Towards Global Monitoring of Forest Cover

摘要

The open Landsat archive offers the opportunity to employ a global, moderate spatial resolution for monitoring forest change. A method for systematically processing Landsat imagery was implemented for four test areas: Indonesia, the Democratic Republic of the Congo, European Russia and Quebec. All Landsat imagery with metadata less than 50% cloud cover was used as inputs. For Indonesia and the Democratic Republic of the Congo, forest extent and loss were mapped from 2000 to 2010. For European Russia and Quebec, forest extent and loss were mapped from 2000 to 2005. All results were compared with samples of expert-interpreted Landsat samples, chosen within a probability-based sample framework. The method can be implemented genetically at the biome scale with results illustrating a capability for large-scale processing of Landsat imagery for global forest extent and change assessments.