A Convex Hull-Based Estimator of Home-Range Size

This paper considers the problem of estimating the home-range size of an animal from locational data, which are usually collected by use of radio telemetry, with convex hull-based estimators. It is shown that the widely used minimum convex polygon has some undesirable characteristics as an estimator, and thus should be used with caution. A method based on convex hull peeling of the data to estimate convex density contours is studied as an alternative simple home-range size estimator. Monte Carlo simulations show that this estimator has very similar statistical properties if the point fixes are generated from a bivariate normal distribution, a uniform on a disc distribution, and two other distributions with intermediate tail behaviour. These standard distributions are used as a reference to define a sample size bias-corrected estimator of home-range size, and to provide an approximate method for calculating the precision of this estimator. The technique is applied to the estimation of the home-range sizes of deer.