Estimation of Forest Stand Structure Attributes from Aerial Photographs

Abstract

Awareness of potential inaccuracies in stand structure data derived from remotely sensed imagery is important in landscape-level analyses. Attributes of forest stand structure were estimated from aerial photographs using standardized photointerpretation procedures and accuracy assessments were conducted using error matrices by comparing the estimates to plot data. Over 500 stands were photo-interpreted by a single interpreter (1:15,840 nominal scale, normal color) for six forest structure attributes (canopy cover, stand height, DBH, cover type, crown diameter, and canopy layers). Percentage accuracy adjusted for chance agreement ranged from 29 percent to 59 percent; percentage accuracy varied according to the techniques used to evaluate individual attributes. Potential means for correcting biases and rating landscape analyses are discussed.