Geographic Information Systems provide the means to explore the spatial distribution of insect species across various land-use types to understand their relationship with shared or overlapping spatiotemporal resources. Blow fly species richness and total fly abundance were correlated among six land-use types (residential, commercial, waste, woods, roads, and agricultural crop types) and distance to streams. To generate multivariate models of species richness and total fly abundance, blow fly trapping sites were chosen across the land-use gradient of Windsor-Essex County (Ontario, Canada) using a stratified random sampling approach. Sampling occurred in mid-June (spring), late August (summer), and late October (fall). Spring species richness correlated highest to residential (-), woods (-), distance to streams (+), and tomato fields (+) in models across all three land-use buffer scale distances (0.5, 1, 2 km), with waste (+/-), roads (-), wheat/corn (-), and commercial (-) correlating at only two of the three scales. Spring total fly abundance correlated with all but one land-use variable across all buffer scale distances, but the distance to streams (+), followed by orchards/vineyards (+) exhibited the greatest importance to these models. Summer blow fly species richness correlated with roads (-) and commercial (+) across all buffer distances, whereas at two of three buffer distances wheat/corn (-), residential (+), distance to streams (+), waste (-), and orchards/vineyards (+) were also important. Summer total fly abundance correlated to models with distance to streams (+), orchards/vineyards (+), and sugar beets/other vegetables (+) at the 2 km scale. Species richness and total abundance models at the 0.5 km buffer distance exhibited the highest correlation, lowest root mean square error, and similar prediction error to those derived at larger buffer distances. This study provides baseline methods and models for future validation and expansion of species-specific knowledge regarding adult blow fly relationships with spatiotemporal resources across land-use types and landscape features.
Read full abstract