Micro-temperature
Variations of Surface Karst
By M.D. Hall
June 2003
Figures
1)
Cave Entrance and Snow Melt due to warm Air coming from Cave
2)
Illustration of Thermal Conductivity
3)
Annual Temperature Distribution
4)
Theoretical and Measured /Estimated Temperature Distribution in Soil
covering Cave
5)
Aerial Photo of Sinkhole
6)
Sinkhole Topography
7)
Noise Survey
8)
Base Station
9)
Noise and Sinkhole Survey
10) Map of Sinkhole Temperatures December 1990 at 6”
Depth
11) Map of Sinkhole Temperatures December 1990 at 12” Depth
12) 3-D View of
Sinkhole Temperatures December 1990 at 12” Depth
13) Map of Sinkhole Temperatures August 1991
14) Noise Survey showing Thermal Drift of
Background Temperatures
15) Depth Estimate to Cave
16) Picture of Discovered Cave Entrance
17) Picture of Enlarged Cave Entrance
18) Map of Cave
19) Depth of Cave Detection
Introduction
Caves
in shallow carbonate terrains have air temperatures that mimic the mean yearly
temperature of the surrounding surface.
Rock and soil overburden act as a thermal insulator and prevent significant
temperature variations inside most shallow caves. This thermal insulation creates a stable
temperature (boundary condition) within a planar (horizontal) cave and causes
the temperature to remain relatively constant over time. The idea I’m presenting is this: that by the careful collection, processing,
and interpretation of micro-temperature data over surface karst, one can assess
the likelihood that surface karst is connected with a cave and an estimate of
the depth to the cave can be made. The
method is prone to misuse and misinterpretation. I’ll present the obvious and the theoretical
for your evaluation as well as some pitfalls and problems using the method.
We
have all felt the warm air coming from a cave in the winter and the cool air in
summer. The following picture
illustrates this point (Figure 1). Here
the warm air has melted the snow a few feet from the cave opening. The dark band around the entrance is where
the snow had fallen, but then melted because of warm air coming from the
cave. The temperature variation between
a cave and outside temperature can be large.
The temperature of this cave is 56 degrees F while the outside air
temperature was 25 degrees F.
This
temperature contrast of 31 degree F (Delta T) is easily felt and seen. What happens if the cave is covered by a
collapse or is filled in with dirt and debris?
Depending on the time of year, the depth of the cave, and the material
filling the opening, the cave may still have a thermal signature, although not
visually apparent or easily detectable.
Figure
1.
Radiation
and Convection
There
are three different mechanisms by which thermal energy is transported: radiation, convection and conduction. I would propose that radiant and convective
heat flow are important only if the cave has an opening(s) to allow for
sun-light to enter the cave and/or if there is significant air circulation in
the cave that causes convection to occur within the cave. Natural convection does occur in all caves,
but is not the primary mechanism of heat flow in covered, planar caves. These two types of thermal transport
mechanisms (radiation and convention) interfere with temperature measurements
on the surface and will be discussed latter.
If the simplifying assumption is made that a buried cave transports
thermal energy out of the cave primarily by conduction, then estimates of the
buried cave entrance size and depth can be made.
Conduction
Heat
is like water. Like water flowing
downhill (or to a lower potential energy position), heat will naturally flow
from higher temperature to lower temperature.
The material that heat flows through determines the distribution of
temperatures in that material and the time necessary for the temperatures across
the material to equilibrate. One can tell by looking at roofs of homes in the
wintertime which roofs have higher thermal conductivity. Roofs with snow are less conductive (more
insulating) than roofs that have had the snow melt from them. We are looking for evidence of the leaking
‘roof’ over a cave: material, which is
more thermally conductive than surrounding bedrock.
This
picture illustrates the idea of thermal conduction (Figure 2).
Figure
2.
The
blue block represents the material blocking the cave entrance. Heat flow through
the material covering the cave can be described as:
Q
= K * (Delta T/L)*A
Where:
Q=
the magnitude of heat flow (discharge)
K=
thermal conductivity of the material covering the cave entrance
Delta
T= Cave air temperature T1 – Surface of ground temperature T2
A=
Cross section area of buried cave entrance
L=
Depth to cave or thickness of material covering the cave entrance.
The
heat flow equation implies that the following are associated with greater heat
flow through material covering the cave opening:
O
Large cross section (surface area) of buried opening (A)
O
Large temperature difference between the cave and surface temperature (Delta T)
O
Shallow depth to cave (less depth of material covering cave) or thin cave fill.
(L)
O
Damp material versus dry material. Water
has a thermal conductivity 23 times that of dry air; therefore damp cave-fills will conduct heat
away from the cave better that dry cave-fills (air acts more like an insulator
than a conductor of heat) (K).
The
ideal buried cave entrance to be located thermally would be one associated with
greater heat flow. The highest heat flow
though a buried entrance would be associated with a large opening covered by a
thin veneer of damp soils being investigated in the late summer or late winter
(depending on local climate) when the temperature difference between the cave
and outside would be the greatest.
Conversely, caves, which have small openings, covered with thick, dry
insulating soils would be difficult to detect thermally, especially in the
spring or fall months when temperature differences between the cave and outside
are small.
If buried caves are seen at all thermally, it
is likely because there are lateral variations in thermal conductivity of the
material covering the cave. Bedrock is
more thermally conductive than dry soil. Wet soils are more thermally
conductive than dry soils. So how can
the area above or alongside a buried cave entrance be more thermally conductive
than the surrounding bedrock? More than
likely it is because of the tendency of sinkholes to be wet. Wet soils and broken rubble tend to be more
thermally conductive than dry bedrock.
To
test the theory that temperature variations in debris covering a cave can be
estimated, a simulation program was used to determine heat flow through a
hypothetical cave entrance. Thermal
conductivity and the heat capacity of the debris and soil covering the cave
were estimated from engineering tables.
The interior boundary condition of the cave was held at 56 degree F and
the exterior surface temperature was varied based on monthly temperature
readings in a nearby city (Figure 3). 
Figure
3.
The
simulation was run over one year of historic temperature measurements. The simulation shows the distribution of
temperatures in the soil covering the cave with a covering of 12’ of soil, clay
and rock (Figure 4).
Figure
4.
Note
that at 4 foot depth, on December 31, the temperature was estimated to be 47
degrees F above the cave. Actual measured temperatures on this date over the
buried entrance of an air-filled cave were estimated to be 44 degrees at 4 feet
from the surface, surprisingly close to the model predictions considering all
the assumptions used in this analysis.
There does appear to be a valid theoretical basis for the use of
temperature surveys in the analysis of conduction of heat through a buried cave
opening.
Before
you grab a thermometer and start sticking it in sinkholes all over creation,
let’s talk about the problems using this method.
Even
though there maybe a theoretical basis for using temperature readings to
indicated caves in connection with surface karst, in most cases anomalies will
be found (temperatures above or below background soil temperatures), that are
not associated with thermal conduction through a buried cave entrance (it just
doesn’t work!) Why?
Pitfalls
and Problems
0
Surface radiation by the sun and convection by the wind make any temperature
interpretation at the surface difficult.
0
Variations in temperature readings are more likely due to a sum of various
errors in measurement (including poor instrumentation and calibration) and
minor changes in heat flow due to radiation and convection, not lateral
variations in heat flow through a cave entrance (thermal noise).
0
Surface temperatures vary systematically during the day and night and make the
differentiation of a thermal anomaly associated with a cave difficult. (Thermal
drift or change in near surface temperature readings unrelated to a cave).
0
Since small changes in temperatures may be significant, sensitive
instrumentation is required.
0
In most areas, only the shallowest of buried caves have a prayer of being found
thermally. In the area of Missouri I have tried this method, it appears that
some caves can be located down to a maximum depth of around 20 feet.
0
If the material covering the cave entrance has the same or similar thermal
properties as the surrounding bedrock, the cave will be invisible thermally.
0
The greatest heartache is this: a valid
thermal anomaly may be detectable over a cave, yet the joint, sinkhole, etc. is
too small to be entered. The surface karst may be in thermal connection with a
cave, but not a human sized connection.
In other words, temperature surveys may help in picking or prioritizing
the spot to dig, but there is never a guarantee that the cave you find is
bigger than a breadbox.
As
in most geophysical methods, the successful use of thermal surveys is greatly
enhanced by the integration of other data sources such as climatic,
topographic, remote sensing, aerial photos, hydrogeologic (both surface and
subsurface) and existing cave temperature data.
Nothing can surpass local experience and knowledge. Thermal surveys may not work at all in your
area if your caves are vertical, with small, dry-soil covered entrances.
There
are definitely problems with recording and interpretation of surface thermal
data and this method of investigation has received some criticism.
One
PhD I mentioned this idea to said that different temperatures in sinkholes are
nothing more than variations in differences in sunlight and wind conditions
around the sinkhole. For this reason I
make my surveys on cloudy, windless days.
The other point is that he had looked at thermal video of a cave and saw
no measurable difference in air temperature, but did see thermal variations in
the groundwater coming from a cave entrance.
He does raise an important issue:
water and air temperatures in a cave are often similar, but not always
identical…meaning a thermal anomaly may exist in the air of a cave but not the
water in the same cave and visa versa.
The other point he says is that in his experience, he has not seen
thermal anomalies associated with cave entrances or buried caves. Here is a great frustration of using the
method…what works in August probably won’t work in May. There can be such
changes in surface temperature conditions that what works one day won’t work
even on the next, or even the same day at a different time. Thermal anomalies are highly dependant on the
temperature contrast between the cave and outside surface temperature. As an example of using the method, I’ll use a
sinkhole temperature survey conducted in south-central Missouri.
Case
History
Climate
Cave
temperatures in the area range from 52 degree F to 56 degree F. The spring associated with the surface karst
that was investigated has a temperature of 56 degree F and was assumed to
represent the temperature in the cave that was being sought. Rainfall in the
area averages 40 inches per year.
Geology
and Karst Morphology
The
collapse sinkhole that was surveyed is developed in the Gasconade formation
(Figure 5). This dolomite is characterized by being massively bedded,
coarse-grained and chock full of solution features including dissolution vugs
and enlarged joints. The dolomite is
cherty, and is present both as nodules and discrete chert beds.
Figure
5.
The
geologic structure of the area is simple, with less that 1 degree of structural
dip to the northeast. The prominent
joint pattern is observed to have an orientation of N 70 degree East in the
area. There are no major faults in the
immediate survey area.
Caves
and springs in the area commonly form in the Gasconade dolomite. The caves in the area are typical Ozark
developments being meandering stream passageways with several side-passages
common. Groundwater flows northeast from the sinkhole towards a spring resurgence. The water table elevation in the area around
the sinkhole was estimated to be 847’ based on water table elevations in nearby
caves, mines, springs and wells. Most caves in the area don’t have much
water…usually a small stream passage with the cave being filled primarily with
air and mud.
Survey
Candidate
Figure
6 shows the topography in the area around the candidate sinkhole. The sink is 200 feet long, by over 100 feet
wide and 60 feet deep. The volume of the
sinkhole was estimated to be over 6,000 cubic yards. The sinkhole trends along the prominent
joint direction of N 70 degree E.
Figure
6.
.
The
sinkhole historically had never been filled with water, regardless of the
amount of rainfall. There was what
appears to be sponge work (speleothems) on the cliff face on the southeast side
of the sinkhole. The sink is covered
with deep, damp soils, brush and trees of 3-5” in diameter. Tree rings were used to estimate the age of
several trees at the base of the sink at between 50-60 years in age, suggesting
that the sink had collapsed at least that long ago.
The
presence of sponge work, the steep cliff face and the lack of water retention
suggested that the sink had formed by a cave’s roof collapse and that a cave
existed below the sink. The elevation of
the base of the sinkhole was 872’ prior to excavation. Because most caves are at or near the water
table of the area (847’), based on the local hydrology, the cave should be
within 25’ of the base of the sinkhole. But where would be the best place to
dig within the sink? Several
micro-temperature surveys were conducted to select the dig location within the
sink.
Date
of Survey
The
first survey was conducted in December 1990, and repeated in August 1991, when
there were large contrasts between the local cave temperatures and the surface
temperatures.
Time
of Day
The
December 1990 survey was conducted on a late afternoon in cloudy cold
conditions with very little wind. The
air temperature was 23.9 degree F. The August survey was on a hot, cloudy day
with the air temperature of 90.0 degree F.
Instrumentation
An
Omega 450 AKT type K Digital Thermometer-thermocouple with a 12” long probe was
used in the December 1990 and August 1991 survey. The relative and absolute accuracy of this
instrument is +/- .1 degree F. This
thermometer was used along with a Fischer’s Digital Thermometer with a 6” probe
to continuously record temperatures at a base station in the sinkhole. This thermometer has a relative accuracy of
+/-.5 degree F, with an absolute accuracy of +/- 1 degree F. Both probes were calibrated prior to
use. Both thermometers can be read to
the nearest .1 degree F.
December
1990 Survey
Noise Survey
A temperature profile was taken at 1’ depths and 5’
spacing near, but not in the sinkhole.
The purpose of this survey was to establish the background soil
temperature near the sink and to establish the variability in temperatures
unrelated to a buried cave. 30
temperature readings were taken along this profile. The temperatures ranged from 32.1 to 33.9
degree F, a range of +/- .9 degree F from the overall average background
temperature of the 30 measurements of 32.7 degree F. (Figure 7).
Figure 7.
Thermal Drift
Temperatures were measured continually during the
noise and both sinkhole surveys at a nearby base station inside the sinkhole
and were found to vary only +/- .1 degree F throughout the time of the survey.
The purpose of the base station is to make sure that soil temperatures were not
changing radically during the time the surveys were conducted (Figure 8).
Figure 8.
Sinkhole Surveys
Since heat would be flowing out of a buried
cave entrance in the winter, I was looking for a “hot spot” in the
sinkhole. Portions of a 12-foot x
12-foot area in the base of the sinkhole were surveyed. The sink was surveyed at both .5 and 1-foot
depth. A thermal “hot spot” was found in
the south corner of the sinkhole near the lowest topographic position in the
sink. The maximum temperature recorded
was 36.2 degree F, at 12”, 3.5 degree higher than the average background
temperature of 32.7 degree F (Figure 9).
Figure
9 (Sinkhole Location-Traverse is marked
on Figure 11).
The
“hot spot” was seen both at the 6” depth and 12” depth of investigation
(Figures 10 and 11).
Figure
10.
Figure
11.
A
3-D view of the temperature distribution of the 12” readings show the anomalous
warm temperature in the southwest corner of the sinkhole (Figure 12).
Figure
12.
Measurements were taken at 3 different depths at
position A (see Figure 9) in the sinkhole:
Table 1
Depth 1 .25”
- 32.0 degree F
Depth 2 6” -
35.0 degree F
Depth 3 12” -
36.2 degree F
It is the rate of change in thermal gradient with
depth that will help in determining the probable depth to a cave.
August
1991 Sinkhole Survey
No
formal noise survey was conducted. Base
station readings inside the sink varied +/- .1 degree F.
Since
it was summer time, I was looking for a “cool spot” inside the sink. The ambient air temperature was 90.0 degree
F. The background soil temperature was
estimated to be 68.0 degree F. A 12- foot
x 12- foot area was again surveyed at a depth of 6”. The lowest temperature inside the sinkhole
was 65.5 degree F, or 2.5 degree lower than the average background reading:
location B (Figure 13). The location of
the December 1990 “hot spot”(Location A) and the August 1991 “cool spot’’
(Location B) were nearly the same in the southwest quadrant of the sinkhole.
Figure
13.
Analysis
and Interpretation
Noise Survey
The standard deviation of the noise survey data
is .53 degree F. Background temperatures at 1-foot depth were
estimated to be 32.7 degree F +/- 1.1 degree F (or +/- 2 standard deviations -
Figure 9). Temperatures fluctuations
within this range were not significant and probably represented random changes
in temperature and measurement. At +/- 2
standard deviations, there is over a 95% chance that data within +/- 1.1 degree
F of the mean (average) temperature still represent thermal noise. Temperatures outside this range suggest
anomalous readings that could be of interest.
Thermal Drift
Thermal drift was not seen to be a concern because
of the base station temperature stability (+/- .1 degree F see Figure 8). However when the noise survey was evaluated
for thermal drift by fitting a linear function to the noise survey data, a +. 4 degree F drift was seen in the
interpreted background temperature from 32.5 degree F to 32.9 degree F (Figure
14).
Figure 14.
The
interpolated drift during the sinkhole survey was another +.1 degree F. This
implies that the soil temperatures at a 1-foot depth were warming by .5 degree
F throughout the duration of both the noise and sinkhole surveys (125
minutes). A technique called residual
analysis was done to assure that the thermal drift (trend) was removed from the
data. There was a decrease of the +/-2
standard deviations to +/- 1.0 degree F when the thermal drift was removed from
the noise data (improvement of +/- .1 degree F), implying that the thermal
drift did impact the sinkhole survey, but not significantly.
December 1990 Sinkhole Survey
A
comparison of the noise survey and a profile view of the temperature in the
sinkhole are shown in Figure 9. The
maximum temperature is + 6.6 standard deviations above background temperature: it is statistically improbable that the
maximum temperature reading is random thermal noise.
Magnitude Above Background
Another
indicator of the 36.2 degree F temperature being significant is the magnitude
of this temperature above background temperatures and then the comparison of
this difference with thermal noise magnitude.
A =
Max Sink Temperature = 36.2 degree F
B =
Background Temperature= 32.7 degree F
C =
Thermal Noise =+/- 1.1 degree F
Magnitude
above Background (MAB) = (A-B) / C = 3.2 or the December 1990 temperature
anomaly is 3.2 times as strong as the estimated thermal noise.
August 1991 Sinkhole Survey
The August 1991 sinkhole survey confirmed the
general location of the thermal anomaly
that was seem in the December 1990 sinkhole survey. Both the 1990 and 1991 surveys suggested that
the cave was located in the southwest portion of the sinkhole were the
anomalies overlapped each other. The
opening was thought to be about 6 feet across, based on the shape of the
thermal anomaly. The depth of the cave
(thickness of material covering the cave) was estimated to be 9’ from the
surface, based on plotting of the depth temperature measurements (Table 1)
inside the sinkhole and projected the temperature gradient with depth to the
likely cave temperature (Figure 15).
Figure 15.
Test of Interpretation
In May of 1992, permission was obtained from the
landowner to start an excavation in the sinkhole between locations A and
B. First a shallow test well was drilled
with a Hydradrill fitted with a posthole drill bit that encountered several small
voids in the sinkhole at 5’. Then in
June 1993, an excavation contractor dug a 3’ by 6’ trench in the south part of
the sinkhole over the location of the thermal anomalies. The cave was discovered at a depth of 12’
(Figure 16).
Figure 16.
Figure 17.
The opening
has subsequently been enlarged to allow for easier access (Figure 17).
Description of Cave
The original
width of the opening to the cave was about 8-10’ by 2 feet tall. The water table in the cave was at 849’. The cave is over 220 feet long, and up to 40
feet in height (Figure 18).
Figure 18.
The cave was scuba dived several years ago and the
trip report is on
www.cavediver.com/hall/morris.htm
Depth Detection Limit
An estimate of the depth of detection of a cave in
this area is seen in Figure 19. This
figure plots the magnitude above background (MAB) values with depth. The MAB at depths of 0’, 11’ and 11.5’ over the known cave are
plotted. The theoretical delectability
of a cave is the depth at which the MAB is equal to or greater than 1. When the December 1990 survey was run, the
depth at which the MAB was equal or greater than 1 was estimated to be 16’: the maximum depth of detection of a cave in
this area on this date.
Figure 19.
Lessons Learned
What was learned from the micro-temperature surveys?
(1)
The best candidate for thermal surveying is one that shows signs of
cavern collapse
(2)
Avoid surveying on sunny or windy days when radiant and convective heat
will interfere with the surveys
(3)
Maintain a base station to check that temperatures are not changing too
much
(4)
Repeat surveys in both winter and summer seasons to verify the location
and validity of the anomaly
(5)
Thermal drift may occur even if base station readings don’t suggest it
is occurring
(6)
Use the longest and most accurate temperature probe possible
(7)
Keep the survey time as short as possible to minimize thermal drift
(8)
The noise survey is critical for determining the validity of an anomaly
(9)
If only one survey is possible, do it in wintertime. The reason is that most investigations will
occur in topographic lows. Density
variations in air will cause cold air to drop into these topographic lows. A warm anomaly in the wintertime has a better
chance of being valid, because it has to overcome the natural tendency for
sinkholes to be cooler.
Summary
The
micro-temperature surveys helped pick the preferred dig area within this large
sinkhole. The cave in Figure 1, is the
excavated cave of this case history. The
effect of the melting snow in proximity to the cave entrance is the exaggerated
effect of having the warm cave air in contact with the cold outside air. The sinkhole fill muted the temperature
differential across the fill: it
disguised the cave, but still allowed thermal evidence to be gathered to
support the interpretation of the cave’s existence.
With a depth of 9’ originally estimated to dig, an excavator
was used to get to the cave. The
temperature surveys not only helped find the cave, but also helped save money
by putting the excavator to work on the part of the sinkhole that had the best
chance of finding the cave ($100/per hour for the tracked excavator). It is
estimated that at least $ 800 was saved by digging in the spot predicted to yield
the cave, rather than a general dig in the sinkhole.
Recommendation
I recommend that micro-temperature surveys should be
tried and evaluated for their predictive capability for locating caves. With
proper initial selection of the candidate for surveying, selecting an
appropriate time and date to survey, use proper instrumentation, careful
repeatable surveys, recording of base temperature readings and thermal drift
(trend) analysis to account for thermal drift, the integration of all available
relevant data, and knowledge and experience in an area (including good old
common sense), temperature surveys can be a valuable tool in the prioritization
and placement of your next dig.
If you have any questions about this analysis or
need advice on what kind of equipment to use or how to conduct your own
temperature survey, please e-mail me at