| Quantification of forage availability is necessary for the management of large herbivores and their habitats. Creating detailed species maps covering large extents is not possible using field surveys, aerial photography, or satellite imagery singly, and effective methods that combine these approaches have not been developed. The objective of this research was to determine the feasibility and accuracy of using a hierarchical scaling approach to map moose forage. The approach entailed mapping and classifying plant patches using ground surveys coupled with fine resolution photography. These classifications then provided training sets for supervised and subpixel classifications of Landsat imagery to generate landscape scale maps of forages. The major forages, willow, alder, and dwarf birch, were spectrally distinct in fall imagery. Subpixel classifications of Landsat imagery produced superior results compared to supervised classifications. This hierarchical approach provides more accurate assessments of food availability for wildlife than those currently used by land management agencies. |
