United States
Environmental Protection
Agency
Hazardous Waste Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-85/089 Dec. 1985
Project Summary
Delineating Toxic Areas by
Canine Olfaction
L. D. Arner, G. R. Johnson, and H. S. Skovronek
A research project was undertaken to
learn how the highly acute olfactory
sensitivity of the canine could be ap-
plied with advantage to environmental
problems. The objectives were to deter-
mine how dogs could be trained to
detect hazardous and toxic pollutants in
the environment and how the dogs'
responses could be used by environ-
mental workers to improve sampling
efficiency and to help delineate contam-
inated sites that might be encountered
in spills and improper disposal incidents.
Three dogs were trained to recognize
and locate chemicals selected from the
toxic and hazardous chemical lists. One
of these dogs was trained to respond
upon detection of chemical scents at
extremely low airborne concentrations
such as those that might exist at the
outer perimeter of a disposal site.
Throughout the project, the safety and
health of both dog and handler were
carefully considered.
Over a relatively short period, two
dogs were successfully trained to rec-
ognize toluene, and 2,4,5- and 2.4.6-
trichlorophenol at levels that could not
be detected as quickly or efficiently
using conventional field instrumenta-
tion. These dogs were trained to seek
out and retrieve chemically contami-
nated articles or to dig at the site of a
simulated ground contamination. In a
field experiment, both dogs success-
fully demonstrated their ability by locat-
ing as little as 0.2 g of chemical from
distances as great as 50 ft.
A third dog was acclimated to another
chemical, 1,2,3-trichloropropane, to
prepare for a field test at a nearby
Superfund site contaminated with this
material. This dog was trained to sit
immediately when he detected the
specified odor. This technique allows a
dog to delineate the perimeter of a con-
taminated area without entering the
dangerous zone defined by conventional
instrumentation. A field experiment at
the Superfund site was carried out
under extremely adverse weather con-
ditions; it provided encouraging but
inconclusive results.
This Project Summary was developed
by EPA's Hazardous Waste Engineering
Research Laboratory, Cincinnati. OH,
to announce key findings of'thfresearch
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Containment and cleanup of hazardous
materials at disposal and spill sites re-
quire some determination of the pres-
ence and extent of the pollutant beyond
ground zero. Such a project often callsfor
extensive air, water, and soil sampling
followed by costly and time-consuming
analyses. Such data are needed to delin-
eate the cleanup area and to determine
which areas require protective gear for
workers and exclusion or evacuation of
neighbors.
Even with the most sophisticated equip-
ment and the most skilled personnel,
characterization of such sites is often a
tedious and costly process. In addition,
the lack of information concerning a sus-
pect site or the location of wastes within a
particular site makes it necessary to take
a large number of samples, often on a
random basis, before the scope of the
problem can be defined. Depending on
the nature of the material under investi-
gation, the analytical procedures may be
-------
so complex that the results are generated
too slowly to be of much help in defining
the site. Certainly the large number of
contaminant-free samples required con-
tributes to the slowness of this site char-
acterization process. For example, dioxin
testing in one Missouri program required
more than 10,000 samples. Of these,
approximately 8000 were found to be
negative. Unfortunately, with the current
state-of-the-art in site evaluation, such
testing has been unavoidable.
When a site is known to be heavily and
widely contaminated (such as at a spill or
well-documented disposal site) it is often
necessary to know whether constituents
are leaving the site as dust or vapors and
if so, how far they are travelling and in
which directions. Current procedures call
for environmental workers to establish
sampling stations (e.g., high-volume air
monitors) at various distances and in var-
ious directions to evaluate the movement
of the pollutants. The results of these
tests can be affected by weather and
again introduce an undesirable time lag
between occurrence and availability of
data.
Environmental researchers have made
extensive passive use of animals in var-
ious forms of biomonitoring. Ranging
from the well known LD5o test with fish,
flies, etc. to the use of free-swimming and
caged fish as an indicator of water qual-
ity, these tests have well-established
credibility in the environmental research
community.
Surprisingly, the sensitivity of such
species have yet to be put to more active
use in environmental programs. The dog,
with his acute scenting ability, his high
trainability, and his history of working
closely with man, is uniquely suited for
such an innovative approach. Canines
have already demonstrated the desired
scenting ability in many areas akin to
environmental programs. Explosives and
narcotics discovery and the tracking of
people are fully recognized, daily uses of
the dog. However, applying the dog's
scenting prowess to environmental prob-
lems remains to be developed.
The current project applies to olfactory
acuity of the dog to environmental areas
where it offers time or cost benefits. Spe-
cifically, the project seeks to determine
(1) whether trained dogs can assist envir-
onmental workers in locating specific
chemical contaminants in the environ-
ment, thereby reducing the need for ran-
dom sampling of suspect areas, and con-
currently (2) whether the dog's response
can help delineate the perimeter of known
contaminated areas, thereby helping to
distinguish areas where protective gear
must be worn from those areas safe for
nearby residents.
Search and Retrieval Program
Training
Two dog/handler teams with extensive
backgrounds in scent work were selected
for this phase of the program to save time
and resources. Toluene was selected to
represent a volatile hydrocarbon that
might be found at industrial sites and
gasoline storage tanks. The dogs were
taught to recognize aboutO.5 g of toluene,
which was somewhat higher than the 0.1
g level planned for eventual use. The
chemical was placed on a cotton ball in a
wooden dowel or a perforated 35-mm
film canister (see Figure 1). This method
gave the dogs something to retrieve, yet
protected them from direct contact with
the chemical. The dogs quickly learned to
recognize the odor, and the quantity was
reduced progressively but quickly to the
target level of 0.1 g. The dogs were also
trained simultaneously to locate airborne
vapors of the chemical from greater and
greater distances, both indoors and in the
field, and even after the chemically-scent-
ed articles had been allowed to age for up
to 24 hrs. (when the toluene would pre-
sumably have evaporated). The dogs
exhibited no reluctance in finding and
retrieving the articles, even from distanc-
es of as much as 50 ft. Presumably they
accomplished these feats by detecting
the movement of minor vapor compo-
nents in the air reaching them. In addition
to learning the target odor, the dogs also
had to be taught to disregard other dis-
tracting odors such as those of the articles
themselves and the handlers.
When the dogs had achieved the basic
objective, a second chemical, 2,4,6-tri-
chlorophenol, was introduced. This mate-
rial was chosen as a potential indicator or
simulator for the dioxin class of com-
pounds. Later, 2,4,5-trichlorophenol, the
isomer commonly associated with the
most common dioxin isomer (2,3,7,8-
TCDD) was obtained and also located
successfully in a series of tests. These
results suggested that the dogs either did
not differentiate the two isomers, made
an association between the two com-
pounds, or understood that their task was
to find the new chemical. Since both of
these compounds are solids, they were
applied to the cotton as a 10% solution in
methanol and allowed to air dry before
the dogs were asked to seek them. The
dogs were also taught to ignore methanol.
The protocol used for the training is
essentially that used in training dogs for
narcotics or explosives detection. The
dogs learn to recognize airborne vapors of
the target material and follow them back
along a concentration gradient (presum-
ably) to the source. The protocol, called
operant conditioning, relies on positive
reinforcement with food and/or praise
for all successes (even very minor ones)
Figure 1. Articles used for training dogs.
-------
and negative reinforcement, through
withholding of the reward and/or praise
for incorrect decisions. Discipline is not
normally part of this protocol.
Field Tests
Once the dogs were trained, a simu-
lated field search was carried out at EPA's
Edison, New Jersey area facilities. An
outdoor test was conducted on a half acre
area where various obstacles were placed
in the field to simulate an actual site. To
simulate an indoor search, an area inside
a warehouse was equipped with 55-gal
drums, wooden pallets, tires, concrete
rubble, etc. About 3 hr before the test was
to begin, both areas were planted with
several dowels and film canisters con-
taining 0.25 g toluene or 2,4,6-trichloro-
phenol, unscented articles, and several
ground impregnations with both chem-
icals.
Results
The outdoor test used one dog, who
almost immediately located a toluene-
scented article hidden at the support jack
of a trailer from about 50 ft away (see
Figure 2). From this area, the dog/
handler team moved quickly to a ground
impregnation of toluene and then to a
series of tires on the ground where the
dog indicated one of the tires but did not
retrieve an article. Her handler investi-
gated and found a film canister (toluene)
partially submerged in water in the tire.
Finally, working in a small depression
where several unscented articles were
planted, the dog retrieved a blank dowel
after much investigation. This article had
been disturbed during a 3-hr delay period
and covered with a rock to prevent visual
detection. New odor may have been
introduced at that time.
The indoor test differed primarily in that
there was essentially no air movement in
the large building. The second dog was
used here, and he also uncovered several
toluene-scented articles and a concrete
block impregnated with a drop of 2,4,6-
trichlorophenol before ending with a
blank article that was only about 1 ft away
from a scented one.
No trichlorophenol-scented articles
were retrieved during the tests. Since this
result was inconsistent with the training,
the procedures were reexamined and it
was found that all the 2,4,6-trichloro-
phenol articles had been prepared with
only 0.05 g of the chemical, or 25% of the
expected amount. Nevertheless, the re-
sults still raise some questions, since the
dogs consistently retrieved aged 2,4,6-
Figure 2. Dog locating toluene-scented article hidden at the support jack of a trailer.
trichlorophenol articles that probably
contained no more than this amount of
the compound.
Attempts to measure airborne levels of
toluene using a Foxboro Century 128
Organic Vapor Analyzer (gas chromato-
graph) were largely unsuccessful, both at
the field test and in laboratory exper-
iments. No instrument readings were
obtained in the field or the laboratory
unless the probe was adjacent to the
sample or unless very large samples were
used and a draft was induced with a fan.
Perimeter Delineation
The second goal of the project was to
learn whether a dog could indicate the
presence of chemical odors at a distance
from a source while NOT moving forward
to the source. Since the two dogs used in
the earlier work had been trained to
locate the source, another untrained dog
was used for this work. To prepare for a
field test at an actual Superfund site, this
test used a chemical expected at that
site—1,2,3-trichloropropane.
Training
The initial training of the dog followed
the protocol described earlier. The dog
was trained to recognize and distinguish
the 1,2,3-trichloropropane odor with as
little as 0.1 g. Because the chemical was
expected to have a continuous source at
the field site, the training did not incorpo-
rate an aging period beyond about 1 hr.
Once the dog understood that he was to
search for 1,2,3-trichloropropane, he was
required to sit immediately on detecting
the odor to receive his reward, which was
food. Because the instinct to move to the
source is so strong and so useful in rein-
forcing the training, and because the
handler often needed some confirmation
that the dog had really detected the chem-
ical, the dog was occasionally allowed to
move further into the scent cone. This
step increased the dog's confidence and
also helped the handler to understand or
"read" the dog.
A larger, nonpoint source was then
simulated by replacing the small samples
of pure compound with 8 in. diameter
pans containing a layer of dilute (25 ppm)
aqueous solution of 1,2,3-trichloropro-
pane. With practice, the dog could detect
such scented "puddles" from as far as 25
ft and immediately alert his handler by
sitting. The odor dissipated from such
sources in 0.5 to 1 hr, and the dog could
no longer locate the pans reliably.
Field Test
The field experiment took place on
March 28,1984, at the Tyson's Wastesite
near King of Prussia, Pennsylvania. As a
result of cleanup work, run-off was now
channeled through an activated carbon
treatment unit before it was discharged to
the nearby Schuylkill River. The site con-
sisted of a narrow (100- to 200-ft) strip of
brush-and tree-covered lowland between
the river and a railroad right-of-way.
Beyond the railroad tracks, the land rose
-------
sharply to a ridge that was perhaps 200 ft
high.
The weather on the test day was
extremely poor, with heavy rain, strong
gusty winds, and a temperature near
40°F. Though the team was concerned
that the dog might be able to detect the
target odor even far downwind under
these conditions, access problems forced
entry from that direction and about 0.25
to 0.5 mile from the treatment plant.
Immediately upon initiating the test, the
dog appeared to be tracking enthusiasti-
cally into the wind and toward the treat-
ment plant. As the team moved forward
on the track right-of-way at the edge of
the undergrowth, which was too thick to
follow the dog into, the dog's interest
appeared to diminish, although he con-
tinued to lead the team forward. Some
renewed interest occurred when the dog
and handler came upon the treatment
plant. After passing the plant, the dog
changed direction for the first time and
returned to the fencing surrounding the
facility. The dog was then taken to the
outfall pipe and several ground seeps in
the area where the trichloropropane was
expected to be present, based on an anal-
yses made several months ago. The dog
showed no interest or response until he
was brought to one of the major seeps—a
puddle about 3 ft across, where he finally
did sit for the first time. His handler asked
for and received a confirmatory sit before
rewarding the dog. Additional searching
of the area failed to elicit any further re-
sponse from the dog.
An organic vapor analyzer had been
brought along for this trial and was
occasionally used to sample the air and
the surface of puddles. Though readings
of as much as 5 ppm of organic vapor (not
necessarily 1,2,3-trichloropropane) were
observed, these results did not coincide
with the dog's reactions or with the
team's own fleeting detection of gas-like
odor. No positive readings were obtained
at the outfall or at the seep where the dog
had sat.
The results of this test are, at best,
ambiguous. Perhaps the dog was always
in the presence of the chemical, even as
the test started, 0.25 mile or more from
the source. If so, he may not have known
how to respond to this new situation, he
may not have been able to get the needed
ON-OFF stimulus for a reaction, or he
may simply have become desensitized by
continued low-level exposure as the team
moved forward. Other possibilities in-
clude (a) inadequate training of the dog to
cope with very adverse weather condi-
tions, (b) masking of the chemical odor by
4
other air constituents, (c) inadequate
airborne concentration because of high
winds, or (d) disappearance of detectable
levels of the chemical in the surface waters
in the months since the site cleanup had
been completed. Other than samples
taken with the organic vapor analyzer, no
air or water samples were taken on the
test day to determine the presence or
absence of the chemical.
Conclusions
This project demonstrates the feasi-
bility of using trained dog/handler teams
to locate small quantities (i.e., pockets) of
pollutant sources or low-level discharges
(airborne or waterborne) from toxic or
hazardous sites. The work also indicates
that such olfactory detection can assist
environmental workers in the early char-
acterization of such sites.
The dog can detect and locate very
small quantities of chemicals from con-
siderable distances, even where instru-
ments are unable to detect residual
vapors.
Once a dog has been trained in the
general protocols of search and retrieval,
he can be trained to locate a specific
pollutant in the environment with relative
ease and speed.
Because of the dog's ability to move
about an area and find and follow scents
carried by the wind at concentrations not
detectable by instruments, the dog and
his handler can be used to search large
areas more quickly and more efficiently
than a person with a portable instrument.
Consequently, the dog can localize pock-
ets of pollutants for more effective use of
subsequent sampling and analytical pro-
cedures.
Considering the extremely low levels
detectable by the trained dog (probably far
below those usually defined as toxic or
hazardous), the dog/handler team can
carry out screening programs with min-
imal risk, even while free of cumbersome
protective gear.
Indications are that under certain con-
ditions, a dog can indicate the presence of
toxic or hazardous material vapors eman-
ating from a disposal site or an accident at
quite a distance. By interpreting the dog's
behavior and the weather conditions, the
handler can quickly estimate the direction
and distance such pollutants have trav-
eled.
Recommendations
Though this project has established the
preliminary feasibility of using dog/han-
dler teams to locate pollutants in the
environment and to define an outer, safe
perimeter for contaminated areas, con-
siderably more work is needed to make
this knowledge a practical tool that envi-
ronmental workers can use in emergency
situations.
Clearly, a great deal more must be
learned about the range of the dog's
abilities in terms of acuity, uniformity
(from dog to dog and day to day), sensitiv-
ity to different chemicals, and selectivity.
A key question is whether the dog can
recognize and associate classes of com-
pounds (e.g., chlorinated hydrocarbons).
Another great need is to determine
whether the dog can be trained to react to
levels other than his minimum detection
level. In other words, there is a need to
know whether the dog's olfactory ability
can be made more quantitative, as sug-
gested by some of the results observed in
this study.
A variety of field situations need to be
considered as potential applications for
the dog. I n this way, the scope of the dog's
abilities and applications can be better
defined. Specific areas to consider include
using dogs to test the decontamination of
equipment and personnel at cleanup
sites, and to locate pockets of specific
pollutants at or near suspect sites. Detec-
tion of dioxins (or trichlorinated phenols)
near manufacturing sites, polychlorinated
biphenyls (PCBs) in the vicinity of dam-
aged or leaking transformers, and gaso-
line leaking from underground storage
tanks are three situations in which dogs
could effect large savings in time and
analytical costs.
Careful and extensive analytical sup-
port should be provided in parallel with
future canine programs, both to provide
information on the levels being tested
and to provide workers with insight into
the relative time and cost factors involved
with the two approaches.
The use of the dog should also be
considered in areas other than that of
hazardous and toxic materials. For ex-
ample, it should be feasible to use dogs to
locate fugitive volatile organic carbon
emissions from valves, fittings, pumps,
etc. at manufacturing facilities. Similarly,
dogs may be able to track the contamina-
tion of surface waters back to their
sources and thus assist enforcement
personnel in locating illegal discharges.
Though the dog is ideally suited for
working with man on environmental
problems, the concept of using other
animals (whales, dolphins, seals, birds)
should also be considered as innovative
means of tracking contamination of the
oceans or the atmosphere.
. S. GOVERNMENT PRINTING OFFICE:1986/646-l 16/20726
-------
L. D. Arner is with Biosensors, Inc., Westmoreland, NY 13490; G. Johnson is with
Guardian Training Academy, Windsor, Ontario; and H. S. Skovronek is with
Environmental Services, Morris Plains, NJ 07950.
Hugh Masters is the EPA Project Officer (see below).
The complete report, entitled "Delineating Toxic Areas by Canine OH'action,"
fOrder No. PB 85-235 596/AS; Cost: $8.50, subject to change) will be available
only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-85/089
------- |