G.S. Springer, H.D. Rowe, B. Hardt, F.G. Cocina, R.L. Edwards, and H.

Cheng – Climate driven changes in river channel morphology and

base level during the Holocene and Late Pleistocene of southeastern West Virginia. Journal of Cave and Karst Studies, v. 71, no. 2, p. 121–129.

CLIMATE DRIVEN CHANGES IN RIVER CHANNEL

MORPHOLOGY AND BASE LEVEL DURING THE
HOLOCENE AND LATE PLEISTOCENE OF
SOUTHEASTERN WEST VIRGINIA
GREGORY S. SPRINGER* , HAROLD D. ROWE , BEN HARDT , FRANK G. COCINA ,
,1 4 2 1,3

R. LAWRENCE EDWARDS , AND HAI CHENG 2 2

Abstract: Rivers commonly respond to climate change by aggrading or incising. This is

well documented for North American rivers in arid and proglacial regions, but is also
true of rivers in unglaciated, humid-temperate regions. Here, we present a record of
Holocene hydroclimatology for a humid, temperate watershed in the Appalachian
Mountains of eastern North America. We use stable isotope geochemistries of a
stalagmite and clastic cave sediments to reconstruct Holocene climate and ecology in the
Greenbrier River catchment (3,600 km2) of southeastern West Virginia. Independently,
we use river-deposited cave sediments to construct a history of incision, aggradation, and
morphological change in the surface channel. The clastic cave deposits display enriched
(less negative) values of sedimentary d13Corg during the Holocene Climatic Optimum
(HCO), which regional pollen records indicate was warm compared to later climes. The
river channel had aggraded by .4 m during or prior to the HCO and adopted an alluvial
morphology, probably due to the mobilization of hillslope sediments accumulated during
the colder, drier full-glacial conditions of the Late Pleistocene. As climate moistened
during the Holocene, the Greenbrier River incised through channel-filling sediments and
back onto bedrock, but not until ,3,500 cal. years B.P. Therefore, the bedrock
morphology of many streams in the Appalachian Mountains may not have existed for
much of the Holocene, which highlights the effect of climate variability on channel
processes. The base-level rise is more evidence that bedrock incision by rivers is often
episodic and that slow, long-term incision rates reported for Appalachian Rivers are
probably not representative of short-term incision rates.

INTRODUCTION alluvial fans and valley bottoms are remobilized and

transported into the stream (Schumm, 1973). In such a
Average global temperatures have been rising in recent generalized scenario, channel aggradation raises base level,
time and climate change is expected to measurably affect which leads to increased flooding on terraces and
stream hydrology (Kundzewicz et al., 2007). As such, floodplains (e.g., Bull, 1988). However, corresponding
societal interactions with rivers will change, including those changes in stream morphology may change stage-discharge
related to flooding. The magnitudes and directions of these relationships and thereby increase or decrease peak flood
changes are uncertain and new, more quantitative data are stages (e.g., Stover and Montgomery, 2001). Thus,
needed to predict future responses and prepare society for predicting changes in base level and channel morphologies
the changes ahead (Wood et al., 2002; Maurer et al., 2004; are important steps toward understanding future stream
Lettenmaier et al., 2006). High-resolution studies are behaviors and risks.
especially needed because of the scale mismatch between Here, we reconstruct Holocene changes in channel
climatic models and catchments, as the latter are the morphology and base-level elevation in a mountainous
functional units of the terrestrial hydrosphere (Kundzewicz watershed draining the Eastern Continental Divide (ECD)
et al., 2007). Climate change affects a host of stream of North America (Fig. 1). The sedimentology of river-
variables, including channel morphology, which is a good
predictor of stream behaviors and processes (Montgomery * Corresponding author
1
and Buffington, 1997). In the simplest case, a decrease in Department of Geological Sciences, Ohio University, Athens, OH, 45701,
springeg@ohio.edu
net precipitation can cause channels to infill with sediment 2
Department of Geology and Geophysics, University of Minnesota, Minneapolis,
(aggrade) and change from an incised channel to a broader, MN, 55455
3
shallower channel (Knighton, 1998). Conversely, an Present Address: Qinetiq North America, 1350 Central Ave. Los Alamos, NM,
87544
increase in precipitation, stream discharge, or gradient 4
Department of Geology, Box 19049, University of Texas at Arlington, Arlington,
can also cause channels to infill if sediments stored on TX, 76019-0049

Journal of Cave and Karst Studies, August 2009 N 121

CLIMATE DRIVEN CHANGES IN RIVER CHANNEL MORPHOLOGY AND BASE LEVEL DURING THE HOLOCENE AND LATE PLEISTOCENE OF SOUTHEASTERN
WEST VIRGINIA

Figure 2. The examined stream reach contains three large

Figure 1. Colonial Acres Cave (CAC) and the associated caves developed in the river-right bank of the Greenbrier
surface channel are located along the Greenbrier River in River (inset). Colonial Acres Tubes contain no usable
southeastern West Virginia. The Greenbrier River watershed sediments. Greenbrier River Cave (GRC) contains slack-
abuts the Eastern Continental Divide along its north and east water sediments deposited by historic and ancient floods, but
boundaries. The CAC/Greenbrier River sedimentary record no evidence of Holocene base level fluctuations has been
is compared to paleoclimate records drawn from Buckeye found GRC. This probably reflects its vertical position
Creek Cave (BCC), Cranberry Glades (CG), and Mountain relative to the river; much of GRC lies 7 m above the low
Lake, Virginia (ML). water surface of the Greenbrier River. The slackwater
sediments discussed in this study are from a modest-size
room and excavated trench in Colonial Acres Cave (CAC).
deposited slackwater sediments is used to construct a
Greenbrier River floodwaters can reach the room from
history of stream hydrology and morphology, which is
impassable pathways within a voice-transmitting collapse at
compared to a record of Holocene climates. We reconstruct the south end of the trench room, a series of tubular passages
climate using stalagmitic d18Ocalcite and d13Ccalcite values, originating at three side-by-side entrances, and a small
bulk organic values of d13Corg in clastic cave sediments, passage emanating from a sump hydrologically connected to
and previously published palynological data. Ours is the the river (all shown).
first detailed simultaneous examination of the climate and
river hydrology within the source region of large Appala- response to changes in temperature, precipitation amount,
chian rivers (Fig. 1). and moisture source (Dansgaard, 1964; Rozanski et al.,
1993). While speleothems in tropical regions exhibit a
METHODS AND SAMPLE LOCATIONS dominant response to the amount effect (i.e., Wang et al.,
2001), mid-latitude sites have a larger temperature com-
STALAGMITE BCC-002 ponent (i.e., Dorale et al., 1998). Here, we assume that 1)
A calcite stalagmite (BCC-002) from Buckeye Creek changes in d18Ocalcite largely reflect past changes in above-
Cave (BCC) (Fig. 1) is used to develop a stable isotope cave air temperature and precipitation source, and 2)
record of climate for interpretation of the causes of d13Ccalcite and d13Corg are proxies for floral and soil
significant changes in slackwater stratigraphy in Colonial community composition, productivity, and relative mois-
Acres Cave (CAC) (Fig. 2) (Springer et al., 2008). BCC is ture levels (Kirby et al., 2002; McDermott, 2004). Post-
within the Greenbrier River watershed of southeastern 2,000 cal. years B.P. values of d13Ccalcite are not interpreted
West Virginia. The watershed is a major tributary of the because of anthropogenic disturbances to the BCC
westward-flowing New River and across the ECD from the watershed (White, 2007).
eastward-flowing Potomac and James Rivers. The 200- Fourteen 234U/230Th age estimates were obtained along
mm-long stalagmite was fed by a soda straw stalactite with the growth axis of BCC-002 using U/Th dating techniques
a slow, but incessant drip rate. Temperature and humidity developed for carbonates (Broecker, 1963) and adapted for
are stable in the cave passage and air movement is slow. measurement on a mass spectrometer (Edwards et al.,
Springer et al. (2008) provide a picture of the sawn, 1987). Calcite powders were milled with a dental drill,
polished stalagmite in their online supplemental material. dissolved, and spiked with a 233U-236U-229Th tracer.
The d18O values in stalagmites vary with changes in the Uranium and thorium were separated using anion ex-
meteoric water (Hendy, 1971), which predictably varies in change resin, and the clean U and Th were run separately
122 N Journal of Cave and Karst Studies, August 2009
 G.S. SPRINGER, H.D. ROWE, B. HARDT, F.G. COCINA, R.L. EDWARDS, AND H. CHENG

60-m-wide river, which is perennial. Locally, mean annual

discharge is 57 m3 s21 (n 5 111) with a peak historic
discharge of 2,600 m3 s21 (Springer, 2002).
CAC lies behind a vertical bank of the river. We
examined a 2-m-thick package of slackwater sediments
deposited by the river in a small room (Fig. 2). The room
ends in collapse against the riverbank. Voices and,
therefore, water readily pass through the collapse, although
light does not. The Greenbrier River is the only significant
source of water and sediment to CAC and completely fills
the room during and above medium recurrence-interval
floods. In addition to the collapse, floodwaters enter the
cave room via a 210-m-long, 1.5-m-diameter passage
originating at the cave entrance and via a 0.75-m-diameter
passage connected to a sump (Fig. 2) that is a permanently
flooded connection to the river. Sump depth varies with
river stage.
A 2-m-deep trench was excavated in CAC slackwater
sediments. Seven AMS-14C dates were obtained from
charcoal in the upper 70 cm of the trench. The ages of
underlying sediments are extrapolated using a sedimenta-
tion rate calculated from the upper 70 cm (175 mm ka21).
Only general conclusions are drawn from sediments below
70 cm because of age uncertainties. Each identifiable
stratigraphic unit was sampled for grain size analysis
(sieving). The sediments were contiguously sampled at a 2-
cm interval for determination of bulk sedimentary d13Corg.
Herein, the fraction of sand present is reported and used to
evaluate energy levels of the formative floodwaters. To
Figure 3. The Greenbrier River at Colonial Acres Cave obtain stable isotopic values for organic carbon (d13Corg),
(CAC) flows in a narrow valley containing narrow or no dry sediment powders were weighed into silver capsules,
floodplains, but many limestone cliff banks. Flow is toward repeatedly acidified with 6% sulfurous acid, and analyzed
the viewer (out of the page). Channel morphology is pool- using a Costech 4010 elemental analyzer coupled via a
riffle or pool-rapid with boulders and large cobbles Conflo III Device to a ThermoFinnigan DeltaPlusXP
dominating bedload. Bedrock is exposed in many riffles isotope-ratio mass spectrometer (IRMS). Values for
and most pools. CAC lies 300 m downstream of the lower- d13Corg are reported relative to V-PDB, and precision for
right corner of this picture. the isotopic standard (USGS-24) and unknowns is ,0.1%.
Temporal variations in sedimentary d13Corg represent a
on an inductively coupled plasma mass spectrometer (Shen watershed-integrated record of ecological change upstream
et al., 2002) along with a chemical blank. Ages were of the deposit. The d18Ocalcite and d13Ccalcite were also
calculated using the decay constants determined by Cheng determined using the ThermoFinnigan DeltaPlusXP iso-
et al. (2000). tope-ratio mass spectrometer, although standard carbonate
sample preparations were performed.
CLASTIC SLACKWATER SEDIMENTS
Clastic slackwater sediments deposited by the Green- HOLOCENE CLIMATE OF GREENBRIER RIVER WATERSHED
brier River in CAC are used to infer past hydrological
relationships between the cave and river (Fig. 2). As The d18Ocalcite and d13Ccalcite values from stalagmite
demonstrated by Springer et al. (1997), Springer (2002), BCC-002 show pronounced variations during the Holocene
and Bosch and White (2003), the hydrological relationships (Fig. 4A, B). The closest published palynological record is
are used to determine relative (to the cave) base-level from Cranberry Glades (Watts, 1979), located 56 km north
elevations, paleoflood stages, and surface channel hydrol- of BCC-002 and only 5 km from the Greenbrier River
ogies. Presently, the surface channel has a bedrock watershed (Fig. 1). The bog lies at an elevation of 1024 m,
morphology (.50% rock-lined), with coarse cobbles and which is 350 m higher than the land surface above BCC-
small boulders composing the bedload (Fig. 3). The 002. Comparing BCC-002 to the bog, d18Ocalcite values are
drainage area upstream of the cave is 3,600 km2. There heaviest during the mid-Holocene when hardwood pollen
are no indications of significant subsurface piracy of the abundance increases (notably Carya (hickory), Fagus
Journal of Cave and Karst Studies, August 2009 N 123
CLIMATE DRIVEN CHANGES IN RIVER CHANNEL MORPHOLOGY AND BASE LEVEL DURING THE HOLOCENE AND LATE PLEISTOCENE OF SOUTHEASTERN
WEST VIRGINIA

Holocene climate, which is recognized elsewhere as the

Holocene Climatic Optimum (HCO) or Hypsithermal
(Kaplan and Wolfe, 2006). The HCO began .6,500 cal.
years B.P. and ended at 4,200 cal. years B.P. when d18Ocalcite
values abruptly decrease by 0.4% (Fig. 4A). Reported
starting and ending dates of the HCO vary by region, but
termination of the HCO at Buckeye Creek Cave coincides
with purported mega-droughts and climatic shifts through-
out the upper Midwest of the USA, northern Africa, the
Middle East (see review by Booth et al., 2005), and New
Jersey (Li et al., 2007).
Mid-Holocene values of d13Ccalcite are enriched (less
negative) (Fig. 4B), which suggests that soils overlying
BCC-002 were drier or less productive than those of the
Late Holocene (McDermott, 2004). In fact, clay mineral-
ogies and weathering profiles of paleosols in the floodplain
of a Greenbrier River tributary (41 km NNE of BCC-002)
record warmer, drier conditions during the HCO (Driese et
al., 2005). Pinus (pine) pollen abundances reach their
Figure 4. Slackwater stratigraphy is compared to indepen- Holocene low during the HCO, despite the existence of a
dent paleoclimate proxies from Buckeye Creek Cave strong, positive correlation between the percentage of
stalagmite BCC-002 (A and B; Springer et al., 2008) and Pinus pollen in modern, mid-Atlantic sediments and
Cranberry Glades pollen (C–E; Watts, 1979). The d18Ocalcite temperatures of the preceding January (R2 5 0.91) (Willard
values are heaviest (most enriched) between 4,200 and 7,000 et al., 2005). Pinus pollen abundances are also correlated
cal. years B.P. At approximately the same time, Holocene with moisture and increase post-HCO, which is observed
temperatures peaked during what is known as the Holocene elsewhere along the eastern margin of North America and
Climatic Optimum (HCO) (Kaplan and Wolfe, 2006). is attributed to a general moistening (Watts, 1979; Webb,
Locally, enriched d18Ocalcite values record a warm HCO 1987; Willard et al., 2005). Presumably, (seasonal?) soil
climate at BCC and Colonial Acres Caves (CAC). Declining moisture was the major HCO control on Pinus abundance
d18Ocalcite values after 4,200 cal. years B.P. record an abrupt or its pollen productivity. Post-HCO values of d18Ocalcite
cooling, after which temperatures were comparatively stable and d13Ccalcite are comparatively depleted (more negative)
for the remainder of the Holocene (A). The d13Ccalcite values than those of the HCO, and the Cranberry Glades pollen
decrease between 7,000 and 2,000 cal. years B.P., which record is dominated by tree species that favor cooler,
records a general moistening or increasing precipitation (B). wetter climates (Fig. 4C–E).
The abrupt enrichment of d13Ccalcite values after 2,000 cal. Summarizing the stalagmite, paleopedological, and
years B.P. records landscape disturbances attributable to palynological data, the HCO was warmer and (seasonally?)
Native Americans (White, 2007; Springer et al., submitted). drier than the Late Holocene in the Greenbrier River
The warm temperatures inferred from d18Ocalcite values are watershed. Climate was comparatively stable throughout
consistent with low spruce (Picea) and pine (Pinus) pollen the Late Holocene (Fig. 4A), but Holocene values of
abundances at that time (C). However, increasing abundanc- d13Corg from alluvium in BCC and CAC are typical of
es of hickory (Carya), beech (Fagus), and oak (Quercus) are forested landscapes (224.5 6 1%) (White, 2007). Although
attributable to the moistening recorded in the d13Ccalcite forest composition evolves with time (Fig. 4C–E), d13Corg
record (D–E). Figure modified from Springer et al. (2008). values indicate that the watershed remained heavily
forested throughout the Mid- to Late-Holocene. Observed
(beech), and Quercus (oak)), pollen abundance from sub- changes in d13Ccalcite and d13Corg post-HCO are attribut-
boreal trees (principally Picea (spruce)) reach a Holocene able to changes in soil productivity and respiration, but not
low (Fig. 4), and pollen counts of the moisture sensitive attributable to major changes in C3 or C4 abundances
Family Cyperaceae (sedges) decline by several orders of (e.g., Dorale et al., 1998). However, the absence of pre-
magnitude and to zero for short intervals (not shown). HCO data prevents interpretation of vegetation types
Coincident with changes in Cyperaceae abundance is the during the Early Holocene and near the onset of the HCO.
appearance of Nyssa (tupelo) pollen in lacustrine sediments
found 100 km to the east of BCC-002 (Watts, 1979; Kneller HOLOCENE HYDROLOGY OF THE GREENBRIER RIVER
and Peteet, 1993). Tupelo prefers warm temperatures and is
mostly absent from regional pollen records, except during The Greenbrier River deposits slackwater sediments in
the mid-Holocene, when values peak. Collectively, d18Ocalcite CAC during intermittent floods. As a result, CAC contains
values and pollen abundances indicate a warm mid- a sedimentological record of stream hydrology. Sediments
124 N Journal of Cave and Karst Studies, August 2009
 G.S. SPRINGER, H.D. ROWE, B. HARDT, F.G. COCINA, R.L. EDWARDS, AND H. CHENG

Figure 5. Three major units can be recognized in the

slackwater sediments excavated in Colonial Acres Cave.
These consist of sands (pale lower unit), which are overlain
by dense, clayey silts (47 to 78 cm). The black, horizontal
line indicates their contact. The white line highlights the Figure 6. Sieving and pipette-withdrawal were used to
contact between the clayey silts and overlying sandy silts and determine the particle sizes of each significant bed in the
silty sands. All units are laminated and have undergone Colonial Acres Cave (CAC) trench, and these results are the
minimal bioturbation, although the lower sands underwent basis for this stratigraphic column. A clayey silt (48 to
soft sediment deformation as they subsided into an underly- 75 cm) separates the lower sands (75 cm to 118 cm) from the
ing clay unit. The latter is not visible in this photograph. younger sandy silts and silty sands (0 to 48 cm). Based upon
Contrast has been digitally enhanced to make the lamina- AMS-14C ages of charcoal (positions as indicated), deposi-
tions and bed contacts more easily visible. tion of the clayey silt ended at ,3,500 cal. years B.P., which
we attribute to incision of the Greenbrier River below CAC.
are excellently preserved and consist of laminated silts and Ultimately, the trench was dug to a depth of 200 cm, which
sands, which have undergone minimal bioturbation revealed soft clays (.118 cm) below the sand unit and a
(Figs. 5 and 6). However, the sedimentological record is basal layer of river cobbles (not depicted). Unfortunately,
censored. At present, the slackwater deposit is only repeated collapses of the lower trench face prevented detailed
inundated after river stage exceeds ,4 m and there are investigation of the clays and cobbles.
no high water marks correlative to individual slackwater
beds because the most recent flood obliterated the high 7). Short horizontal lines demarcate the boundaries
water mark of the previous flood, as that flood did to its between the units in Figure 5. The upper two units record
predecessor’s, ad infinitum. However, as will be shown, the very different hydrologic regimes. The capping unit of silts
censored record can be used to determine whether the cave and sandy silts contains partially articulated bat bones along
was permanently or only intermittently flooded at partic- with other vertebrate bones and organic detritus. Deposition
ular points in time. of these sediments began at ,3,500 cal. years B.P. and has
The slackwater sediments can be divided into three continued to present. The bat bones demonstrate that the
principle units: 0–47 cm, 47–78 cm, and .78 cm (Figs. 5– Greenbrier River has been below the observed slackwater
Journal of Cave and Karst Studies, August 2009 N 125
CLIMATE DRIVEN CHANGES IN RIVER CHANNEL MORPHOLOGY AND BASE LEVEL DURING THE HOLOCENE AND LATE PLEISTOCENE OF SOUTHEASTERN
WEST VIRGINIA

cave passages are permanently flooded and where sedi-

mentation rates are slow (Reams, 1968; Springer et al.,
1997). The corresponding depositional environment exists
below low-water river surfaces and its products are known
as backswamp or quietwater sediments (Springer et al.,
1997; Bosch and White, 2003).
The clayey silts were deposited while the low-water
surface of the Greenbrier River was higher than the cave
room. In addition to the backwater sedimentological
features, this conclusion is supported by the absence of
terrestrial animal bones, desiccation features (e.g., mud-
cracks), and organic detritus such as has been carried into
the cave by post-3,500 cal. years B.P. floodwaters. As an
alternative hypothesis, all conduits connecting the cave and
river could have been plugged during deposition of the
clayey silts. This could have produced sustained back-
flooding. However, we reject this idea because of the
improbability of every possible opening in a karstified
riverbank being plugged such that a flooded cave could exist
within meters of an open cliff. The simplest interpretation is
Figure 7. Paleoclimate proxies (d18Ocalcite and d13Ccalcite) that all openings were themselves permanently flooded,
from stalagmite BCC-002 (A and B) are compared to data which would only be possible if base level were higher than
obtained from the Colonial Acres Cave (CAC) slackwater present. Base-level rise is most commonly accomplished by
sediments (C and D). Note that sample ages from CAC are sediment infilling a channel (aggradation).
estimated below 4,000 cal. years B.P. using a sedimentation Laminated and cross-bedded sands containing some
rate calculated for the upper, age-dated portion of the trench. visible organic detritus underlie the clayey silts (Figs. 6 and
And only six d13Corg values were obtained from CAC 7). Cross-bed orientations indicate that flow entered the
sediments below 5,000 cal. years B.P. Ignoring the period of room from the 210-m-long entrance passage and sediments
land disturbance, d13Ccalcite and d13Corg values were most prograded across the room floor. Currently, the floor of the
enriched during the mid-Holocene and values decrease first 100-m of the entrance passage consists of cross-bedded
thereafter as the climate moistened (B and C). Submergence sands largely devoid of organic detritus, but overlain by
of CAC began prior to 5,000 cal. years B.P. and, assuming a woody flotsam. Presumably, the sands underlying the clayey
low sedimentation rate for the clayey silts (see text), silt in CAC were deposited in a manner similar to the
submergence probably began before the local Holocene modern sands, and were the product of episodic flooding
Climatic Optimum (HCO)(A). Deposition of the clayey silt above the low-water surface of the Greenbrier River.
ended ,3,500 cal. years B.P. (D), and the overlying silts and In conclusion, the low-water surface of the Greenbrier
sands contain abundant bat bones indicating incision of the River was below CAC during deposition of the lower
river below the cave. sands, but subsequently aggraded above the cave (Fig. 8).
The river re-incised below CAC at ,3,500 cal. years B.P.
and has remained below the cave since then. BCC-002 and
sediments since ,3,500 cal. years B.P. Today, there are CAC values of d13Ccalcite and d13Corg, respectively, decrease
extensive bedrock exposures in the channel banks and bed, after ,4,000 cal. years B.P. The enriched d13Ccalcite and
and this bedrock channel morphology has probably existed d13Corg values continue beyond the period of record, which
for much of the last 3,500 years. suggests that deposition of the lower sands occurred during
The clayey silts deposited prior to ,3,500 cal. years B.P. or prior to the HCO. Given the order of magnitude of the
are devoid of bones and visible organic matter is rare. previously calculated sedimentation rate, deposition of the
However, the clayey silts do contain insoluble cm-scale clayey silt probably began during the Early Holocene or
bedrock chips derived from thin claystone dikes in the cave the earliest part of the HCO. Thus, CAC slackwater
ceiling. Paleozoic, coarsely crystalline marine fossils, such sediments record a major aggradational event ($4 m) in
as blastoid thecas, are also present and derived from the the Greenbrier River during the Early Holocene.
host limestone. The coarsely crystalline fossils dissolve
slower than the fine-grained limestone and stand in relief SUMMARY DISCUSSION
on existing cave walls and ceilings. Shale fragments and
fossils found in the slackwater sediments fell or sank after The Greenbrier River aggraded no later than the Early
the surrounding limestone was dissolved. Concentration of Holocene and may have begun aggrading during the Late
low solubility particles is a slow process and occurs when Pleistocene. The implied increase in the sediment supply
126 N Journal of Cave and Karst Studies, August 2009
 G.S. SPRINGER, H.D. ROWE, B. HARDT, F.G. COCINA, R.L. EDWARDS, AND H. CHENG

climate was very dry (Kneller and Peteer, 1993). Therefore,

it is entirely possible that the enhanced sediment produc-
tion was not matched by an increase in the sediment
transport capacity of regional streams and rivers. Under
such a scenario, excess colluvium would have accumulated
throughout much of the Greenbrier River watershed prior
to the Holocene. The observed post-Pleistocene moistening
of the region would have necessarily led to low-order
streams remobilizing the accumulated regolith (cf. Eaton et
al., 2003) and a large influx of sediment into the Greenbrier
River. This sediment influx is probably the cause of the
aggradational event recorded in CAC sediments.
We lack the means to establish the depth of infilling of
the Greenbrier River channel at CAC. But the distinctive
clayey silts associated with submergence are found 4 m
above the low water surface of the modern Greenbrier
River, which establishes a minimum depth of aggradation.
Such infilling would have resulted in an alluvial channel
morphology (Fig. 8), although bedrock exposure in pools
Figure 8. The sequence of events we interpret to have cannot be ruled out. Significantly, the river began to incise
occurred begins with Colonial Acres Cave (CAC) lying through the accumulated sediments sometime during the
above the Greenbrier River, which presumably had a bedrock Holocene Climatic Optimum (HCO) and incised below
channel morphology (upper diagram). Subsequently, the CAC by 3,500 cal. years B.P. Climate has moistened since
channel aggraded as a wetter Holocene climate allowed the HCO (Figs. 4 and 7), which has presumably led to an
Pleistocene-age colluvium to be remobilized and carried into increase in sediment transport capacity.
the river where transport capacity was initially less than So, how will the Greenbrier River respond to Global
sediment input, resulting in aggradation. The channel Warming? Climate modelers predict (locally) decreasing
infilling may have created an alluvial channel morphology runoff as Global Warming advances (Arnell, 2003). If we
(bottom diagram). Eventually, the sediment wave was assume that decreasing runoff will lessen sediment trans-
transmitted downstream and the river incised below CAC port capacity in the Greenbrier River, we must conclude
during the Late Holocene (upper diagram). This sequence of that there is the potential for aggradation and morpholog-
events is known to have occurred elsewhere in the ical change. However, it is possible that heavy or extreme
Appalachians (Eaton et al., 2003). precipitation events may become more common as a result
of Global Warming (Emori and Brown, 2005). If so, this
could result in a net increase in transport capacity and no
may be the result of the Late Pleistocene-Holocene climatic aggradation. Thus, it is premature to predict aggradation
transition. The Wisconsin Glaciation peaked ,20,000 cal. or incision.
years B.P., and evidence for periglacial activity during the Uncertainties about the future aside, the history of the
Wisconsin of the Appalachian and Blue Ridge Mountains Greenbrier River offers important insights into the long-
is widespread (Gardner et al., 1991; Eaton et al., 2003; term behavior of Appalachian rivers and hillslopes. Late
Mills, 2005; Nelson, et al., 2007). The associated pre- Pleistocene climate variability resulted in significant
Holocene sedimentary deposits are commonly preserved at channel infilling. The accumulated sediments were not
the bases of hillslopes where they are susceptible to fully excavated from the Greenbrier River channel until as
remobilization by streams. The deposits reached valley late as 3,500 cal. years B.P., or roughly 16,000 years after
bottoms via slope wash, solifluction, and gelifluction while the last glacial maxima (Webb, 1987). Interestingly, the
climate was considerably cooler than present. The lower Greenbrier River was already incised to its present
elevation limit for such periglacial activity at Mountain elevation well before the last glacial maxima, and it is
Lake, Virginia (50 km southeast of CAC) was ,800 m probable that very little net incision has occurred in tens of
amsl (Nelson et al., 2007). Most ridges and mountaintops millennia. Incision may occur episodically during and
within the Greenbrier River watershed exceed that immediately following interglacials, when precipitation
elevation, and well-developed periglacial landforms are totals are high or heavy precipitation events more
present at 650 m amsl in the Buckeye Creek watershed. common. This raises many questions concerning the
Appalachian hillslope sediment production was en- usefulness of long-term incision rates calculated using
hanced by periglaciation during the Late Pleistocene paleomagnetic and cosmogenic isotopic data. These latter
(Gardner et al., 1991; Eaton et al., 2003; Mills, 2005), but estimates require averaging incision over many hundreds to
regional palynology indicates that the Late Wisconsin thousands of millennia (e.g., Springer et al., 1997; Anthony
Journal of Cave and Karst Studies, August 2009 N 127
CLIMATE DRIVEN CHANGES IN RIVER CHANNEL MORPHOLOGY AND BASE LEVEL DURING THE HOLOCENE AND LATE PLEISTOCENE OF SOUTHEASTERN
WEST VIRGINIA

and Granger, 2004; Anthony and Granger, 2006), so their Dorale, J.A., Edwards, R.L., González, L., and Ito, E., 1998, Climate and
vegetation history of the midcontinent from 75 to 25 ka: A speleothem
estimates of actual millennial-scale incision rates may be record from Crevice Cave, Missouri, USA: Science, v. 282, p. 1871–
off by many orders of magnitudes. 1874.
Driese, S.G., Li, Z-H., and Horn, S.P., 2005, Late Pleistocene and
Holocene climate and geomorphic histories as interpreted from a
FUTURE DIRECTIONS 23,000 14C yr B.P. paleosol and floodplain soils, southeastern West
Virginia, USA: Quaternary Research, v. 63, p. 136–149.
Stable isotopic records of paleoclimate from spe- Eaton, L.S., Morgan, B.A., Kochel, R.C., and Howard, A.D., 2003,
Quaternary deposits and landscape evolution of the central Blue
leothems are rapidly becoming available throughout the Ridge of Virginia: Geomorphology, v. 56, p. 139–154.
world. As our study demonstrates, such records can be Edwards, R.L., Chen, J.H., and Wasserburg, G.J., 1987, 238U–234U–230Th
combined with traditional studies of slackwater stratigra- systematics and the precise measurement of time over the last 500,000
years: Earth and Planetary Science Letters, v. 81, p. 175–192.
phy to infer past effects of climate change on rivers. As
Emori, S., and Brown, S.J., 2005, Dynamic and thermodynamic changes
such, there will soon be many opportunities to directly in mean and extreme precipitation under changed climate: Geophysical
determine relationships between climate and hillslope and Research Letters, v. 32, no. L17706 p. doi: 10.1029/2005GL023272.
fluvial processes. However, such studies will be dependent Gardner, T.W., Ritter, J.B., Shuman, C.A., Bell, J.C., Sasowsky, K.C.,
and Pinter, N., 1991, Stratified slope deposit in the Valley and Ridge
upon fortuitously situated caves capable of recording the Province of central Pennsylvania, USA: Sedimentology, stratigraphy,
behavior of adjacent surface channels because surficial and geomorphic evolution: Permafrost and Periglacial Processes, v. 2,
deposits are rapidly destroyed in many climatic settings p. 141–162.
Hendy, C.H., 1971, The isotopic geochemistry of speleothems 1: The
(Kite et al., 2002; Springer, 2002). Combined with the calculation of the effects of the different modes of formation on the
subterranean origin of stalagmitic paleoclimate records, we isotopic composition of speleothems and their applicability as
foresee karst studies achieving prominent roles in geomor- palaeoclimatic indicators: Geochimica et Cosmochimica Acta, v. 35,
p. 801–824.
phology and climatology and encourage others to advance Kaplan, M.R., and Wolfe, A.P., 2006, Spatial and temporal variability of
our methodology. Holocene temperature in the North Atlantic region: Quaternary
Research, v. 65, p. 223–231.
ACKNOWLEDGMENTS Kirby, M.E., Mullins, H.T., Patterson, W.P., and Burnett, A.W., 2002,
Late glacial–Holocene atmospheric circulation and precipitation in
the northeast United States inferred from modern calibrated stable
The authors thank Tom Gardner and an anonymous oxygen and carbon isotopes: Geological Society of America Bulletin,
reviewer for helpful suggestions that improved the manu- v. 114, p. 1326–1340.
Kite, J.S., Gebhardt, T.W., and Springer, G.S., 2002, Slackwater deposits
script. We thank Gene Turner for access to Buckeye Creek as paleostage indicators in canyon reaches of the central Appala-
Cave and the West Virginia Association for Cave Studies chians: Reevaluation after the 1996 Cheat River Flood, in House, K.,
for fieldwork assistance. Webb, R., Baker, V., and Levish, D., eds., Ancient floods, modern
hazards: Principles and applications of paleoflood hydrology: Water
Science and Application Volume 5 (monograph), Washington, D.C.,
REFERENCES American Geophysical Union, p. 257–266.
Kneller, M., and Peteet, D.M., 1993, Late-Quaternary climate in the ridge
Anthony, D.M., and Granger, D.E., 2004, A Late Tertiary origin for and valley of Virginia, U.S.A.: changes in vegetation and depositional
multilevel caves along the western escarpment of the Cumberland environment: Quaternary Science Reviews, v. 12, p. 613–628.
Plateau, Tennessee and Kentucky, established by cosmogenic 26Al and Knighton, D., 1998, Fluvial forms and processes: A new perspective, New
10
Be: Journal of Cave and Karst Studies, v. 66, p. 46–55. York, Oxford University Press, 400 p.
Anthony, D.W., and Granger, D.E., 2006, A new chronology for the age Kundzewicz, Z.W., Mata, L.J., Arnell, N.W., Döll, P., Kabat, P., Jiménez,
of Appalachian erosional surfaces determined by cosmogenic nuclides B., Miller, K.A., Oki, T., Senand, Z., and Shiklomanov, I.A., 2007,
in cave sediments: Earth Surface Processes and Landforms, v. 32, Freshwater resources and their management, in Parry, M.L., and
p. 874–887. doi: 10.1002/esp.1446. Canziani, O.F., et al., eds., Climate Change 2007, Impacts,
Arnell, N.W., 2003, Effects of IPCC SRES emissions scenarios on river Adaptation and Vulnerability. Contribution of Working Group II
runoff: a global perspective: Hydrological Earth Systems Science, v. 7, to the Fourth Assessment Report of the Intergovernmental Panel on
p. 619–641. Climate Change, Cambridge, UK, Cambridge University Press,
Booth, R.K., Jackson, S.T., Forman, S.L., Kutzbach, J.E., Bettis, E.A. p. 173–210.
III, Kreigs, J., and Wright, D.K., 2005, A severe centennial-scale Lettenmaier, D.P., de Roo, A., and Lawford, R., 2006, Towards a
drought in midcontinental North America 4200 years ago and capability for global flood forecasting: WMO Bulletin, v. 55,
apparent global linkages: The Holocene, v. 15, p. 321–328. p. 185–190.
Bosch, R.F., and White, W.B., 2003, Lithofacies and transport of clastic Li, Y.X., Yu, Z.C., and Kodama, K.P., 2007, Sensitive moisture response
sediments in karstic aquifers, in Sasowsky, I.D., and Mylroie, J.E., to Holocene millennial-scale climate variations in the Mid-Atlantic
eds., Studies of cave sediments, New York, Kluwer Academic and region, USA: The Holocene, v. 17, p. 1–6.
Plenum Press, p. 1–22. Maurer, E.P., Lettenmaier, D.P., and Mantua, N.J., 2004, Variability and
Broecker, W.S., 1963, Preliminary evaluation of uranium series inequili- potential sources of predictability of North American runoff: Water
brium as a tool for absolute age measurement on marine carbonates: Resources Research, v. 40, no. W09306
Journal of Geophysical Research 68, p. 2817–2834. McDermott, F., 2004, Palaeo-climate reconstruction from stable isotope
Bull, W.B., 1988, Floods; degradation and aggradation, in Baker, V.R., variations in speleothems: A review: Quaternary Science Reviews,
Kochel, R.C., and Patton, P.C., eds., Flood Geomorphology, New v. 23, p. 901–918.
York, Wiley, p. 157–165. Mills, H.H., 2005, Relative age dating of transported regolith and
Cheng, Hai, Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., and application to the study of landform evolution in the Appalachians:
Asmerom, Y., 2000, The half-lives of uranium-234 and thorium-230: Geomorphology, v. 67, p. 63–96.
Chemical Geology, v. 169, no. 1–2, p. 17–33. Montgomery, D.R., and Buffington, J.M., 1997, Channel-reach morphol-
Dansgaard, W., 1964, Stable isotopes in precipitation: Tellus, v. 16, ogy in mountain drainage basins: Geological Society of America
p. 436–468. Bulletin, v. 109, p. 596–611.

128 N Journal of Cave and Karst Studies, August 2009

 G.S. SPRINGER, H.D. ROWE, B. HARDT, F.G. COCINA, R.L. EDWARDS, AND H. CHENG

Nelson, K.J.P., Nelson, F.E., and Walegur, M.T., 2007, Periglacial Springer, G.S., Rowe, H.D., Hardt, B., Edwards, R.L., and Cheng, Hai,
Appalachia: Palaeoclimatic significance of blockfield elevation gradi- 2008, Solar forcing of Holocene droughts in a stalagmite record from
ents, eastern USA: Permafrost and Periglacial Processes, v. 18, West Virginia in East-Central North America: Geophysical Research
p. 61–73. Letters, v. 35, no. L17703, doi: 10.1029/2008GL034971.
Reams, M.W., 1968, Cave sediments and the geomorphic history of the Stover, S.C., and Montgomery, D.R., 2001, Channel change and flooding,
Ozarks [PhD thesis], St. Louis, Mo., Washington University, 167 p. Skokomish River, Washington: Journal of Hydrology, v. 243,
Rozanski, K., Araguás-Araguás, L., and Gonfiantini, R., 1993, Isotopic p. 272–286. doi: 10.1016/S0022-1694(00)00421-2.
patterns in modern global precipitation, in Swart, P.K., et al., eds., Wang, Y.J., Cheng, Hai, Edwards, R.L., An, Z.S., Wu, J.Y., Shen, C.-C.,
Climate Change in Continental Isotopic Records, Geophysical and Dorale, J.A., 2001, a high-resolution absolute-dated Late
Monograph, Series, v. 78, Washington, D.C., American Geophysical Pleistocene monsoon record from Hulu Cave, China: Science,
Union, p. 1–36. v. 294, no. 5550, p. 2345–2348.
Schumm, S.A., 1973, Geomorphic thresholds and complex response of Watts, W.A., 1979, Late Quaternary vegetation of central Appalachia and
drainage systems, in Morisawa, M., ed., Fluvial geomorphology, the New Jersey coastal plain: Ecological Monographs, v. 49, p. 427–469.
Binghamton, New York State University Publications in Geomor- Webb, T. III., 1987, Climate change in eastern North America during the
phology, p. 299–310. past 18,000 years: comparisons of pollen data with model results, in
Shen, C.C., Edwards, R.L., Cheng, Hai, Dorale, J.A., Thomas, R.B., Ruddiman, W.F., and Wright, H.E., Jr., eds., The Geology of North
Moran, S.B., Weinstein, S.E., and Hirschmann, M., 2002, Uranium America, v. K-3, Boulder, Colo., Geological Society of America,
and thorium isotopic and concentration measurements by magnetic p. 447–462.
sector inductively coupled plasma mass spectrometry: Chemical White, D.M., 2007, Reconstruction and analysis of Native American land
Geology, v. 185, p. 165–178. use during the late Holocene [MS thesis], Athens, OH, Ohio
Springer, G.S., 2002, Caves and their potential use in paleoflood studies, University, 164 p.
in House, K., Webb, R., Baker, V., and Levish, D., eds., Ancient Willard, D.A., Bernhardt, C.E., Korejwo, D.A., and Meyers, S.R., 2005,
floods, modern hazards: Principles and applications of paleoflood Impact of millennial-scale Holocene climate variability on eastern
hydrology, Water Science and Application Volume 5 (monograph), North American terrestrial ecosystems: pollen-based climatic recon-
Washington, D.C., American Geophysical Union, p. 329–344. struction: Global and Planetary Change, v. 47, p. 17–35.
Springer, G.S., Kite, J.S., and Schmidt, V.A., 1997, Cave sedimentation, Wood, A.W., Maurer, E.P., Kumar, A., and Lettenmaier, D.P., 2002,
genesis, and erosional history in the Cheat River Canyon, Long range experimental hydrologic forecasting for the eastern U.S.:
West Virginia: Geological Society of America Bulletin, v. 109, Journal of Geophysical Research, v. 107, no. D20, p. 4429, doi:
p. 524–532. 10.1029/2001JD000659.

Journal of Cave and Karst Studies, August 2009 N 129