Located near Montreal in Camden County, Missouri, Carroll Cave (Figure 1.1) is the largest cave in south-central Missouri and one of the most prominent in both Missouri and the United States. Originally known as Traw Cave, about 19km of Carroll Cave has been charted, though it is thought that there are over 32km of passages. First discovered in 1957, Carroll Cave recently became a Registered U.S. Natural Landmark. Within the cave are many areas of scientific interest including a 'bat graveyard', schools of blind fish and speleothems (a secondary mineral cave deposit formed by chemical precipitation, similar to stalagmites [Australian Speleological Federation Inc. 2002]) yet to be described in scientific literature (Weaver 1980). Plans are in place to survey the cave as a whole as well as undertake studies into the biology and geology of the cave.
Figure 1.1 Location of Carroll Cave in Missouri (Yahoo 2002)
There is only one natural opening to Carroll Cave, a low arch 3.5m high and 18m wide to the east of the cave. A small stream flows out of the opening, which ponds due to the flat lying nature of the topography. A boat is needed to traverse the first 400m of the cave. In times of heavy rains, water levels can rise to the ceiling making it a dangerous entrance in stormy weather (Helwig 1965). The natural opening is located on private property and for the past 15 years, the landowner has refused entry to his property and hence into the cave. To prevent trespassing, the entrance has been barred over. Near the entrance, there are biologically and geologically sensitive areas that would be at risk with heavy traffic through the area (Hines 2002a). Consequently, an alternative entrance is required.
1.2 Carroll Cave Conservancy (CCC)
A group of cavers came together in November 1995 in an attempt to gain an alternative entrance to the cave. Searches for another entrance were in vain so a dig began in a sinkhole above the cave on a separate landowners property (Figure 1.2). The dig was done via mechanical means, following natural openings in the rock through which there was airflow. This continued slowly with the broken rock being mucked out of the hole to the surface. In January, 1998, this group of cavers formed the Carroll Cave Conservancy (CCC), dedicated to the conservation and 'perpetual protection' of Carroll Cave and its associated structures as well as construction of a new entrance to the cave. After over 1000 man-days of digging and reaching a depth of over 118m, the first dig was cancelled in 2000 due to safety concerns (Hines 2002b).
Figure 1.2 Cross section of first dig (Hines 2002, unpub.)
It was decided that the best way to access the cave was to sink a shaft through the overlying rock. A location was selected that was central to the cave system and located on property whose owner approved of the workings. The shaft is located near the intersection of the three major passages in the cave system giving the name, 'T-Junction' (Figure 1.3). Initially, the plan was to break the rock through the use of a jackhammer as drilling a hole of sufficient dimensions was beyond the finances of the CCC. A 23cm (9in) borehole was drilled as a pilot hole for the rock chips to pass through, which would eventually be used to construct a landing in the cave. The dolomite, however, proved to harder than anticipated and a series of blockages in the pilot hole for the CCC to rethink their strategy (Hines 2002b).
Blasting was seen as the preferred method to sink the shaft as the CCC has a very limited budget and they wished to undertake the work themselves. An approach was made to Dr. Paul Worsey, Senior Explosives Research Investigator at the University of Missouri-Rolla's Rock Mechanics and Explosives Research Centre, to assist in the planning, preparation and execution of the project.
Figure 1.3 Location of shaft in Carroll Cave (Carroll Cave Conservancy 2002)
The aim of the project is to optimise the blast design for a 33m (110ft) deep, vertical shaft no smaller than 76cm (30in) in diameter with minimal disturbance to the natural formations within the cave system. As mentioned, a 23cm (9in) pilot hole from the surface to the cave has previously been drilled and can be used as a conduit to remove the blasted material from the shaft. Though no major geologic formations are within approximately 15m, there are stalagmites, stalactites and soda straws within this distance. Larger formations are near by and ground vibrations coupled with air blast and flyrock do pose a threat. As the shaft is to have an extended lifetime, back break in the shaft walls should also be minimised.
Southern Missouri forms part of the Ozark Plateau, a 750m to 900m thick, uplifted sequence of calcareous rocks that extends into northern Arkansas and eastern Oklahoma. Deposited in the Upper Cambrian to Early Ordovician Periods, the gently dipping structure lies unconformably overly Middle Proterozoic rhyolites and granites. There are three major units in the Ozark Plateau; the Ozark Aquifer, the St. Francois Confining Unit and the St. Francois Aquifer showing a series of interbedded limestones and calcareous sandstones (Figure 2.1). Outcropping as rolling hills for hundred of kilometres, the plateau is a large structural dome with only small scale, localised deformation. The region plays an important role economically as the Ozark Plateau is host to the world's largest lead-zinc mining district, the Viburnum Trend (Orndorff, Weary and Sebela 2002).
Figure 2.1 Stratigraphic column of the Ozark Plateau (Orndorff, Weary and Sebela 2002)
2.2 Karst formation in the Ozark Plateau
The Ozark Aquifer is host to the majority of karst features in Missouri with roughly 94% of them falling within three zones of its stratigraphy as shown in Figure 2.1. Immediately above each of these zones is a sandstone unit with low porosity. There are two hypotheses put forward to explain this. As the water seeps through the sandstone, its chemistry changes, increasing the dissolution of limestone and dolomite below these sandstones (Orndorff, Weary and Sebela 2002). Alternatively, these sandstones were once a confining unit, increasing the pressure of the groundwater. This, in turn, increased the partial pressure of the CO2 in the water lowering the pH levels, resulting in increased dissolution (Arakaki and Mucci 1995). It is likely that the caves are a result of a combination of these suggestions.
Orndorff, Weary and Sebela (2002) show that the major controls over karst development in the Ozarks is bedding and stratigraphy. Vugs (small cavities in rocks aligned with minerals of different composition [ed. Bates & Jackson 1984]) within the stromatolitic dolomite layers give the rock a higher porosity than the surrounding rock. These vugs as well as joints within the rocks form the opening in which dissolution is initiated. Once these openings widen, bedding planes then take over as the controlling structure as they are more continuous (Figure 2.2). The importance of the sandstone horizons is highlighted by the fact that the stromatolitic dolomites occur throughout the stratigraphic sequence and not just below the sandstones.
Figure 2.2 Influence of bedding planes on karst development at Camp Yarn Cave, MO (Orndorff, Weary and Sebela 2002)
Carroll Cave is located within the 91m thick Gasconade Dolomite, within the Ozark Aquifer (Helwig 1965). This Early Ordovician formation shows a basal interbedded sandstone and dolomite member, the Gunter Sandstone Member, overlain by medium to thick beds of fine and coarse grained, light-grey dolomite with several cherty and pelagic horizons (Orndorff, Weary and Sebela 2002). Stromatolites can be found throughout the dolomite layers. The local structure is sub-horizontal, dipping slightly to the east at seven to eleven metres per kilometre due to the domal structure of the region (Helwig 1965).
Carroll Cave is divided into three units, the Carroll River passage, the Upper Thunder River passage and the Lower Thunder River passage that all meet at a location known as the 'bridge' (Figure 2.3). As previously mentioned, Carroll Cave was formed due to the dissolution of limestones resulting from lowering pH levels and changes in groundwater chemistry due to the overlying Roubidoux Sandstone. This led to the formation of the subterranean Carroll River, which included both the Carroll River passage and the Upper Thunder River passage. In its early stages of development, the Lower Thunder River would have been a tributary stream to the Carroll River. Enlargement of this passage coupled with a lowering of the water table, lead to the water flow being diverted away from the Carroll River passage and the formation of the subterranean Thunder River (Helwig 1965).
Figure 2.3 Plan view of the Carroll Cave System (Helwig 1965)
Today, the three main passages are over 10.5km in length, however, the cave system is believed to create over 32km of caves (Carroll Cave Conservancy 2002). Due to the piracy of the Thunder River, the Carroll River has no major source of water so there is negligible flow. Flow in the Thunder River has been measured at 5.2ML of water per day (Helwig 1965). This value, however, is not constant throughout the year. Figure 2.4 shows the monthly variations in water depth and compares them to the seasonal rainfall at a well near West Plains, Missouri. Though this well is about 150km south of Carroll Cave, it shows that flow rates in caves within the Ozark Plateau vary seasonally with the rainfall. As it is not known when the flow measurement was taken at Carroll Cave, it is not known if the measured rate is high, low or average (USGS 2002).
Figure 2.4 Seasonal variations in precipitation and groundwater levels (USGS 2002)
Lying conformably above the Gasconade Dolomite is the Roubidoux Formation. It is composed of fine to coarse grained, poorly sorted sandstones, interbedded with thin to medium beds of fine to medium grained dolomite and chert (Orndorff, Weary and Sebala 2002). Of particular interest is the basal sandstone member known colloquially as the 'Roubidoux Sandstone'. This member is the overlying sandstone believed to have controlled the formation of Carroll Cave.
Helwig (1965) suggests that the upper Gasconade changes in thickness from 12-25m with Carroll Cave being 18m from the Roubidoux-Gasconade contact. At the blast site, Carroll Cave is 36m below the surface (33m to be blasted, 3m to be stripped) suggesting the top 18m of the blasting will be in the Roubidoux Sandstone. In his report of 1965, Helwig suggests that the cave lies in the upper Gasconade Dolomite then contradicts himself by saying it is entirely within the lower Gasconade Dolomite so the locations of these contacts are, at this stage, hypothetical. Although over 23m of the shaft has already been blasted, no-one with sufficient expertise has been in the shaft to locate the contacts, however, a geologist from the University of Missouri-Rolla is expected to map the shaft's geology as part of the reconciliation process.
The major consideration in developing an initial blast design for the blast is the presence of a 23cm (9in) borehole at the site of the blast. This gives the design flexibility as the borehole can be used as a conduit for removal of the rock. The theory behind this is similar to that applied to the development of ore passes. Both Priest (pers. conv. 13 June 2002) and Worsey (pers. conv. Nov. 2001) agree that a 'rule of thumb' can be applied to the material flowing through the borehole. For hangups to be avoided, fragmentation must reduce the largest particles to less than one-third the diameter of the borehole; in this instance, less than 8cm (3in).
There are many shaft sinking methods that are currently being used in the mining industry, though their scale is generally much larger than that being dealt with in this project. With a diameter of only 76cm (30in), there may be some difficulty in miniaturising conventional shaft sinking methods. Analougous to this project are escapeways in underground mines and the methods used to construct these. Traditionally, rises going from one level to another would be done using alimak raising. Today, however, raise boring is the preferred option. Both of these methods are not feasible due to their high costs and the need to get equipment underground.
As the shaft is vertical and connecting to an existing cavern, vertical crater retreat (VCR) should be considered. Small drillholes would be required to drill the blastholes due to the shaft's small scale. Worsey (pers. conv. 3rd Nov 2001) suggests that small diameter drilling is not accurate over distances larger than 30m (100ft). As VCR requires very accurate drilling to maintain a constant shaft diameter, no further consideration is given to this method.
Most shaft blasts are designed similar to development headings, with a central blast (either burn cuts, V-cuts or fan cuts) opening up a free face for the rest of the face to be blasted into. Often cuts are made as two separate blasts, in a process known as benching or sump cuts, to increase the free face, improving blast efficiency and drainage (Gregory 1984). The small scale requirements of this shaft mean that a central cut would be sufficient to open the shaft to the required diameter. The preferred central cut design is a burn cut for a number of reasons. Firstly, drillholes are parallel, making easy drilling for inexperienced drillers. Secondly, the borehole can act as a reamer forming the basis for the blast design. Finally, work by Bullock has shown that it is an effective method of blasting in Missouri limestone.
Richard Bullock undertook a study into the performance of burn cuts in 1958. Tests were carried out in the silver-lead-zinc deposits of the Viburnum Trend in southeastern Missouri. Figure 3.1 shows a number of different designs that were tested by Bullock and it was found that number 30 performed best in the Ozark Aquifer rocks (Worsey 2001). Of particular interest to this project are the patterns using a single, central reamer hole and produce a circular-shaped blast as these can be simply applied to the requirements of the shaft (5, 11, 12, 16, 17, 18, 21, 22 and 31).
Figure 3.1 Burn cut designs tested by Bullock (Worsey 2001)
3.2 Development of initial blast design
There are many methods that can be used to design a burn cut. As the reamer is already present in the form of a borehole, this is used as the basis of the blast design. It is assumed that there is only one central reamer with all drilled holes to be loaded. This will improve fragmentation and improve pull. Figure 4.2 shows the influence that the diameter of the reamer has on the distance between the centre of the drillholes holes. For a clean blast, the distance between centres of drillholes (a) must be less than 1.5 times the reamer diameter (Φ), though breakage will occur when a is up to 2.1Φ (Langefors & Kihlstrom 1978).
Figure 4.2 Relationship of reamer diameter, distance between centres and blast performance (Langefors & Kihlstrom 1978)
As mentioned, the reamer in this instance is 23cm (9in) and the shaft must be at least 76cm (30in) in diameter. Using the relationship of 1.5Φ gives a distance between centres of 343mm (13.5in) and a shaft diameter of 686mm (Figure 4.3). This falls less than 8cm short of the required diameter. It was assumed that 6 drillholes would be used in the blast pattern as it results in a circular shape, with adequate spacing between drillholes. All drillholes would be fired on number 12 LP delays to simplify the blast design. It is assumed that the inherent scatter of these 6.4 second delays would be sufficient to provide millisecond delays between each hole.
Figure 4.3 Initial blast design
To increase the blast pattern diameter, the distance between centres can be expanded out to the required radius (380mm), however, it is unlikely that this would lead to good pull as it would not blast cleanly. The difference is, however, minor at less than 4cm so it may not have a significant impact on the blast performance. Alternatively, new drillholes could be added to reduce the distance between centres. Two drillholes, each with a distance of a=Φ on opposite sides of the drill pattern will open up the reamer before the rest of the blast is set off (Figure 4.4). Drilling times would, however, be increased to compensate for the additional holes and alternative delays would need to be introduced to ensure that the two inner holes blast first. A ready supply of number 6 LP delays (1.8 seconds) are available so these will be used for the inner holes. Increasing the holes will result in an increase in charge concentration, increasing fragmentation and hence reducing the likelihood of hangups in the borehole.
Figure 4.4 Blast Design with two inner holes