Paleokarst reservoir architecture derived from forward modeling of modern cave systems—A case study

In some of the world’s largest hydrocarbon provinces, infilled and collapsed cave systems are an essential component for many paleokarst reservoirs. Conceptual modeling methods for this reservoir type are relatively sparse in the literature and often fail to simulate collapse processes realistically or forecast sediment infill. This study proposes a new concept-driven approach to paleokarst reservoir modeling. A novel interactive forward collapse modeling tool is developed to evaluate cavern stability and simulate collapse and infill processes of recent karst systems. The proposed tool uses a combination of cave survey data, conventional field measurements, and geomechanical data to simulate post-collapse morphologies and to generate new spatial data suitable for generating point-clouds. The collapse propagation, and the affected volume, is controlled by user-defined paleokarst facies proportions and associated average porosities. The point-clouds representing the various facies are used for geometric modeling and subsequent geocellular rendering. Inherent clastic sediments are discretized by horizon mapping and geometric modeling using electrical resistivity data as input. The volumetric and geometric accuracy of the reservoir models is compared to the simulated values and morphologies of the forward collapse modeling. Additionally, the interlinkage between bedding dip and cavern stability is appraised. This study demonstrates that our proposed method can be used to simulate collapse and infill processes of recent caves systems and generate paleokarst reservoir model analogues. Post-collapse morphologies can be simulated and subsequently rendered in a geocellular framework with a relatively high geometric and volumetric precision by employing easily accessible data. In turn, this could offer better constraints to forecast fluid flow behavior and seismic responses of paleokarst reservoirs given different collapse scenarios.