Sadly after 20 years,
eecs.utk.edu will no longer host these web pages.
:-) Pages are now maintained at TNlandforms.US
There are innumerable sinkholes in Tennessee, and the number is growing. The table above lists only those sinkholes with depths greater than 100'. Initial depths were estimated from USGS digital elevation data (10m DEM, see example or another). The first 34 were discovered by visual examination of online topo maps. Chuck Sutherland later (2013) provided a digital analysis (see below) from 10m DEM using Spatial Analyst in ArcMap that identified 9 more 100'+ sinkholes and provided depth, area, and volume. Big Sink in White County is the deepest (261.4'), but Grassy Cove, a National Natural Landmark, has the biggest area (5 square miles) and largest volume (see below for top 10 in area and volume). Additions and corrections welcome.
There are serveral impact craters in Tennessee, but they are not visually significant.
Visual analysis -- topo reconnaissance
In the 2001 Shofner study, geology students were assigned the task of scanning the state's topo maps (7.5' quadrangles) for sinkholes (hachured contours). A grid was provided for a quadrangle, and the students counted the number of sinkhole contour lines that crossed the grid lines. So rather than just a raw count of sinkholes, the metric accounted somewhat for depth and area. Here are density maps gray scale or color from the study.
Digital analysis 2013
Chuck Sutherland located Tennessee sinkholes in the USGS 10m DEM digital elevation data using Spatial Analyst with ArcMap. The nominal accuracy of USGS topo elevation data is half the contour-interval. Most of Tennessee topos use 20' contour, so the digital analysis summarized here only considers sinkholes with a depth of at least 3m. The ArcMap analysis provides sinkhole location (boundary path), area, perimeter, depth, and volume.
The raw data located 21,365 sinkholes (or depressions) with a depth of at least 3m. It is necessary to review the raw data and remove false sinkholes ("fauxholes"). Outliers include man-made depressions (quarries, gravel pits, clay pits, strip-mine holding ponds, levees, road impoundments) and non-karst depressions such as stream-channel depressions and depressions in the sandy soils of west Tennessee. There were also digital anomalies at some quad boundaries, perhaps, reflecting older DEM data based on 7.5' quadrangles. At least 15.4% of the data were false sinkholes -- scubbing is ongoing.
Not surprisingly, the spatial correlation coefficent is 0.93 for sinkholes per quad for the scrubbed DEM data versus the visual counts per quad (54,0000+). The table below shows that counties with the most sinkholes have a slightly different order if only sinkholes of depth 3m or more are counted from the DEM data. For all sinkholes, Robertson County had 3602, but for depths of at least 3m, White county has the most. A table at the bottom of the page show the regional distribution of sinkholes and caves.
|Similary, the underlying sinkholes per quad are roughly the same. For all sinkholes visible on topo, the Talbott quadrangle has the highest count at 1,284, but for DEM sinkholes with a depth of at least 3m, the Livingston quad has the highest (391) and Talbott is second (383). (Based on scrubbed 3m+ DEM data). The CDF shows the density of sinkholes (caves) per quad -- 36% of the quads have no visual sinkholes, 42% have no sinkholes of depth 3m+, and 43% have no caves. To see differences in detail of visual analysis sinkholes (black pin) versus the 3m+ DEM sinkholes (red circle) view the Anderson county data, topo sinkholes. There were 625 visual sinkholes on topo map, and the digital analysis reported 239 sinkholes of depth 3m or more (288 unscrubbed). The spatial distribution of visual vs digital sinkholes over Anderson county is roughy the same (hence the high spatial correlation).|
The table below lists caves and sinkholes per county and the sinkholes of distinction for each county -- deepest, largest area, largest volume. The last four columns are based on the digital analysis of sinkholes with a depth of at least 3m (scrubbed data).
|City||Sinkholes||per sq |
The table at right shows the number of topo sinkholes in a few metropolitan areas. For a coarser state-wide view, see choropleth map.
Here are various Google maps of digital analysis for sinkholes with depth of 3m+ on 3x3 grid over topo quad:
The following tables summarize the 3m+ digital data and give the top 10 sinkholes in volume and area.
The following table shows the distribution of depth or greater over the collection of sinkholes.
The following table lists the 10 biggest sinkholes by area (minimum depth 3m).
The following table shows the cumulative distribution of area over the collection of 3m+ sinkholes (raw and scrubbed data).
The following table lists the 10 biggest sinkholes by volume.
|The Big Sink||4.0|
Most of the Tennessee sinkholes are on private property, but you can visit the following:
Spatial density of Tennessee caves and sinkholes
The Tennessee Cave Society (TCS) has coordinates for the state's 9,800+ caves (in 2013: 9,839 caves). Using Sutherland's cumulative data for caves-per-county and caves-per-quadrangle, we can compare topo sinkhole and cave density. (The Shofner study also compared cave and sinkhole density.) The table at right highlights the county cave and sinkhole density (see full table above) and the following maps illustrate the density.
|see ESRI map|
The Shofner paper reports spatial correlation (0.20) between caves and sinkholes. Using our sinkholes/quad count (visual analysis) and TCS caves/quad count, we calculate a spatial correlation of 0.35 for the whole state. The spatial correlation drops to 0.30 if we exclude quads with no caves and no sinkholes. The correlation of caves to sinkholes of 3m+ depth (digital analysis) is 0.26. About 36% of the state 7.5' quads and 10 of the 95 counties have no caves and no sinkholes.
|Highland Rim W||14%||11%||7%|
|Highland Rim E||15%||16%||33%|
GSMNP landforms (TN/NC) BSF landforms (TN/KY)
Revised Sat, 08 Mar 2014 19:34:25 -0500