Preliminary Topographic/Sinkhole Analysis - Montreal 7.5 Minute Digital Elevation Model:
Implications for the Exploration of the Carroll Cave System
Preliminary analysis of the topographic surface of the Montreal 7.5 minute quadrangle is complete. A total of 170,752 elevation data points were evaluated at 30 meter spacing over the Montreal 7.5 Minute Quadrangle. The elevations are estimated accurate to within 4 vertical inches. Statistics of the topographic grid include: mean elevation- 937.2’, minimum elevation- 693.2’ and maximum elevation- 1188.9’.
Original Topographic Data- Montreal 7.5 Minute Digital Elevation Model
(USGS) Contour Interval = 10 meters (32’)
Filtering was done on the digital topographic data to identify sinkholes. All contours that closed (forming a depression) were retained while contours that didn’t close were eliminated. The next map show the locations of sinkholes on the Montreal quad based on defining the sinkholes at varying depths of 5’ and 10’. All sinkholes located were at least 30 meters by 30 meters or .2 acres in area. The minimum resolvable sinkhole is therefore 900 square meters (.2 acres) in area. Any sinkholes less than 900 square meters in area will not be detectable on these maps.
The sinkholes are color-coded: red sinkholes are over 900’ elevation above sea level while blue sinkholes are under 900’elevation.
There are 486 sinkholes defined as being at least 5’ in depth, and 264 sinks at 10’ in depth.
Sinkhole Distribution – Montreal Digital Elevation Model – 5’ minimum depth
Note that many of the possible sinkholes are within Traw Hollow, Barnett Hollow, Garman Hollow and Mill Creek. A number of aerially large sinks are located west of Traw Hollow and south of Montreal. The problem in using the 5’ depth cutoff for identifying sinkholes is that many farm ponds could be identified as sinks. However, it is relatively common that farmers’ use sinks as ponds, so a reason exists for using a 5’ cutoff. The other problem that is seen on this 5’ sink map is that many “sinkholes” are located in creek valleys that have alluvium. It is probable that many of these “sinkholes” are artifacts of the 30-meter sampling interval, and simply represent sharp bends in valleys walls. This problem is apparent on the 10’ sink map as well.
Sinkhole Distribution – Montreal Digital Elevation Model – 10’ minimum depth
Eliminating sinkholes within the valleys would cut out the majority of sinkholes on all sinkhole location maps. Hopefully with some additional work on aerial photos, these types of “sinkholes” can be verified. Located on valley floors, many of the located sinks would be in alluvium, rather than bedrock, and therefore possibly not valid.
Shaded Topographic Relief Map with Sinkholes = or >5’ depth (Carroll Cave Area)
Shaded Relief Map of DEM – Montreal Quad- with 5’ sinkholes
Shaded Topographic Relief Map with Sinkholes = or >10’ deep (Carroll Cave Area)
Shaded Relief Map of DEM – Montreal Quad – with 10’ sinkholes
Note that there appears to be two basic types of sinks: plateau sinks (usually red) and valley sinks (usually blue). The deeper sinkholes located in the valley areas have a better chance of directly connecting to Carroll while the plateau sinks are drains into the upper part of the cave system and may not provide human passage. Note that Barnett and Traw Hollows have many valley sinks while Davis Hollow doesn’t have as many.
Sinkhole Validation Techniques Using Digital Data Sets- Residual Method
As a means of helping to isolate the valid sinks from the artificial sinks, I fit the topography with a series of high order polynomial functions to mimic the local geologic structure of the area. The idea is that the topography that we see today is dependent on two basic elements: the geologic horizons and their individual erosion characteristics and the geologic structure, which has folded the originally flat-lying carbonates. If we can determine the geologic structure and remove that influence from the topography, we are left with the eroded and karsted profile of flat lying layers of Gasconade and Roubidoux dolomites. Then by filtering the contours on these restored flat layers, the valid sinkholes should be more obvious.
A scan of polynomial functions was run over the topo data. The polynomial that best fits the known geologic dip and strike orientation and dip magnitude was given by a 22nd order polynomial fit as seen below. This structure approximates Helwig’s structure map as published in his 1965 report on Carroll Cave.
22nd order polynomial fit to Topo Map (Top of Lower Gasconade Structure)C.I.=20’
The effect of the Decauterville structure is seen to the southwest portion of the map as a structurally high with dips to the northeast.
The following map shows the residual contours: the difference in elevations between the original topographic data and the lower Gasconade geologic structure as derived from the polynomial surface fit.
Residual Contours- Montreal Quad – Red Showing Positive Residuals Contours, Blue Showing Negative Residual Contours – Contour Interval = 10 feet.
This map represents the surface of the land as if there had not been any geologic structure in the area: both the Gasconade and Roubidoux formations would be flat lying, and the resulting topography would be a combination of surface weathering and karst formation. If this residual map is then filtered for closed depressions, it could possible help delineate karsted features from erosion features.
Using the residual contour method as described above, I filtered the residual map to reflect just closed residual contours. The contour interval is 10’ so the sinks that show-up must be at least 10’ deep. The surprising observation is that Mill Creek-Garman Hollow become a large blind valley, implying that not only does the cave drain rainfall from the west (plateau sinks around Montreal), but also from the southwest, opening the potential for the cave to go in that direction. I’m not sure I believe that, but the large closed residual area is such a pronounced feature that it can’t be ignored. The validation of sinkholes using the residual technique is this: sinks that show up on both residual and topo maps are more likely to be real.
Residual Sinkholes – 10’ depth - Montreal Quad
Sinkhole Validation Techniques Using Digital Data Sets- Surface Geology
Another means of helping to determine the validity of sinkholes is the use of surface geologic maps. Alluvium and hillside debris tend to cover and mask karst features. Sinks proximal to alluvium and talus fills should be suspect. One needs to be somewhat careful in making generalities however, as blind valleys or sinking streams can be a closed depression (sinkholes) and be filled with alluvial gravels and sands. Sinkholes have a better chance of being significant if they are located where the Gasconade outcrops at the surface. Since the majority of Carroll is formed in the lower Gasconade, sinks formed here are even of greater interest.
To start with, an alluvium and hillside debris map is created from the shaded relief topographic map.
The following map delineates the sands, gravels and debris that fill the valleys in the area.
Alluvium and Hillside Talus Map – Blue Outlines non-cavernous zones
The following map is the surface geologic map for the Montreal 7.5 minute quad.
Alluvium and hillside talus is shaded blue, the lower Gasconade is yellow, the upper Gasconade is brown, the lower Roubidoux is green and the upper Roubidoux is red. The structural dip has been taken into account in the making of this map.
Surface Geologic Map – Montreal Quad – Without Overlays
A detailed view of the Carroll Cave area shows several 5’ or deeper sinks. Several of the sinks are developed in the upper Gasconade formation. One of the larger sinks
Surface Geologic Map with 5’ sinkhole overlays (Carroll Cave Area)
is formed at the contact of the upper Gasconade and lower Roubidoux.
The regional geologic map shows that many sinks are developed in the Gasconade formation with fewer sinks developed in the Roubidoux.
Surface Geologic Map – Montreal Quad – With 5’ Sinkhole Overlay
The regional view of the area shows the upper and the lower Gasconade formations are the geologic horizons most likely to form sinkholes.
The next map shows the surface geology, topography, and location of the 5’ sinks and the map of Carroll Cave. Note that the contacts between formations are displaced in the direction of structural dip, thus reflecting structural deformation on those contacts.
Surface Geologic Map with Topography, 5’ Sinkholes and Carroll Cave Outlined
Note that Toronto Spring rises in alluvial fill directly above the lower Gasconade.
A detailed view of the Carroll Cave area shows that the Upper Thunder River and Carroll River Passage lie under a Roubidoux capped, east-west ridge while the Lower Thunder River and D-7 side passage cuts across topography to terminate at the intersection of the cave and Barnett Hollow. Note also that the entrance to the cave is at the boundary between Mill Creek alluvium fill and the lower Gasconade.
Detail Surface Geologic Map with Topography, 5’ sinkholes, and Carroll Cave
Implications for the Exploration of the Carroll Cave System
One of the major questions to ask is what has controlled the formation of the cave and its location in the Gasconade? It appears that the cave is developed primarily in the lower Gasconade formation: many of the sinkholes are in this formation also. A secondary control of Carroll’s location is the geologic structure that has been previously discussed in this report.
Overlay of Geologic Structure / Top of Lower Gasconade and Carroll Cave Map
The cave roughly follows geologic dip in the Upper Thunder River and Carroll River sections, but follows geologic strike in the Lower Thunder River and D-7 side passageway. If the assumption is made that the cave developed at a specific zone within the lower Gasconade (the initial bedding plane that focused the cave’s location vertically in the lower Gasconade), and that this bedding plane was deformed like the entire Gasconade formation, then an general estimate can be made of the elevation of the cave, and the amount of overburden that covers the cave can be estimated.
The following map shows the estimated elevation of the cave. Subtracting 69 feet from the top of the lower Gasconade structure map makes this map. If the map is correct, then a zone about 69 feet below the top of the lower Gasconade was the location of the initial bedding plane that developed into the Carroll Cave system.
Estimated Elevation of Carroll Cave Derived from Geologic Structure Map C.I.=20’
The amount of overburden above the cave can then be estimated and compared with known overburden measurements over the cave. The following map shows the overburden over Carroll Cave using the difference between the estimated elevation of the cave and the surface topography.
Overburden over Carroll Cave- Contour Interval = 20’
The overburden map and the sinkhole location map can be used separately and together to highlight areas of the cave that are close to the surface.
Montreal Quad-Topographic Map showing location of 5’ sinkholes and Carroll
A close up of the cave and sinkhole locations show a group of potential sinks in Traw Hollow, west of the end of the cave in Upper Thunder River and at the termination of the D-7 side passage of Lower Thunder River.
Topographic Map, 5’ sinkholes and Carroll Cave Outlined
Detailed Sinkhole Map, Carroll Cave area showing 5’ sinks and topography:
Red Oval- Upland, plateau sinks
Grey Ovals – Sinks close to cave passageway, which might indicate, cave extensions
Grey Square – Sinks over cave passageway, which might provide direct connection to known cave
There are only three sets of sinkholes over the known cave:
(1) A series of small sinks at the end of Upper Thunder River, which are too small to show on any of these maps (includes old sinkhole dig)
(2) A sink downstream of the T-Junction in Lower Thunder River and
(3) A sink which marks the termination of the D-7 side passageway.
Other potential sinkholes sets are seen on the above map.
Detailed Overburden Map - With 5’ sinks - Red line- 0’ overburden
The following are measured and estimated depths to the cave from the surface at different points in the cave:
T-Junction - Actual 118.5’, Estimated 119’
Collapse End of Cave - Estimated 80’
Round Room - Estimated 100’
Sump @ Lower Thunder River - Estimated 125’
Terminus @ D-7 - Estimated 15’
Lunch Room - Estimated 105’
This map will be most incorrect where there is a dome or collapse in the cave like in the Mountain Room. I estimate the cave to be actually only about 20-30’ from the surface in this room. Toronto Spring appears to be covered with 9’of fill. Also the lower Thunder River area might be deeper by the amount of Thunder Falls (10-15’+).
After doing this map, I think it is more essential than ever to have an accurate map of the cave to verify some of these estimates. According to this map, the D-7 termination is very close to the surface. It appears that the D-7 passageway has been eroded and the cave has been exposed north of the end of the passage. It is quite possible that a hillside opening back into the cave exists along Barnett Hollow north of the end of D-7.
Initial Water Table Map of Gasconade Aquifer – Toronto Springs Area
After doing some additional work on the water-table map, I overlaid the new map onto the cave, surface topography, 5’sink, and alluvium-hillside debris map.
Without accurate elevations of the cave, it is really impossible to determine the exact configuration of the water-table/cave intersection, but using the sump at the end of Lower Thunder River as a guide suggests that there will not be much air-filled cave in the D-7 side passageway termination area (basically to the north and north east of the known cave). However there may be a fairly significant area where the cave will have both air and water filled chambers (shown in black pattern on next map). The blue patterned area shows where the cave will likely be totally water-filled.
Area of Probable Water-Filled Cave –Black Pattern (Air/Water), Blue Water
The next map shows the areas recommended for special consideration.
The following are some recommendations:
Push Upper Thunder River Passage west towards the plateau sinks. A minimum of 35’ of overburden exits to the west of the collapse end.
Look for side passageways going south from about 1/3 up the Upper Thunder River Passage past the Round Room (left going into cave). The overburden gets pretty thin at a minimum of 20’ west of the Round Room.
Exploration in this area may help push the cave past Traw Hollow on the south and into another drainage divide.
Look for side passageways on the left side (west) in D-7, about 2/3 the length down the passage. These might connect with the sinkholes located northwest in Barnett Hollow.
Look for side passageways to the north along the right side of the Carroll River Passageway, between the Mountain Room and the Turnpike Passage. One of these might connect with Perkins Cave, a cave about 1 mile in length located northeast of Carroll.
The most promising potential, but the most dangerous, is the water passage to Toronto Spring from the sump at the “end” of Lower Thunder River. It is likely that the passage becomes larger going towards the spring. This one segment could double the length of the known cave system.
There is a possible extension of the D-7 side passage towards Toronto Springs.
Much work remains to be done. The next step in the sinkhole evaluation is to examine aerial photos and check the validity of the sinkholes, particularly the “sinks” the computer has identified in Traw and Barnett Hollows. I will try and validate the residual sinks the computer has identified also using aerial photos.