Northern Prairie Wildlife Research Center
D. Erran Seaman, USGS, Biological Resources Division, Forest and Rangeland Ecosystem Science Center, Olympic Field Station, 600 E. Park Ave., Port Angeles WA 98362-6798 USA.
Kernel methods for estimating home range are increasing in popularity, but the effect of sample size on their accuracy is not known. I used computer simulations of 10 to 200 points per home range, and compared the accuracy of home range estimates produced by fixed and adaptive kernels, using the "reference" and least squares cross validation (LSCV) methods for determining the amount of smoothing. Simulated home ranges varied from simple to complex shapes, created by mixing normal distributions. I used two measures to assess the accuracy of kernel home range estimates: 1) the size of the 95% home range area; and 2) the relative mean squared error of the surface fit. For both measures, the bias and variance approached an asymptote at about 50 observations per home range. The fixed kernel with LSCV smoothing provided the least biased estimates of the 95% home range area. All kernel methods produced similar surface fit for most simulations, but the fixed kernel with LSCV had the lowest frequency and magnitude of very poor estimates. These results will assist researchers in determining necessary sample sizes for home range studies and in selecting a kernel method for data analysis.