Forest decline model development with LANDSAT TM, SPOT, and DEM DATA

The relationships between percent defoliation and digital near-infrared reflectance data detected by the Landsat thematic mapper and SPOT sensors were investigated. These data were both found to be negatively correlated with defoliation data collected within the boreal montane spruce-fir ecosystem of the Black Mountains, North Carolina. Correlation coefficients were significant at the 0.05 level. Linear regression analysis demonstrated that neither source of satellite-based remotely-sensed data is an accurate predictor of defoliation. The addition of digital elevation data, however, as an independent variable to the regression equations significantly improved the predictive reliability of the models.