Spatial Methods for Line Transect Surveys

Abstract

Line transect surveys have traditionally been used only to estimate the mean intensity of an underlying spatial point process. In this paper, we propose a test for complete spatial randomness and a method of estimating the clustering parameters in a simplified Neyman-Scott Poisson model, using observations collected in a line transect survey with a Gaussian detection function. These procedures use only the coordinate in the direction of the transect line. In simulation studies we compare the performance of a one- dimensional test for complete spatial randomness and a test based on the traditionally most powerful two-dimensional test. We also examine the bias and variance of the clustering parameter estimates obtained using the one- dimensional procedure. We apply the methods to data collected in a line transect survey for minke whales in the northeastern Atlantic.