Technical note: A spectral-angle methodology for mapping net forest cover change in northeastern China

Abstract

As part of the Global Observations of Forest Cover/Land Cover Dynamics (GOFC-GOLD) project, forest-cover change in northeastern China is being mapped using Landsat-5 and Landsat-7 imagery. This technical note details some of the challenges associated with mapping net forest-cover conversion in the region, and presents a new algorithm that overcomes some of these challenges. The mapping approach uses the concept of spectral angle to establish change between images, and implements spectral thresholds to determine the type of change. Validation results give an overall accuracy of 84-96% and a detection accuracy of 72-83%.