Can coarse-grain patterns in insect atlas data predict local occupancy?

摘要: Aim Species atlases provide an economical way to collect data with national coverage, but are typically too coarse-grained monitor fine-grain patterns in rarity, distribution and abundance. We test the performance of ten downscaling models extrapolating occupancy across two orders magnitude. To a greater challenge models, we extend previous tests plants highly mobile insect taxa (Odonata) life history that is tied freshwater bodies for reproduction. investigate species-level correlates predictive accuracy best performing model understand whether traits driving spatial structure can cause interspecific variation success. Location Mainland Britain. Methods Occupancy 38 British Odonata species were extracted from Dragonfly Recording Network (DRN). Occupancy at grains ≥ 100 km2 was used as training parameterize models. Predicted 25, 4 1 km2 grains compared observed corresponding grains. Model error evaluated grains. Main conclusions The Hui gave most accurate predictions 114 species:grain combinations 14 species, despite being only using information single grain. The occupancy–area relationship sigmoidal shape species. Species' type dispersal ability explained over half level. Species climatic range limit Britain poorly predicted other types, high associated relatively poor predictions. Our results suggest widely available coarse-grain atlas data, reasonable estimates occupancy, even strong structure. Linking reveals general principles about when will be successful.