In marine oil and gas exploration and development, accurately identifying the location of reservoir fluids is crucial for enhancing detection accuracy and development efficiency. In recent years, the study of fluid factors by geophysicists has significantly propelled the development of this field. Fluid factors, as critical parameters for describing reservoir fluid properties, can optimize drilling locations, design extraction procedures, and accurately estimate reserves and predict production performance. The Russell fluid factor, owing to its distinct physical meaning and superior performance, especially in complex geological structures, has become a focus of research. However, the acquisition of fluid factors often depends on indirect calculation or inversion through approximate reflection coefficient formulas, which may easily introduce errors and reduce the accuracy of oil and gas predictions. This article proposes a nonlinear inversion scheme for reservoir fluid factors based on the analytical solution of the exact Zoeppritz equations. The Newton-Raphson-based optimizer (NRBO) algorithm is introduced to address complex nonlinear problems in the inversion process. The inversion test results using the East China Sea ocean bottom node (OBN) seismic data validate the effectiveness of this method in identifying reservoir fluids.

Rock physics and borehole data indicate the presence of directionally aligned fractures with moderate dip angles within the formation. Accurate identification of these fractures is crucial for the efficient exploration and development of unconventional hydrocarbon reservoirs. To enhance the prediction accuracy of fracture parameters in fractured reservoirs, this study derives an approximate reflection coefficient equation for tilted transversely isotropic (TTI) media under the assumption of weak anisotropy. The proposed method is based on the anisotropy parameters (A-parameter) and incorporates linear-slip theory to establish a relationship between the A-parameter and fracture weakness parameters in VTI media. This relationship is further extended to TTI media through coordinate rotation. To achieve stable and reliable estimation of fracture-related parameters, we develop an amplitude-difference-based AVAZ inversion (ADI) workflow within a Bayesian framework. Tests on synthetic models demonstrate that the proposed method is robust to noise and provides accurate inversion results. Furthermore, applying the method to OBN seismic data from a field area in the East China Sea confirms its feasibility and effectiveness, with inverted results showing strong agreement with well log data. This study offers a novel, efficient, and reliable framework for fracture prediction and reservoir characterization in geologically complex settings.

The assumption of transversely isotropic media with a vertical symmetry axis vertically transverse isotropic (VTI) has been widely utilized in the exploration of shale reservoirs, with its reflection response and pre-stack inversion attracting significant attention from geophysicists. However, accurately estimating anisotropic parameters using existing approximate reflection coefficient formulas for VTI media remains challenging. In theory, inversions based on exact reflection coefficients for VTI media could overcome these limitations. Nevertheless, such inversions present a highly nonlinear problem. The simultaneous inversion of five parameters further exacerbates the ill-posed nature of the inversion, making it difficult for conventional linearized algorithms to handle effectively. To address these challenges, this article proposes a nonlinear pre-stack inversion scheme based on the exact reflection coefficient equations for VTI media. Specifically, we establish an inversion objective function within a Bayesian framework and introduce a heuristic Newton-Raphson-based optimization (NRBO) algorithm for solving the nonlinear objective function. This approach mitigates the inaccuracies introduced by conventional linearized algorithms, enhancing the precision of the inversion results. Tests on 1-D single-well synthetic data, 2-D Hess VTI model synthetic data, and field data demonstrate that the proposed inversion scheme can stably and effectively estimate both elastic and anisotropic parameters simultaneously. Moreover, the inversion accuracy of the proposed approach surpasses that of inversions based on an approximate formula.

Prestack seismic data retain the amplitude variation with offset (AVO) characteristics, providing more geophysical information reflecting lateral reservoir variations, thus facilitating the identification of oil and gas reservoirs. However, due to the band-limited nature of seismic data, the precision of forward modeling, and the accuracy of algorithms, traditional prestack approaches suffer from ambiguity and uncertainty. With the development of deep learning and big data, an increasing number of deep learning methods have been proposed. We integrate the extended long short-term memory (xLSTM) modules with the Unet framework, and design a novel neural network architecture, that is, xLSTM-Unet, for elastic parameter inversion ($V_{\text {P}}$,$V_{\text {S}} $, and$\rho $) from prestack seismic gathers. Through testing on synthetic seismic records and field data, the proposed xLSTM-Unet outperforms both the traditional Unet and LSTM-Unet models in predicting elastic parameters from prestack seismic data. The xLSTM-Unet proposed in this article provides a stable and effective approach for predicting prestack elastic parameters, offering new insights for the intelligent development of seismic exploration.

Russell’s fluid factor plays a crucial role in predicting hydrocarbon reservoirs. Numerous studies have been conducted on the inversion of Russell’s fluid factor. However, most of these studies exploit the approximate formulas of Zoeppritz as the forward equation. The approximate formulas affect the accuracy of the estimated Russell’s fluid factor, particularly in the moderate to large incidence angle range. To address this issue, we propose a nonlinear inversion scheme for Russell’s fluid factor based on the exact Zoeppritz equation. Initially, we derive a new form of the Zoeppritz equation that incorporates Russell’s fluid factor and Poisson’s ratio. Then, in the Bayesian framework, we establish a joint PP-PS objective function that represents by the Russell’s fluid factor. To solving the nonlinear objective function, we enhance the traditional whale optimization algorithm (WOA) by incorporating the Lévy flight strategy and stochastic learning theory, resulting in an optimized version named LSWOA (Lévy Flight and Stochastic Learning Whale Optimization Algorithm). The Russell’s fluid factor estimated by the new inversion method could provide a theoretical basis for the identification of oil and gas reservoirs, and also provide a new way for unconventional reservoirs prediction and sweet spot identification. To validate the accuracy of our method, we test it using both synthetic and actual field seismic data. Our test results demonstrate that the PP-PS wave joint inversion method of the Russell’s fluid factor based on the LSWOA algorithm exhibits high inversion accuracy and can be effectively applied in production.