This study conducts a detailed seismic hazard assessment of the Himalayan region. It focuses on studying how b-values, based on the Gutenberg-Richter law, vary throughout location and time. These fluctuations assist measuring tectonic stress and provide insights into the region’s seismic activity. This research focuses on five Himalayan sub-regions: Far Western, Western, Central-I, Central-II, and Eastern. It incorporates earthquake data spanning 1964 - 2023 obtained from the International Seismological Centre. The data were de-clustered using the Reasenberg method and examined by Maximum Likelihood Estimation. The results demonstrated considerable spatial variability in b-values across the Himalayan sub-regions. The Far Western Himalayas displayed the greatest b-value (0.93 ± 0.02), indicating frequent smaller earthquakes and lesser tectonic stress. In contrast, the Eastern (0.68 ± 0.02) and Central-I (0.69 ± 0.03) regions had the lowest b-values, implying more stress accumulation and a greater risk of future strong earthquakes. Temporal fluctuations, as a decrease in b-values preceding to the 2015 Gorkha earthquake (Mw 7.8) and a subsequent increase in Central-II (1.19 ± 0.03), highlighted the retention and release cycles. The Eastern Himalayas, particularly the Dhubri-Chungthang fault zone seismic gap in Bhutan, are considered a key high-risk zone. This region, with b-values ranging from 0.65 to 0.75, has remained unruptured since the 1934 Bihar-Nepal earthquake (Mw 8.4). The findings showed the influence of the continual convergence of the Indian and Eurasian plates (~20 mm/year) on strain heterogeneity. This study underlines the vital demand for intensive seismic monitoring, resilient infrastructure, and disaster readiness in low b-value areas to alleviate catastrophic risks in one of the globe’s most tectonically active regions.

Atoll structures formed in complex geological settings act as stratigraphic hydrocarbon traps and are typically circular or elliptical reef structures with a large lagoon at the center. Initially, the circular reef with flat limestone serves as a potential reservoir rock and holds significant importance in the petroleum industry, as it forms hydrocarbon-bearing traps. Therefore, identifying these structures in seismic sections is crucial. To understand the seismic behavior of atoll structures, seismic shot gathers of a geological model were generated, and migration sections were obtained. In this study, artificial data modeling of an atoll structure containing oil traps was carried out using the two-dimensional acoustic finite difference method due to its practicality and the flexibility to select different trap models as needed. Seismic data modeling was performed in a pre-stack shot domain, and two different data processing stages were applied to the shot data to obtain pre-stack and post-stack Kirchhoff time migration sections. The spatial location and size of hydrocarbon traps in the migration sections were determined and compared with the initial atoll model. In this way, the seismic response of hydrocarbon trap structures in the atoll model was analyzed. The importance of the two different data processing methods was also examined. As a result, it was observed that the pre-stack Kirchhoff time migration method provides better results than the post-stack time migration method for the atoll model.

This study proposes an enhanced method for natural earthquake and artificial explosion recognition, which comprises two parts, namely the multiscale fuzzy entropy (MFE) feature extraction of complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and the non-dominated sorting genetic algorithm III (NSGAIII) optimization of the one-dimensional convolutional neural network (1D-CNN). CEEMDAN decomposes earthquake signals into initial functions (intrinsic mode functions) and extracts fuzzy entropy features to construct a discriminative time-frequency representation. The hyperparameters of 1D-CNN (minimum batch size, initial learning rate, and learning rate drop factor) were optimized by NSGAIII, using a dual objective function to minimize mean squared error and maximize R2. Tests on 1000 earthquake events (883 earthquakes and 117 explosions) showed that the model has an accuracy of 97.82%, which is better than traditional networks (1D-CNN, generalized regression neural network, probabilistic neural network, back propagation neural network, and radial basis function neural network) and has better regression indicators (mean absolute error = 0.0795, root mean squared error = 0.1302, R2 = 0.7361). The Adam optimization algorithm achieved peak performance (99.50%), significantly surpassing SGD-M and RMSprop. This framework effectively solves the small sample and high-dimensional classification problems in earthquake monitoring and improves the automatic event detection capability of the early warning system.

Seismic data quality frequently deteriorates due to random noise contamination, substantially impeding subsequent processing and geological interpretation. While deep learning approaches have emerged as powerful tools for noise suppression, conventional single-stage architectures exhibit inherent limitations in handling complex seismic features while preserving subtle geological details. These challenges motivate the development of advanced multi-stage neural networks for seismic data enhancement. The proposed multi-stage progressive U-shaped convolutional network (MPU-Net) architecture addresses these limitations through supervised cross-stage attention mechanisms that maintain feature connectivity throughout the network. Building upon this foundation, group enhanced convolutional blocks (GEB)-MPU-Net introduces GEB to specifically counteract the progressive attenuation of shallow features in deep networks. This dual-stage enhancement strategy combines hierarchical feature preservation, adaptive information fusion, and stable gradient propagation. Comprehensive evaluation using both synthetic and field datasets demonstrates GEB-MPU-Net’s superior performance compared to conventional time-frequency analysis methods and established networks, such as U-Net, residual dense network, residual dense block U-Net, and MPU-Net. The architecture consistently achieves enhanced reflection continuity, improved geological feature resolution, and robust noise suppression. These advancements provide more reliable input for seismic interpretation, better preservation of subtle stratigraphic features, and increased applicability to challenging field conditions.

The formation of tectonic fractures is primarily influenced by stress distribution during the tectonic period. Therefore, in situ stress plays a crucial role in predicting fracture development zones. It significantly impacts the effectiveness of fractures by determining the size, orientation, and distribution pattern of fractures, thereby affecting stimulation results. Existing seismic methods for in situ stress prediction utilize seismic data to estimate stress parameters and calculate the horizontal stress difference ratio or the orthorhombic horizontal stress difference ratio (DHSR). These methods are based on the horizontal transverse isotropy or the orthorhombic anisotropy medium models. However, shale formations are often subject to tectonic movements that can rotate the symmetry axis of a transversely isotropic medium, leading to the formation of a tilted transversely isotropic (TTI) medium or a monoclinic medium with an inclined symmetry plane. Based on the TTI and monoclinic medium assumptions, this paper proposes new formulas for calculating the DHSRs (tilted transverse isotropy DHSR and monoclinic DHSR). The formulas are further validated through sensitivity analyses. Finally, this study demonstrates the effectiveness of the in situ stress seismic prediction method, grounded in TTI, and monoclinic medium theory through model-based examples.