Williams Fork Formation Reservoir Characterization at An Overview of the Williams Fork Formation Reservoir Characterization at Mamm Creek Field, Piceance Basin, Colorado Matthew J. Pranter, Alicia Hewlett, Sait Baytok, Rachel Shaak University of Colorado at Boulder

Outline Research Objectives Study Area 3-D Reservoir Characterization and Modeling of Matrix Properties Seismic Interpretation, Fracture Analysis, and Fracture Modeling Conclusions

Research Objectives Within the southeastern Piceance Basin and Mamm Creek Field, how does the lower Williams Fork Formation vary in terms of stratigraphic architecture, shoreline stacking patterns, and lithology? For the Williams Fork Formation at Mamm Creek Field, what is the stratigraphic variability of sandstone-body type and distribution, matrix reservoir quality, and static connectivity? For the Williams Fork Formation at Mamm Creek Field, what are the main fault types and their distribution and how does fracture distribution vary with faulting, lithology, architectural elements and other parameters?

Study Area and Focus N Fracture intensity and seismic attributes Regional distribution of coastal plain and marine deposits Fracture intensity and seismic attributes 3-D reservoir modeling of tight-gas sandstones N Discrete fracture network modeling 3-D seismic analysis of fault and fracture distribution

Study Area and Focus data from 1,400 wells 3-D seismic survey 8 cores Measured Section Reservoir Model Area Seismic Survey data from 1,400 wells 3-D seismic survey 8 cores 12 borehole image logs outcrop sandstone-body statistics measured section Image Logs Cores

Reservoir Characterization and Modeling 2 mi2 91 wells irregular 10-acre density

Stratigraphic Interval for Reservoir Modeling (~2200 ft) Coastal /Alluvial Plain Shoreface/Deltaic Marsh, Mire, Swamp, and Estuarine Marine Shelf/Ramp Modified from Carroll and others (2004); Cole and Pranter (2008)

Model Framework 91 wells 15 zones 40’ x 40’ x 1’ 89.4 million cells Upper Williams Fork Formation Middle Williams Fork Formation V.E. = 1X Lower Williams Fork Formation Paonia Sh. Mbr. 91 wells 15 zones 40’ x 40’ x 1’ 89.4 million cells Upper Ss. Middle Ss. Cameo Rollins V.E. = 4X

Model Inputs and Constraints Calculated Lithology logs Interpreted Architectural Element Logs crevasse splay sandstone/ clean sandstone shaley sandstone point bar Variography Outcrop dimensional statistics and object shapes Experimental Variogram Azimuth: 39˚ Variogram Map

Architectural Element Object Shapes Channel Bar / Point Bar Crevasse Splay

Architectural elements Vertical Constraint: Vertical Proportion Curves Basic Lithology 0 0.2 0.4 0.6 0.8 1 Refined Lithology 0 0.2 0.4 0.6 0.8 1 Architectural elements 0 0.2 0.4 0.6 0.8 1 Upper Williams Fork Middle Williams Fork Lower Williams Fork (Paonia) Upper sandstone Middle sandstone Cameo-Wheeler Top Rollins

3-D Spatial Constraint: Seismic Probability Volume Middle Williams Fork Middle Sandstone Provided by iReservoir.com

3-D Lithology Modeling and 3-D Architectural-Element Modeling Sequential-indicator simulation (SIS) of basic lithology Sandstone, mudstone, coal modeled With 3-D seismic-based spatial probability constraint Sequential-indicator simulation (SIS) of refined lithology Clean sandstone, shaley sandstone, mudstone, and coal modeled Object-based simulation constrained to lithology Constrained to outcrop dimensional statistics for fluvial architectural elements (Pranter et al., 2009) Object-based simulation constrained to architectural elements

Modeling Examples Basic Lithology w/ 3-D seismic constraint Refined Lithology Basic Lithology Refined Lithology Architectural Elements

Future Modeling work Petrophysical modeling Porosity Permeability (conventional core data) Channel cluster analysis Static sandstone-body connectivity

3-D Seismic Interpretation Uninterpreted Interpreted vertical slices through seismic data (amplitude) Depth slice through ant tracking volume without interpretation Depth slice through ant tracking volume with interpreted faults Noise on survey edge

3-D Seismic Interpretation Fault interpretation based on seismic amplitude, ant-tracking results, and curvature attributes Two near-vertical fault sets (shallow faults strike N60W; deep faults strike N45E) N Uninterpreted Interpreted Picture on the left hand side shows uninterpreted seismic horizon surface map of middle sandstone with extracted ant-tracking attributes on it, red outline is the model area, dominant nne-ssw strike orientation can be seen Picture on the right hand side shows interpreted faults cutting this surface N Deep Faults Middle Sandstone

3-D Seismic Interpretation Fault interpretation based on seismic amplitude, ant-tracking results, and curvature attributes Two near-vertical fault sets (shallow faults strike N60W; deep faults strike N45E) N N Uninterpreted Interpreted Picture on the left hand side shows uninterpreted seismic horizon surface map of Top mesaverde formation with extracted ant-tracking attributes, red outline is the model area, dominant wnw-ese strike orientation can be seen Picture on the right hand side shows interpreted faults cutting this surface Shallow Faults Top Mesaverde Group

Fracture / Seismic Attribute Analysis Relationship between ant tracked seismic attribute and fracture intensity in fractured zones Fractured Zone Upscaled Fracture Intensity Ant-tracking Attribute Correlation coefficient: 0.73 Relationship b/w ant tracked attributes and fracture intensity exists within fractured zones, therefore ant tracked attributes can be used to constrain distribution of fractures Cross plots, which has 0.62-0.73 correlation coefficient, show good relationship b/w ant tracked attributes and fracture intensity logs within 5 fractured zones Only intensely fractured areas has seismic response, fracture swamps disturb seismic waves Fracture properties (intensity) vs. seismic attribute relationship is still being investigated, more attributes to look at curvature(most negative, most positive, maximum, minimum, gaussian)instantaneous frequency, t attenuation etc.

Fracture Analysis and DFN Modeling Fracture Intensity and Lithology + DFN (discrete fracture network) model Ki Kj Kk Fracture Dip Angle and Dip Azimuth Seismic attributes can be re-sampled into a 3-d grid and be used as fracture driver, in this case ant tracked attributes will be used to constrain fracture distribution; therefore, uncertainty can be reduced, distribution constraint to lithology and intensity property from fmi logs and seismic attributes Given the availability of 3-d seismic data and 10 fmi logs, more accurate reservoir models that predict fracture density can be created. Secondary properties can be used in modeling of intensity. Distribution constraints are: distance to faults, seismic attributes, lithology Orientation constraints are: dip azimuth, dip angle, and concentration properties Geometry will be constrained based on outcrop data Scale up fracture properties