STRUCTURAL AND STRATIGRAPHIC INFLUENCES 
ON THE DEVELOPMENT OF 
SOLUTION CONDUITS
IN THE UPPER ELK RIVER VALLEY, 
WEST VIRGINIA

The upper Elk River Valley is located in northern Pocahontas and southern Randolph Counties, West Virginia. For over 8 km (5 miles), this valley is floored with Mississippian Greenbrier Group limestones dipping gently to the west. Upon reaching the Union Limestone near the top of the Greenbrier Group, the Elk sinks and then rises at two sets of occluded riverbank springs, also at the top of the Union Limestone and 8 km (S miles) to the north. Over 29 km (18 miles) of surveyed cave passages, seen in and adjacent to the upper Elk River Valley, contain streams flowing for up to 12.9 linear kilometers (8 linear miles) and 244 vertical meters (800 feet) between sink and rise. Half of these passages are developed along joints that follow a NE-SW trending fracture trace carrying drainage from an adjacent river basin to the upper Elk River. Where the fracture trace crosses the Elk River, the river sinks, drops 38 vertical meters (125 feet) through both the upper Greenbrier Group limestones and the shaley Taggard Formation below (normally a major aquitard). The remaining passages are seen in several caves that parallel the Elk River Valley to the north of the fracture trace. These caves, extending over a 3.2 km (2 mile) linear distance along and beneath the Elk’s Valley, underdrain it and consist of solutionally enlarged beds found at several distinct stratigraphic horizons. The underground Elk, seen in two of these caves, flows beneath the Taggard Shales until the elevation of the shales passes beneath that of the lower set of springs. This paper discusses the relationships between these two patterns of caves, outlines a sequence of cave development for this area, and discusses the nature of the underground flow paths of the Elk River and its tributaries. 

Document Outline

Introduction And Background
Methods Used
Physiographic And Geological Setting
Stratigraphy 
Local Karst Hydrology
The Sinking Of The Elk River
Caves In The Elk River Valley
        Simmons Mingo Cave
        Falling Springs Cave
        Elk River Cave
        Bradshaw Run Cave
        Left It Pit
Cave Development
Hydrological Relationships
Acknowledgements
References
Publication History   

INTRODUCTION AND BACKGROUND 

 The Elk River rises in northern Pocahontas and southern Randolph Counties in eastern West Virginia (Fig. 1), flows to the west, and drains into the Kanawha River at Charleston, West Virginia. Speculation about the existence of caves beneath the bed of the upper Elk River has occurred since the late 19th century. In 1898, Hu Maxwell, a West Virginia historian wrote: 

‘‘Theory and all known facts lead to the conclusion that a cave of enormous dimensions exists in Randolph County under or near the course of the Elk River between the Pocahontas County line and the mouth of Valley Fork six miles below. But no one has ever yet found an entrance into the cave, and its existence cannot be positively affirmed. The facts which are explained on the theory of a vast cave are these: Elk River, except in time of freshet, flows into a crevice at the foot of a mountain, or when very low, disap­pears among the boulders of its channel . . . and six miles below, the water rushes to the surface. Its underground course is through limestone and it must flow through galleries of large size. In 1896, near the point where the water sinks, a portion of the river bottom dropped down, leaving an opening about 15 feet square into which the whole river plunged and disappeared. No bottom was visi­ble, and no one attempted to enter or examine. The next flood filled the opening with boulders.” (Maxwell, 1898).

Figure 1

In spite of years of speculation and searching, no caves were known to exist beneath the bed of the Elk River. In­deed, only two sizeable caves approached the valley: the downstream end of the 12.9 km (8 mi) long Simmons Mingo I My Cave complex and Falling Springs Cave, devel­oped beneath Falling Springs Run, an infeeder to the Elk River. Since 1981, over 9.7 km (6 mi) of passage have been explored and surveyed in three newly discovered caves which parallel the surface bed of the Elk and which are developed up to 40 vertical meters (130 ft) beneath it. Also, the length and depth of Falling Springs Cave have been ex­tended and, in one of the newly discovered caves, a segment of the underground Elk River has (finally) been found flow­ing in accessible passage for a distance of about 800 m (0.5 mi).  In this paper, we discuss stratigraphic and structural in­fluences on both the development and orientation of the Elk River valley caves, postulate a sequence of cave develop­ment for this area, and discuss the nature of the under­ground flow path taken by the underground Elk under both low and high flow conditions.

METHODS USED

 As part of the work carried out in preparing this paper, over 8 km (5 mi) of surface surveys have been carried out to determine relative positions of cave entrances, the bed of the Elk River, springs, the top of the Union Limestone and other karst features with respect to the locations of several U.S. Coast and Geodetic Survey benchmarks in the Elk River valley. These surveys, carried out with handheld and tripod mounted Brunton and Suunto compasses and clino­meters, and fiberglass tapes, were repeated as necessary to obtain consistent results. Elevations were also obtained with an altimeter (accuracy ± 2 feet) calibrated at one of the benchmarks. Changes in atmospheric pressure, obtained with a recording barograph at the benchmark, were taken into account in adjusting the altimeter readings. The mean deviation between elevation differences obtained via surveys and altimeter readings was 4.0 feet. Surveys conducted in the caves were with handheld Suunto compasses and clinometers and fiberglass tapes. Survey loops were closed using the Survey Manipulation, Analysis and Plotting System (SMAPS) software package developed for cave sur­veys. The mean closure error for survey loops in the caves was 0.8 percent.
Limestone thicknesses were measured both on the surface and in caves as part of the surveys. Where contacts at the top and bottom of major members of the Greenbrier Group were crossed in the caves, thicknesses were computed taking strike and dip into account. Numerous strike and dip readings were taken throughout the study area. Estimates of regional strike and dip were also obtained by taking the surveyed coordinates of widely spaced points at the top of the Union Limestone, and, using least squares technique, fitting a plane to these points. The strike of the fitted plane differed from the mean of the observed strike readings by 3 degrees while the dip of this plane differed from the mean of the observed dip readings by 0.25 degrees. 
Stream tracing was carried out using sodium fluorescein dye and activated charcoal detectors. Adsorbed dye was elutriated from the charcoal using a 10 percent solution of KOH in ethanol. Testing for the presence of dye was carried out using a Turner model 111 fluorimeter.

PHYSIOGRAPHIC AND GEOLOGICAL SETTING 

The Elk River rises on clastic rocks in the Allegheny Plateaus Province in eastern West Virginia. Relief in the area is 500-700 m (1700-2300 ft) with ridge tops reaching elevations of 1280-1380 m (4200-4500 ft) and consisting of rocks of the (Upper Mississippian) Mauch Chunk and (Lower Pennsylvanian) Pottsville Formations. In the study area, the Elk River ranges from 760-690 m (2500-2260 ft) in elevation, flows north-northwest through a narrow, steep-sided valley and has a gradient of 7.6 meters per kilometer (40 feet per mile), both on the Greenbrier Group carbonates and on the clastics above and below.

Fig. 2a. Limestone pavement in bed of Elk River

Fig. 2b Dry bed of Elk River

Fig. 2c Elk River (same vantage point) in flood

The oldest rocks exposed in the area are those of the (Mid­dle Mississippian) Greenbrier Group, a sequence of limestones containing thin interbedded layers of shales and sandstones. Locally, the Greenbrier Group varies in thick­ness from 92-104 m (300-340 ft) although only the upper 25-30 m (80-100 ft) of the Greenbrier are exposed in and im­mediately adjacent to the bed of the Elk River and its tributaries. A general view of the dry bed of the Elk River is shown in Figure 2. In the study area, the measured dip varies between 1.0 and 1.5 degrees and the strike varies between due north and N.SE. A map of that part of the Elk River valley that is discussed in this paper is given in Figure 3. The shaded area represents exposures of the Union Limestone, the highest major cave-forming unit in the Greenbrier Group. The letters refer to those major karst features in this area that are discussed in the text of this paper. Faulting in the valley consists of low angle thrusts striking north-south and having minimal displacement. These faults modestly influence the directional orientation of the caves but have a greater influence on passage morphology. Where faults are encountered, passage walls! ceilings generally follow the fault plane. A substantial fracture zone/linea­ment trending N58E crosses the Elk River valley at the loca­tion where much of the Elk sinks. This feature (Fig. 3, L) as will be noted below, plays a major role in influencing the location of the underground Elk in the Greenbrier Group limestones.

Figure 3. Limestone exposures and other major features in the Elk River area.

STRATIGRAPHY 

In the Elk River valley, the stratigraphic sequence of limestones and interbedded clastics in the Greenbrier Group influences the placement of cave passages in the limestones. In descending stratigraphic order, these members are described below. Descriptions and thicknesses are based on Reger (1931), Wells (1950), and the authors’ measurements in the area. A column is given in Figure 4.

LOCAL KARST HYDROLOGY 

 For a distance of almost 10 km (6 mi), the valley of the Elk River is floored with Mississippian limestones of the Greenbrier Group. Where it crosses into the upper Union Limestone, the Elk River sinks in its bed (Fig. 3, ERB and BH) and for much of the year the entire flow of the Elk is underground. The Elk rises 8.8 km (5.5 mi) to the north at a series of alluviated springs at river level where the top of the Union Limestone passes beneath the streambed (Fig. 3, MS). Under higher flow conditions, some of the Elk also rises at another set of springs (Fig. 3, HS), 1.6 km (1 mi) up­stream of the main springs and 12 m (40 ft) above them. Draining an area of 238 sq km (92 mi2) and having a measured discharge varying between 0.3 and 7 m3 I sec (11 and 250 cubic feet per second) at the springs, depending on flow conditions, the upper Elk River may be the largest sinking stream in West Virginia. 

A detailed picture of karst hydrology in the entire Elk River basin was given in an earlier paper (Medville, 1977) and is briefly summarized below. The carbonate sequence in the upper Elk River valley is a free-flow aquifer with cap­ping and perching beds as described by White (1969). This aquifer is anisotropic with subsurface flow concentrated in enlarged bedding plane partings and joints. The direction of subsurface flow is controlled by regional structure with the hydraulic gradient following the regional dip to the north­west. In very general terms, streams in the area sink in the upper 9 m (30 ft) of the Union Limestone, drop through the Union and lower limestones, and then rise stratigraphically back to the top of the Union where this horizon passes beneath the bed of the Elk River at the downstream (northern) end of the study area. A more detailed examination of the nature of the subsurface flow of the Elk River between its sink point and its rising is one of the subjects of this paper.

THE SINKING OF THE ELK RIVER 

At the town of Slaty Fork, West Virginia, two streams; Old Field Fork and Big Springs Fork, join, the combined flow designated as “Elk River.” At this locality, the Elk is flowing on the Greenville Shale and higher clastics and con­tinues to do so for 4 km (2.5 mi) as it flows north. The river then reaches the upper Union Limestone and becomes a los­ing stream. Within 800 m (a half mile) of reaching the Union and near the mouth of Blackhole Run, an infeeder from the west, the Elk crosses what has been previously described as a lineament or fracture zone (Medville, 1977) trending NS8E. At this point the north-flowing Elk turns and follows the lineament to the northeast for a few hundred meters before resuming its northerly course. 

Where the river crosses the lineament (Fig. 3, BH), a substantial amount of its remaining water flows into open bedding plane partings and joints along its east bank at an elevation of 760 m (2490 ft). The largest of these openings, Black Hole Cave, is only a few meters long and is almost en­tirely choked with logs and other surface debris. The quanti­ty of water sinking here varies with flow conditions and with the changing ability of Black Hole Cave to accept water. This small cave may be the “crevice” noted by Maxwell (1898) although other ephemeral swallets downstream of Blackhole Run have been seen. In drought, all of the Elk River will sink in its bed 400 meters upstream of Black Hole Cave (Fig. 3, ERB), while in flood, the cave’s entrance is submerged beneath several meters of water and the inflow to this cave, while substantial, will be unnoticeable. 

The water sinking at Black Hole Cave flows to the north­east for 180 m (600 ft) along the lineament and is seen again at the downstream end of the Simmons Mingo/My Cave System (described below) where it appears as a 6 m (20 ft) waterfall emerging from the top of the Upper Taggard Shale. The water then enters a sump (The Crayfish Pool, elev. 721 m (2366 ft) in the upper Patton Limestone and is 38 m (124 ft) lower in elevation than the Elk River bed at Black Hole Cave. The waterfall, Taggard Formation, and Crayfish Pool below it are shown in Figure 5. Dye placed at the Black Hole entrance appears at this waterfall within 15 minutes. 

Based on observations of brecciated zones in passage ceil­ings and occasional slickensides, the lineament along which the Simmons Mingo / My Cave System is developed is hy­pothesized to be a right lateral. strike-slip fault (Mylroie, personal communication, 1986). The lineament/fault is significant in that it permits solution to occur at a substan­tial vertical distance beneath the bed of the Elk River and indeed, permits the Elk to flow beneath the Taggard Shales and into the upper Patton Limestone. The hydrogeological problem encountered involves the subsequent flow regime and passage morphology of the underground Elk River be­tween the Crayfish Pool in My Cave and its rising at the springs 8.7 km (5.4 mi) to the northwest. With a spring elevation of 689 m (2260 ft), the vertical separation between the two points is only about 30 m (100 ft). 

With respect to the flow path taken by the underground Elk River, four possibilities exist:

(a) The underground Elk River, flowing at a uniform gra­dient of about 4 in/km (20 ft/mi), gradually climbs stratigraphically, passing back through the Taggard Shales from below within 1600 m (1 mile) of the Crayfish Pool. All passage to the north of this point would be found in the Union and Pickaway Limestones. 

(b) The underground Elk remains below the Taggard Shales, paralleling the surface gradient of the river bed; about 8 m / km (40 ft/mi). At a point where the plane of the Taggard Shales passes beneath the elevation of the Elk River springs (about 4 km [2.5 mu upstream of the springs), the underground Elk crosses the Taggard and then flows with a negligible gradient toward the springs. 

(c) An intermediate pattern exists in which the under­ground Elk passes through the Taggard Shales several times; i.e., a “bumpy path” hypothesis. 

(d) The underground Elk River flows beneath the Taggard Shales to an elevation lower than that of the springs and then rises up as a phreatic lift in the vicinity of the springs. 

Since the Elk River is the major base level stream for almost a 260 sq km (100 sq. mi.) area and since its subsurface gra­dient may define the top of the saturated zone, the nature of this gradient is of some interest. A case could be made for each of the four hypotheses given above. An associated ex­ploration problem is to find, if possible, the underground Elk River in one or more places between the downstream end of My Cave and the springs. This problem is com­pounded by the facts that the underground Elk flows at depth beneath its bed and thus may be inaccessible, and that it may flow in recently developed, low gradient conduits, which, if not entirely water filled, could be prone to flooding. 

In the remainder of this paper, we discuss the nature of the caves which have been found in the Elk River valley, stratigraphic and structural influences on the locations of these caves in the Greenbrier Group limestones, the relationships between these caves, and the observed subsurface flow of the Elk River where it is seen in these caves.

CAVES IN THE ELK RIVER VALLEY

To date, over 29 km (18 mi) of passage have been surveyed in five major caves in and adjacent to the Elk River valley (Fig. 6). These caves are significant in that they con­tain current and paleoflow routes for the underground Elk River. In this section, the basic patterns of these caves are discussed with emphasis placed on stratigraphic and struc­tural controls on passage orientation, placement of passages in the Greenbrier Group limestones, and passage mor­phology.

SIMMONS MINGO/MY CAVE

The Simmons Mingo/My Cave System (Fig. 7) is the longest and deepest of the Elk River area caves, extending along the lineament for 4300 m (14,000 ft) and having about 210 m (700 ft) of relief. At least 13 km (8 mi) of passage have been surveyed in this cave by members of the Potomac Speleological Club since the mid1960s. Systematic explora­tion associated with the surveying has resulted in the connec­tion of Simmons Mingo and My Caves by divers in February, 1978 and in the discovery of new entrances to the cave. A comprehensive account of the history of exploration of this system is given by Swicegood (1982). 

Although only the downstream (My Cave) end of this system approaches the Elk River valley, the entire Simmons Mingo /My Cave System is hydrologically significant in that it intercepts the underground Dry Branch (Fig. 6, DB), the largest infeeder to the Elk River in the study area, draining 32.4 sq km (12.5 mi2); and that it diverts water from the Tygart River drainage, 6.4 km (4 mi) to the east of the Elk River. 

Fig. 5. Waterfall breaching Taggard formation at downstream end of Simmons Mingo/ My Cave System.

The Elk River entrance to My Cave (Fig. 7, ME) is 6 m (20 ft) below the top of the Union Limestone and over 15 m (50 ft) higher than the bed of the Elk River. This entrance opens to the top of a 12 m (40 ft) paleotrunk, the floor of which descends to the elevation of the Elk River bed. Solution scallops along the walls of this passage indicate that the Elk once flowed into this entrance and then to the northeast, following the joint set along which this part of the cave is developed (W. White, pers. comm.). This large entrance passage and an extension at its ceiling level can be followed for several hundred meters to the northeast to the top of a 12 m (40 ft) wide canyon with its floor over 24 m (80 ft) below. A stream at the base of this canyon, having a base flow of 0.1 cms (3 cfs), flows from the northeast and represents the Simmons Mingo drainage, the underground Dry Branch, and local infeeders to My Cave. As noted above, this canyon breaches the Taggard Formation and terminates downstream at a sump where the water from Black Hole Cave also enters.

The canyon, up to 30 m (100 ft) high, can be followed upstream for about 300 m (1000 ft) to junctions with active local infeeders and fossil drainage routes from beneath the surface Dry Branch. In this area, the cave stream drops rapidly for almost 30 m (100 vertical ft), from the lower Union Limestone to the bottom of the Pickaway Limestone.   
Continuing upstream in the Union, the combined Simmons Mingo / underground Dry Branch flow can be followed for over 300 m (1000 ft) before terminating at a series of sumps, the other side of which is the historical downstream end of Simmons Mingo Cave. The entire length of Simmons Mingo Cave is developed along the lineament (except for a short north-south trending section where the underground Dry Branch enters from the south via a deep sump). The main passage in this cave, following the lineament, has over 150 m (500 ft) of relief and is stratigraphically significant in that it extends from the top of the Union Limestone to the bottom of the basal Sinks Grove Limestone. The cave’s profile (Fig. 7) illustrates the influence of various members of the Greenbrier Group as perching and capping beds for the cave’s higher level passages and the gradation of the lowest (stream level) passage toward the Elk River valley.

These passages as well as the one containing the cave’s stream then climb stratigraphically into the Union Limestone near the cave’s historical downstream end at the Simmons Mingo / My Cave sump beneath the Dry Branch valley. While the cave stream in this vicinity flows through comparatively small and submerged passages, a much larger, low-gradient, paleotrunk parallels this stream, 20-27 m (70 to 90 ft) above it. This passage also climbs stratigraphically, crosses the aquitards from below, and grades toward the upper Union Limestone where it is ex­posed in the lower Dry Branch Valley. A more detailed discussion of stratigraphic controls on the development of the Simmons Mingo I My Cave System and of the significance of the paleotrunk is presented in the Cave Develop­ment section of this paper.

FALLING SPRINGS CAVE 

Falling Springs Cave (Figs. 8 and 9), is one of the most complex of the caves described in this paper and has a surveyed length of over 4200 m (14,000 ft). In contrast to the more recently discovered Elk River valley caves, Falling Springs Cave was known and partially explored in the late 1800’s. The cave was described by Maxwell (1898) who wrote:

“This interesting series of pits, galleries and rooms is a combination of a cave and sinkhole. Falling Spring Run heads against Mingo Knob and Elk Mountain and after flowing one and a half miles and receiving numerous tributaries which makes it a stream of considerable size, it approaches within a quarter mile of Elk River where it plunges into a yawning gulf 200 feet in circumference and 40 feet deep, and the water is seen no more. It enters a gallery from the bottom of the pit and is supposed to reach the subterranean channel of Elk River, but exploration has not yet established this as a fact.”

More recent exploration was carried out in the mid 1960’s by Schmidt (1965), who partially mapped the cave and by the Potomac Speleological Club in the early 1970’s. While Maxwell speculated that Falling Springs Cave could lead to the hypothetical underground Elk River, Schmidt’s map gave little indication of its depth, lateral extent or complex­ity. In order to determine these and to understand the rela­tionship, if any, between this cave and the other caves in the Elk River valley, a resurvey was undertaken by the authors and others in 1981. Because of the complexity of the cave, Falling Springs is still not completely explored or surveyed.

Figure 8. Plan view of Falling Springs Cave

Figure 9. Profile view of Falling Springs Cave looking along strike

Falling Springs Run is an eastern infeeder to the Elk River, draining about 6.5 sq km (2.5 sq mi) on the west side of Mingo Knob. This stream reaches the top of the Union Limestone 550 m (1800 ft) east of the Elk and flows into a 30 m (100 ft) diameter, 10 m (30 ft) deep, vertical-walled sink. The top of the sink is about 5 m (15 ft) below the top of the Union Limestone. The cave’s entrance at the base of the sink is 6 m (20 ft) wide and 1 m (4 ft) high and leads to a passage 5-6 m (15 to 20 ft) wide and high. This passage ex­tends to the north for a short distance but then turns to the southwest; the main orientation of the cave. After 400 m (1300 ft), this passage opens to a 12 m (40 ft) diameter room, then narrows and crosses the Union! Pickaway contact via a 7 m (22 ft) deep pit. The passage below the pit doubles back beneath the upper passage and drops rapidly through the Pickaway via enlarged joints and then bedding planes. This passage terminates in a gravel and mud-choked crawl at a point 56 vertical meters (185 ft) below the cave’s entrance and near the base of the Pickaway Limestone (Figs. 8, 9: H). The passage terminus is the traditional “end” of the cave in the sense that this represents a local low point and marks the end of the earlier exploration. This point (elev. 2396) is only 183 m (600 ft) northwest of the cave’s entrance and is also 21 m (70 ft) lower than the bed of the Elk River at the mouth of Falling Springs Run.

Fig. 10. Contact between Lower Taggard Shale (ceiling) and Patton Limestone in Falling Spring Cave. Photo by Ron Simmons.

Also found in the historical part of Falling Springs Cave is a chamber called Vic’s Room (Figs. 8, 9: V) reached via crawls on the left side of the entrance passage about 330 m (1100 ft) from the cave’s entrance. This chamber, in the lower Union Limestone, is 23 to 30 m (75 to 100 ft) wide, 60 m (200 ft) long and up to 12 m (40 ft) high. While the room terminates at a bedding plane crawl near the north side of the Falling Springs Run valley, there is no evidence that Falling Springs Run flowed into this room from the surface. The apparent floor of Vic’s Room consists of clastic fill and large breakdown blocks. Two unobvious routes through the breakdown lead downward to fragments of two separate flowing streams, each of which is about 12 m (40 ft) below this apparent floor. Both of these streams breach the Union/Pickaway contact, can be followed downstream for a few meters, and then are lost in breakdown and mud chokes in the upper Pickaway. Neither of these streams is related to the cave’s entrance stream although one can be followed upstream for several hundred meters toward the cave’s entrance area. These streams, flowing at depths of about 43 m (140 ft) below the entrance, indicate the ex­istence of deeper drainage within the cave. During flood conditions, constrictions in the lower part of the cave’s entrance passage result in substantial backflooding into Vic’s Room with subsequent deposition of organic material in the room’s floor. This water then migrates downward through the breakdown and fill in the floor to the hypothetical lower levels.

More significant than Vic’s Room is a substantial paleo­trunk (Figs. 8, 9: P) offset from the room and 6-10 m (20-30 ft) above its local floor. This passage also trends to the southwest and ends in breakdown along the hillside about 230 m (750 ft) downstream from the cave’s entrance and just below the elevation of the valley floor. This higher level passage ultimately joins the cave’s entrance passage just above the Union/Pickaway contact and, based on solution scallop orientation and graded fills in its floor, appears to represent a former flow path used by Falling Springs Run when it sank farther down-valley than it does at present.

In October 1981, an unobvious route in the floor of the paleo-passage above Vic’s Room was found leading to a low, muddy, southwest trending passage formed at the Union/Pickaway contact. After about 300 m (1000 ft) of tight, wet and drafting crawls, this passage abruptly opens at the top of a 6 m (20 ft) pit and steep mud slope into a 60 m (200 ft) long, 15 m (50 ft) wide room (Figs. 8, 9: B) which penetrates the thickness of the Pickaway Limestone. A low crawl at the base of this room leads to a second room of similar size and then to passages which penetrate the Taggard Shales and extend downward into the Patton Limestone. Here, at the cave’s lowest level, 75 m (250 ft) below the top of the Union Limestone at the cave’s entrance sink, is found part of the underground Elk River (Figs. 8, 9:R). This water, seen through open joints in the passage floor, consists of a deep pool with no obvious outlet but flowing to the north. The pool is 40 m (130 ft) lower than the bed of the Elk River, is 150 m (500 ft) east of the river bed and is at an elevation of 712 m (2335 ft). Taking the dip component of the limestone between the cave entrance and the pool into account (about 1.2 degrees), the pool is 70 m (230 ft) below the top of the Union Limestone. Using the observed thicknesses of the Union (40 m—130 ft), the Pickaway (20 m—65 ft), and the Taggard (6 m—20 ft), the pool should then be about 6 m (20 ft) below the top of the Patton Limestone and, as observed in the cave, this is the case. As with the Crayfish Pool seen 1.6 km (1 mi) to the south in My Cave, the Taggard Shales act as a capping bed, rather than the more usual perching bed, for the underground Elk. While the exploration leading to the pool represents the culmination of the search begun in the late 1800’s by Maxwell and continued by others, it is, in a sense, anticlimactic in that having reached the underground Elk River, there is, in this cave, no way to follow it.

Above the pool is one other passage of interest. Here, a 12 m (40 ft) climb leads back up through the Taggard Shales and into the base of the Pickaway Limestone. This is il­lustrated in Figure 10. The ceiling is the bottom of the lower Taggard Shale and the person is standing on the Patton Limestone. A low passage at the top of this climb extends 180 m (600 ft) to the east-northeast, following the dip of the limestone back up-valley and toward the historic part of the cave. This passage then opens to yet another substantial chamber, 75 m (250 ft) long, 12 m (40 ft) wide and 15-24 m (50-80 ft) high (Figs. 8, 9: L). Large piles of leaves found in the floor of this room are evidence of inflowing water at times when access to this part of the cave is not possible. While the survey indicates that this room is only 75 m (250 ft) to the west of Vic’s Room and that the high points in its ceiling approach the elevation of the floor of Vic’s Room, there is no evidence of a traversable connection between the two.

The known extent of Falling Springs Cave lies almost en­tirely beneath the valley of Falling Springs Run with the ma­jor passages following joints which trend N55-70E. Although some passages do trend north-south, the cave is not developed for any significant distance in this direction and, in this sense, Falling Springs Cave is more similar to the Simmons Mingo I My Cave complex than it is to the other Elk River valley caves to the north.

Internal drainage in Falling Springs Cave is complex and circuitous. The cave’s entrance stream sinks into mud and! or joints in the floor of its passage at various places, depending on flow conditions. The cave’s internal streams, as noted above, sink within a few meters of the points at which they enter negotiable passages. With the exception of the cave’s entrance passage and one of its internal streams, every major passage in the cave has been abandoned by per­manently flowing water. Rather than having a more com­monly found dendritic pattern with infeeder streams flowing toward a master conduit, Falling Springs Cave consists of several stacked layers of passages generally found at specific stratigraphic horizons. The crossing of these horizons via pits and climbs occurs at widely spaced and unobvious loca­tions in the cave, making the cave’s exploration a somewhat haphazard and unpredictable process.

Exploration has also been impeded by evidence that the cave floods from below. That is, pool level, representing the underground Elk River, rises for substantial vertical distances (in excess of 20 m [70 ft] ) in response to precipita­tion in the Elk’s recharge area. The farthest points in the cave take over three hours to reach and lie beyond low passages that flood if the pool rises only a few meters. Since the rapidity of response of the underground Elk to storm events occurring up to 32 km (20 mi) away is not known, ex­ploration and survey of the lower parts of the cave have been inhibited.

ELK RIVER CAVE  

The Elk River Cave (Fig. 11) was found during a search of infeeders to the Elk River during 1981. As with Falling Spr­ings Cave the entrance is in a high-gradient eastern infeeder to the Elk River; Rough Gap Run. This stream drains about 3.6 sq km (1.4 sq mi) above its sink point at the cave entrance. The entrance is a narrow joint in the streambed and is about 20 m (60 ft) below the top of the Union Limestone.

Fig. 11. Elk R. Cave plan view.

In September and October of 1981, about 2750 m (9000 ft) of passage were mapped with the survey ending at a point where the underground Elk River is first seen in the cave. This exploration was described in an earlier article (Storage, 1981).  In March, 1982, severe flooding completely plugged the entrance to this cave with silt and rocks. The cave was reopened in September of 1982 after several excavation at­tempts were made. At that time we discovered that major infilling had occurred in the cave’s entrance area including gravel and mud fills up to 4.5 m (15 ft) in depth. In addition, a total collapse of part of the entrance room had taken place where several breakdown blocks, the largest measuring 3 x 3 x 4.5 m, had fallen from the ceiling.  The surveyed length of this cave is over 4200 m (14,000 ft). An additional 610 m (2000 ft) of river level passage has been explored upstream (south) to a sump. Access to this unsurveyed river passage is through an area characterized by very low air space in drought. It has not been possible to enter this part of the cave since the autumn of 1983.  For descriptive purposes, the Elk River Cave passages can be grouped into four reasonably distinct levels correspond­ing to various stratigraphic horizons in the upper Greenbrier Group as shown in Figure 12A. A similar profile for caves on the west side of the Elk River is shown in Figure 12B and will be discussed in the next section of this paper. In both profiles, the view is perpendicular to the major axis of the caves; facing toward N53E. The cave entrances have the proper apparent lateral separation (about 1070 m (3500 ft)) when viewed in this direction. Finally, the general locations of major contacts, taking dip into account, is also shown.

 The first major level encountered in the Elk River Cave (but not the highest level) is developed 18-24 m (60-80 ft) below the entrance and consists of abandoned trunk frag­ments in the Pickaway Limestone. A series of pits and joints beneath the cave’s entrance drops through the Union! Pickaway contact to this level; developed on the uppermost of two shale beds in the Pickaway. Here, a major trunk rem­nant, the Happy Maggot passage (Figs. 11, 12A: HM), is en­countered. This passage trends south for 240 m (800 ft), terminating in a mud choke. A continuation of this passage 
trending north from the entrance, as well as the upper por­tions of Reverse Canyon, described below, are also formed at this level. These passages are generally rectangular or elliptical in cross section and are frequently filled with silt or mud to a depth of several meters. In much of the Happy Maggot passage, it appears that the Union/Pickaway con­con­tact and the upper shale bed in the Pickaway Limestone act as capping and perching beds, respectively, for the large volume of water occasionally flowing in this passage. 

Union Limestone near Elk R. Cave entrance.	Passage perched on Bethel Sandstone in Elk. R. Cave	Phreatic tube at Elk R. Cave base level

The second major level encountered (the cave’s highest level) is directly above the Union! Pickaway contact at a depth of about 12 m (40 ft) below the cave’s entrance. This level includes the Upper Trunk and the remainder of the passages found to the north of the entrance. This latter sec­tion is similar to the Happy Maggot passage in cross section and ends in breakdown beneath a shallow sink in the field above. 

Scallops indicate that the paleoflow direction at this level was to the north, paralleling the current flow direction of the Elk River. Several small infeeders enter this passage at locations cor­responding to sinking surface streams. These infeeders flow for a few meters in the Upper Trunk before exiting through small openings in the passage floor. 

Reverse Canyon (Figs. 11, 12A: RC) begins beneath the south end of the Upper Trunk, descends through the Pickaway Limestone, and ends 300 m (1000 ft) to the south where it joins the Elk River Passage. Reverse Canyon is rec­tangular in cross section, is 4.5-9 m (15-30 ft) wide and 3-8 m (10-25 ft) high, and contains a small stream (under .1 cfs) which flows to the south for several hundred meters. This is the only stream seen in any of the Elk River valley caves that flows in this direction. Several pits, found in the first few hundred meters of Reverse Canyon, drop through the lower Pickaway Limestone to the third significant level, on top of the Upper Taggard Shale. Although major passage development does not take place at this stratigraphic horizon, it is significant in that it terminates pits and narrow joints originating in the Upper Trunk. Only short passage segments are found here and, as with the passages beneath Reverse Canyon, these resemble inverted “T” ‘s in cross section, 1.5-3 m (5-10 ft) wide, 15-30cm (6-12 in) high at the base, and 1.5-6 m (5-20 ft) high in the center. In both this and the other Elk River valley caves, the Pickaway / upper Taggard contact serves as an aquitard for drips and seeps within the cave, although for larger volumes of water, e.g., the underground Elk River, it is breached from above and below.

Fig. 13. Underground river in Elk River Cave. Photo by Dave Black

At its south end, Reverse Canyon is perched on the Tag­gard Shales for a short distance and then crosses the Tag­gard before joining the Elk River Passage (Figs. 11, 12A: UGE). This fourth (and lowest) level of cave development is 43-46 m (140-150 ft) below the cave’s entrance and is 3-9 m (10-30 ft) below the top of the Patton Limestone. The nature of the Elk River Passage differs from that of the rest of the cave. This passage is generally smaller and more irregular in cross section, averaging 3 m (10 ft) in width and height, and containing large amounts of breakdown derived from the less competent Taggard Shales above. Formed in the upper Patton Limestone, the passage con­taining the underground Elk River (Fig. 13) represents base level for the Elk River valley caves. To the north (down­stream), this passage can be followed for about 30 m (100 ft) to an end in breakdown. This point is 5200 linear meters (17000 ft) from the main Elk River springs (Fig. 3: MS) and is only 3-6 m (10-20 ft) higher in elevation than these spr­ings. While the likelihood of additional traversable passage in this direction and at this elevation is considered to be small, given the low hydraulic gradient, such passage may exist, especially if the underground Elk climbs back into the Union and Pickaway Limestones along the gradient to the springs.

In the upstream direction, the Elk River Passage can be followed for at least 900 m (3000 ft) through several regions of low airspace, before terminating at a sump. All of the river-level passage is in the Patton Limestone with short side passages and occasional high ceilings extending upward through the Taggard Shales. At river level, the passage is al­most choked at several points by breakdown.

With the exception of the Upper Trunk, every major pas­sage in the Elk River Cave floods to the ceiling. It is ap­parent that in times of high flow, the underground Elk fills its passage, overflows into Reverse Canyon and then flows north (evident from the orientation of small sand dunes [scallops] in the passage floor) to a series of unenterable pits and drains. During such occurrences, the water level rises at least 12 m (40 ft) into the normally southward flowing stream in Reverse Canyon. We have also seen evidence of severe flooding in upper Reverse Canyon, 24 vertical meters (80 ft) above base (river) level. It is not known whether this results from the rising river, or is the result of a large in­crease in the flow of the Reverse Canyon stream. The latter possibility seems unlikely since there is no sign of distur­bance of floor cobbles after floods.

While the Upper Trunk does not seem to flood, the Hap­py Maggot passage below it appears to flood at least once per year. Green leaves, live plants and other recent debris are often seen in ceiling cracks. At its upstream end, water in this passage rises at least 3 m (10 ft) into the joints connect­ing it with the Upper Trunk. This flooding results from the inability of the sediment choked drains at the downstream end of the Happy Maggot passage to carry water sinking at the entrance. For example, in June, 1981, after a heavy shower, water was pooled about 3 m (10 ft) below the top of the 12 m (40 ft) pit leading into the Happy Maggot passage near the cave’s entrance. At this point this passage is about 6 m (20 ft) high. The entire section of cave north of the en­trance, which normally has no active stream, floods under similar circumstances.
Because of the frequency of very local storms, the aerial extent (over 230 sq km-90 sq mi) of the drainage basin, and the high infeeder stream gradients, it is not unusual for some tributaries of the Elk River to flood while others are com­pletely dry. Thus the lower cave may be inaccessible when the entrance is dry and vice versa.

BRADSHAW RUN CAVE 

Bradshaw Run Cave (Figs. 12B and 14) was discovered and surveyed in 1982. The cave’s entrance is in the bed of a high gradient (76 m I km—400 ft/mi) western infeeder to the Elk River draining about 2.6 sq km (one square mile). The en­trance, at an elevation of 767 m (2516 ft), is about 4.5 m (15 ft) below the top of the Union Limestone. Because of channeling in a 24 m (80 ft) wide alluvial fan, the entrance accepts about three fourths of the surface stream regardless of flow conditions. 

At the entrance the stream falls 4.5 m (15 ft) through sandstone boulders and flows horizontally for 12 m (40 ft) before intersecting a large canyon passage near its ceiling. A 11 m (35 ft) waterfall is formed by the entrance stream dropping down the west wall of this north-south trending canyon formed in the Union Limestone. As with the caves described above, the entrance stream is lost almost immediately in breakdown and mud. To the south, the large canyon passage can be followed only 15 m (50 ft) to a massive breakdown choke beneath the valley containing the entrance. To the north, this passage can be followed as a dry trunk, 8-12 m (25-40 ft) wide and 3-9 m (10-30 ft) high, occa­sionally interrupted by breakdown (Fig. 12B: t). The passage gradient, wall scallops and sediment and cobble orientation indicate northward flow in this and all other passages in the cave. After 300 m (1000 ft), the trunk ends abruptly at the top of a mud slope with 12 m (40 ft) of relief (Fig. 12B: in). The presence of organic debris and fresh mud indicates that this cave floods to a level midway down this slope. This point is at an elevation of 739 m (2425 ft) and corresponds to the elevation of the surface bed of the Elk River, 240 m (800 ft) to the northeast. At the base of the mud slope and in the lower Union Limestone, an abandoned stream passage is encountered (Fig. 12B: a). This passage parallels the upper canyon and is about 300 m (1000 ft) long. At its southern (upstream) end, it approaches the valley and terminates in breakdown. Several small streams, emerging from cross joints, flow perpendicular to this passage and exit through mud chokes. Fifteen feet (4.5 m) down the mud slope, a 2 m (8 ft) diameter passage continues to the north. This passage is perched on a sandy bed in the lower third of the Union Limestone. This bed may correspond to the Bethel Sand­stone, found further to the north in West Virginia, where it is a prominent marker bed in the Union.

After several hundred meters, a 15 m (50 ft) pit (Fig. 12B: p) drops through the lower Union Limestone, leads to a short section of low passage on the Union/Pickaway contact, and then intersects a still deeper, large trunk fragment (Fig. 12B: lt). The unnamed upper shale layer in the Pickaway Limestone forms the floor of this passage for about 300 m (1000 ft). After crossing this shale bed, the passage then drops steeply through the remainder of the Pickaway and terminates at a deep pool, perched on the upper Taggard Shale.

This pool, 63 m (208 ft) below the cave’s entrance and at an elevation of 703 m (2308 ft), is 8 m (27 ft) higher than the downstream end of the underground Elk River passage, seen in the Elk River cave, 300 m (1000 ft) to the northeast. Stratigraphically, the pool is 12 m (40 ft) higher than the underground Elk, taking the slight westward dip of the limestone into account. During the only visit to this part of the cave, water depth in the surface bed of the Elk River was at least 2 m (6 ft). Thus, it is possible that the observed pool level was higher than at times of low flow. No current was present in this pool when observed. Given this fact and the pool’s stratigraphic horizon, we conclude that this pool does not represent base level elevation for the underground Elk River.

LEFT IT PIT 

The southern termination of Bradshaw Run Cave beneath the valley where the entrance is located and strong airflow from breakdown beneath this valley led to speculation con­cerning the existence of a continuation of this cave to the south. By projecting the dominant joint trend of Bradshaw Run Cave across the valley and as a result of extensive dig­ging, the entrance to Left It Pit was opened. This entrance is a narrow joint dropping 6 m (20 ft) into a small streamway. After 45 m (150 ft), this passage terminates at the top of a 12 m (38 ft) deep shaft with a 12 m (40 ft) diameter passage below. The floor of the lower passage is 26 m (85 ft) below the cave’s entrance and is slightly lower in elevation than the surface bed of the Elk River, 240 m (800 ft) to the northeast. This passage floor is also accordant with the part of Brad­shaw Run Cave that is near the base of the Union Lime­stone. To the north of the pit, the lower passage ends within 30 m (100 ft) in breakdown beneath the valley and about 45 m (150 ft) from Bradshaw Run Cave. The entrance stream flows north into this passage, sinks in boulders, and is prob­ably seen in Bradshaw Run Cave at point a in Figure 12B.

To the south of the 12 m (38 ft) pit, the lower passage trends southeast for over 2200 m (7500 ft) (LIP Trunk in Figures 12B and 14). This is a major paleo-conduit, modi­fied along its eastern (valley facing) wall by breakdown and infilling due to sinking streams and sinks along the side of Gauley Mountain. While several substantial rooms, up to 45 m (150 ft) in diameter, are found along this passage, these chambers do not appear to be hydrologically significant.

The entire Left It Pit trunk passage is developed at a depth of 24-34 m (80-110 ft) below the cave’s entrance. Solu­tion scallops along the walls and floor indicate flow direc­tion from south to north, as also seen in the other Elk River valley caves. Due to the eastward component of the trend of the cave as one proceeds upstream (southeast) in this trunk and the modest (1.0-1.5 degree) westward dip of the lime­stone, this passage drops stratigraphically in this direction. Consequently, while the northern (downstream) end of this passage is in the lower Union Limestone, at its southern ter­minus, in breakdown beneath a small valley, the upstream end of the passage is stratigraphically lower; about 9 m (30 ft) below the Union/Pickaway contact.

Treated as a single conduit, the combined linear extent of the Bradshaw Run / Left It Pit complex exceeds 2700 m (9000 ft), paralleling the Elk River Valley beneath its western flank. Strong airflow emerging from the breakdown at the south end of this complex may indicate a continuation of the paleo-conduit.

CAVE DEVELOPMENT

The caves described in this paper have a combined surveyed length of over 29 km (18 mi) and are developed in all members of the Greenbrier Group. While a detailed analysis of the development of each cave is beyond the scope of this paper, it is possible to assess the relationships which exist between the caves and to outline in general terms, the sequence of development of these caves.

Complex interactions between dropping base level, structure, stratigraphy and lithology have produced the caves found in the Elk River basin. White and White (1983) have described two primary patterns of cave development in West Virginia, corresponding to valley and plateau karst. They note that thickness of limestone in West Virginia is much less than relief of valleys. The upper Elk River basin, while having over 670 m (2200 ft) of relief, is uncommon in that it
represents an early stage of karst valley development with only the upper 20 m (70 ft) of an almost 120 m (400 ft) thick carbonate rock sequence exposed in the valley floor and adjacent hillsides.

Figure 15. Profile of the Elk River valley caves looking along strike.

Two fairly distinct cave patterns are seen in the Elk River valley: those caves in which N55-65E jointing predominates (the Simmons Mingo I My Cave System and Falling Springs Cave) and those in which the major passages are developed along solutionally enlarged beds and which generally trend in a north-south direction (The Elk River Cave and the Bradshaw Run complex). The former type, developed sub-parallel to the dip, serve as conduits for eastern tributaries to the Elk River and tend to have substantial vertical extent. The latter, paralleling the strike, consist of current or abandoned low gradient conduits for the subsurface Elk River and parallel the Elk River valley.

The dip-oriented caves contain streams flowing in solutionally enlarged joints. These joints drop stratigraphically in a stairstep pattern (noted by White and White [1983] and by Medville and Werner [1977] ) before reaching the base level controlled strike oriented conduits.

Aquitards exert significant vertical control on passage development in these caves. In Falling Springs Cave, for example, the stairstep pattern results from the influence of aquitards in the Union Limestone, at the Union / Pickaway contact, within the Pickaway Limestone and at the Taggard Formation.

Evidence of lithologic influence in all of the caves is given by the rapid rate at which some passages descend through the Pickaway Limestone. This is shown in Figure 15; a profile of the Elk River valley caves looking north, along the strike. In Falling Springs Cave, the mud rooms in the back of the cave (Fig. 15: m) are developed through the entire thickness of the Pickaway. Streams enter these rooms from passages on the Union / Pickaway contact and exit in passages on or below the Taggard Shales. In Bradshaw Run Cave (Fig. 15: b), My Cave, and the upper sections of the Elk River Cave (Fig. 15: e), large, low gradient passages, upon reaching the Pickaway, descend completely through it over relatively short distances (120 m-400 ft or less). While the calcareous portion of the Pickaway is soluble, this limestone also contains several thin beds of clay and shale. This material tends to fill passages in the lower Pickaway. In Falling Springs Cave, for example, a number of small streams flow downdip along the Union/Pickaway contact, descend almost vertically into the Pickaway, and are then lost in deep mud and gravel plugs resulting from the inability of these streams to transport clastic material to base level drainage below.

Fig. 17. Waterfall at Union/Pickaway contact in My Cave.

The clastic plugs and funnels in the lower Pickaway Limestone (Fig. 16) are derived from sediments deposited by the underground Elk River when it rises into these passages from below and, to a lesser extent, from clastic materials in the Pickaway and Union Limestones. In an analysis of the shaley bed 7 m (22 ft) below the top of the Pickaway, for example, Worthington (personal communication) found that this bed consists of 54.3 percent carbonates and 45.7 percent clastics, mostly clays. The relatively high passage density at the Union / Pickaway contact in all of the valley caves and the shafts formed just below this contact may also be an indication of lithologic control. A 0.6-3 m (2-10 ft) thick area at this contact is apparently much less soluble than the surrounding rock although neither sandstone nor shale beds are visible. In Figure 17 for example, the Union/Pickaway contact is seen at the top of a 8 m (25 ft) deep pit with the cave's stream; the underground Dry Branch, flowing into this pit. Similar shafts, found just below this contact, are seen in the other valley caves.

Simmons Mingo Cave provides the best example of the stairstep pattern observed in the Elk River valley caves (Fig. 7). The stream flowing through most of this cave follows a high gradient route along the cave's major axis. This stream drops rapidly within the first few hundred meters of the cave, in this case, completely through the Greenbrier Group. (Paleotrunk passages are found perched on the Union/Pickaway contact for over 300 m [1000 ftj and on the Upper Taggard Shale for over a thousand meters.) After reaching the bottom of the Sinks Grove Limestone; the basal member of the Greenbrier in this area, the stream climbs stratigraphically with negligible gradient for over 1500 m (5000 ft) in a manner similar to the low gradient, strike-oriented passages in the caves beneath the Elk River valley. Where this stream joins the underground Dry Branch beneath the Dry Branch valley, both streams are in the lower Union Limestone. The perched Simmons Mingo stream, eventually dropping rapidly to the pool level in My Cave, represents an underground hanging valley feeding the subsurface Elk River.

A projection of the low gradient upper passage in downstream Simmons Mingo Cave to the Southwest indicates that it will intersect the lower Dry Branch valley about 21 m (70 ft) higher than the streambed, at an elevation of about 782 m (2565 ft). This elevation is just below the top of the Union Limestone in the part of the Elk River valley where the greatest vertical exposure of limestone is seen. We believe that this area, in the vicinity of the Elk River / Dry Branch junction, was the location of the initial point of contact between the downcutting Elk River and the top of the Union Limestone. Here, the Elk flows somewhat farther to the east than in the rest of the valley and thus, would have contacted the west-dipping limestones at an earlier time than elsewhere. We conclude that this upper passage, accordant with the former Elk River elevation when it first reached the top of the Union Limestone, represents the earliest resurgence of the stream in Simmons Mingo Cave; probably in the lower Dry Branch valley.

As the aerial extent of the limestone exposure in the Elk River valley increased, the river intersected both the Simmons Mingo lineament at the top of the Union Limestone, about 800 m (a half mile) upstream of Dry Branch, and exposed limestone farther down valley. Over a period of time, a majority of the river, enlarging joints along the lineament (e.g., the passage inside the Elk River entrance to My Cave), sank and continues to sink in these joints, developing the lowest portion of the Simmons Mingo/My Cave System. As a result of the rapid dropping of local base level to a significant (over .30 m-100 ft) depth below the elevation of the Elk's riverbed, the Simmons Mingo stream was able to abandon its former passage and resurgence site and to join the underground Dry Branch.

The caves paralleling the Elk River valley and seen farther downstream contain a substantial amount of passage in the vicinity of the current elevation of the Elk River. These passages, at elevations of 722-738 m (2370-2420 ft) (Fig. 15), are higher than the pool at the downstream end of the Simmons Mingo/My Cave System (elev. 721 m-2366 ft), and, stratigraphically, are over 27 m (90 ft) above it. While the underground Elk River occasionally rises into some of these passages from below and uses them as overflow routes (this is discussed in the next Section), we believe that these long, valley-aligned paleo-passages were developed prior to the Elk's capture at the lineament. At such time, the Elk, flowing in the upper Union Limestone for several kilometers, would have been able to sink in its bed and to flow beneath the hillsides paralleling its valley, underdraining the riverbed before rising in springs farther down valley. Ample evidence exists for this having been the case. Several segments of large diameter, paleo-passages are seen in the valley, both in the caves described in this paper (e.g., the Upper Trunk in the Elk River Cave) and in several smaller caves as well. One of these; Conrad Cliff Cave, contains a 210 m (700 ft) long segment of dry passage, 6-9 m (20-30 ft) wide and high, ending in mudchokes and rockfall where hillsides curve around and intersect it. This and other similar passages parallel the Elk, are either at the elevation of, or up to 6 vertical meters (20 ft) higher than the riverbed and contain solution scallops indicating former flow to the north.

The nearly complete absence of speleothems in the Elk River valley caves makes a comprehensive program of radiometric dating difficult to achieve. It is possible, however, to develop a general chronological sequence of passage formation on a cave-by-cave basis. Using the available evidence, we conclude that: (a) the Simmons Mingo cave stream, flowing in the highest passages in that cave, originally rose near the top of the Union Limestone in the Elk River! Dry Branch area, at a time when limestone was first exposed in the Elk River valley, (b) the upper levels of the major valley caves, found at and above the current riverbed elevation, were formed relatively rapidly, slightly thereafter, as more limestone was exposed, and (c) upon capture of the Elk River at the lineament, a relatively rapid drop in local base level took place in all of the caves with both the underground Elk River and other sinking streams passing through the Taggard Shales.

HYDROLOGICAL RELATIONSHIPS

We have noted that the Elk River sinks in its bed in the upper Union Limestone and rises at a series of springs 10 km (6 mi) to the north where the top of the Union passes beneath the river bed. Earlier in this paper four hypotheses were presented concerning the nature of the flow path of the underground Elk between these points: (a) The path has a uniform gradient of about 4 m /km (20 ft / mi) to the springs, (b) the path remains beneath the Taggard Shales until the elevation of the lower springs is reached (gradient of 8-10 m /km [40-50 ft/mi]) and then flows horizontally to the springs, (c) the path is “bumpy”; passing through the shales from above and below several times before reaching the springs, and (d) the path remains beneath the Taggard Shales for a significant distance below the elevation of the lower springs and then rises as a phreatic lift to the springs.

Figure 18. Vertical relationships between Elk River valley caves, Elk River bed, and springs.

Although over 16 km (10 mi) of surveyed cave passage exists beneath and immediately adjacent to the Elk River valley, the Elk River itself is observed flowing in one cave for about 900 m (3000 ft) and is possibly seen in two other caves as deep pools at the lowest levels of these caves (the Crayfish Pool in My Cave where some of the Elk River enters from above and the pools in the lowest part of Falling Springs Cave). Thus, only scattered observational data exist for drawing conclusions. These data are summarized in Table 1 and illustrated in Figure 18. Using this information, we may summarize the characteristics of the underground flow path of the Elk River as follows:

(a) Depth beneath riverbed. For a 2.6 km (1.6 mi) distance, from where it is first seen entering My Cave at the Crayfish Pool, to the downstream end of the river passage in the Elk River Cave, the subsurface Elk is 38-43 m (125-140 ft) lower than the surface bed and has a gradient that slightly exceeds that of the riverbed: 10 in/km (53 ft per mile) vs. 8 in/km (44 ft per mile).

(b) Stratigraphic Location. The Elk River, sinking at and above Black Hole, drops through the Union and Pickaway Limestones, the Taggard Formation below, and then into the upper Patton Limestone. Where it is seen in the lowest levels of Falling Spring Cave and for its entire length in the Elk River Cave, the pool / river level remains in the upper 6 m (20 ft) of the Patton but climbs stratigraphically to the north. Only at the downstream terminus of the Elk River Cave does the river approach the Patton! lower Taggard contact. Thus, we conclude that the lower Taggard Shales do not function as a capping bed upstream of this point.

(c) Relationship to springs. The elevation of the down­stream terminus of the underground Elk River passage is 692 m (2270 ft). This is only a few meters higher than the elevation of the main Elk River springs 5 km (3.1 mi) to the north (Fig. 3: MS). Thus, while the subsurface Elk River has about the same gradient as the riverbed upstream of this point, its downstream gradient is negligible. It is not merely coincidental that it is at the downstream end of the traversi­ble portion of the underground Elk River where the eleva­tion of the west-dipping plane of the lower Taggard Shales approaches that of the Elk River springs and it is only at this point where the shales begin to act as a capping bed. To the north of this point, the Elk breaches the shales from below and climbs through the Pickaway and Union Limestones to the springs. It is possible that the shales will force the river to flow at elevations lower than that of the springs (hypothesis (d) above), and that the Elk will then rise under artesian conditions at the springs. We have no evidence, however, that this is the case. The Taggard Shales are easily breached in several places in the Elk River Valley caves and are probably only locally important as an aquitard. Rather, these are gravity, occluded bluff springs as described by Mylroie (1977). The rising of the Elk takes place over a 180 m (600 ft) distance at three such springs, no more than 3 ver­tical meters (10 feet) apart. Under low flow conditions, the upper two springs are dry and only the lowest in elevation discharges water (0.3 cms—10.5@@@ cfs measured). Under seasonally high flow conditions, all three springs discharge (over 7 cms—250 cfs measured). Under conditions of very high flow, an additional set of four springs, 1.6 km (1 mi) upriver and 12 m (40 ft) higher in elevation, also discharge. These high springs (Fig. 3: HS) are active only after extreme precipitation events (e.g., following over 13 cm (5 in) of rainfall in 48 hours during the period Nov. 3-5, 1985, when much of eastern West Virginia experienced severe flooding) and may be evidence for backflooding in the lowest levels of the caves. The Elk River, sinking farther upstream, has been traced to both sets of springs under such conditions.

In general terms, the hydrological relations which exist beneath the upper Elk River valley appear to be straight­forward. When examined in greater detail, however, these are complex with fairly subtle interactions taking place be­tween the conduits, the various sink points of the Elk and its tributaries, and discharge points. It is apparent that two ma­jor conduits exist beneath the Elk river valley; the Bradshaw Run complex paralleling the valley on the west for almost 3 km (2 mi) and the Elk River cave beneath the east side of the valley. The Bradshaw Run system appears to represent an older, now abandoned flow path developed entirely in the Union and Pickaway Limestones. The Elk River Cave in contrast, while primarily consisting of abandoned passages at well defined stratigraphic horizons, also contains a con­duit in the upper Patton Limestone which carries the underground Elk River. We do not rule out the possibility that another active conduit exists in inaccessible lower levels of the Bradshaw Run System. The volume of water observed in the Elk River Cave under fairly average conditions (0.4-0.6 cms; 15-20 cfs) is, however, about the same as that seen at the lower Elk River Springs, and we conclude that a separate, parallel base level conduit does not exist.

We have noted that the northern, downstream terminus of the Bradshaw Run conduit is in the lower Pickaway Limestone at an elevation of 703 m (2308 ft). The trend of this cave, if projected to the north-northwest for another 850 m (2800 ft), will pass 21 vertical meters (70 ft) beneath the bed of the Elk River (Fig. 3: 51). At this location, a substantial volume of water (over 1.4 cms—50 cusecs in high flow) sinks in boulders in the riverbed over a 90 m (300 ft) section. If further projected to the north-northwest for another 1500 m (5000 ft), this path will again intersect the riverbed, this time at river level (elevation 698 m—2290 ft). It is at this point where the upper set of Elk River springs are located (Fig. 3: HS).  
  
Given this circumstantial evidence, we can speculate that at some time in the past, this flow path, formed entirely above the Taggard Shales and terminating at the upper springs, was independent of the conduit seen in the Elk River Cave and containing the underground Elk River. The latter conduit, apparently more recent, is stratigraphically lower and terminates at the lower set of springs. 

While at some time in the past, two separate conduits may have existed, the current situation is more complex. Evidence exists for the integration of the two conduits upstream of the upper set of springs and indeed, for hydrological connections between the valley caves. All stream traces conducted while both sets of springs are active result in dye emerging at all springs. This holds for dye placed in the bed of the sinking Elk River (Fig. 3: Si and S2), as well as the entrances to Bradshaw Run Cave (Fig. 3: 
B) and the Elk River Cave (Fig. 3: E). We conclude that at these, and by inference, other sinkpoints in the valley, water drops vertically to the currently used Elk River conduit and resurges at the lower springs with overflow rising at the up­per springs. Even if a continuation of the Bradshaw Run flowpath does continue to the north-northwest as hypothe­sized above, we believe that any flow through this is pirated by the conduit containing the underground Elk and conse­quently, when the upper springs are dry, all such water flows to the lower springs. 

Connections between the two conduits are also used by the underground Elk when its conduit is full. At such times, backflooding will occur with the overflow rising at the upper springs and filling the lower passages in the valley caves. In effect, these connecting conduits are subsurface estavelles with water both descending and ascending in them, depend­ing on the flow regime. 

Even though it is not physically possible to traverse these connecting conduits, all of the observational evidence in the Elk River valley caves supports this conclusion. For exam­ple, backflooding occurs at several widely spaced locations in the western Bradshaw Run system. At the extreme up­stream end of Left It Pit, a passage descends 9 vertical meters (30 ft) beneath the elevation of the cave’s major con­duit, becoming too narrow to follow near the base of the Pickaway Limestone. While normally a downstream route for a small volume of water entering an adjacent dome, a substantial volume of water sometimes rises in this passage from below. The orientation of cobbles and sand scallops in the passage floor, the orientation of wall scallops in the limestone and the complete absence of sediment in this part of the cave all indicate that the passage serves as a phreatic lift and that the cave’s major conduit is now used only as an overflow route. At the opposite (north) end of the Bradshaw Run System, a similar situation occurs. Again, while the passage leading to the terminal sump at this end of the cave normally carries a small stream flowing down toward the sump, under some conditions the pool level rises substantial­ly, backflooding this passage. It is not possible, unfortunately, to quantify the volume of water flowing in the lower conduits when such backflooding takes place since at such times, these parts of the caves are inaccessible. 

As a result of internal obstruction in the caves such as rockfall and sediment-choked passages, the model presented above: of the underground Elk River filling its conduit and then uniformly rising into successively higher levels in the caves (in essence, the elevation of the underground Elk defining the top of the saturated zone), is subject to some modification. The downstream end of My Cave for exam­ple, is a sediment-choked sump 38 m (124 ft) lower than the surface Elk River bed a few hundred meters away. This stream and that part of the Elk River which sinks at the en­trance of the nearby Black Hole Cave merge at this pool. Because of the inability of the pool to accept large quantities of water, it occasionally rises to the elevation of the Elk River bed. At the same time, it is possible to enter other Elk River valley caves and to descend in open passages to eleva­tions that are well below the riverbed. 

A related situation exists with respect to the sinking of the Elk at Black Hole Cave. Due to obstructions in the narrow cave entrance and armoring of the riverbed by sediments and clastic rocks, the rate of inflow of water into this cave is limited. Consequently, some of the river will often flow beyond its sink point at Black Hole Cave and the riverbed farther downstream will be bank full. Under such conditions open air passages, immediately adjacent to the riverbed and extending for considerable vertical distances beneath it, are enterable. This occurs until the cave’s internal obstructions prevent discharge from taking place as rapidly as the inflow­ing water and local backflooding takes place in the caves. As a result, the Elk River karst can modify seasonal flood behavior as the conduits beneath the valley accept water and buffer the impacts of flood pulses. We should note, however, that for a variety of reasons, the aquifer has a limited ability to do this. First, flowthrough time in the conduits is quite rapid. Dye placed at the Black Hole Cave sinkpoint is detected at the lower Elk River springs, 8.9 km (5.5 mi) downstream, within 24 hours. Also, the springs, while alluviated, are still capable of discharging considerable quantities of water. Under fairly high flow conditions for example, discharge of over 7 cms (250 cfs) has been gauged at two of the three lower springs. Finally, because of the limited ability of the swallets in the riverbed to accept water, the majority of the flow in the river under flood conditions remains on the surface. At the time when the 7 cms (250 cfs) discharge at the springs was re­corded for example, the Elk was in full flood and it was not possible to measure discharge in the river (estimated discharge at that time was over 28 cms (1000 cfs) based on the 18 m (60 ft) width of the river, average water depth in ex­cess of 2 m (6 ft) and velocity of 1-2 in/sec (3-5 ft/sec). 

Under such circumstances, any contribution by the aquifer to modifying flood behavior is overwhelmed by the combination of large surface flow above it and rapid flow through it. This is similar to the findings of E. White (White, B. L., 1975) who, in a statistical analysis of 62 car­bonate basins in the Appalachians, found no relationship between the karst and runoff properties (specifically, be­tween basin area and mean annual flood) for the 12 most highly karsted basins. She notes that “dampening (of runoff) does take place but the presence of carbonate rock alone does not automatically mean there will be damped floods.” This is entirely consistent with our observations in the upper Elk River valley.

ACKNOWLEDGEMENTS

We are indebted to a variety of people who have given significant amounts of their time and energy to the Elk River valley and its caves over the past several years. Many miles of surface surveys were carried out by Hazel and Susan Medville, Charlie Plantz and Bob Thrun. These surveys, between benchmarks, cave entrances and the Elk’s riverbed, were conducted to establish relative positions of key features in the valley. Bob Thrun and Dick Sanford provided plots and profiles of Simmons Mingo Cave. John Ganter pro­vided us with survey data, critical elevations, and geological observations from his resurvey of My Cave. Traps from stream traces were analyzed by Barry Chute.

John M. Hall at Goodyear Aerospace provided valuable assistance in writing computer programs tQ reduce survey data and generate cave profiles. Additional plots and pro­files were prepared using the Survey Manipulation, Analysis and Plotting System (SMAPS) software developed by Doug Dotson at Frostburg State College. Geological comments and advice received from Roy Jameson, John Mylroie, Will White and Steve Worthington, all of whom we managed to lure into the caves, were greatly appreciated as were our in-cave debates with them concerning passage evolution, paleohydrology and the location of contacts in the Green­brier Group.

Finally, we thank those people who worked with us in ridgewalking, digging open new entrances, and surveying all cave passages as they were found. The Elk River caves are particularly muddy, breezy, flood prone and unpleasant. Consistent support in surveying these caves was provided by Andrea Dakoski, Mike Dyas, John Ganter, Hazel Medville, Kathy Nutter, Charlie Plantz, Ron Simmons, Cady Soukup and Roberta Swicegood. Additional surveying was carried out by Dave Black, Jim Borden, Stan Carts, Sandy Flint, Keith Goggin, Dick Graham, Bob Gulden, Roy Jameson, Ben Johnson, Tony Knaus, Bill Liebman, Susan Medville, Rod Morris, John Mylroie, Dick Sanford, Tommy Shifflett, Ed Strausser, Ron Tilkens, Bob Thrun, Steve Worthington and John Zidian.

REFERENCES

Maxwell, Hu (1898)- History of Randolph County, pp. 281-285, Morgantown.

Medville, D. (1977)- Karst Hydrology in the Upper Elk River Basin, West Virginia, NSS Bulletin, 39:18-26.

Medville, D. and E. Werner (1977)- Karst Hydrology and Water Chemistry in a Mixed Sedimentary Terrain, IN Tolson, J. and F. L. Doyle, (eds.)- Karst Hydrogeology, pp. 443-457. 

Mylroie, J. (1977)- Speleogenesis and Karst Geomorphology of the Helder­berg Plateau, Schoharie County, NY, New York Cave Survey Bulletin 2, 336 pp. 

Reger, D. B. (1931)- Randolph County: West Virginia Geological Survey, 989 pp. 

Schmidt, V. et. al. (1965)- The Elk River/Dry Branch Region of West Virginia, Netherworld News, 14:2. 

Storage, W. K. (1981)- Elk River Cave, NSS News, 19:253-254. 

Swicegood, R. (1982)- Simmons-Mingo and the Elk River System, IN: Anderson, R. and L. Baker (eds.)- Capital Area Cavers Bulletin No. 1, 137 pp. 

Wells, D. (1950)- Lower Middle Mississippian of South-Eastern West Virginia, American Association of Petroleum Geologists Bulletin 34:892—922. 

White, E. L. (1975)- Role of Carbonate Rocks in Modifying Extreme Flow Behavior, Pennsylvania State University, Dept. of Civil Engineer­ing, 164 pp. 

White, W. B. and E. L. White (1983)- Patterns of Cave Development and Speleogenesis in West Virginia IN: Medville, D., G. Dasher and E. Werner, (eds.)- 1983 NSS Convention Guidebook, 146 pp.

PUBLICATION HISTORY

This document first appeared in The NSS Bulletin, Volume 48, No. 1, July 1986, ISSN 0146-9517. This version, formatted for web publication, includes additional photos, graphics, and minor revisions and updates.