Effects of aircraft disturbance on behaviour of harlequin ducks (Histrionicus histrionicus)

Abstract

I studied behaviour of Harlequin Ducks (Histrionicus histrionicus) at Fig River (53° 03' N, 63° 09' W) and Crooked River (54° 06', 60° 48' W) in Labrador in spring and summer of 1999 to 2002. Observations indicated that paired adults devoted modest proportions of time (30 to 40%) to foraging. This, and the lack of variation in time budgeted to feeding across time and space, indicated that Harlequin Ducks were not limited by food on their breeding habitat in Labrador. -- Most pair bonds dissolved by early June indicating nest initiation by the females. All adult females that were examined following capture and/or followed by radio signal appeared to be reproductively active. Low annual productivity on the study areas appeared to be due to predation of nests, and I concluded that Harlequin Ducks breeding in Labrador may be limited by predators. -- Noise is a significant stressor for animals, and the non-auditory effects of noise are considered whole body stress responses. In addition to overt responses, behavioural effects of noise on adult Harlequin Ducks were subtle and protracted. I interpreted this as evidence that a larger 'whole body' or physiological response was occurring. Animals challenged repeatedly develop high circulating levels of stress hormones in the bloodstream that can ultimately lower survival and inclusive fitness. -- Alert responses by adult Harlequin Ducks in central Labrador occurred to noise (75- 120 dBA) generated from low-level (30- 100m agl) military jet over-flights. Alert (startle) was generally <1% of time budgets, and increased in a dose-response manner, accelerating above a threshold of approximately 80 dBA. Protracted effects included increased inactivity, and decreased time out of water. There was evidence of residual effects of increased agonistic behaviours up to 1 h and decreased courtship up to 1.5 h following over-flights by military jets. The protracted and residual effects have the potential to negatively affect time-activity budgets of individuals. Multivariate statistical analyses demonstrated the importance of considering behaviour holistically because the inclusive modeling of covariance among behaviours was superior to traditional univariate approaches. Important effects may easily be overlooked because of bias by researchers in defining 'behavioural responses' a priori (e.g. Startle or Alert) that are easily observed because effects can be subtle and protracted well beyond the actual disturbance.