Deep gas well encounters ultramafic kimberlite-like material in the Sauk Sequence of southeastern Ohio, USA

We report our preliminary findings on the newly discovered Murray ultramafic zone near Zanesville, Ohio, based on limited subsurface data. We use the rock term “kimberlite(?)” informally, until further geochemical analyses can be performed. The discovery well was drilled in 1994 by Clinton Oil Company (permit 27892¹) in search of natural gas, based on a seismic reflection anomaly beneath the Murray lease in Muskingum County, Falls Township (see Figs. 1 and 2). The well reached a total depth of 2065 m in the Middle Cambrian Conasauga Group and was completed as a gas well in a 20-m-thick Middle Ordovician sandstone overlying the Knox unconformity. Examination of mud-logging well cuttings from the Murray well using a 30× binocular microscope indicated abundant limonite staining of Cambrian Knox and Conasauga Group dolomite within a 200 m interval. Microscopic examination also revealed a very unusual 20-m-thick zone of what appeared to be biotite-rich arkose. A limited volume of the total drill cuttings was available for microscopic study, ∼31 g from the 18-cm-diameter borehole of this 20 m interval. Other available cuttings, examined from other deep wells in the area, did not contain anomalous lithologies. Grain mounts were made to petrographically analyze samples from this anomalous zone. Reconnaissance scanning electron microscopy (SEM) with energy-dispersive spectrometry was performed on selected points within polished grain mounts. Subsurface mapping with available deep geophysical wire-line logs was done using computerized hand-contouring verified coincidence with a reported seismic reflection anomaly described in unpublished industry data (Peter McKenzie, 1998, personal commun.). As part of a graduate student project, gravity and magnetic data were acquired at the surface above the seismic anomaly.

Kimberlites are rare, small ultramafic intrusive bodies that were first named for diamond-bearing peridotite rocks of Kimberly, South Africa (Lewis, 1887). According to Dawson (1984), kimberlite is a hybrid rock, consisting of a complex assemblage of altered high-temperature minerals and megacrysts, including a wide range of wall-rock material. Field studies indicate that a wide range of mineralogy and rock types is typical with rocks reported as kimberlites in the Appalachian Basin. These extremely rare rocks, thought to be mantle derived, have been reported from surface exposures in the Appalachian Basin states of Kentucky, Maryland, New York, Pennsylvania, Tennessee, Virginia, and West Virginia (Bolivar, 1982; Parrish and Lavin, 1982; Southworth et al., 1993; Watts et al., 1992) (Fig. 1). Kimberlites of Elliott County, Kentucky, are associated with calcite of igneous origin and occur in vertical pipes (diatremes) as well as dikes, which weather rapidly at the surface (Bolivar, 1982). Parrish and Lavin (1982) noted the difficulty in locating kimberlite intrusions by magnetic methods because of weathering, low magnetic susceptibility, and small size.

The Precambrian unconformity surface in eastern Ohio is defined as the top of Grenville Province metamorphic and igneous rocks, and in western Ohio by the top of East Continent Rift Basin sedimentary and volcanic rocks and Granite-Rhyolite Province igneous rocks (Fig. 1). Bass (1960), and subsequently many others, interpreted a Precambrian magnetic boundary in Ohio, referred to as the Grenville front or Grenville front tectonic zone, separating the Grenville Province on the east from East Continent Rift Basin and older Granite-Rhyolite Province on the west. The Grenville front is thus a magnetic lineament or boundary.

The Precambrian crystalline rocks of Ohio have not been U/Pb age dated. Calculated whole-rock Rb/Sr dates for Ohio range from ca. 0.9 to 1.3 Ga (Lucius and von Frese, 1988). Consequently, Ohio Rb/Sr dates only indicate the presence of Precambrian rocks and do not show relationships to currently accepted provinces or boundaries. Drahovzal et al. (1992) used seismic reflection data and limited petrology and geochronology from deep wells to determine the timing of the Ohio Precambrian Provinces, i.e., the Eastern Granite-Rhyolite, the East Continent Rift Basin, and the Grenville Province. The Appalachian Basin architecture was controlled largely by subsidence of Grenville Province basement rocks east of the Grenville front during the Phanerozoic. Preexisting structural weakness zones in the Precambrian are hypothesized as an important factor in controlling Phanerozoic geology (Beardsley and Cable, 1983; Riley et al., 1993). Kimberlites, considered to be emplaced along deep-seated faults and fractures (Bolivar, 1982; Parrish and Lavin, 1982; Shultz, 1999), are well documented in the Appalachian Basin along a regional fairway east of Ohio and above the Cambrian Rome Trough rift system (Parrish and Lavin, 1982). Shultz (1999, p. 217) speculated, on the basis of proprietary seismic reflection data, that there might be subsurface kimberlites in eastern Ohio and western Pennsylvania.

The Murray kimberlite(?) is on the western flank of the Appalachian Basin, 80 km northwest of the Rome Trough. The Appalachian Basin strata in this area dip <1° east-southeast on mapped Ordovician and Cambrian surfaces. The Murray kimberlite(?) is not near known local or regional Precambrian structures. The nearest known structures are the Cambridge cross-strike structural discontinuity and the Starr fault system (Baranoski, 2002) (Fig. 1).

The total thickness of Paleozoic sedimentary rocks at the Murray well is estimated to be 2100 m. Only the Cambrian and Ordovician rock units are discussed in this paper. The Murray well penetrated 57 m of Trenton Limestone, 152 m of Black River Group (limestone), 43 m of Wells Creek Formation (shale, limestone, dolomite, and sandstone), 173 m of Knox Dolomite (Rose Run sandstone and Copper Ridge dolomite), and 152 m of Conasauga Group (dolomite and sandstone) (Fig. 3). The well did not penetrate Precambrian rocks. Detailed stratigraphic correlation and structural analyses with wells in the area did not reveal significant structural or thickness variations above the Ordovician Wells Creek Formation. The structure map on the Wells Creek shows east-dipping contours at ∼20 ft/mi (∼3.5 m/km) (Fig. 4A).

Two anomalous lithologies were encountered beneath the Black River Group: a localized 20-m-thick sandstone in the lower portion of the Wells Creek Formation at ∼1500 m below sea level and a 21-m-thick arkosic interval in the Conasauga Group at ∼1720 m below sea level (Fig. 3). Sample cuttings from available wells in the area were examined, and anomalous or exotic lithologies were not identified. The localized sandstone is gas bearing in the Murray well and overlies the Sauk Sequence or Knox unconformity. In this area, Rose Run sandstone of the Knox Dolomite is typically ∼30 m thick and consists of interbedded dolomitic subar-kose, quartz arenite, and dolomite (beneath the Knox unconformity). The sandstone directly overlying the unconformity is white, very fine grained, moderately to well rounded, moderately to well sorted, pyritic, and very friable. The unit is characterized on geophysical logs by relatively low photoelectric and gammaray response compared with the underlying Rose Run. It correlates to stratigraphically similar sandstones encountered in wire-line–logged wells within ∼1.6 km (Fig. 4B) of the Murray well. This thick sandstone was not observed during well drilling in adjacent areas.

A structure map of the Murray well area (Fig. 4A) depicts the Ordovician units overlying the Knox unconformity and indicates the expected north-south strike and eastward dip of ∼3.5 m/km. Structure mapping of the Cambrian Copper Ridge dolomite in the area, however, indicates a localized structural low dipping eastward at ∼19 m/km. The position of the structural low coincides with both the occurrence of the thick sandstone (Fig. 4B) overlying the Knox unconformity and a small fault-bounded graben. The graben originates in the faulted Precambrian Grenville basement complex and was interpreted from industry seismic data near the Murray well (Fig. 2). The Murray well reached total depth in the Middle Cambrian Conasauga Group at 2070 m. Structure mapping of the Conasauga is the logical next step; however, there have not been enough drill sites to the Conasauga Group near the Murray well. The nearest wells drilled are more than 16 km away, and are thus too distant to use as control points for contouring this localized graben. As a result, mapping beneath the Copper Ridge dolomite is not possible.

Kimberlites are mineralogically complex rocks that are derived from the mantle and violently emplaced within stable, usually ancient, cratons. They are volatile-rich, potassic, ultramafic alkaline rocks that occur as dikes, sills, and diatremes (Winter, 2001). They can vary widely mineralogically, but usually contain some olivine, phlogopite, ilmenite, magnetite, garnet, diopside, enstatite, and chromite (Mitchell, 1995). Deuteric calcite, serpentine, and chlorite are often present. Kimberlite suites also characteristically contain xenocrysts and xenoliths of other mantle-derived alkaline rocks as well as other wall-rock material picked up on their rise through the mantle and crust.

The ultramafic kimberlite(?) material described in this study consists of a few very small rock fragments ranging from 0.5 to 4 mm that were handpicked from ∼31 g total drill cuttings for the 20 m interval. Epoxy grain-mount glass slides were ground to 3 µm and polished for optical analyses and SEM. Phlogopite, Nafeldspathoids, titaniferous magnetite, ilmenite, apatite, and calcite were positively identified. In addition, Ca-Mg pyroxenes (probably diopside) and secondary amphiboles were identified in the groundmass by SEM. Although olivine, serpentine, and garnet were not positively identified in this preliminary study, they may be present in the groundmass or in other sample cuttings not mounted on the grain mounts. In one rock fragment, the geometric arrangement of opaque granules suggests alteration from another mafic phase (Fig. 5A).

The phlogopite is conspicuous and abundant in some of the rock fragments, varying from 0.01 to 1.0 mm (Fig. 5B). It occurs primarily as relatively fresh, subhedral to nearly euhedral phenocrysts, and to a lesser extent as smaller grains in the matrix. Optically, phenocrysts are biaxial (−) with a 2V angle of 2°–5°. They exhibit parallel extinction, poor (nearly absent) to good cleavage, and are reddish-brown. A few grains exhibit bending or kinking. Reconnaissance SEM analyses of the phlogopite indicate an Mg:Fe ratio of ∼3:1 with a relatively high titanium content (Fig. 6).

Overall, the mineralogy of the ultramafic zone in the Murray well is similar to that of other kimberlites described in the region (Bolivar, 1982; Alibert and Albarede, 1988; Shultz, 1999; Watts et al., 1992). However, we have not identified mantle-derived xenoliths containing olivine, garnet, and other mafic minerals in the well samples. Such an identification is also conditional because of the low volume of recovered borehole samples. Nevertheless, we interpret the phlogopite, apatite, and magnetite, and the secondary alteration minerals amphibole, chlorite, Fe-oxides, and possibly serpentine, as representative common constituents found in kimberlite-like suites. Unequivocal comparison of Murray well ultramafic material cannot be made to other well-studied regional kimberlites without core samples. Nevertheless, it is significant that the presence of ultramafic material has been described in the Conasauga Group, >90 m above Ohio's Grenville basement rocks.

Deep-seated faulting in the Precambrian Grenville basement complex is evident, based upon industry seismic reflection data collected directly over the Murray well (P. MacKenzie, 1998, personal commun.; G. Mason, 1999, personal commun.). The seismic reflection data indicate a highly faulted structural sag at depth, suggesting localized extension of the crust (Fig. 2). The basement faulting is roughly coincident with two features on regional contour maps by Hildenbrand and Kucks (1984a, 1984b), i.e., a north-south–trending total magnetic intensity, and a gravity high. Structure and isopach mapping of deep gas-well data indicates no significant anomalies above the Ordovician Wells Creek Formation. A thick sandstone unit overlying the Knox unconformity surface coincides with a localized structural low area mapped on the Copper Ridge dolomite and faulting in Grenville basement. The deep-seated extensional faulting of the Grenville basement associated with the graben may have provided sufficient conduits for transfer of mantle-derived ultramafic material into the Conasauga Group dolomite. Subsurface mapping with deep well control indicates that the faulting in the basement occurred prior to deposition of the Middle Ordovician Wells Creek Formation. Structure and isopach mapping of units above the Wells Creek shows only the expected regional strike and dip trends.

The Murray ultramafic zone is recognized only in sample cuttings. The geophysical well log (Fig. 3) shows poor clay and porosity development, which indicates that the zone is very thin and may not be weathered. The wire-line logs for the Murray well are not highly anomalous in this ultramafic zone. Gamma-ray and density log response in this zone are typical of Cambrian carbonate rocks in eastern Ohio and do not indicate washout zones, anomalous porosity, or lithology zones. In the absence of core samples, mineralogical verification of spiky wire-line log response is tenuous; however, Murray well wire-line logs suggest possible correlation where the gamma ray increases slightly to the right at points A and B, with photoelectric increase at point C (Fig. 3).

Because kimberlites are typically enriched with mantle-derived carbonates, we speculate that the Murray ultramafic rock may have passed through faulted Grenville-aged marble prior to being intruded into Cambrian carbonates. Two deep wells drilled in central and eastern Ohio recovered marble from the Precambrian Grenville rocks (Bass, 1960). In this scenario, the additional carbonate country wall rock further enriches kimberlites within carbonate; thus, distinguishing small intrusive bodies on wire-line logs is difficult. Alternative scenarios include horizontal tabular layers or sills; a faulted tectonic slice of Grenville metasediments; erosional deposits of materials shed from a nearby exposed igneous body; explosive tuffaceous sediments from a pipe; or an impact site. The dominance of K-rich phlogopite in the sample, and the correlation of this phlogopite-rich zone to a spiky gamma-ray log response (at points A and B; see Fig. 3), could support either thin vertical dikes or horizontal sill scenarios. The fresh-appearing phlogopite recovered in the well cuttings supports an intrusive or faulted structural scenario. Erosional or explosive sediments scenarios are not likely, because phlogopite-rich sediments probably would have weathered rapidly prior to burial on this portion of the stable craton. A slice of Grenville basement thrusted upward into the Conasauga could only be explained by localized strike-slip compressional stresses, versus what appears in the seismic reflection data to be dominantly tensional stress. Planar deformation features, one of the diagnostic criteria for impact sites, were not observed in well cuttings. Thus a bolide impact interpretation for the seismic anomaly, as presented by Mason et al. (2004), is not supported.

The Murray ultramafic zone has not been isotopically dated. Based on subsurface mapping, we speculate timing of original emplacement prior to Middle Ordovician time, which is much older than the well-defined eastern North America Mesozoic kimberlite magmatism ages (Heaman and Kjarsgaard, 2000). The Murray ultramafic zone is significant structurally because it is located on the stable craton on the Appalachian Basin and 80 km northwest of the fairway of shallow kimberlites in the Rome Trough (Parrish and Lavin, 1982). It follows that other kimberlites in southeastern Ohio may have been penetrated by drilling, but have not been identified because they are not easily distinguished from country rock on the wire-line logs, and may have been overlooked in well cuttings. We will not know if the Murray ultramafic zone is a kimberlite or another exotic rock until further mineralogical and geochemical work is completed. The presence of this ultramafic zone on the stable craton is exciting, and leads us to question when it occurred, and what is its mineralogy and geometry. It is possible that there are more exotic zones in the deep subsurface lacking surface expression; if so, other oil and gas and mineral deposits may be associated with these zones, and locating them is a further challenge.

We thank Jim Templin of Columbia Natural Resources, Charleston, West Virginia, for making well cuttings available; Greg Mason of The Energy Cooperative, Granville, Ohio, for sharing seismic data; Robert Smith and John Barnes of the Pennsylvania Geological Survey for scanning electron microscope analyses; Doug Pride of Ohio State University for analytical discussions; Lisa Van Doren for digital artwork; Larry Wickstrom, Dennis Hull, and Chris Tickle of the Ohio Department of Natural Resources, Division of Geological Survey, Chris Laughrey of the Pennsylvania Geological Survey; an anonymous reviewer for critical reviews of the manuscript.