Geostatistical Modelling of Hydrothermal Dolomite by PluriGaussian Simulation From Digital Outcrop Dataset (Latemar, N-Italy)

Abstract

A number of carbonate reservoirs show hydrothermal dolomite imprint, however little information exists on the geometry and distribution of dolomite bodies in these reservoirs. The spectacularly exposed Latemar platform (Northern Italy) represents an outcrop analogue for carbonate reservoirs affected by hydrothermal dolomitization genetically linked to igneous activity. Detailed lithological information obtained from digital outcrop dataset and outcrop observations of a hydrothermal dolomitized carbonate platform has been used to constrain the detailed modelling of the 3D dolomite body architecture. PluriGaussian Simulation (PGS) has proved useful in other instances to model carbonate reservoirs and is an improvement of previous effort to model the Latemar dolomites by Sequential Indicator Simulation because it allows to model multiple lithologies and preserve their connectivity. Six detailed pseudo-wells were “drilled” in the digital outcrop, and used as the data to constrain the modelling. The quantitative architectural information derived from the digital outcrop consists of the variograms of the observed lithologies (high-permeability dolomite, low-permeability limestone and magmatic dikes) and the cross-variograms between these three. The input to PGS consists of vertical proportion curves of the lithologies, and truncation diagrams or lithotype rules which are used to constrain the transitions between lithologies. The lithotype rules are a graphical representation of facies or lithology relationships. The models obtained with PGS satisfy the vertical proportion curves, truncation diagram and (cross-)variograms. In this study different lithotype rules are tested and compared to determine the most representative rule for modelling hydrothermal dolomite bodies. The resulting 3D model conforms outcrop observations and shows dolomite bodies which are elongated in dike-parallel directions and a preferential transition from dike to dolomite, with some degree of randomness allowing direct limestone to dikes transitions. The 3D model, after truncation by the digital outcrop topography, shows a satisfying resemblance with the real outcrop. This study shows that PluriGaussian simulation has the potential to be applied more generally to subsurface cases of hydrothermal dolomitization. Also other plays where diagenetic fluids come up along faults and alter the surrounding rocks would benefit from a similar approach.

AAPG Datapages/Search and Discovery Article #90259 ©2016 AAPG Annual Convention and Exhibition, Calgary, Alberta, Canada, June 19-22, 2016