In this post, Paul Lukacs discusses a paper he recently handled by Mathias Tobler and colleagues “Spatiotemporal hierarchical modelling of species richness and occupancy using camera trap data

Setting up a camera trap.
Dr. Mathias Tobler installing a camera trap in the Peruvian Amazon

Flipping through the pages (or scrolling through the links) of almost any applied ecological journal, you might begin to think that there is a remote camera placed in almost every patch of forest or on every fencepost. Do we really need all of these remote camera studies? In my opinion, the answer is a resounding “Yes”. A new paper by Tobler et al. in the Journal of Applied Ecology highlights the advantages that remote cameras bring to ecological research while also providing advancements to analysis of camera data. The authors expand their view to encompass multiple species and multiple surveys.

Understanding the mechanisms driving species richness remains a central theme in ecology. Determining how those mechanisms play out in real-world situations frames the problem that many wildlife managers face in their jobs. Tobler et al. demonstrate how remote cameras allow us to simultaneously estimate species richness and test hypotheses about factors affecting richness.

Caught in the camera trap: jaguar (top); nine-banded long-nosed armadillo (middle); anteater (bottom). Pictures courtesy of Mathias Tobler.
Caught in the camera trap: jaguar (top); nine-banded long-nosed armadillo (middle); giant anteater (bottom). Pictures courtesy of Mathias Tobler.

While remote cameras allow insight into multiple species in ways that classic trapping studies often cannot, cameras are not a panacea. Tobler et al. demonstrate strong variation in detection probability across species as well as substantial variation across camera sites. One might guess that behavioural and size differences between jaguars and armadillos would lead to varying detection probability. Incorporating the differences in detection probability across camera sites caused by variation in abundance at the site scale is more complicated. Tobler et al. build on work by Royle and Nichols (2003) and Yamamaura et al. (2011) to directly incorporate variation in their species richness models while sharing information to maintain parsimony.

Biologists across the globe deploy remote cameras, but rarely are the data from different studies combined into a single analysis. Tobler et al. demonstrate the power of combining data from multiple camera studies to gain broader insights than any single data set could have provided. Combining data further emphasizes the need to account for variation in detection probability. Tobler and colleagues’ extension provides a nice link to handle this variation. Combining studies broadens the scale of inference and allows resource managers to examine how land use may impact species richness.

Literature Cited

Royle, J.A. & Nichols, J.D. (2003) Estimating abundance from repeated presence-absence data or point counts. Ecology, 84, 777-790.

Tobler, M., Zúñiga Hartley, A., Carrillo-Percastegui, S., & Powell, G. (In press). Spatiotemporal hierarchical modelling of species richness and occupancy using camera trap data. Journal of Applied Ecology, doi: 10.1111/1365-2664.12399.

Yamaura, Y., Royle. J.A., Kuboi, K., Tada, T, Ikeno, S. & Makino, S.I. (2011) Modelling community dynamics based on species-level abundance models from detection/non-detection data. Journal of Applied Ecology, 48, 67-75.