Skip to Main Content
U.S. Forest Service
Caring for the land and serving people

United States Department of Agriculture

Home > Search > Publication Information

  1. Share via EmailShare on FacebookShare on LinkedInShare on Twitter
    Dislike this pubLike this pub
    Author(s): Suzanne M. Joy; Robin M. Reich; Richard T. Reynolds
    Date: 2000
    Source: In: Proceedings of the eighth Forest Service remote sensing applications conference: Remote sensing and geospatial technologies for the new millennium; 2000 April 10-14; Albuquerque, NM. Bethesda, MD: American Society for Photogrammetry and Remote Sensing. 11 p.
    Publication Series: Miscellaneous Publication
    PDF: View PDF  (1.9 MB)

    Description

    We used field data, topographical information (elevation, slope, aspect, landform), and Landsat Thematic Mapper imagery to model forest vegetative types to a 10-m resolution on the Kaibab National Forest in northern Arizona. Forest types were identified by clustering the field data and then using a decision tree based on the spectral characteristics of a Landsat image and topographical information to predict the forest types. Significant variables in the models included raw basal area and proportion of basal area by species. Use of additional variables (canopy closure, understory vegetative height, seedling/sapling presence, and proportion of ground covered by vegetation) did not improve the model. Forest types described by the model included pinyon-juniper, ponderosa pine, ponderosa pine-fir mixes, spruce-dominated mixes, deciduous-dominated mixes, and clearings. Sample-based accuracy assessment accounted for 92.9% of the variability in the vegetation model. Error rates (post-stratified) were weighted by the proportion of area each forest type occupied. Independent validation using double sampling with post-stratification accounted for 74.5% of the estimated variability in the model. Ponderosa pine comprised the largest proportion (55.5%) of vegetative area and contributed the highest accuracy estimate (sample-based: 98.0%; cross-validation: 90.8%) to the overall forest model. Identified sources of error included (1) differentiating between pine-fir and spruce-dominated forest types (sample-based assessment) and (2) distinguishing openings in the forest from deciduous-dominated mixes (double sampling). This model has been used to describe forest structure (basal area, canopy cover, maximum understory vegetative height, presence of seedlings and saplings, and proportion of pine, aspen, spruce and fir basal areas) on the Kaibab National Forest to a 10-m resolution. Models of forest composition and structure will be linked with point-process models and a ranking of territories of northern goshawks with the purpose of identifying determinants of goshawk habitat quality.

    Publication Notes

    • You may send email to rmrspubrequest@fs.fed.us to request a hard copy of this publication.
    • (Please specify exactly which publication you are requesting and your mailing address.)
    • We recommend that you also print this page and attach it to the printout of the article, to retain the full citation information.
    • This article was written and prepared by U.S. Government employees on official time, and is therefore in the public domain.

    Citation

    Joy, Suzanne M.; Reich, Robin M.; Reynolds, Richard T. 2000. Modeling small-scale variability in the composition of goshawk habitat on the Kaibab National Forest. In: Proceedings of the eighth Forest Service remote sensing applications conference: Remote sensing and geospatial technologies for the new millennium; 2000 April 10-14; Albuquerque, NM. Bethesda, MD: American Society for Photogrammetry and Remote Sensing. 11 p.

    Keywords

    classification, Landsat Thematic Mapper, double sampling, forest composition, northern goshawk habitat, Accipiter gentilis, Kaibab National Forest

    Related Search


    XML: View XML
Show More
Show Fewer
Jump to Top of Page