Kernel Methods in Line and Point Transect Sampling

Abstract

We consider kernel estimation of population density based on size and distance data in line and point transect sampling. Asymptotic normality of the estimators is established in each setting, with very weak assumptions imposed on the random sample size. It is pointed out that the kernel approach does not require specifying a horizon, a severe shortcoming in the Fourier series method. Some numerical examples based on simulated and real data are presented. The results suggest that the kernel method is a viable alternative to other existing estimators.