Network-based Home Range Analysis Using Delaunay Triangulation

The home range is the fundamental measurement of fish and wildlife space-use patterns. Kernel density estimation (KDE) is the most widely applied home range estimator, although its poor performance has recently been documented. In this paper, we suggest that KDE is inappropriate for home range estimation, because it assumes Euclidean-based space usage. Because animal space-use patterns show characteristics of network-based movement, we develop a network-based home range estimator. First, we use Delaunay triangulation (DT) to approximate a network of travel paths from a set of animal point locations. Then, we adapt KDE to estimate home ranges as a function of that network. Preliminary results suggest that network-based home range estimation using DT has the potential to improve the way ecologists measure animal space-use patterns.