Evaluation of biodiversity value based on wildness: A study of the westernNorthwoods, Upper Great Lakes, USA

Regional conservation plans are developed and implemented to protect areas from human land uses that adversely affect biological diversity. Several di fferent criteria have been used to identify locations with high biodiversit y value; among these is "wildness." We evaluated the extent of wildness of the Upper Great Lakes region of the United States (the "Northwoods") to ide ntify areas for inclusion in a regional conservation plan. Using a geograph ic information system to organize data, we evaluated wildness using six dif ferent data layers in raster format: land cover classification, forest cove r type, land ownership, human population density, road density, and wolf (C anis lupus L.) habitat suitability. For each data layer, we quantified wild ness (more vs, less wild) based on threshold criteria. We classified pixels within data layers as more wild if: (1) forest is present, (2) the forest has has the same general species composition since European settlement, (3) the area is in public ownership, (4) human population density is < 1 perso n per km(2), (5) road density is < 0.5 km of road per km(2), and (6) the pr obability that wolves would colonize the area exceeds 75%. We found all the se measures of wildness to be significantly associated with each other base d on chi-square tests of independence (adjusted P << 0.001). We created a n ew data layer by combining five existing layers (all except the wolf coloni zation Layer) using an overlay procedure. This new layer shows a gradient o f wildness across the region, with only 5.7% of all pixels meeting all five wildness criteria. The probability of wolves colonizing an area increases with its wildness. The methods we use to evaluate wildness here can be used in conjunction with other methods of regional biodiversity assessment to g uide acquisition and easement efforts, zoning policies, and land management activities as opportunities arise.