Appropriate assessment of the accuracy of burned area products is required to assess greenhouse gas (GHG) emissions in a reliable way. This paper provides validation results for three burned area products with different pixel sizes: MCD45 (500 m), GlobCarbon (1 km), and L3JRC (1 km) for the Orinoco River basin, which is widely affected by fires, many of them oriented towards the conversion of Amazonian forest to cattle pasture. Accuracy measurements were based on commission and omission errors computed from 15 confusion matrices, as well as in the Pareto Boundary concept, which evaluates the impact of different pixel sizes in omission and commission errors. An edge metric was used to analyze the impact of shapes of burned patches on global accuracy. It was shown that all products underestimate burned-area and that an increase in pixel size or burned patch elongation, results in larger estimation errors. © 2012 American Society for Photogrammetry and Remote Sensin.

• Fecha de publicación: 2012
• Institución: Universidad de Medellín
• Repositorio: Repositorio UDEM
• Idioma: eng
• OAI Identifier: oai:repository.udem.edu.co:11407/1413
• Acceso en línea: http://hdl.handle.net/11407/1413
• Palabra clave: Accuracy assessment; Accuracy measurements; Amazonian forests; Burned areas; Burned patches; Confusion matrices; Estimation errors; Omission errors; Pareto boundary; Pixel size; River basins; Validation results; Agriculture; Errors; Greenhouse gases; Pixels; Rating; accuracy assessment; cattle; error analysis; forest ecosystem; forest fire; greenhouse gas; pasture; pixel; Amazonia; Orinoco Basin; Bos
• Rights: restrictedAccess
• License: http://purl.org/coar/access_right/c_16ec