Unconventional reservoir characterization using geochemical signatures: Examples from Paleozoic formations, Saudi Arabia

The soaring domestic energy consumption in the Kingdom of Saudi Arabia (KSA) continues to create a demand and supply gap. Paleozoic tight sand and shale reservoirs in KSA are considered potential sources of natural gas that could meet the growing domestic demand. To enhance the exploration and exploitation of these unconventional systems, research focusing on their reservoir quality has tremendously increased in the past few decades. A challenging aspect of these Paleozoic reservoirs is that they have similar lithologies and are poorly constrained biostratigraphically. Therefore, facies analysis, stratigraphic correlation, and geosteering are problematic. This paper describes how chemostratigraphic markers are used to identify and separate individual lithologies. Three hundred and twenty (320) data points were acquired by scanning both end caps of 165 core plug samples from continuously cored sections of three Paleozoic formations. Raw elemental spectral data were acquired using a high-resolution tabletop μXRF system. Multivariate statistical analysis including principal component analysis (PCA), pairwise correlation, and hierarchical clustering of principal components (HCPC) were used to characterize the acquired geochemical values. Eleven (11) chemofacies, within three lithofacies with clearly defined boundaries, were identified. Stratigraphic distribution patterns of the geochemical signatures (e.g., Si, Ca, Mn, Al, K, and Fe), such as sharp truncation at the upper and lower bounds, clearly defined each lithofacies. Thus, key marker elements for each lithofacies and chemozones were identified, even where the lithologies appeared homogeneous. The overall results indicate sediment deposition under varied depositional settings, from fluvial to shallow-marine environments. The results demonstrate that elemental compositional analysis can be used as a discriminating tool for resolving stratigraphic uncertainties. Such data can document subtle variations among samples that appear homogeneous using other techniques.