spatial locations of trench boundaries. This assistance would
provide the needed information for site remedial investigations.
The following is a discussion of the techniques and results of this
effort.
2.2 Data Collection
Remotely sensed data were collected through several DOE-HQ
sponsored programs that involved multiple U.S. government
agencies. Primary data collection programs included the
Government Applications Task Force (GATF), the Environmental
Task Force (ETF), and the Strategic Environmental Research and
Development Program (SERDP) Waste Site Study. Each had a
mission to demonstrate the utility of remotely sensed imagery to
detect and locate buried waste trenches under a variety of
conditions and to analyze the phenomenology underlying the
signatures observed on thermal imagery. Data collected included
a combination of historical imagery, multispectral remote sensing
data, and "ground truth" data to evaluate the accuracy of remotely
sensed data and to understand the thermodynamics of trench
cooling/heating.
Historical data were obtained from Federal photographic archives
and included both high and low spatial resolution aerial
photography and remote sensing data. This data did not provide
a complete historical perspective, but did include random
coverage of the period 1942 (prior to DOE occupation of the site)
to the present.
Thermal imagery collection was used to detect differential thermal
patterns that would be indicative of the differences in soil density
and moisture that are characteristic of trench disposal.
Multispectral remote sensing from instrumentation, such as the
Daedalus 1268, which was used to obtain thermal imagery.
Daedalus 1268 data was obtained during 1992 and 1994 by
EG&G Energy Measurements using DOE-owned equipment.
Spatial resolution of 1.5 to three meters per pixel were obtained.
"Ground truth" experiments were conducted during 1994 and
involved extensive measurements of soil temperature, soil
moisture, and meteorological conditions. These data were
collected and analyzed to understand the physical processes that
produced the trench signature observed on thermal imagery.
Arrays of ground sensors were set up at the SWSA-4 study site to
collect data on the surface and near-surface conditions in the
trench and in an adjacent control location. The choice and
configuration of sensors were:
° Soil Temperature: Measurements of soil temperature were
recorded at several points inside and outside the trench at
depths of one inch and a vertical profile at four, six, and 10
inches.
° Thermal Radiance: Infrared transducers were used to assess
the emitted energy in the longwave Infrared band (8-14
microns) over the trench and non-trench areas.
° Soil-Moisture: An array of instruments recorded soil water
potential (i.e., moisture) at two-inch depth, both inside and
outside the trench.
2.3 Results
Through analysis and fusion of the combination of historical
photography, thermal imagery, and “ground truth” data, the Oak
Ridge Operations Remote Sensing Program was able to derive an
accurate trench map of SWSA-4 that could be used by the
92
Remediation Manager to delineate trench boundaries. A
reproduction of this map is shown in Figure 1.
Statistical analysis of the “ground truth” data indicated the nature
of the thermal signature of the trench areas. Both the thermistor
and radiometer data showed that the differences between a trench
area and the control area (i.e., non-trench area) was most evident
at night, with the trenches typically cooler. Soil moisture
measurements showed that the trench area generally exhibited
greater soil moisture than the control area, which may account for
the observed thermal differences. The data also showed that the
spatial variation in temperature within a trench was larger than the
variation between trenches, suggesting the need for multiple
observations. During the daytime hours, the thermal difference
was not readily discernible.
The SWSA-4 results indicate that nighttime thermal imagery can
be successfully employed to help delineate waste trench areas at
known sites under a variety of conditions; however, one must
consider site ground cover when planning a thermal survey.
SWSA-4 is characterized by mowed grass fields and similar
ground cover for both the trench and non-trench areas. This
simplifies data interpretation. Similar studies at other DOE sites
with non-homogeneous surface conditions (varying species of
vegetation) indicate that analysis of thermal behavior is more
complex and that thermal image signatures are difficult to predict.
Consistent thermal signatures are associated with sites where the
surface conditions are more homogeneous, whereas sites with
mixed and complex vegetation can exhibit different behavior.
Furthermore, even if no buried waste was present, disturbed soil
typically exhibited the thermal signature observed at the trenches.
Thermal imagery must be co-analyzed with other historical and
“ground truth” information to provide positive confirmation of
trench presence or absence.
2.4 Summary
The results of this study have benefited DOE in two ways. First,
the study provided information that will be used directly in the
remediation and monitoring of SWSA-4. Second, the project
demonstrated a method for using thermal imagery to assist in the
detection and location of buried waste trenches. The methods
demonstrated in this project should be applicable to buried waste
at, potentially, hundreds of similar sites in the coming years.
Previous ground-based field investigation at SWSA-4 identified
two individual seep areas that contribute over 90-percent of the
radioactive strontium releases from the study area. The trench
map derived from remote sensing data was the key factor in
pinpointing localized sources feeding these major seeps. Current
remedial actions are focusing on controlling these few sources and
will provide a cost-effective interim action to reduce Strontium-90
releases and off-site risk. Without the remote sensing results, the
ability to quickly and effectively pinpoint the locations of the
individual sources would have been lost. The alternative for
controlling releases (i.e., cap the whole site, collect and treat
surface water) would cost in excess of $5 million more than the
current action of directly controlling the sources.
The procedures demonstrated here can be applied to numerous
other waste sites where remedial action is necessary to stop the
migration of contaminants from burial trenches. Candidate sites
of this type exist at numerous DOE reservations and military
facilities. The alternative to employing remote sensing technology
is to rely on extensive and costly ground sampling to precisely
locate the hazardous material. Direct boring into contaminated
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
