The Minnesota Breeding Bird Atlas (MNBBA) was a five-year project (2009-2013) completed by over 800 volunteers and paid researchers with the goal of documenting the distribution, abundance, and diversity of Minnesota’s breeding birds. The researchers used georeferenced MNBBA point count and general atlas data to derive spatially explicit species distribution models for 164 breeding bird species. They included three modeling strategies:
- General Linear Models (GLMs) were used for species with >75 detections of singing males (75 species) with an offset to model density.
- GLMs were also used for species with >75 observations but with inadequate detailed data to provide an offset for density calculations (31 species).
- For species that did not fit the first two methods (58 species), Maximum Entropy was used to model relative habitat suitability.
In each case, the researchers used a suite of GIS variables including land use/land cover (e.g., LANDFIRE), disturbance (e.g., road density), land cover structure (e.g., vegetation height), landscape metrics (e.g., patch richness), and climate (e.g., annual precipitation). Model selection was based on AICc using branching forward model selection and bootstrap aggregation. After model selection, all models were projected across the state to produce predicted species distribution maps for each species. The team also estimated statewide breeding populations for the 75 species that used GLMs with an offset.
These models are being developed for a public website and eventually a book on Minnesota’s breeding birds. They will also be useful for a variety of applications including bird watching, improved land management, consideration on climate change issues, and for conservation.