The Erlian basin is one of the most important basins in northern China to host sandstone-type uranium deposits (SUDs), in which Bayanwula, Saihangaobi, and Hadatu are under development, to name a few. Issues such as the metallogenic mechanism and mineralization of these deposits need to be addressed throughout the mining process. Over the past several decades, 2D and 3D seismic reflection surveys have been carried out to study these typical SUDs. The seismic technique has become the most effective geophysical tool of uranium (U) exploration, and it is used to develop our understanding of the stratigraphic configuration, faults, and sandstone contents of target layers in uranium environments. In addition, seismic interpretation could yield useful suggestions regarding the subsequent drilling program in the work area. There are two seismically predictable patterns of SUDs, named “Big depression + fault” and “Large-angle unconformity + fault”, which have been established following detailed seismic research in this basin. The characteristics of these faults are as follows: (1) the “‘U’-shaped formation” is conducive to the inflow of O-U-bearing groundwater into the target sandstone; (2) the “Big depression of reductive formation” provides plenty of organic matter (containing reducing media and U pre-enrichment) to promote redox reaction mineralization; (3) “Large-angle unconformity” is favorable to the migration of reducing substances, consequently leading to an enhancement in redox U mineralization; (4) “faults with long-term activity” become rising channels for reducing the presence of fluids and gases at depth; and (5) “sandstone and its scrambled seismic facies”. The results also offer indirect evidence of a connection between hydrothermal fluids and U mineralization; a hypothesis of “hydrothermal effusion” mineralization is proposed accordingly. In conclusion, seismically produced images of geological structures and sandstone distribution could yield important information for U prospecting and mine planning; it is worth considering seismic technologies in the future exploration of SUDs.

The primary exploration objective of Qiharigetu Sandstone-type Uranium (SU) deposit in the Erenhot Basin of China is to understand the stratigraphy, lithology and fault distribution of the target layer. Various geophysical techniques including gravity, magnetic, electromagnetic and 2D seismic have been proved not very effective due to its shallow depth of burial and small geophysical differences between layers. A specific 3D seismic survey of the Qiharigetu SU deposit has been carried out, and the inferred geological features of the uranium mineralization could provide important references for subsequent drilling programs. It is deduced that the target layer in the central part of the study area has a local "depression" through 3D seismic imaging and interpretation, which is conducive to river runoff and formation of sand body, as the sand body is a necessary precondition for mineralization. There are faults (5–50 m in throw) striking north–south for around 2 km, and it is an important ore-controlling factor, which could provide channels for the rise of reducing agents (such as gas, H₂S, and coalbed methane) at depth as they can promote the redox reaction for mineralization. The 3D seismic inversion is used to estimate the impedance, lithology, sand content and porosity of the target layer. With the help of drilling verification, it is found that the area with sand content over 75% and porosity over 10% is favorable for the SU deposit. Compared with other geophysical techniques, 3D seismic method is able to provide high-resolution images for unraveling SU mineralization and thereby reduce the drilling risk of the SU deposit.

Seismic prospecting has been accepted as one of the most widely available methods for exploring sandstone-hosted uranium deposits (SUDs). However, conventional seismic interpretation faces a challenge in the identification and characterization of a uranium reservoir’s complexity. How to characterize in detail a uranium reservoir’s physical complexity and effectively improve uranium reservoir prediction accuracy remain unresolved problems. To address this, we develop a novel combination of petrophysical modeling and prestack simultaneous inversion to understand in detail the physical properties of uranium-bearing reservoirs and efficiently predict favorable SUD sites. First, we develop a workflow of rock-physics modeling for SUD logs using the Xu-White method to calculate the modulus of elasticity of the grain matrix; subsequently, we extend the Walton model for the modulus prediction of the dry rocks and the Gassmann equation for one of the saturated rocks after a massive calculation test; and then, we predict the S-wave data used for the following inversion. Second, we execute a prestack simultaneous inversion to obtain the petrophysical parameters (e.g., P-impedance, density [ρ], shear modulus [μ], Lamé coefficient [λ], and Young’s modulus) that can provide insights into the physical properties of a uranium metallogenic environment. Accordingly, we discover that sites bearing uranium mineralization strongly correspond to areas with low elastic-parameter values (especially μ*ρ and λ*ρ), whereas nonuranium anomalies occur in high-value sites. This indicates that weakened elastic characteristics are caused by the enhancement of the total organic content and total clay mineral volumes of the uranium-bearing layers. In summary, the developed combination approach can yield an effective and accurate characterization of the geologic properties of uranium-bearing formations, and it can provide prediction factors (e.g., parameters related to the shear modulus) for uranium mineralization.

Seismic fault interpretation is a critical task for subsurface modeling and resource exploration. Although data-driven deep learning methods have shown considerable potential in automating fault segmentation, they often suffer from poor generalizability, limited interpretability, and physical inconsistency. To address these challenges, we present the first adaptation of the Physics-Informed Neural Network (PINN) paradigm to the seismic fault detection domain, proposing a modular loss function framework that encodes domain-specific constraints into network training. Specifically, the framework includes: (i) a physics-informed loss that uses the steady-state Allen–Cahn residual as a curvature-regularization (surface-tension) prior to sharpen and smooth fault interfaces; (ii) a topology-informed loss that penalizes departures from target Betti numbers via persistent homology, promoting a single connected, hole-free surface; and (iii) a geology-informed loss combining total variation, dip-angle smoothness, local connectivity, and branch suppression to encourage structurally coherent, geologically plausible predictions. This modular loss design is architecture-agnostic and can be seamlessly applied to any segmentation backbone. Experimental results on synthetic 3D seismic datasets demonstrate consistent improvements across four representative architectures (CNN, U-Net, ViT, and a Transformer-embedded U-Net), yielding higher accuracy, precision, recall, and F1-score. Qualitative analyses further confirm the generation of more continuous fault structures with reduced artifacts. Application to field seismic data validates the practical effectiveness of the proposed framework in enhancing geological interpretability. These findings underscore the value of embedding domain knowledge into loss functions as a principled and generalizable approach for improving the reliability of deep learning–based seismic fault interpretation.

The Erlian Basin is one of the most significant uranium-producing basins in China, mainly yielding two types of uranium deposits: roll-front (e.g., sandstone-related) and sedimentary-diagenetic (e.g., mudstone-siltstone-related). Although geophysical research has provided valuable insight into sandstone-hosted uranium deposits, there is a paucity of findings regarding mudstone-siltstone-related uranium mineralization. Therefore, the geophysical characteristics of mudstone-siltstone-related uranium deposits remain unclear, and a cost-effective prediction pattern on geophysical data has yet to be created. To address these challenges, we used integrated geophysical exploration methods (e.g., gravitational surveying, ground magnetic prospecting, radon surveying, electromagnetic surveying, and seismic exploration) to study the Nuheting syn-sedimentary-diagenetic uranium deposits (SDUD), which represent the largest mudstone-siltstone-related deposits in this basin. The results revealed the essential geophysical response characteristics associated with uranium-favorable sites. These areas tend to exhibit low-medium magnetic and gravity readings; a distinct pattern of soil radon anomalies was noted, following a “high-low-high” trend at these sites; the geophysical properties of the upper member of the target formation indicate low resistivity and unstable seismic facies, whereas the lower layers show high resistivity and scrambled seismic facies. Furthermore, this type of uranium mineralization often occurs in a “U”-shaped stratigraphic formation, typically situated in regions underlying faults. Based on the findings, we established a prediction pattern of the SDUD for the first time. In conclusion, a combination of gravitational, magnetic, radon, electromagnetic, and seismic prospecting techniques proves to be a method for effectively estimating U environments, and it is more efficient than any single approach in exploring the SDUD. Furthermore, the geophysical prediction pattern we develop allows for the economical and efficient discovery of similar SDUD in sedimentary basins. By optimizing the use of integrated geophysical techniques, this approach enhances uranium exploration while reducing the time and costs associated with traditional methods.