Deploying Geostatistical Stochastic Inversion Modeling In The Characterization Of Reservoirs: A Case Study From An Onshore Niger Delta Field

Three-dimensional (3D) seismic data and well logs were conjointly used to characterize reservoirs in the Onshore field Niger Delta Field of South-eastern Nigeria. The goal was to model and improve on the limitations of seismic resolution in conventional inversion techniques. This is necessitated by the quest to better off understanding of the distribution of lithology and pore fluid. Both geo-statistical and deterministic inversion techniques were applied in this study to appreciate the impact of each process on reservoir recognition and demarcation. The methodology began with well log conditioning and well-to-seismic tie. Rock-Physics feasibility analysis prior to the seismic inversion was done to estimate the elastic attributes appropriate for distinguishing between different fluid types and lithologies. The rock Physics results revealed the separation of the Velocity ratio (Vp/Vs) versus Acoustic impedance (Ip) cross plot into three regions, which are hydrocarbon, brine and shale zones. This suggests that the inversion results would be capable of distinguishing hydrocarbon sands from shale. Cross-plots of Mu-Rho (μρ) versus Lambda-Rho (λρ) showed that clusters of data, which are separated into three different zones inferred to be potential hydrocarbon, oil, brine and shale zones. The geo-statistical stochastic inversion was done by integrating variogram models and probability density functions (PDFs) to identify high frequencies or low periods in the output inversion outcome. The deterministic inversion results show that hydrocarbon saturated-sands are identified from the following attributes: low values of acoustic impedance (2.11 – 2.24 ×104 ft/s*g/cc), low velocity ratio (1.70–1.82), low Lambda-Rho (14.4 –19.7 Gpa*g/cc). In all cases, the results of geo-statistical inversion provided more detailed and increased resolution than deterministic inversion, allowing for detailed reservoir characterization. In the geo-statistical inversion, some regions in deterministic inverted sections with low acoustic impedance, velocity ratio, and Lambda-Rho attribute values inferred to be hydrocarbon sands appeared as either shale or brine. The inversion shows that the lambda-Rho quality is more useful in determining fluid classification whereas the acoustic impedance attribute is a good lithology discriminator. The overall results demonstrate the workflow’s ability to accurately map the rock properties with higher resolution and the delineation of new forecast as an effective, cost-effective and decision-making instrument.