Predicting 3-D Connectivity From Core and Seismic: An Example From Outcropping Deep-Water Slope Channels

Abstract

Outcrop studies provide valuable analog data for prediction of deep-water channelized reservoir extent, architecture, and performance. However, analog data produced is often qualitative and presented in the form of annotated photomosaics and/or conceptualized models intended to illustrate subseismic-scale architecture in the subsurface. Statistics generated from outcrop analogs can be applied directly to subsurface modeling workflows. However, usage of statistics for guiding predictions of subsurface deposits beyond wells (e.g., channel-belt width and stacking) prior to analog application and model building is commonly understated. For example, elucidating a priori flow connectivity from statistics derived from cores and low-resolution seismic-reflectivity data can be useful in optimizing injector-producer well locations and preventing early water breakthrough. The effects of intra-channel facies relationships and stratigraphic architecture on the connectivity of a series of stacked channel-fill deposits has been investigated using exposure from the Late Cretaceous Tres Pasos Formation, southern Chile. A high-resolution digital outcrop model of stacked, deep-water slope-channel fills was generated based on > 1,600 m of cm-scale measured section and differential GPS points (10 cm accuracy) from an outcrop belt ~2.5 km long and 130 m thick. Spatial static connectivity was calculated stratigraphically from base to top of the modeled channel section and down depositional-dip to capture planview variability in connectivity. Results of this analysis show a clear correlation between internal channel architecture statistics derived from 1-D measured sections, channel stacking patterns interpreted from seismic-scale observations, and static connectivity. A strong relationship exists between channel mobility ratio (i.e., lateral vs. vertical offsets between successive channel fills) and static connectivity metrics. Furthermore, vertical trends derived from 1-D measured sections track vertical trends in connectivity metrics. Finally, the relationship between static and dynamic connectivity is captured using simple 2-channel sector models. The methods developed herein can be used to quantify risks of connectivity prediction from a single logged core and low-resolution seismic-reflectivity data.

AAPG Datapages/Search and Discovery Article #90291 ©2017 AAPG Annual Convention and Exhibition, Houston, Texas, April 2-5, 2017