Characterizing the Development of North American Source Rock Reservoirs from the Ordovician-Jurassic: A Proxy-Based Multivariate Geochemical Approach

Abstract

Source rock reservoirs (SRRs) are characterized by fine-grained facies that contain adequate organic matter (OM; typically >2 wt%) to generate and retain significant hydrocarbon accumulations. Though SRRs may be deposited in a range of settings, OM content and quality is largely controlled by the interplay of three factors; a source, flux to the sediment, and efficiency of preservation. Much work has been done using modern analogs to develop models for application to the ancient record. From these studies, it is clear that O2-deficient conditions favor preservation of qualitatively- and quantitatively-enhanced OM, with appreciable productivity and OM flux to the sediment required to establish and maintain these conditions; yet the relative dominance of these factors remains disputed. Here, we present detailed multi-proxy sedimentary geochemical studies of major North American SRRs to elucidate their depositional conditions. This is the first study focused on SRR geochemistry to use Fe-speciation – presently the only independent proxy able to distinguish anoxic conditions as ferruginous (H2¬S-limited) or euxinic (H2S-replete, Fe-limited) – coupled with total organic carbon (TOC) and redox-sensitive trace element proxies. Further, we apply ensemble machine learning techniques (multivariate random forest analysis) to define the relative importance of redox setting vs. productivity for a more complete view of the biogeochemical factors at play in the development of these units as SRRs. We show that no single factor (i.e., preservation, production, sedimentation) predominates control of these units across geologic time. Rather, our results suggest that while deposition under euxinic conditions characterizes most major SRRs, there is evidence for intervals of persistent ferruginous anoxia in the Paleozoic, and deposition in these settings does not preclude development of a SRR. Additionally, our results imply that predictive modeling techniques to can be used effectively on geochemical data collected during the drilling process to define both redox setting and TOC distribution within a formation. Together, these insights can improve targeting and steering as well as inform decisions in the completion and production stages. Ultimately, we propose that beyond a broad framework for redox evolution in which to place these observations, universal models should be abandoned in favor of more basin- or formation-specific perspectives.

AAPG Datapages/Search and Discovery Article #90373 © 2019 AAPG Eastern Section Meeting, Energy from the Heartland, Columbus, Ohio, October 12-16, 2019