Measurement of non-circular home range.

Abstract

Abstract We examine the use of conventional sampling data to estimate the area extent of an animal's home range. Popular methods of estimation based on polygons and capture radii are discussed. A simple example shows they are not directly comparable. Furthermore, when home ranges lack circular symmetry, methods based on capture radii suffer large biases. On the other hand, methods based on polygons have sample size biases. A new method, based on the determinant of the covariance matrix of the capture points, is introduced. We show that this method, under specified assumptions, is free from both forms of bias. Tables that identify the biases in the polygon methods are provided. An important use of the tables, that of making comparisons between values obtained by the various methods of estimation, is illustrated. The statistical stability of each method is given by means of a table or formula for its relative variance as a function of sample size. Finally, the methods are evaluated. For purposes of estimating home range area, the determinant method is recommended, the old convex polygon method is defended, and the use of methods based on capture radii is discouraged.