Conservation detection dogs are one of the most effective methods of non-invasive wildlife monitoring there is. In a 2021 review of 1,220 publications featuring wildlife detection dogs, dogs were found more effective than other methods in 88.71% of cases, and worse in only 0.98%. A past K9 Conservationists article discussed the strengths and weaknesses of detection dogs as a method of non-invasive wildlife monitoring. It highlighted circumstances in which detection dogs thrive, and those in which other methods may be more suitable. This article will build on that by explaining how detection dogs can pair with and complement other methods. First, it is important to have a baseline for how detection dogs compare to human searchers.
Detection Dogs and Human Searchers
In a 2019 study of scat detection and discrimination amongst related species with identical diets, expert humans distinguished between American mink and Eurasian otter scats with 89% accuracy on average while detection dogs achieved 95% accuracy on average. Additionally, dog-human search teams found 3–4 times more scats than human searchers alone. This indicates that dogs are more adept at not only species discrimination, but total detection.
In a 2020 study of bird and bat carcass detection among dog and human searchers on wind farms, dogs detected 77.3% of carcasses while humans detected 21.5%. Additionally, the performance of dogs was less impacted by carcass size and vegetation structure. This indicates that dogs work well even in certain environments that hinder human searchers.
In a 2020 study comparing the abilities of a dog versus human searcher to detect bilby scat, the dog found 89 out of 90 scats (98.9%) while the human found six out of 90 (6.7%). This demonstrates a large gap in the detection abilities of dog and human searchers.
All in all, detection dogs have proven to be leaps and bounds ahead of human searchers alone in terms of efficacy. They have also proven more effective than other methods in many cases. One such method is camera traps.
Camera Traps
What Are Camera Traps?
A camera trap is an automated camera with a motion sensor that captures a photo or video when triggered by movement.
Camera traps are a favored method of non-invasive wildlife monitoring because they can be left for weeks to months with no human interference. Studies featuring camera traps are highly replicable, as research teams across the world can use the same camera model. Camera traps provide 24/7 data on species behavior, location, population size, etc. Many are equipped with infrared flash, which allows them to capture clear images in low light without disturbing the subjects, making them highly effective for observing nocturnal species. Additionally, they can capture rare events such as wolf predation on sea otters.
On the other hand, camera traps often capture irrelevant “bycatch” images, including those triggered by wind. They also have high upfront costs. According to WWF, a typical mid-range camera costs between $300–500 while a typical 50-camera survey costs between $15,000–40,000. Camera traps can become damaged in extreme conditions such as high humidity and/or heavy rainfall, decreasing their efficacy. Theft, vandalism, and wildlife interference are ever-present risks. Additionally, camera traps are minimally effective for observing small, ectothermic, and aquatic species. This indicates that the efficacy of camera traps varies widely from study to study. Even when they are highly effective, other methods may be used to supplement them.
For example, one of K9 Conservationists’ co-founders Kayla Fratt uses both detection dogs and camera traps to study wolf and deer occupancy across islands in Alaska. Camera traps allow her to passively monitor many islands, gathering information on the animals that use a given space; this helps her estimate prey availability for the wolves. The detection dogs allow her to identify individual wolves and their diets. Together, the prey availability from camera traps and the diet items identified in scats found by dogs paint a fuller picture of what wolves eat in relation to what is available in an environment.
How Do Camera Traps Compare to Detection Dogs?
In a 2015 study comparing the cost and efficacy of camera traps and detection dogs for bobcat detection, dogs took only 2 days to achieve the 90% probability of detection that took camera traps 7-8 weeks. However, month-long dog surveys cost 33% more than 4-month camera surveys. The study concluded that the best method for individual projects was goal-dependent. This highlights the facts that all methods have their strengths and weaknesses, and no single one, detection dogs included, is perfect for every project.
Detection Dogs and Camera Traps: Mesoamerican Carnivores
A previous K9 Conservationists article discussed the role of detection dogs in a project collecting carnivore scat in the Maya Biosphere Preserve of Guatemala. Ellen Dymit, the PhD student at Oregon State University behind the project, continues to study Mesoamerican carnivores, and detection dogs continue to help. However, they raise unique considerations as living animals. Their biological needs, including plenty of breaks to prevent overheating, constrain working hours, especially in hot, humid environments like Guatemalan rainforests. Additionally, they are forbidden from some protected areas. For this reason, Dymit uses camera traps to gather data that detection dogs can not. Some she places high in rainforest canopies where dogs can not go. In addition to not having the same constraints as dogs, camera traps provide unique visual data on metrics such as prey availability.
Detection Dogs and Camera Traps: Cheetahs
In a February 2025 episode of the K9 Conservationists podcast, Kayla talks with Tim Hofmann of the Cheetah Conservation Fund about his research using detection dogs and camera traps to monitor cheetah populations in Namibia and beyond. His team started using dogs to collect more scats for genetic, hormonal, dietary, and parasite analysis, particularly near prominent landscape features such as boulders, termite mounds, and play trees. At play trees, groups of males engage in social behaviors such as grooming, chasing, and marking via defecation. These sites are jackpots for the large numbers of scats they contain. However, they produce biased samples because while female cheetahs visit, only males defecate at them. This is where dogs and camera traps intersect. Dogs are highly effective at finding these sites where camera traps can be placed to obtain behavioral information about all individuals onsite. In this way, Hofmann’s research is similar to Dymit’s in that camera traps gather visual data that dogs can not. Additionally, dogs can find scats in hidden places that humans easily overlook.
In Summary…
Camera traps
- Can be left for long periods of time in remote places with little to no interference
- Allow for highly replicable studies
- Provide 24/7 visual data, highly effective for observing nocturnal species, especially when equipped with infrared flash
- May capture “bycatch” images, which may be mitigated with AI image recognition softwares
- Typically have high upfront costs and may have high maintenance costs
- At risk of damage in extreme weather conditions
- At risk of theft and vandalism by animals and people
- Minimally effective for observing small, ectothermic, and aquatic species
Detection dogs
- Often able to detect targets more quickly and efficiently than camera traps
- Collect a different sort of data than camera traps; ie, scat versus images
- Detection leads to tangible samples for researchers to analyze
- Can locate sites where camera traps will be most effective
- Often more expensive than camera surveys
- Biological needs create working constraints
- Certain areas are inaccessible to dogs (i.e. rainforest canopies)
GPS Collars for Monitoring Animal Movement
What is GPS Tracking?
In the context of wildlife conservation, global positioning system (GPS) tracking is a method in which a tracking device, often a collar, is placed on an animal. This device sends signals to electronic receivers and a satellite, providing researchers with data on the location of tracked individuals. In addition to providing valuable insights into species movement and behavior patterns, GPS tracking can help reduce human-wildlife conflict. For example, by tracking GPS-collared elephants on the outskirts of Hwange National Park in Zimbabwe, experts at the International Fund for Animal Welfare (IFAW) help mitigate the risk of negative human-wildlife encounters such as crop raids. However, GPS tracking is not without downsides.
As fixing wildlife with tracking devices can be difficult and time consuming, it is generally unrealistic to track every member of a population. For this reason, only a subset is selected. While this subset can be carefully curated to reflect the broader population, the risk of bias remains. For example, there may be data gaps in areas where tracking devices are under-deployed or there could be bias in the subpopulations of animals that are more readily captured and collared than others. This means that GPS tracking alone does not provide a comprehensive look at wildlife behavior. Even for tracked individuals, the data is somewhat limited, though some trackers offer video, gyrometers, and/or accelerometers. Luckily, GPS tracking combines well with other methods for more detailed data collection.
Detection Dogs and GPS Tracking: Terrestrial Carnivores
In a 2021 study, researchers paired detection dogs with various spatial capture-recapture models to estimate the population abundance of black bears, bobcats, cougars, and coyotes. Estimates like these are tricky, especially for cryptic, low-density species like the terrestrial carnivores listed. The researchers used a combination of spatial capture-recapture (SCR), generalized spatial mark-resight (gSMR), and spatial count, or spatial presence-absence, models. While SCR requires that all individuals in an encounter be identified, gSMR only requires identification of a subset, and spatial absence-prescence requires no individual identification. In addition to these methods, detection dogs were used to collect scat for genotyping, and a subset of individuals from each species was affixed with GPS collars and resighted by remote cameras. The results indicated that hybrid models using all these methods were most effective, and that for future studies, at least a subset of the population should be individually marked.
Detection Dogs and GPS Tracking: Gray Brocket Deer
In a 2023 study, researchers monitored six Gray Brocket Deer with GPS collars for one year while simultaneously collecting fecal samples along transects using detection dogs. The purpose was to assess habitat selection and utilization among brocket deer. Because brocket deer are solitary, elusive, and mostly nocturnal, GPS/telemetry studies relying on opportunistic captures have historically been challenging. While scat sampling is generally a reliable method of non-invasive wildlife monitoring, factors such as fecal detectability, degradation, and habitat-dependent defecation behavior may present challenges when studying habitat use. It can be difficult to determine the age of feces, so fecal sampling lacks the temporal specificity of methods such as GPS tracking. For these reasons, researchers combined the methods and analyzed the convergence of data between them.
Both GPS and fecal sampling data indicated a preference among Gray Brocket Deer for forest and cerrado ecosystems, likely for rest, protection, and thermal comfort. Nonlinear movement patterns in these areas suggested foraging. In contrast, longer, straighter movements in open grasslands suggested passing through. Nonetheless, both environments hold ecological significance for the deer, demonstrating a need for conservation and sustainable land use. In this case, combining methods increased the soundness of the data, improving the validity of land use decisions.
In Summary…
GPS tracking
- Provides insight into species movement
- Can help reduce human-wildlife conflict
- Allows for individual identification
- Provides temporal data
- Pairs well with other methods
- Potential for sampling bias, as it is generally unrealistic to track every member of a population
- Provides somewhat limited data
- Challenging for species that are solitary, evasive, highly aggressive, etc.
Detection dogs
- Detect samples that provide data on metrics GPS can not measure
- Can help reduce the potential sampling bias of GPS
- Do not necessitate direct contact with wildlife
- Pair well with other methods
- Can not easily obtain temporal data
- Can not track individuals over large areas and long periods of time
Thermography
What Is Thermography?
Thermography is the process through which a camera, drone, or similar device picks up heat signatures in the environment and configures them into an image. Generally, cooler areas appear darker and more muted while warmer areas appear brighter. The efficacy of thermography is heavily influenced by factors such as weather conditions, vegetation, distance and size of object, and in the case of carcasses, stage of decomposition and larval density. Additionally, thermal imaging technology comes with high upfront costs. A drone can cost between $6,150 and $7,800 on the low end, and upwards of $20,000 on the high end. This does not include potential pilot service, training, and/or repair costs. For this reason, other methods may be used to supplement thermography.
Detection Dogs and Thermography: Brown Hare Leverets
In a 2020 study, researchers combined a handheld thermal imaging camera, thermal drone, and wildlife detection dog to determine each method’s efficacy in detecting brown hare leverets in northwestern Switzerland. Brown hares are an endangered species, but their small size, minimal movement, and cryptic nature make data collection challenging, especially when live animals are needed for radio- or GPS- collar fitting and behavior observation. For this reason, the researchers chose methods that did not require direct contact with the animals.
Results indicated that while thermal imaging cameras are best used in low to no vegetation, thermal drones can be used in medium vegetation, and detection dogs can be used in high vegetation where thermography is unsuitable. This is because vegetation obscures thermal signatures. However, all methods were found suitable for gathering data at night or when subjects are camouflaged, as they rely on heat and smell rather than sight.
More broadly, the success of each method depends largely on the goals of the project. Summarized from the study, considerations for each include:
Thermal imaging
- Most effective in habitats with minimal obstructions
- Most effective in habitats with low temperatures and little direct sunlight, as thermal contrast between the target species and their background is critical
- Excess humidity can blur images, and wind speeds of more than 5 km/hr can disrupt drone flight
- Daily activity patterns of the target species determine when they will be most visible
Detection dogs
- The study found that while the detection dog covered a smaller area than both thermal imaging methods, he found targets faster
- While detection dogs are not as directly limited by weather conditions and tricky terrain as thermal imaging, these factors can affect scent detectability.
- Special care must be taken in hot weather to avoid heat stress
- Must have low or controllable prey drive, especially when targeting live animals
- Training a dog on a live target can put stress on said target
- Training samples must be accessible
- Targets must have a detectable odor. Some animals use tactics to reduce their scent, such as lowering metabolism and metabolic odorants, and reducing surface contact with the air. Furthermore, wind can cause unpredictable flow of odors.
eDNA
What is eDNA?
eDNA, short for environmental DNA, is any DNA an organism leaves behind in their environment, including skin cells, scat, etc. eDNA analysis is typically cost effective, especially for species that are highly challenging to detect otherwise. It requires relatively little training to conduct, and in the case of metabarcoding, can detect multiple species at once. However, it is not without potential drawbacks. For example, it can linger for decades in soil and sediment, or be transported by wind or water, causing uncertainty about exact origins. This along with sequencing and taxonomic errors can increase the risk of false positives and false negatives. Determining species abundance is also a challenge. Still, eDNA detection is largely effective for confirming the presence of species, especially when paired with other methods to bolster the data. This makes it ideal for projects such as rediscovering a lost species.
Detection Dogs and eDNA: Golden Moles
In a June 2024 episode of the K9 Conservationists podcast, Kayla talks with Esther Matthews and Samantha Mynhardt of the Endangered Wildlife Trust about using eDNA and a detection dog to rediscover De Winton’s golden mole, a species previously thought lost, near the west coast of South Africa. The species hadn’t been seen for over 80 years prior to the study. It was classed as possibly extinct by the IUCN (now critically endangered on the Red List) and a most wanted lost species by re:wild.
In search of De Winton’s golden mole, the team analyzed eDNA in soil samples from other golden mole burrows. Matthews trained her dog to help find burrows for eDNA collection. Because no De Winton’s samples were available, the dog was trained to indicate the presence of Grant’s golden moles. When golden mole burrows were found and the dog didn’t indicate, this suggested that the burrows belonged to a non-Grant’s species. The researchers analyzed the soil samples and confirmed the presence of De Winton’s golden moles. This highlights another way that detection dogs can be paired with other methods for maximum efficacy in study goals.
In Summary…
eDNA
- Requires relatively little training to conduct
- Metabarcoding allows for detection of multiple species
- Can linger for long periods of time, or be transported by wind or water, leading to uncertainty about exact origins
- Potential for sequencing and taxonomic errors, risk of false positives and false negatives
- Highly challenging to determine species abundance or density
- Overall effective at confirming the presence of species, even given the potential for errors
Detection dogs
- Can detect data for eDNA analysis
- If conclusions reached by detection dogs and eDNA analysis validate each other, the risks of false positives and false negatives are reduced
- Can indicate areas where eDNA collection would be most effective
- Can help determine species abundance and density by the number of samples found in given locations
- Both dogs and handlers require significant training
Conclusion
When it comes to something as critical as wildlife conservation, a choice of methodology for a particular project can shape the future of an entire species or ecosystem. For this reason, it can be tempting to view methods of non-invasive wildlife monitoring as though they were competing with each other. The right method is crucial. However, that method varies from project to project. No singular method is perfect, and combining multiple that complement each other’s strengths and weaknesses may be the best option. When these methods are analyzed through a critical and nuanced lens, great conservation success stories— like the rediscovery of a lost species— can be written.