E, 34 degrees-39 degrees N) of 2,833 local natural earthquakes above magnitude 1.0 recorded by 128 seismograph stations provided by the China Earthquake Networks Center covering the period from January 2008 to October 2023. We extracted the first arrival P- and S-wave arrival time data and obtained 26,351 P- and 26,349 S-wave absolute arrival times and 99,627 P-differential and 99,625 S-differential arrival times. Then, we determined 3-D P and S wave velocity structures under the Luxi region by using the double-difference tomography method. The results show lateral heterogeneities under the Luxi region, and the distribution of surface basins and uplift zones is associated with seismic velocities within the crust. The vertical section shows the widespread presence of vertical high-velocity bodies within the crust of the Luxi region, indicating seismological evidence of mantle upwelling in the region. Meanwhile, our imaging results also clearly indicate the presence of a significant low-velocity anomaly at depths of 13-17 km, which corresponds to the presence of a high-conductivity layer at that depth obtained from electromagnetic bathymetry. The transition zone between high and low velocities exists at a depth of 7-13 km, which plays an important role in the transfer of lithospheric stresses from the deep to the shallow part of the lithosphere. Combined with the results of the previous deep seismic wave velocity structure, petrology, and geological investigations, the deep dynamical background of the formation of slip tectonics and seismic mechanism in the Luxi area can be attributed to the go-slip activity of the Tan-Lu fault Zone, the subduction of the Pacific Plate, and the magmatism generated by mantle upwelling.

The 2019 Ridgecrest MW7.1 earthquake has received significant attention due to its complex fault activity. It is also noticeable for its MW6.4 foreshock sequence. There are intricate dynamic relationships between earthquakes in such vigorous sequences. Based on the relocated catalogue, we adopt the nearest neighbour algorithm to analyze its foreshock and aftershock sequences. Detailed links and family structures of the sequence are obtained. The results show that a MW5.0 event at 03:16 (UTC) on 6 July is a direct foreshock of the MW7.1 mainshock. It is likely related to barriers on the northwest-striking fault. The MW6.4 event on 4 July is characterized as a complex conjugate rupture. Notably, a magnitude 4.0 event occurred on the northwest-striking fault before the MW6.4 event, establishing it as a direct foreshock. The Ridgecrest sequence is predominantly influenced by northwest fault activity. It first caused small fractures on the northwest-striking fault. Then, it triggered conjugate slips on the southwest-striking fault. Lastly, it led to larger ruptures on the northwest-striking fault.

Seismic activities can be seen as the composition of background and clustering earthquakes. It is important to identify seismicity clusters from background events. Based on the Nearest Neighbour Distance algorithm proposed by Zaliapin, we use the Gaussian mixture model (GMM) to fit its spatiotemporal distribution and use the probability corresponding to clustering seismicity in the GMM model as the clustering ratio. After testing with synthetic catalogues under the ETAS (epidemic-type aftershock sequence) model, We believe the method can discriminate cluster events from randomly occurring background seismicity in a more physical background. We investigate the seismicity and its clustering features before the M6.6 Jinggu earthquake in Yunnan Province, China on 7 October 2014. Our results show the following: 1) The seismogenic process of this strong earthquake has three stages, which are already described by the IPE model (the model is similiar to dilatancy diffusion model, growth of cracks is also involved but diffusion of water in and out of the focal region is not required); 2) The main shock might have been caused by the breaking of a local locked barrier in the hypocentre, and the meta-instability stage was sustained for about 1 year on the fault. From this study, we conclude that the evolution of seismicity clustering features can reflect changes in stress in the crust, and it is closely connected to the seismogenic process of a strong earthquake.

On 21 May 2021, the MS 6.4 earthquake struck Yangbi County, Dali City, Yunnan Province. Minor seismic activities were common both before and after the earthquake; the foreshock sequence activity characteristics of this earthquake were studied to gain a better understanding and more perspective. First, between May 18 and 5 July 2021, we collected data of the seismic events from the Yangbi MS 6.4 earthquake sequence, determined the minimum magnitude of completeness of the sequence based on the magnitude-frequency relationship, filtered out the more complete earthquake sequences, and performed double-difference earthquake relocation using the HypoDD method. Then, we improved the nearest-neighbor distance algorithm and used it in conjunction with the Gaussian mixed model fitting method to conduct a comprehensive multiple factor analysis of the Yangbi MS 6.4 earthquake sequence. The main findings of our preliminary analysis are as follows: 1) We distinguished the foreshocks, mainshocks, and aftershocks of the Yangbi earthquake sequence as well as their higher-order aftershocks, using the method described in this paper, and we can obtain clear intergenerational relationships between them. We identified eight “foreshocks” with statistically greater physical significance than the others and found that majority of the shocks occurred shortly after the mainshock were direct aftershocks, with secondary or higher-order aftershocks gradually increasing in the later stage of the sequence. 2) Combining the double-difference earthquake relocation and the event distribution on the fault plane, we found that the b-value of the Yangbi foreshock sequence was clearly on the lower end of the spectrum, with an obvious nonlinear amplification process, and can be divided into three foreshock sub-sequences: sub-sequence I with the lowest b-value and a concentrated spatial distribution, which is an F-value foreshock sequence; subsequence II with an overall shallower source depth and an obvious rebound in b-value, which is an explosive aftershock sequence; subsequence III with a lower b-value and a rapid rupture spread, which is a typical U-F-ρ foreshock. 3) We constructed a topological tree of the Yangbi foreshock sequence in Yunnan Province using the nearest-neighbor distance algorithm, combining the correlation between earthquake sequence type and fault rupture intensity. We analyzed the distribution and topology of the three sub-sequences (combined) of the Yangbi MS 6.4 foreshock sequence and found the following: the topology of subsequence I was more linear than the sequence II, each was the parent event of the subsequent event, and its rupture mode was similar to the fluid intrusion rupture in a specific channel; sequence II had a relatively simple topology, exhibiting a spray topology, and independently formed a main-aftershock mini-sequence, probably triggered by its shallow brittle rupture, which caused the fluid intrusion rupture to spread faster and eventually triggered the Yangbi MS6.4 mainshock.

In this paper, we collected the seismic phase arrival data of 14,033 local natural earthquakes above magnitude 3.0 recorded by 435 seismograph stations in the study area provided by the International Seismological Centre (ISC), covering the period from January 2011 to April 2020. We selected the first arrival P- and S-wave arrival time data and obtained 281,859 P- and 112,926 S-wave absolute arrival times and 528,250 P-differential and 207,968 S-differential arrival times. Then we determined 3-D P and S wave velocity structures from the Japan Trench to the back arc area under the Japan Islands by using double-difference tomography method. The results show strong lateral heterogeneities under the forearc region. The subducting Pacific slab is imaged clearly as a high-velocity (high-V) faster than the surrounding mantle. Low-velocity (low-V) zones are imaged in the mantle wedge with significant along arc variations under the volcanic front. The vertical section shows that the low velocity zone extends to a depth of approximately 70–150 km below the arc. The crust and mantle wedges beneath the front and back arcs of the volcanoes reveal a low-V anomaly, the likely main source of which is the partial melting of plate and mantle wedge material. Referring to the previous results of plate dehydration, mineral composition and thermal state, it is found that fluids play a crucial role in the arc magmatism and plate melting of mantle wedge behind Japan Trench. The fluids brought down by the Pacific subducting plate are released into the mantle wedge by dehydration and subsequently transported upward by the upwelling flow in the mantle wedge. The present results obtained using new and advanced imaging methods enrich the understanding of the velocity structure beneath Japan Islands, which may improve the understanding of the dynamic processes of subduction zones and mantle upwelling.

On 21 May 2021, the MS 6.4 earthquake struck Yangbi County, Dali City, Yunnan Province. Minor seismic activities were common both before and after the earthquake; the foreshock sequence activity characteristics of this earthquake were studied to gain a better understanding and more perspective. First, between May 18 and 5 July 2021, we collected data of the seismic events from the Yangbi MS 6.4 earthquake sequence, determined the minimum magnitude of completeness of the sequence based on the magnitude-frequency relationship, filtered out the more complete earthquake sequences, and performed double-difference earthquake relocation using the HypoDD method. Then, we improved the nearest-neighbor distance algorithm and used it in conjunction with the Gaussian mixed model fitting method to conduct a comprehensive multiple factor analysis of the Yangbi MS 6.4 earthquake sequence. The main findings of our preliminary analysis are as follows: 1) We distinguished the foreshocks, mainshocks, and aftershocks of the Yangbi earthquake sequence as well as their higher-order aftershocks, using the method described in this paper, and we can obtain clear intergenerational relationships between them. We identified eight “foreshocks” with statistically greater physical significance than the others and found that majority of the shocks occurred shortly after the mainshock were direct aftershocks, with secondary or higher-order aftershocks gradually increasing in the later stage of the sequence. 2) Combining the double-difference earthquake relocation and the event distribution on the fault plane, we found that the b-value of the Yangbi foreshock sequence was clearly on the lower end of the spectrum, with an obvious nonlinear amplification process, and can be divided into three foreshock sub-sequences: sub-sequence I with the lowest b-value and a concentrated spatial distribution, which is an F-value foreshock sequence; subsequence II with an overall shallower source depth and an obvious rebound in b-value, which is an explosive aftershock sequence; subsequence III with a lower b-value and a rapid rupture spread, which is a typical U-F-ρ foreshock. 3) We constructed a topological tree of the Yangbi foreshock sequence in Yunnan Province using the nearest-neighbor distance algorithm, combining the correlation between earthquake sequence type and fault rupture intensity. We analyzed the distribution and topology of the three sub-sequences (combined) of the Yangbi MS 6.4 foreshock sequence and found the following: the topology of subsequence I was more linear than the sequence II, each was the parent event of the subsequent event, and its rupture mode was similar to the fluid intrusion rupture in a specific channel; sequence II had a relatively simple topology, exhibiting a spray topology, and independently formed a main-aftershock mini-sequence, probably triggered by its shallow brittle rupture, which caused the fluid intrusion rupture to spread faster and eventually triggered the Yangbi MS6.4 mainshock.

大量研究认为,日本地区强震活动与华北地区的中强地震活动之间存在一定相关关系.通过梳理太平洋板块西北缘俯冲带的大震活动,在搜集历史资料的基础上,结合构造背景,重新对日本与华北地区强震活动的相关性进行了分析,最终认为:日本地区8级以上大震与华北地区6级强震之间不存在一一对应的直接相关关系;日本和华北地区大的地震活动期和稳定期具有某种对应性,华北地区活动期的强释放阶段与东西两个构造边界活动存在联系;2011年日本"311"9.0级地震的余震活动基本结束,日本地区可能已经恢复背景的活动节律;当前华北地区可能仍处于活动期末尾的调整期,未来几年,华北地区仍存在进入剩余释放阶段最后一个活跃幕的可能.

Earthquakes can occur in one or several periods, and by analyzing the period spectrum and its synthetic probability, the trend of seismic activity in a certain area in the future period can be studied. In this paper, namely, the seismic period spectra and their linear synthetic probability values were calculated in three regions using earthquake data of magnitude https://www.w3.org/1998/Math/MathML"> M ≥ 4 3 3 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003308584/bebb6986-3d8f-445c-9ce0-332068e7a6e9/content/inline-math111_1.tif" xmlns:xlink="https://www.w3.org/1999/xlink"/> (all magnitudes are M in the text without special notes) from 1480-2000 and M≥4 since 1970 in Shandong and nearby areas (34–38.5°N, 114–123°E, no further hints below), and a certain time period in each region according to the characteristics of different subregions, The actual seismic prediction tests were conducted for each partition according to the characteristics of each partition. The results show that the method has good results in retrospective verification of moderate to strong earthquakes in Shandong and neighboring areas, except for the earthquake period spectrum of magnitude https://www.w3.org/1998/Math/MathML"> M ≥ 4 3 3 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003308584/bebb6986-3d8f-445c-9ce0-332068e7a6e9/content/inline-math111_2.tif" xmlns:xlink="https://www.w3.org/1999/xlink"/> since 1480 in the Yishu zone, which shows no dominant period, and makes a reasonable analysis of seismic activity trends in Shandong based on synthetic probability out-test predictions, there is a possibility of a magnitude 4 earthquake in Shandong and nearby areas in the next year.

2020年2月18日山东济南长清发生M4.1(Ms 4.1)地震,基于中国地震台网中心的观测报告,使用双差定位法对济南长清地震序列进行重定位,利用重定位后的地震参数拟合了断层面的走向及倾角.利用P波初动方法求取了 5个震源机制解,并反演了震源区的局部应力场,根据局部应力场和断层面参数计算了滑动角.得到以下初步结论:长清地震序列从SE向NW延伸、倾向NE;断层面拟合得到断层面走向NW(320.5°),倾角接近垂直(84.8°),滑动角为15.7°,本文推断发震断层为走向NW、倾角近垂直且具有左旋走滑性质的隐伏次级断裂.主震震源机制解是正走滑型地震,其节面Ⅰ走向111°、倾角65°、滑动角-45°,5个震源机制解按时间顺序呈正断走滑型与走滑型地震相互交替出现.震源区局部应力场呈正断兼少量走滑的应力机制,与山东背景应力场不同.本文认为在太平洋板块自东向西推挤作用背景下,鲁中隆起与其西北凹陷区垂向运动差异,是济南长清发生地震的根本原因.

以最邻近事件距离算法为基础,结合混合概率密度分布对胶东半岛地区的小震丛集特征进行分析.结果表明:1)该地区存在明显的地震成丛现象,且近年来丛集率显著提高;2)乳山震群和莱州地震前,半岛地区均存在b值降低和丛集率升高的现象.分析认为,低b值异常配合高丛集率可能是该区域地震活动较好的预测指标.

An M(s)6. 4 earthquake occurred on May 21st, 2021 in Yunnan Yangbi country. In order to analyze the seismogenic fault and structure deeply, and discuss the difference of seismogenic structure and background between the 2013 Eryuan, 2017 and 2021 Yangbi earthquakes. The earthquake sequence is relocated by double-difference method with the observation report from the China Earthquake Networks Center (CENC). And the tectonic stress field is calculated with nine focal mechanism solutions from the Global Centroid Moment Tensor (GCMT) catalogue and the United States Geological Survey (USGS) in seismic source region. The initial conclusions are as follow: (1) The Yangbi earthquake sequence is distributed in the direction from NW to SE. More aftershocks are located in the part of SE segment than that in NW segment of earthquake cluster. Few aftershocks occurred around the earthquakes with magnitude larger than 5.0 because of heterogeneity release of crustal stress. Initial rupture depth of four events with magnitude larger than 5.0 are deeper than their centroid depth, their rupture process of seismogenic faults are beginning from fault bottom to fault upper. (2) Two unknown faults F2 and F3 are the seismogenic faults of the Yangbi earthquake sequence, which are lying down southwestern of Weixi-Qiaohou-Weishan fault. Both of them are similar with nearly vertical dip, tending to SW direction, trending from NW to SE direction and sliding in right-lateral strike-slip. The length of F2 is about 30 km, another one (F3) is shorter with the length of nearly 11 km and intersecting with F2 fault in the middle of earthquake cluster. (3) The tectonic stress field is strike-slip regime in seismic source region. The principal compressive stress axis is tended in SSE direction (174.57 degrees) with lower plunge (18.79 degrees). And the principal extensional stress axis is tended in SWW direction (-93.65 degrees) with nearly horizontal plunge (5.21 degrees). The seismogenic structure in seismic source region is controlled by right-lateral strike-slip boundary between Sichuan-Yunnan Block and Southern Yunnan Block. (4) All the three earthquakes occurred under the background of strike-slip regime in southwestern boundary of Sichuan-Yunnan Block with right-lateral strike-slip movement. The seismogenic structure of the 2013 Eryuan earthquake is controlled mostly by local structure with vertical movement. The 2017 Yangbi earthquake is controlled only by Sichuan-Yunnan Block boundary with right-lateral strike-slip movement. The 2021 Yangbi earthquake sequence is controlled mostly by right-lateral strike-slip movement on the boundary of Sichuan-Yunnan Block and little vertical movement from local structure crust.

1 研究方法 地震活动性是表征某地区地震活动强弱的重要指标,具有地域性和时域性的特点.为将地震活动定量化,引入一个新的模糊量S,该参数是对地震活动性时、空、强诸因素的综合表征,将各隶属函数归一化并以加权组合的方式给出,基本公式为S=1/ n∑n i=1W(Xi)μ～A(X i) (1)式中,Xi是隶属函数的参数,μA～(为隶属函数,W为权因子.

通过限定震群集中区范围,剔除同一台站记录的震中位置相近且具有相同路径的地震波,使用双差波速比两次差分的方法,对2017～2018年长岛地区发生的2个震群的波速比值开展特征分析.通过对震相数据敏感性分析得出,双差波速比方法对Pg波到时较为敏感,资料处理时需要将Pg、Sg波震相误差分别控制在0.02 s和0.20 s以内.最终分析结果显示,长岛震群波速比值在1.69～1.78范围内波动,波速比变化与震群活动过程密切相关;震群几次较大余震发生前,波速比呈现低值状态;主震和余震的波速比变化存在差异,可能反映二者发震机理有所不同.

采用高频截止(High-Cut)震源模型,以均方根误差最小原则稳健地求解震源谱参数,并由此推算震源尺度和静态应力降.实际应用显示,该模型的理论谱对观测谱有很好的拟合,可明显改善拐角频率识别准确度.计算了长岛震群内71次ML≥2.5地震事件的震源参数,结果表明:①拐角频率处于2～10 Hz范围,与震级大小存在一定的相关性,截止频率范围处于10～30 Hz之间,与地震大小的相关性不明显;②地震矩M0分布在1012～1014 N·m,与震级ML存在正相关关系:logM0=0.977ML+10.186;ML与矩震级Mw之间的关系为:Mw=0.651ML+0.766;③根据相对应力降时域演化发现,自2017年3月3日ML4.5地震之后应力快速释放,应力降水平在均值附近波动,而且多数ML≥3.5地震发生于应力降下降之后的回升过程中;④应力降范围在0.01～1 MPa之间,应力降与震级的统计关系为:logσ=1.158 ML—6.591.长岛震群应力降水平明显偏低,反映了震中区域构造应力水平较低,但是中小地震活跃,推测主要有两方面的原因:一方面,1548年发生的渤海7级地震引起区域性地震矩释放,导致长岛震源区内介质相对破碎;另一方面,长岛海域断层易受到海水渗入,引起介质孔隙压力增大,同时内摩擦系数降低,从而导致断层内剪切应力降低.综合而言,长岛地区虽然处于较低的构造应力背景之中,但是流体入侵造成的断层内剪应力降低可能是造成长岛地区发生大量中小地震的主要原因.

1 研究背景 波速比是特征化描述岩石成分和流体饱和度的重要参数,可用来反映地下介质性质改变以及间接反映断层活动情况.波速比研究是地震学领域重要的研究方向,常用多台或者多震方法计算平均波速比(黎明晓等,2004;张小涛等,2012;王林瑛等,2014).但在震群活动过程中,震中较为集中,传统波速比计算方法不能准确反映震源区地下介质情况.Lin等(2007)提出,基于P波、S波的双差思想,将距离同一台站相近的2个地震组成一组,通过扣除这2个地震射线的相同路径,获得震群震源区波速比.该方法首次讨论和检验了应用2次差分来计算震源区波速比的可行性,随后诸多学者做了相关研究( Dahmet al, 2014;Bachura et al,2016;贾漯昭等,2017;郑建常等,2018).双差波速比方法是,基于误差分布和概率密度,借鉴双差定位程序中2次差分技术,使用台站S与P震相到时差和台站对的2次差分,扣除地震射线相同路径, 实现震群活动波速比值求解.

大地震造成的地面运动可能在远场触发微震.通过对比检视胶东半岛地区应变仪和强震仪记录,估算2021年9月16日四川泸县MS 6.0地震在该区造成的动态应力变化,从而认为其后在乳山海域发生的ML2.9地震可能是受到此次泸县MS 6.0地震触发的一次小震活动.

基于时间相依的地震复发间隔混合概率模型,开展山东地区中、短期尺度上的中小地震的概率预测实践,1年的检验结果显示,3、4级中小地震基本发生在此前给出的地震危险性高概率区.研究认为,该方法在日常地震会商中应用效果较好,并有望为破坏性地震的概率预测提供参考.

本文利用基于波形互相关的双差定位方法对2020年2月18日长清Ms4.1地震序列进行了精定位计算,共得到33个地震事件的精定位结果.结果显示,地震序列主要沿NW向分布,在水平方向上具有自NW向SE迁移,在深度上具有由浅向深迁移的特征;序列震源深度主要集中在2～7 km,其中,主震的震源深度约2.8 km.由于长清地震序列的地震数量较少,为了更准确地了解长清地震序列的发震构造、探索该序列的发生和发展过程,本文采用CAP方法反演了主震的震源机制解,其中,节面Ⅰ走向223°、倾角42°、滑动角-160°,节面Ⅱ走向117.9°、倾角76.8°、滑动角-49.8°,最佳拟合震源矩心深度约2.8 km,矩震级Mw4.2.结合区域构造特征分析认为,长清Ms4.1地震的发震断裂为孝里铺断裂和东阿断裂之间发育的一条浅层次生断裂.在ENE向区域应力场作用下,发震断裂产生高角度正断滑动,并伴有左旋走滑分量,从而引发长清地震序列.