I am working with a REDD methodology that calls for measuring transitions between dryland forest degraded dryland forest going back 6+ and 10+ years. For the 2005-2011 period, I can classify degradation using landsat TM/ETM imagery and field reference data from Feb-11 but have nothing for the prior period. In a more densly forested area one might be able to detect the degradation visually (e.g., using ndvi, band 5s from multiple dates), but here I can’t in most cases so I am flying blind. How to proceed? Perhaps one could start with the assumptions that degradation is likely to occur a. in a subset of the areas where it was observed in Feb-11 and b. on the periphery of observed deforestation areas and try to detect greenness/wetness changes there. Seems sketchy. Any insights are most welcome. Thank you.