Demonstration of Semi-Automated Approaches for Monitoring National Tropical Deforestation

International agreements of the United Nations Framework Convention on Climate Change (UNFCCC) seek to reduce national forest-related emissions via activities and policies on Reduced Emissions from Deforestation and Degradation, plus the role of conservation, sustainable forest management and enhancement of forest carbon stocks (REDD+). These agreements require countries to develop systems to monitor and report changes in forest cover and carbon stocks via systems of Measurement, Reporting and Verification (MRV) that include forest monitoring. We demonstrate three methodologies proposed for semi-automated monitoring of deforestation at the national level: ClasLite, ImgTool, and Global Land Analysis and Discovery (GLAD). Methods are based on Landsat satellite images and compared for two common study areas in Mexico and Colombia. High-resolution images from RapidEye were interpreted for comparison with the methods’ results in deforestation estimates. Comparisons with samples of interpreted points from the high-resolution images indicate overall high levels of accuracy and agreement; however, an evaluation of the deforestation rates and patterns over the entire study area indicate significant differences, especially in Mexico, which contains dry forest and a large portion of secondary forest fallow. We provide a discussion of implications for monitoring and recommendations for further study.