Unconventional reservoir characterization and sensitive attributes determination: A case study of the Eastern Sembar Formation, Lower Indus Basin, Pakistan

Abstract

The Sembar Shale formation in Lower Indus Basin Pakistan is thought to contain significant potential of unconventional resources; however, no detailed study has yet been carried out to quantify its potential. In conventional oil and gas exploration, reservoir rocks have been the main focus therefore, limited number of wells target the Sembar Formation. To explore its regional view, the seismic characterization of these shale is required. Generally, a poor correlation is generally observed between P-wave impedance and the reservoir and geomechanical properties of rocks, making it challenging to characterize them using seismic data. We present a workflow for characterizing the seismic derived unconventional prospect of the Sembar Shale using prestack seismic data along with well logs. The logging results of the two wells show that organic matter richness of well A is in high to very high values while, well B is in low to very low values. Considering the mineral composition and brittleness index evaluation the Sembar Shale in well A is brittle to less brittle in nature. The organic content, porosity, and brittleness index results in well A makes the Lower Cretaceous Sembar Formation favorable to be considered as a potential organic shale reservoir. Four sensitive attributes, derived through integration of the rock petrophysical, geochemical and geomechanical parameters, are correlated with P-wave impedance. The correlation of each sensitive attribute has been applied to characterize the Sembar Shale potential. These attributes are first-order indicators to depict organic matter, porosity and geomechanical properties. This attribute approach is further validated through rock physics modeling. The workflow presented in this study can be employed to assess unconventional reservoir potential of the Sembar Formation in other parts of the basin.

A deep learning approach to separating scattered direct waves from reflection seismic records on rugged surface: Synthetic data examples

“…Great progress has been made in the applications of deep learning technology for seismic data processing and interpretation, such as travel time computation (Waheed et al., 2021), seismic trace interpolation (Wang et al., 2020), first‐break picking (Yuan et al., 2018), velocity model building (Chen & Schuster, 2020; Chen & Saygin, 2021; Yang & Ma, 2019; Yu & Ma, 2021), geological body recognition (Huang et al., 2017), seismic facies classification (Ross & Cole, 2017), reservoir characterization (Abid et al., 2021), passive seismic event detection (Othman et al., 2022) and microseismic monitoring (Shaheen et al., 2021). For seismic data denoising, the convolutional neural network (CNN), a commonly used deep learning network (LeCun et al., 2015), has made good performance in random noise attenuation (Liu et al., 2020; Wang & Chen, 2019; Xie et al., 2018; Yu et al., 2019; Zhao et al., 2019b), such as white noise (Wu et al., 2019) and swell noise (Zhao et al., 2019a; You et al., 2020).…”