How Accurate are Search Dogs? – Part 1: Area Detection Dogs

“The canine sense of smell is legendary and, as with many legends, it is sometimes hard to find the sources and separate scientific facts from urban legends”  – Goldblatt et al. (2009)

A well trained search dog should not make frequent mistakes, but how accurate are they?  How often do search dogs miss locating what they are trained to find?  How frequently do they make a mistake and indicate something is there when it is not?

There are several different types of search dogs including area detection, tracking and trailing dogs.  This article explores the scientific research on the accuracy of different types of area detection dogs.  This type of search dog is trained to search an area (determined by the handler) and indicate if they find one or more target scents.  Most area detection dogs are not trained to find the scent of a single individual (i.e. scent specific).  Dogs may be trained to find many different scents including humans (search and rescue dog), cats (cat detection dog), bombs (explosives detection dog), cancer (medical detection dogs), household pests (bed bug or termite detection dogs) or certain species of wildlife scat (conservation dog).

Scent Detection Dog Research Studies

Several research studies on search dog accuracy measured two factors: 1) how often did the search dog miss a target scent and 2) how often did the search dog misidentify a scent (i.e. give a false positive).  Helton (2009) reviewed the results of 12 different detection tasks including cancer detection, scat detection, termite detection and bomb detection.  He found that search dogs found and correctly identified the target scent 91.61% of the time (range = 75.00% to 100.00%).  Looked at another way, the search dogs missed indicating a target scent in an average of 8.39% of trials (range = 0% to 25%).  Helton (2009) also found that the search dogs gave false alarms (e.g. indicated a scent was present when it was not present) an average of 3.42% (range = 0% to 18.20%).

Additional studies conducted in the field provide more insight into the accuracy of search dogs.  Three field research studies (Harrison 2006, Long et al. 2007 and Smith et al. 2003) found that the search dogs located the correct species of wildlife scat 89% of the time (range = 72% to 100%).  While it is often difficult to determine how frequently search dogs missed locating scat in the field, Wasser et al. (2004) found that search dog handlers (rather than search dogs) located 0% to 25% of bear scat samples.  Reindle-Thompson et al. (2006) used search dogs to determine the presence of ferrets in prairie dog colonies.  One dog was 100% accurate at finding ferrets while the second dog only correctly indicated the presence of ferrets in four out of seven (57%) prairie dog colonies.

Possible Factors Affecting Scent Detection Dog Accuracy

As these studies indicate, the majority of detection dogs are accurate most of the time, but they are not infallible.  The training and experience of the search dog team (that is both the dog itself and the dog’s handler) are probably the biggest factors affecting how often a search dog makes mistakes.  Environmental factors can also have a strong influence on whether the search dogs misses locating a target scent such as a buried bomb or wildlife scat.  Some environmental factors that may make a search more challenging are hot dry air (Smith et al. 2003), gusty winds, lack of wind, and topography of the search area.  MacKay et al. (2008) estimated that scat detection dogs were able to locate a scat anywhere from 100 meters away to only 1 meter away depending on environmental conditions.

Still other factors that can influence a search dog’s accuracy are physical fitness and health (Altom et al. 2003 and Gillette 2004).  A hot or tired dog that is panting heavily is not able to sniff as effectively and fewer scent particles are detected.  Certain medications such as steroids (Ezeh et al. 1992) and some diseases can also effect a dog’s ability to identify and process scents (Furton and Myers 2001).

Accuracy versus Success Rates for Scent Detection Dogs

Whether or not the search dog is accurate is important, but that does not mean that they find their intended target 80 to 90% or more of the time.  In order to have a successful find, the target scent must be present in the search area, and the search area is determined by the search dog handler or another person.  Therefore, a large part of their success is determined by the knowledge and experience of the search dog handler, especially when finding lost people or lost cats.  In human search and rescue multiple search dog teams are often used and each is assigned a separate quadrant to search.  In lost cat searches, often a single dog is used to search the most likely areas where the cat will be found.  Typical success rates for cat detection dogs may be as high as 20% or as low as 5%.  In both human and lost cat searches, many other tools besides the search dog are also used to improve the chances of a successful find.

Recommended Reading

Helton, William, editor.  2009.  Canine Ergonomics: The Science of Working Dogs.  CRC Press.

Long, Robert, Paula MacKay, William Zielinksi, and Justina Ray, editors.  2008.  Noninvasive Survey Methods for Carnivores.  Island Press.

Literature Cited

Altom, E.K., G.M. Davenport, L.J. Myers, and K.A. Cummins.  2003.  Effect of dietary fat source and exercise on oderant-detecting ability of canine athletes.  Research in Veterinary Science 75: 149-155.

Brooks, S.E., F.M. Oi, and P.G. Koehler.  2003.  Ability of canine termite detectors to locate live termites and discriminate them from nontermite material.  Journal of Economic Entomology 96: 1259-1266.

Ezeh, P.I., L.J. Myers, L.A. Hanraham, R.J. Kemppainen, and K.A. Cummins.  1992.  Effect of steroids on the olfaction of dogs.  Physiology and Behavior 51: 1183-1187.

Furton, K.G. and L.J. Myers.  2001.  The scientific foundation and efficacy of the use of canines as chemical detectors for explosives.  Talanta 54: 487-500.

Gazit, I. and J. Terkel 2003.  Domination of olfaction over vision in explosives detection by dogs.  Applied Animal Behavior Science 82: 65-73.

Gazit, I, A. Goldblatt, and J. Terkel.  2005.  Formation of olfactory search image for explosives odours in sniffer dogs.  Ethology 111: 669-680.

Gillette, R.L. 2004.  Optimizing the scenting ability of the dog.  Athletic and Working Dog Newsletter, May.

Goldblatt, Allen, Irit Gazit, and Joseph Terkel.  2009.  Chapter 8 – Olfaction and Explosives Detector Dogs.  Canine Ergonomics: The Science of Working Dogs.  CRC Press.

Harrison, R.I. 2006.  A comparison of survey methods for detecting bobcats.  Wildlife Society Bulletin 34: 548-552.

Helton, William.  2009.  Chapter 5 – Overview of Scent Detection Work: Issues and Opportunities.  Canine Ergonomics: The Science of Working Dogs.  CRC Press.

Kauhanen, E., M. Harri, A. Nevalainen, and t. Nevalainen.  2002.  Validity of detection of microbial growth in buildings by trained dogs.  Environment Internationl 28: 153-157.

Kiddy, C.A., D.S. Mitchell, D.J. Bolt, and H.W. Hawk.  1978.  Detection of estrus-related odors in cows by trained dogs.  Biology of Reproduction 19: 389-395.

Long, R.A., T.M. Donovan, P. MacKay, W.J. Zielinski, and J.S. Buzas.  2007.  Effectiveness of scat detection dogs for detecting forest carnivores.  Journal of Wildlife Management 71: 2007-2017.

MacKay, P. D.A. Smith, R.A. Long and M Parker.  2008.  Chapter 7 – Scat Detection Dogs.  Noninvasive Survey Methods for Carnivores.  Island Press.

McCulloch, M. et al. 2006.  Diagnostic accuracy of canine scent detection in early- and late-stage lung and breast cancers.  Integrative Cancer Therapies 5: 30-39.

Pickel, D. et al.  2004.  Evidence of canine olfactory detection of melanoma.  Applied Animal Behavior Science 89: 107-116.

Reindl-Thompson, Sara A., John A. Shivik, Alice Whitelaw, Aimee Hurt, Kenneth F. Higgins.  2006.  Efficacy of Scent Dogs in Detecting Black-Footed Ferrets at a Reintroduction Site in South Dakota.  USDA National Wildlife Research Center – Staff Publications.  Paper 438.

Smith et al. 2003.  Detection and accuracy rates of dogs trained to find scats of San Joaquin kit foxes (Vulpes macrotis mutica).  Animal Conservation 6: 339-356.

Wasser, S.K., B. Davenport, E.R. Ramage, K.E. Hunt, M. Parker, C. Clarke, and G. Stenhouse.  2004.  Scat detection dogs in wildlife research and management: application to grizzly and black bears in the Yellowstone ecosystem, Alberta, Canada.  Canadian Journal of Zoology 82: 475-492.