Plate convergence, foreland development and fault reactivation: primary controls on brine migration, thermal histories and trap breach in the Timor Sea, Australia

Abstract

During the latest Miocene and Early Pliocene (∼5.5. MaBP), the collision of the Australian and Eurasian plates resulted in proto-foreland development and significant structural reactivation in the Timor Sea, north-western Australia. Flexural extension, resulting from the down-warping of the Australian plate into the developing Timor Trough, resulted in the dilatation of the major Jurassic and older extensional faults and the formation of shallow Mio-Pliocene fault arrays. An integrated, multidisciplinary study of hydrocarbon traps from this region using 2-D and 3-D seismic data, stable isotope geochemistry, fluid inclusion measurements and apatite fission track data has revealed that this fault reactivation produced three categories of traps: high (HIT), moderate (MIT) and low (LIT) integrity traps. These have characteristic hydrocarbon fill–spill, fluid flow and thermal histories. In MITs and LITs, the dilatation was moderate to intense respectively, and allowed hot (90–130°C), highly saline (200,000 + ppm salinity) brines from deep Palaeozoic evaporites to migrate up the reactivated faults and chemically and thermally affect the reservoir and shallower intervals. Apatite fission track data suggest that fluid migration lasted for between 100,000 and one million years in the case of the MITs, but for only 10,000–100,000 years in the LITs. This major fluid flow event resulted in the development of a prominent, localised Late Tertiary heating spike in the MITs, which can significantly affect the accuracy of modelled thermal histories. In the LITs, the thermal effect is less marked, due to the more transient nature of the fluid flow event. HITs were largely unreactivated and hence conduits for brine migration from depth were absent. Consequently, these traps are the most representative of the thermal histories of the source rock depocentres. Where MITS or LITs were charged, the associated loss of fault seal integrity facilitated hydrocarbon loss from the Mesozoic reservoirs, which co-migrated with the brines up through the Mio-Pliocene fault network. Upon entering a shallow, clastic aquifer system (the Eocene Grebe Formation), bacterial oxidation of the hydrocarbons liberated CO 2 which, in turn, resulted in significant and very isotopically light carbonate cementation. This cementation produces sufficient acoustic impedance with the surrounding uncemented sands that it allows these hydrocarbon-related diagenetic zones (HRDZs) to be mapped seismically. Since both the size and acoustic response of the HRDZs are directly proportional to the amount of hydrocarbons that have leaked from the traps, their presence or absence provides a powerful indicator, predrill, of both trap integrity and the likely thermal regime that that traps have experienced. An important observation is that the leaky fault segments over partially breached traps typically only extended for 200–1000 m, whereas over the breached traps, leaky segments extended for 3000–5000 m. Consequently, exploration programs acquiring remote sensing geochemical data (such as geochemical sniffer and airborne laser fluorosensor (ALF) techniques), should have closely spaced line spacings if leaky, potentially commercial fields are to be detected reliably. Potential analogues exist between the processes documented during HRDZ formation, namely the mixing at shallow depths of basinal brines, hydrocarbons and connate waters, and processes occurring during the formation of Pb–Zn and other, low temperature ore deposits.