Fracture zones are geological formations resulting from the strong movement of the Earth’s crust, often manifesting as fragile and sensitive areas. These zones are closely linked to natural disasters such as earthquakes and landslides. Accurate extraction of fracture zones is crucial for quantitative studies of earthquake faults, providing a scientific basis for risk assessment and decision-making in earthquake prevention and mitigation. Thus, an in-depth study to determine their distribution patterns and surface geometry is essential for understanding earthquake dynamics and mechanisms.

This paper addresses the shortcomings of existing methods in extracting fracture zones from LiDAR point clouds, which often suffer from incomplete extraction, poor continuity, and high error rates. We propose a method based on a multi-scale neural network with RS-Conv to improve the automatic extraction of fault zones in complex terrain regions. Fracture zones exhibit complex morphologies and scale features; therefore, single-scale neighborhood point sets fail to reveal their intrinsic structural information fully. Our approach begins by constructing neighborhood point sets at different spatial scales to comprehensively examine geometric features at various levels within the point cloud. The RS-Conv operator effectively portrays the spatial relationship between the center point and neighboring points. We then build a multi-scale neural network model using the RS-Conv operator as the convolution module. This model captures the spatial relationships in the point cloud, efficiently extracting deep features at different scales. The extracted multi-scale features are concatenated to form a richer and more comprehensive feature representation, which is inputted into a fully connected layer to classify the centroid and solve the fracture zone extraction problem. We compared our method with the Tensor Decomposition and Deep Neural Networks(DNN)methods using the ISPRS point cloud dataset, the Sichuan-Yunnan point cloud dataset, and the Xianshuihe dataset. Results show that our method achieves the highest classification accuracy across all three datasets. Specifically, our method’s total classification error is only 0.3%, a reduction of 0.91% -2.79%compared to other methods. This significant error reduction demonstrates the accuracy, stability, and reliability of our proposed method in handling complex point cloud data. The main conclusions of this study are as follows:

(1)The construction of neighborhood point sets at different scales reveals that the combination of these scales significantly impacts the model’s classification performance. Selecting appropriate scale combinations is crucial for optimizing the model’s classification accuracy, facilitating better distinction between fracture zone points and non-fracture zone points.

(2)Compared to traditional and machine learning methods, the deep learning network model developed in this study shows significant advantages in extracting fracture zones from point clouds. The model can automatically learn deep features from point cloud data and process large-scale, high-dimensional point cloud datasets, thereby achieving more accurate fracture zone extraction in complex terrain conditions.

(3)Comparative experiments on different datasets further demonstrate the proposed method’s generalization ability. It is effective not only in extracting fracture zones under single terrain conditions but also in maintaining stable performance across multiple terrain conditions. This adaptability enhances the extraction of fracture zones in various terrain scenarios.

In conclusion, the method proposed in this paper offers a novel approach to fracture zone extraction. It achieves higher classification accuracy compared to existing traditional and machine learning methods, effectively addressing the challenge of fracture zone extraction in complex terrain areas.

MODIS land surface temperature(LST)products are of great value in the exchange of atmospheric matter and energy, climate change research, and detection of thermal anomalies as earthquake precursors. However, due to the influence of the cloud, there are a large number of missing values in the MODIS LST data products, limiting its wide application. Therefore, in this study we propose a method of surface temperature reconstruction based on mixed model: SCLSTM(SSA-CLSTM). Compared with traditional methods, this method does not need to establish a complex regression relationship model. In addition, since CNN can fully extract local features of one-dimensional time series data, and LSTM can fully learn long-term time series features of data, the combination of CNN and LSTM is capable of fully learning potential features of data.

Firstly in this study, the trend value of LST time series is extracted by SSA model to fill the missing pixel, and the initial reconstruction of LST is realized. Then, CLSTM(that is, 1DCNN, three-layer stacked LSTM)model is used to learn the local temporal characteristics and long-term dependence of the data, and the iterative prediction of the surface temperature of the missing pixel is realized to complete the fine reconstruction of the data. Based on the experimental results in Hotan region of Xinjiang and Wenchuan region of Sichuan, it can be proved that compared with the other two existing reconstruction methods based on mixed models, the reconstructed return data error is minimum, and the consistency with the original data is the highest. The RMSE of this method can be reduced to 0.712K, the minimum is 0.695K, and the correlation between the reconstructed return data and the original data can reach more than 0.95. In addition, the reliability of the method is further verified by the measured surface temperature data of the meteorological station. In summary, the proposed method provides a new technical means and ideas for deep learn-based reconstruction work, and also provides a solid data foundation for the research of surface processes and seismic thermal anomalies.

Using MODIS MYD11A2 remote sensing data and based on the proposed new method, the reconstruction experiments were carried out in Hotan region of Xinjiang Uygur Autonomous Region and Wenchuan region of Sichuan Province with the strategy of “remove-construction-contrast”, and the results were compared with other two mixed model methods. In addition, the reliability of the reconstruction accuracy of the proposed method is verified based on the 0cm measured surface temperature data of the weather station. The main conclusions are as follows:

(1)Through the reconstruction experiment of LST in Hotan, Xinjiang, it is concluded that compared with the existing two hybrid model reconstruction methods, the new method can better capture the time series features of LST data, so that the reconstructed image can not only better maintain the texture features of the original image, but also improve the accuracy of data reconstruction. The reconstruction error is the smallest among the three methods, RMSE can be reduced to 0.712K, and the correlation with the original data can reach more than 0.95 after reconstruction.

(2)In order to prove the regional applicability of the new method, a reconstruction experiment was carried out in Wenchuan region of Sichuan Province, and the missing values were reconstructed using the proposed method. Through this reconstruction experiment, we found that the reconstruction method in this paper can achieve better data reconstruction effect even in areas with more cloud and fog coverage, poor weather conditions, and complex land cover types, which proves the reliability and regional universality of the method in this paper.

(3)To further verify the reliability of the new method, the accuracy of the new method was evaluated by using the measured 0cm surface temperature data of 6 meteorological stations in Hotan region of Xinjiang. Based on the temporal variation characteristics of the MODIS return data from 2015 to 2019, the return values of 2020 are reconstructed, and the reconstructed results are compared with the measured data. By comparing the 0cm measured data of the meteorological station and the data before and after reconstruction, it can be concluded that the correlation and average deviation of the returned data and measured data after reconstruction based on SCLSTM method are closer to the correlation and average deviation of the original data and measured data. Therefore, the reconstructed data based on the new method can maintain a good consistency with the original data.

(4)By reconstructing the missing value regions of Hotan region of Xinjiang in 2008 and Wenchuan region of Sichuan in 2020, we found that the texture of the images after the supplementary value is fine and natural, without obvious boundary effect. Therefore, it can be proved that the method in this paper can realize the data reconstruction of a large area with missing values.

In summary, the method proposed in this paper provides a new idea and technique for MODIS surface temperature reconstruction work based on deep learning, and also provides a solid data foundation for the ground process research and seismic thermal anomaly information extraction based on MODIS LST.

As one of the most serious geological disasters, earthquake is of sudden and destructive characteristic. Therefore, it is of great significance to earthquake monitoring and early warning. The phenomenon of surface thermal infrared radiation enhancement is a common precursor of moderate and strong earthquakes and has been used as an important reference information for early warning and short term prediction. A variety of explanations have been given to understand internal mechanism of the above phenomenon, in which the stress-induced heating hypothesis is widely accepted and has been confirmed in the laboratory rock mechanical loading experiments, that is, under ideal conditions in the laboratory, the rock heats up when it is pressed and cools down when it is stretched. Under field conditions in practice, however, weak seismic precursors of thermal anomalies are often interfered by various environmental factors(solar radiation, atmospheric movement and human activities, etc.), and it has not been investigated whether the corresponding relationship between the above crustal compression-extension motions and thermal radiation anomalies can be observed under field conditions. The earth's gravity field, as one of the basic physical fields of the earth, contains the density distribution of crustal structure, which can be served to study the migration of the earth's material, the deformation of the crust and the change of the stress field. In this paper, we use GRACE gravity and MODIS thermal infrared remote sensing data to verify the stress-induced heat hypothesis in the field with Wenchuan earthquake as the time node. Firstly, the crustal mass density obtained by GRACE satellite was compared with thermal infrared radiation. Then, the gravity anomalies extraction method based on maximum shear strain and in-situ temperature method were used to obtain the gravity anomalies and thermal anomalies respectively. Furthermore, the correlation between the two anomalies before the earthquake was detected from the time scale and space scale respectively, and the consistency analysis between the above anomalies and the spatial distribution of the tectonic fault zone was carried out. For this purpose, two important indicators i.e., anomaly intensity and anomaly distribution, were established in time domain and space domain, respectively. The following conclusions could be drawn: 1)The stress-induced heating hypothesis can be verified by remote sensing in field conditions. The warming zone of the crust(positive thermal offset index)corresponds to the compression zone, and the cooling zone(negative thermal offset index)corresponds to the stretching zone. The consistency of positive and negative variation between the crustal mass density and thermal offset index is 88.9%, which provides field observation evidence for the stress-induced heating hypothesis. 2)The spatio-temporal variation of gravity anomalies and thermal anomalies before earthquake has strong correlation. In the time domain, there is a strong correlation between the gravity anomalies and the thermal anomalies, which shows that the intensity of the two anomalies suddenly increases synchronously and reaches the maximum simultaneously three months before the earthquake. In the spatial domain, gravity anomalies mostly occur at the junction of positive and negative values of thermal offset index, which indicates that the spatial distribution of gravity anomalies and thermal anomalies also has a certain correlation. In addition, the two anomalies appear to be distributed along the fault zone for many times, which shows that they are closely related to tectonic activities.

The occurrence of earthquakes is closely related to the crustal tectonic movement and the migration of earth mass, which consequently cause the changes of the earth‘s gravitational field. Global time-varying gravity field data obtained by GRACE gravity satellite can be used to detect pre-seismic gravity anomalies. For example, gravity signals caused by several large earthquakes, such as the 2005 M_W8.6 Indonesia earthquake, the 2010 M_W8.8 Chile earthquake and the 2011 M_W9.0 Japan earthquake, have been successfully extracted using GRACE data. However, previous studies on GRACE satellite-based seismic gravity changes focused more on the dynamics of the co-seismic gravity field than on the pre-seismic gravity anomalies which are of great significance for the early warning of earthquakes. Moreover, the commonly adopted difference disposal of the gravity field with the gravity field of adjacent months or the average gravity field of many years when obtaining gravity anomalies cannot effectively remove the inherent north-south stripe noise in GRACE data. On the contrary, it is more likely to cause the annihilation of the medium-high order information in GRACE gravity field model, which results in the loss of some gravity information related to tectonic activities. To explore the pre-seismic gravity anomalies in a more refined way, this study proposes a method of characterizing gravity variation based on the maximum shear strain of gravity, inspired by the concept of crustal strain. In other words, the gravity strain tensor is obtained by further calculating the second-order gradient of the increment of disturbance potential after the removal of hydrological disturbance, and then the maximum shear strain of gravity is ultimately generated to characterize the pre-earthquake tectonic activities. Then, to better understand the seismogenic process of the fault zone by further extracting the pre-earthquake anomalous changes, the data of the maximum shear strain time series are analyzed in this study by means of the offset index K to describe the gravity anomaly. Because the maximum shear strain is calculated by the second-order gradient of GRACE gravity field, this method can suppress the stripe noise better than the difference disposal, thus effectively improving the sensitivity of gravity anomaly detection. The exploratory experiments are carried out in the Tibetan plateau and its surrounding area, which locates among the Pacific Ocean, the Indian Ocean and Eurasia, with the highest altitude, most complex topography and frequent strong earthquakes. Ultimately, the Wenchuan earthquake and Nepal earthquake were used as an example to complete the extraction of pre-earthquake gravity anomaly information by the above method, and the pre-earthquake tectonic activity of the fault zones was analyzed. The results show that a large area of gravity anomalies consistent with the spatial distribution of the fault zone appeared on the Longmenshan fault zone during the half a year before the earthquake, and the maximum anomalous value appeared within 50km from the epicenter, while no anomalies appeared during the non-earthquake period. In addition, compared with the traditional methods, the proposed method has a better ability to extract anomaly information of gravity field, which provides a new idea for understanding the dynamic mechanism of large earthquakes using GRACE data.

There are many factors related to the variations of TEC, and the changes of TEC caused by earthquake only occupy a small portion. Therefore, it is vital how to exclude the ionospheric interference of non-seismic factors accurately in the process of seismic ionospheric anomaly extraction. This study constructed a TEC non-seismic dynamic background field considering the influence of solar and geomagnetic activities. Firstly, the TEC components of half-year cycle and annual cycle are extracted by wavelet decomposition. Then, it establishes a regression model between TEC in which periodic factors are removed and solar activity index, geomagnetic activity index with SVR method(support vector regression)in non-seismic period. Finally, based on the constructed model, the solar activity index and geomagnetic activity index is used to reconstruct aperiodic components of TEC in earthquake's period. From the reconstructed aperiodic components of TEC plus the half-year periodic components and annual periodic components of TEC in the same period, the non-seismic dynamic background field is obtained. Comparing the residuals relative to original TEC values in non-seismic dynamic background field and traditional sliding window background, there are apparent monthly periodic change and semi-annual periodic change in the residuals of sliding window background, which can have obvious impacts on the subsequent seismic ionospheric anomaly detection. In order to test the validity of seismic TEC anomaly detection based on the background field construction method, this paper investigated the long time series TEC anomalies near Wenchuan city(30°N, 100°E)from March 1 to September 26 in 2008. It is found that under the condition of non-seismic disturbance such as solar activity and geomagnetic activity, TEC abnormal disturbance is rarely detected by non-seismic dynamic background field method, when compared with the traditional sliding time-window method. And before the earthquake, more TEC anomalies were detected based on the proposed method, also, they were more intense than those extracted by sliding window method. Therefore, the TEC background field construction method based on SVR(support vector regression)has superiorities in both system errors elimination, which are caused by solar, geomagnetism, the non-seismic ionospheric disturbance events and periodic fluctuations of TEC, and in reducing the false alarm rate of seismic TEC anomaly. Moreover, it can also improve the seismic TEC anomaly detection ability. In addition, this paper analyzed the time-spatial distribution of TEC anomaly before three earthquakes on May 12, August 21 and August 30, 2008. They were mainly negative abnormal perturbations and often distributed on the equatorial side of epicenter.

Earthquake events are one of the most extraordinarily serious natural calamities, which not only cause heavy casualties and economic losses, but also various secondary disasters. Such events are devastating, and have far-reaching influences. As the main disaster bearing body in earthquake, buildings are often seriously damaged, thus it can be used as an important reference for earthquake damage assessment. Identifying damaged buildings from post-earthquake images quickly and accurately is of real importance, which has guidance meaning to rescue and emergency response. At present, the assessment of earthquake damage is mainly through artificial field investigation, which is time-consuming and cannot meet the urgent requirements of rapid emergency response. Markov Random Field(MRF)combines the neighborhood system of pixels with the prior distribution model to effectively describe the dependence between spatial pixels and pixels, so as to obtain more accurate segmentation results. The support vector machine(SVM)model is a simple and clear mathematical model which has a solid theoretical basis; in addition, it also has unique advantages in solving small sample, nonlinear and high-dimensional pattern recognition problems. Thus, in this paper, a Markov random field-based method for damaged buildings extraction from the single-phase seismic image is proposed. The framework of the proposed method has three components. Firstly, Markov Random Field was used to segment the image; then, the spectral and texture features of the post-earthquake damaged building area are extracted. After that, Support Vector Machine was used to extract the damaged buildings according to the extracted features. In order to evaluate the proposed method, 5 areas in ADS40 earthquake remote sensing image were selected as experimental data, this image covers parts of Wenchuan City, Sichuan Province, where an earthquake had struck in 2008. And in order to verify the applicability of this method to different resolution images, an experimental area was selected from different resolution images obtained by the same equipment. The experimental results show that the proposed method has good performance and could effectively identify the damaged buildings after the earthquake. The average overall accuracy of the selected experimental areas is 93.02%. Compared with the result extracted by the widely used eCognition software, the proposed method is simpler in operation and can improve the extraction accuracy and running time significantly. Therefore, it has significant meaning for both emergency rescue work and accurate disaster information providing after earthquake.

In this work, using the method determining the earthquake magnitude from coda wave continuous time, the intensity characteristics of acoustic emission in rupture preparation process of the sample including hard inclusion is studied. The results show that the intensity of AE is steady at the early pressuring stage. When the pressure reaches to 89% the rupture pressure stress, the intensity obviously increases; when the pressure comes up to 92% the rupture pressure stress and the intensity reaches the maximum value greater 20% than the value at the initial pressuring stage. Then the intensity falls, drops to the minimum value and relative quiet activity of AE appears at this time with the pressure increasing. With the pressure further increasing, the intensity rises again, following closely the main rupture occurs. As viewed from the space distribution of AE, AE occurs in the periphery of hard inclusion in the process from the steady value at early loading stage to the maximum and roughly appears in the hard inclusion and above it from the low value to recovering of the high value. At last the main rupture occurs and the inclusion rupture locally too, which is similar to the time progress in phenomenon before the Lancang-Genma earthquake.

During the preparing process from compression to fracture,the elastic waves of gneiss have obvious various characteristics along the σ3 direction( obliquely crossing with main fracture plane): the various state of the first arrival time of P,S wave shows a shape of relatively high value～stability～high value～decrement,and the ratio of trave time has shape of stabiltity～low value～increment. The amplitude of P,S waves shows a shape of low value～stability high value decrement(low value) and their ratios have a shape of relatively stabllty～very high value～decrement. The variation state of attenuation coefficient of the wave trian shows a shape of stability～low value～increment.The main reason for those variations is the openingand closing of micro-cracks,the nonsynchronous variation of P, S wave and the difference of their variations.