The circumpolar Arctic is predicted to be particularly affected by global change, including climate change. Likely ecological vulnerabilities include flora and dependent fauna. Reindeer (Rangifer tarandus), a keystone Arctic herbivore, are of particular importance, ecologically, as well as culturally and socio-economically. Many Arctic regions, particularly in Russia, lack critical data thereby limiting current understanding and the ability to assess consequences of change.

This research develops new botanical and spatial vegetation data for a much understudied region, increases understanding of the landscape and its relationships to reindeer habitat and climate change through multi-scale assessments, and creates the potential for analysing consequences of change. Three primary scales of vegetation analysis and their limitations, uses and value are compared within the Barents region. More detailed examinations are undertaken within study regions in the Nenets Autonomous Okrug, Russia. Plant communities, community structure and environmental influences are described.

The most detailed level of analysis uses new data derived from traditional field-based, species-level botanical assessment. Statistical analyses show distinct community divisions and the importance of particular vegetation species or species groupings in defining these communities. The intermediate scale of analysis defines vegetation communities through development of a thirty-metre resolution map, based on Landsat ETM+ satellite sensor imagery. The study region is unusually heterogeneous, complicating classification development. The third and coarsest scale comprises existing regional and global land-cover assessments, with resolutions of one to five kilometres.

Multi-scale comparison shows that the intermediate level is the most suitable for reindeer pasture and habitat assessment, and that climate-related shifts could be observed on any of the three scales, depending on the objectives, enabling assessment of patterns of change in species, communities or landscape. Long-term monitoring, however, is currently limited to coarser-scale analysis. Replication and broad regional expansion of finer-scale analysis, as developed here, are restricted by data availability, computing power, logistics and cost, amongst other factors.

The thesis ends with a discussion of the potential for habitat assessment that includes additional environmental factors critical to reindeer habitat.

All botanical and vegetation-mapping data and about 500 vegetation species, community and landscape photographs, geo-referenced and with metadata, should be archived in DSpace@Cambridge under the Scott Polar Research Institute archive.