To address the industry’s demand for sustainable oilfield development and safe production, it is crucial to enhance the scientific rigor and accuracy of monitoring ground stability and reservoir parameter inversion. For the above purposes, this paper proposes a technical solution that employs two-dimensional time-series ground deformation monitoring based on ascending and descending Interferometric Synthetic Aperture Radar (InSAR) technique first, and the composite dislocation model (CDM) is utilized to achieve high-precision reservoir parameter inversion. To validate the feasibility of this method, the Liaohe Oilfield is selected as a typical study area, and the Sentinel-1 ascending and descending Synthetic Aperture Radar (SAR) images obtained from January 2020 to December 2023 are utilized to acquire the ground deformation in various line of sight (LOS) directions based on Multitemporal Interferometric Synthetic Aperture Radar (MT-InSAR). Subsequently, by integrating the ascending and descending MT-InSAR observations, we solved for two-dimensional ground deformation, deriving a time series of vertical and east-west deformations. Furthermore, reservoir parameter inversion and modeling in the subsidence trough area were conducted using the CDM and nonlinear Bayesian inversion method. The experimental results indicate the presence of uneven subsidence troughs in the Shuguang and Huanxiling oilfields within the study area, with a continuous subsidence trend observed in recent years. Among them, the subsidence of the Shuguang oilfield is more significant and shows prominent characteristics of single-source center subsidence accompanied by centripetal horizontal displacement, the maximum vertical subsidence rate reaches 221 mm/yr, and the maximum eastward and westward deformation is more than 90 mm/yr. Supported by the two-dimensional deformation field, we conducted a comparative analysis between the Mogi, Ellipsoidal, and Okada models in terms of reservoir parameter inversion, model fitting efficacy, and residual distribution. The results confirmed that the CDM offers the best adaptability and highest accuracy in reservoir parameter inversion. The proposed technical methods and experimental results can provide valuable references for scientific planning and production safety assurance in related oilfields.

The accuracy and reliability of soil moisture retrieval based on Global Positioning System (GPS) single-star Signal-to-Noise Ratio (SNR) data is low due to the influence of spatial and temporal differences of different satellites. Therefore, this paper proposes a Random Forest (RF)-based multi-satellite data fusion Global Navigation Satellite System Interferometric Reflectometry (GNSS-IR) soil moisture retrieval method, which utilizes the RF Model’s Mean Decrease Impurity (MDI) algorithm to adaptively assign arc weights to fuse all available satellite data to obtain accurate retrieval results. Subsequently, the effectiveness of the proposed method was validated using GPS data from the Plate Boundary Observatory (PBO) network sites P041 and P037, as well as data collected in Lamasquere, France. A Support Vector Machine model (SVM), Radial Basis Function (RBF) neural network model, and Convolutional Neural Network model (CNN) are introduced for the comparison of accuracy. The results indicated that the proposed method had the best retrieval performance, with Root Mean Square Error (RMSE) values of 0.032, 0.028, and 0.003 cm³/cm³, Mean Absolute Error (MAE) values of 0.025, 0.022, and 0.002 cm³/cm³, and correlation coefficients (R) of 0.94, 0.95, and 0.98, respectively, at the three sites. Therefore, the proposed soil moisture retrieval model demonstrates strong robustness and generalization capabilities, providing a reference for achieving high-precision, real-time monitoring of soil moisture.

Given the complexity of near-surface soil moisture retrieval, a single machine learning algorithm often struggles to capture the intricate relationships among multiple features, resulting in limited generalization and robustness. To address this issue, this study proposes a multi-satellite fusion GNSS-IR soil moisture retrieval method based on heterogeneous ensemble machine learning models. Specifically, two heterogeneous ensemble learning strategies (Bagging and Stacking) are combined with three base learners, Back Propagation Neural Network (BPNN), Random Forest (RF), and Support Vector Machine (SVM), to construct eight ensemble GNSS-IR soil moisture retrieval models. The models are validated using data from GNSS stations P039, P041, and P043 within the Plate Boundary Observatory (PBO) network. Their retrieval performance is compared against that of individual machine learning models and a deep learning model (Multilayer Perceptron, MLP), enabling an optimized selection of algorithms and model architectures. Results show that the Stacking-based models significantly outperform those based on Bagging in terms of retrieval accuracy. Among them, the Stacking (BPNN-RF-SVM) model achieves the highest performance across all three stations, with R of 0.903, 0.904, and 0.917, respectively. These represent improvements of at least 2.2%, 2.8%, and 2.1% over the best-performing base models. Therefore, the Stacking (BPNN-RF-SVM) model is identified as the optimal retrieval model. This work aims to contribute to the development of high-accuracy, real-time monitoring methods for near-surface soil moisture.

The rugged terrain and dense vegetation in the mountainous area of Luding after the strong earthquake have made geologic hazards hidden and difficult to verify, and there are limitations in the fine-resolution monitoring of small-scale landslides, especially in the area covered by high vegetation. Currently, there is a lack of research on the application of L-band LuTan-1 (LT-1) for landslide detection in the dense vegetation-covered area of the Luding strong earthquake zone, and it is necessary to carry out the analysis of the detection capability of LT-1 for small-scale landslide hazards under the complex terrain and dense vegetation area. In this study, the Stacking-InSAR method was employed using LT-1 and Sentinel-1 satellites to conduct deformation monitoring and landslide detection in the Luding seismic area and to investigate the small-scale landslide detection capability of LT-1 in vegetation-covered areas. The results show that LT-1 and Sentinel-1 identified 23 landslide hazards, and their obvious deformation and landslide characteristics indicate that they are still in an unstable state with a continuous deformation trend. At the same time, through the detection analysis of LT-1’s landslide detection capability under high vegetation cover and small-scale landslide detection capability, the results show that the long wavelength LT-1 can be more effective in landslide hazard identification and monitoring than the short wavelength, and LT-1 with high spatial resolution can be more refined to depict the landslide deformation characteristics in space, which demonstrates the great potential of LT-1 in the refinement of landslide detection. It shows the significant potential of the LT-1 satellite data in landslide detection. Finally, the effects of geometric distortion on landslide detection under different satellite orbits are analyzed, and it is necessary to adopt the combined monitoring method of elevating and lowering orbits for landslide detection to ensure the integrity and reliability of landslide detection. This study highlights the capability of the LT-1 satellite in monitoring landslides in complex mountainous terrain and underscores its potential for detecting small-scale landslides. The findings also offer valuable insights for future research on landslide detection using LT-1 data in similar challenging environments.