Wireless digital camera traps are providing unprecedented and detailed information on some of Jasper Ridge Biological Preserve's (JRBP) biggest unknowns, such as the abundance and behavior of large predators and the nature of biological corridors linking JRBP to other areas. By silently recording whatever passes by—whether a mountain lion, jackrabbit, or even an occasional trespasser—the cameras help address the challenges of the wildland-suburban interface, including the need to convey the role of large predators in ecosystem health while raising awareness of potential encounters. 

The history of camera trapping at the preserve began in 2006, when Biology Professor Rodolfo Dirzo and his colleagues installed a first-generation set of cameras. Over 2.5 years, they recorded more than 15,000 photos but captured no mountain lions. 

A new chapter began in October 2008, when JRBP technology specialist Trevor Hébert started installing a more advanced, wireless digital camera network along open trails and dirt paths. Funded in part by a National Science Foundation grant, this new system consists of solar-powered, motion-activated cameras that use infrared sensors instead of a visible flash. They transmit photos for extended periods without servicing via wireless communication stations to a server that stores photos and metadata. The first cameras in this new network became operational in 2009, and the results were dramatic. On September 10, 2009, an infrared camera captured the first-ever photograph of a mountain lion at the preserve. Today, there are 22 cameras permanently in place for long-term wildlife monitoring. 

By 2016, over 150,000 photos had been collected. Hébert created database and web-based labeling software so that a group of Jasper Ridge docent volunteers could help tag the photos with the correct species. Hébert and the volunteers ultimately labeled 170,000 photos, which were then used as a training set for Deep Learning photo classification software developed by students recruited by Hébert from Stanford's CS341 class. Since then, all new camera trap photos have been labeled using this software with an average of 93% accuracy, although some checking and correction by humans is still necessary.

Related research publications: 

Rodrigo Béllo Carvalho, Chrysanthe Frangos, Chinmay Sonawane, Trevor Hébert, Elizabeth A Hadly, Rodolfo Dirzo, Limited changes to mammal diel activity during COVID-19 lockdown in an urban preserve. 2026. Journal of Mammalogy, Volume 107, Issue 2, April 2026, Pages 358–367, https://doi.org/10.1093/jmammal/gyag012

Kevin Leempoel, Trevor Hebert, Elizabeth A. Hadly. 2020. A comparison of eDNA to camera trapping for assessment of terrestrial mammal diversity. Proc Biol Sci 1 January; 287 (1918): 20192353. https://doi.org/10.1098/rspb.2019.2353

Mendoza, E., Martineau, P.R., Brenner, E. and Dirzo, R. 2011. A novel method to improve individual animal identification based on camera-trapping data. The Journal of Wildlife Management, 75: 973-979. https://doi.org/10.1002/jwmg.120