Submerged aquatic vegetation SAV (GCPO LCC Gulf Assessment Version)

We obtained multiple primary and supplemental static data sources for the assessment of SAV in GCPO Gulf Coast estuaries. Our primary data source was the Gulf-wide SAV layer produced by Handley et al. (2011) as part of the 1940-2002 Seagrass Status and Trends Report (Handley et al. 2007), which compiled multi-source photo-interpreted and field-verified seagrass data spanning up to 1999 as part of the Gulf of Mexico Ecosystem Pilot Project. This dataset contains strata of seagrass composition (<10%, 15-40%, 45-70%, 75-85% patchy seagrass cover; >90% continuous seagrass cover). We also incorporated the Seagrasses in the continental United States as of March 2015 data layer compiled and posted on MarineCadastre.gov by the NOAA Office of Coast Management in partnership with DOI Bureau of Ocean Energy Management MarineCadastre.gov. This dataset combines submerged, rooted vascular species and submerged or rooted floating freshwater tidal vascular species data in the Gulf Coast and other Eastern U.S. geographies, including the Mobile Bay National Estuary Program, Florida Fish and Wildlife Research Institute, Florida Minerals Management Service, NOAA National Centers for Coastal Ocean Science, USGS National Wetlands Research Center, University of Southern Mississippi, and NOAA Office for Coastal Management among others. We supplemented the Gulf-wide SAV and Continental Seagrass layers using 2014 National Wetland Inventory data, extracting aquatic bed rooted vascular classes (estuarine subtidal [E1AB3], estuarine intertidal [E2AB3], marine subtidal [M1AB3], marine intertidal [M2AB3]). We also incorporated seagrass (gained and unchanged categories) from local efforts in the waters adjacent to the Mississippi Sound barrier islands (Cat Island, Horn Island, Ship Island, Petit Bois Island), and SAV mapping efforts of the Grand Bay National Estuarine Research Reserve. We also explored the feasibility of using NOAA Coastal Change Assessment Program (C-CAP) land cover data estuarine aquatic bed class (class 23), though this class includes rooted and floating vascular vegetation, in addition to floating algal mats. We also examined the Florida Cooperative Land Cover (CLC) V3 submergent aquatic vegetation class (2150), but found no SAV polygons classified within the GCPO LCC portions of the western Florida panhandle. All layers were assessed within the GCPO LCC geographic extent and extending out to state seaward boundaries.

After examining all aforementioned data layers we decided to use Gulf-wide SAV, NOAA Continental Seagrass, NWI, Mississippi Sound Islands, and Grand Bay NERR data for this assessment (excluding C-CAP and Florida CLC data). We first converted each vector layer to a 10m pixel raster layer, clipped to a 10km buffer around the GCPO LCC. We then reclassified to a binary pixel layer (seagrass present = 1, absent = 0) and summed overlays of each data product to examine areas where datasets were in agreement, mosaicing summed products where necessary. We next reclassified to two binary data sets: 1) where all summed pixels >1 were included (i.e., all unique pixels where SAV was classified equally, regardless of presence in a single or multiple datasets); and 2) only where pixel sums ≥2 existed, such that 2 or more data layers must support evidence of SAV presence to be counted. After examining both binary pixel sum datasets, it was clear that the conservative process of eliminating all pixels with sums of <2 resulted in an underrepresentation of SAV in areas where pixels of SAV cover were non-overlapping, but that including all pixels resulted in a gross overrepresentation of SAV. Given that neither of these situations was ideal, we chose to summarize the data conservatively (requiring pixel sums ≥2, or two or more data layers must indicate a pixel is SAV to be included. Note that this does limit the application of this assessment for areas of patchy SAV distribution where non-overlapping datasets occurred (particularly in the Mississippi Sound). For simplicity and to reflect the dynamic nature and resource value of seagrass systems we assessed patchy and continuous seagrass concomitantly.