Description of spatial pattern in seabird distributions along line transects using neighbour K statistics

Description of spatial pattern is an important step towards understanding the underlying, pattern-generating processes, but few statistical techniques adequately describe both the grain and intensity of a spatial distribution. Counts of seabirds collected along line transects are particularly problematic because the data may contain a high proportion of zero counts, and the pattern at increased spatial scales is dependent on how the counts are grouped. Statistics based on Ripley's K function seem well-suited to characterising scale-dependent spatial structure in such data. The spatial distributions of seabirds and fish schools off the Otago Peninsula, New Zealand, are described to illustrate the application of these statistics. Aggregation characteristics of sooty shearwaters Puffinus griseus, red-and black-billed gulls Larus novaehollandiae and L. bulleri and black-backed gulls L. dominicanus were similar. Sitting seabirds were highly clustered, with patch lengths between 145 and 2850 m. The distributions of flying seabirds were less crowded, with patch lengths from 135 to 5290 m. Patches of krill-eating red-and black-billed gulls were longer along transects with high salinity structure. This may reflect a change in foraging strategy in response to visible surface fronts. There was no clear association between the distribution of sitting birds and the presence of schools of predatory, pelagic fish.