Ground Penetrating Radar and Global Positioning System Data Collected from Central Florida Gulf Coast Barrier Islands, Florida, February-March 2021

Navigation data acquisition - Position and elevation data for each GPR line were recorded at the time of collection using a Spectra Precision SP90M GPS receiver and geodetic antenna. GPS data were recorded concurrently throughout the survey using a SP90M receiver and Trimble Zephyr-3 Base antenna at base stations set up on local benchmarks at the Fort DeSoto flagpole (NGS identifier AG0484; CK, MK, and SJK lines) or along the Dunedin Causeway at Hog Island (NGS identifier AL0220; CI, HI, and AK lines). The rover unit was mounted 1.320 m above the GPR antenna and was connected to the GPR's digital control unit. Raw GPS positions were output to the GPR control unit in real time as a National Marine Electronics Association (NMEA) GGA string at 0.5-second (s) intervals using coordinated universal time (UTC) stamps and saved as a .PLT data file.

GPR data acquisition - A total of 120 GPR lines (lines 001-019 and 048-053 for SJK; 020-024 for CK; 025-047 for MK; 054-084 for HI; 085-106 for CI; and 107-120 for AK), representing a linear distance of approximately 40.5 km, were acquired during USGS FAN 2021-308-FA within developed and undeveloped beach, dune, and back-barrier environments at the CFCG study sites. Data were collected along established roads, beach accesses, and along hiking trails through back-barrier scrub/shrub vegetation. The GPR data were collected with a Geophysical Survey Systems, Inc. (GSSI) TerraSIRch SIR System-3000, which acquires single-channel GPR data and includes a Digital Control Unit (model DC-3000) and 200-Megahertz (MHz) antenna. A 16-inch survey wheel (GSSI model 620) was attached to the back of the antenna and calibrated prior to surveying to ensure an accurate revolution-of-wheel to distance ratio. Distance mode was used to record the raw GPR data and the raw GGA position string was recorded in real time via an Acumen SDR data logger. GPR traces were collected at a maximum rate of 64 scans per second with a vertical resolution of 1,024 samples per scan and an Infinite Impulse Response (IIR) filter was used (Lowpass = 600 MHz, Highpass = 50 MHz) to increase the signal-to-noise ratio of the recorded GPR data (Forde and others, 2018). Data were acquired using the SIR 3000 in TerraSIRch mode and saved in GSSI's proprietary Data Zero Time (DZT) format.

Navigation processing - Base station data were post-processed through the NGS Online Positioning User Service (OPUS, https://geodesy.noaa.gov/OPUS/). The time-weighted coordinates calculated from all base-station occupations at AG0484 and AL0220 were used for post-processing. The base station coordinates were imported into GrafNav, version 8.9 (NovAtel Waypoint Product Group), and the data from the rover GPS were post-processed to the concurrent base-station session data. The final, differentially corrected, GPS positions were computed at 0.5-s intervals for each rover GPS session. For lines 001-003, 007-008, and 014, the post-processed navigation data were exported in American Standard Code for Information Interchange (ASCII) text format as a NMEA GGA string, which replaced the uncorrected real-time rover positions recorded during acquisition. Along most GPR lines, however, there were significant dropouts (possibly attributed to atmospheric conditions and (or) the height of the back-barrier vegetation relative to the height of the rover antenna) that resulted in large gaps in the post-processed navigation data, especially in back-barrier environments (Forde and others, 2018). Along these lines, elevations at the raw rover x,y coordinates were interpolated from lidar data collected by the U.S. Army Corps of Engineers (USACE) in June 2015 (USACE JALBTCX, 2016) or, for SJK only, lidar data collected by the National Geodetic Survey (NGS) in January 2021 (NGS, 2022) using Esri ArcMap version 10.8.1. Visual and quantitative comparison of post-processed and lidar elevation profiles along each GPR line showed satisfactory agreement and acceptable 2-sigma values, please see the positional accuracy report for additional details. The elevation values in the raw navigation string were replaced with post-processed (beach and dune environments) or lidar (everywhere else) elevations; the NMEA checksum string was recalculated for each navigation string in Matlab version R2021A. Processed navigation data are provided in the North American Datum of 1983 (NAD83) (2011) coordinate system and processed orthometric elevations are reported relative to the North American Vertical Datum of 1988 (NAVD88), derived using the GEOID18 geodetic model. The final navigation files provided in the data release were saved as ASCII text format (.gga) data files.

GPR processing - Reflexw Version 7.2.2 (Sandmeier Scientific Software) geophysical near-surface processing and interpretation software was used to process the GPR data. GPR data were acquired in DZT format and later imported into Reflexw, where they were converted into a DAT file. For archival purposes, a non-proprietary version of the raw data was created by exporting the DAT file from Reflexw and saving it in ASCII 4-column format (.asc). The data were processed in a consistent order: (1) static correction was applied; (2) the mean value was subtracted (dewowed); (3) header gain applied during acquisition was removed; (4) manual Automatic Gain Control (AGC) gain was applied; and (5) post-processed navigation data were imported into the trace headers. Data were visually inspected after each step listed above and before elevation-corrected subsurface profiles were exported; all profiles were analyzed for errors or data gaps in the navigation and trace data to ensure data quality was maintained throughout. Hyperbola analyses were performed on all profiles to estimate the radar-wave velocities through the sediment. Calculated velocities at AK (N=13) ranged from 0.06 to 0.60 meters/nanosecond (m/ns) and, excluding three statistical outliers, averaged 0.10 m/ns. Calculated velocities at HI and CI (N=51) ranged from 0.05 to 0.30 m/ns. Velocities at HI-CI clustered into two groups (v < 0.10 m/ns and v >= 0.10 m/ns) with average values of 0.06 and 0.19 m/ns, respectively. Calculated velocities at CK, MK, and SJK (N=50) ranged from 0.05 to 0.20 m/ns. Velocities at CK-MK-SJK clustered into two groups (v <= 0.10 m/ns and v >= 0.15 m/ns) with average values of 0.07 and 0.20 m/ns, respectively. The processed profile data were re-imported into GSSI RADAN version 7.5.18 software and the surface normalization processing algorithm was applied, adjusting the profile to both the measured terrain and the site-specific radar-wave velocities. The elevation- and velocity-corrected profiles are provided as DZT-formatted files and were also exported as Joint Photographic Experts Group (JPEG) images. Some GPR lines were mostly or completely attenuated; for those lines elevation-corrected profile images are not provided in the data release. The processed trace data (.asc, .dzt, .dzx) for all 120 lines have been provided in the data release. Processed GPR profile trackline files were exported from RADAN 7 in Keyhole Markup Language (KML) format, projected to the Universal Transverse Mercator (UTM) 17 North (17N) coordinate system, and appended with profile-specific data acquisition and processing parameters in Esri ArcMap version 10.8.1.