Abstract
The development by ornithologists of the variable circular plot mehtod of estimating bird densities from observer-to-bird distances has led to estimation problems similar to those of line transects sampling. However, to apply line transect methods without modification requires that the distance data be transformed to areas by squaring. The implications of this transformation on the assumed shapes for the detection curve have not been investigated in the literature. This paper illustrates that application of line transect models to area data maylead to poor model performance and erroneous conclusions. Developments based on distance data are discussed briefly, and binomial models are derived for the special case in which birds are categorised simply according to whether they are within or beyond a specified distance.

This publication has 3 references indexed in Scilit: