{ "layers": [ { "currentVersion": 10.91, "cimVersion": "2.9.0", "id": 0, "name": "RSR", "type": "Raster Layer", "description": "
The Map of Biodiversity Importance (MoBI) consists of a series of raster maps that combine habitat information for 2,216 imperiled species occurring in the conterminous United States, using weightings based on range size and degree of protection to identify areas of high importance for biodiversity conservation. Species included in the project are those which, as of September 2018, had a global conservation status of G1 (critical imperiled) or G2 (imperiled) or which are listed as threatened or endangered at the full species level under the United States Endangered Species Act. Taxonomic groups included in the project are vertebrates (birds, mammals, amphibians, reptiles, turtles, crocodilians, and freshwater and anadromous fishes), vascular plants, selected aquatic invertebrates (freshwater mussels and crayfish) and selected pollinators (bumblebees, butterflies, and skippers).<\/SPAN><\/SPAN><\/P> This specific raster depicts summed range-size rarity for vertebrates, vascular plants, selected aquatic invertebrates (mussels and crayfish) and selected pollinators (bumblebees and butterflies and skippers) based on habitat data for 2216 species.<\/SPAN><\/SPAN><\/P> NatureServe developed a habitat map for each species. Habitat maps for most species are model outputs generated using the random forest algorithm. Data inputs into the models include forty years of NatureServe Network element occurrence and observation data supplemented with records from USGS BISON and other sources and environmental predictors representing terrain, climate, land cover, soils, and hydrology. A national library of over 200 terrestrial and aquatic predictor variables supported the project. The resolution of modeling was 30m for most terrestrial species and 330m resolution for some wide-ranging species. Habitat models for aquatic species used a similar approach, but the medium resolution National Hydrography Dataset (NHD) was the modeling unit rather than a raster grid. If an existing, range-wide, and vetted model was easily obtainable for a species, we used that model instead of developing a new one. For species not amenable to inductive habitat modeling, we developed or obtained alternative habitat maps built using deductive approaches. We upscaled or otherwise converted habitat maps for all species to a 990-m raster to provide a consistent unit of measurement across species and avoid revealing the precise location of highly sensitive species. <\/SPAN><\/SPAN><\/P>