Delineation of a buried volcanic system in Kora prospect off New Zealand using artificial neural networks and its implications

Abstract

The Kora volcano, a submarine Miocene andesitic stratovolcano, is documented to be buried below ~1700 m sedimentary strata in the northern Taranaki basin off New Zealand. The buried volcano and enclosing older sedimentary strata, structurally modulated the subsurface architecture leading to the formation of structural and stratigraphic traps for hydrocarbon reservoirs. The Kora field is known for accumulating sub-commercial hydrocarbon resources within the volcanogenic deposits. Here, we attempt to image such a complex geological system from 3D time-migrated seismic data using state-of-the-art artificial neural networks coupled with interpreter's acquaintances. For this, we have computed several attributes; optimally amalgamated these and trained over interpreter's intelligence. This has resulted into a single new attribute, defined as the intrusion cube (IC) meta-attribute, to produce the best possible image of the subsurface. The resultant IC meta-attribute has successfully brought out the extension and distribution of volcanic edifice within the buried volcanic system along with several structural elements such as the sill networks, dyke swarms, forced folds, drag folds, jacked up strata and pinch-outs (along flanks of the volcano) in the host sedimentary successions, which are very essential in understanding the petroleum system of the Kora field. This interpretational approach, based on a blended output of neural intelligence of artificial networks and interpreters' knowledge, can be suitably employed for imaging any complex volcanic system from 3D seismic data.