Authors
Dr. Jennifer Verdolin
University of Arizona
Objective
Apply contingency analysis and false discovery rate (FDR) for an environmental conservation scenario.
Background
Monitoring wildlife is critical for managing wilderness areas and guiding conservation efforts. Traditionally, wildlife monitoring involved manually surveying individual animals along a transect of some width and length. Manual surveying is costly and time-consuming, especially when conducted over large areas, such as those of the U.S. National Parks. In addition, manual surveying often misses rare, hard-to-observe species. For example, snow leopards are notoriously difficult to observe in the wild, making it hard to estimate the population size of this elusive and endangered feline.
Technological advances have fueled innovations in wildlife monitoring. The use of camera traps – motion-sensitive, remotely triggered cameras – has increased substantially, improving wildlife biologists’ ability to monitor remote areas and rare species.
Camera traps can do more than reveal which and how many species live in a given area. Camera trap data can also be used to evaluate how species influence each other, whether human recreational activity is impacting species, how species are distributed across habitat types, and if living close to the edge of a park boundary presents challenges to certain species.
Individual camera trap observations such as these can yield important insights in themselves. Going further, statistical analysis of large-scale camera trap data sets can reveal important patterns in the distribution of species and interactions or dependencies among them, which in turn can guide further research and conservation efforts.
The Task
- Perform contingency analyses to uncover associations between observation of black bears with observation of the other 20 species in the data.
- Apply false discovery rate control to account for conducting a family of 20 hypothesis tests.
- Visualize species co-occurrences using an interactive geographic map.