A forester's look at the application of image manipulation techniques to multitemporal Landsat data

Registered, multitemporal Landsat data of a study area in central Pennsylvania were analyzed to detect and assess changes in the forest canopy resulting from insect defoliation. Images taken July 19, 1976, and June 27, 1977, were chosen specifically to represent forest canopy conditions before and after defoliation, respectively. Several image manipulation and data transformation techniques, developed primarily for estimating agricultural and rangeland standing green biomass, were applied to these data. The applicability of each technique for estimating the severity of forest canopy defoliation was then evaluated. All techniques tested had highly correlated results. In all cases, heavy defoliation was discriminated from healthy forest. Areas of moderate defoliation were confused with healthy forest on northwest (NW) aspects, but were distinct from healthy forest conditions on southeast (SE)-facing slopes.