Tomorrow I get to dump 10G of Landsat scenes into a 16T super-computer that was built specifically for processing all of the data: 33 scenes each for three study areas, each being unpacked from the tar file, geometrically corrected, radiometrically corrected, atmospherically corrected etc, then processed into NDVI trends omg I want to WATCH. Apparently when all the data are  unpacked and processed each study area will have ~10T of data. All of the processing @_@ 


quick example showing the importance of surface reflectance correction for Landsat 8 imagery (and most other satellites’ imagery too).

at the top, surface reflectance-corrected NDVI along a swath of western Israel. productive vegetation really pops out here.

at the bottom, NDVI without any surface reflectance correction. highs are washed with mediums, and lows look even lower.