Monitoring seabirds on fixed ferry-routes: an evaluation

An analysis of 71 ship-based seabird surveys on fixed ferry routes in the southernmost part of the North Sea revealed that parameters indicating the periodicity of counts (year, season, month) were more influential than those caused by unwanted side-effects (route, observer, wind). With the GLIM tool, we formed minimal adequate log-linear models of the densities of thirteen abundant species/taxa and quantified the impact of route, observer and wind. It was found that a relatively small shift in route (c. 15 km) resulted in significant differences in densities in half of the species. The difference could be as high as a factor 2.4-14.3 times more. Observer effects were considerable in auks (differences in densities of 1.5-3.2x), and really far out for several gull species. These differences might have been caused by slight changes in methodology and demonstrate clearly that interobserver differences need more attention. Finally the wind force affected the observed densities of Kittiwake and Common Gull in such a way that it is recommendable to avoid surveying at strong winds when surveys are part of a long-term monitoring programme. It is concluded that seabird counts on fixed ferry routes are a useful tool to monitor long-term changes in seabird abundance, provided the counting methods are further standardized and the analysis takes the most relevant side-effects into account (e.g. with GLIM)