Potential contact statistics for measuring scale-dependent spatial pattern and association:: an example of northern cod (Gadus morhua) and capelin (Mallotus villosus)

Analysis of simulated data showed that potential contact statistics could be used to describe spatial pattern in sample density data. Potential contact is a new method, analogous to Ripley's K function for mapped point pattern analysis. Potential contact can be used to describe spatial pattern and association over a range of scales without grouping data and is robust against the presence of zeros. The statistical output is ecologically interpretable, as a measure of the degree of contact between individuals. This new technique was applied to examine changes in the spatial distribution of Atlantic cod (Gadus morhua) off Newfoundland, Canada, from 1985 to 1994, a period that encompassed a collapse of the cod stock. Sample data from bottom-trawl surveys indicated that cod were aggregated in patches with dimensions of 100-250 km. During the period of cod decline in the 1990s, spatial structure changed in three ways: the number of patches decreased, patch size shrank, and contact with conspecifics at small (10-20 km) scales fell. Cod were broadly associated with capelin (Mallotus villosus), a major prey species. Spatial distribution of capelin changed over the same time period as changes in cod distribution, and there was no evidence that contact between cod and capelin decreased.