A TM-based hardwood-conifer mixture index for closed canopy forests in theOregon Coast Range

The purpose of this study was to develop and evaluate a multi-spectral vege tation index for quantifying relative amounts of hardwood and conifer cover from Thematic Mapper ( TM) imagery. We focused on closed canopy forests in the Oregon Coast Range, where hardwood, conifer, and mixed stand condition s are prevalent. An approach based on the Gramm-Schmidt orthogonalization p rocess was used to derive three different hardwood-conifer mixture indices (HCMIs). Using correlation and regression analyses, the capacity of these i ndices to predict closed canopy hardwood percentage was compared with three other groups of spectral variables: (1) the untransformed TM reflectance b ands, (2) the Tasseled Cap indices of brightness, greenness, and wetness, a nd (3) the first three principal components of closed canopy forest reflect ance. Results show that while similar amounts of information were explained by HCMI, TM band, Tasseled Cap, and principal component models, only predi ctions derived from the HCMI1 and HCMI2 variables were unbiased with respec t to topographic effects.