Couvillion, Brady Ganju, Neil K. Defne, Zafer 2021 Raster Digital Data Set USGS Data Series doi:10.5066/P97DQXZP doi:10.5066/P97DQXZP https://doi.org/10.5066/P97DQXZP Prior research has shown that sediment budgets, and therefore stability, of microtidal marsh complexes scale with areal unvegetated to vegetated marsh ratios (UVVR) suggesting these metrics are broadly applicable indicators of microtidal marsh vulnerability. This effort has developed the UVVR metric using readily available satellite imagery for the coastal areas of the contiguous United States (CONUS). These datasets provide annual averages of 1) developed, 2) vegetated, 3) unvegetated ratios and 4) an unvegetated to vegetated ratio (UVVR) at 30-meter resolution over the coastal areas of the contiguous United States for the years 2014-2018. Additionally, multi-year average values of vegetated ratio, its standard deviation and a UVVR based on the annually-averaged vegetated ratio are provided for the coastal wetlands of the contiguous United States based on the National Wetland Inventory delineation. These datasets are provided as objective and consistent means to help evaluating geomorphic status and vulnerability of coastal wetlands at a national scale. Specifically, the unvegetated to vegetated marsh ratio (UVVR) is useful for establishing vegetative cover status and for tracking changes in the status of salt marshes at the national scale annually. Author ORCIDs: Couvillion, Brady(0000-0001-5323-1687); Ganju, Neil K. (0000-0002-1096-0465); Defne, Zafer (0000-0003-4544-4310) Author ORCIDs: Couvillion, B.R.(0000-0001-5323-1687);Ganju, N.K.(0000-0002-1096-0465);Defne, Z.(0000-0003-4544-4310) 20140101 20141231 observed Complete As needed -82.35539 -66.01728 45.39367 25.20179 Alexandria Digital Library Feature Type Thesaurus wetlands Marine Realms Information Bank (MRIB) keywords wetlands None wetland change wetland loss wetland gain USGS Metadata Identifier USGS:5e4d5ed3e4b0ff554f6d1833 Common geographic areas Atlantic Coast Conterminous United States coastal wetlands Florida Georgia South Carolina North Carolina Delaware Maine Maryland New Jersey New York Virginia It is strongly recommended that this data is directly acquired from the distributor described above or from another United States Geological Survey (USGS) Biological Resources Division server and not indirectly through other sources which may have changed the data in some way. The distributor makes no claims as to the data's suitability for other purposes. These data are intended for annual, coarse-scale analyses. It may not be appropriate for analyses at finer spatial or temporal scales. Acknowledgement of the U.S. Geological Survey as a data source would be appreciated in products developed from these data, and such acknowledgment as is standard for citation and legal practices for data source is expected by users of these data. Sharing new data layers developed directly from these data would also be appreciated by USGS staff. Users are advised to read the data set's metadata thoroughly to understand appropriate use and data limitations. Users are advised to contact the author of this research with questions regarding its appropriate use. The distributor shall not be liable for improper or incorrect use of these data, based on the description of appropriate/inappropriate uses described in this metadata document. These data are not legal documents and are not to be used as such. US Geological Survey Brady Couvillion Research Geographer mailing and physical

700 Cajundome Blvd.

Lafayette LA 70506 USA 225-578-7484 couvillionb@usgs.gov Environment as of Metadata Creation: Microsoft Windows 7 Version 6.1 (Build 7601) Service Pack 1; Esri ArcGIS 10.5 (Build 6491) Service Pack N/A (Build N/A) To assess the accuracy of the dataset, Landsat derived unvegetated, vegetated, and water percent cover estimates for a 2014-2019 time-period were reclassified into 1% bins and a random stratified sample was taken from each bin from 1% to 99%. For the purposes of this analysis, 0% and 100% cover were excluded from this random sample, as we are most interested in the fractional estimates of cover. In general, 100% water (for example) is easy to quantify, and the inclusion of large areas of 100% water (for example) would bias the statistics and give the appearance that the fractional cover estimates are more accurate than they are in the fractional (or mixed) categories. This sample consisted of 500,000 random points which were intersected with the corresponding vegetated, unvegetated, or water percent derived from National Agricultural Inventory Program (NAIP) images composited over a 2014 to 2019 time-period. For the purposes of this accuracy assessment, the NAIP derived percent cover estimates were used as truth, and compared to the Landsat derived estimates. The root mean squared error (RMSE) and bias of the Landsat derived percent cover estimates were calculated. The root mean squared error (RMSE) of the percent unvegetated estimates was 12.95% with a bias of +0.81%. This indicates that overall, the Landsat derived percent unvegetated land overestimates unvegetated land compared to the NAIP analyses. It is important to note however that the bias and RMSE varies through the percent cover estimate range. A higher overestimate of percent unvegetated land was observed in the lower portions of the percent cover range (i.e., NAIP derived Percent Unvegetated Cover from 1%-30%). This is to say that Landsat derived percent vegetative cover overestimates substantially at the low end of the range. At the higher end of the range, (i.e. NAIP derived Percent Unvegetated Cover from 70%-99%), the Landsat derived percent vegetation and NAIP derived percent vegetation were in very close agreement. The root mean squared error (RMSE) of the percent vegetated estimates was 17.75% with a bias of +1.5%. This indicates that overall, the Landsat derived percent vegetated land overestimates vegetated land compared to the NAIP analyses. It is important to note however that the bias and RMSE varies through the percent cover estimate range. A higher overestimate of percent vegetated land was observed in the lower portions of the percent cover range (i.e., NAIP derived Percent Vegetated Cover from 0%-30%). This is to say that Landsat derived percent vegetative cover overestimates substantially at the low end of the range. At the higher end of the range, (i.e. NAIP derived Percent Vegetated Cover from 70%-100%), the Landsat derived percent vegetation an NAIP derived percent vegetation were in very close agreement. The root mean squared error (RMSE) of the percent water estimates was 16.76% with a bias of -0.87%. This indicates that overall, the Landsat derived percent water land underestimates water compared to the NAIP analyses. It is important to note however that the bias and RMSE varies through the percent cover estimate range. A higher underestimate of percent water land was observed in the higher portions of the percent cover range (i.e., NAIP derived Percent Water Cover from 70%-99%). At the lower end of the range, (i.e. NAIP derived Percent Water Cover from 1%-30%), the Landsat derived percent vegetation an NAIP derived percent vegetation were in very close agreement. It is important to note that the NAIP imagery is being designated as truth due to its superior spatial resolution; however, it also contains error. The inconsistency in spectral information contained within the NAIP imagery leads to errors in its ability to estimate percent cover, and as such the true accuracy of the Landsat derived percent cover estimates is probably greater than these results indicate. It is also important to note that the Landsat derived estimates are taking into account greater temporal frequency than the NAIP derived estimates. Finally, geo-rectification errors may lead to some differences in the ground area covered by the NAIP and Landsat derived percent cover estimates, which may lead to further discrepancy between the two datasets and explain some of the inaccuracy. UVVR data has gone through extensive QA/QC to ensure a reasonable level of accuracy. This is not to say, however, that inaccuracy does not remain in the dataset, and users of the dataset should be aware of said potential for inaccuracy. Data set is considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. Refer to details regarding source imagery horizontal accuracy. A formal accuracy assessment of the vertical positional information in the data set has either not been conducted, or is not applicable. U.S. Geological Survey 2019 Version 2.0 publication Sioux Falls, South Dakota USGS EROS Digital and/or Hardcopy 2019 observed Landsat 8 (LASRC) input data Collin Homer Jon Dewitz Suming Jin George Xian Catherine Costello Patrick Danielson Leila Gass Michelle Funk James Wickham Stephen Stehman Roger Auch Kurt Riitters 202004 publication ISPRS Journal of Photogrammetry and Remote Sensing vol. 162 n/a Elsevier BV ppg. 184-199 https://doi.org/10.1016/j.isprsjprs.2020.02.019 Digital and/or Hardcopy 2001 2016 observed NLCD input data NOAA 2020 raster digital data Charleston, SC: NOAA Office for Coastal Management National Oceanic and Atmospheric Administration, Office for Coastal Management. 2016. www.coast.noaa.gov/htdata/raster1/landcover/bulkdownload/30m_lc/ Digital and/or Hardcopy 1975 2016 ground condition CCAP input data Datasets were created using Landsat 8 Operational Land Imager (OLI) Surface Reflectance imagery. Landsat was chosen as the rich historical record of imagery provides for the ability to create future datasets covering a time period back to 1984. Surface Reflectance data are generated from the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) and Landsat 8 Surface Reflectance Code (LASRC). Additionally, the LEDAPS methodology provides masks for clouds, cloud shadows, and adjacent clouds, which were used in the next step of the methodology. These images contain atmospherically corrected surface reflectance data from Landsat 8. The images consist of 5 visible and near-infrared (VNIR) bands and 2 short-wave infrared (SWIR) bands. The thermal bands were not used for the purposes of this effort. Landsat 8 (LASRC) 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 United States 225-578-7484 337-266-8616 couvillionb@usgs.gov Cloud recognition and exclusion was conducted using the “pixel_qa” bands. Values representing clouds or possible clouds as represented in Tables 6-2 and 6-3 of the following document: U.S. Geological Survey. 2019. Landsat 8 Surface Reflectance Code (LASRC) Product Guide, Version 2.0. EROS. Sioux Falls, South Dakota Using these values, the collection of all Landsat images was first filtered to include only pixels which were free of clouds or other sources of contamination. Landsat 8 (LASRC) 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 United States 225-578-7484 337-266-8616 couvillionb@usgs.gov The following indices were calculated for each image in the collection: mNDWI = (Green-SWIR)/(Green+SWIR) NDVI = (NIR-Red)/(NIR+Red) NDBI = (SWIR – NIR) / (SWIR + NIR) 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 United States 225-578-7484 337-266-8616 couvillionb@usgs.gov While cloud recognition and exclusion were implemented in a previous step to exclude sources of contamination, occasionally contaminated pixels are incorrectly identified. To further reduce the impact of these erroneous values, as well as to isolate a pixel value which represented somewhat normal conditions for the year being analyzed, a yearly median value was calculated for each pixel. Median is used rather than mean as the mean would still be adversely affected by contaminated pixels. 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 United States 225-578-7484 337-266-8616 couvillionb@usgs.gov High-resolution analyses were used to develop endmembers (to be used in Linear Spectral Unmixing) in Landsat 8 data which represented unvegetated, vegetated, and water conditions. These aerial imagery-based products were aggregated from 1-meter resolution to 30-meter resolution to match the Landsat pixels and the percent unvegetated/vegetated/water in each 30-meter pixel was recorded. Mean and standard deviation values of the NDVI, mNDWI, and NDBI were recorded in 1% intervals from 1% to 99% of each category. Endmember values were calculated from the line which best fit the calibration data. Initially, spectral unmixing consisted of separating land and water categories. The values of these lines at 0% land (100% water) and 100% land (0% water) were 0.1538883359 and -0.2729483920, respectively. Following unmixing into land and water categories, a second iteration of Linear Spectral Unmixing was used to quantify the unvegetated and vegetated components of Landsat pixels. For this process, NDVI, and NDBI indices were used. The values of the endmembers for those indices respectively were as follows: 0% vegetated land (100% unvegetated) 0.0712670576 and -0.04015064775 respectively, and 100% vegetated land were 0.5574777074 and -0.15864157625, respectively. 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 United States 225-578-7484 337-266-8616 couvillionb@usgs.gov Targets containing aquatic vegetation (AV) have strong vegetation signals, a characteristic usually indicative of vegetated land. Aquatic vegetation is often times ephemeral however, and it was the desire of this assessment to quantify vegetated and unvegetated land categories, and the change to those parameters in those land locations. Unless FAV is identified correctly and recoded to water, a transient vegetation signal can be misinterpreted as change to the UVVR. Aquatic vegetation has a strong vegetation signal, at least some water signal, and more importantly, those signals are variable. We therefore created an aquatic vegetation mask created by querying pixels that contained a variable NDVI signal as well as a variable mDNWI signal through the period of record. The resulting mask was then used in conjunction with a spectral signature of aquatic vegetation and Linear Spectral Unmixing was used to estimate the portion of each pixel comprised by aquatic vegetation. These values were recoded to water for the purposes of UVVR calculations. 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 US 225-578-7484 337-266-8616 couvillionb@usgs.gov Once fractional estimates of the percent unvegetated, vegetated, and water components of each pixel were calculated, the final UVVR was calculated. As the name suggests, this ratio is calculated as UV/Veg. 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 US 225-578-7484 337-266-8616 couvillionb@usgs.gov The final UVVR dataset was subset to a boundary extending from approximately a 10-meter elevation contour inland, to a seaward boundary of the federal waters boundary, and further subset to include only the Atlantic Coast of the conterminous United States. 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 US 225-578-7484 337-266-8616 couvillionb@usgs.gov As the purpose of this dataset is to provide a UVVR for coastal wetlands, ancillary datasets including NLCD and CCAP were used to identify wetland areas. Final datasets were masked to include only pixels identified as wetland in one of these two classifications. NLCD CCAP 2020 Brady Couvillion U.S. Geological Survey, Southeast Region Geographer mailing address

Louisiana State University

Baton Rouge LA 70803 US 225-578-7484 337-266-8616 couvillionb@usgs.gov Raster Grid Cell 74925 60625 4 0.00026949458523585674 0.0002694945852358567 Decimal seconds WGS_1984 WGS 84 6378137.0 298.257223563 L8_2014_ATL_Unveg_Veg_Water_UVVR.tif 4 band raster geospatial data file. Producer Defined Band 4 UVVR Producer Defined This ratio is calculated as UV/Veg L8_2014_ATL_Unveg_Veg_Water_UVVR.tif 4 band raster geospatial data file. Producer Defined Band 3 Water Fraction Producer Defined 0.0 1 Fraction L8_2014_ATL_Unveg_Veg_Water_UVVR.tif 4 band raster geospatial data file. Producer Defined Band 2 Vegetated fraction Producer Defined 0.0 1 Fraction L8_2014_ATL_Unveg_Veg_Water_UVVR.tif 4 band raster geospatial data file. Producer Defined Band 1 Unvegetated fraction Producer Defined 0.0 1 Fraction The entity and attribute information provided here describes the tabular data associated with the data set. Please review the detailed descriptions that are provided (the individual attribute descriptions) for information on the values that appear as fields/table entries of the data set. The entity and attribute information was generated by the individual and/or agency identified as the originator of the data set. Please review the rest of the metadata record for additional details and information. U.S. Geological Survey - ScienceBase mailing and physical

Denver Federal Center, Building 810, Mail Stop 302

Denver CO 80225 USA 1-888-275-8747 sciencebase@usgs.gov Distributor assumes no liability for misuse of data. Digital Data https://doi.org/10.5066/P97DQXZP None. No fees are applicable for obtaining the data set. 20210205 Brady Couvillion US Geological Survey Research Geographer mailing and physical

700 Cajundome Blvd.

Lafayette LA 70506 USA 225-578-7484 couvillionb@usgs.gov FGDC Content Standard for Digital Geospatial Metadata FGDC-STD-001-1998 Record created using USGS Metadata Wizard tool. (https://github.com/usgs/fort-pymdwizard)