抖阴探探

ances that were detected in Landsat time-series imagery by Landtrendr on Google Earth Engine (LT-GEE) under three different algorithm configurations or 'tunings'. The 'combo6' and 'rt75' tunings included were developed via this project, and a 'defaults' tuning (which we slightly adjusted from actual LT-GEE defaults) is provided for reference. All were based on LT-GEE segmentation of the Normalized Burn Ratio (NBR) spectral index as derived from Landsat c2. These tunings were selected to provide the 'best' overall range of results, based on an iterative cross-validation ('ground-truthing') procedure using extensive land use/harvesting records, time-series imagery (TimeSync) and field reconnaissance. Stacks are rasters (.tif files) with six bands, each band representing an output variable, including year of disturbance, magnitude, duration, rate of change and others (see metadata for detailed variable definitions). As such, they should be visualized and analyzed one variable or band at a time; combining two or more bands for GIS-based visualization is not recommended. Each of the three sets of LT-GEE outputs was further parsed into four geographic subregions within the FEMC service area: 1) NY, 2) ME, 3) MA, CT and RI, 4) VT and NH. This resulted in a total of 12 tif files.