Coastal Land-Cover Data Derived from Landsat Satellite Imagery, Delaware Bay, New Jersey to Shinnecock Bay, New York, 2008-2022

The regional AOI, which was derived from the National Oceanic and Atmospheric Administration (NOAA) US Coastal Zone Management Act Boundary, was subset into four scene-specific AOIs for processing and analysis: WRS-2 path 13 row 32 New York (nfwf_1332_ny), WRS-2 path 13 row 32 New Jersey (nfwf_1332_nj), WRS-2 path 13 row 33 (nfwf_1333), and WRS-2 path 14 row 33 (nfwf_1433).

For each image acquisition date, Collection-2, Level-2 surface reflectance (SR; reflective bands), surface temperature (ST; thermal infrared [TIR] bands), and SR-derived NDVI images were downloaded from the EROS ESPA On Demand Interface (https://espa.cr.usgs.gov/).

All images were batch-processed using Spatial Model Editor in ERDAS IMAGINE. The SR and ST bands were stacked to create 7- (TM) or 8- (OLI) band multispectral images and clipped to the scene-specific AOI. From these composite images, two additional spectral indices, mNDWI (Xu, 2006) and NBLI (Li and others, 2017), were calculated.

Land-cover classification was modified from the workflow described by Bernier and others (2021) using multilevel (2 thresholds per image) automatic thresholding of spectral indices and applying a rule-based classification scheme. First, Otsu's method (Otsu, 1979) for automatic histogram thresholding was applied to mNDWI images to create a binary "land"-water raster for each image acquisition date. Second, water area was masked from NBLI and NDVI images, and Otsu's method was applied to the masked images to create binary bare earth-"unclassed" and vegetated-"unclassed" rasters, respectively. The following rule-based classification was then applied, where T1 and T2 represent the first and second multi-Otsu thresholds, respectively, for each spectral index: If mNDWI > T2 then class (1) = water If NBLI > T2 then class (3) = bare earth (sand) If T1 < NBLI < T2 and UVVR = NoData then class (12) = unclassed If (T1 < NBLI < T2 and NDVI > T1 and UVVR > 0) or (NBLI < T1 and NDVI < T1) then class (4) = vegetated If NBLI < T1 and NDVI < T1 then class (11) = unclassed Where UVVR is the 2014-2018 composite Unvegetated to Vegetated Ratio for the U.S. Atlantic Coast (Couvillion and others, 2021) subset to the regional AOI. Finally, the binary land-cover images were converted to thematic rasters, merged, single-pixel "clumps" were removed using a 3x3 majority filter, and a standard colormap was applied to create a final land-cover raster dataset for each image acquisition date. The resulting land-cover rasters use the naming convention YYYYMMDD_lt05_AOI_lcr_ce.img (Landsat 5) or YYYYMMDD_lc08_AOI_lcr_ce.img (Landsat 8), where YYYYMMDD denotes the image-acquisition date (4-digit year, 2-digit month, 2-digit day) and AOI denotes the scene-specific AOI. All steps were batch-processed using the Image Processing toolbox in MATLAB (Otsu thresholding and binary image creation) or Spatial Model Editor in ERDAS IMAGINE (spectral index masking and generation of classed land-cover rasters).

For 35 datasets, thresholding resulted in misclassification of some nearshore water areas as "land" after mNDWI thresholding. Visual analysis showed that most misclassed open-water pixels correspond to areas of breaking waves, and misclassed extents were manually cleaned from the final datasets by selecting and re-classing the pixels in ERDAS IMAGINE. The cleaned land-cover rasters use the naming convention YYYYMMDD_lt05_AOI_lcr_ce_cl.img (Landsat 5) or YYYYMMDD_lc08_AOI_lcr_ce_cl.img (Landsat 8), where YYYYMMDD denotes the image-acquisition date and AOI denotes the scene-specific AOI.