Line-transect distance sampling compared with fixed-width strip-transect counts for assessing tomtit (Petroica macrocephala) population trends

Distance sampling and fixed-width strip-transect counts were compared as methods for estimating population trends of the tomtit (Petroica macrocephala) from late September / early October to early November 2004, before and after aerial 1080-poisoning for control of the brushtail possum (Trichosurus vulpecula). Two observers independently recorded the distance and compass bearing to tomtits detected along 36 transect lines in each of two forests (one prefed with non-toxic bait and one not prefed). From these data we calculated (a) male tomtit density in the program DISTANCE and (b) the number of male tomtits within 50 m of each transect line. The number of male tomtit detections increased from September/October to November, probably as a result of changes in bird behaviour. Detections peaked at 21–25 m from the transect lines, most likely indicating that some closer birds stopped singing on approach of an observer and were undetected. The mean detection distance did not differ significantly between observers in September/October but did differ in November, suggesting changes over time in the relative performance of the observers. Thus, several key assumptions underlying distance sampling were not met. The two methods produced similar population trends, indicating no difference in bias. Also, both indicated that poisoning with prefeeding had either no adverse impact on the tomtit population or no greater adverse impact than poisoning without prefeeding.