Effect of observer differences on abundance estimates of seabirds from ship‐based strip transect surveys

Data from four ship‐based strip transect surveys of seabirds, in which six observer teams from several countries cooperated, were used to estimate observer differences. The numbers of birds per sampled area as reported by the various observer teams were used to estimate differences among observer teams. As overlap in effort was not complete (all teams did not work in exactly the same areas at the same times), a log‐linear model was used to correct for differences due to survey and region effects and the state of the sea. Considerable observer differences were found. For example, three teams out of five observed less than 16% of GuillemotUria aalgeand even less than 10% of KittiwakeRissa tridactylarelative to the team that reported the highest densities.