New Zealand Journal of Ecology (2007) 31(2): 223- 231

Distance sampling techniques compared for a New Zealand endemic passerine (Philesturnus carunculatus rufusater)

Research Article
Phillip Cassey 1*
John L. Craig 2
Brian H. McArdle 3
Rosemary K. Barraclough 4
  1. Centre for Ornithology, School of Biosciences, University of Birmingham, Edgbaston B15 2TT, Birmingham, UK
  2. School of Geography and Environmental Science, University of Auckland, Private Bag 92019, Auckland, New Zealand
  3. Department of Statistics, University of Auckland, Private Bag 92019, Auckland, New Zealand
  4. Ecology and Conservation Group, Institute of Natural Resources, Massey University, Albany Campus, Private Bag 102-904, North Shore Mail Centre, Auckland, New Zealand

The effectiveness of line- and point-transect distance sampling methods was compared for estimating the density of a conspicuous endemic passerine, the North Island saddleback Philesturnus carunculatus rufusater, in two forest habitats on Tiritiri Matangi Island, Hauraki Gulf, New Zealand. The reference population abundance in each habitat was calculated through an intensive capture, colour-banding, and resighting effort. Line transects consistently produced unbiased estimates of density for both sites at two time periods (morning and afternoon), and proved to be the most efficient of the distance sampling methods tested. Point-transect methods proved to be biased in this instance, consistently producing overestimates of density for this species, and were unreliable as indices of the small differences in density between habitats. Transect counts conducted from established walking tracks underestimated both density and the variance of estimated density. We conclude that in situations requiring estimates of absolute density for North Island saddleback, a series of randomly positioned line transects will accurately represent the density of individuals. Furthermore, this study highlights the need to pilot distance survey techniques on the species of interest to detect possible violation of the assumptions underlying these methods. We question the widespread use of point-transect counts for censusing avian species without such preliminary investigation, and recommend further comparisons of distance sampling methods in New Zealand with reference populations where opportunity provides.