The Benzilan-Tangke deep seismic sounding profile in the western Sichuan region passes through the Songpán-Garze orogenic belt with trend of NNE. Based on the travel times and the related amplitudes of phases in the record sections, the 2-D P-wave crustal structure was ascertained in this paper. The velocity structure has quite strong lateral variation along the profile. The crust is divided into 5 layers, where the first, second and third layer belong to the upper crust, the forth and fifth layer belong to the lower crust. The low velocity anomaly zone generally exists in the central part of the upper crust on the profile, and its integrates into the overlying low velocity basement in the area to the north of Ma’erkang. The crustal structure in the section can be divided into 4 parts: in the south of Garze-Litang fault, between Garze-Litang fault and Xianshuihe fault, between Xianshuihe fault and Longriba fault and in the north of Longriba fault, which are basically coincided with the regional tectonics division. The crustal thickness decreases from southwest to northeast along the profile, that is, from 62 km in the region of the Jinshajiang River to 52 km in the region of the Yellow River. The Moho discontinuity does not obviously change across the Xianshuihe fault based on the PmP phase analysis. The crustal average velocity along the profile is lower, about 6.30 km/s. The Benzilan-Tangke profile reveals that the crust in the study area is orogenic. The Xianshuihe fault belt is located in the central part of the profile, and the velocity is positive anomaly on the upper crust, and negative anomaly on the lower crust and upper mantle. It is considered as a deep tectonic setting in favor of strong earthquake’s accumulation and occurrence.

In this paper, the differences of precursory anomalies between strong earthquakes with M≥7 and moderate earthquakes with M=5~6 were compared and analyzed. The results indicate that there are not any essential differences, but obvious quantitative differences between them. Prediction efficacies for increase of seismicity with M 5~6 before and earthquake with M≥7 and increase of seismicity with M 4~5 before and earthquake with m=5~6 were also statistically examined and were passed, respectively.

Correlation among various seismological prediction indexes related to earthquake frequencyN or energyE for describing seismicity changes is studied by seismic data of 11 seismic zones (areas) in the Sichuan-Yunnan region. A statistical test for correlativity shows that seismicity indexes obtained from transformation of earthquake frequencyN or energyE have correlation to a certain extent while they are used to describe seismicity changes. If the transformation is linear in all cases, the correlation coefficient is equal to 1. The smaller the transformation deviation from linearity, the bigger the correlation coefficient is.

1 问题与思路 "八五"地震短临预报攻关号召大家在"新"和"实"字上下功夫, 尤其是80年代末、 90年代初, 适逢地震界兴起非线性热, 于是, 引进、开发了许多地震学地震预报新指标和新方法, 促进了地震预报研究的发展. 可是, 地震预报的实际水平并未因此发生质的飞跃. 因此, 在"九五"立项时, 不少专家提出一个重要问题: 这些新指标与老指标是否相互独立?能否筛选出相互独立, 又有较好预报效能的指标和方法?

The real-time processing system of CTSN performs following: A/D conversion; automatic event detection; event data saving; automatic measure of P and S arrivals; event location and print out the calculated results. It is corrected at ny moment by using the off-line system. Since December 1993, this system has been operating normally in the CTSN. More than 4 000 earthquakes have been recorded in the system. It has high accuracy in automatic picking P and S arrivals. The location of the earthquakes determined by on-line system are close to those given in published catalogues which are determined by manual procedure. This system can finish locate event in three minutes. It also gives satisfactory epicenter locations for distant events by inputting manually S arrivals in the off-line system. The operation of this system had brought the technical superiority of the CTSN. It not only reduces the labor intensity and simplifies the working procedure, but also makes our research facility into the superior ranks in this field of our country.

Based on the single scattering model of coda power spectrum analysis, digital waveform data of 50 events recorded by the real-time processing system of the Chengdu telemetry network are analyzed to estimate the Q c values of earth medium beneath the Chengdu telemetry network for several specified frequencies. It is found that the Q c shows the frequency dependency in the form of Q c = Q 0 f n in the range of 1.0 to 20.0Hz. Estimated Q 0 ranges from 60.83 to 178.05, and n is found to be 0.713 to 1.159. The average value of Q 0 and n are 117 and 0.978 respectively. This result indicates the strong frequency dependency of the attenuation of coda waves beneath the Chengdu telemetry network. Comparing with the results obtained in other regions of the world, it is found that Q 0 −1 value and its change with frequency are similar to those in regions with strong tectonic activity.

Temporal curves of algorithmic complexity C(n) are calculated by using a time window of one year, a sliding step of half a year and three magnitude levels, i.e., 2.5 �� M < 3.5, 3.5 �� M < 4.5 and M �� 4.5 on the basis of data from 11 seismic zones in the Sichuan-Yunnan Area. The possibility to predict whether an M �� 5 earthquake will occur in the next half a year in this seismic zone is discussed.

In this paper, temporal curves of capacity dimension D0 and information dimension D1 are calculated with time window of one year and sliding step of half-a-year by using data of 11 seismic zones in the Sichuan-Yunnan area. The possibility of using the change rates of D0 and D1 to predict whether an M≥5 earthquake will occur in the next half-a-year in the seismic zone is discussed. Generally speaking, it is possible to obtain grading values R which are different from those of random prediction. It is demonstrated that the use of the change rates of capacity dimension and information dimension for earthquake prediction has certain prospects.

In this paper, the methods of pattern recognition and fuzzy cluster are applied to estimate comprehensive whether there would be an earthquake which magnitude is commensurate with (N=M 0±0.3) or greatM>M 0) thanM 0 within 3 months using sequential data in 3 days immediately after a moderate or strong earthquake (M 0≥4.7) occurred, The result of retrospective test indicates that the comprehensive judgment effects by using these applied mathematics methods are better than that of using every individual method obviously. TheV mark value of algorithm CORA-3 in pattern recognition is the highest among them. There are 44 known earthquake sequence data (5 type- I and 39 type- II) used to learn and train, the results of internal coincidence test show that all 44 sequence samples could be distinguished correctly. Extrapolation test by using other 4 known earthquake sequences (2 type- I and 2 type- II) shows that all 4 extrapolation samples could be distinguished correctly also. In the process of study, these methods have been applied to judge the post-earthquake tendency of 2 moderate earthquakes occurred recently, one is distinguished correctly and the other wrong. The algorithm Hamming in pattern recognition and fuzzy cluster method have been applied to judge the early post-earthquake tendency after a moderate or strong earthquake, too. TheV mark values of internal coincidence tests could get above 0.8 most cases of extrapolating are correct.

The precursory swarm, three mainshocks (M = 7.2,6.7, 7.2), and aftershocks of the Songpan earthquakes have been reanalyzed using both local and teleseismic data. The three mainshocks of this sequence occurred on the Huya fault over a 7‐day period. Relocations of the aftershocks using local arrival times show that three fault strands were activated during this sequence. Each mainshock occurred on a separate strand, each one south of the strand activated in the previous mainshock, and the aftershock zones of each mainshock appear to abut rather than overlap. Fault plane solutions determined by matching teleseismic P waveforms at World‐Wide Standard Seismograph Network stations with synthetic seismograms are consistent with the observed aftershock zones. The first and third mainshocks (M0 = 1.3 ×1 019 and 8.4 × 1018 N m, respectively) showed almost identical senses of motion, a combination of reverse and left‐lateral strike‐slip motion, on parallel strands, striking N15°W, that were separated by a large rightstepping en echelon offset. The second mainshock (M0 = 4.0 × 1018 N m), occurred in this offset on a fault at a steep angle (∼125°) to the other two strands and showed almost pure reverse motion. Differences in the orientations of the slip vectors of the three mainshocks show that the first mainshock increased the normal and shear stresses on the fault segment that moved in the second mainshock and that the second mainshock decreased the normal stress on the fault segment activated by the third mainshock. These changes in normal stresses may have given rise to the longer time between the first and second events (5 days) as compared with the time between the second and third events (30 hours). A precursory swarm that preceded the Songpan sequence by 3 years occurred in a volume that surrounded the northernmost part of the planar aftershock zone. The time between the start of the swarm and the mainshocks and the magnitude of the largest event in the swarm are similar to those seen for precursory swarms in Soviet Central Asia.