Spatiotemporal analysis of animal movement and interactions

The spatiotemporal aspects of animal telemetry data, collected using technologies such as global positioning systems (GPS) and acoustic sensors, has become a critical area of ecological research. This dissertation contributes to the computational analysis of animal movement by analyzing and introducing methods to facilitate the analysis and interpretation of animal movement, specifically their use of space and their interactions. In Chapter 2, time-based home range methods were evaluated for accuracy in terms of area, shape, and location in response to sample size and common wildlife GPS-point patterns. These characteristics of home range estimation are important for inferring animal habitat and resource use. Despite the improved accuracy of time-based methods compared to traditional point-based methods, location was often inaccurate for all GPS-point patterns, as were shape and area for GPS-point patterns with perforations (common for areas with large physical barriers like mountains or lakes). Therefore, future research should focus on methods aimed at improving home range accuracy. Animal interactions are an important aspect of animal behavioral ecology that impacts processes like predation, mating, and disease spread. Interactions are spatially and temporally variable and can occur either through direct contact or at asynchronous spaces and/or times. Chapter 3 introduces a method, termed the temporally-asynchronous joint potential path area (ta-jPPA), that maps potential areas of spatially-synchronous-temporally-asynchronous interactions in order to link interactions to landscape features or other environmental variables. Given the spatiotemporal nature of animal movements, analysis metrics that quantify interaction often use subjective thresholds to subset or classify telemetry data. In Chapter 4, an unsupervised learning technique is presented as an exploratory data analysis tool that uses the spatiotemporal properties of two animal trajectories to uncover spatiotemporal analysis thresholds. These thresholds are not meant to replace expert knowledge of a species being studied, but to be used when no knowledge exists, as well as to explore different spatiotemporal scales of interaction. Using simulated data, the method was able to distinguish interactions from random movement, found appropriate thresholds based on derived models, and with empirical data the method was able to find patterns that were consistent with previously observed behavior.