The objective of this study was to assess the applicability of using TM imagery for estimating above ground biomass (AGB) in a small forestland. Based on a single Landsat TM image, spectral reflectance from six TM bands and three vegetation indices was correlated to ground-based AGB measurements using regression analysis to develop AGB models. The fit of 2models was evaluated using the coefficient of determination (R). An accuracy assessment using independent test points was performed for the best model. Overall this study found that lower 22numbers of training points resulted in higher R; however, models with high R did not show high accuracy levels when validated against an independent sample of data. Predicted values for AGB models were consistent among models that modeled: all tree species; all tree species in interior forest stands; hardwood stands; or pine stands.
