Visibility Bias of Waterbirds During Aerial Surveys in the Nonbreeding Season

Abstract

ABSTRACTAerial surveys for waterfowl and other waterbirds provide abundance estimates that are commonly used by state and federal agencies for waterfowl and wetland management. However, most existing surveys provide an index of abundance and are uncorrected for visibility bias, which may limit their use in accurately determining local population size. We used concurrent ground and aerial surveys to estimate visibility bias during cruise‐style waterfowl surveys from September through January 2014–2017 along the Illinois River Valley, Illinois, USA. Immediately before an aerial survey, a ground observer counted waterfowl and other waterbirds by species within a distinct survey area and counts were compared with aerial abundance estimates to quantify detection rate and count bias (i.e., visibility bias). Overall, waterfowl guilds had high detection rates and low count bias, resulting in low overall visibility bias (ducks, −11% [SE = 5%]; geese, −8% [SE = 3%]; swans, −5% [SE = 3%]). At the species level, visibility bias varied greatly (e.g., green‐winged teal [Anas crecca], 182% [SE = 221%]; American wigeon [Mareca americana], −74% [SE = 20%]). Group size, species prevalence, cloud cover, and temperature influenced visibility bias, but direction and magnitude of effects were variable among taxa. Low visibility bias for most guilds indicated that well‐designed, cruise‐style aerial surveys with bias estimation techniques provided reliable estimates of population size and have great utility for monitoring waterfowl and other waterbirds as part of an adaptive management framework. © 2021 The Wildlife Society.