Bias in Aerial Survey of Feral Goats in the Rangelands of Western Australia.

Abstract

Broad-scale aerial surveys of feral goat populations in the aridsemi-arid rangelands of Western Australia have used strip transect methods without correction for visibility bias to estimate minimum abundance. Improved accuracy of these minimum estimates is necessary for more effective control. Line transect methodology was used to assess visibility bias in aerial survey of feral goats in a 375,000 km2area of north-west Western Australia. Visibility of goat groups decreased with increasing vegetation cover. In open country with little or no cover there was no decline in visibility within a 200 m strip from the transect line. In country with high vegetation cover visibility declined markedly with distance from the transect line, and it was estimated that only half the goat groups within the 200 m strip were seen. No size-bias in the estimation of mean group size was detected, The line transect results, and consideration of possible violations to the line transect method, suggest that previously published minimum estimates of feral goat abundance may be negatively biased by at least 30-40%.