Monitoring seabirds population in marine ecosystem: The use of strip-transect aerial surveys

In the case of marine systems, monitoring seabird populations at sea require adequate, low cost protocols that maximize data resolution, accuracy and survey effort. This paper examines bias associated with pelagic aerial strip-transects, an easily designed and low cost survey method, to validate its use for monitoring and mapping seabird populations at sea at a very large scale. We used data collected on the entire seabird community between October 2001 and March 2002 in the Bay of Biscay (100000 km(2) on the French Atlantic coast, Northeast Atlantic) to determine whether aerial strip-transect seabird survey were subject to temporal (i.e. variations of detection probability during the survey) or distance (i.e. variations of detection probability across the strip) bias. To detect these biases, two seabird taxa were contrasted, the highly conspicuous northern gannet versus three cryptic species of auks pooled as one taxonomic group. We reported that the only temporal bias was attributable to sun glare. By contrast, distance bias did not occur in a strip of 150 m for both species, and the effect of distance up to 230 m was very weak. We conclude that visibility bias have a limited effect on pelagic aerial surveys for most seabird species. Particular attention should be paid to the constant record of sun glare, while altitude and speed can be fixed in a relatively wide range (140-180 m, 150-200 km/h) without affecting detection probability. Since distance bias did not occur across a 150 m strip, there is no need to add complexity to survey protocol by using line-transect method. This method takes distance effects into account but requires distance estimates between the bird and the transect line for every sightings. These estimates are usually easily collected in the case of marine mammals but become problematic in the case of flying animals such seabirds, sometimes encountered in very high density and aggregated into groups of several hundred individuals. (C) 2008 Elsevier Inc. All rights reserved.