We analyzed the relationship between Landsat ETM+ reflectance values and commercially managed loblolly pine (Pinus taeda L.) stand characteristics in east Texas. Multivariate regression techniques were used to predict age and tree density for all stands. A linear combination of NDVI, ETM4/ETM3 and the tasseled cap wetness index was a better predictor of stand age (R2 = 78%) than other combinations of original bands and derived indices. However, models involving transformed bands did not improve the overall predictability of stand density (R2 = 60%). Results from the principal component analyses (PCA) conducted on mature stands (age > 18 years) yielded valuable information about the relationship between stand structure and reflectance values recorded by the ETM+ sensor. The first principal component was interpreted as a measure of stand complexity. A linear regression model with infrared bands 4 and 5 as independent variables was able to account for 76% of the variability in stand structure. A second model with transformed bands did not increase the amount of variability in stand structure. Results obtained from this study demonstrate the relationship between loblolly pine stand characteristics and ETM+ reflectance values and the utility of certain transformed bands. Forest managers could use ETM+ data for gaining insights about stand characteristics and this information would be also useful for generating maps required for developing forest management plans.
All Science Journal Classification (ASJC) codes
- Nature and Landscape Conservation
- Management, Monitoring, Policy and Law