Abstract Shale reservoirs host ubiquitous multi-scale natural fractures created by factors such as pore pressure build-up after petroleum generation and palaeo crustal stress changes during tectonic episodes. Natural fractures are among postulated drivers of stimulated shale well productivity and related variability. Quantifying the aforementioned role and optimizing recovery call for litho-structural assessments. In this interdisciplinary study, natural fracture density (fractures per unit length) for North and South American shales (Marcellus, Eagle Ford, Haynesville, Barnett, Fayetteville and Vaca Muerta) was estimated from published observations of outcrops, cores and borehole images. Associated production for the latest horizontal wells, drilled in the most productive locations (sweet spots), was normalized by lateral length and reservoir thickness. It was found to correlate positively with the density of small-scale natural fractures. Durations of transient linear flow, diagnosed from production data, were play-specific, negatively correlated with small-scale natural fracture density and led to realization of picodarcy matrix permeability. Conversely, large-scale (tectonic) fractures limit stimulation efficiency and pose environmental/induced seismicity risks. Therefore, stimulation-driven reactivation of small-scale fractures facilitates drainage and enhances well productivity. Relatedly, reservoir flow regimes and production decline curves are intricately controlled by interplay of natural fracture density and matrix permeability. Variability of these parameters calls for acreage-tailored stimulations.

Abstract Hyperspectral core imaging as a commercial technology is fairly new to the oil and gas industry, providing a rapid and non-destructive method for mapping mineralogy of drill core and cuttings. The technology makes use of infrared spectrometers that collect imagery across the visible, near infrared, short-wave infrared and thermal (long-wave) infrared range that contains information related to a variety of mineral species, and in some cases including mineral chemistry and texture. In this paper we illustrate how the mineralogical information can be used directly for a variety of applications relevant to reservoir quality, including mineral mapping, sedimentological mapping and upscaling of micro information to the well scale. Statistical analysis allows the data to be used to segment cores into different rock types, and via calibration to quantitative mineralogical techniques, the data can be used to construct continuous curves of quantitative mineralogy, total organic content and production parameters.