Avian Conservation and Ecology (Dec 2022)
Comparing waterfowl densities detected through helicopter and airplane sea duck surveys in Labrador, Canada
Abstract
Sea ducks are inadequately monitored because traditional waterfowl surveys omit most of their breeding range and may be conducted too early for these species which typically nest late in the season. The gap in monitoring is particularly concerning for scoters (genus Melanitta) because the limited available data suggest that the abundance across the three species of scoters in North America has declined since the 1980s. We conducted trial sea duck surveys in central Labrador, Newfoundland, and Labrador, Canada, using both helicopter plot and fixed-wing line-transect surveys (10–19 June and 17–19 June 2009, respectively) to assess the feasibility of conducting specialized surveys for late-nesting waterfowl during the scoter breeding season. We present the results of the fixed-wing line-transect component, which we analyzed in a distance-sampling hierarchical framework to calculate and correct for imperfect detection. We found that the breeding density of Black (Melanitta americana), Surf (Melanitta perspicillata), and White-winged (Melanitta deglandi) Scoters combined was 0.15 (90% credible interval: 0.12–0.18) indicating breeding pairs per km², one of the highest breeding densities of any waterfowl species in the area at this time. Estimates of waterfowl density approximately doubled for all species after accounting for detection because observers only detected between 20% to 40% of all groups depending on the species or genus. Though there was a slight male bias in the sex ratios, groups observed were two individuals (i.e., a breeding pair), suggesting that timing the survey in mid-June captured the breeding window for sea ducks. Despite correcting for detection, breeding densities estimated by the fixed-wing transect component of the survey remained substantially lower than those estimated from the previously published helicopter component, suggesting there were differences in availability bias between the two survey platforms.
