Introducing Mineral Magnetic Methodology and Fuzzy C-Means Clustering to Determine Controls of Reservoir Quality

Mineral magnetic techniques sensitively detect depositional and post-depositional changes in iron oxides and iron sulfides as a consequence of even minute changes in the thermal history or, for example, fluid migration in hydrocarbon plays. So-called remanent magnetic minerals are ubiquitously present in trace amounts in all sedimentary environments. We demonstrate, that established rock magnetic techniques provide distinct lithological and diagenetic zonations, hence, correlation surfaces in sediments, especially when integrated with element mineralogical approaches such as chemostratigraphy and a statistical evaluation of the data. We use fuzzy c-means clustering (FCM) combined with non-linear mapping (NLM).

Commonly, a detrital suite of magnetic minerals undergoes some (early) diagenetic alteration as a function of sediment depth due to increasingly reducing conditions, involving dissolution of magnetite and/or new formation of magnetic minerals, particularly sulfides. The extent of diagenetic overprint on the natural remanent magnetization (NRM) ranges from (virtually) absent to (nearly) complete. The degree of alteration is a function of the geochemical environment and (mostly) bacterially mediated organic matter degradation (e.g. reactivity of organic matter, sedimentation rate). Even minute changes in the geochemical environment will result in alteration of parts of the magnetic signal. Therefore, the sensitive mineral-magnetic techniques have the potential to reveal incipient alteration processes or diagenetic or lithological surfaces that otherwise go unnoticed if using more conventional geochemical or sedimentological approaches.

To arrive at a better understanding of the subtleties contained in the mineral-magnetic signal, the behavior of individual magnetic parameters serves as a basis for a joint interpretation of magnetic parameters and for example element mineralogical data (QEMSCAN, XRF, XRD) using multivariate statistical techniques. Fuzzy c-means clustering (FCM) and non-linear mapping (NLM) were used to identify groups within the data that could be assigned environmental and sedimentological significance (e.g. Urbat et al. 1997, 1999).

Several study examples from a variety of depositional marine settings (e.g. Urbat et al. 2000, Brandau and Urbat 2003, Urbat and Pletsch 2003) are reviewed with respect to their value for hydrocarbon exploration.

AAPG Search and Discovery Article #90142 © 2012 AAPG Annual Convention and Exhibition, April 22-25, 2012, Long Beach, California