Sometimes it's just a grainy black blur, the pixilated remnants of a fast-moving creature prowling the forest in the dark of night. But to Fabiola Iannarilli, it's evidence of a carnivore on the move. With luck there's also a telltale sign, such as the distinctive black legs of a red fox or the white patch on the back of a bobcat's ear, that matches the blur to a species and provides new insight into the habits of a secretive animal.
For the past year, Iannarilli, a doctoral candidate at the University of Minnesota, has been tracking bears, wolves, fishers, American martens, and other carnivores by searching hundreds of thousands of electronic photographs, downloaded from tree-mounted cameras that capture images remotely in the north woods.
DNA provides another avenue
to census carnivores
Genetic analysis of wolf scat can provide researchers with valuable information about Minnesota's wolf population. Ron Moen, associate professor at the Natural Resources Research Institute at the University of Minnesota Duluth, says genetic population analysis could complement data from camera and track surveys and provide an independent assessment of their accuracy. This is particularly important for wolf management and for considering effects wolves might be having on other species, such as moose and deer, Moen says.
Like the DNR's camera survey, Moen's first goal is to develop a standard procedure for collecting biological information. In the first year, his team identified scat rather than hair as the best source of DNA for this project, primarily because it's easier to obtain. They also identified scat collection routes and began collecting scat samples, which they used to determine the species, gender, and individual identity of the wolf, Moen says.
By genetically and statistically analyzing the samples, Moen's team will eventually be able to measure genetic diversity among the wolf population and estimate total wolf population size. Both of these measurements are significant indicators of the health of the population. "A species with a large census population may still have low genetic variation," Moen says, which can reduce a population's ability to resist new diseases and evolve with changing environmental conditions.
In addition to measuring the size and genetic diversity of the population, genetic data sampled throughout the range of the species will allow Moen's team to identify possible migration patterns of wolves as well as dispersal from established territories and formation of packs.
She is part of a team looking at replacing the Department of Natural Resources' decades-old track surveys with a camera survey to monitor populations of 14 of the state's common carnivores. If it's successfully implemented, the camera survey could produce more accurate data at a lower cost than track surveys, and it could provide the opportunity to engage students and citizen scientists in collecting and identifying images of wild animals.
The study comes at a time when critical carnivore population data is increasingly difficult to obtain through track surveys. The DNR conducts one statewide track survey in the fall and another in northern Minnesota during the winter. "Both these surveys are usually conducted off remote roads and trails. Over the years, we've seen more traffic and more paved roads, which may disrupt animal movement and reduce survey accuracy," says John Erb, the DNR research biologist who is overseeing the work. "Early-season snows have also been less reliable, and tight budgets are making it more difficult to find qualified people who can conduct track-based surveys. Cameras may offer easy and more reliable species identification, are fun, and are increasingly popular among outdoor enthusiasts. They may allow for more reliable and collaborative wildlife monitoring."
Data derived from these surveys is increasingly important to biologists and policymakers due to ongoing controversies surrounding wolf management as well as efforts by biologists to pinpoint why Minnesota populations of fisher and pine marten have declined in the past decade.
Iannarilli and her colleagues aim to develop a standard procedure for using cameras to monitor carnivore populations. They hope to find the best time of year to conduct camera surveys, how to best position cameras, how many cameras to use, how to select sites for cameras, which lures are most effective, and how best to statistically analyze camera data. If these logistical and technical issues can be addressed, Erb says, "Camera-based monitoring has potential to offer significant advantages over track surveys."
Four Decades of Tracks
Since 1975, staff from the DNR and various county or federal agencies have fanned out across the state twice each year—fall and winter—to record the type and number of carnivore tracks they see at specific locations. Track counts can be useful for detecting changes in the distribution or population trend of a species over time. Along with other data, such as annual harvest, these population trends are used to make habitat management or policy decisions, including setting seasons and bag limits for those carnivores that are legally hunted or trapped.
For the fall track survey, biologists make a series of tracking stations—circles of sifted soil about 1 yard in diameter with a scent tablet placed in the middle. The tablets contain synthesized fatty acids, which mimic the scent of food. "It's like that restaurant smell, except for animals," Erb says. The stations, placed on routes established when the survey was first developed, are spaced at ⅓-mile intervals on alternating sides of a gravel road. Biologists visit each station the next morning to identify and record the tracks they find. Track identification can be as much art as science, and accuracy is influenced by soil conditions, animal behavior at the track station, and the experience of the surveyor.
The winter track survey was developed mainly to gather data on fishers and American martens, secretive carnivores that tend to shy away from open areas where scent stations are located. Survey participants travel an assigned road or trail, identifying and recording carnivore tracks the day after a fresh snowfall. Because the sign of other carnivore species is readily detected in fresh snow, participants also record tracks for selected species such as bobcat, wolf, and coyote.
One reason that track surveys are valuable is the length of time the DNR has been conducting them, which allows biologists to examine population trends across longer time periods.
Because they can operate remotely and continuously, cameras work well for detecting fishers, martens, bobcats, and other stealthy predators that compete for food with larger carnivores like bears and wolves.
Recent improvements in camera technology, plus reduced cost and new methods for analyzing camera data, have bolstered their value even more, Erb says. However, he notes that there has been relatively little study of how to best deploy and use cameras to survey multiple species of carnivores simultaneously.
For the first year of the study, researchers set out 100 cameras in the spring and fall, which recorded approximately 680,000 and 370,000 images, respectively. Because the cameras were set for high sensitivity, shadows caused by bright spring sun along with sprouting vegetation accounted for a large number of false triggers—images with no animal—in the first spring, Iannarilli says. Reducing the sensitivity would mean fewer false triggers, but might also mean missed animals. Researchers continue to refine ways to reduce false triggers and still capture animal images.
Bears were detected at four times more sites in the spring than in the fall. Coyotes, red and gray foxes, raccoons, striped skunks, martens, and fishers were detected twice as often in the fall compared to spring. Wolves were detected equally often in both spring and fall.
Knowing when certain species are most or least likely to be active allows researchers to tailor their camera surveys to the questions they are trying to answer. "For instance, we might not use cameras in spring to get data on female fishers because they're locked down in dens with young and move very little," Erb says. "And of course winter is not a good time to detect bears because they are hibernating."
Cameras vs. Tracks
One of the biggest differences between track and camera surveys is the potential for improved confidence in population data from one year to the next. As Erb explains, track surveys provide very good long-term population trends, whereas camera surveys should be better at assessing annual fluctuations in a population. The species being monitored dictates the type of data needed.
"For species that are common and abundant, such as coyote or raccoon, the long-term trend is adequate for management," Erb says. "However, with species that occur at low population density such as marten or fisher, an accurate estimate, or index, of next year's population will give us more information for setting bag limits and season lengths and understanding effects of changes in habitat, weather, or hunting and trapping season regulations."
Monitoring a location continuously for many days via a camera allows researchers to develop a "detection probability," which is a statistical term that defines whether an animal may be present at a survey location, but not detected by the surveyor. It removes some of the randomness that can occur when a survey route is visited only once.
"If you are driving and see no deer at a location one day, then four deer at the same location the next day, there's a good chance that the additional deer were nearby the first day, but you didn't see them," Erb says. "Having a history of site visits and detections gives us a clearer picture of the overall population."
Developing a detection history for track surveys would require multiple site visits by survey participants, which would drive up the cost of the survey, Erb says. Camera survey sites are visited only to deploy and retrieve the camera. This also reduces the amount of human scent in the area, which can affect animal behavior.
Another benefit and possible cost savings from camera surveys will be the opportunity to involve volunteers from around the state. "If camera surveys are eventually implemented, it may be possible to rely on citizen scientists to deploy and download images from cameras," Erb says. "I could see this being an excellent project for biology and math classes, as well as clubs, conservation organizations, and individuals."
Once the current fieldwork is complete, possibly after this spring season, Erb and his colleagues will make formal recommendations to the DNR on how best to proceed with the future of carnivore surveys. While there are a lot of advantages, using cameras for research raises a number of issues.
The survey may improve reliability and reduce costs over the long term, but the up-front costs of purchasing enough cameras could be too steep, says Erb. Other considerations include the large number of images that would need to be processed and analyzed as well as coordinating and training citizen volunteers and collecting and distributing equipment across the state.
While cameras may someday become the DNR's primary tool for monitoring carnivores, Erb says the use of cameras to monitor other species may not be appropriate. "Ubiquitous species, like deer in forested areas, are likely to be present at most sites even when populations are down. That makes it difficult to detect changes," he says. However, wildlife researchers in Minnesota and elsewhere will continue to find new uses for remote cameras and more sophisticated ways to analyze the data they collect.
It's nearly certain that remote cameras will continue to open more windows into the private world of wild animals that we've only begun to glimpse. For researchers, citizen scientists, and those who just appreciate the wild, it's yet another tantalizing opportunity to explore, learn, and ask new questions about the creatures who share our resources.