Geology Reference
In-Depth Information
Another early study of coal fires by the US Geological Survey (Greene et al., 1969) investigated 22 fires in the
anthracite fields of Northeastern Pennsylvania and in the bituminous coal fields near Pittsburgh in Western
Pennsylvania. A Reconofax IV instrument was used for these surveys and imagery was acquired just before
dawn and during the daytime. The predawn TIR imagery proved to be more effective because of greater contrast
(Greene et al., 1969).
Greene et al. (1969) suggested that TIR could detect coal fires at shallow and intermediate depths (i.e., those less
than 30m below the surface). They suggest that TIR was effective for detecting deeper coal fires only if they had
been burning for more than 10 years. The authors suggested that this amount of time was necessary for the heat
from the deeper fire to be conducted to the surface. However, based on anecdotal information, most coal fires have
been burning for extended periods and so the 10-year requirement should not be a problem. For example, the
Centralia coal fire has been burning for 40+ years, the Laurel Run coal fire has been burning for 90+ years, and the
Burning Mountain coal fire in Australia (Ellyett and Fleming, 1974, see below) has been burning for at least 180
years.
The US Bureau of Mines reported a study of coal fire control projects (Dierks et al., 1971) where TIR imaging was
used to survey the Cedar Avenue coal mine fire on the southeastern edge of Scranton, Pennsylvania. The authors
noted that thermal anomalies were detected in a landfill and in nearby burning banks; however, these were only part
of the area affected by the coal fire. Dierks et al. (1971) suggested that the lack of detection was due to the depth of
the coal fire and that TIR was ineffective for detecting coal fires deeper than 30m. However, the authors gave no
indication of the type of TIR instrument used for the survey, the time of day the survey was flown, or the time of
year the survey was flown, which are all factors that can affect whether the TIR instrument would detect deeper
areas of the coal fire.
Ellyett and Fleming (1974) describe a TIR study conducted over The Burning Mountain coal fire in New South
Wales, Australia. This coal fire has been known since at least 1829, and evidence (e.g., slag and baked sediments)
suggested that the fire had burned over a distance of 6 km and so may have been burning for several thousand
years. Hot gases were escaping to the atmosphere from a moving chimney at this site, which is similar to the
moving fronts at Centralia (Nolter and Vice, 2004). The authors conducted three separate TIR surveys (during mid-
day, evening, and dawn) using a Daedalus scanner over the Burning Mountain and were able to delineate some of
the features of the coal fire but not others. They did not give an IFOVor temperature sensitivity for the instrument
they used.
The authors Ellyett and Fleming (1974) concluded that the solar heating made detection of any anomaly associated
with the coal fire difficult in the mid-day TIR survey. Ground measurements by the authors indicated that there was
little lateral conduction of heat from the chimney. The limited aerial extent of the TIR anomaly associated with the
coal fire meant that it could easily be overpowered by any solar anomaly. The evening TIR survey also had
extensive anomalies created by solar heating (Ellyett and Fleming, 1974), which tended to mask the anomaly
associated with the chimney. Only the dawn TIR survey had an easily detected anomaly associated with the
chimney (Ellyett and Fleming, 1974, p. 83). This is comparable to the author
s experience in conducting TIR
surveys for geothermal surface manifestations in the Washington Cascades and in Western Montana (Vice, 2007).
The TIR surveys needed to be conducted in the early morning hours in order to avoid solar anomalies (Vice, 2007).
Greene et al. (1969) and Hong et al. (1996) also noted that predawn airborne thermal IR surveys were best for
detecting coal fires.
'
Bakker et al. (1978) reviewed some uses of a custom-built TIR scanner with a temperature sensitivity of 0.15°C
and note that imagery obtained during the African dry season was better for most uses because the vegetative cover
was less during this time. As part of this review, the authors described a coal fire at Witbank in the Republic of
South Africa. Most of the warm areas in the TIR imagery are associated with subsidence areas on the land surface
(the coal bed is at a depth of 30m and was mined using room and pillar). The authors suggest that the general area
of the fire is visible in the imagery but the fire
'
s extent cannot be determined.
Miller and Watson (1980) described the use of TIR imaging to detect coal fires in the Sheridan area of Northeastern
Wyoming. This area is part of the Powder River basin, which contains large deposits of subbituminous coal. The
study was based on surveys in July of 1975 and October of 1978 using a Texas Instruments RS14A scanner with an
IFOVof 3 milliradian. The authors noted that a number of surface factors including (1) meteorological conditions,
(2) physical property differences (e.g., such as differences in the composition of soil and rocks), (3) topographic
differences, and (4) near-surface geothermal effects (which probably refers to the movement of groundwater) could
