Arc-to-craton: Devonian air-fall tephras in the eastern United States

Air-fall volcanic tephra beds in sedimentary successions are key sources of varied geological information. A product of explosive silicic volcanism, tephra is forcefully ejected from an underlying magma chamber, thrust upward into the troposphere or even the stratosphere, and transported, settling across all environments downwind of the source volcano. Dependent on several factors, some tephras are deposited across a significant geographic area. When preserved in strata, they provide key, distinct marker beds for local to long-distance correlation, and their radiometric dates allow for refinement of the geologic time scale. In addition, they comprise a key source of information for reconstructing a history of paleovolcanism, which is especially crucial adjacent to deeply eroded magmatic/volcanic arcs.

Following the settling and deposition of tephra materials (extruded ash, pumice, phenocrysts, and rock fragments) onto land or through water to the sediment floor, a primary tephra layer is exposed to various active environmentally dependent processes (Ver Straeten, 2004a). Rapid burial under relatively low-energy and biologically inactive conditions best preserves a primary air-fall tephra. Low sedimentation rates under quiet conditions may lead to stacking of multiple eruptive event layers. In most settings, a range of physical, biological, and chemical processes act on exposed tephra sediments, mixing them with subsequently deposited tephra and/or background sediments, sometimes leading to complete mixing and obliteration of the eruptive event. Studies on the postdepositional history of tephras include Baird et al. (1994), Huff et al. (1999), Königer and Stollhofen (2001), Ver Straeten (2004a, 2008), Benedict (2004), Ver Straeten et al. (2005), and Püspöki et al. (2005, 2008).

Various descriptive names are applied to air-fall volcanic tephras in the sedimentary record and scientific literature. These sometimes imply grain size (e.g., tuff); often, they reflect the diagenetic history of a tephra (e.g., bentonite, K-bentonite, metabentonite, tonstein), or a generic, common name for such layers (e.g., ash). If shown to be of an air-fall volcanic origin, these layers are, independent of applied descriptive terms, tephras, i.e., layers of extruded igneous materials (ash, pumice, phenocrysts, and rock fragments). Following the practice of Cenozoic volcanologists, and as recommended to Ver Straeten in discussions with Andrei Sarna-Wojcicki (2007, personal commun.), the term “tephra” is utilized throughout this paper, with the other terms retained as descriptive modifiers. Images of some eastern U.S. Devonian tephra beds are shown in Figure 1.

During the Devonian, large portions of eastern Laurentia (North America) were at times flooded by shallow seas. Subduction and collisional tectonic processes resulted in uplift of an orogenic belt extending from east Greenland to the southeastern United States, with widespread deformation, metamorphism, magmatism, and flexural downwarping of a retroarc foreland basin system. The Appalachian Basin region has been variously interpreted to have been located 25°S–40°S of the equator (van der Voo, 1988; Witzke and Heckel, 1988; Scotese and McKerrow, 1990). In this setting, explosive silicic volcanism resulted in deposition of numerous air-fall volcanic tephras in Devonian strata across the region.

In this paper, we review the occurrence of known Devonian air-fall tephras from the eastern United States (Figs. 2 and 3), followed by a discussion of correlation issues of various Eifelian Stage (lower Middle Devonian) Tioga tephras, and dating of the Eifelian Stage base of Marcellus strata. In addition, reexamination of Tioga Tephra phenocryst maps led to reinterpretation of Tioga paleovolcanic sources. Finally, we present a first-time comparison of the ages of eastern U.S. Devonian tephras with recently dated Devonian igneous rocks in the Appalachians (northeastern United States). Throughout this paper, Devonian tephra dates utilized here are from U-Pb thermal ionization mass spectrometry (TIMS) analyses on zircons, which are from the literature cited herein or are more recently determined, except for a monazite date for the Tioga B Tephra from Roden et al. (1990).

During the Late Silurian to Early Carboniferous, oblique continent-continent collisions of multiple terranes resulted in uplift of an orogenic belt along the eastern margin of North America. Commonly known as the Acadian orogeny, it has recently been subdivided into separate Acadian and Neoacadian orogenies (van Staal 2007; van Staal et al., 2009). Along with widespread Acadian–Neoacadian structural and metamorphic activity, subduction-related melting beneath Laurentian crust led to extensive plutonic and volcanic activity, much of it silicic in composition. Throughout the Devonian Period, a shallow epicontinental sea covered much of eastern North America, spread across the Michigan, Illinois, and Iowa cratonic basins, as well as the larger retroarc Acadian foreland basin system. The Appalachian Basin is a preserved, nonmetamorphosed portion of the greater Acadian–Neoacadian foreland basin (Ver Straeten, 2010).

Prior to 1960, only a few air-fall tephra beds were known from Devonian sedimentary successions in the eastern United States (Fettke, 1952; Oliver, 1954, 1956). John Dennison (1960, 1961) was the first to document a greater number of Devonian tephras within the Middle Devonian Tioga Bentonite interval. His efforts, along with those of his collaborator Daniel Textoris and James Conkin, represented the next stage of documenting Devonian tephras across the region. Beginning in the mid- to late 1980s, another stage of documentation and interpretation in the eastern United States began, which continues to the present.

Over 100 Devonian-age air-fall volcanic tephras are now known from the eastern United States. They occur scattered through the succession, sometimes occurring in relatively closely spaced clusters of eight or more beds; other tephras occur as more isolated beds. Some stratigraphic intervals lack documented tephras altogether (Fig. 3).

Two clusters and several stratigraphically isolated, altered air-fall volcanic tephra beds are reported from the Lower Devonian of the Appalachian Basin. The clusters comprise the Lochkovian-age Bald Hill K-bentonites of Smith et al. (1988; dated by Tucker et al. [1998] at 417.6 ± 1.9 Ma) and the apparent lower Emsian Stage Sprout Brook K-bentonites of Ver Straeten (2004a, 2004b; dated by Tucker et al. [1998] at 408.3 ± 1.9 Ma).

The Bald Hill Tephras cluster (Bald Hill Bentonites of Smith et al., 1988) consists of up to 15 clay-rich K-bentonites distributed through lower to middle Lochkovian strata in the Appalachian Basin (Ver Straeten, 2004a). Reported from New York, Pennsylvania, Maryland, Virginia, and West Virginia, they variously occur in the Kalkberg and New Scotland formations in the north, and the Corriganville and Mandata formations to the south. Tucker et al. (1998) reported a zircon TIMS age of 417.6 ± 1.0 Ma for the original tephra layer described from the Bald Hill cluster, an approximately 8-cm-thick bed from Cherry Valley, New York. Key references for the Bald Hill Tephras include Smith et al. (1988, 2003), Shaw et al. (1991), Metcalf (1993), Hanson (1995), Ver Straeten (2004a), and Benedict (2004).

The younger Sprout Brook Tephras cluster has only been documented in the northeastern part of the Appalachian Basin, in eastern New York (Ver Straeten, 2004a, 2004b). Up to 15 K-bentonite layers occur interbedded with cherts, siliceous siltstones, and shales in lower strata of the Esopus Formation (lower part of the Spawn Hollow Member). Tucker et al. (1998) reported a mean TIMS age of 408.3 ± 1.9 Ma, based on analyses of zircons from two altered tephras of the Sprout Brook cluster. These lower Esopus tephras overlie sandstones of the Oriskany Formation or correlative limestones of the Glenerie Formation. The base of the Emsian Stage has been tentatively placed at the base of the Esopus Formation in New York (Rickard, 1975); however, no detailed biostratigraphic data are available to confirm that.

The Sprout Brook Tephras cluster has been documented across eastern New York from south of Catskill to at least Cherry Valley, eastern New York, over a distance of ~125 km. The Esopus Formation pinches out at a disconformity 25 km west of Cherry Valley. South of the Catskill area in New York, the Sprout Brook interval is not well documented. The tephra beds were not recognized in a western New Jersey outcrop or westward into central Pennsylvania. Several Sprout Brook beds were found to geochemically correlate between outcrops in the Hudson Valley, but not to the Cherry Valley section to the west (Hanson, 1995; Ver Straeten et al., 2005).

Conkin and Conkin (1979, their fig. 22) reported mixed volcanic-detrital strata in the same position near the Virginia–West Virginia border at Williamsville, Virginia. A search for indicators of volcanogenic input at this same time-position across Pennsylvania to Virginia and West Virginia (Ver Straeten, 2004b) indicated mixed volcanogenic-detrital zircons at some, but not all, localities beyond eastern New York (Ver Straeten, 2004b). Key references on the Sprout Brook Tephras include Ver Straeten (1996a, 2004a, 2004b, 2010), Hanson (1995), Benedict (2004), and Ver Straeten et al. (2005).

Additional Lower Devonian tephra beds, not associated with clusters, occur in the northern Appalachian Basin, in the upper Emsian Schoharie Formation of New York (Hanson, 1995; Ver Straeten 2004a; this report). Two of these are associated with erosional and/or condensed hiatuses, with a mix of volcanogenic and detrital grains and authigenic glauconite ± phosphate. These two beds occur locally at the base and near the top of the formation (north of Cherry Valley, New York). At least four additional silty micaceous beds of apparent volcanic origin occur within the upper Emsian Schoharie Formation near Kingston, eastern New York. Forming decimeter-scale recessions, these beds consist of micaceous, finely laminated siltstones in their lower portions, capped by typical K-bentonitic–appearing, light-tan, greasy-feeling clays in their upper parts. Further investigation of these beds is anticipated.

James Hall (1843) noted an “unctuous” clay layer in upper strata of the Onondaga Limestone in New York, for which Luther (1894) first suggested a volcanic origin. Fettke (in Ebright et al., 1949) and Fettke (1952) first applied the term “Tioga Bentonite” to a biotite-rich, brown to brownish gray “bentonitic” shale in subsurface well cuttings in Pennsylvania. For several years, only one such altered tephra layer was thought to occur in upper Onondaga and time-equivalent strata (Oliver, 1954, 1956).

John Dennison was the first researcher to focus closely on the “Tioga Bentonite” interval. Initially on his own, and later in a long collaboration with Daniel Textoris, they documented multiple Tioga air-fall volcanic layers within a greater Tioga Bentonite zone, eventually shown to be up to 61 m thick (Dennison and Textoris, 1978). They also addressed the distribution and isopach thicknesses, potential volcanic sources, and even Devonian paleowind directions based on the Tioga ash-fall record (Dennison, 1960, 1961, 1986; Dennison and Textoris, 1970, 1978, 1987). Penecontemporaneous with the later work of Dennison and Textoris, microstratigraphic tephra studies by James and Barbara Conkin (Conkin and Conkin, 1979, 1984b; Conkin, 1987) contributed significantly to Tioga studies, recognizing and correlating numerous bentonites in the Tioga interval from New York cratonward into Ohio and Indiana.

Within the greater Tioga interval, Dennison and Hasson (1976) proposed a Tioga “Middle Coarse Zone” (Tioga MCZ). It occurs as “a bundle of three or four recognizable tuff beds within a span of usually two feet (0.6 m), coarser than any other portion of the Tioga tuffaceous material. This … is apparently the portion of the Tioga which forms the principal marker horizon which can be traced farthest from the volcanic source” (Dennison and Textoris, 1978, p. 167). The middle coarse zone thickens eastward to 2.4 m (Dennison and Textoris, 1978, p. 167).

Subsequent work by Smith and Way (1983) and Way et al. (1986) correlated seven Tioga Ash Beds (Beds A–G) across 280 km of the Valley and Ridge Province in central Pennsylvania. Their work was subsequently extended into New York by Ver Straeten (Brett and Ver Straeten, 1994), and then basinwide, from New York to southwest Virginia and into Ohio (Ver Straeten, 1996a, 1996b, 2004a, 2007; Fig. 4).

At Williamsville, Bath County, Virginia, Ver Straeten (2007) noted a second cluster of tephras above the Middle Coarse Zone within the greater Tioga interval, a short distance above the change to Marcellus black shales. The same cluster of beds occurs just above initial black shales and the Middle Coarse Zone at Bluefield, West Virginia, and at Wytheville, Virginia. These latter occurrences were noted by Dennison and Textoris (1978) to extend as far as Clinch Mountain in northeastern Tennessee (Dennison and Boucot, 1974). They currently are unknown north of Williamsville, Virginia, and have not been seen in southern Pennsylvania or northernmost West Virginia outcrops. Along the outcrop belt, this “basal Marcellus Tephras” cluster falls within strata assigned to the Marcellus, Millboro, Chattanooga, and Wildcat Valley formations.

Beyond the Appalachian Basin, Tioga-age beds of air-fall volcanic origin from upper Onondaga–correlative strata are reported from the Michigan and Illinois basins (Meents and Swann, 1965; Collinson, 1967; Droste and Vitaliano, 1973; Baltrusaitis, 1974; Droste and Shaver, 1975). Meents and Swann (1965) also reported a Tioga bed from a single outcrop in Iowa.

Within the greater Tioga Tephras interval of Dennison and Textoris (1978), numerous tephra beds occur below the Tioga Middle Coarse Zone, Tioga A–G Tephras, and basal Marcellus clusters in eastern U.S. sedimentary strata, chiefly in the Appalachian Basin. These include beds both below the Tioga A–G Tephras and below the Middle Coarse Zone in strata correlative with the Edgecliff, Nedrow, and lower to middle Moorehouse members of the Onondaga Formation of New York. Several of these can be correlated widely along the Appalachian Basin outcrop belt, even apparently into the Columbus Formation on the western edge of the basin (central Ohio; Conkin and Conkin, 1979, 1984b; Ver Straeten, 2007).

Key references on the Tioga Tephras in the Appalachian Basin include Dennison (1961, 1983), Dennison and Textoris (1970, 1978, 1987), Conkin and Conkin (1979, 1984b), Conkin (1987), Smith and Way (1983), Way et al. (1986), Waechter (1993), Brett and Ver Straeten (1994), Ver Straeten (1996a, 1996b, 2004a, 2007, 2010), and Shaw (2003).

Above the lowest Marcellus Tioga Tephras, two additional, younger tephras were reported from Eifelian strata by Ver Straeten (2004a, 2007). The lower of these two tephras occurs in the middle of the Union Springs Formation in New York, in the lower part of the Marcellus subgroup in New York. Termed the “mid-Union Springs tephra,” it is widely correlatable within the Appalachian Basin, including into the Delaware Limestone in central Ohio (Ver Straeten, 2007). In intermediate- to shallow-depth marine strata, the bed closely underlies a shift to shallower, relatively coarser-grained strata associated with the onset of a falling stage systems tract of sequence stratigraphy, and progradation of coarser, sometimes more calcareous strata. In basinal settings, it occurs within distal black to dark-gray shales and mudstones (Ver Straeten, 1996a, 1996b, 2004a, 2007). The second post–lowest Marcellus tephra bed occurs in two outcrops in eastern New York, where it underlies the Cherry Valley Member (as defined in New York State) of the coeval Oatka Creek–Mount Marion formations, in the middle of the Marcellus subgroup (Ver Straeten, 2004a).

Compared to the Lower and lower Middle Devonian (Eifelian) strata, tephra beds are less common in the upper Middle Devonian (Givetian) and Upper Devonian (Frasnian and Famennian) strata of the Appalachian Basin, likely due to the changes in Acadian orogenic activity and increased coarse siliciclastic sedimentation into the basin. The two best-documented, and only named units, are the Belpre Tephra in the middle Frasnian, first described from eastern Ohio in the Ohio Shale (Collins, 1979), and the Center Hill Tephra in the uppermost Frasnian, first described in central Tennessee by Hass (1948). Other Givetian and Upper Devonian tephra beds have been reported in the East Berne (Gerwitz, 2009), Ludlowville (Batt, 1996a, 1996b), Moscow (Wilcott and Over, 2005), Cashaqua, Rhinestreet, Angola, Pipe Creek, and Hanover (Over et al., 1998; Over at el., 2013), and lower Olentangy, upper Olentangy, Chattanooga, Brallier, Foreknobs, and Huron formations, correlated with the Belpre or Center Hill, or considered unnamed (Roen, 1980; Conkin, 1989; Lanik et al., 2016).

Tephra beds in the Givetian strata of the Appalachian Basin are very thin beds to laminae that are recognized by a plastic clay-rich nature that typically weather recessively with an orangish (rusty) color. These “gummy” horizons are often found at the base of organic-rich shale beds and may not represent a discrete volcanic event, but the accumulation of volcanogenic material during intervals of low siliciclastic input, or accumulation on a disconformity surface or paraconformity in the terminology of Conkin and Conkin (1984a). These thin strata may not be composed of minerals typically associated with altered volcanic ash beds—montmorillonite or mixed-layer illite-smectite clays—but are often characterized by euhedral crystals of apatite, biotite, muscovite, quartz, sanidine, and zircon, which are less susceptible to alteration or erosion and transport. Conkin and Conkin (1984b) speculated that diagenetic processes, winnowing, and transport of volcanic ash and tuffs in marine settings preferentially removed clay minerals and necessitates recognition of tephra beds by the presence of resistant phenocrysts of the pyroclastic material. The interpretation that all of the “gummy” beds are of volcanic origin is equivocal, as not all have yielded phenocrysts or clay minerals typical of volcaniclastic origin.

We have noted numerous recessive partings and gummy beds in the calcareous mudstones and shales of the Hamilton and Genesee groups, most of which have not been investigated in detail. Of note, a gummy bed in the dark shales of the East Berne Member of the Mount Marion Formation—the eastern equivalent of the Oatka Creek Formation in New York State—was composed primarily of illite, quartz, and minor pyrite and did not yield any phenocrysts (Gerwitz, 2009). Ver Straeten (2004a) reported a thin gummy bed from close above the base of the overlying Otsego Member from one locality in the Hudson Valley, eastern New York. Another such bed also occurs in the lower part of the Butternut Member of the Skaneateles Formation south of Syracuse, central New York. Batt (1996a, 1996b) reported one distinct clay-rich bed, presumably of volcanic origin, and suggested the occurrence of numerous others in the Wanakah Member of the Ludlowville Formation. The gummy bed at the base of the Windom Member in the Genesee River Valley yielded zircon phenocrysts, but their origin is uncertain, as this horizon also contains phosphate grains and rounded quartz sand, indicative of a disconformity (Wilcott and Over, 2005). Several other gummy beds in the Windom Member, when processed, did not yield any phenocrysts.

The Belpre Tephra is the lowest prominent volcaniclastic unit in the Givetian and Upper Devonian of the Appalachian Basin, originally described and named by Collins (1979) as a tephra suite in the Ohio Shale from numerous cores in eastern Ohio. This bed is distinct from the Middle Devonian Tioga cluster, with speculation that it was equivalent to the Center Hill Tephra of Hass (1948) and Conant and Swanson (1961), known from the upper Dowelltown Member of the Chattanooga Shale in central Tennessee. Subsequently, the Belpre suite was recognized in outcrop in the lower Chattanooga Shale in eastern Tennessee and western Virginia, including tephra beds initially described by Harris and Miller (1958), and at Little War Gap (identified as Tioga) by Dennison and Boucot (1974; see Filer et al., 1996). Tephra beds and gummy horizons in the Cashaqua Shale and the lower Rhinestreet Shale (West Falls Group) in western New York (Levin and Kirchgasser, 1994) have also been associated with the Belpre Tephra suite, although conodonts and goniatites in the tephra-bearing strata in New York led to this correlation being questioned (Baird et al., 2006). Lanik et al. (2016) produced stratigraphically consistent dates of 375.55 ± 0.10 Ma from “tephra 01” and 375.25 ± 0.13 Ma from “tephra 06” at Little War Gap, Hancock County, Tennessee. The later date is analytically identical to 375.14 ± 0.12 Ma from “tephra 7.67” at Eighteenmile Creek, Erie County, New York. While the younger two ages provide an apparently isochronous marker horizon for stratigraphic correlation, the conodont zonation for the two localities is disjunct, although not widely separated, where the younger date is biostratigraphically older.

A decimeter-scale, greenish-gray, micaceous claystone bed occurs in marine strata near a marine to nonmarine transition southeast of Sydney, Delaware County, New York (D. Bishuk, 2014, personal commun.; and field study). The bed weathers to a recessive sticky, greasy clay, in sharp contrast with all surrounding strata. The locality is mapped at a position interpreted to be correlative with the contact of the Cashaqua and Rhinestreet formations (Fisher et al., 1970; Rickard, 1975). This correlation is, however, not well constrained.

Gummy laminae in the upper Angola and Pipe Creek formations from New York have been described, but analysis is incomplete. The gummy tephra laminae in the middle of the Pipe Creek Formation at Eighteenmile Creek is characterized by abundant fine phyllosilicate grains (Over et al., 2013). It is unclear if this horizon corresponds to the tephra described by Kepferle (1993) from the basal Pipe Creek Shale in the subsurface of Kentucky, which was tentatively correlated with the stratigraphically higher Center Hill Tephra.

The Center Hill Tephra is characterized by a single thin bed, ~3 cm thick, in Tennessee that contains abundant biotite, as well as phenocrysts of apatite, sanidine, orthoclase, quartz, and zircons, which yielded a date of 371.91 ± 0.15 Ma using high-precision chemical abrasion (CA) TIMS U-Pb zircon methods determined at the facility at Boise State (M. Schmitz, 2014, personal commun.). The tephra was first described by Hass (1948) from outcroppings in the central uplift area of Tennessee from the upper portion of the Dowelltown Member of the Chattanooga Shale and shown to be very close to the Frasnian-Famennian boundary (Over, 2007). The Center Hill Tephra thickens slightly eastward (Conant and Swanson, 1961) and has been tentatively identified within the Pound Sandstone Member of the Foreknobs Formation in Montgomery County, Virginia (Brame, 1999, personal commun.). The bed thickness has suggested a distant source, and prevailing winds would have been from the southeast. In the upper Hanover Shale in New York State, several gummy beds are present in the same stratigraphic interval near the Frasnian-Famennian boundary (Over et al., 1998), notably at the base of the Point Gratiot Bed, which is just below the Frasnian-Famennian boundary, and at several horizons lower and higher in the strata near the boundary.

Within the long Famennian succession of the northern Appalachian Basin, no purported tephras have been described in the literature, but little apparent concerted effort has been made to look for such layers. It is not surprising that tephras have not been reported from the New York and Pennsylvania region, given the low number and generally low resolution of reconnaissance stratigraphic studies in this area. With the exception of slope-to-basin deposits of the Canadaway Group in western New York, northwest Pennsylvania, and northeast Ohio, as well as parts of the outer shelf–slope deposits of the Chadakoin Formation in northwest Pennsylvania and Ohio, the Famennian succession is dominated by closely spaced, coarse, tempestitic siltstone and sandstone beds as well as extensive tracts of nonmarine facies, which render any search for tephras all the more difficult (Caster, 1934; Tesmer, 1963; Baird et al., 2013a, 2013b). More promising strata for tephra discoveries are coeval lower slope/ramp and basin facies in Ohio and Kentucky, which are dominated by finer-grained, offshore marine deposits.

The most promising unit to examine for tephras is the Cleveland Member of the Ohio Shale Formation, which occurs immediately below the level of greatest seismic disturbances, and which includes strata within and above the Palmatolepis gracilis expansa conodont zone and the VH-LN miospore biozones of the late Famennian (Streel et al., 1987; Zagger, 1993). This black shale unit records largely quiet depositional conditions in an epicontinental basin with preservation of fine laminations at many levels. Moreover, the Cleveland Member is particularly distinctive for the occurrence of calcareous cone-in-cone concretions, which occur at multiple levels within the unit. Although cone-in-cone concretion formation is probably unrelated to tephra deposition, per se, it is possible that crystal growth within these tabular to lenticular features may occasionally nucleate off of thin, preexisting tephra layers that may be otherwise cryptic within shale. Thin coarse beds of sand-size sediment that have served as planar nucleation surfaces within, or flooring, Pennsylvanian-age cone-in-cone concretions have been observed; such occurrences have a curious resemblance to bands of prismatic calcite flooring tephra layers in the Ordovician Utica Shale, which are a product of structural displacement along the plane of the tephra band (Ver Straeten et al., 2012; Zambito and Baird, 2006; Zambito et al., 2005). Acid dissolution of certain cone-in-cone beds from the Cleveland Member and from underlying Canadaway, Conneaut, and Conewango group divisions may prove useful for finding levels of concentrated pyroclastic constituents in the long Famennian sedimentary succession.

In contrast with Tioga interval tephras in the northern to central Appalachian basin, Ver Straeten (2004a) distinguished two clusters bounding the contact of the Needmore-Marcellus and correlative strata along the central to southern Virginia–West Virginia border area. The lower cluster, capped by the Tioga Middle Coarse Zone (Dennison, 1983, 1986; Dennison and Textoris, 1970, 1978), occurs in calcareous strata of the upper part of the Needmore Formation and in correlative cherty strata of the upper, post–Bob’s Ridge Sandstone part of the Huntersville Formation, northeast to southwest along the outcrop belt (Figs. 3 and 4; Wytheville, Virginia, section in figs. 3, 8, and 12 of Ver Straeten, 2007). The upper cluster occurs in lowermost black to dark shales variously assigned to the overlying Marcellus, Millboro, Chattanooga, and Wildcat Valley Formations between Bath County, Virginia, and Mercer County, West Virginia, to Wythe County, Virginia. Unpublished Tioga cross sections by Dennison, given to Ver Straeten in 2001, show that Dennison documented beds of the upper cluster in some outcrops in Virginia and West Virginia, but not in others. Wherever he found beds of the upper cluster (e.g., Williamsville and Wytheville, Virginia, and other sites), he included them in his “Tioga Interval.”

Ver Straeten (2004a, 2007) correlated the upper cluster with the Tioga A–G zone of Pennsylvania and New York (Fig. 4), based on the following: (1) at study sites such as Williamsville and Wytheville, Virginia, and Bluefield, West Virginia, the number, thickness, and spacing of the individual beds in the upper cluster mimicked that of the beds of the Tioga A–G zone in Pennsylvania and New York (Wytheville, Virginia, section in Fig. 3 versus Fig. 4; overall character at Williamsville, Virginia in Fig. 5); (2) the closely spaced beds of the Tioga Middle Coarse Zone did not match the noted geometry of the Tioga A–G zone to the north (Wytheville, Virginia, section; Figs. 3 and 4); and (3) Tucker et al. (1998) documented a date of 391.4 ± 1.8 Ma for a sample in the Tioga Middle Coarse Zone from Wytheville, Virginia; this latter point is, however, a weak line of evidence, as Roden et al.’s (1990) date of 390 ± 0.5 Ma for the Tioga B was within the range of error in Tucker et al.’s date. Overall, these data supported a possible interpretation of the middle coarse zone being older than the Tioga A–G zone. The fact that the upper cluster occurred in black shale along the Virginia–West Virginia border did not seem implausible; in a basinward outcrop of the Needmore Formation near Frankstown, Pennsylvania, the Tioga A–G zone clearly occurs in black shale well above any carbonate beds, in upper Selinsgrove Member–/Onondaga Formation–equivalent black shale facies. Finally, no similar consistent series of tephra beds that appear to mimic the patterns of the Tioga A–G Tephras of Pennsylvania and New York appeared to occur along the Virginia–West Virginia outcrop belt.

Nevertheless, additional work through Eifelian strata from central Pennsylvania to southwestern Virginia and southeastern West Virginia by Ver Straeten, beginning in 2007, raised questions about his 2004a correlations. Despite the lines of evidence laid out above, it now also seems plausible that the Tioga A–G zone could correlate with the Tioga Middle Coarse Zone and a few tephras closely below it (see Fig. 5). This would agree with previous interpretations by Dennison (unpublished cross sections from 1983), Dennison and Textoris (1970, 1978), and Conkin and Conkin (1979), i.e., that the Tioga Middle Coarse Zone is correlative with the Tioga F of Smith and Way (1983), which occurs at the top of the Onondaga Limestone in central to western New York. This is the same bed termed the “Tioga Bentonite (restricted)” of Conkin and Conkin (1979). In some sections along the Virginia–West Virginia borderlands, a second, lower and relatively thick, clay-rich tephra occurs a short distance below the Tioga Middle Coarse Zone; this may represent the Tioga B bed (e.g., Williamsville section in Fig. 5). The apparent absence of a Tioga B correlative, along with the other, thinner Tioga A–B beds at some localities, may be related to local faulting out of these beds, or the lateral tectonic squeezing out of less resistant, more ductile, clay-rich altered tephra material. That may not explain its absence at all sites, however. Possible A, B, C, D, E, and G correlatives may occur at Williamsville, Virginia, bounding the Tioga Middle Coarse Zone (Fig. 5). A single thick tephra, apparently the Tioga F/middle coarse zone is underlain by deformed black shales at Hayfield, Virginia (Fig. 5), pointing to possible absence of the Tioga A through F beds related to structural causes. At the Wytheville section in southwest Virginia (Fig. 5), no deformation was noted between the Bobs Ridge Sandstone and overlying Tioga Middle Coarse Zone. A centimeter-scale measured section dug out through deeply weathered strata in that interval appeared to indicate an absence of tephras. If the Bobs Ridge Sandstone is the correlative of the Edgecliff Member of the Onondaga Limestone, as projected by Ver Straeten (2007), Onondaga-age strata would be very highly condensed at the top of the Huntersville Chert at Wytheville. This is clearly seen in the thinning of the post–Bobs Ridge Sandstone strata between there and Frost, West Virginia (Fig. 4, “Fr” on map). Under such highly condensed sedimentation, the seeming absence of the Tioga A–F beds may be because they are stacked up in a continuous tephra succession including the middle coarse zone, owing to a lack of background sedimentation in this portion of the basin at this time. This would be analogous to findings of Püspöki et al. (2005, 2008) in the Miocene of Hungary, where the only sediment deposited through multiple small-scale (Milankovitch-band) cycles was air-fall volcanic tephra.

In summary, the Tioga A–G zone of Smith and Way (1983) of Pennsylvania and New York (Way et al., 1986; Brett and Ver Straeten, 1994; Ver Straeten 2004a, 2007) may now appear to correlate with the Tioga Middle Coarse Zone and a set of underlying tephras in the southern Appalachian Basin (Fig. 5). This would agree with older correlations of the Tioga Middle Coarse Zone with the Tioga F bed to the north. However, Ver Straeten’s (2004a) correlation of the A–G zone with a basal Marcellus Tephras cluster (Fig. 4) is not fully discounted. It is suggested that geochemical fingerprinting of phenocrysts from key Tioga beds, initiated by Waechter (1993), Shaw (2003), and Benedict (2004), and applied to other non-Devonian tephras (e.g., Samson et al., 1988; Carey et al., 2009; Sell et al., 2015), should be applied to test these Tioga Tephras correlations.

Recent studies by Hayward (2012) and Parrish (2013) have dated air-fall tephras in and around the Tioga tephras interval from cores in Pennsylvania and West Virginia. They interpreted these tephras to fall within the lower part of Marcellus-age black shales, correlative with the basal Marcellus Tephras cluster discussed above. These dates for supposed basal Marcellus strata, obtained using the less accurate sensitive high-resolution ion microprobe (SHRIMP) method, range in age from 403.8 ± 4 Ma to 380.9 ± 2 Ma from Hayward (2012), and from 394 ± 5 Ma to 389 ± 3 Ma from Parrish (2013).

On first reading, these results are anomalous, with most lying outside of radiometric ages for their interpreted strata (slightly younger than 391 ± 1.8 Ma—Tucker et al., 1998; 390 ± 0.5 Ma—Roden et al., 1990), and largely outside of the very well-constrained relative age of lower Marcellus strata basinwide (uppermost part of the P. costatus costatus conodont zone, middle Eifelian Stage). Biostratigraphic data show that the Marcellus Shale in the type area of New York (Danielsen et al., 2016) is late Eifelian to early Givetian in age. Using the time scale from Gradstein et al. (2012), the base of the Marcellus, as defined in the type section in New York, is just slightly younger than 390 Ma; the top of the Marcellus should fall at ca. 388–387 Ma. The older age of 403.8 ± 4 Ma of Hayward (2012) lies within the upper Lower Devonian, within the Beaverdam Member of the Needmore Formation (lower Emsian Stage), likely correlative with the middle or upper third-order depositional sequence of the Esopus Formation of New York (Devonian Sequence Ib₂ or 1b₃, Quarry Hill or Wiltwyck Member; Ver Straeten, 2007, 2009; Becker et al., 2012). The youngest dated tephra of Hayward (2012), 380.9 ± 2 Ma, is ~10 m.y. younger than lower Marcellus strata. Comparing the SHRIMP date to the 2012 Time Scale (Becker et al., 2012), this sample is from strata correlative with lower Upper Devonian Frasnian Stage strata of conodont FZ zone 4; this would be correlative with the West River Formation at the top of the Genesee Group in New York, well above the classic Tully Limestone of New York and correlatives. Parrish’s (2013) data and interpretations are more conservative, with dates ranging from 394 ± 5 Ma to 389 ± 3 Ma. These dates roughly span the latest Emsian Stage through much of the Eifelian Stage (Becker et al., 2012), with the youngest date falling in lower Marcellus strata, as would be expected.

From dating results and stratigraphic interpretations, Hayward (2012) and Parrish (2013) interpreted the base of the Marcellus to be diachronous by as much as 23 m.y. (Hayward, 2012) or 5 m.y. (Parrish, 2013), which is not plausible. It appears that, using the SHRIMP dates from tephras deposited in highly condensed distal reaches of the basin, the authors dated air-fall tephras from not only basal Marcellus strata, but also from significantly older and younger deposits. Hayward apparently dated variously aged black shale facies from the middle Needmore Formation up to correlative strata of the lower Trimmers Rock, lower Brallier, middle Harrell, or upper Millboro formations (name dependent on geographic locality from central Pennsylvania to southwestern Virginia and adjacent West Virginia; Patchen et al., 1985). Parrish’s (2013) SHRIMP ages for lower Marcellus strata span a shorter interval of time (5 m.y.); her dates lie largely between uppermost Lower Devonian strata of the unnamed middle member of the Needmore Formation (Ver Straeten, 2007), correlative with the upper sequence of the Schoharie Formation of New York, to lower Middle Devonian basal Marcellus-age strata basinwide.

What Hayward (2012) and Parrish (2013) do show is that multiple air-fall volcanic tephras occur in basinward cores through the Emsian to Frasnian Stages, including in stratigraphic intervals where we may have no apparent record of them as yet in better-studied, shallower marine deposits. Further searches in less basinward facies across the basin may yield previously undocumented Devonian tephra beds.

In the various publications of Dennison (1960, 1961, 1986) and Dennison and Textoris (1970, 1978, 1987), they interpreted a single volcanic source for the Tioga Tephras. Initially, Dennison (1960, 1961) suggested a location slightly east of Lexington or Staunton, Virginia. In subsequent studies, the authors investigated known igneous rock bodies in the central Appalachians (e.g., felsic rocks near Monterey, western Virginia; Columbia granodiorite, central Virginia; Berea Pluton, northern Virginia; and the Petersburg granite plutons, southern Virginia; Dennison and Textoris, 1978, 1987). Subsequent radiometric dates, however, indicated that these were not plausible Tioga sources. Retaining an interpretation of a single source, Dennison and Textoris (1987) settled into an unknown locality in the vicinity of Fredericksburg, northern Virginia, buried under Paleozoic crystalline rocks in the Appalachian Mountains or Cretaceous coastal plain sediments farther east.

During their decades of research, Dennison and Textoris also built a large database of Tioga petrology and isopach thickness from outcrops and wells spread across New York, Virginia, Illinois, and Michigan. This included data on the distribution and size of various volcanogenic phenocrysts, apparently from all Tioga interval beds they sampled and analyzed. These included biotite (data from 82 localities), quartz slivers (68 localities), feldspar (59 localities), zircon (18 localities), apatite (16 localities), and pumice fragments (7 localities). Dennison and Textoris (1987) published these data as a set of palinspastic maps that cover the extent of Tioga beds across the eastern United States. The three maps constructed from the larger data sets, for maximum biotite diameter, maximum quartz sliver length, and maximum feldspar diameter, are reproduced here as Figures 6, 7, and 8.

A closer viewing of these maps indicates a more complex distribution of the Tioga Tephras than interpreted by Dennison and Textoris (1987). Four or more tongues of relatively coarser phenocryst fallout, from northeast to southwest, project rather directly outward toward the west-northwest from the Appalachians. The distribution and directional trends of the tongues cannot be easily explained by shifts in wind direction during transport of the eruption clouds.

It therefore appears that the four or more prominent tongues of phenocrysts in Figures 6–8 point back to multiple volcanic sources located in southeastern Pennsylvania, Maryland or northern Virginia, central Virginia, and northern North Carolina. The relatively straight-line projection of these different tongues argues against a single eruptive center for the Tioga Tephras.

Dennison and Textoris (1987) did not delineate the phenocryst data by individual Tioga beds. The maps lump all beds together, not allowing us to distinguish which layers or set of layers were erupted from which source area. Examination of a series of unpublished Tioga interval cross sections by John Dennison indicated that discovery and collection of individual samples from a variety of apparently discreet Tioga beds were likely inconsistent from outcrop to outcrop. It is likely tephras from throughout the greater Tioga Tephras interval, below and above the A–G/middle coarse zone, may have been included in the data set.

Nevertheless, the phenocryst maps in Figures 6–8 clearly distinguish multiple sources for the lower Middle Devonian (mid–Eifelian Stage) Tioga Tephras interval. The Tioga A–G cluster, Middle Coarse Zone, and lowest Marcellus air-fall tephras were erupted from at least four volcanic sources in the central to southern Appalachians. At this time, it is not possible to determine which tephras were erupted from which volcanic center.

Reconstruction of the history of felsic Devonian paleovolcanism during the Acadian and Neoacadian orogenies will rely on data from both the foreland and hinterland. In the hinterland, recent high-resolution geochronologic dates for an increasing number of igneous rock bodies (e.g., Appendix A2 of Bradley et al., 2015) provide one key line of evidence. In the foreland, stratigraphically constrained relative and numerical radiometric ages of air-fall tephras in sedimentary rocks (this paper) provide another key set of perspectives. There are data gaps, biases, and other issues with each database. However, their comparisons do provide insights into Devonian silicic igneous activity in the eastern United States, and possible gaps in the sedimentary tephra bed record.

In the long-eroded Acadian–Neoacadian orogenic belt/magmatic arc, most volcanic rocks are now missing. More commonly preserved silicic plutonic rocks yield ages related to cooling, coarsely approximating a broad interval of eruptive activity.

Preserved air-fall volcanic tephras in adjacent sedimentary successions provide a more detailed record of individual eruptive events. However, a range of sedimentary and paleobiological processes that act on a primary tephra layer after deposition may preserve, alter, or destroy it, leaving behind an incomplete record (e.g., Ver Straeten, 2004a).

In this section, we present an initial comparison of foreland and hinterland records of regional explosive volcanism during the Devonian Period. To this end, we plot data from this paper and from Appendix A2 of Bradley et al. (2015) in Figure 9. While using an obviously incomplete single data source from the orogen, Figure 9 outlines these initial results. In doing this, we invite hinterland researchers to flesh out and fine tune this comparison with all available, reasonably well-constrained silicic igneous rock ages from the U.S. Appalachians, and we invite Canadian researchers to pursue a similar two-pronged study and comparison along the rest of the Acadian and Neoacadian orogen and foreland.

Bradley et al.’s (2015) 101 numerically dated igneous rock bodies from the New England Appalachians and the southeast townships of Quebec are plotted in Figure 9, against air-fall tephras in the eastern United States (this paper). As previously noted, most of the air-fall layers from the foreland have not been radiometrically dated. Relative ages of varying resolution are documented for them, however, via biostratigraphic and sequence stratigraphic methods, pending more radiometric age dating.

The Bradley et al. (2015) compilation of dates provides an overview of igneous activity in New England through the Devonian. Potentially biasing issues from the data include the use of varied dating methods and the use of different minerals. However, with the probable exception of late-stage pegmatites and dates from a small number of mafic ± intermediate igneous rocks, the Bradley et al. (2015) data do provide a coarse overview of felsic igneous activity in the region. Not all rocks have been recently dated in the northeastern region, and, as discussed in the previous section, extensive felsic volcanism occurred beyond New England in the central to southern Appalachians in the United States, and in maritime Canada.

Of the 101 dated igneous rock bodies in New England compiled by Bradley et al. (2015), the vast majority (perhaps >85%) consist of granites and granodiorites, with lesser tonalites, felsic tuffs, and rhyolites. These magmas would have been capable of Plinian and similar explosive eruptions, potentially resulting in long-distance transport and deposition of tephra across a broad region. Only a minor number of the dated rocks have an intermediate to mafic composition and would be less likely to generate widespread tephra deposits. Some dated plutons have a mix of felsic and/or intermediate and/or mafic rocks, and it was not always clarified which were dated in Bradley et al.’s (2015) Appendix A2. Four pegmatites are also included in the database; these may have crystallized and isotopically closed long after the main phase of volcanic activity, and therefore their dates may not be applicable to this comparison.

A visual comparison of the two columns shows that the age of foreland tephras and hinterland igneous rocks are similar in some intervals, such as during deposition of the Lochkovian Bald Hill, approximately Pragian–Emsian Sprout Brook, Frasnian Belpre clusters of tephra beds and additional lower Eifelian and upper Frasnian tephras in the foreland (for previous approximately Pragian–Emsian foreland-hinterland comparisons, see Ver Straeten, 2004b, 2010). In other intervals, there are significant gaps in the record of foreland tephras, while igneous activity continued in New England (e.g., mid-Lochkovian, lower to middle Emsian, lower Frasnian, Famennian stages).

Some of the mismatches can be related to silicic volcanic activity beyond New England at those times. This appears to strongly apply to the Tioga and basal Marcellus Tephras clusters, which, as previously discussed, appear to have been largely sourced from the central to southern U.S. Appalachians. In other intervals, such as in the Famennian Stage, the foreland strata are less exposed and little examined, and they are geographically situated in the foreland further from volcanic sources.

A different issue arises in the lower to middle Emsian and lower Frasnian stages, where foreland basin deposits have been reasonably well studied and explored, and few tephras are documented. Hypothetical reasons for the apparent absence of foreland tephras in the northern Appalachian Basin (adjacent to New England) at these times include (1) stronger environmental biases against tephra layer preservation; (2) less comprehensive searches of some intervals within in these strata; (3) temporal change in wind directions; (4) magmatic sources of a less silicic composition; or (5) other unrecognized factors.

Out of these possible biasing factors, numbers 2 through 4 would not appear to be involved in this apparent volcanogenic “signal loss.” We have scouted strata of both the lower to upper Emsian and Frasnian stages in eastern and western New York, respectively, and beyond. As to wind direction, a comparison of the occurrences of air-fall tephras between eastern and western U.S. Devonian strata (Appalachian Basin and Nevada), including by us, shows that while a magmatic arc was located adjacent to the former, and a volcanic island arc was approaching the latter, no tephras have been found in Nevada. According to paleogeographic models (e.g., Witzke and Heckel, 1988; Scotese and McKerrow, 1990), both areas were at the same approximate latitude through the Devonian. So it would appear that through the Devonian, winds were west-directed easterlies in both regions. It seems unlikely that prevailing wind direction would have reversed for several million years, and then reverted back to the previous trends. So, it seems that a bias of temporary shift in prevailing wind direction would not seem to apply to the Appalachian basin and the Acadian/Neoacadian orogen. This west-directed tephra transport appears to imply that the Appalachian Basin and Nevada were apparently within the belt of easterly winds, at or north of 30°S latitude.

Finally, compositions of the dated hinterland rocks through those intervals are also largely silicic, which should lead to explosive eruptions and tephra deposition. So, is the absence of foreland tephras through these times due to environmental-related preservational biases? This seems unlikely, especially in lower Frasnian deposits in western New York, which represent relatively offshore, deep-water, low-energy, and often anoxic to dysoxic settings, which should enhance preservation of air-fall tephra layers.

Questions arising from this initial comparison point to new research challenges. These include more concentrated field searches for additional layers, especially in intervals with hinterland igneous rocks, yet with few-to-no known tephras. These efforts should also consider other possible primary or diagenetic lithification modes beyond that of the typical recessed, easily weathered K-bentonite clay beds. Tephras that are preserved as more resistant phenocryst-rich “tuffs,” carbonate-, silica-, or pyrite-cemented beds, and other less obvious modes of preservation may be less easily recognized, but would still need to be documented.

We would like to thank numerous people for their discussions, insights, and other assistance over many years, including first the late John Dennison. He was the first to document many Devonian air-fall tephras and, along with Daniel Textoris, began using the tephra data to reconstruct regional paleovolcanism. We would also like to thank James (Jim) Conkin and his wife Barbara, whose great detailing of Devonian air-fall tephras also led the way for subsequent work. Thanks also go to Warren Huff, George Shaw, Carlton Brett, Doug Rankin, and R. Smith III. Insightful reviews from Warren Huff, Paul Karabinos, David Bailey, and Editor K. Lee Avary strengthened the paper. Finally, we are grateful to the editors, who invited us to participate in this volume, dedicated to one of the great “Deans of the Appalachian Basin Devonian,” John Dennison.

Through the efforts of Priscilla Dennison, John’s wife, and Daniel Textoris, Tioga tephras samples collected across the eastern United States by Dennison and Textoris are reposited in a subcollection of Paleozoic tephra samples, as part of the New York State Museum’s Sedimentary Rock Collection.