Identifying conservation-priority areas in a fragmented minnesota landscape based on the umbrella species concept and selection of large patches of natural vegetation

Two common but relatively untested approaches to prioritizing conservation areas involve identification of umbrella-species habitat and selection of l arge blocks of remaining natural vegetation. We tested an umbrella-species approach to identification of conservation areas in the Agassiz Beach Ridge s landscape of Minnesota based on the habitat needs of the Greater Prairie Chicken (Tympanuchus cupido pinnatus). All natural vegetation patches that fell within 1.6 km of prairie chicken booming grounds were selected as cons ervation priorities and were compared with randomly selected vegetation pat ches. Patches selected by the umbrella-species approach encompassed slightl y more biologically important land (59%) than selection of the largest patc hes (56%), and both these sets contained more biologically important land t han randomly selected patches (47%). The set of largest natural vegetation patches encompassed the greatest number of natural community types (12 of 1 6 possible). Patches selected by the umbrella-species approach encompassed more occurrence (individual locations) of species and communities than both other approaches. Using the habitat requirements of prairie chicken to ide ntify conservation area adequately addressed other grassland species and co mmunities but not uncommon forested habitats. In many cases, the large-patc h approach did equally well or better at protecting various biodiversity co mponents, indicating that the umbrella-species and large-patch approaches a re complementary in this landscape. Because we had long-term information on our umbrella species, we were able to explore the setting of conservation priorities using different levels and types of data. As expected, increasin g the number of booming grounds by combing years of data (n = 25 to n = 337 ) generally increased the number and area of selected vegetation patches an d the number of occurrences and amount of biologically important land conta ined within them. We also found that data sets with the fewest booming grou nds often produced the most efficient results, such as most biologically im portant land per unit area selected. We observed that stable booming ground s were more efficient in incorporating other components of biodiversity per unit area of selected habitat than were temporary booming grounds. These r esults demonstrate that the quantity and quality of data on umbrella specie s will greatly affect planning results.