Forest changes at large spatial scales can be analyzed effectively using remotely sensed data and geographic methods. In my work I introduced a metric using distance to closest forest as an indicator for qualitative assessment of forest changes related to forest spatial patterns change (and by proxy, forest consolidation) in ecological regions across the U.S. The metric was first aggregated in each ecoregion to show the general geographic variations of forest consolidation and help to identify forest loss of high spatial uniqueness. In addition, consolidation affected by forest changes in neighboring ecoregions was studied and significant geographic variability was found between ecoregion behaviors. Furthermore, observed consolidation was contrasted with random simulations by introducing forest loss of complete spatial randomness, with results indicating significantly higher observed that expected consolidation on the western ecoregions. Finally, consolidation sensitivity was assessed in different landscape segments allowing comparison between consolidation potential of all non-forested areas versus exclusive urban or water areas. Water and urban areas generally had higher consolidation sensitivity than overall non-forested areas. Furthermore, from the geographical perspective the highest sensitivities of water and urban areas were distributed along the Appalachian mountain range. The findings of this thesis assist towards more effective management strategies to meet specific objectives in ecoregions as well as to facilitate management plans at the national level.

Keywords: forest consolidation, forest loss, spatial randomness, geographic information science, remote sensing