Kelp-bed dynamics across scales: Expanding distributional analysis capacity with spatial pattern metrics and modelling

Spatial pattern metrics are used to investigate the links between a species' distributional patterns and underlying ecological drivers. These metrics, in combination with remotely acquired imagery, offer a novel approach for the study of kelp bed-urchin barren systems and enable the study of fully submerged kelp distribution at broad spatial scales (km(2)) difficult to achieve through traditional scuba-based methods. This study uses aerial imagery of 2.85 km(2) of shallow (<7 m deep) seabed around five islands in the Mingan Archipelago (northern Gulf of St. Lawrence, Canada) to (1) quantify kelp distribution patterns using spatial pattern metrics at large, intermediate, and small spatial scales and (2) examine correlations between kelp presence and physical and biological variables over a broad spatial scale. Imagery was classified visually and divided into two benthic classes: kelp and non-kelp. Variability in spatial pattern metrics and kelp coverage across spatial scales indicated that kelp distribution is not uniform among islands, suggesting that the spatial extent over which observations are obtained strongly influences the patterns detected. Kelp covered 62 % of the entire study area but with substantial variation (46 % to 87 %) among islands. Kelp and non-kelp patches varied in size from 225 to 891,225 m(2), while exhibiting considerable variation in geometric complexity. Over 80 % of the kelp patches were relatively small (<1350 m(2)), although the fewer, larger patches contained most (98 %) of the kelp-covered seabed and were located in shallower water, near the coastline. Both the kelp and non-kelp benthic classes were highly aggregated as suggested by clumpiness indices of 0.67 and 0.59, respectively. Modelling results showed that increasing depth, urchin density, and relative exposure to waves independently led to a decrease in kelp presence, with depth having the strongest correlation, followed by urchin density and relative exposure. This study expands knowledge of kelp distribution patterns at broad spatial scales largely unexplored for fully submerged species and is a step forward towards a comprehensive understanding of scale-dependent processes regulating their distribution.