Learning What We Can Learn from Camera Trap Photos: Part 1

Others have said it before and it’s true: New forms of technology such as camera traps make it possible for field researchers to collect information we could previously only dream about. Technology advances so quickly that we’re still evaluating what we can do with these tools and what questions we can and cannot answer by using them. We have many basic questions that are still unanswered even for species as large as the Andean bear. Here are two of the most basic: How many are there? Are the populations increasing, stable, or decreasing? The answers to these questions and others would help researchers, conservationists, and governments decide how much of their limited resources to invest in research efforts and conservation actions, in the hopes that 100 years from now there will still be bears roaming the forests of South America. Unfortunately, there are still no answers to these questions.

How would you answer these questions? How do you count animals that live in dense forests in rugged habitat, when those animals avoid contact with humans? It’s been said for decades that the markings of individual Andean bears vary so much that you can identify each individual bear just by looking at it. If that’s true, then maybe we could use camera traps to identify individual Andean bears in photographs and then estimate population sizes and densities. However, how do you test whether individual bears can be reliably identified in photographs? In order to test this you’d need photos of a lot of different bears whose identity you definitely knew. That means you can’t just use photos of wild bears from camera traps, because you don’t know how many bears walked in front of the camera traps.

Compare the markings of this bear, Tommy, with the bear above (named Turbo) and notice the differences.

Compare the markings of this bear, Tommy, with the bear above (named Turbo) and notice the differences.

The only way we could think of to test this was to take photos of different bears from captivity, so we’d know the identify of the bears, ask people to compare those photos, and keep score of whether bears were correctly identified, or not. When a group of international collaborators and I did this we were surprised to discover that people were initially not very good at this task. In fact, they would have done just as well if they’d flipped a coin! That’s really not the kind of result we were expecting, or hoping, and it led us to consider whether we were over-confident in our own abilities to identify individual bears. However, it turns out that with a little practice and training, participants became better at identifying bears from their photographs. After thorough review and discussion by other scientists, this work has been published in the journal Wildlife Biology and you can read all the details and see more photos here.

So, the good news is that, if we’re careful, we and other field researchers can use photos from camera traps to identify individual Andean bears, estimate the sizes of their populations, and compare populations densities. Now, we “just” need to get the cameras into the forests where there are bears!

Russ Van Horn is a scientist at the San Diego Zoo Institute for Conservation Research. Read his previous blog, Ambassador Mi Ton Teiow Receives a New Posting.