Comparing farmer-based and satellite-derived deforestation estimates in the Amazon basin using a hybrid classifier

摘要: The Amazon basin remains a major hotspot of tropical deforestation, presenting clear need for timely, accurate and consistent data on forest cover change. We assessed the utility hybrid classification technique, iterative guided spectral class rejection (IGSCR), accurately mapping Amazonian deforestation using annual imagery from Landsat Thematic Mapper (TM) Enhanced Plus (ETM+) 1992 to 2002. mean overall accuracy 11 classifications was 95% with standard deviation 1.4%, z-score analysis revealed that all were significant at 0.05 level. IGSCR thus seems inherently suitable monitoring in Amazon. resulting sufficiently assess preliminarily magnitude causes discrepancies between farmer-reported satellite-based estimates household level sample 220 farms Rondonia mapped field field-and satellite-derived significantly different only 0.10 studied, 8.9% higher than derived situ survey methods. Some this difference due tendency farmers overestimate amount within their property our survey. Given objectivity reduced expense monitoring, we recommend it be an integral part household-level causes, patterns processes deforestation.