--> Karst – Is there a magic way to model it?

The 1st AAPG/EAGE PNG Geosciences Conference, PNG’s Oil and Gas Industry:
Maturing Through Exploration and Production

Datapages, Inc.Print this page

Karst – Is there a magic way to model it?

Abstract

Karstification plays an important role in the reservoir effectiveness of carbonate reservoirs. It is estimated that 38% of the world’s oil and gas volumes are associated with fields with karst-related features. Predicting the properties and distribution of karst is very challenging and commonly requires the integration of high definition 3D seismic data. Various techniques for karst modelling have been published but are generally field specific and not readily transferable to other carbonate reservoirs. This paper presents an innovative approach to karst modelling of the Antelope gas field in Papua New Guinea utilising 2D seismic data, modern-day analogues and appraisal well data. The Antelope gas field is situated within the Eastern Foldbelt of the Papuan Basin, in Papua New Guinea. The field is a part of a sequence of isolated, Miocene, platform carbonates, possibly related to structural relief generated during early stages of regional compression, with widespread dolomitization, dissolution and karstification. Evidence of karsts can be identified from well logs, core and drilling events, which can also be broadly linked to 2D seismic character. Given only moderate quality 2D seismic lines were acquired over Antelope the main challenge was modelling the spatial distribution of the dolomite and karst-impacted reservoir. Antelope has been interpreted as an atoll-shaped reef build-up overlying a basal platform carbonate with lagoonal facies located within the central low area of the atoll. A significant amount of dolomite is predicted – and observed on well penetrations – within the upper part of the reefs related to karst processes that enhance the reservoir quality. An exposed Miocene carbonate build-up from Mallorca, Spain was used as an analogue, representing a Carbonate Island Karst Model. This model suggests confining extensive karsts to the reef front facies, with karst caverns most intense in reef facies, but less frequently penetrating the tighter lagoonal facies. Karst channels also generally ran parallel to the reef front orientation, then normal to reef front into the back reef forming palm-like-shapes. It is observed that transparent and chaotic seismic features indicated porous dolomite/limestone within the reef build-up and were often associated with karsts/reef collapse. Continuous and strong reflectors indicated poor limestone/lagoon deposits with less karst influence. Such seismic character was used to define the outer boundary of karstification. Karst isopach maps were built from the analysis of well logs, seismic attributes, analogues and the conceptual geological model. These maps define the karst envelopes controlling the lateral karst distribution and approximate karst proportion between wells. Morphic Rock Facies based on Resistivity Image logs, core data, other log data and drilling events were used to guide development of karst flow units. The isopach maps and flow units were combined to develop the 3D karst network using Truncation Gaussian Simulation conditioned by depo-facies, zones and karst envelopes. Finally, the karst model was overlayed on the depo-facies model, before populating with petrophysical properties ready for dynamic simulation. By creating the karst envelopes and karst flow unit indicators, the conditioned Truncation Gaussian Simulation approach was utilised to model the Karsts. The model was validated and tuned by extensive well testing results giving confidence in prediction of Karst formations within the Antelope field.