Search

The normal modes (i.e. Earth's free oscillations) are long-period low-frequency seismic signals, which are excited by a variety of factors, such as earthquakes, volcanic eruption, landslide, avalanche and so on, are an essential vehicle for global seismic tomography to elucidate large-scale heterogeneities within the deep Earth. Accurate extraction of signals on normal mode spectrum is a prerequisite for the imaging inversion, providing the differences between the observed and synthetic normal mode spectrum. However, the normal mode spectrum has great complexity due to many structural factors within the Earth, so unacceptable false and dismissed selections of the signals always occur, which hinder the development of exploration of the deep Earth’s deep interior based on normal mode data. To address these problems, we build a deep-learning based neural network, named ModeNet, which is capable of precisely and efficient selecting the frequency windows to cover the target normal modal signals on a noisy spectrum, which could outperform the conventional spectrum-FLEXWIN method without relying on comparisons with synthetics. We also define our own method to evaluate the performance of ModeNet on the testing set and obtain a precision as high as ~0.98. Moreover, ModeNet achieves good generalization in processing seismograms of different events with different noise levels, components, and time window data, as well as superconductivity-gravimeter observations. Therefore, ModeNet could be implemented as a valuable tool for the future deep Earth inversion., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

A typicalforeshock-mainshock-aftershock sequenceoccurred atRidgecrest, California, USA.TheMw 7.1mainshockhappenedon6thJuly,2019. The correlation betweenthis typicalsequence and tidesis investigated withstatisticalmethods.The resultsformSchuster’s test show that high correlation between tides and earthquakesis foundfor the foreshocksanddisappeared for the aftershocks.Numerical results show thatthe b value decreased by about 25%comparedwith the background valuefor the foreshocks,anditslowlygoesback to the backgroundlevel in several monthsafter themainshock.The decrease of the b valueindicatesthat thestress of thefaultbecomes much higher than thebackgroundstress.Therefore,the b-value can beregarded asanindicatorfor assessing the potential earthquake risk.As a conclusion, it is found thattides play apredominantrole in the triggering or modulation ofthe foreshocksbut not the aftershocks with thistypical earthquake sequence., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

In recent decades, the magnitude and harm of injection-induced earthquakes (IIE) due to anthropogenic activities have gradually increased, which has evolved into a global problem. However, the triggering mechanism of induced earthquakes is complicated, and there are significant spatial differences and time delays between the pumping areas and the areas where the induced earthquakes are concentrated, not all injections induce earthquakes, and it is not uncommon for regions with strong injections to remain seismically silent. To explore the controls on IIE, we compare geodetically derived tectonic strain rates with seismicity in Oklahoma where injections are prevalent. Results demonstrate that induced earthquakes are high in areas with moderate tectonic strain rates, while areas with high seismic moment release are consistent with areas with moderate tectonic strain rates, indicating that the regional tectonic is the fundamental cause of IIE. Moreover, the moment rates released by induced earthquakes approach even exceed the tectonic moment rate in the IIE-rich zone. We thus argue that injections indeed causes a large number of induced earthquakes and local seismic hazard in the short term, but limits the potential for larger future earthquakes by depleting the shallow tectonic moment., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

Based on the status of the distribution and measurements in the national gravity network, a piece-wise linear dynamic adjustment model is introduced and applied to the analysis of the relative gravity observations in order to study the temporal gravity variations in mainland China. Compared with the traditional static adjustment model, more reliable temporal gravity variation characteristics can be given by the new model. To verify the validity of the model, it is processed that the gravity data from the national network and simulated data using the two methods, respectively. For the national gravity network, the mean difference of the rates of gravity changes obtained by the two adjustment methods is 13.4×10-8 m·s-2/a, with a maximum of 50×10-8 m·s-2/a. The precision of the dynamic adjustment is obviously better than the traditional static adjustment. For the simulated data, the rates of gravity change are compared with the theoretical ones at the same points. It is found that over 80% of the differences are less than 1×10-8 m·s-2/a from dynamic adjustment, and only two differences are larger than 2×10-8 m·s-2/a. In contrary, there are only 44.4% of the differences are less than 1×10-8 m·s-2/a from static adjustment, and 21% of the differences are larger than 2×10-8 m·s-2/a. Therefore, the piece-wise linear dynamic adjustment model can provide more reliable information of the temporal gravity changes compared with the traditional static adjustment model.

In geographical information engineering, height anomalies must be known in order to convert GPS ellipsoid heights into geodetic heights. There are many conversion models, such as polynomial, BP neural network and multi-quadrics fitting. Because the quasi-geoid is an irregular geometric object, every method has both advantages and disadvantages, and is appropriate to different conversion patterns. It is difficult to identify which conversion model is most suitable for a particular area. In order to obtain a more precise and reliable analytical result, the combined model based on least squares support vector machines (LSSVM) has been approached. Derived from statistical learning theory, LSSVM are learning systems that use a hypothesis space of linear function in a high dimensional feature space, trained with a learning algorithm from optimizations theory. As a result, the fitting ability of single models is significantly improved. The combined model still possesses the same important features as the single models. Examples are presented and the results are analyzed in detail to demonstrate the efficiency of the proposed methodology.