Allegheny Mountain Rescue Group

Search-and-Rescue Dogs Function and Deployment

 by Ken Chiacchia, PhD, WEMT-B

 A good dog team can be one of the most valuable assets to a SAR operation. Using estimates based on AMRG's investigations into canine sweep width, an airscent dog can do the job of as many as six human searchers. Unfortunately, a poorly trained or tasked team becomes a liability. And when winter strips the foliage or a search subject is wearing high-visibility colors, ground searchers can in fact be more effective than the best dog team.

The proper place for a dog team is as an integrated part of an incident-command-system run SAR operation that makes full use of the strengths of many different resources. That said, I'd like to outline the current understanding that the SAR dog community and researchers have created regarding dogs and their proper deployment. In this article, I'm going to talk about "wilderness" SAR dogs. These dogs find lost or injured people in a wilderness, rural, or undeveloped suburban property setting.

Types of SAR dogs

There are two types of SAR dog. The trailing dog follows the human scent trail on the ground. The air-scenting dog follows airborne scent directly to the search subject. (See the figure.)














Each type of dog has its strengths and weaknesses; each can play a crucial role in a successful search, if used properly. Some dogs are cross-trained in both specialties.   In our experience, there are few absolutes regarding which breed of dog is best suited for either type of SAR work. While a number of air-scent dog handlers prefer herding breeds such as German shepherds or border collies, and many trailing handlers field bloodhounds, we've seen people effectively work numerous breeds, including mixed-breed dogs.

It's important to remember that humans invented the distinction between trailing and airscenting. In fact, a trailing dog will air scent if that's how the strongest scent is coming to her; and an air-scenting dog will follow a hot ground trail. But each dog is most reliable when performing the task that she's been specially trained to do, and handlers tailor their tactics in the field to take advantage of the dog's strengths.

Another, overlapping, type of specialty training exists: the cadaver dog. Scent can escape from water, earth, or even cracked or porous concrete. Dogs can thus be used to find human bodies that have sunk or been buried, as well as those on the surface. Many air-scenting and trailing dogs are cross-trained on human cadaver scent so that they can find human remains if they come across them.  In our team, we use the term "cadaver dog" to refer to SAR dogs that have been cross-trained to detect cadaver scent. We refer to dogs that have trained intensively with cadaver scent and are more accurate and reliable for finding trace human remains, or those that are very old, as "forensics dogs."

Trailing Dogs

Generally, a trailing dog begins working at the search subject's last known location -- called the point last seen (PLS) or last known point (LKP). From there, the trailing dog follows the ground scent left behind by that person, traveling more or less along that person's route of travel until either she finds him or loses the trail for some reason.   Trailing dogs must be trained to scent discriminate. When given a scent article that only the subject has touched, the dog can follow that person's trail -- even if it crosses trail scent left by other people. Trailing dogs sometimes, but not always, work on leash.

Trailing dogs are a powerful SAR resource. If a ground trail is relatively fresh, if searchers have a good PLS or LKP, and if they can find an uncontaminated scent article, a good trailing dog can be the fastest, most direct way of getting to the subject.  But hand in hand with those strengths come corresponding weaknesses. While moisture and warmth are generally good (they help bacterial production of odor molecules, as we'll discuss later), rain and the passage of time can destroy a ground trail. So can ill-considered search efforts at the LKP -- a common mistake is to set up a command post on the LKP, thus destroying any chance of finding ground trails.  Often a scent article isn't available -- or just as bad, has been contaminated with other people's scent. In these cases, an uncritical reliance on the dog can lead to disappointment -- and in an unfair assessment of the dog's abilities. To minimize such contamination, scent articles should be handled as little as possible before being given to the dog.

A word of warning about scent discrimination: it is an extremely powerful technique, but only if done properly. Research studies have revealed that scent-discriminating dogs can be "skunked" if presented with a task that is sufficiently different from the problems they've trained on.  As with any other facet of SAR dog work, we're making use of dogs' innate abilities -- but we can't be sure we have the ability to direct and take advantage of those abilities unless we've trained under realistic conditions. Scent discrimination work should not be undertaken lightly, and must be done well if it's done at all.

I should probably make a brief mention of tracking dogs, which are technically different from trailing dogs. Competition tracking dogs follow a person's tracks, step by step. These dogs probably follow the smell of crushed vegetation and other disturbance rather than the subject's body smell. SAR trailing dogs probably both track and trail when they're working -- a dog needs the body scent to discriminate between individuals, for instance, but may follow crushed vegetation as well.

Air-scenting dogs

Air-scenting dogs detect human scent on the wind. They tend to pick up scent coming directly from the search subject, and so may take a more direct route to the subject than a trailer would. But first the dog has to be downwind of the subject, and thereby hangs the tale.

Methodically "sampling" an area's smells as a dog team moves through it is one of the challenges of air-scenting dog handling. Because of the need for the dog to range unrestricted by the handler to find airborne scent plumes, air-scenting dogs work off-leash.  Early in a search, experienced SAR incident commanders generally will employ air-scenting dog teams for "hasty search" patterns, designed to get the "easy" find quickly and collect information for later in the search. Late in a search, air-scent teams are useful for sweeping, or gridding, large areas.

Air-scent handlers generally don't do discrimination/trailing work with their dogs, instead letting the dogs find and report on any human who is in the area. In practice, this is not usually a drawback: anyone in the area is a potential witness with information about the search subject, and so needs to be interviewed. On the other hand, discrimination-trained air-scenting dogs can be a particularly flexible, useful SAR resource.

An air-scenting dog's lack of specificity can be a problem in a highly contaminated area. If a search area isn't cleared of hunters, hikers, or other searchers (the time necessary is highly dependent on local conditions), a dog team can waste a lot of time finding many wrong people, or even pockets of residual scent. More experienced dogs, especially if they've done some training in semi-crowded park and suburban environments, have some ability to compensate for these problems.

An air-scenting dog can often detect and localize human scent from 100 yards or more -- the longest refind in AMRG's database is 270 yards. The longest such detection on record is two miles, set by a dog on the Alaskan tundra. But these results are highly dependent on weather conditions. AMRG's investigations in canine sweep width suggest dog probabilties of detection (PODs) at 100 yards ranging from about 80 percent or more on cloudy, windy days and clear nights, to 65 percent or less on still, sunny days.

It's important to remember that, while air-scenting dogs can and do find people, much like human searchers they more often find clues. A dog may catch a whiff of a person's scent that's too brief to follow -- but a watchful handler will notice this "alert," and search commanders can note the location of the alert and the wind direction on their maps. Another search team, dog- or human-based, can then check in that direction to see if the person is there.   To be honest, this level of dog team/command team integration is probably not realistic until both groups have worked with each other and become comfortable with each other's capabilities over time. The clue value of reported dog alerts is highly variable, depending on the amount of human traffic in and adjacent to the search area, previous scent contamination, and perhaps most importantly, the experience level of the dog's handler.

I've been both the dog handler trying to convince incident staff of the importance of an alert and the operations officer trying to make sense of mutually contradictory alerts from different dog teams. Both points of view are valid, and the only way to get past uncertainty is to work together and find a balance with which everyone is comfortable. We dog handlers also have to accept that the value of scent clues has never been thoroughly proved, and may not be as useful as we'd estimate from our memory and experience.

Scent and atmospheric conditions

The study of how weather affects scent is that of a lifetime. But dog handlers have learned a few simple tricks for making nature work for them -- or at least for preventing nature from completely disrupting their efforts.

At daybreak, both the air and the ground tend to be relatively cool. But sunrise makes two changes in this more-or-less stable situation: it serves as the engine for wind movement, and it heats the ground.  Sunlight passes right through the air without heating it much. But the sun does heat the ground relatively quickly. The result, early in the morning, is cool air over warm ground. The ground warms the air directly above it; once warmed, that air tends to rise past the still-cool air higher up.  Even in the absence of true wind, these "convection currents" will tend to produce a gentle uphill breeze. For this reason, early in the morning dog handlers like to be at the tops of hills or ridges, where their dogs can pick up any scent rising from the valleys below. Experienced SAR incident commanders will, therefore, try to task dogs before the sun starts warming the ground, so that they are in place to take full advantage of the updrafts when they begin.

Later in the day, especially sunny days, these uphill convection currents can get too strong. Warm air rising through cool air tends to be chopped into little pieces -- watch how the air seems to crinkle over hot pavement in the summer -- and any scent carried by that air gets chopped up as well. Aside from the fact that it's difficult to keep a dog safely watered on a warm, sunny day, convection can often destroy air scent so completely that a team of ground searchers would be able do a better job than a dog. This can even happen on clear, sunny, but cold winter days.

Around sundown, though, this picture changes enormously. Now the air is thoroughly warmed. But just as the air doesn't gain heat easily, it doesn't lose heat easily either. The ground gives off heat in the form of infrared radiation, cooling quickly, and now we have warm air over cool ground. Now the ground cools the air slowly, and the convection currents work in reverse, with air tending to sink downhill. This is a great time for dog searching: a dog at the bottom of a valley can get a good smell of the slopes on either side.   On clear nights, eventually much of the day's heat escapes into space, with the air and ground reaching equally cool temperatures. At this point -- called stable conditions by meteorologists -- scent travels with no loss to convection, and very long-range detection by scent is possible. For this reason, dogs are an extremely efficient resource for night searching.

Clouds trap heat and reflect it back to the ground, leading to intermediate conditions. While winds (the long-range movement of air with weather fronts, as opposed to convection currents) are generally good because they make scent move faster and farther, they cause other complications. Wind can overwhelm convection currents; it can also can combine with them to produce complicated spiraling air currents. Generally, air-scenting dog handlers try to get downwind of an area to work; but isolated convection cells or turbulent eddies can can trap scent on the downwind side of buildings or sharp downhill slopes, making detection downwind difficult.  Rain can wash scent out of the air, forcing normally air-scenting dogs to keep their noses low or even washing out scent completely. Falling snow can also scavenge scent components from the air; we've found that dry, powdery snow (generally rare in western Pennsylvania) is especially "scent sticky" and can magnify the effects of any scent contamination in the area.

The nature of scent   

Human beings are messy. As we walk, we shed a continuous stream of gaseous molecules. In addition to gaseous molecules, people trail behind them microscopic flakes of dead skin. These "skin rafts" are tiny scent generators. The bacteria, perspiration, and skin secretions they carry generate more scent molecules with time. The different components of the cloud of debris that we emanate act very differently. Larger flakes of dead skin tend to fall to the ground relatively quickly. They, and the scent molecules they continue to generate, become a trail of ground scent.

Other scent components -- smaller skin raft particles and gasses -- can travel indefinitely on the wind, becoming the air scent.  Some of the smelliest of human scent molecules are created when bacteria create carboxylic acids out of skin oils, and when the body releases "panic scents" -- steroid-like molecules produced by the apocrine glands, found in the armpits, groin, and in smaller numbers elsewhere on the skin. Research in the 1980s and 1990s revealed that this combination of carboxylic acids and odiferous steroids seems to play a major role in the human body odor that humans can smell. Interestingly enough, these same chemicals have also been implicated in studies of how animals (mostly rodents) discriminate between other individual animals by scent. I suspect that these substances are what SAR dogs smell when they're looking for people, but it hasn't been proved yet.

A research group from the University of Florida and the U. S. Department of Agriculture recently cataloged at least 277 airborne compounds that come off human skin; it will be interesting to see which of these, and which combinations of them, trained SAR dogs will respond to. Many exciting and useful discoveries about the nature of human scent may be just over the horizon. For a more complete discussion of this issue, see In Search of Human Scent, and Who Goes There? The Body's System for Generating Individual Scent.

*Parts of an earlier version of this article appeared in Advanced Rescue Technology,  October/November 2000, pages 29 and 31. Copyright 2000 by Summer Communications, Inc. Reproduced with the permission of the publisher.   

Further reading

American Rescue Dog Association, Search and rescue dogs: Training methods, Howell Book House, 1991

Axel, R., "The molecular logic of smell," Scientific American 10:154-9, 1995.

Graham, Hatch, "Probability of Detection for Search Dogs: or How Long is Your Shadow?"

Pearsall, M. D., and Verbruggen, H., Scent, Alpine Publications, 1982.

Stern, A. C. et al., Fundamentals of air pollution, Academic Press, 1974 (especially chapters 16 to 20).

Syrotuck, W. G., Scent and the scenting dog, Arner Publications, 1972.

Wysocki, C. J., and Preti, G., "Human body odors and their perception," Japan Journal of Taste and Smell Research 7(1):19-42, 2000.