Single-Beam Bathymetry Data Collected in 2022 from Point Aux Chenes Bay, Mississippi

GNSS Data Acquisition: Two Global Navigational Satellite Systems (GNSS) base stations were established on NGS benchmarks in which the roving vessels would be within 10-15 kilometers (km) of the base station at any given time. The primary benchmark B166 was located to the north-east of the survey extent at the entrance to Bayou Heron Boat Ramp. The secondary base station NER1 was also located to the north-east on the Grand Bay National Estuarine Research Reserve. The base stations were equipped with Spectra Precision SP90M GNSS receivers recording full-carrier-phase positioning signals from satellites via a Trimble Zephyr 3 Base GNSS antennas recording at a rate of 0.1 second (s). The roving vessels were each outfitted with Spectra Geospatial SP90M GNSS receivers and GNSS antennae recording data at the same rate. The GNSS data files were converted from Spectra Precision proprietary G-file formats (for example, GB166A22.158, GWVR1A22.158, GWVR2A22.158) into the non-proprietary RINEX raw data file format (version 2.11) using the RINEX Converter program. The base and rover RINEX files were post processed together in a subsequent step to obtain differentially corrected navigation data and is described later in the metadata. Example RINEX filename WVR11580.22O, where 'WVR1' is the four-letter site name, '158' is the day of year, '0' is the session 0-#, '17' is the year, 'O' is the observation file, 'N' is the navigation file, 'M' is meteor logical file, and 'G' is the GLONASS navigation message file. The GNSS data acquisition was completed for both survey months (June and November) using identical equipment.

Single-Beam Bathymetry Acquisition: The single-beam bathymetric data collected under the USGS FAN 2022-320-FA, include two separate survey platforms/subFANs. The first survey occurred in June 2022, collected by the R/V SHARK (WVR1 - subFAN 22CCT09), and the R/V CHUM (WVR2 - subFAN 22CCT10), both 12-ft Yamaha personal watercraft (PWC) which collected 242.44 line-km and 216.39 line-km, respectively. In November 2022, WVR1 (subFAN 22CCT11) collected 68 line-km (20 lines) and the WVR2 (subFAN 22CCT12) collected 68 line-km. Boat motion was recorded at 50-millisecond (ms) using a SBG Ellipse A motion sensor. HYPACK A Xylem Brand (version 21.0.2.0, only subFAN 22CCT09 used version 21.3.1.0), a marine surveying, positioning, and navigation software package, managed the planned-transect information and provided real-time navigation, steering, correction, data quality, and instrumentation-status to the boat operator. Depth soundings were recorded at 50-ms intervals using an Odom echotrac CV100 echosounder with a 4-degree, 200-kilohertz (kHz) transducer. Data from the GNSS receiver, motion sensor, and echosounder were recorded in real-time aboard all vessels independently and merged into a single raw data file (.RAW) in HYPACK, with each device string referenced by a device identification code and time stamped to Coordinated Universal Time (UTC).
Sound velocity profile (SVP) measurements were collected using SonTek Castaway Conductivity, Temperature, and Depth (CTD) instruments. The instruments were periodically cast overboard to record changes in water column speed of sound (SOS). In the June 2022 survey, 78 successful sound velocity casts were collected and ranged in depth from 0.15 to 3.79 m (average 3.70 m), and in sound velocity from 1522.17 to 1544.71 meters per second (m/s) (average 1532.73 m/s). In the November 2022 survey, 9 successful sound velocity casts were collected and ranged in depth from 0.15 to 2.21 m (average 0.77 m), and in sound velocity from 1524.21 to 1524.36 m/s (average 1523.03 m/s)

Differentially Corrected Navigation Processing : The NGS Permanent Identifier (PID) DO5987, stamping 0166 B 2010 (https://www.ngs.noaa.gov/OPUS/getDatasheet.jsp?PID=DO5987), was the primary base station location for this dataset. For this dataset remain consistent with the 2015 and 2021 bathymetry datasets (DeWitt and others, 2017; Stalk and others, 2021), the same base location (DO5987) and base value were used for data collection and post-processing. However, to be current with the WGS84 (G2139) realization during acquisition, the base position used to post-process navigation data during the 2015 survey (DeWitt and others, 2017) was converted from ITRF2008 (equivalent to WGS84) into WGS84 (G2139) using the Horizontal Time-Dependent Positioning (HTDP) online transformation utility. The latest realization (G2139) was introduced on 01/23/2021. The survey acquisition dates succeed this date; therefore, it is the appropriate datum realization for post-processing the navigation data. The resultant base position used for post-processing was 30°24’46.75374 North, 88°24’12.11420 West (WGS84-G2139), and -29.032 meters (m) ellipsoid height (WGS84-G2139). A secondary base station was erected on PID DP1533, stamping 5 T 2 2013, but it was not needed for post-processing. The kinematic trajectories (rover to base) were processed using Novatel’s Waypoint GrafNav software.
The rover RINEX and the base RINEX files were imported into GrafNav, converted into proprietary files, and the kinematic (rover) GNSS data session from the survey vessel was post-processed to the concurrent base GNSS data session. Analyzing the data plots, trajectory maps, processing logs, satellite health plots, and other viewing utilities that GrafNav produces, provided measures to attain trajectory solutions (between the rover and the base) free of erroneous data that resulted in fixed positions. Some analytic examples include 1) excluding satellites flagged by the program as having bad/poor health or cycle slips, 2) Excluding poor satellite time segments that have a negative influence toward a fixed solution, and 3) adjusting the satellite elevation mask angle to improve the position solutions. This process was repeated for every GNSS data session. The final differentially corrected, precise DGPS positions were computed at 0.1 s and exported in American Standard Code for Information Interchange (ASCII) text format (.txt) in WGS84 (G2139) UTM 16N geodetic datum. One text file is produced for each roving GNSS session using the example file naming convention: 22CCT09_158_2022_B166_TWVR1_A.txt where the convention '22CCT09' is the subFAN, '158' is day of year, '2022' is the year, 'B166' is the base station used, 'WVR1' is the rover, and 'A' is the GPS session from A to Z.
During the November survey on WVR2 (22CCT12), the GNSS unit stopped recording internally so that data was not available to post-process from the rover to the base station. GPS was recorded to the HYPACK .RAW files, but only as Satellite Based Augmentation System (SBAS) which does not produce centimeter accuracy. The lines affected were not included in this data release. Please refer to the Quality Control, Quality Assurance (QA/QC) and Uncertainty Analysis process steps in this metadata record for more information.

Bathymetry Processing: All data were processed using CARIS HIPS and SIPS version 11.4.14. First, a vessel file *.vhf unique to the platform, was created that contains the surveyed offset measurements between the sensors. A CARIS project was created using the *.vhf file, and then the HYPACK *.RAW and the SVP profiles *.SVP were loaded. Next, the differentially corrected navigation files were imported using the generic data parser tool. Then, the differently corrected navigation files overwrote every non-differential position in the HYPACK *.RAW file based upon time. The bathymetric data components (position, motion, depth, and SOS) were then georeferenced and geometrically corrected in CARIS to produce processed xyz positional data. The data points were further analyzed using CARIS single-beam editor and subset editor tools providing 2-dimensional (2D) and 3-dimensional (3D) data views respectively. At this point, these tools help highlight any remaining echosounder artifacts or issues commonly associated with shallow water single beam surveys. Some examples include the following 1.) erroneous or false values, 2) patchy, inconsistent depth soundings usually a result of a vessels limitations in extremely shallow water, 3) the seafloor sediment composition can absorb or deflect sound, 4) the transducer measurement threshold, and 5) instrument malfunction. Outlier data was either filtered out or manually removed. Any questionable areas were assessed for repair against their surrounding data and ultimately removed from the dataset if a solution was not found. Once editing was considered complete, the point data were then exported as an xyz ASCII text file referenced to WGS84 (G2139) UTM 16N (DeWitt and others, 2017; Hansen and others, 2017).

Quality Control, Quality Assurance (QA/QC) and Uncertainty Analysis: The single-beam data exported from CARIS HIPS and SIPS (two vessel/subFAN xyz ASCII text files for both surveys) were transformed in ArcMap to a point shapefile (.shp) utilizing the "Create Feature Class From XY Table" geoprocessing tool. The projection was set to WGS84 UTM 16N. The generated shapefile was visually scanned for any obvious outliers or problems. Next, a polyline shapefile (representing tracklines) was produced from the point shapefile using XTools Pro "Make Polylines from Points" geoprocessing tool for each survey platform (subFANs 22CCT09, 22CCT10, 22CCT11 and 22CCT12). Utilizing both the xyz (point) and trackline (polyline) shapefiles, a Python script was used to evaluate elevation differences at the intersection of crossing tracklines by calculating the elevation difference between points at each intersection using an inverse distance weighting equation with a search radius of 3 m (DeWitt and others, 2017; Hansen and others, 2017). After each vessel was reviewed individually and initial crossing values assessed, the shapefiles were merged by geographic area and time. June subFANs 22CCT09 and 22CCT10 were merged to create the following point shapefiles: Point_Aux_Chenes_June_2022_SBES_WGS84_UTM16N_xyz.shp, Point_Aux_Chenes_North_Reef_June_2022_SBES_WGS84_UTM16N_xyz.shp, and Point_Aux_Chenes_South_Reef_June_2022_SBES_WGS84_UTM16N_xyz.shp. November subFANs 22CCT11 and 22CCT12 data were merged together to create Point_Aux_Chenes_North_Reef_November_2022_SBES_WGS84_UTM16N_xyz.shp, Point_Aux_Chenes_North_Reef_November_2022_SBES_WGS84_UTM16N_xyz.shp, and Point_Aux_Chenes_South_Reef_November_2022_SBES_WGS84_UTM16N_xyz_pp.shp. These point shapefiles are intermediate datasets (not published in this release) used to produce the trackline shapefiles. These point shapefiles were later re-converted to an ASCII text file (.txt) for inclusion in this data release, refer to the additional process steps below for more information. The trackline shapefiles were produced from these point shapefiles as previously described, and the Python script was applied again an inverse distance weighting equation with a search radius of 3 m.
The maximum crossing error for Point_Aux_Chenes_June_2022_SBES_WGS84_UTM16N_xyz.shp was 28 cm between the WVR1 and WVR2 files. The final crossing file had 1847 crossings with a range of 0.00 to 28 cm and standard deviation 0.05 cm, in which 72.00% were 10 cm or less; 26.08% were between 10-20 cm; 1.92% were between 20-30 cm.
The maximum crossing error for Point_Aux_Chenes_North_Reef_June_2022_SBES_WGS84_UTM16N_xyz.shp was 16 cm between the WVR1 and WVR2 files. The final crossing file had 399 crossings with a range of 0.00 to 16 cm and standard deviation 0.04 cm, in which 94.74% were 10 cm or less; 5.26% were between 10-20 cm. Two tracklines could not be merged with the corresponding differential navigation data during post-processing due to a processing bug. This resulted in an offset from the surrounding lines. To correct the data, the ellipsoid value of the line was adjusted by the average value of all the crossings that went through that particular line. Line 22CCT09_174_0031_1700 was adjusted by -0.24 m and 22CCT09_174_0032_1702 by -0.41 m. This adjustment is represented in the Point_Aux_Chenes_June_2022_SBES_WGS84_UTM16N_xyz.txt file as well.
The maximum crossing error for Point_Aux_Chenes_South_Reef_June_2022_SBES_WGS84_UTM16N_xyz.shp was 17 cm between the WVR1 and WVR2 files. The final crossing file had 316 crossings with a range of 0.00 to 17 cm and standard deviation 0.02 cm in which 99.05% were 10 cm or less; 26.38% were between 10-20 cm.
The maximum crossing error for Point_Aux_Chenes_North_Reef_November_2022_SBES_WGS84_UTM16N_xyz.shp was 28 cm between the WVR1 and WVR2 files. The final crossing file had 1847 crossings with a range of 0.00 to 28 cm and standard deviation 0.05 cm in which 72.00% were 10 cm or less; 26.08% were between 10-20 cm.
The maximum crossing error for Point_Aux_Chenes_South_Reef_November_2022_SBES_WGS84_UTM16N_xyz_pp.shp was 14 cm between the WVR1 and WVR2 files. The final crossing file had 274 crossings with a range of 0.00 to 14 cm and standard deviation 0.01 cm in which 99.64% were 10 cm or less; 0.36% were between 10-20 cm. All the tracklines for WVR2 (22CCT12) in this area did not have post-processed GPS, as explained in the Differentially Corrected Navigation Processing process step. The WVR1 (22CCT11) survey did have post-processed GPS data and was used as the metric to adjust the WVR2 data to correct the WVR2 lines that crossed WVR1. The following tracklines were adjusted based upon the mean of the crossings the trackline had with WVR1. The tracklines and the applied adjustment are 22CCT12_WVR2_0001_1746 by -0.21 cm, 22CCT12_WVR2_0002_1749 by -0.25 cm, 22CCT12_WVR2_0003_1751 and 22CCT12_WVR2_0004_1754 by -0.13 cm, 22CCT12_WVR2_0005_1757 by -0.19 cm, 22CCT12_WVR2_0006_1759 by -0.28 cm, 22CCT12_WVR2_0007_1802 by -0.27 cm, 22CCT12_WVR2_0008_1805 by 0.19 cm, 22CCT12_WVR2_0009_1808 by 0.10 cm, 22CCT12_WVR2_0010_1811 by 0.02 cm.
These point shapefiles were then exported from ArcMap as xyz ASCII text (.txt) files using the XTools Pro "Table to Text" function. The resultant ASCII text file can be downloaded from this data release. Point_Aux_Chenes_2022_SBES_xyz.zip. The trackline shapefiles were exported from ArcMap and are included in Point_Aux_Chenes_2022_SBES_tracklines.zip.

Datum Transformation: NOAA/NGS's VDatum was used to transform the single-beam data points' horizontal and vertical datums (xyz) with a reported vertical uncertainty of 0.06429 m. VDatum reports a nationwide standard deviation of 2.0 centimeters (cm) for ellipsoid - NAD83 transformations, and a nationwide standard deviation of 5.0 cm for NAD83 to NAVD88 transformations in the coastal regions of the continental U.S. For more information about the positional accuracy for these datum transformations, visit the Estimation of Vertical Uncertainties VDatum webpage (https://vdatum.noaa.gov/docs/est_uncertainties.html). The resultant ASCII text data files can be downloaded from this data release in Point_Aux_Chenes_2022_SBES_xyz.zip.

WorldView (WV) 2 imagery from 20220511 was downloaded from Maxar’s Global Enhance Geoint Delivery (G-EGD). The imagery was radiometrically and atmospherically corrected and then pansharpened using ERDAS IMAGINE 2020 to obtain measures of ground reflectance. The output image was reprojected into the WGS84 UTM 16N coordinate system. The image was then co-registered to high-resolution aerial imagery (NAIP) using the AutoSync-workstation in ERDAS Image pixels. First, an NAIP image mosaic was created for an extent larger than the WV image coverage. Ground control points coincident on both images were used to adjust the WV image to the corresponding location on the NAIP. Control points with an error value greater than 1 were removed to generate vector shorelines from WV images, and a similar methodology described by Maglione (2014) was followed. All WV images were classified into land-water rasters using tools within Arc Map. First, normalized difference vegetation index (NDVI) was calculated using WV band 5 in the visible red spectrum (RED) and band 7 in the near-infrared spectrum (NIR1) using the following formula: NDVI=(NIR1-RED)/(NIR1+RED). Vegetation was classified as high values (above 0.2), water represented low values (usually less than -0.2), and soil was classified between the values of -0.2 and 0.2. The value of 0.21 provided an adequate representation of the shoreline (wetland-water boundary) and that value was used to reclassify the NDVI to a land-water raster. The land-water raster was processed to reduce isolated pixels by using expand-shrink procedure (Smith and others, 2021). The resulting image classified vegetated shorelines, however some shorelines within the study area were composed of sandy beach (bright, sand and shells). To improve the shoreline classification for beach shorelines, a threshold using band 8 (NIR2) was selected, which showed high reflectance for the beach and urban areas (a value of 5000 was used) and created a new land classification that was merged with the vegetated land class to create a binomial land-water raster. In ArcMap, the classification boundaries were smoothed to remove any extraneous cells missed in the expand and shrink steps. The raster was then converted into polygons, and the polygons were then converted into lines. The lines, however, still contained the ridged cell edges, so the lines were smoothed using the Polynomial Approximation with Exponential Kernel (PAEK) algorithm and a 2-meter smoothing filter (Smith and others, 2021). Shorelines not in the study area or shorelines improperly classified (such as nearshore sand shoals) were manually removed or edited. Shorelines were projected into the NAD83 UTM 16N coordinate system.

Shoreline Preparation: To provide the most accurate representation of the bathymetry within the nearshore areas, shoreline data was prepared and used in conjunction with the xyz data that was produced during previous processing steps. Replicating the steps from Stalk and others (2021), the shoreline polyline data was first converted to points utilizing the XTools Function "Polylines to Points". Utilizing the "Buffer" geoprocessing tool in ArcMap, 100 m and 200 m buffers were generated around the Point Aux Chenes single-beam survey tracklines from June 2022. The shoreline points were then clipped using the "Clip" geoprocessing tool to each generated buffer file to ensure only shoreline points that fall within 100 m or 200 m from a bathymetry point were utilized for further processing. The 100 m buffer was utilized as the primary distance. However, on the westernmost extent of the Bay, 200 m was a more appropriate buffer distance due to the existing data coverage. Some manual editing of the resultant shoreline point file was necessary to ensure proper representation of the bathymetry extent. Utilizing the attribute table editor and Field Calculator function, an integer field was added to the shoreline points attribute table named "NAVD88_G12A", and the field was populated with a zero-value to represent a zero elevation at the shoreline extent. The resultant point shapefile has been made available in the Point_Aux_Chenes_2022_SBES_NAD83_NAVD88_GEOID12A_UTM16N_10m_DEM.zip. The shoreline points were then converted to a polygon utilizing the XTools Pro "Points to Polygons" function to establish an extent file to be used in later processing steps. Polygon vertices were manually added using the Editor tool to the western and eastern portions of Point Aux Chenes Bay, placed at the approximate land-water intersection using provided shoreline (Smith and others, 2021), as well as along the offshore extent of the xyz data. The resultant coverage file was utilized in subsequent processing steps and has also been made available in Point_Aux_Chenes_2022_SBES_NAD83_NAVD88_GEOID12A_UTM16N_10m_DEM.zip.

Create Digital Elevation Model: The transformed xyz ASCII text dataset from the June 2022 survey (generated in the previous processing step) was imported into ArcMap using the "Create Feature Class From XY Table" tool. That dataset, along with the generated shoreline points, were used to create a triangulated irregular network (TIN) using the "Create TIN" tool. The TIN was subsequently converted into a raster file using the "TIN to Raster" tool utilizing floating point with a cell size of 10 meters. The resultant raster was then clipped utilizing the "Extract by Mask" tool, utilizing the coverage file produced in the previous step. The final digital elevation model (DEM) was then exported from ArcMap as a 32-bit floating GeoTIFF (Geographic Tag Image File Format, .tif), utilizing the Natural Neighbor algorithm, and made available within the data downloads section of this data release (Point_Aux_Chenes_2022_SBES_NAD83_NAVD88_GEOID12A_UTM16N_10m_DEM.zip). As a part of standard QA/QC procedures, each DEM was compared to the associated xyz data, and associated root-mean-square error (RMSe) values were populated. The Point_Aux_Chenes_June_2022_SBES_NAD83_NAVD88_GEOID12A_UTM16N_10m_DEM has a RMSe value of 0.039 m.