Research on completeness of earthquake data in the Chinese mainland（Ⅱ）──The regional distribution of the beginning years of basically complete earthquake data

Based on the concrete conditions of earthquake data in the west of China, East China and SOuth China, we studied the completeness of data in these regions by suitable methods to local conditions. Otherwise, we roughly estimated monitoring capability of local networks in China since 1970 and some outlying regions where the data is lack. Finally, we gave the regional distribution of the beginning years since which the data for different magnitude intervals are largely complete in the Chinese mainland.

Research on the completeness of earthquake data in the Chinese mainland（I）──North China

In terms of the temporal-spatial distribution features of earthquakes, we study the completeness of historical data in North China where there is the most plenty historical data and with the longest record history by some meth ods of analysis and comparison. The results are obtained for events with Ms≥4 are largely complete since 1484 in North China (except Huanghai sea region and remote districts, such as Nei Mongol Autonomous region), but quakes with Ms≥6 are largely complete since 1291 in the middle and lower reaches of the Yellow River.

Anomaly detection of earthquake precursor data using long short-term memory networks

Earthquake precursor data have been used as an important basis for earthquake prediction.In this study,a recurrent neural network(RNN)architecture with long short-term memory(LSTM)units is utilized to develop a predictive model for normal data.Furthermore,the prediction errors from the predictive models are used to indicate normal or abnormal behavior.An additional advantage of using the LSTM networks is that the earthquake precursor data can be directly fed into the network without any elaborate preprocessing as required by other approaches.Furthermore,no prior information on abnormal data is needed by these networks as they are trained only using normal data.Experiments using three groups of real data were conducted to compare the anomaly detection results of the proposed method with those of manual recognition.The comparison results indicated that the proposed LSTM network achieves promising results and is viable for detecting anomalies in earthquake precursor data.

Co-seismic fault geometry and slip distribution of the 26 December 2004, giant Sumatra–Andaman earthquake constrained by GPS, coral reef, and remote sensing data

We analyze co-seismic displacement field of the 26 December 2004, giant Sumatra–Andaman earthquake derived from Global Position System observations,geological vertical measurement of coral head, and pivot line observed through remote sensing. Using the co-seismic displacement field and AK135 spherical layered Earth model, we invert co-seismic slip distribution along the seismic fault. We also search the best fault geometry model to fit the observed data. Assuming that the dip angle linearly increases in downward direction, the postfit residual variation of the inversed geometry model with dip angles linearly changing along fault strike are plotted. The geometry model with local minimum misfits is the one with dip angle linearly increasing along strike from 4.3oin top southernmost patch to 4.5oin top northernmost path and dip angle linearly increased. By using the fault shape and geodetic co-seismic data, we estimate the slip distribution on the curved fault. Our result shows that the earthquake ruptured ＊200-km width down to a depth of about 60 km.0.5–12.5 m of thrust slip is resolved with the largest slip centered around the central section of the rupture zone78N–108N in latitude. The estimated seismic moment is8.2 9 1022 N m, which is larger than estimation from the centroid moment magnitude（4.0 9 1022 N m）, and smaller than estimation from normal-mode oscillation data modeling（1.0 9 1023 N m）.

On Application of Big Data Mining in Earthquake Precursor Observation

Research and application of big data mining,at present,is a hot issue. This paper briefly introduces the basic ideas of big data research, analyses the necessity of big data application in earthquake precursor observation,and probes certain issues and solutions when applying this technology to work in the seismic-related domain. By doing so,we hope it can promote the innovative use of big data in earthquake precursor observation data analysis.

Studies of Earthquake Hazard Using Microseismicity Data in Modern Times

This paper selects some representative regions to obtain their G-R relation curves according to their seismicity characteristics,by using ML≥2.0 microseismicity data(1970~1993)in North China.The annual occurrence rate of events of each magnitude can be inferred from the G-R relation.At the same tune,the actual annual occurrence rate of earthquakes of higher magnitudes can be calculated from historical earthquakes(1300-1993)recorded in the same region.It seems that both results are almost the same.Therefore,the rate of events of higher magnitudes can be obtained by using microseismicity data when the proper region is selected.However,two points should be noticed:(1)The method can only give the annual occurrence rate in a seismicity system and estimate the whole situation of the system.(2)When there is a very large earthquake in and near the period in which the microseismicity data are applied,the actual occurrence rate of the system,including this larger earthquake,cannot be obtained by this method.

Anomalous signals before 2011 Tohoku-oki Mw9.1 earthquake,detected by superconducting gravimeters and broadband seismometers

The 2011 Tohoku-oki earthquake,occurred on 11 March,2011,is a great earthquake with a seismic magnitude Mw9. 1,before which an Mw7. 5 earthquake occurred. Focusing on this great earthquake event,we applied Hilbert-Huang transform（ HHT） analysis method to the one-second interval records at seven superconducting gravimeter（ SG） stations and seven broadband seismic（ BS） stations to carry out spectrum analysis and compute the energy-frequency-time distribution. Tidal effects are removed from SG data by T-soft software before the data series are transformed by HHT method. Based on HHT spectra and the marginal spectra from the records at selected seven SG stations and seven BS stations we found anomalous signals in terms of energy. The dominant frequencies of the anomalous signals are respectively about 0. 13 Hz in SG records and 0. 2 Hz in seismic data,and the anomalous signals occurred one week or two to three days prior to the event. Taking into account that in this period no typhoon event occurred,we may conclude that these anomalous signals might be related to the great earthquake event.

The principle of coupled stress release model and its application

oupled stress release model is proposed in the paper considering the interaction between different parts based on stress release model by VereJones, and is used to historical earthquake data from North China. The results by this model are compared with the results by original stress release model using AIC criterion. The results show that coupled stress release model is better than original model.

Frequency band-dependence of S-wave splitting in China mainland and its implications

Butterworth band-pass filter has been applied to S-wave data recorded at 8 stations in China mainland, and S-wave splitting at different frequency bands is analyzed. The results show that the delay time and the fast polarization directions of S-wave splitting depend upon the frequency bands. There is an absence of S-wave splitting at the station of Urumqi (WMQ) for the band of 0.1-0.2 Hz. With the frequency band broadening, the delay time of S-wave splitting decreases at the stations of Beijing (BJI), Enshi (ENH), Kunming (KMI) and Mudanjiang (MDJ); the fast polarization direction changes from westward to eastward at Enshi (ENH), and from eastward to westward at Hailaer (HIA). The variations of delay time with bands at Lanzhou (LZH) and Qiongzhong (QIZ) are similar, and there is a coherent trend of fast polarization directions at BJI, KMI and MDJ, respectively. Initial interpretations to the results of frequency band-dependence of S-wave splitting are also presented.