We used the 2011 National Land Cover Database (NLCD) U.S. Forest Service Tree Canopy (analytical) product (USDA Forest Service Remote Sensing Applications Center 2014) combined with forested wetland mask derived above for assessment of overstory canopy cover within forested wetlands in the Mississippi Alluvial Valley (MAV) and other GCPO subgeographies. The USFS forest canopy layer contains values representing the unmasked proportion of each 30x30m pixel covered by tree canopy (0 to 100%) produced using random forest regression algorithms. To align with resolution of the forested wetland mask we sought to generate an average proportion of tree canopy across 30 m pixels within each 250 m forested wetland cell in the GCPO geography. To calculate average canopy cover we first aggregated 30 m canopy cover cells to 240 m using a mean function and a cell factor of 8 (the aggregate function in ArcGIS only allows for cell factor aggregation, not aggregation to a desired pixel size). We then resampled the 240 m cell aggregate to 250 m resolution using a nearest neighbor algorithm. This produced an approximation of the average tree canopy cover within each 250 m forested wetland pixel. We next reclassified the average tree canopy layer (0-60% = 0; 60-70% = 1; 70-100% = 0) to identify 250 m cells with 60-70% forest canopy cover.
USDA Forest Service Remote Sensing Applications Center. 2014. NLCD 2011 USFS Percent Tree Canopy (Analytical Version).