DESCRIPTION / INTRODUCTION
Dogs have an extraordinary ability to detect and distinguish odors. With specialized training, they can locate explosives and drugs, find missing people or human remains, detect pests and wildlife, and even recognize odor changes associated with some medical conditions. Yet canine detection is not simply a matter of having a “better nose.” Performance depends on the dog, handler, training, target odor, environment, and way in which the team’s performance is tested. These 20 questions explain how canine olfactory detection works, where it is used, what can affect its reliability, and how detection results should be interpreted.
At a Glance: Common Uses of Canine Olfactory Detection
| Type of detection | What the dog searches for | Examples |
|---|---|---|
| Explosives detection | Odors associated with explosive materials | Airports, public events, military and police work |
| Narcotics detection | Odors associated with specific controlled substances | Law enforcement, border and customs work |
| Human search and rescue | Human odor associated with living persons | Missing persons, disasters, wilderness searches |
| Human-remains detection | Odors produced during human decomposition | Crime scenes, disasters, forensic searches |
| Tracking/trailing | Odor associated with a particular person’s movement or scent | Missing-person and law-enforcement searches |
| Pest detection | Odors associated with insects or other pests | Bed bugs, agricultural pests |
| Conservation detection | Odors from animals, plants, scat, nests, or other biological materials | Endangered species, invasive species, wildlife surveys |
| Medical detection | Changes in human odor associated with physiological or disease processes | Research involving cancer or infectious disease; medical-alert applications |
| Other specialized detection | A trained target odor | Currency, firearms, agricultural products, electronic-storage devices, and other targets |
Dogs have been used in an increasingly wide range of biological, security, forensic, conservation, and medical applications. The underlying principle is the same: the dog learns to recognize an odor pattern associated with a target and to communicate detection of that odor to its handler. (Kokocińska-Kusiak et al., 2021).
1. What is canine olfactory detection?
Canine olfactory detection is the use of a dog’s sense of smell to locate or recognize a particular target odor.
A trained dog searches its environment for the odor it has learned to recognize. When the dog detects that odor, it communicates the finding to its handler, usually through a trained behavior such as sitting, standing still, staring at the source, lying down, barking, or another response.
The dog is therefore not merely “smelling something.” Dogs smell many odors continuously. Detection work requires the dog to distinguish a particular trained odor from many unrelated odors and then respond appropriately.
Detection dogs work together with human handlers, so it is often more accurate to speak of a canine detection team rather than of the dog alone. (ANSI/ASB, 2020; Kokocińska-Kusiak et al., 2021).
2. How does a dog’s sense of smell compare with a human’s?
A dog’s olfactory system is much more specialized for detecting odors than a human’s.
Estimates vary by breed and by how receptors are counted, but dogs are commonly reported to have roughly 125–300 million olfactory receptors, compared with about 5–6 million in humans. Dogs also have a much larger area of olfactory tissue and specialized nasal anatomy that helps direct odor-containing air toward that tissue. Their sniffing behavior further improves the collection and analysis of odor molecules. (Jenkins et al., 2018; Kokocińska-Kusiak et al., 2021).
It is tempting to turn these differences into a statement that a dog’s nose is a certain number of times “better” than a human’s. That is misleading. Sensitivity varies enormously with the chemical being detected. Dogs may detect some substances at extremely low concentrations, sometimes approaching parts-per-trillion levels, but there is no single number describing how much better a dog’s sense of smell is for every odor.
The dog’s advantage comes from a combination of receptor number, nasal structure, airflow, brain processing, sniffing behavior, and the ability to learn and discriminate odor patterns, not from receptor count alone.
3. What types of substances or targets can detection dogs be trained to find?
The range is remarkably broad.
Dogs have been trained to detect explosives, narcotics, firearms-related odors, currency, human scent, human remains, agricultural products, insects such as bed bugs, wildlife, invasive species, animal scat, microorganisms, and odors associated with certain diseases or changes in human physiology.
What matters is not whether humans think the target has a noticeable smell. The target must produce (or be consistently associated with) volatile chemicals that the dog’s olfactory system can detect and that can be presented appropriately during training.
A dog usually does not understand that it is searching for “a bomb,” “cancer,” or “a bed bug.” It learns an odor or odor pattern that has been associated with a reward. (Kokocińska-Kusiak et al., 2021).
4. How are detection dogs trained?
Most detection training is based on reward and conditioning.
First, the dog learns that finding a particular target odor leads to something it values, such as a favorite toy, play, food, or praise. Over repeated training sessions, the dog learns to search for that odor and ignore irrelevant odors.
Training usually progresses from relatively simple problems to increasingly difficult ones. The target may be hidden in different locations, mixed with distracting odors, reduced in concentration, or presented in unfamiliar environments. Dogs also learn a specific response that communicates a find to the handler.
Good training includes both positive samples, in which target odor is present, and blank or negative searches, in which it is absent. Otherwise, a dog may learn that it is always expected to find something.
Training does not end when the dog begins operational work. Detection teams require continuing practice, testing, documentation, and exposure to realistic search conditions. (ANSI/ASB, 2020).
5. What is an “indication” or “alert,” and why does it matter?
The terminology can vary among organizations, but it is helpful to distinguish several behaviors.
A dog may first show a change of behavior when it encounters an interesting odor (for example, slowing down, changing its breathing, turning its head, or searching more intensely).
A trained final response, often called an indication or alert, is the behavior the dog has specifically been trained to perform when it locates its target odor.
That response may be passive, such as sitting, staring, standing still, or lying down. It may also be active, such as barking or scratching, although passive responses are preferred in situations where disturbing the target could be dangerous or damaging.
The distinction matters because normal interest in an odor is not necessarily the same thing as a trained alert. The handler must learn to read the dog’s behavior without turning every change in behavior into a reported detection. (ANSI/ASB, 2020).
6. How accurate are detection dogs?
There is no scientifically defensible single accuracy percentage that applies to all detection dogs.
Performance depends on what the dog is detecting, how the dog was trained, odor concentration, search conditions, handler behavior, and, very importantly, how accuracy was measured.
Researchers commonly evaluate at least two kinds of error:
- A miss or false negative occurs when target odor is present but the dog fails to identify it.
- A false alert or false positive occurs when the dog gives its trained response even though target odor is absent.
Performance may also be described using sensitivity, the proportion of target-positive samples correctly detected, and specificity, the proportion of target-negative samples correctly rejected.
Very high performance can occur under carefully controlled conditions, but results from one laboratory study should not automatically be applied to an airport, crime scene, hospital, wilderness search, or other operational setting.
Medical-detection studies illustrate the problem particularly well: published sensitivity and specificity values have varied widely, and differences in study design and blinding make direct comparison difficult. (Bauër et al., 2022).
The best question is therefore not simply, “How accurate are dogs?” but “How well has this particular dog-and-handler team been shown to perform this particular task under conditions similar to those in which it will be used?”
7. What factors affect a dog’s detection performance?
Many factors can affect performance.
Environmental conditions influence the way odor molecules move. Wind may carry odor away from its source or create an odor plume that the dog follows. Heat, cold, humidity, rain, surfaces, ventilation, and air currents can also change how odor becomes available.
The dog’s own condition matters as well. Illness, nasal inflammation, fatigue, dehydration, physical exertion, heat stress, some medications, age, and other physiological factors can affect olfactory performance. (Jenkins et al., 2018).
Training history, motivation, experience, unfamiliar surroundings, distracting odors, and the complexity of the search also matter.
Finally, detection is a team activity. The handler can help the dog conduct an efficient search, but can also unintentionally influence the result.
8. Which breeds are commonly used for detection work, and does breed matter?
Labrador Retrievers, German Shepherd Dogs, Belgian Malinois, Springer Spaniels, and several other breeds are commonly used in detection work. Bloodhounds have long been associated with human scent trailing.
Breed can influence characteristics such as body size, endurance, search style, motivation, and some aspects of olfactory anatomy. However, breed alone does not determine whether a dog will make a good detector.
Programs usually select individual dogs for characteristics such as strong motivation to search, willingness to work for a reward, physical health, confidence, persistence, trainability, environmental stability, and ability to work cooperatively with a handler.
A highly suitable individual dog from a less commonly used breed may be a better candidate than an unsuitable individual from a traditional working breed. Standards for detection-dog selection therefore emphasize the individual dog’s physical and behavioral suitability rather than breed alone. (ANSI/ASB, 2021).
9. How long does it take to train a detection dog, and how is proficiency maintained?
There is no universal training period.
Training time depends on the dog’s previous experience, the type and number of target odors, the complexity of the search task, certification requirements, and whether the handler is being trained at the same time.
Initial odor recognition may be learned relatively quickly, but reliable operational performance requires much more. The dog must learn to search unfamiliar environments, ignore distractors, work at different odor concentrations, respond correctly when no target is present, and communicate reliably with its handler.
Once operational, teams need maintenance training throughout their careers. Records should document training, problems encountered, corrective training, testing, and operational performance. Periodic proficiency testing or recertification may also be required by the agency or discipline. (ANSI/ASB, 2020).
10. What is the difference between detection dogs, tracking dogs, and trailing dogs?
These terms describe related but somewhat different uses of scent, and terminology is not completely uniform among organizations.
A detection dog usually searches for a particular target odor or category of odor (for example, explosives, drugs, human remains, or bed bugs) without necessarily following a route traveled by the source.
A tracking or trailing dog follows human scent associated with a person’s movement through an environment.
Traditionally, tracking often refers to following a relatively precise path or track, sometimes including scent and disturbances on the ground. Trailing generally refers to following an individual person’s scent and may allow the dog to work farther from the person’s exact footsteps as odor moves through the environment.
In practice, methods and terminology overlap. Human-scent dogs may also be trained for area searches, location checks, scent discrimination, or other specialized tasks. NIST/OSAC therefore maintains separate standards for several human-detection disciplines rather than treating all human-scent work as one activity.
11. Can detection dogs be used for medical or disease detection?
Yes, but two different applications should be distinguished.
Medical-alert assistance dogs may work directly with an individual and respond to changes associated with conditions such as hypoglycemia or certain seizure-related events.
Biomedical detection dogs used in research are trained to distinguish samples (such as breath, urine, sweat, or other biological materials) from people with and without a particular condition.
Researchers have investigated canine detection of cancers, infectious diseases, and other conditions. Many studies have produced promising results. However, results vary widely among studies, and important questions remain about sample selection, training methods, blinding, confounding odors, and exactly which volatile chemicals the dogs are detecting.
A systematic review of cancer and infectious-disease studies found considerable variation in sensitivity and specificity and noted that relatively few studies used fully double-blind, screening-like designs. The authors concluded that the evidence was not yet sufficient for canine scent detection to replace established clinical diagnostic methods. (Bauër et al., 2022).
Medical detection is therefore a promising field of research, but a dog’s alert should not by itself be considered a medical diagnosis.
12. Are there health or welfare concerns specific to detection dogs?
Yes. Detection dogs are working animals, and their welfare directly affects both their quality of life and their performance.
Potential concerns include heat stress, dehydration, fatigue, musculoskeletal injury, nasal or respiratory disease, environmental hazards, and exposure to potentially harmful substances. Operational programs should consider whether the dog might contact or inhale hazardous materials and should use appropriate safety procedures.
Adequate rest, conditioning, veterinary care, hydration, nutrition, safe housing, and appropriate limits on working time are important. Programs should also monitor behavioral welfare. Chronic stress, poor handling, excessive workload, or inappropriate training can affect both the dog and the reliability of the work.
Modern canine standards therefore address not only detection performance but also selection, kenneling, health, conditioning, and veterinary care. (ANSI/ASB, 2021; Jenkins et al., 2018).
13. How can handlers influence or unintentionally bias detection results?
Dogs are extremely attentive to human behavior. This is useful in normal dog-handler communication, but it creates a potential problem during detection work.
A handler who expects the target to be in a certain place may unknowingly change walking speed, leash tension, posture, gaze, or other behavior. The handler may also interpret an ambiguous canine behavior as an alert because of an expectation about where the odor should be.
This is sometimes compared with the “Clever Hans effect,” in which an animal responds to unintended human cues.
In a well-known experiment involving experienced drug- and explosive-detection teams, handlers were deliberately given incorrect information suggesting that target odor might be present at marked locations. No target odor was actually present. Handler beliefs significantly influenced where alerts were reported. (Lit, Schweitzer, & Oberbauer, 2011).
For this reason, blind testing, in which the handler does not know the target location, provides a stronger test of the dog-handler team’s ability. In some research settings, double-blind procedures are preferable so that neither the handler nor the person interacting with the team can unintentionally reveal the answer.
14. What certifications or standards exist for detection dogs and teams?
There is no single universal certification that covers all detection dogs.
Police departments, military organizations, search-and-rescue groups, private organizations, and other agencies may use different testing and certification systems.
In the United States, the National Institute of Standards and Technology’s Organization of Scientific Area Committees for Forensic Science (OSAC) maintains a Dogs & Sensors standards area. Consensus standards published through organizations such as the Academy Standards Board address subjects including canine selection and health, general training and certification, explosives detection, human-remains detection, and various forms of human-scent detection.
Certification should not be viewed as proof that a dog will always be correct. It shows that the team met a specified performance standard under particular test conditions. Continuing training, quality assurance, record keeping, and periodic reassessment remain important afterward.
15. What are the limitations of canine olfactory detection compared with technology?
Dogs and instruments have different strengths.
Dogs are mobile, fast, highly sensitive, trainable, and capable of searching complicated environments. They can follow odor through space and often locate an odor source rather than merely reporting that a chemical is present in a collected sample.
But dogs are biological detectors. They become tired, differ from one another, require training and care, and may be affected by health, environment, motivation, or handler influence.
Analytical instruments can measure and identify particular chemicals objectively and repeatedly and can produce stored numerical data. However, instruments may require collected samples, laboratory equipment, more time, or a known chemical target.
For many applications, dogs and technology should therefore be considered complementary rather than competing tools. A dog may be excellent at rapidly locating an odor of interest, while another test or instrument may then help confirm or identify what is present. (Kokocińska-Kusiak et al., 2021).
16. What is a detection dog actually smelling?
Usually, the dog is not smelling the abstract thing that humans have named.
Odor results from volatile chemical compounds released into the air from a substance, organism, person, or biological process. A complex target may release many compounds at once. Together these can form an odor pattern, or “odor signature”, that a trained dog learns to recognize.
For example, a biomedical detection dog trained to distinguish samples from people with a disease may not be smelling the disease itself. It may recognize a pattern of volatile organic compounds produced directly or indirectly by changes in metabolism, inflammation, microorganisms, medication, diet, or other biological processes.
Likewise, a dog searching for an explosive may sometimes respond to volatile compounds associated with the explosive formulation rather than to every ingredient equally.
This is one reason careful research and training controls are necessary: investigators must make sure the dog has learned the intended target odor rather than an unrelated odor that happens to occur with it.
17. Can a dog detect an odor after the object or person that produced it is gone?
Yes. This phenomenon is often called residual or lingering odor.
Odor molecules may remain on surfaces, in soil, clothing, containers, rooms, vehicles, or other environments after the original source has been removed. A dog’s nose may be sensitive enough to detect that remaining odor.
This creates an important interpretation problem. Suppose a dog alerts at a location but investigators cannot find the target. One possibility is that the dog made an error. Another is that the target had previously been present and left detectable residual odor.
For this reason, professional terminology distinguishes a confirmed false alert from an unconfirmed alert or non-productive response. An unconfirmed alert means that the target cannot be verified; it does not necessarily establish whether the dog was right or wrong. NIST specifically notes that residual odor is one possible explanation.
18. How should the accuracy of a detection dog be tested?
A meaningful test should resemble the task the team is expected to perform.
Testing should include known positive targets, negative or blank searches, realistic distractor odors, appropriate target concentrations, and different locations. The team should not simply be tested repeatedly under conditions it has memorized.
Whenever possible, the handler should be blind to target placement. Otherwise, conscious or unconscious handler cues may influence the result.
A good evaluation should report more than the percentage of searches called “correct.” Important measures include:
- Sensitivity: How often does the team correctly find a target when it is present?
- Specificity: How often does the team correctly reject a search when the target is absent?
- False-positive rate: How often does it alert when target odor is absent?
- False-negative rate: How often does it miss target odor that is present?
The exact performance standard depends on the discipline and consequences of error. Missing an explosive, for example, has different implications from making an unnecessary additional inspection.
Standards therefore emphasize documented, objective testing rather than assuming that one generic accuracy figure describes every detection dog. (ANSI/ASB, 2020).
19. What are false alerts, missed detections, and unconfirmed alerts?
These terms describe different outcomes and should not be treated as interchangeable.
A false alert occurs when the dog gives its trained final response even though the target odor is absent.
A miss, or false negative, occurs when target odor is present but the dog fails to alert.
An unconfirmed alert occurs when the dog alerts but investigators cannot determine whether trained target odor was actually present. The dog may have made a mistake, but residual odor or another unverified source could also explain the response.
This distinction is especially important outside controlled testing. In an experiment, researchers know where target odor was placed and can classify an alert accurately. In a real-world search, the true history of a location may be unknown.
Calling every unconfirmed response a “false positive” can therefore make a canine team appear less accurate than the evidence supports; automatically assuming that every alert reflects a real target creates the opposite error. (NIST/OSAC).
20. Does a detection dog’s alert prove that the target is present?
No. A canine alert is evidence that the trained dog detected an odor it associated with its target, but the meaning of that alert depends on the circumstances.
There are several reasons for caution. The dog could make an error. The handler could misinterpret its behavior. The dog might detect residual odor from something no longer present. It might also have learned an unintended odor during training.
For many applications, the best practice is therefore to use canine detection as one part of a larger decision process. An alert may direct investigators toward a location, sample, package, person, or object that deserves further examination. When independent confirmation is available, it can help determine what actually produced the odor.
This limitation does not make canine detection unhelpful. The dog’s great strength is precisely its ability to rapidly screen large, complex environments for extremely faint odors. The important point is to distinguish detection from confirmation.
That distinction is particularly important in medical detection, forensic work, scientific research, and other settings in which conclusions can have major consequences. Medical-detection reviews, for example, caution against treating experimental canine performance as equivalent to an established diagnostic test.
REFERENCES
American Academy of Forensic Sciences, Academy Standards Board. (2020). ANSI/ASB Standard 088-20: General guidelines for training, certification, and documentation of canine detection disciplines. Colorado Springs, CO.
American Academy of Forensic Sciences, Academy Standards Board. (2021). ANSI/ASB Standard 085-21: Standard for detection canine selection, kenneling, and healthcare. Colorado Springs, CO.
Bauër, P., Leemans, M., Audureau, E., Gilbert, C., Armal, C., & Fromantin, I. (2022). Remote medical scent detection of cancer and infectious diseases with dogs and rats: A systematic review. Integrative Cancer Therapies, 21, 15347354221140516. doi:10.1177/15347354221140516.
Jenkins, E. K., DeChant, M. T., & Perry, E. B. (2018). When the nose doesn’t know: Canine olfactory function associated with health, management, and potential links to microbiota. Frontiers in Veterinary Science, 5, 56. doi:10.3389/fvets.2018.00056.
Kokocińska-Kusiak, A., Woszczyło, M., Zybala, M., Maciocha, J., Barłowska, K., & Dzięcioł, M. (2021). Canine olfaction: Physiology, behavior, and possibilities for practical applications. Animals, 11(8), 2463. doi:10.3390/ani11082463.
Lit, L., Schweitzer, J. B., & Oberbauer, A. M. (2011). Handler beliefs affect scent detection dog outcomes. Animal Cognition, 14, 387-394. doi:10.1007/s10071-010-0373-2.
National Institute of Standards and Technology, Organization of Scientific Area Committees for Forensic Science. (2023). Dogs & Sensors terminology: False alert, non-productive response, and unconfirmed alert. NIST OSAC Lexicon.
National Institute of Standards and Technology, Organization of Scientific Area Committees for Forensic Science. (2026). Forensic Science Standards Library: Dogs & Sensors standards. National Institute of Standards and Technology.