Abstract: Template-based modeling: Bridging the gap between quantitative outcrop studies and subsurface reservoir characterization

Reservoir modeling workflows that test sensitivity to fluid flow and/or seismic response are typically performed at a sector-scale (i.e., grid cells at 10’s meters areally and ~1 m vertically). Even at this “fine scale”, there is a gap between observations gathered from core and analogue outcrops at the bed-scale and the model scale. Furthermore, the link between the sector model and subsurface seismic responses are often difficult to understand. Outcrop modeling studies designed to investigate realistic heterogeneity through the construction of single deterministic 2-D or 3-D interpolated models are often limited in direct application to subsurface modeling workflows.

Reservoir modeling workflows that test sensitivity to fluid flow and/or seismic response are typically performed at a sector-scale (i.e., grid cells at 10’s meters areally and ~1 m vertically). Even at this “fine scale”, there is a gap between observations gathered from core and analogue outcrops at the bed-scale and the model scale. Furthermore, the link between the sector model and subsurface seismic responses are often difficult to understand. Outcrop modeling studies designed to investigate realistic heterogeneity through the construction of single deterministic 2-D or 3-D interpolated models are often limited in direct application to subsurface modeling workflows.

This work introduces the use of model templates derived from outcrop analogue studies that explicitly capture bed- to geobody-scale architecture (i.e., grid cells 2 m areally and 0.25 m vertically). The model templates serve as foundational depositional system building blocks and are used to gain a fundamental understanding of the impact of sedimentologic architecture in deep-water channels on subsurface fluid flow and seismic responses. Specifically, the objectives of this work are to use template-based models to investigate how intra-channel (bed-scale) and inter-channel (geobody-scale) stacking patterns control 1) both static and dynamic connectivity, and 2) seismic reflectivity response.

Template-based models are constructed from the Laguna Figueroa section of the well-exposed Upper Cretaceous Tres Pasos Formation in Chilean Patagonia. The models are derived from observations and statistical analyses from > 1,600 meters of cm-scale measured section from an ~2.5 km long by 130 m thick outcrop belt. A single representative deep-water channel element is constructed as a base case. Fluid flow and seismic amplitude responses are compared from a single channel element to those from a model composed of two stacked channel elements. This simplified approach fosters the ability to differentiate influences of stacking patterns from changes in internal architecture on connectivity and imaging.

Results show that flow connectivity and seismic responses are a function of both intra-channel architecture (bed-scale) and inter-channel architecture (stacking patterns) at the scale of sub-seismic and subsequently, sub-sector grid scale heterogeneity. Under waterflooding conditions, axial channel element deposits act as thief zones to injected water with vertically stacked templates, while marginal channel element deposits baffle and block flow between elements in laterally-offset stacked templates. Bed-scale architecture and smaller model grid-cell sizes are therefore more critical in channel systems (or parts of the channel systems) that predominantly contain laterally-offset stacked elements. Seismic amplitude and apparent thickness responses show discernable differences between single element and dual element models at realistic subsurface frequencies, particularly in vertically stacked templates and laterally-offset stacked templates. Two channel elements are more difficult to differentiate from one channel element in intermediate stacking patterns, but clues to multiple elements are revealed in interpreted horizon rugosity and in the length-scale of the seismic amplitude ( > than 1 channel element width). These models produce guidelines for coarse-scale model building (e.g., optimal grid-cell size and degree of fine-scale detail) and more informed seismic interpretation. This work also provides foundational models for seismic- and/or flow-based machine learning approaches.

Distinguished Lecturer

Lisa

Lisa Stright

Assistant Professor, Geosciences

Colorado State University

Video Presentation

Contacts

Heather Hodges Programs Coordinator +1 918-560-2621
Susie Nolen Programs Team Leader +1 918 560 2634