An open-source, low-cost sensor developed by Colorado State University and US Forest Service researchers pairs air-quality readings with GPS and physiological data to reveal how wildfire smoke affects free-ranging animals.
As wildfires grow larger, more frequent, and more severe, the smoke they produce can degrade air quality hundreds of kilometers away, and for days to weeks after a fire is active.1 People can retreat indoors when the air turns hazardous, but free-ranging wildlife cannot, and researchers have had few tools to measure what those animals are actually breathing.
A team led by researchers at Colorado State University (CSU) and the US Forest Service has developed a low-cost, open-source air-quality monitoring device designed to close that gap. The smoke logger can be mounted on a GPS collar alongside existing sensors that track location, activity, and physiological data, allowing investigators to connect individual smoke exposure with changes in wildlife health and behavior.1,2 Initial results were published October 7, 2026, in Methods in Ecology and Evolution, and the device’s design and programming are now freely available.1,2
“Wildfire is now a chronic problem in North America. We all are experiencing these massive smoke plumes that can negatively affect air quality for weeks at a time,” said George Wittemyer, PhD, a professor in CSU’s Warner College of Natural Resources, who led the project with Mark Ditmer, PhD, a US Forest Service wildlife ecologist, and Adam Parlin, a CSU postdoctoral researcher.2
“We’re interested in how wildlife adjusts its behavior relative to smoke exposure, and it wasn’t possible to know until now,” he continued.
Despite public health and safety concerns surrounding wildfire smoke, little is known about how free-ranging wildlife experience or respond to it, a shortfall the study authors attributed to sparse air-quality monitoring in the remote landscapes animals occupy.1 Air quality typically is not monitored in wilderness areas, Wittemyer noted, and extended smoke events represent a new stressor for wildlife.2
Animal-borne sensors have transformed wildlife ecology by allowing researchers to quantify how individuals respond to environmental stressors in real time, but an equivalent tool for measuring smoke exposure had been missing.1
“We know relatively little about the impacts of wildfire smoke on the behavior of animals in the field, especially how free-roaming wildlife respond to severe smoke events,” said Parlin.2 “These units provide one avenue to better understand the consequences from wildlife’s perspective—such as identifying refugia during smoke events—that can provide actionable insights for conservation and management.”2
The device uses a low-cost, commercially available laser-scattering sensor, similar to those in the PurpleAir monitoring network, to measure fine particulate matter 2.5 µm in diameter or smaller (PM2.5), particles that are detrimental to cardiovascular and respiratory health.1,2
Designed to be worn by medium and large animals, the collar-mounted version runs on a solar-charged battery and records sub-hourly measurements throughout entire fire seasons, storing an air-quality reading every 30 minutes.1,2 Rather than transmitting data, which would be costly, would drain the battery, shorten the device’s life span, and is often impossible given the lack of connectivity in wildlands, the logger stores data on board, so collars must be recovered for data to be downloaded.2
In addition to collar mounting, the logger can be deployed as a stationary unit within an animal’s home range to characterize ambient exposure to particulate matter associated with smoke and vehicle combustion.1 Deployed collar units weighed 135 g, though further miniaturization has reduced that to 100 g.1 Researchers can assemble their own units for about $100 in materials by downloading the free design and programming.2
In initial deployments along an urban-to-rural gradient, GPS-collared mule deer (Odocoileus hemionus) fitted with smoke sensors reported PM2.5 values that varied with proximity to roads, demonstrating the sensor’s ability to resolve fine-scale pollution gradients.1 The team fitted loggers to collars on 5 mule deer in Colorado and 3 in Utah during 2025 captures.1 Deer monitored in Colorado and Utah were not exposed to unhealthy levels of wildfire smoke, but testing showed the sensor was sensitive enough to register minimal air pollution spikes from traffic among urban-dwelling deer.2
In remote Utah backcountry, where the closest PurpleAir stations were 20 to 22 km away across the Wasatch Range, 1 logger recorded continuously for 38 days before the collared deer was killed by a cougar.1 PM2.5 concentrations at that deer’s location were approximately an order of magnitude lower than those measured by the nearest stations.1 According to the researchers, the discrepancy underscores how localized air quality can be and the limits of relying on existing monitoring networks to estimate wildlife exposure.1,2
“That was eye-opening for us and shows how variable air quality is at a microscale,” Wittemyer said.2
The loggers also picked up fine-scale differences during a regional smoke event. On September 5, 2025, before collar deployment, 2 loggers placed outside in Fort Collins, Colorado, recorded a distinct local plume lasting about 2 hours; some nearby PurpleAir stations captured it, while others missed it entirely.1
By pairing collar-measured PM2.5 with location, activity, and physiological measurements, ecologists can test mechanisms that previously could only be inferred, the authors wrote, including behavioral responses to degraded sensory environments during smoke events and sublethal health effects that persist long after visible smoke clears.1 One example is sensory disruption, in which smoke may mask or distort the scents prey species rely on to detect predators. That question is driving the team’s deployments on cougars in northern New Mexico and elk in Washington.1 Many collaring studies already track health measures such as heart rate in addition to GPS coordinates, and combining those data with smoke logger readings is expected to reveal new information about wildlife welfare and activity.2
The authors also noted that disagreement between logger readings and regional smoke products may itself be ecologically informative, potentially flagging smoke refugia that coarser monitoring approaches cannot detect.1 The open-source design can be extended to additional pollutants and microclimate variables, such as temperature, humidity, and volatile organic compounds.1
Ditmer said a stronger understanding of how smoke affects wildlife health will improve conservation planning and allow prescribed burns to be timed to reduce risk to animals.2
“Prescribed fire is an important tool for managing forests in the face of increasingly severe wildfires, and understanding when and where smoke actually reaches wildlife could help managers plan burns that meet forest-management goals while reducing exposure during sensitive periods or in important habitats,” he said.2
Current deployments include collar-mounted units on mule deer in Colorado, California, and Utah; elk in Washington; and cougars in New Mexico, plus 8 stationary units in the habitat of the federally endangered Southern Sierra Nevada fisher, a small mammal related to weasels, in Yosemite National Park.1,2 In some Colorado and Utah deer, smoke loggers are being paired with implanted physiological biologgers to test whether smoke exposure predicts changes in movement and acute physiological responses.1 Those results will be published in future studies.2
The researchers are encouraging investigators whose projects involve recovering wildlife collars to partner with them and add smoke loggers to their sensor packages.2 Funding for the study came from the Utah Division of Wildlife Resources, Utah Archery Association, USDA Forest Service Rocky Mountain Research Station, Sportsmen for Fish and Wildlife, Brigham Young University, Mule Deer Foundation, Colorado State University, and Safari Club International.1
References
