| The forest cover in settled landscapes, such as southern Ontario, has been shaped by agriculture, forestry and urban development for more than two hundred years. In the Region of York, used as the study area, the existing forest fragments are heavily utilized remnants that have either survived land clearing for agriculture; are secondary woodlands on abandoned land; or are forest plantations. These forests, along with lands that were once forested and that have the potential to be forested again, are all threatened directly or indirectly by land-use conversion processes and development. Present forest management practices focus primarily on current status of the forest remnants and promote the protection of selected ones. These practices base conservation and management on their current state, with no regards to their past or future distributions and characteristics.; This study introduces a new approach to forest conservation instructing that effective landscape level forest management and conservation should be carried out in the context of the vegetation types that could be supported by the site conditions—potential natural vegetation (PNV). PNV maps show the most likely distribution of vegetation types in a landscape. This distribution is the reference condition that, in part, can inform and direct decisions around forest conservation, management and planning, and broader land use planning. The research provides a methodology to predictively map dominant tree distribution using historic land surveyors' records and from these to derive PNV maps. The PNV and existing vegetation maps were used to identify gaps in the distribution of natural vegetation and their habitats, which provide baseline information for directing forest conservation, restoration, and management in settled landscapes.; PNV maps were derived from the individual species models following developed forest classification rules. The 13 individual species models were developed using the classification tree methodology, vegetation data from the land surveyors' records, and readily available environmental data.; This study demonstrates how readily available information on vegetation and environmental data may be integrated to model potential tree species and PNV distributions and to define conservation and restoration priorities in settled landscapes. It also introduces a new method for PNV mapping. |
