Geological Assessment of Leakage Potential at the Michigan Basin CO2 Pilot…pages.geo.wvu.edu/~wilson/netl/MBReport-08.pdf · 2008-11-01 · pertinent to possible leakage of CO2

Geological Assessment of Leakage Potential at the Michigan Basin CO2 Pilot: Ostego Co., Michigan

Tom Wilson, Institute Fellow NETL-IAES and West Virginia University Arthur Wells, MMV Team Leader, National Energy Technology Laboratory Abstract Evaluation of the near-surface geology at the MRCSP’s Ostego Co., Michigan CO2 Pilot site reveals the presence of about 600 feet of glacial till covering lower Mississippian and upper Devonian shale intervals that includes the Antrim Shale. The injection zone is in the Bass Island Group approximately 4500 feet beneath the surface. The Antrim is an organic rich fractured Devonian Shale reservoir. Any CO2 entering the Antrim would likely migrate upward through extensive fracture systems in this reservoir. Leakage is known to occur in the area. Geochemical sampling reveals anomalies over Niagaran reef trends to the southwest. Hydrocarbon microseepage tends to be absent in produced areas. A study conducted in the vicinity of the pilot site by Toelle et al. (2007) revealed that a prevalent source of leakage in the area is associated with corroded well casing. The Niagaran reef in the area of the injection well was inadvertently water flooded by migration of water disposal from the shallow Dundee Formation along corroded well casing. The water flooding was extensive and terminated production in from the reef in some wells. Recommendations for additional sampling were based primarily on the results of the study presented by Toelle et al. (2007), but also included consideration of the regional fracture systems and local structural dip. Background Characterization of Near-Surface Geology at the Site A background geological characterization effort was conducted at the Michigan Basin pilot site with the intent to provide some perspectives and insights directly to the National Energy Technology Laboratory’s Measurement Monitoring and Verification team. The database developed as part of this effort covered an approximately 7 square mile area surrounding the pilot site in Ostego Co., Michigan. The following maps provide some background information on the subsurface geology down to the level of the Antrim “Dark” Shale. The Antrim is a Devonian shale equivalent to the Devonian shales of the Appalachians and an equally pervasive gas producer throughout the Michigan basin. A relatively large number of wells with information on this shallow zone (approximately 70) were available in the area. Very few deep wells were encountered in the area. Most of the wells obtained in this study were completed in the Antrim Shale. The surface is covered by a layer of glacial till between 500 and 600 feet thick (see Figure 1). The till thins to the north of the injection well and is locally thicker beneath the injection well.

1

Figure 1: Thickness of glacial cover in the region. The injection well is located roughly in the center of grid of CATS near the southwest corner of the middle section (Section 30) in this map. The depth to the till/bedrock interface at the injection well is about 640 feet. The stratigraphic column for the basin (Figure 2) shows the relative positions of strata in the sequence overlying the injection zone in the Bass Island Dolomite. The Bass Island injection zone lies at a depth of between 3000 and 3400 feet below sea level in the area of the injection well. The surface elevation at the injection well is 1188 feet above sea level putting the injection zone at a depth of about 4200 to 4600 feet subsurface. The glacial till in this area covers the lower Mississippian Coldwater Shale (Figure 2). The Antrim Shale is separated from the base of the glacial till by an interval of shale including the Sunbury and Bedford shale intervals (Figure 2).

2

Figure 2: General stratigraphic column for the Michigan basin area. Depth to base of till in the vicinity of the injection well is approximately 600 feet. The Bass Island lies about 3000 feet below the Antrim Shale Bedrock elevations at the base of the glacial till (Figure 3) are relative to the sea level datum and thus do not include the influence of variations in topographic relief which are superimposed on the total depth to the base of the till shown in Figure 1. On this sea-level referenced view, the pilot well is located over a local bedrock high. Groundwater drainage at least at deeper levels within the till will be away from the injection well to the northwest. The regional scale view

3

provided by maps indicates that at the level of the bedrock/till interface, the area surrounding the injection well is in a bedrock low.

Figure 3: Elevation variations on the base of the till relative to the sea level datum. The structure on the Antrim Shale (Figure 4) dips gently to the south-southwest in the vicinity of the injection well and in general throughout the surrounding area. There do not appear to be local structures related to faults in the area. Small perturbations in the contours are probably due to variations in the density of well coverage in the vicinity of the injection well. Some anomalous features in the surrounding area have been checked. A few wells in the data base are deviated which leads to some differences in measured versus true vertical depths.

4

Figure 4: Structure on the top of the Antrim Shale. The Antrim Shale is an organic rich shale that produces gas from open fracture systems. As the depth to the top of the Antrim map reveals (Figure 4) comes to within about 1000 feet of the surface and may contribute to the soil gas anomalies that are pervasive in the basin (Wood et al., 2004a and b). Since the Antrim is a fractured reservoir any CO2 that might make its way up through the strata overlying the Bass Island injection interval will migrate through the fracture systems of the Antrim in the updip, north-northeast direction in this area (Figure 4). The updip direction also brings the unit closer to the surface (Figure 5).

5

One possible suggestion for placement of additional CATS was to place one to the north where the thickness of the interval separating the base of the till and the Antrim is thinnest. As shown in Figure 5, the interval separating the base of the till and Antrim is thinnest to the northwest (about 400 feet thick). This also happens to be in the area where the Antrim is nearest the surface (within about 600 feet; see Figure 6).

Figure 5: Depth (measured depth) to the top of the Antrim shale.

6

Figure 6: An isopach map of the thickness of the interval between the base of the till and the top of the Antrim Shale. Based on the near surface geology and extensive drilling history at the site our primary concern would be that CO2 might possibly migrate out of existing wells in the area. We would expect any CO2 making its way up into the Antrim to migrate updip. However, the dip is slight in the area (just over 100 feet per mile (a little over a degree to the south-southwest). Nearby updip wells are at considerable distances to the east and west (about 1600 feet and 1500 feet respectively). The cluster of 3 wells to the southeast lie at a depth of 10 feet or so less than that at the injection well. Ryder (2003) indicates that fractures observed in outcrop about 10 to 15 miles north of the pilot site have dominant orientations of N52E and N46W with subordinate sets oriented north-south and east-west. These trends are inferred from approximately 5000 measurements across the northern part of the basin. Ryder (2003) reports that fracture trends remains fairly consistent throughout the area.

7

The north-northeast rise in the Antrim structure combined with fracture orientation data suggest that additional CATS could be placed to the northeast along the N52E fracture trend (see Figure 7). Ryder (2003) provides a comprehensive summary of the general details of fracture observations and literature review. Observations of more than 600 fractures taken from oriented core reveal consistency with the outcrop observations (Richards, Waters, and others, 1994). Based on limited information restricted to near-surface geology it is possible to make some general recommendations for placement of additional monitors. First, it has to be determined which if any of the surrounding wells in the area are going to be monitored for CO2 and how that monitoring will be performed. If local fracture systems in the Antrim shale control flow of escaped CO2 then two additional CATS could be placed in the up dip direction along an azimuth associated with the major systematic fractures noted by Ryder. A secondary location would lie to the northwest of the injection well along the azimuth corresponding to the secondary NW trending fracture set.

Figure 7: Possible locations for a couple additional CATS, Michigan Basin site are shown as black outlined diamonds. Surface Geochemical Evidence of Leakage in the Area Extensive geochemical sampling conducted by Wood et al. (2004 a and b) reveal the presence of hydrocarbon seeps associated with Niagaran production. Reefs exhibited propane geochemical anomalies that formed haloes surrounding one reef, with highs surrounding the reef and lows

8

directly above the reef. Pentane anomalies were also observed in some areas. Variable results were obtained in their studies and the reader is referred to the papers by Wood et al. (1004a and b) for the details of their study. Their studies do suggest that light hydrocarbon microseeps are present in the area associated with deeper hydrocarbon reservoirs. These anomalies tend to be absent over produced fields. Their observations suggests the presence of migration pathways facilitating short term migration of light hydrocarbons from the Niagaran reef trend which lies at greater depth than the Bass Island injection zone. Well Integrity and the History of Well Leakage at the Site Our evaluation incorporates results and observations from an earlier enhanced gas recovery (EGR) operation conducted by Core Energy in the deeper Niagaran reef trend. Schlumberger is cooperating with Core Energy in a 3D seismic analysis of the reef complex that includes a focus on the 4D or time lapse seismic response resulting from CO2 injection into the reef. A paper by Toelle et al. (2007) provides some interesting perspectives on well history in the area that are pertinent to possible leakage of CO2 injected in the shallower Bass Island Formation as part of the MRCSP carbon sequestration pilot test. Several wells are noted in the Toelle et al. (2007) study (Figure 8).

Figure 8: Structure on the Antrim shale showing the location of key wells in the Toelle et al. (2007) study.

9

Toelle et al. (2007) specifically mention the Charlton 4-30 as the injection point for the Bass Island sequestration test. They indicate that the well is expected to become available for operation in the deeper Niagaran reef following CO2 injection. The broader study serves purposes that go beyond the objectives of the MRCSP Bass Island pilot study. Toelle et al. note that corroded casing is common to all wells. Disposal of produced water in the shallower Dundee Formation helped facilitate degradation of the casing and eventual leakage. Eventual well corrosion caused disposal water injected into the Dundee to drain into deeper intervals. This led to indirect water flooding of the deeper Niagaran reservoir. The 2-30 well (Figure 9, east of the C2-30 well) began to produce water in 1985. A 100% water cut arrived at the C2-30 well, about 20 feet down dip, in 1997. This resulted in the termination of primary production in the field. Appearance of CO2 in the Charlton 1-30 well injected from the end of the C2-30 lateral (Figure 9) indicates a nearly east-west connection between these two points.

Figure 9: The end of the C2-30 injection lateral extends approximately 800 feet north-northwest of the C2-30 surface location. CO2 injected into the C2-30 migrated into the Charlton 1-30 well.

10

Core Energy injected CO2 into the Niagaran through the deviated C2-30 well and temporarily produced oil from the 1-30 well to the north. The plan was to produce from the 1-30 well until it began cycling unacceptable amounts of CO2. At that point, the 1-30 well would be converted to an injection well with the idea of pushing the remaining oil to the south. The history of the 1-30 included 5 months of water production with no oil. The well began to produce CO2 in the production stream only a month after oil production occurred.

Figure 10: Suggestions for additional CATS locations. Note that the NETL MMV team had already decided to monitor some of these points. Based on the history of well corrosion, water dumping and inadvertent flooding of the deeper Niagaran Reef in the area, recommendations for CATS placement were revised as shown in Figure 10 (blue diamonds). Given the history of significant casing corrosion, CATS placements were recommended near wells noted in the study by Toelle et al. (2007). Although at some distance, the presence of well corrosion along with interconnection over a distance of approximately 1 km would make this a good location to monitor.

11

Conclusions: Geologic characterization of the site reveals the presence of a thick layer of glacial till in the area surrounding the CO2 injection well. The thickness of the glacial till near the injection well is approximately 600 feet. The injection well is located in a bedrock low where till thickness increases to about 640 feet. The till was deposited on the lower Mississippian Coldwater Shale. The upper Devonian shales including the Antrim Shale lie within a few hundred feet of the surface. The Antrim Shale is a tight naturally fractured gas reservoir that lies a few hundred feet below the till. The Antrim is a prolific gas producer in the Michigan Basin. The Antrim rises gently with a dip of about one degree to the northeast across the site. While this dip is quite small, if CO2 were to escape from the injection zone and reach the Antrim it might move through regional fractures systems in the up-dip direction. Geologic characterization of the area did not suggest the presence of any faults. Some early recommendations for additional sample placement were made on the basis of dominant trends in the Antrim natural fracture systems reported by Ryder (2003) and on the local structural dip. Geologic characterization of the site provides the group with several perspectives regarding the potential for CO2 leakage and likely leakage mechanisms. Previous geochemical studies conducted in the area by Wood et al. (2004 a and b) reveal that light hydrocarbon seeps are prevalent in the area. These seeps have been associated with production from the Niagaran reef trend which underlies the Bass Island CO2 injection zone. The presence of these seeps reveals that migration pathways are present in the area and can facilitate rapid migration of light hydrocarbons to the surface from intervals deeper than the CO2 injection zone. Extensive leakage associated with corroded wells in the immediate vicinity of the CO2 injection well was reported in detail by Toelle et al. (2007). Based on their study, additional recommendations were made in direct support of the NETL plan to place samplers near wells close to the CO2 injection well. References Richards, J.A., Walter, L.M., Budai, J.M., and Abriola, L.M., 1994, Large and small scale structural controls on fluid migration in the Antrim Shale, northern Michigan basin, in Advances in Antrim Shale Technology: Gas Research Institute in cooperation with the Michigan section SPE [Mount. Pleasant, Michigan, Dec. 13, 1994], 23 p. Ryder, R.T., 2003, Fracture patterns and their origin in the Upper Devonian Antrim Shale gas reservoir of the Michigan Basin: A review. USGS Open File Report 96-23. Toelle, B., Pekot, L., and Mannes, R., 2007, CO2 EOR from a north Michigan Silurian reef: Procedings paper, Spcietyof Petroleum Engineers SPE-111223-PP, 6p. Wood, J. R, Wylie, A., and Quinlan, W., 2004a, Surface geochemical results complement conventional development approaches: in WorldOil Magazine (WorldOil.com) - http://www.worldoil.com/magazine/MAGAZINE_DETAIL.asp?ART_ID=2456&MONT -Online Magazine Article Features – Dec-2004, 10 pages. Wood, J. R, Wylie, A., and Quinlan, W., 2004b, Using recent advances in 2DE seismic technology and surface geochemistry to economically redevelop a shallow shelf carbonate reservoir” Vernon Field, Isabella County, MI: DOE Quarterly report, Award number DE-FC26-00BC15122, 23p.

12