Title: Applying Reservoir Models to Effective Exploration for Subtle and Unconventional Traps: Examples from the Permian Basin
Speaker: Louis J. Mazzullo, AAPG CPG, Mazzullo Energy Corp., Geological Consulting, Golden, Colorado
Date: February 5, 2009
Publication: The Outcrop, February 2009, p. 18-19
Mapping and identification of reservoir facies is a critical first step to the understanding of reservoir development and trend analysis. The use of 3D seismic has benefitted these tasks greatly, enabling us to visualize the stratigraphic sequences within an exploration area to more refined levels than ever before. However, oftentimes our initial evaluation of a new exploration trend area or initial offset development involves picking formation tops and sequence boundaries using well log correlations, and, later on, trying to force these correlations to synthetic seismic ties. This type of approach at times could lead to seismic mis-ties, failure to accurately predict precise formation tops and reservoir trends, and missed opportunities in recognizing more subtle stratigraphic or combination traps. A multivariate approach to subsurface analysis should therefore include detailed sample evaluation that calibrates log responses to actual rock sequences. Several examples of correlation pitfalls from the Permian Basin are presented here. These include applications to the Permian Clear Fork of the Midland Basin, Upper Pennsylvanian Cisco-Canyon and Lower Pennsylvanian Atoka-Morrow sandstones (and associated Chesterian carbonates) in southeastern New Mexico, and the Siluro-Devonian and Ellenburger basin-wide. Each of these formations presents unique challenges to reservoir prediction because of their complex depositional and post-depositional histories.
Some of the problems related to misidentification of sequences or correlations from the Permian Basin have relevance to other basins. Well-to-well log correlations will often cross time units and lead to an improper interpretation of the juxtaposition of depositional environments. Log responses, especially gamma ray signatures, are often similar from sequence to sequence because of the repetitive nature of deposition in cyclic carbonate and elastic regimes. Karsted horizons often show higher gamma ray responses, and are reported on mudlogs as shales, but can be, in actuality, shalesupported karst and cave-fill facies with locally high interstitial porosity. Some of these karst-related facies can contribute substantial hydrocarbon reserves, but are generally overlooked as potential reservoir facies because they appear pessimistically shaly on well logs. There are many well-known structures where Atoka-Morrow sands lay unconformably upon lower Mississippian (Osagian) or older carbonates, absent the entire section in between. In order to effectively explore for the downthrown reservoir trends, it is essential to know the precise stratigraphic sequence and the degree to which Mississippian and younger beds were removed. But it is often difficult to tell the difference among the carbonate formations simply by correlating well logs. Simply mapping a structure on the Silurian-Devonian or Ellenburger is not enough to define a play; it is important to determine the diagenetic and stratigraphic attributes of these rocks below the conventional mapping horizons (e.g., Woodford Shale) in order to more accurately map a prospect and determine offset locations. Karsting is prevalent throughout the lower Paleozoic section, and further confuses the picture. Effective exploration for more subtle, yet lucrative, traps in all these rocks will require more “out-of -the-box” thinking as larger, structured fields are played out.