Detailed compositional analysis of gas seepage at the National Carbon Storage Test Site, Teapot Dome, Wyoming, USA

Abstract

A baseline determination of CO₂ and CH₄ fluxes and soil gas concentrations of CO₂ and CH₄ was made over the Teapot Dome oil field in the Naval Petroleum Reserve No. 3 (NPR-3) in Wyoming, USA. This was done in anticipation of experimentation with CO₂ sequestration in the Pennsylvanian Tensleep Sandstone underlying the field at a depth of 1680 m.

The baseline data were collected during the winter, 2004 in order to minimize near-surface biological activity in the soil profile. The baseline data were used to select anomalous locations that may be the result of seeping thermogenic gas, along with background locations. Five 10-m holes were drilled, 3 of which had anomalous gas microseepage, and 2 were characterized as “background.” These were equipped for nested gas sampling at depths of 10-, 5-, 3-, 2-, and 1-m depths. Methane concentrations as high as 170,000 ppmv (17%) were found, along with high concentrations of C₂H₆, C₃H₈, n-C₄H₁₀, and i-C₄H₁₀. Much smaller concentrations of C₂H₄ and C₃H₆ were observed indicating the beginning of hydrocarbon oxidation in the anomalous holes. The anomalous 10-m holes also had high concentrations of isotopically enriched CO₂, indicating the oxidation of hydrocarbons. Concentrations of the gases decreased upward, as expected, indicating oxidation and transport into the atmosphere. The ancient source of the gases was confirmed by ¹⁴C determinations on CO₂, with radiocarbon ages approaching 38 ka within 5 m of the surface.

Modeling was used to analyze the distribution of hydrocarbons in the anomalous and background 10-m holes. Diffusion alone was not sufficient to account for the hydrocarbon concentration distributions, however the data could be fit with the addition of a consumptive reaction. First-order rate constants for methanotrophic oxidation were obtained by inverse modeling. High rates of oxidation were found, particularly near the surface in the anomalous 10-m holes, demonstrating the effectiveness of the process in the attenuation of CH₄ microseepage. The results also demonstrate the importance of CH₄ measurements in the planning of a monitoring and verification program for geological CO₂ sequestration in sites with significant remaining hydrocarbons (i.e. spent oil reservoirs).

Introduction

The proposal for deep subsurface disposal of CO₂ initially seems to be an excellent approach. However, gases are buoyant relative to ground waters, and will tend to migrate vertically. Over-pressured systems are particularly susceptible to leakage as demonstrated by pressure-manipulated gas storage reservoirs used to meet seasonal variation in gas demand of metropolitan areas. Relatively little is published on this potential problem as the data tend to be sequestered as part of legal settlements. A few publications demonstrating surface leakage are in the public domain (Coleman et al., 1977, Araktingi et al., 1982, Arp, 1992, Jones and Burtell, 1996). Seepage rates rapidly respond to reservoir pressure manipulation as demonstrated in the Santa Barbara channel, California (Quigley et al., 1999).

Since the behavior of CH₄ is much more “ideal” than CO₂ in the geologic environment, inclusion of CH₄ in studies of potential microseepage is critically important. This is because CO₂ is soluble in, and reactive with water, whereas CH₄ is not. Possible migration of CO₂ will thus tend to be attenuated by these processes. Faults and fractures are also important in the transmission of buoyant fluids toward the surface. This study at Teapot Dome has the objectives of refining surface detection methods for microseepage, then characterizing it in an underpressured field prior to the injection of large amounts of CO₂. The first phase of this research was reported in Klusman (2005). An earlier study at Rangely, Colorado USA reported results for a field that had been under a CO₂ flood for approximately 15a (Klusman, 2003a, Klusman, 2003b, Klusman, 2003c).

The Teapot Dome Field is an asymmetric anticline on the southwestern corner of the Powder River Basin, Wyoming, USA (Fig. 1).

The surface stratigraphy on the flanks of the Teapot Dome is dominated by the Parkman Member of the Upper Cretaceous Mesaverde Formation. The Parkman Member has been removed by erosion from the central part of the Teapot Dome, leaving the upper part of the Steele Shale at the surface. There are extensive bentonite beds in the Steele Shale that undoubtedly control to some extent the hydrology of the area, and serve as a seal on the shallow hydrocarbon-producing Shannon Sandstone member.

Underlying the Steele Shale are the Niobrara and Carlile Shales, then the Frontier Formation which contains at least 3 hydrocarbon-producing members. Of these, the Second Wall Creek Member is the most productive and was under pressured at the time of the baseline survey.

In the Paleozoic, the Pennsylvanian Tensleep Formation has two sandstone members which also produce hydrocarbons at Teapot Dome. Tensleep A and Tensleep B are relatively clean sands, producing petroleum and low-salinity water. The Tensleep is at a depth of 1600–1700 m, and is the primary candidate for CO₂ injection and EOR experimentation. A surface geologic map and a stratigraphic column are contained in Klusman (2005).

McCutcheon (2003) developed a structural map from a 3D seismic survey. Numerous tensional faults developed at approximately right angles to the structure during uparching of the Teapot Dome structure. These are primarily recognized at the surface by offsets in the Parkman Member, and by near-vertical calcite veins in some of the faults. McCutcheon (2003) has used these data and 3D seismic data to determine the degree of extension of these faults into the subsurface, as well as their surface expression. This analysis was extended by M. Milliken (pers. comm.) with aerial photography. Recent mapping by Milliken of the outcrop of the Sussex sandstone member within the Steele shale has also been used to map faulting. A thorough study of faulting in the Teapot Dome is important in the evaluation of its integrity as a CO₂ sequestration site. A seismic cross section of the Teapot Dome and seismic panels illustrating the geometry of faults from McCutcheon (2003) is contained in Klusman (2005).

The flux and soil gas measurements were carried out over the period of January 7–January 28, 2004, and are summarized in Table 1 (Klusman, 2005). The similarity of the mean and median reflect a near-normal distribution.

A parametric and non-parametric statistical analysis of the flux data support a small positive CH₄ flux of 0.137 mg m⁻² day⁻¹, with a 99% probability of being positive. The CO₂ flux in Table 1 is dominated by biological sources, as indicated by stable C isotopes, with a few exceptions indicating either an atmospheric or geological contribution.

Table 2 summarizes concentration data for CO₂ and CH₄ in soil gas, and isotopic data for CO₂. Concentrations of CO₂ increase downward as would be expected based on CO₂ production from normal soil respiration activity. The mean and median values of δ¹³C for CO₂ do not vary significantly with depth, but the range suggests a slight ¹³C enrichment with depth. Methane shows very little variability with depth, indicating that CH₄ was in a small background range with very little evidence for the processes of microseepage or methanotrophy in the summary data.

A gas flux into the atmosphere is normally derived from a soil pore space in the case of a positive flux, and from the atmosphere to the soil in the case of a negative flux. This should be reflected in the relationships between flux and soil gas concentrations. Carbon dioxide flux correlates positively with concentration at all 3 measured depths, but only significantly at 100 cm (Table 3). The opposite behavior is observed for CH₄, giving a first indication of methanotrophic oxidation (Table 3). Since the soils were frozen to a depth of approximately 60 cm, most residual winter-time biological activity is at a depth below the frost line.