Manipulation of Landsat Spectral Characteristics to Classify Vegetation and Soil Wetness in the Rainforest of Bolivia,
Abstract
This study attempts to classify tropical region soils and vegetation by moisture content from multispectral imagery. Identified wet areas were used to determine the percentage of wetness in the study area, by evaluating the spectral response of tropical rainforest vegetation and soils. Supervised classification, unsupervised classification, and manipulation of spectral band techniques were used to determine percentages of wetness in the study area. Using these methodologies, vegetation and soil units associated with wet conditions were classified. Soil types were categorized into the following: (1) areas that are moist, (2) forested areas that are moist or wet depending upon the season, and (3) forested/swamp/land subject to inundation, areas that are wet a good portion of the year. Vegetation was classified into the following: (1) marsh, (2) tropical swamp forest, and (3) tropical moist forest. Results demonstrate that digital spectral data from Landsat imagery can be used to locate and evaluate varying degrees of wetness in soils and different vegetation types associated with wet conditions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1992
- Accession Number
- ADA257847
Entities
People
- Harry B. Puffenberger
- Michael G. Barwick
Organizations
- Army Geospatial Center