The Plains CO2 Reduction (PCOR) Partnership, led by the Energy & Environmental Research Center (EERC), is working with Denbury Onshore (Denbury) to evaluate the effectiveness of large-scale injection of carbon dioxide (CO2) into the Bell Creek oil field for CO2 enhanced oil recovery (EOR) and to study long-term incidental CO2 storage. A technical team that includes Denbury, the EERC, and others are conducting activities to determine the baseline reservoir characteristics for development of a geologic model for predictive simulations and to serve as a comparison to time-lapse data as they are acquired. One of the activities was the acquisition and interpretation of a baseline three-dimensional (3-D) surface seismic survey acquired over a major portion of the field which is the subject of this report. The interpretation will be used to advance the field characterization effort. The geophysical data will be integrated with the EERC’s geologic model to improve its accuracy. In the future, when paired with at least one subsequent 3-D surface seismic survey, the data difference will provide a direct indication of where the CO2 has migrated within the reservoir and aid in monitoring, verification, and accounting goals.
A brief overview of the source testing, data acquisition, and processing is given. A seismic source configuration consisting of two heavy vibrators operated in tandem was chosen after a series of tests conducted in August and December of 2011. The data acquisition for the main survey occurred in August and September 2012. The data were processed by a contractor in Houston and delivered to Denbury early in 2013. The EERC received a stacked data set in April with redactions where Denbury did not have mineral rights or leases.
Seismic interpretation efforts began with making well ties to the data and identifying the Bell Creek reservoir reflector. The polarity of the dataset was established to be such that an increase in acoustic impedance (AI) would cause a negative deflection on the data. The reservoir is thin and of higher AI than the encasing shale layers, so with this polarity, it presents on the seismic data as a trough–peak combination representing the entering and exiting reflections at a two-way time of ~1150 msec at the 05-06 OW monitoring well location.
The origin of the reservoir reflection is due to a large increase in AI at the top of the Springen Ranch and a similar decrease in AI at the top of the Skull Creek. The measured thickness of the Springen Ranch-to-Skull Creek interval in the field varies from about 50 to 75 feet. Spectral analysis of data in the zone of interest reveals a bandwidth of 10–48 Hz, which together with the average interval velocity mathematically limits the vertical resolution of the seismic data at reservoir depth to just under 60 feet. Therefore, the reservoir is a thin-bed reflector with thickness near or below the limit of resolution. A reflector of this type would be expected to exhibit possible tuning effects, but they are not evident.
Tuning is an effect that occurs on thin beds when the top and bottom reflections interact to partially add in phase resulting in high reflection amplitudes. An important assumption is that the
ix
interval is lithologically homogeneous. The tuning thickness for this data’s bandwidth is 76 feet, so high amplitudes would be expected in thicker Springen Ranch-to-Skull Creek sections and lower amplitudes expected where the interval is thinner. The data exhibit the opposite character. This implies a nonhomogeneous interval where internal interactions due to lithology affect the composite reflection amplitude. The character of the Bell Creek sand meaningfully contributes to this effect.
The Bell Creek reservoir sand is a subset of the reflection interval. Typically 20 to 30 feet thick, its top and bottom cannot be directly resolved on the seismic data set as delivered. When the sand is thick and clean and juxtaposed against the harder, fine-grained components of the Springen Ranch at the top and the Rozet below, these impedance contrasts internal to the overall composite reflection appear to cause interference effects which result in a low-amplitude reflection. When the sand is thinner or fines upward with less AI contrast, less internal interference results, and the composite reflection maintains a higher amplitude. Because of this, a map of seismic amplitudes generated from the reservoir reflection differentiates between areas of 1) thick clean sand and 2) thinner clean sand or sand with a fine-grained component or shaley matrix. The effect is illustrated in cross sections paired with well logs and also by overlaying the map on existing isopach maps of sand, with good agreement.
Several geologic features are visible on the amplitude map and are briefly examined in cross section with the seismic data and well logs. These features include:
1. A fluvial channel feature in the northern part of the field, trending roughly north–south, has a higher amplitude than surrounding areas and is shown to be shale-filled and acts as a flow boundary.
2. A flow boundary between Phases 1 and 2 is also shown to be shale-filled and is a possible extension of the fluvial channel feature to the north.
3. A linear erosional valley trending northwest–southeast which predates the fluvial channel feature and serves as a flow boundary between Phases 1 and 3 and between Phases 2 and 4. The fill at the erosional surface is characterized by coal and bentonitic shales.
4. An erosional valley on the south end of the field which exhibits a dendritic outline. The fill at the erosional surface is characterized by coal and bentonitic shales.
Three structural aspects of the field are briefly explored:
1. Thinning of the Bell Creek sand which forms the updip boundary and trap to the southeast is not directly discernible on the seismic data.
2. Polygonal faulting is shown to be prevalent within the overlying Belle Fourche Formation. The faults are thought to originate from dewatering of thick bentonite layers at the bottom of the Belle Fourche and top of the Mowry. Faults do not extend below the top of the Mowry or above the Belle Fourche and likely do not impact the containment integrity of the reservoir.
3. A possible basement faulting system thought to control the southwest to northeast trend of the reservoir paleo-high was identified and is shown on a section.
Future work will include integrating the 3-D data and interpretation with the geocellular model and 3-D VSP data acquired in 2013 and 2014. Geomechanical properties will be computed across the field guided by the seismic data. Future time-lapse surface seismic data will use the baseline survey to generate difference displays to image and verify the progress of CO2 that has been injected into the reservoir. Modeling performed by Denbury has shown that injected CO2 will induce a detectable amplitude reduction on the reservoir reflection at the current bandwidth.
