Understanding how species interact with their environment is the first step in being able to craft effective conservation and management plans. One of the most common ways to evaluate species-habitat relationships is to use Resource Selection Functions (RSFs). RSFs work on the assumption that animals prefer landscape features that are used more often than expected and avoid landscape features that are used less often than expected, based on their availability. Therefore by comparing areas animals used versus those that were available, but not used, we can estimate the strength of preference or avoidance of different landscape features and, across a study area, predict the probability that an individual will use a specific location.
My research has focused on using GPS telemetry data from collared animals for estimating RSFs. GPS telemetry data provides a wealth of information that allows us to differentiate habitat use based on different behavioral states of individual animals. For example, my colleagues and I have shown that preference for a habitat type can change depending on whether an animal is in a ‘resource-use’ state versus when an animal is in a ‘movement’ state. GPS data can also be used to examine multi-scale RSFs, incorporating the fact that species respond to different landscape features at different spatial scales. We have shown that multi-scale models outperform single-scale models and have developed novel methods for estimating multi-scale RSFs to examine habitat use during animal movement.
Our research has indicated RSFs are extremely sensitive to how we select and define the environmental input layers. Typically geospatial, or GIS, layers are used to define the used and available areas for a species. If we select incorrect layers or define them inappropriately, biases can be introduced into the results, making our analyses ineffectual and clouding our knowledge species-habitat use. This finding is concerning and emphasizes the need to examine multiple geospatial layers and landscape definitions in order to accurately model species-habitat use.