Mapping wildness for protected area management: A methodological approach and application to the Dolomites UNESCO World Heritage Site (Italy)

Wildness maps may provide valuable information for the management of natural and protected areas (e.g. Recreation Opportunity Spectrum). This requires the adoption of mapping methods that can handle the relative nature of wildness, providing consistent evaluations for any context of analysis and supplying outputs that can be directly applied by park managers. To this purpose, a novel mapping approach is introduced that uses unsupervised classification to automatically cluster land parcels sharing similar wildness characteristics, as described by a set of spatial indicators. Wildness maps of the Dolomites UNESCO World Heritage Site (Italy) were generated by considering seven indicators of remoteness, perception and naturalness, and assigning each pixel of the study area to one of three classes (i.e. wild, semi wild, non wild), based on their values for the above-mentioned indicators. Results of our application showed a good degree of concordance with wildness maps obtained through Multi Criteria Evaluation (MCE) and emphasized how the class-based output may directly inform zoning activities and the identification of recreational opportunities. While lack of user's control is an obstacle to incorporating the views of multiple groups, as it is allowed by MCE-based methods, the proposed approach supports the idea that land characteristics should define the context of wilderness and drive management decisions. Further applications to a wide set of different contexts can help validate this approach. © 2013 Elsevier B.V. All rights reserved.