Camera grid systems are being increasingly used to monitor threatened and declining wildlife species – a method that is essential for understanding wildlife habitats, behavior, and population estimates.
Placing these systems involves a tremendous amount of time and effort. To monitor grizzly bears in the South Chilcotin Mountains of British Columbia, researchers need to fly into a remote lake and bike through miles of wilderness, all so they can hike through difficult terrain to place and service camera equipment. Each day on a trip like this includes 6-8 hours of hard work and allows researchers to reach about ten cameras (on a good day).
Conservation Science Partners (CSP), a California-based nonprofit that advances quantitative analysis and planning services for global conservation projects, and their collaborators at the University of Southern California and Microsoft, are working to make the most of limited conservation resources by developing an open-source algorithm to optimize camera placements within any given landscape.
The project, funded by the Paul G. Allen Family Foundation, will use data from existing camera systems in Kenya, British Columbia and Washington state to develop optimized monitoring designs for Grevy’s zebras, grizzly bears and cougars, respectively.
"Monitoring sensitive wildlife populations demands a huge effort and eats up time and resources by conservation organizations,” said Justin Suraci, senior scientist and director of wildlife conservation science with Conservation Science Partners. “Monitoring wildlife through on-the-ground systems such as cameras and bioacoustics is often done in remote areas, making it difficult and expensive to regularly access the equipment. We’re working to maximize the efficiency of camera-based studies by determining the best spatial layouts to estimate animal abundances while accounting for the cost and difficulty of deploying camera traps.”
Working with conservation organizations that are already using monitoring equipment, this work will develop special plans for cameras and other sensors to most effectively sample wildlife targets in each geography. For instance, in Washington state, CSP is working with Panthera, a wild cat conservation organization, and the Lower Elwha Klallam Tribe to optimize their larger, collaborative Olympic Cougar Project that includes five other tribes and natural resource organizations in the Pacific Northwest.
“Our goal is to understand how the ecosystem is functioning by keeping an eye on six key wildlife species across the Olympic Peninsula,” said Kim Sager-Fradkin, Wildlife Biologist with the Lower Elwha Klallam Tribe. “Doing that over a vast area isn’t easy or cheap. By collaborating with CSP, we can strategically place our camera grids to work smarter and more effectively. This shared multi-species approach gives us more insight into how animal populations and behaviors are shifting over time.”
The project includes more than 500 established camera traps across the Olympic Peninsula to monitor the species and understand population trends and movement over time. Data captured from the camera grid will be used by the project team to build the optimization plan – and then used to test the performance of their algorithms to make sure efficiencies are real.
“Some of these cameras are so remote it takes an hour just to get to one. That kind of fieldwork adds up fast,” said Vanessa Castle, Panthera's Cultural Conservation Science Coordinator for the Olympic Cougar Project. “This is especially true when studying elusive animals like cougars. Knowing where to find efficiencies means we can better illustrate their role in the ecosystem – and do more with the time and resources we have.”
CSP is also working with The University of British Columbia that has extensive camera systems in place that monitor grizzlies in the Canadian wilderness – and with the Massachusetts Institute of Technology and the Mpala Research Centre that have a similar system for Grevy Zebras in Kenya. These three different sites and the diversity of terrains will provide ground-truthing for the simulated optimization.
“Once we know our simulations produce the right design to optimize sensor placements, we will develop a free interactive tool for wildlife scientists and conservation practitioners to use,” said Suraci. “The tool will allow practitioners to explore optimized sensor designs that maximize the effectiveness of wildlife monitoring studies under budgetary constraints.”