GEOLOGICAL EVOLUTION OF THE COBLESKILL PLATEAU,NEW YORK STATE, USA
PAUL A. RUBIN
HydroQuest, PO Box 387, Stone Ridge, NY, USA 12484
Abstract
The geological evolution of the Cobleskill Plateau, in eastern New York State, USA, is interpreted from geologic field investigations. The Cobleskill Plateau contains four of the State’s longest caves (Secret- Bensons, Howe Caverns, McFail’s, Barrack Zourie). Major caves in Upper Silurian-Lower Devonian Helderberg Group carbonates, some with strike-oriented segments, developed progressively from east to west, coincident with headward incision of the Cobleskill Plateau by Cobleskill Creek. Recognition of Guenther Resurgence Cave as the relict outlet of Howe Caverns permits an estimate of initial plateau cave development on the order of 1.5 to 5 my. The interaction between karst and multiple glaciations, as interpreted based on the superposition of cave deposits, is significant. The morphology and superposition of cave passages, sediment deposits, flowstone, and breakdown enable interpretation of late Cenozoic geomorphology and climatic change. Regional geologic history is recorded by fills in plateau caves and is especially well preserved in Howe Caverns. A moderately rounded, chert-rich, pebble melt-out till present as canyon fill, on ledges, in braided anastomotic tubes, and in two phreatic conduits, provides evidence of one of the oldest glacial advances or retreats in New YorkState. They indicate rapid infilling of till via sinkhole and joint inputs beneath warm-based glacial ice. Deeply weathered chert pebbles indicate an age of more than 1 my. This is further supported by younger cave fills and massive flowstone over this till. Some fill removal required large flows of water, probably glacial meltwater. Floodwater dissolution of speleothems and canyon walls provide evidence that repeated long-term influx of glacial meltwater increased Plateau cave dimensions. Many Plateau caves were later twice inundated by Glacial Lake Schoharie which, in places, left behind thick rhythmites. Pebble till and breakdown-rich horizons within glaciolacustrine sediments provide evidence of a ~ 87 m lake level change around the Middleburgh readvance. Around 14 ka, an ice tongue impounding this lake thinned sufficiently that Plateau caves drained.
1. Introduction
Cobleskill Plateau caves have repeatedly been covered by continental scale ice sheets. While researchers have recognized and determined that these and other area caves predate Wisconsinan glaciations (Lauritzen and Mylroie, 2000), we have, to only a limited degree, been able to elucidate earlier regional landscape and climatic conditions. Surficial erosion and valley filling by sediments have served to remove and mask the geologic record. However, in-cave sediments and material have faithfully preserved portions of New YorkState’s geologic history. Glaciologists believe that the Pleistocene Laurentide ice sheet made four major advances into New YorkState (Isachsen et al., 2000). Some had multiple advances and retreats, as indicated by moraines and marine oxygen isotope data. In New YorkState almost all sediments not associated with the most recent glacial advance (i.e., Woodfordian Substage of the Wisconsinan Stage) were removed, leaving few clues as to sediment type and climatic conditions associated with pre-Wisconsinan advances and interstades. Sediments preserved in Howe Caverns, isolated from typical terrestrial erosive processes, provide a window into New YorkState’s Cenozoic history. The trunk passage is ideal for the preservation of sediments because of its location near a valley flank and a number of sinkhole inputs that deposited sediments downstream of a major passage collapse area. Here, a first approximation of the interaction between caves and glacial processes in the late Cenozoic Era is presented.
2. Geologic Setting
Caves and karst features in the Cobleskill Plateau have developed in approximately 40-50 m of the Helderberg Group carbonates. A gentle south-southwesterly bedrock dip controls vadose cave development. Many sinkholes drain low-permeability glacial soils into preglacial caves, sometimes around drumlins. McFail’s Cave, the longest and deepest surveyed cave in the plateau (7 km, 96 m; Palmer, 1976), has high canyon passages and a 2+ m high phreatic passage. Other Plateau caves have similar conduits, and domes to heights of 33 m. Howe Caverns and Secret Caverns are the only caves in the area open to the public. A number of caves and a mine in the Cobleskill Plateau reveal faulting. The importance of faulting and its impact on cave development has been debated by researchers. All downstream Howe Caverns and the transition zone between phreatic levels follow a thrust fault zone. In places, fault planes are displaced some 2 m and beds are repeated. Laterally continuous, slickensided, white calcitic horizons commonly mark multiple bedding thrusts. Widely different slickenside orientations revealed on opposite sides of a single thick calcite bed provide evidence for at least two distinct faulting episodes probably associated with the Acadian Orogeny (400-380 mya). A series of offset extension veins in downstream Howe Caverns reveals rotational motion along multiple fault planes. Examination of fault planes relative to cave development reveals that it is only in the transition zone between the cave’s two phreatic levels that influence of the fault zone has been significant (Mylroie, 1977; this author).
3. Development and Age of CobleskillPlateauCaves
Cave development in Helderberg Group carbonates has adjusted to multiple regional base-levels coincident with the westward incision of the plateau by the Cobleskill Creek and valley aggradation via glacial sediments. Initial cave development progressed in a down-dip direction until reaching a former water table elevation graded to the Cobleskill Creek (Fig. 1). Here, a large phreatic conduit connected McFail’s Cave with Howe Caverns prior to occlusion by glacial sediments (e.g., Egemeier, 1969; Palmer, 1976; Mylroie, 1977; Palmer et al., 1991; Dumont, 1995; Palmer and Rubin, 2007). Subsequent vadose entrenchment of McFail’s Cave occurred as Cobleskill Creek downcutting continued. McFail’s Cave now resurges upward through glacial sediments at Doc Shaul’s Spring. Howe Caverns, a beheaded relict portion of McFail’s NW-SE passage, documents regional history and climate change spanning multiple glaciations. A 12-m-diameter phreatic conduit that formerly conducted groundwater through Howe Caverns to GuentherResurgenceCave (Palmer and Rubin, 2007) is now 45 m above the invert elevation of Cobleskill Creek. The large size of this conduit documents a long-term stable regional base level. Measurements of chemical equilibria in McFail’s Cave by Palmer (2007) permit a possible extrapolation of a cave age of 1.5-5 my based on his estimated entrenchment rate of 0.001-0.003 cm/yr. Vadose conduits in the Plateau, including a Howe Caverns canyon passage graded to this higher phreatic tube (Winding Way, Fig. 2), provide evidence of flow from both active and now-abandoned sources. A drop in regional base level of 6 m is documented in a lower strike-oriented phreatic conduit, representing a second stable base level in the region. Later vadose canyon incision followed base-level lowering. Cave deposits provide material that can be used to further refine age and geologic events.
4. Chronology of Late Cenozoic Events Interpreted from Howe Caverns
The relative superposition of cave sediments and materials, and their subsequent alteration, allow a first interpretation of glaciogenic sequences, events, associated climatic conditions, and hydrologic conditions. Absolute dating of these events via uranium-series disequilibrium dating and other methods has not been conducted. For ease of reference, the geologic sequence has been broken into chronologic intervals based on superposition and relative age, sometimes grouping several events together. Future research may help refine this working model and may add to these geologic intervals. Furthermore, age dating of speleothems and sediment deposits may identify depositional gaps.
Interval 1: Flowstone and small stalagmites were deposited over small wavelength scallops on exposed bedrock walls of the highest level phreatic tube. These small stalagmites may be some of the oldest speleothems in the northeastern United States. Formation precipitation is interpreted as having occurred during interstadial, ice-free, conditions some 1-5 mya.
Interval 2: A subglacial, pebble-rich (~50%), melt-out till (Fig. 3) flowed into Howe Caverns primarily through a large open joint near the Vestibule and via several sinkhole inputs. The ~ 1m wide joint trending N14°E is located near a valley wall. Melt-out till formation and flowage occurs commonly at the margins of glaciers and ice sheets, where ice is nearly stagnant (Hambrey, 1994; Sugden and John, 1976) and where thick unfrozen subglacial till exists beneath temperate ice (Boulton, 1972). Unstable, saturated, till (yellowish brown sandy clay loam to clay), injected under pressurized conditions, was squeezed throughout the cave. It is assumed that the cave outlet was open in an unconfined hydrologic setting in order for the till to flow completely through the cave. Whether or not melt-out till exited the cave subaerially or under glacial ice is not clear. Till filled the cave’s vadose canyon to unknown depths, as well as much of both phreatic conduit levels. Its hard-packed, over-consolidated, form indicates drying in-situ vs. after emplacement by fluvial action. Till also fills much of a high-level phreatic conduit exposed in a quarry wall near the cave’s original downstream outlet at Guenther Resurgence Cave, situated above the new Cave House Museum (Palmer and Rubin, 2007). Nearby, excavation of the quarry floor within a de-roofed portion of Howe Caverns produced massive quantities of this melt-out till. The vertical distribution of this deposit, emplaced throughout all levels of the cave under subglacial conditions, documents that the cave was open and actively functioning as a zone of low hydraulic head. Therefore, at least locally, it is likely that preglacial conduits served to control the potentiometric surface beneath continental or wasting ice. Moderately well-rounded chert pebbles in this deposit suggest subglacial transport from beyond the small watershed present today. The deeply weathered condition of insoluble pebbles indicates an age in excess of one million years (pers. comm. with Campbell, in Palmer, 2007). This interpretation is further supported by deep weathering rinds within chert pebbles (Flint, 1957). The melt-out till is interpreted as evidence of glacial recession. (Alternately, if this till was not emplaced under pressure, it would reflect deposition coincident with either the earliest phreatic cave level or at a time of valley filling and related rise in the regional base level.) If indeed the age of this pebble till exceeds 1 million years, it would provide evidence based on marine oxygen isotope stage dates that Cobleskill Plateau caves may have survived more than 10 glacial advances. This till may represent the oldest documented terrestrial Pleistocene or Pliocene deposit in New YorkState.
Interval 3: Following emplacement, the melt-out pebble till was deeply excavated throughout much of the cave. Remnants of this deposit occur on ledges, in canyon exposures, in infeeders, and in floodwater tubes. The over-consolidated nature of this till probably required considerable time and water quantity to excavate, likely substantially more than that measured during a March 2007 flood (6 cms). Interval timing is somewhat constrained by Interval 7 flowstone deposits that would have substantially reduced stream power and caused back flooding in the main cave conduit had they been present. This would have precluded till excavation. It is interpreted here that till excavation occurred via invasion of subglacial meltwater during glacier thinning and retreat (i.e., warm-based ice under either polar tundra or microthermal climatic conditions). The Cobleskill Creek valley may have been partially or wholly ice-free at this time. It is likely that Interval 3 represents many hundreds of thousands of years and multiple glaciations, with little or no remaining sedimentary evidence.
Interval4: Climatic conditions were warm during this time interval. Surface water infiltration into the cave slowly began to increase, resulting in significant formation development in the Winding Way, near the Vestibule area, above the LongBridge, and elsewhere. Saturated cave pool conditions and related calcite growth occurred in natural cave segments on both ends of the Manmade Tunnel. Sediments in the Silent Chamber area were cemented together. Also, calcite wall coatings were precipitated in conduits that, during wet climatic conditions, conduct floodwater (e.g., Fat Man’s Misery/Devil’s Gangway). In the absence of age dating, speleothems in this interval are interpreted as pre-dating massive flowstone growth discussed in Interval 7. Formations developed in this time interval were subjected to floodwater dissolution, whereas the massive flowstone cascades of Interval 7 remain largely intact.
Interval5: Under ice-margin conditions, a second melt-out till flowed into Howe Caverns. This flow was less massive than that in Interval 2 and is predominantly pebble-free brown clay. This till includes cobble and boulder (25+ cm) sized rocks probably derived from in-cave breakdown, indicating that it flowed in rapidly in a saturated state. Remnants of this till occlude side passages in the Winding Way and several large side conduits along the up-dip side of the Howe Caverns master conduit upstream of the Lake of Venus. The relative timing of this till interval is inferred to be after removal of the Interval 2 pebble till and before massive invading subglacial flows (Interval 6) removed most of this deposit from the Winding Way. The location of sediment exposures provides evidence that this clay-rich till formerly filled much of the Winding Way and part of the main cave. Its limited exposures in lower cave levels outside of the Winding Way suggest that the cave served as an open conduit drain to a wholly or partially ice-free Cobleskill Valley. Thus, flowage into the cave probably occurred at or near a wasting ice margin. A small sediment exposure in the Lake of Venus depicts a basal contact with the Interval 2 melt-out till and overlying glaciolacustrine rhythmites (Interval 9).
Interval6: This interval was marked by the influx of vast quantities of subglacial meltwater that far exceeded historic floods. This required a watershed larger than that of today, indicating a far-reaching subglacial source area (Rubin, 1991). During this stage, almost all Interval 5 melt-out till was removed from the Winding Way and lower levels of Howe Caverns. In addition, short wavelength wall scallops attest to aggressive, high flows in the Winding Way that continued to expand canyon width. Concurrently, flood water dissolution partially dissolved Interval 4 flowstone formations in the Winding Way and elsewhere, leaving small wavelength scallops behind as evidence. During this interval, meltwater continued to expand the main cave and furthered development and enlargement of tap-off passages extending down-dip toward the Cobleskill Creek. Today, some of these tap-off passages are active while others remain as small relict caves.
Interval 7: A continental glacier again advanced over the region. Massive flowstone cascades and formations developed in area caves over time. The partial cross-section of a Lake of Venus cascade (Fig.4) exposes 2 m of thinly-layered flowstone. If van Beynen et al.’s (2004) minimum annual speleothem growth rate of 0.019 mm/yr is assumed, a growth time of about 100,000 years is possible. Other flowstone deposits are more than twice as thick, indicating that quiescent climatic conditions may have persisted for at least 250,000 years (Fig.5). Slow growth is likely because massive flowstone extends downward into what are today active stream passages (e.g., River Styx, ReynoldsRiver). U/Th dating of this flowstone may be possible. Lauritzen and Mylroie (2000) correlated U/Th terrestrial-based dates obtained from speleothems with marine oxygen-isotope stages, thus documenting that many regional caves survived multiple glacial advances. Sample ages analyzed by Lauritzen and Mylroie (2000) and Smart (see Dumont 1995) from Barrack Zourie and CabooseCaves reveal flowstone growth in marine oxygen isotope stages 6 and 8, indicating speleothem growth during glacial conditions. It is possible that U/Th dating of Interval 7 flowstone cascades may reveal formation coincident with oxygen isotope stages 6 to 8 (130-300 Ka). Pebbles from the Interval 2 melt-out till are cemented in the base of some of these deposits. Similarly, a large formation in the center of a nearby Barton Hill stream passage (Schoharie Caverns above waterfall) also indicates long-term quiescent regional conditions, possibly during the same time period. Alternately, Interval 7 flowstone may predate oxygen isotope stages 6-8 (e.g., a speleothem in Schoharie Caverns yielded a date greater than 350 Ka; Lauritzen and Mylroie, 2000). Massive flowstone growth throughout the main stem of Howe Caverns is interpreted as having initiated beneath a continental glacier, with partial growth under warmer conditions. U/Th dating is warranted.
Interval8: This long-term interval is characterized by the break out of large and small bedrock blocks from conduit ceilings and walls. Stream erosion continued and new formations developed. This interval may include one or more glacial advances and retreats. Additional field work may find that this interval predates Interval 7.