The Geology of the Union Furnace Area

Previous Studies

Kay (1944a, 1944b) developed the basic stratigraphic framework in use today, and Faill and others (1989) provided excellent stratigraphic descriptions of the Ordovician carbonates for field mapping. Thompson (1963) examined the stratigraphy and petrography of the Salona and Coburn Formations in central Pennsylvania. Rones (1969) did the same for the Hatter, Snyder, Linden Hall, and Nealmont Formations. Both Thompson’s (1963) and Rone’s (1969) reports include measured sections for exposures located along the Pennsylvania Railroad cut near Union Furnace. Wagner (1966) correlated the Middle and Upper Ordovician rocks of central Pennsylvania with those in the subsurface of western Pennsylvania, south central New York state, northern West Virginia, eastern Ohio, and southwestern Ontario. Chafetz (1969) reported on the stratigraphy and sedimentology of the Loysburg Formation in this region.

Berkheiser and Cullen-Lollis (1986) summarized the history of mining and geologic investigations of the Ordovician carbonate rocks at Union Furnace. They provided a detailed measured section of the Loysburg through Coburn Formations where they are exposed along PA 453. Berkheiser and Cullen-Lollis (1986) developed the first modern facies interpretation for these rocks, and were the first to suggest a carbonate ramp depositional environment for these sediments in central Pennsylvania (Figure 13). We found their work invaluable during our field and core studies.

Gardiner-Kuserk (1988) identified centimeter-scale cyclic sedimentation patterns in the Upper Ordovician carbonates of central Pennsylvania, and she attributed the patterns to storm events on a progressively deepening ramp. The cycles she identified are made up of repetitive successions of micrite, bioclastic or intraclastic limestone, and siliciclastic shale. The bioclastic or intraclastic limestone occurs as a lag, and the other lithologies are arranged in fining-upward beds reflecting waning storm conditions. Gardiner-Kuserk (1988) identified five such cyclic patterns on the basis of textural attributes, carbonate allochems, sedimentary structures, and bed thickness.

Slupik (1999) investigated the sedimentology and stable isotope chemostratigraphy of the Nealmont, Salona, and Coburn Formations in central Pennsylvania. She too suggested that the rocks originated in a ramp environment. Slupik’s (1999) interpretation departs from those of Berkheiser (1986) and Gardiner-Kuserk (1988), however, in that she interprets the Salona Formation carbonates as deposited in the deep ramp or slope environment below fair-weather wave base, but slightly above storm wave base. She interprets the overlying Coburn formation as deposited in shallower water than the Salona, still below fair weather wave base, but well above storm wave base. Thus she places the Coburn in a more proximal ramp position, implying that it is a regressive sequence. Slupik (1999) also documented a positive inorganic carbon isotope excursion of approximately 3 o/oo (PDB) across the Nealmont and Salona Formations boundary. She correlated this isotope shift with a previously recognized δ13C positive isotope excursion observed in Upper Ordovician petroleum source rocks in eastern Iowa, and attributed it to regional increase in organic productivity and preservation.

Carbonate Ramp Facies Models

Carbonate ramps (Figure 14) comprise a category of platform characterized by gentle slopes (typically less than one degree) on which shallow water, wave-agitated sediments of the near shore zone pass down slope into deeper water, low-energy deposits (Read, 1985). The transition from shallow to relatively deeper water facies occurs without a marked break in slope. Ramps differ from rimmed carbonate shelves in that continuous reef trends usually are absent, high-energy carbonate sands accumulate near the shoreline, and deeper water breccias, if present at all, lack clasts of shelf-edge lithofacies.

Carbonate ramps are subdivided on the basis of profile into homoclinal ramps and distally steepened ramps (Figure 14). Homoclinal ramps have comparatively uniform, mild slopes (<1°) into the basin. Carbonate facies include 1) peritidal and lagoonal facies; 2) shoal-water bank complexes; 3) open marine, deeper ramp, muddy sediments; and 4) slope and basin lime muds and interbedded shales (Read, 1985) (Figure 14). The Persian Gulf and Shark Bay, Australia are examples of modern homoclinal carbonate ramps (Figure 15).

Distally steepened ramps are defined by an increase in gradient in the outer, deep ramp environments (Read, 1985; Tucker and Wright, 1990) (Figure 14). Sediment gravity flow deposits are common in this setting. Since the break in slope occurs in relatively deep water on the ramp, the resedimented deposits consist of outer ramp and upper slope carbonate sands and muds, and clasts of this material. Peritidal, lagoonal, and shoal water facies occur well back on the platform, and deep ramp carbonate mud blankets (low energy) or broad lime sand blankets (high energy) occur seaward of the shoal complexes (Read, 1985) (Figure 14). The deep ramp lithofacies consist of argillaceous, burrowed, nodular skeletal wackestone/mudstone. The rocks contain an open marine biota, and may have slumps, breccias, and turbidities (Read, 1985). The northeastern Yucatan and western Florida are modern distally steepened ramps (Figure 16).

The Trenton and Black River Group successions contain elements of both ramp types. The Black River to upper Linder Hall Formation at Union Furnace contains a facies succession readily interpreted as an ancient barrier-bank type homoclinal ramp (Figure 14). These rocks exhibit an upward-deepening trend comprised of, in ascending order, peritidal, lagoonal, shoal complex, and deeper ramp facies (Figure 17) (See Figure 18 for the symbol legend for Figure 17). The uppermost Linden Hall and the Trenton Group, however, contain facies characteristic of an ancient distally steepened ramp (Figure 14). The top of the Linden Hall and the Nealmont Formations exhibit an upward deepening trend consisting of middle to deeper ramp facies. The Salona, and Coburn Formations exhibit an upward deepening, then upward-shallowing trend comprised of low-energy deep ramp facies overlain by slope and basin facies, which are, in turn, overlain by more proximal deep ramp facies (Figure 17). We present the details of our descriptions, interpretations, and arguments in the Stop 3 discussion below.


Hardgrounds are synsedimentarily lithified carbonate seafloors that become hardened in situ by the precipitation of carbonate cement in the primary pore space (Wilson and Palmer, 1992, p.3). They are sedimentary horizons in marine carbonates that exhibit evidence of exposure on the sea floor as lithified rock (Figure 19). Hardgrounds form under a reasonably consistent set of physical conditions; thus geologists use hardgrounds to estimate ancient sedimentation and erosion rates, oceanic geochemistry, and both tectonic and eustatic sea level changes. Hardgrounds are often thought to mark the tops of regressive sea level changes, or at least to have been followed by transgression (Fursich, 1979; James and Bone, 1991; Jones and Desrochers, 1992; Lehrmann and Goldhammer, 1999) (Figure 20). Other hardgrounds form rapidly, however, and are not associated with sea level changes (Shinn, 1969; Dravis, 1979; Brett and Brookfield, 1984; Hillgartner and others, 2001).

Paleontologists study hardgrounds to determine the evolutionary patterns of different organisms that adapted to life on these hard surfaces (Brett, 1988). Well-preserved hardground biotas provide an opportunity to analyze a paleoecosystem with accuracy not possible in most fossil assemblages (Wilson and Palmer, 1992). 

Hardgrounds are one of the most conspicuous and important carbonate rock features at Union Furnace. Yet they apparently have gone unrecognized here throughout decades of study at this outcrop. They have been misidentified as mud cracks by some workers (see Rones, 1969), a significant mistake because these hardgrounds signify submarine cementation, exposure, and erosion rather than subaerial exposure. Although Faill and others (1989) did not designate hardgrounds in these rocks, they did note, “…light colored, very irregular bands, 0.5 to 2 cm thick and 3 to 10 cm apart, that extend for 2 meters or more” in the Snyder Formation. They estimated that these bands (which are hardgrounds) constitute at least 15 percent of the entire Snyder Formation.

James and Choquette (1990) listed numerous criteria for recognizing ancient submarine hardgrounds, and Demicco and Hardie (1994) added to that list. Fursich, 1979 suggested a combined morphological-genetic classification of hardgrounds. Brett and Brookfield (1984) published an excellent paper on the morphology, paleontology, and origin of carbonate hardgrounds in the Trenton and Black River Groups of southern Ontario, Canada. They modified Fursich’s classification, and we find this most applicable for describing the hardgrounds exposed at Union Furnace. We recognize the following types of hardgrounds described by Brett and Brookfield (1984) (Figure 21):

  •  Smooth hardgrounds: developed on uniformly cemented beds by removal of unlithified sediment without significant erosion; lack burrow systems.
  • Rolling hardgrounds: formed by slight erosion of a nodular bed, or a bed with dwelling burrows.
  • Hummocky and undercut hardgrounds: form through further erosion of rolling hardgrounds accompanied by differential removal of unlithified sediment from burrows and between nodules.
  •  Pebbly hardgrounds: formed by erosion and collapse of hummocky and undercut hardgrounds.
  • Reworked hardgrounds: formed by removal of pebbles to other depositional sites.
  • Planar hardgrounds: formed when the sediment filling burrows lithified to the same hardness as the rest of the bed; involved significant erosional planning.
  • Composite hardgrounds: closely adjacent hardgrounds formed by renewed sedimentation and erosion, and closely spaced hardgrounds that feather into one another.

These morphologies can be further modified by bioerosion to create pitted and extensively bored hardground beds. All of these hardground types are present at the Union Furnace outcrop. Hardgrounds occur in all of the formations exposed there. They are extensively developed as composite hardgrounds in the Snyder and Linden Hall Formations. Figures 22, 23, and 24 illustrate some common hardground morphologies found at Union Furnace.

Nodular limestones often form through early diagenetic processes on the sea floor, and might be related to early hardground formation. For example, nodules of limestone produced by differential cementation of slope deposits off the Bahamas bank may be a modern analogue for nodular bedding found in ancient distal ramp carbonates (Mullins and others, 1980). When nodule growth continues in these environments, a partially continuous to continuous three-dimensional framework takes shape, which might eventually form an incipient hardground. Nodular bedding in limestones at Union Furnace might reflect initial hardground development (Figure 25). Compaction and pressure solution, however, have severely modified both the depositional and early diagenetic fabrics of nodular-bedded limestones at Union Furnace, particularly in the Nealmont Formation. It is difficult to discriminate between depositional, diagenetic, and compactional fabrics in many of the nodular bedded limestones.

Sequence Stratigraphy

Holland and Patzkowsky (1996) established a well-developed sequence stratigraphic framework for the southern Appalachians, the Cincinnati Arch, and the Nashville Dome. They recognized fourteen depositional sequences in the Middle and Upper Ordovician strata of these regions. Six of these sequences begin in the Mohawkian Series, and three of these (M3, M4, and M5) include the Turinian (former Black Riverian) and Chatfieldian (former Trentonian or Rocklandian-Kirkfieldian) Stages.

Cornell and Brett (2001) suggested that the Black River Group of central New York state, and adjacent Ontario, comprises most or all of the M3 and M4 sequences. The also suggest that the Watertown, Selby, and Napanee Formations of central New York comprise M5, and that the base of the Watertown Formation should be used as the Trenton –Black River Group boundary.  

It is far too early in our work here in Pennsylvania to argue how the sequences here will fit into the regional sequence stratigraphic framework of the Appalachian basin. We can share some of our thoughts and speculations, however, and we invite comments and observations from all field trip participants.

We discern four third-order sequences in the strata that we measured along PA PA 453 at Union Furnace (Figure 17). The first begins in the lowermost Hatter Formation. We interpret the contact between the low-energy, shallowing-upward peritidal successions and the distinctive lagoonal facies (just above bentonite B0 in Unit 2) as a combined sequence boundary/transgressive surface, as described by Holland and Patzkowsky (1996) for Ordovician carbonate sequences in the eastern United States. This boundary marks the first flooding surface beginning a set of retrogradational parasequences, i.e., an overall deepening upward trend. Holland and Patzkowsky (1996) recognized common two-part transgressive systems tracts in the Ordovician carbonate sequences of the region, and these are apparent here in this sequence. The lagoonal facies comprise the lower transgressive systems tract, and the oolitic and mixed grainstones of Units 6 and 7 in the Hatter Formation, and Units 9 – 12 in the Snyder Formation, comprise an upper transgressive systems tract. The upper transgressive systems tract consists of backstepping parasequences (see Rones, 1969, Pl. 6 and figs. 21 – 24). The base of Unit 13, where the distinctive bryozoan bioherm is located, marks the turn around from retrogradational to progradational stacking. This is the maximum flooding surface. All of the Snyder and most Linden Hall strata above this offlap (see Rones, 1969), and make up the high stand systems tract. The unconformity near the top of the Linden Hall is the top of this first sequence.

The second sequence is more difficult to discern, and we propose it as tentative at best. It begins at the unconformity near the top of the Linden Hall Formation. This is another sequence boundary/transgressive surface, and the start of another overall deepening upward trend. The transgressive systems tract consists of the restricted middle ramp carbonate facies in Units 24 and 25, and slightly into Unit 26. The maximum flooding surface is just above bentonite B8, where the nodular bedding of the Rodman Member become evident. Rones (1969) mapped the Rodman as an offlapping succession, and Arens and Cuffy (1989) documented wave-induced shoaling and bioherm development in this member at other locations. The Rodman Member is more fossiliferous and coarser grained than subjacent Nealmont limestones at this location. For these reasons we tentatively note the start of the highstand systems tract at the base of the nodular-bedded Nealmont facies (Rodman Member). We place the upper boundary of this second sequence at the base of the rhythmites in the lower Salona Formation, just above bentonite B13 and a little more than 2.5 m (8.2 ft) below the Millbrig K-bentonite.

The third sequence begins at the base of the rhythmites in the lower Salona Formation. This contact represents an abrupt change to a deeper water environment. It is another combined sequence boundary/transgressive surface, and we interpret it as a drowning unconformity separating a highstand systems tract from a transgressive systems tract. This unconformity is what Schlager (1999) proposes to call a Type 3 sequence boundary. The Salona rhythmites continue to deepen upward for about 30 m (9.8 ft) to the base of the Roaring Springs Member. Slupik (1999) placed the maximum flooding surface here on the basis of the change from mudstones and argillaceous mudstones to laminated quartz-rich mudstones, an increase in terrigenous clay, and the evident changes in bedding thickness. The upper Salona Formation comprises the highstand systems tract.

The abrupt appearance of very fossiliferous, wave-agitated strata at the base of the Coburn Formation marks a distinct basinward shift of facies, and begins a new sequence.

The reported age of the Diecke K-bentonite is 454.6 ma (Leslie and Bergstrom, 1997). The Diecke is 0.5 m (1.6 ft) above the base of the Salona Formation at the Union Furnace outcrop along PA 453, and 0.55 m (1.8 ft) above the base of the Salona in core from the Wallace Farm just south of the outcrop. This places the lowermost Salona Formation, as recognized and mapped in central Pennsylvania, near the top of Holland and Patzkowsky’s (1996) M3 sequence. The drowning unconformity at the base of the Salona Formation rhythmites might be the M3/M4 sequence boundary of Holland and Patzkowsky (1996). The Black River Group, Nealmont Formation, and lowermost Salona Formation, as mapped in central Pennsylvania, probably represent the M2 and M3 sequences. The Salona Formation contains the M4 sequence, and the Coburn Formation encompasses the M5 sequence.

We repeat that these ideas are conjecture, and await further measurements, descriptions, and correlations before they can evolve into a useful concept of sequence stratigraphy in this part of the basin.