Shipboard line transect surveys of Crabeater seal abundance in the pack-ice off eastern Antarctica: evaluation of assumption

We examine the extent to which the assumptions underlying line transect sampling are satisfied in shipboard surveys of crabeater seals (Lobodon carcinophaga) hauled out on the Antarctic pack-ice. Measurement of the perpendicular distance of seal groups from the ship with an electronic inclinometer fitted to a rifle stock was unbiased. Crabeater seals showed little movement in response to the approaching ship. Movement away from the ship by seals close to the ship's track was partially responsible for a relative lack of sightings close to the transect-line, but otherwise had little effect on the sighting histogram. Minor deviations from the transect direction to avoid running over seals violated the assumption of uniform distribution of groups, and contributed to a relative lack in sightings close to (<40 m) the transect-line. We estimate that 5%-10% of seal groups close to the transect-line were not sighted by bridge observers prior to passing abeam of the ship, but most of these missed groups were likely to have been sighted some distance behind-abeam. Shipboard transects provided a biased sample of four environmental features known to be related to crabeater seal abundance because of logistical difficulties in the ship traversing along straight transects through thick ice. Calculation of transect length L from successive GPS locations was mildly sensitive to the frequency of locations. We provide analytical recommendations to reduce or eliminate the effect of assumption violation when present and hence minimize bias in abundance estimation.