Hydrocarbon Retention in Clastic Reservoirs of NW Borneo - Examples of Hydrocarbon Trap, Reservoir, Seal and Implications on Hydrocarbon Column Length

Abstract

Several siliciclastic fault-trap settings of the NW Borneo margin, from the West Baram Delta covering offshore Sarawak and Brunei to offshore NW Sabah, are compared with each other to assess the intricate relationship between hydrocarbon retention, seal capacity and reservoir parameters. Hydrocarbon column length is found to be a derivative of several parameters potentially affecting the integrity of a hydrocarbon trap. The presence of an effective and laterally continuous top seal is perhaps the most important success parameter; though relatively thin top seal can be surprisingly efficient. Seal capacity is featured by parameters such as mineralogy, grain size, contiguity, diagenesis and lateral continuity. Contiguity of hydrocarbon reservoirs is also important as discontinuous reservoir bodies commonly lead to very short and variable columns. The overall sand-to-shale ratio governs to some extent clay gouging capacity, with hydrocarbon columns tending to be longer in overall clay-prone environments, such as those found in outer shelf and deepwater turbidite environments. In these depositional settings, P50 columns are in the order of 250 m because sand-to-clay juxtaposition is more likely in fault-controlled traps. A better fault seal is often realized due to good shale gouge. However, hydrocarbon columns tend to be short (P50 of around 30 m) in sand rich shallow marine to deltaic settings given the discontinuity of reservoirs, leaky top seal, abundant yet discontinuous reservoir sand bodies and poor fault sealing capability. It is also observed that there are patterns of parameters such as sealing, reservoir, pressure and drive that have been identified, and combinations that appear viable (probabilistic success patterns) and non-viable for hydrocarbon retention (probabilistic failure patterns). The authors thus suggest to develop plausible patterns/scenarios and apply probabilistic simulations to each of the various combinations to assess the likely outcomes for column length predictions.