November 2011 LIP of the Month

The Cambrian Wichita bimodal large igneous province in the Southern Oklahoma rift zone

Richard E. Hanson
School of Geology, Energy and the Environment,
Texas Christian University, Fort Worth, TX (r.hanson@tcu.edu)

Robert E. Puckett, Jr.
Oklahoma City, OK (bpuckett@priceedwards.com)

David A. McCleery
School of Geology, Energy and the Environment,
Texas Christian University, Fort Worth, TX (d.mccleery@tcu.edu)

Matthew E. Brueseke
Department of Geology, Kansas State University, Manhattan, KS (brueseke@ksu.edu)

Casey L. Bulen
Department of Geology, Kansas State University, Manhattan, KS (cabulen@ksu.edu)

Stanley A. Mertzman
Department of Earth and Environment,
Franklin and Marshall College, Lancaster, PA (stan.mertzman@fandm.edu)

Introduction

Opening of the Iapetus Ocean was preceded and accompanied by Neoproterozoic to Cambrian intraplate magmatism at numerous sites along the eastern margin of the Laurentian craton (Aleinikoff et al., 1995; Cawood et al., 2001; Tollo et al., 2004). There is less evidence for similar magmatism along the southeast margin of Laurentia. One notable exception occurs in southern Oklahoma and adjacent parts of Texas, the location of the Southern Oklahoma rift zone (Fig. 1). The rift zone contains an extensive bimodal igneous assemblage that is typically considered to be Early Cambrian in age but may also extend back into the late Neoproterozoic. This bimodal assemblage represents one of the single largest magmatic episodes associated with the opening of Iapetus anywhere along the eastern or southern Laurentian margins. In spite of this, the rift zone has received only limited attention in the LIP literature. Here we summarize important aspects of the rift-related bimodal rocks and present some results of our work in progress on the volcanic fill of the rift.

Geological Setting

The Southern Oklahoma rift zone penetrates into the craton interior at a high angle to the Paleozoic continental margin (Fig. 1). Hoffman et al. (1974) considered the rift zone to represent an aulacogen or failed rift arm of a rift-rift-rift triple junction, with the other two arms evolving into sites of seafloor spreading as Iapetus began to open. More recently, Thomas (1991, 2011) has argued against the aulacogen model and has instead interpreted the magmatism in the rift zone to have occurred in a leaky transform fault zone associated with a pronounced offset in the Paleozoic continental margin. Any model for the Southern Oklahoma rift zone must take into account the fact that the rift lines up to the northwest with a discontinuous belt of early Paleozoic alkaline complexes and tholeiitic dike swarms exposed in parts of Colorado, northern New Mexico, and Utah (Larson et al., 1985).

An older basin containing Proterozoic strata ~12 km thick occurs in the subsurface to the south of the Southern Oklahoma rift zone (Pratt et al., 1992), raising the possibility that the latter feature developed along a pre-existing line of structural weakness in the craton. Felsic igneous rocks of the ~1.4 Ga southern midcontinent granite-rhyolite province form the main subsurface basement units to each side of the Southern Oklahoma rift zone (Van Schmus et al., 1996) and are locally exposed in the Arbuckle Mountains, where they form the northern margin of the rift (Ham et al., 1964).

Inversion of the Southern Oklahoma rift zone occurred in the Pennsylvanian and Early Permian, related either to collisional Ouachita orogenesis along the southeast continental margin, or to far-field stresses transmitted inboard from the Cordilleran margin to the west (Ye et al., 1996). Major Cambrian rift faults were reactivated at this time to form reverse faults or thrusts in a compressional to transpressional regime, resulting in exposure at the surface of parts of the Cambrian bimodal igneous assemblage within the rift. Deep basins developed in association with fault-bounded uplifts along the trend of the earlier rift and became the sites of major petroleum accumulations. The most notable example is the Anadarko basin, one of the deepest continental basins in the world.

Cambrian igneous rocks within the rift form relatively limited exposures in the Wichita Mountains in southwest Oklahoma and in the Arbuckle Mountains farther east (Fig. 1). Elsewhere it is necessary to rely on subsurface data to constrain the distribution and relations of the igneous units in the rift. A surprising number of wells have been drilled into the igneous rocks, because they form the hanging walls of reverse faults that place them over petroleum-bearing Paleozoic strata. The information from these basement wells allows a much better understanding of the subsurface igneous assemblage within the rift than would otherwise be possible, although our knowledge along these lines is still far from complete. Ham et al. (1964) produced the first synthesis of the distribution of the igneous rocks in the subsurface of southern Oklahoma, based on studies of basement wells then available, and much of this pioneering work has stood the test of time.

Following cessation of Cambrian extensional tectonics and magmatism, thermal subsidence of the rift zone occurred during marine transgression across the region. The oldest strata to be deposited unconformably on the Cambrian igneous rocks during the transgression are Upper Cambrian (Ham et al., 1964), providing a younger limit on the timing of magmatism. 

Cambrian igneous rocks in the Southern Oklahoma rift zone

Cambrian igneous rocks present throughout southern Oklahoma and adjacent parts of Texas are grouped together as the Wichita province (Ham et al., 1964). The best exposures of these rocks occur in the Wichita Mountains (Fig. 2), and much of our understanding of the early phases of rift evolution is based on studies of those outcrops. Schematic relations of the main igneous units exposed in the Wichita Mountains are illustrated in Figure 3. More limited exposures in the Arbuckle Mountains consist dominantly of rhyolite.

Raggedy Mountain Gabbro Group. Abundant gabbroic rocks in the Wichita province have been assigned to the Raggedy Mountain Gabbro Group by Ham et al. (1964). The oldest examples of these rocks that crop out in the Wichita Mountains belong to the Glen Mountains Layered Complex. A composite vertical sequence ~1.l km thick of parts of the layered complex is exposed at the surface (Powell, 1986; Powell et al., 1980), but subsurface data show the complex to be more extensive (Ham et al., 1964). The exposed rocks consist of anorthosite, anorthositic gabbro, troctolite and olivine gabbro showing typical rhythmic and cryptic layering. These rocks represent the middle parts of a standard tholeiitic layered complex and require the presence of ultramafic rocks at deeper levels (Powell and Phelps, 1977). Prior to emplacement of younger igneous rocks in the province, the layered rocks were tilted gently to the north, and Powell and Phelps (1977) have estimated that 2-4 km of overlying, more differentiated parts of the complex were removed by erosion, as well as an unknown thickness of roof rock. The uplift has been attributed to block rotation during normal faulting within the developing rift zone (McConnell and Gilbert, 1990). Following this tilting and erosion, a series of smaller intrusions of biotite-bearing tholeiitic gabbro (Roosevelt Gabbros) were intruded into the layered complex and are petrogenetically unrelated to it (Powell et al., 1980; Price et al., 1998a). Subsurface data indicate that similar biotite-bearing gabbros are widespread in the subsurface in western Oklahoma and adjacent parts of the Texas panhandle. Also present in the subsurface in that region are fairly extensive amounts of hornblende diorite and quartz diorite (Ham et al., 1964). The dioritic rocks have not been studied in detail, and their petrogenetic relations to other parts of the Wichita province are unclear (Powell et al., 1980).

Navajoe Mountain Basalt-Spilite Group. This term was introduced by Ham et al. (1964) for a sequence of variably altered basaltic to andesitic volcanic rocks penetrated by eleven basement wells in the western part of the Wichita Mountains province; the thickest penetration was 320 m, and the base of the volcanic rocks was not reached. The exact relations of this volcanic sequence to the Glen Mountains Layered Complex are unknown. Most of the rocks are lavas, but palagonite tuffs were penetrated in one well, suggesting that at least some of the volcanism involved explosive phreatomagmatic eruptions in subaqueous settings or in terrestrial environments where rising magma came in contact with groundwater-rich zones beneath the surface. Chemical studies by Shapiro (1981) and Aquilar (1988) indicate tholeiitic compositions for the lavas. Whole-rock chemical analyses and optical determinations of plagioclase compositions indicate that andesitic rocks are present as well as basalts within parts of the succession (Ham et al., 1964; Aquilar, 1988), and it is most likely that these intermediate-composition lavas were derived by fractionation of basaltic parental magmas.

Carlton Rhyolite and Wichita Granite Groups. The uppermost major igneous unit within the Southern Oklahoma rift zone is the Carlton Rhyolite Group, which can be traced for at least 40,000 km² in the subsurface (Fig. 1; Ham et al., 1964). The rhyolites overlie the Navajoe Mountain Basalt-Spilite Group in the subsurface in the western part of the province, and regional relations indicate they originally covered the erosional surface developed on top of the Glen Mountains Layered Complex. The greatest known thickness of rhyolite occurs at Bally Mountain in the Wichitas (Fig. 2), where a number of different rhyolite flows form a sequence ~2 km thick (Pollard and Hanson, 2000); this is a minimum thickness because the base of the sequence is truncated by a Pennsylvanian fault. The maximum subsurface rhyolite thickness penetrated by drilling is 1.4 km in the Arbuckle Mountains region. Stratigraphic relations in that area demonstrate that the rhyolites were ponded against a major rift fault (Washita Valley fault) along the northern margin of the rift zone (Ham et al., 1964); this fault was reactivated as a compressional to transpressional structure during Pennsylvanian inversion of the rift zone. Elsewhere, subsurface evidence suggests that the rhyolites poured across the rift-bounding faults and extended farther north (Keller and Stephenson, 2007). There is no evidence for the presence of calderas within the Southern Oklahoma rift zone, and it is likely that the rhyolites issued from fissure-type vents (Hogan and Gilbert, 1998).

Major components of the Wichita Granite Group are a series of sheet granites injected along the contact between the layered complex and the overlying rhyolites (Figs. 3 and 4). The most extensive of these sheet granites is the Mount Scott Granite (Fig. 4), which forms a sill that can be traced a minimum of 55 km laterally but is only 0.5 km thick (Hogan and Gilbert, 1995). Granite sills also occur between the rhyolites and the underlying basalts in the subsurface (Ham et al., 1964). Both the rhyolites and granites have typical features of A-type felsic rocks (Myers et al., 1981; Hogan and Gilbert, 1995, 1997). They plot uniformly in fields for within-plate, A-type granites on standard discrimination diagrams using elements resistant to secondary alteration (Fig. 5). High F contents in magmatic amphiboles in the granites and experimentally determined phase relations indicate the felsic magmas had high F contents (Hogan and Gilbert, 1995; Price et al., 1999), and Zr geothermometry on the granites and rhyolites yields temperatures up to ~950° C (Hogan and Gilbert, 1997). The relatively low viscosity of these A-type felsic magmas presumably explains their ability to form extensive granite sills that are surprisingly thin relative to their lateral extents.

Diabase dikes. The youngest igneous rocks in the Southern Oklahoma rift zone consist of a large number of diabase dikes and sills or transgressive sheets that cut all the other igneous units and are exposed both in the Wichita and Arbuckle Mountains. These so-called “late diabases” have tholeiitic to transitional affinities (Cameron et al., 1986; Gilbert and Hughes, 1986; DeGroat et al., 1995). They show variable trends but generally strike northwest, parallel to the trend of the rift (Ham et al., 1964; Denison, 1995).

Similar tholeiitic diabase dikes, typically with northwest trends, intrude the 1.4 Ga basement rocks exposed in the Arbuckle Mountains on the north side of the rift zone, and they increase in abundance as the margin of the rift is approached (Denison, 1995; Lidiak et al., 2005). The dikes tend to be poorly exposed at the surface, but quarries in the Proterozoic basement near the rift margin reveal the dikes to occupy up to ~30% of the rock volume (Fig. 6). These dikes may well have fed flood basalts overlying the basement rocks north of the rift. If so, the basalts were entirely removed by erosion prior to deposition of transgressive Upper Cambrian strata. Denison (1995) has suggested that this episode of dike injection preceded emplacement of the felsic rocks inside the rift zone.

Temporal relations and magma volumes and sources. Considerable progress has been made in developing good isotopic age constraints for the magmatic evolution of the Southern Oklahoma rift zone (Fig. 3), although some units remain poorly dated, and some data are available only in abstract form. There are currently no robust age constraints on the Navajoe Mountain Basalt-Spilite Group. Lambert et al. (1988) obtained an internal Sm-Nd whole-rock–mineral isochron date of 528 ± 29 Ma for the Glen Mountains Layered Complex, the oldest igneous unit exposed at the surface within the rift zone. The Wichita granites and a rhyolite xenolith within the granites have yielded U-Pb zircon ages of 533 ± 1 to 530 ± 1 Ma (Wright et al., 1996). Hanson et al. (2009) reported zircon ages of ~532 Ma for two flows from the base and top of the rhyolite sequence at Bally Mountain, and rhyolites exposed in the Arbuckle Mountains have yielded zircon ages of 539 ± 5 and 536 ± 5 Ma (Thomas et al., 2000). As discussed above, the Glen Mountains Layered Complex underwent a significant episode of uplift and erosion prior to emplacement of the felsic rocks. A more precise age for the Glen Mountains Layered Complex is clearly needed, but the present data are consistent with emplacement of the igneous units in a relatively narrow time frame within a dynamically evolving rift setting.

One of the Roosevelt Gabbros intrusive into the Glen Mountains Layered Complex has yielded ⁴⁰Ar/³⁹Ar hornblende and biotite dates of 533 ± 2 and 533 ± 4, respectively (Hames et al., 1998). Bowring and Hoppe (1982) obtained a U-Pb zircon age of 552 ± 7 Ma for the same body of Roosevelt Gabbro dated by Hames et al. (1998). This result appears to conflict with field evidence that emplacement of the Roosevelt Gabbros was penecontemporaneous with granite intrusion, raising the possibility that the U-Pb isotopic data may be influenced by an inherited component (Hogan and Gilbert, 1998). The diabase dikes in the rift zone have not yielded reliable geochronological results, but they clearly predate deposition of Upper Cambrian sediments across the igneous rocks within the rift zone, and field evidence indicates that their emplacement overlapped in time with the felsic magmatism (Hogan and Gilbert, 1998).

As noted above, the Carlton Rhyolite Group has a subsurface extent of at least ~40,000 km². Assuming an average thickness of one kilometer, which is a conservative estimate, the total volume of magma erupted to form the rhyolitic volcanic field is a minimum of 40,000 km³. The volume of felsic magma emplaced as granite intrusions in the same time frame is more difficult to estimate but is also significant. Recent geophysical studies indicate the presence of massive amounts of mafic rock at deeper levels along the entire axis of the rift zone in Oklahoma and the Texas Panhandle, extending to depths > 10 km (Keller and Stephenson, 2007), and the total volume of mafic magma emplaced during rifting is probably at least an order of magnitude greater than the volume of felsic magma. The most likely explanation for the anomalous production of magma in this region is that it records the impact of a mantle plume during the rifting that led to Cambrian opening of the Iapetus Ocean to the southeast. Better age constraints, however, are needed for parts of the igneous province before reasonable estimates on magma production rates are possible.

Prior to the availability of isotopic constraints on source regions for the magmas, it was commonly thought that the voluminous felsic rocks in the province were generated from partial melting of older crust at depth (e.g., Gilbert and Denison, 1993). However, the felsic rocks have yielded low initial ⁸⁷Sr/⁸⁶Sr ratios and positive ε_Nd values that overlap with values for the late diabases (Hogan et al., 1995), suggesting that the felsic magmas were derived either by differentiation of basaltic parental magmas or by partial melting of a mafic underplate. Trace-element patterns for both the granites and rhyolites point to derivation from ocean-island-basalt-type precursors (Hogan et al., 1995; McCleery and Hanson, 2010a).

Work in progress on the volcanic fill of the Southern Oklahoma rift zone

New subsurface data for basalts in the rift zone. Basement wells drilled since the work of Ham et al. (1964) provide the opportunity to better constrain the distribution of igneous units within the Southern Oklahoma rift zone. To this end, one of us (Puckett) has initiated a study of 41 wells drilled into the volcanic fill of the rift zone in the area within and to the northwest of the Arbuckle Mountains (Puckett, 2011; Puckett et al., 2011). Complimenting this work, Brueseke and students have initiated geochemical and isotopic studies of mafic to intermediate units from the wells, to better understand their tectonic setting and petrogenesis (Puckett et al., 2011; unpublished data). The wells span an along-strike distance of 42 km south of and parallel to the Washita Valley fault that forms the northern margin of the rift zone in this region. Ham et al. (1964) previously documented the extensive amount of Carlton Rhyolite present in the subsurface in the area. A surprising result of the new work has been the recognition of extensive amounts of mafic to intermediate lava flows which both underlie and are intercalated with the rhyolites (Fig. 7).

Thicknesses of volcanic rocks penetrated in the wells range from 475 m to 4.3 km. These wells present a cumulative total in excess of 37 km of non-correlated igneous section available for examination and reveal a much more complete profile of the volcanic section in the Arbuckles than is available on outcrop, where only rhyolites are exposed. Representative lithological logs from some of the wells are shown in Figure 7. These logs were compiled by examination of drill cutting samples taken at 10-foot intervals and correlated with gamma-ray logs, which readily distinguish between the mafic and felsic rock types. Flow boundaries of thick extrusive units, buried weathering surfaces, and important thrust faults have also been identified. The deepest well in Figure 7 represents the thickest penetration of igneous rocks in Oklahoma (Puckett, 2011). None of the wells in the study area penetrated completely through the volcanic section; in most wells the base of the section is truncated by the Washita Valley fault. One well (labeled Sec. 19-T1N-R2W in Fig. 7) contains a thick clastic section of arkose and arkosic sandstone interpreted as a fan delta sequence derived from Proterozoic basement rocks exposed outside the rift zone. The clastic interval also contains thin rhyolitic volcaniclastic units and was subsequently covered by later volcanic eruptions.

The simplest interpretation is that the mafic to intermediate lavas documented in this recent work are correlative with the Navajoe Mountain Basalt-Spilite Group, the main subcrop of which occurs ~120 km to the northwest within the rift zone (Ham et al., 1964). The new data from the Arbuckle region suggest that this volcanic package is probably an important constituent of the rift fill throughout its extent. Possibly a significant proportion of the mafic rocks detected geophysically at deeper levels in the rift (Keller and Stephenson, 2007) also consist of basalt, with some more evolved lavas, rather than plutonic rocks.

In the wells in the Arbuckle region, phenocrysts in the basalts are dominantly plagioclase, with less common olivine (?) replaced by Fe-rich clay minerals. Many cuttings exhibit intergranular groundmass textures (Fig. 8a), consisting of randomly arranged plagioclase microlites and interstitial mafic grains, including clinopyroxene and magnetite. A smaller number of cuttings have a groundmass consisting of dark tachylitic glass (Fig. 8b). These textural variations suggest the wells penetrated a series of lava flows, with intergranular textures representing slowly cooled flow interiors and tachylite being derived from more rapidly chilled flow margins. Some intervals also contain vesicular, ash- and lapilli-sized particles of altered sideromelane glass (Fig. 8c), indicating that explosive hydrovolcanic eruptions played a role in the evolution of parts of the volcanic sequence.

Recent work by Brueseke, his student Casey Bulen, and Mertzman yields new major and trace element whole-rock geochemical results from three of the wells. Sampled drill cuttings were examined via binocular microscope, hand-picked to eliminate any potential contaminants (e.g. zeolites, carbonates, rhyolite and/or granitoid fragments, etc.), powdered, and analyzed by XRF at Franklin and Marshall College, following Mertzman, 2000 (http://www.fandm.edu/earth-and-environment/x-ray-laboratory). This work is in progress and a detailed discussion of these data is beyond the scope of this contribution. However, first order observations shed light on nomenclature, geochemical diversity, tectonic implications, and potential mantle sources of these lava flows. Figure 9 illustrates the general nomenclature of these samples and indicates that they range from basalt to andesite and are dominantly subalkaline to mildy alkaline [e.g., they straddle the subalkaline-alkaline boundary on the Zr/TiO₂ vs. Nb/Y diagram of Winchester and Floyd (1977) and the total alkalis vs. silica diagram of Le Bas et al. (1986)]. That the samples yield the same nomenclature in the classifications of Winchester and Floyd (1977) and Le Bas et al. (1986), coupled with acceptable analytical totals (Brueseke, unpublished), points to the robustness of the major and trace element results. CIPW norm calculations (assuming 70% of Fe is ferrous) for two of the samples yield olivine tholeiite to quartz tholeiite compositions.

Figure 10 is the tectonic discrimination diagram of Meschede (1986). All drill cutting samples plot within a region that straddles the boundary between intraplate tholeiites and E-MORB. Figure 11 illustrates that the drill cuttings have Zr/Nb values of ~7-11 (avg. = 8.7), similar to EMI OIB (Weaver, 1991). Other incompatible trace element ratios show similar relationships for the drill cuttings (e.g. OIB-like Ba/Nb, Nb/Y, etc.). The bulk chemical trends suggest that the more evolved samples are related to the least evolved samples, via fractional crystallization (± assimilation of crust). More work, including our ongoing work and radiogenic isotope data, is needed to fully interpret these data. In summary, the wells penetrate a volumetrically significant package of mafic to intermediate rocks that have OIB-like, tholeiitic to slightly alkaline chemistries, similar to flood basalt packages exposed in other LIPs.

Ongoing work on the Carlton Rhyolite Group. Work on the physical volcanology of the Carlton Rhyolite in the Wichita Mountains has been carried out by Hanson and a series of his students in the Zodletone Mountain, Bally Mountain, Blue Creek Canyon, and Fort Sill areas (Fig. 2; Bigger and Hanson, 1992; Hanson et al., 2009, 2011). More recently, another of Hanson’s students, Amy Eschberger, has begun similar studies of the rhyolites exposed in the Arbuckle Mountains. The rhyolite outcrops are generally fairly limited, but they give us an intriguing window into the extensive, mostly buried Cambrian rhyolitic volcanic field forming the upper part of the igneous fill within the Southern Oklahoma rift zone. We have mapped nearly 30 individual flows in the Wichitas, none of which can be correlated between the four main outcrop areas shown in Figure 2 because of structural complications. Individual flow units are up to 400 m thick and form tabular bodies stacked on top of each other or separated by generally thin rhyolitic volcaniclastic deposits. On outcrop most of these flows can only be traced a few kilometers before going into cover or being faulted out.

Rhyolite flows in the Wichita Mountains generally exhibit a standard vertical zonation (Fig. 12; Pollard and Hanson, 2000; Philips, 2002; Burkholder and Hanson, 2006). Originally glassy zones, now devitrified and altered, define the upper and lower margins to the flows and show relict perlitic texture, as well as well-developed flow banding extending down to the thin-section scale (Fig. 13a, b). Peperite is developed at the base of many of the flows, where lava flowed over, quenched against, and mixed with unlithified, wet tuffaceous sediment (Fig. 13c). The inner parts of the glassy margins in many cases contain distinctive zones rich in lithophysae (Fig. 13d). Flow interiors consist of thick, monotonous, holocrystalline zones that lack flow banding and in thin section are seen to contain quartz paramorphs after tridymite crystals (Fig. 14). The tridymite needles increase in size towards flow centers, indicating that these flows were emplaced as single thick cooling units. Very similar textures shown by tridymite needles in the groundmass of other A-type felsic units have been described by Twist and French (1983) and Trendall (1995).

Textural zonations within the Carlton Rhyolite flows are consistent with the interpretation that the generally limited outcrops present are remnants of more extensive flows similar to those documented from many other, better exposed A-type volcanic provinces (Henry and Wolff, 1992). Injection of some of the Wichita granites as thin, laterally persistent sills (Fig. 4) certainly suggests that the extrusive counterparts of these granites could have flowed for considerable distances on the surface. The most extensive rhyolite flow we have mapped in the Wichita Mountains occurs in the Fort Sill area (Fig. 15). This flow is > 190 m thick and can be traced ~19 km before going under cover or being truncated by intrusive granite (Finegan and Hanson, 2006); the original extent of the flow may have been considerably greater.

Laterally extensive flow units of this type in other A-type felsic provinces are interpreted in some cases to be true lavas, sometimes referred to as flood rhyolites (Henry and Wolff, 1992), whereas others are believed to represent explosively generated pyroclastic flows that became nearly completely homogenized during or after emplacement to form lava-like rheoignimbrites (Branney and Kokelaar, 1992). In the Carlton Rhyolite, pyroclastic-like textures are absent in most of the flows. We conclude that, if any of the Carlton Rhyolite flows represent rheoignimbrites, they must have traveled for some distance in a lava-like state before coming to rest.

Drill cuttings of the rhyolites penetrated by the wells in the Arbuckle region being studied by Puckett exhibit a comparable range of groundmass textures to that seen in Carlton Rhyolite exposed in the Wichita Mountains. In some intervals, the cuttings show relict perlitic texture (Fig. 16a), recording hydration of rhyolite glass; we interpret these cuttings to be derived from chilled, glassy margins to the lava flows. In many intervals, the cuttings contain randomly oriented tridymite crystals (now inverted to quartz), which vary significantly in size between different cuttings (Figs. 16b and 16c). By analogy with rhyolite outcrops in the Wichita Mountains, we interpret these cuttings to record cooling gradients within thick flows after the lava had come to rest. Pyroclastic textures, including unwelded pumice fragments and tricuspate shards, generally occur in a small proportion of the cuttings, which we interpret to be derived from relatively thin pyroclastic interbeds between flows.

So far, we have found no convincing evidence for the presence of welded ignimbrites within the rhyolite succession penetrated by the Arbuckle wells. In most of the wells examined to date, the rhyolites appear to consist dominantly of lava flows ranging from tens of meters to possibly as much as several hundred meters thick. These data suggest that flow units comparable to those documented in the Wichita Mountains are probably a standard feature of the Cambrian rhyolitic volcanic field within the Southern Oklahoma rift zone.

Geochemical studies on the Carlton Rhyolite are also in progress (McCleery and Hanson, 2010a, b), but a detailed discussion of these data is beyond the scope of the present report. Representative data from individual flows within the different rhyolite outcrops in the Wichita Mountains are shown on Harker diagrams in Figure 17. Scatter in the data is partly due to secondary alteration, but overall the rhyolites show a relatively restricted compositional range that overlaps with available data for the coeval Wichita granites. Note that the laterally extensive Fort Sill rhyolite flow shows significant within-flow compositional differences, which is also true of a few other flows shown in Figure 17, suggesting that these flows tapped heterogeneous magma batches during eruption.

Evidence that crystal-liquid fractionation explains some of the compositional variations in the rhyolites is shown in Figures 18a-c, where the trends are qualitatively consistent with fractionation of clinopyroxene, plagioclase and alkali feldspar. However, note that the main Fort Sill rhyolite defines a separate trend from the other rhyolites, raising the possibility that some of the rhyolite flows were derived from separate magma reservoirs. This concept is tested further in Fig. 19, which shows several plots of elements that are resistant to secondary alteration. These diagrams show that, except for a few anomalous points, samples from the main Fort Sill rhyolite define a distinct group (Group 1) that is separate from the other rhyolite flows. The other rhyolites also consistently fall into two distinct groups (Groups 2 and 3). We infer that these groups record derivation of the rhyolite magmas from three distinct sources or magma reservoirs. An interesting point is that another rhyolite flow in the Fort Sill area, which occurs beneath the main Fort Sill flow, falls in Group 2 on the diagrams, along with three units in the Blue Creek Canyon area and one flow at Bally Mountain. The rest of the Bally Mountain flows plot with three of the Blue Creek Canyon flows and the Zodletone Mountain flows to define Group 3. These data indicate that flows exposed within the same area in some cases tapped different magma reservoirs. This may imply a complex subsurface magma plumbing system, or may mean that laterally extensive flows derived from different source vents and magma chambers came to rest on top of each other. These data obviously pertain only to a limited area in the Wichita Mountains, and we are now extending these geochemical studies to other parts of the rift zone to see if similar compositional groups can be recognized in the rhyolites elsewhere in the rift.

References