Influence of scale (grain and extent) on explaining bird distribution in human-dominated landscapes of eastern Ontario, Canada

The appropriate spatial scale for ecological investigation is crucial because patterns and controlling processes may operate at different scales. Using remotely sensed data, I examined the effect of changing grain (4 m - 1 km) and extent (90 m -10 km) of 5 variables on explaining the variation in bird community composition in Eastern Ontario. Variation partitioning using canonical correspondence analysis revealed subtle differences between the amount of variation uniquely explained by the variable grains and extents examined. Forest amount at the finest scale accounted for nearly a third (29%) of the total variation (48%) accounted for in the bird data, but all variable grains and extents were significant. While sampling across a broader environmental gradient will help clarify the roles of variable grain and extent, these results reinforce the importance of fine-scale habitat composition and incorporating variables measured at a range of grains and extents in bird community modelling.