Window Sensitivity Functions for Line Transect Sampling

Abstract

/ The estimation of animal or plant abundance by the line or belt transect sampling method has been investigated by many authors, including biologists, statisticians, biometricians, wildlife scientists, and fisheries and forestry specialists. The method has reached wide acceptance over the last 30 years, and the theoretical development has continued to expand, both from a parametric and a nonparametric point of view. This article examines the line transect sampling method from a slightly different perspective. If the line transect region is regarded as a rectangle with length L and width 2w, then by a selection of viewing window is meant a choice of w, with the intent being to search for optimal viewing windows, with the goal in mind of improving variances of estimators of population density, reducing sampling effort, while maintaining the property of unbiasedness. The notions of increasing window sensitivity (IWS) and decreasing window sensitivity (DWS) are introduced, and a method of deriving confidence intervals is discussed.KEY WORDS: Line transection sampling; Population estimation; Probability detection function; Window sensitivity function; Strip census.