Kentucky is hiding far more beneath the surface than most people realize.
Along Interstate 69, crews discovered an abandoned underground coal mine less than 20 feet beneath the roadway after the ground failed for a second time. The mine wasn’t properly recorded, and falling groundwater levels may have helped weaken the support below the highway.
But Kentucky’s underground story is much bigger. Around two-thirds of the state’s population lives in karst country, where limestone slowly dissolves into caves, underground rivers and sinkholes. Mammoth Cave alone now stretches for more than 426 mapped miles, while newer LiDAR and geological surveys continue revealing features older maps missed.
In this DiscoverTube documentary, we explore the hidden mines, collapsing ground, giant cave systems and underground waterways that make Kentucky one of America’s most unusual geological landscapes.
Subscribe to DiscoverTube for more hidden geology, USGS research, sinkholes, earthquakes and scientific discoveries beneath America.
This video examines the geological instabilities of Kentucky, focusing on the risks posed by karst topography and unmapped coal mines. It explores how limestone dissolution creates sinkholes and vast cave systems, the technological shift toward LiDAR and AI for better mapping, and the paleontological significance of the state's subterranean archives.
It is a massive problem one man says he can't fix by himself. A large sinkhole is nearly swallowing
up his backyard. New Channel 5's Kelsey Gibbs talked to this homeowner. Every time it floods
this hole gets bigger, bigger, and bigger. What you just saw is a single hole at the edge of one
lane on one Kentucky interstate. It's also a symptom of something much larger because the
ground under that road is not solid in the way most people assume. Kentucky sits on two overlapping
hidden worlds. One is natural, a limestone landscape that water has been slowly dissolving
for millions of years. The other is man-made, a century of coal workings dug beneath the western
part of the state. Stories like this are what we dig into on DiscoverTube, where our documentaries
go past the headline and keep you updated as new developments change the picture.
The ground that opened on the northbound lane. The sinkhole opened along the edge of the northbound
driving lane of Interstate 69 near mile marker 111 in Hopkins County on the stretch between
Madisonville and Morton's Gap. It was not the first time that piece of road had failed. It was the
second, and the second one appeared roughly 50 feet from where crews had already worked. That
distance is the important number. A repair that holds for a few months and then fails again just
down the shoulder tells engineers that whatever they fixed was only part of something wider.
The first round began with suspected settlement in the roadway, identified after an extended
dry spell that followed a period of record rainfall. Crews excavated the embankment and
drilled core samples down to 31 feet, looking for the source of the movement. They didn't find it
right away. That's worth sitting with for a moment, because the search itself was the difficult part.
The road surface above looked normal. Traffic was moving over it, and the cause was somewhere below,
invisible from the top. Eventually, they located a void roughly 25 feet beneath the road. It was
filled with concrete, and the fix held for about eight months before the ground moved again. Coal
was mined at this location more than 60 years ago, before the highway that is now I-69 existed here
at all. The road was built across workings that predated it, and the surface gave no warning while
the support underneath was already compromised. Drivers passing at 70 miles an hour had no way
of knowing the difference between solid fill and an arch of soil holding itself up over empty space.
So when the pavement sank a second time, 50 feet from the first repair, the Kentucky
Transportation Cabinet stopped treating it as an isolated defect. A spokesperson for the Cabinet
later described the realization plainly. This was a bigger issue than they'd assumed.
That shift forced a different kind of investigation, not patching the surface,
but going down to find out what was actually there. The mine that wasn't on any map.
Working with the Kentucky Division of Abandoned Minelands, engineers excavated the affected area
and exposed an underground mine roughly 18 to 20 feet below the surface. The agency confirmed
something striking about it. There are no records of permitted mining at that location and that
depth since modern record-keeping began in the 1960s. Available data suggests the workings may
date back as far as the 1920s. In other words, a mine sat under an interstate and officially
did not exist. To confirm what they were dealing with, Abandoned Minelands lowered a camera into
the works. The footage confirmed the old mine workings were there, and it showed something
else that matters far more than the mine's age. Groundwater levels inside the mine had recently
dropped. That observation is the mechanical heart of this whole story, so it's worth explaining
carefully. Abandoned underground mines typically fill with water over time. That water is not
passive. It exerts hydrostatic pressure against the roof and the remaining coal pillars, and that
pressure actively helps hold the structure up. Water in a flooded mine is load-bearing. When the
water table falls, that support falls with it. Pillars and roof sections that were being braced
by pressure from below are suddenly carrying more of the load alone, and the likelihood of
settlement or collapse rises. That collapse then works its way upward through the overlying material
until it reaches the surface as subsidence. Now look back at how the first failure started.
Settlement noticed after a dry period that followed record rainfall. That's the same lever.
The agency has confirmed the groundwater drop as an observation and stated the mechanism it drives,
treating the drought as the single proven cause of everything that happened here,
would go further than the evidence does. But the two sit together uncomfortably well.
In April, the Transportation Cabinet awarded an emergency contract to GeoSpecialties,
a Lexington firm, to stabilize the corridor. Grout injection into the underground voids began
within days. Nat Hughes, the Public Information Officer for the Cabinet's District 2, described
the method directly. Crews drill boreholes through the roadway and the outside shoulders,
then pump grout down, which leads to the obvious question. If a mine sitting 18 feet under a major
highway could go completely unrecorded, what else is missing from Kentucky's picture of its own
underground? Because the state's natural voids have been outrunning its maps for far longer than
its mines have, the cave that keeps getting longer beneath Edmondson County. Six people
from the Cave Research Foundation spent more than 14 hours underground in 1972, pushing through
tight, wet passage until they surfaced on the far side of a ridge and settled an argument
cavers had been having for years. The Flint Ridge system and Mammoth Cave were not two caves. They
were one cave, with a paper boundary running through the middle of it. Flint Ridge had been
surveyed at about 86.5 miles, Mammoth at just under 58. Once the connection was confirmed,
both numbers collapsed into a single system of roughly 144 miles, and the world's longest cave
got substantially longer without anyone digging a new passage. That pattern kept repeating. Proctor
Cave, Morrison Cave, and Roppel Cave were each absorbed into the same system as connections were
found, and Roppel alone ran about 50 miles as an independent cave before it stopped being one.
Today, the surveyed figure stands at 426 miles. That is the long-standing official total published
by the National Park Service, not a fresh measurement, and it describes only the length
that has been walked, crawled, and recorded by hand. For perspective, that surveyed length is
longer than the drive across Kentucky from one end of the state to the other, and the second
longest cave system on Earth does not reach half of it. Stranger still is how the number was produced,
not by remote sensing, but accumulated section by section by volunteers carrying compasses,
measuring tapes, and rangefinders through passage that sometimes requires swimming
or kayaking to reach. The Cave Research Foundation still runs at least 10 expeditions a year inside
the park's 52,830 acres, and the total keeps creeping upward. Nor is Mammoth an isolated freak.
Fisher Ridge, a separate system in the same region, exceeds 100 miles on its own.
Which raises the question the survey data cannot answer. Measuring passage tells you how much cave
exists. It doesn't tell you what carved it, or why the roof above some of it eventually
finds its way to the surface. How limestone eats a state. The answer starts about 340 million years
ago, on the floor of a warm, shallow sea that covered what is now central Kentucky. The
limestones and shales of the St. Genevieve Formation were laid down there as reef sediment,
and that is the rock the entire cave system is cut into. Limestone has one property that
makes everything else inevitable. It dissolves in mildly acidic water. Rainwater picks up carbon
dioxide from the air and soil, turns slightly acidic, and works down through joints and
bedding planes. Over millions of years, those hairline routes widen into conduits, conduits
into passages, and passages stack into a multi-level labyrinth. A resistant sandstone caprock over
parts of the system helped hold the ceilings up, which is why some Mammoth Cave passages are so
unusually large. That landscape has a name. Karst is terrain formed by the dissolution of soluble
rock, and it produces a recognisable set of features. Sinkholes, streams that vanish into
the ground, subterranean drainage, and caves. Its defining trait is speed. Water in karst does not
percolate slowly through soil and gravel. It moves through open conduits, fast, with very little
filtration. If you enjoy this kind of deep dive into how landscapes actually work, subscribing
to DiscoverTube keeps these documentaries coming to you. The collapse mechanism deserves careful
attention, because it explains why sinkholes seem to arrive without warning. The Kentucky
Geological Survey describes the sequence in stages. Soil above a void in the bedrock bridges the gap
and forms an arch. Stress cracks develop inside that arch as it thins. The arch eventually fails,
dropping soil into the cave below, and an underground stream carries the fallen material
away until the banks slump inward, and the surface shows a bowl-shaped depression. The critical
detail is where that failure happens. The void doing the work sits in the soil above the bedrock,
not in the rock itself, held up by nothing more than the soil's own strength. The ground overhead
can pass every visual inspection, right up to the moment it lets go. These features range from one
or two feet across to tens of feet deep and wide. And that same process shaped the visible state.
Much of Kentucky's rolling topography is collapse country, not erosion country. So how well is any
of it mapped? Less well than you would hope, and researchers say so directly. Kentucky's Karst
Potential Index is built on a geologic map at a scale of 1 to 500,000, supported by fewer than
500 carbonate rock data points. At that scale, an entire city fits inside a strip of paper about
as wide as a fingertip. The sinkhole inventory behind it descends from topographic maps with
roughly 10-foot resolution and contour intervals that erase shallow or small features entirely.
A University of Kentucky assessment concluded, the resulting method is broad in scope and
ineffective for actual hazard assessment. That gap has a consequence property owners feel directly.
Sinkholes that were filled and built over can reactivate, because the underground stream
keeps flushing the fill out from below. Two-thirds of Kentucky lives on it.
Bowling Green is the clearest demonstration. The city sits entirely on a well-developed karst
plain, and its natural drainage does not run over land to a river. It runs down through sinkholes,
sinking streams, and underground rivers in the limestone, before reaching the Barren River.
That arrangement is not unusual here. Kentucky emergency management figures indicate roughly
67% of the population lives in a karst region, and separately, about 55% of the state's land area
sits on karst-prone substrate. Those two numbers measure different things, people and acreage,
but they point the same direction. Frankfurt, Louisville, Lexington, Lawrenceburg, and Georgetown
are all built on this geography. Development changes the equation. Parking lots, roofs,
and roads are impervious, so instead of rain soaking in evenly, stormwater gets concentrated
and delivered to a few points. In karst, where water goes decides where the ground fails.
The consequences are practical. Ground movement ruptures sewer lines, gas mains,
and buried utilities. And because karst conduits carry water fast, with little filtration,
contamination entering one sinkhole can reach springs and wells, serving thousands of homes
far quicker than most people assume. All of that drainage has a destination. In the cave region,
it is the Green River, and that river now sits at the centre of a federal decision.
The Line Congress is redrawing above the caves. In March, the House Natural Resources Subcommittee
on Federal Lands held a legislative hearing on the Mammoth Cave National Park Boundary Adjustment
Act, and one of the voices supporting it came from the county sitting directly on top of the cave.
Scott Lindsay, the judge executive of Edmondson County, testified in favour and put numbers on
what the park means locally. Roughly 747,000 visitors last year, spending an estimated $73.2
million in surrounding communities. The bill itself is narrow. It would authorize the acquisition of
about 551 acres in Edmondson and Barron counties, currently held by the Nature Conservancy,
and fold that ground into the park. Precision matters here, because this is not a completed
transfer. The sponsoring congressman's office described the hearing as the first hurdle,
with committee markup and votes still ahead. And even if it passes, the land would still have to
Nothing has changed on a map yet. What makes the acreage interesting is what lies underneath it.
According to Representative Brett Guthrie's office, the proposed addition includes cave
passages containing prehistoric artefacts, and the land falls inside the Green River Watershed,
which testimony described as one of the most biologically diverse aquatic ecosystems in
the country. That river is not a passive feature of the landscape. Flooding on the
Green River tore out the infrastructure of the Green River Ferry, the crossing local
residents used to reach the far side of the park. Repairs required 2,000-foot navigational cables,
four towers standing 20 feet tall, and the replacement of a 15.5-tonne concrete anchor
the water had dislodged from the riverbank. That anchor weighs about as much as a loaded city bus,
and river water moved it. Daily service has since resumed, running six in the morning to six in the
evening. So the same water that carved the passages is still rearranging the surface above them.
Which raises a question worth putting to you directly. Should federal protection extend to
cave passages nobody has fully surveyed, and how would you even draw that boundary? Tell us how
you see it, because that tension between protecting ground and measuring it is exactly what we keep
returning to on DiscoverTube. Reading the ground with lasers and old teeth. Kentucky's mapping
problem is finally being attacked with tools that do not depend on someone walking the ground.
Researchers at the Kentucky Geological Survey built a sinkhole mapping method out of airborne
LiDAR point clouds and tested it in the Floyd's Fork Watershed east of Louisville. LiDAR fires
laser pulses at the ground from an aircraft and times their return, producing an elevation model
detailed enough to resolve features a few feet across. That resolution is the entire point. The
older sinkhole inventory came from topographic maps with contour intervals of three meters,
six meters, or more. And anything shallower than those intervals simply never appeared on paper.
A LiDAR model has no such blind spot, so depressions that were invisible before
show up as measurable closed basins. A team at the survey and the University of Kentucky,
including Junfeng Zhu, Charles Cunningham, and Ryan Koch, has taken that further, presenting work
on extending cast and sinkhole hazard assessment using deep learning image segmentation, training
software to recognize the shape of a sinkhole across huge areas of terrain automatically. No
statewide figures have come out of it yet. It is in progress, not concluded. The survey is also
candid about the limits of what better maps deliver. Its public sinkhole dataset states
plainly that it does not indicate the probability of a future cover collapse. Knowing a depression
exists is not the same as knowing when, or whether, the soil arch under a particular patch of ground
will fail. Other methods fill in what lasers cannot see, because LiDAR reads the surface
and says nothing about where water travels once it disappears. For that, the Kentucky Division
of Water has run dye trace studies, pouring fluorescent dye into sinking streams and
watching which springs it emerges from, sometimes many miles away. Those traces show the plumbing
is not fixed. Flow routes shift during heavy rainfall, sending water to different outlets
than it uses in dry weather. The survey's geologic hazard section also listens from below, running
the Kentucky Seismic and Strong Motion Network's 24 permanent stations, with recordings from 14
shared with the University of Memphis network. Inside the cave, the same story of revision
has been playing out in the rock itself. A paleontological resource inventory began there
several years ago, and teams of geologists, paleontologists, park staff, and volunteers
have since identified more than 40 species of fossil sharks and their relatives from the cave
walls and ceilings. The reason those specimens matter is preservation. Fossils exposed at the
surface are worn down by weather, while these have been sealed inside a cave for millions of years.
One paleontologist described a find as looking like it had just come out of the shark's mouth.
Named species have accumulated steadily from that work. Strigilodus tollisoni, Troglocladodus
trimblei, Glicmanius cariforum, Clavusodons mckinnessi, Macadens ulsoni. Some were tiny.
Clavusodons, nicknamed the chipmunk shark, was only three to four inches long, feeding on
One of them moved a timeline. Evidence from Glicmanius cariforum pushes the origins of
that Tenocanth lineage back more than 50 million years earlier than researchers had assumed.
This is not a fresh announcement. It is the accumulated result of an inventory running
for years. But it makes the same point the LiDAR makes from the opposite direction.
The expected picture was too small. So the answer to why Kentucky's underground keeps
outgrowing its maps is not mysterious. Almost everything known about it was recorded by hand,
inferred from surfaces, or copied from records that stopped being kept. Better maps now decide
where roads, subdivisions, schools, and pipelines get built above the voids. Back in Hopkins County,
the grout wall is still doing its job on Interstate 69, a reinforced perimeter drawn
around workings nobody documented, while groundwater levels beneath the corridor
remain the thing to watch. Committee markup on the boundary bill is the next step to follow,
and the old contour data is steadily being replaced by laser models.
If this is the kind of story you want more of, subscribe, and tell us in the comments,
which stretch of ground in your own region should DiscoverTube investigate next?