Chirp seismic reflection data- shotpoints, tracklines, profile images, and SEG-Y traces for EdgeTech 3400 chirp data collected during USGS field activity 2022-001-FA (point and polyline shapefiles, CSV text, PNG Images, and SEGY data, GCS WGS 84)

PROCESS STEP 1: RAW TRACE IMPORT, NAVIGATION EXTRACTION, SEAFLOOR PICKING
jsf2segy (version 5) and Shearwater Reveal (version 2021) were used to process JSF data and extract navigation data. The processing flows are summarized in the following steps.
(1.) ‘jsf2segy’ was used to create segy files containing the real chirp traces from the JSF files. Option '-x' extracted real values from Analytic subbottom data, '-o' allowed writing real traces to an output .sgy file, and '-c' allowed writing a comma separated values (csv) file containing a JSF header values for ffid, year, Julian day, UTC time and towfish heave, pitch, roll, and pressure for use in additional processing.
(2.) The Reveal flow ImportSegy consisted of the following steps for each seismic file. Input imported the raw real segy files. HeaderMath unscaled and converted navigation data from seconds of arc to decimal degrees (WGS 84), converted towfish depth in meters to milliseconds and heave in seconds to milliseconds, and calculated the linear offset in meters between the towfish and the DGPS antenna via Pythagorean geometry, where the layback astern from the davit block is given by the square root of ((meters of cable out from block)^2 - (meters towfish depth + meters block to water line)^2) and total offset is obtained by adding meters offset from GPS antenna to block. Header Filtering smoothed the towfish depth time series using a 151 trace boxcar filter. ApplyStatic shifted the traces by the smoothed towfish depth and heave static values. ObspyPicking used the threshold algorithm to predict the seafloor time and Header2Picks wrote the times to an output csv file for editing. DBWrite created an output 'nrp-nav' csv file containing shot, navigational reference point GPS coordinates, UTC day/time, and the calculated GPS to towfish offset data. Output wrote the traces to .seis (Reveal’s format) files for additional processing.
These process steps and all subsequent process steps were conducted by the same person - Wayne Baldwin.

PROCESS STEP 2: NAVIGATION PROCESSING AND FEATURE CLASS GENERATION.
(1.) The output 'nrp-nav' csv files for each profile line were inspected for erroneous or absent coordinate data. Profile lines 20220606_170908, 20220610_143851, 20220610_153924, and 20220610_155751 each contained sections of missing navigation due to interruption of the GPS feed to Discover. A python Jupyter notebook utilized the python modules pandas (version 1.5.2), geopandas (version 0.12.1), pyproj (version 3.4.0), numpy (version 1.23.5), shapely (version 1.8.5.post1), and sqlite3 (version 3.40.0) to process the navigation data for each profile line in the following steps. The 'nrp-nav' csv file was imported to a pandas dataframe. Appropriate time periods were extracted from Hypack raw files and merged by UTC time into the pandas dataframe of files with missing navigation sections. In these instances, shot numbers were interpolated linearly between the existing records and the calculated GPS to towfish offset value for the record prior to the gap was repeated over its extent. Navigation coordinates repeated over many traces because the DGPS sampling frequency was 1-Hz, while the chirp shot rate was 4-Hz. To account for this, duplicate latitude and longitude coordinate pairs in the dataframe were replaced by NAN values, while maintaining the first occurrence of a unique pair, as well as the last unique pair. The NAN values were replaced by coordinate pairs calculated through linear interpolation between the remaining unique pairs, resulting in unique navigation throughout. The python function ShotlineLayback (developed by Nathan Miller of USGS-WHCMSC) computed layback navigation coordinates for the 3400 towfish. The algorithm used pyproj to transform the nrp coordinates from geographic (decimal degrees) to WGS84 UTM Zone 19 (meters), interpolated a sail line from the nrp coordinates along with a cumulative distance value, then computed layback positions for the 3400 shots by translating the nrp coordinates back along the sail line by the associated GPS to towfish offset value, and finally transformed the calculated layback UTM 19 coordinates back to geographic and wrote both set of coordinates to new data frame columns. A subset of the resulting dataframe consisting of values for layback easting and northing, layback latitude and longitude, profile line name, shot, UTC date/time, nrp longitude and latitude, as well as new fields populated for profile image name, survey ID, vehicle ID, and device ID was generated. The subset dataframe was written to a new 'layback-nav' csv file to be used in additional trace processing, then Geopandas and shapely were used to create a point geometry column from the layback geographic coordinates. SQL queries were used to load the navigation dataframe into a geospatial SQLite database using the pyspatialite interface. Records were appended to three database tables, the first containing the unique shot navigation, the second maintaining the first and last shots and even 500 shot intervals (which correspond to the annotation interval provided along the top of the seismic-reflection profile images), and the third containing trackline features created from the unique layback navigation point geometries.
(2.) The 500 shot and trackline tables were imported into QGIS (version 3.26) from the SQLite database, then exported (Right click on database feature class > Data > Save Features As) to the new Esri point and polyline shapefiles '2022-001-FA_ET3400_sht500.shp' and '2022-001-FA_ET3400_Tracklines.shp', respectively.

PROCESS STEP 3: SEAFLOOR PICK EDITS, LAYBACK NAVIGATION MERGE, FINAL TRACE PROCESSING
(1.) Using the Reveal 2D profile viewer, the seafloor predictions generated in step one were overlain on their associated seismic traces for quality control. The pick editor was used to eliminate or appropriately adjust erroneous times and edits were saved.
(2.) The Reveal flow GainSegy consisted of the following steps for each seismic file. Input imported the raw real .seis file. DBMerge imported layback navigation from the 'layback-nav' csv files, using shot as the merge field to populate new headers for layback latitude and longitude. Table2Header imported the edited seafloor picks replacing unedited header values. Despike was applied to remove noise bursts in trace samples caused by sea state or acoustic interference. TraceMath applied a time varying gain to trace samples after the seafloor time (trace sample = trace sample * (trace sample time * 1000) ^ 0.7), and muted trace samples before the seafloor time. Output wrote the processed real traces to seis and segy files. Complex Trace Analysis (CTAN) computed envelope traces from the real, and finally Output wrote the processed envelope traces to seis and segy file.
The Seismic Unix script Plot512i created variable density postscript plots of the seismic profiles showing two-way travel time (seconds) along the y-axis (left margin) and shots along profile (labeled at 500 shot intervals) on the x-axis (along top of profile). The postscript images were converted to 200 dpi portable network graphic (PNG) images using ImageMagick (version 6.9.9-40). The plots are labeled along the bottom of the X axis.

PROCESS STEP 4: DIAGNOSTIC SEISMIC PROFILE PLOTTING
The Seismic Unix script PlotChirp created variable density postscript plots of the seismic profiles showing two-way travel time (seconds) along the y-axis (left margin) and shots along profile (labeled at 500 shot intervals) on the x-axis (along top of profile). The postscript images were converted to 300 dpi portable network graphic (PNG) images using ImageMagick (version 6.9.11-60). The plots are labeled along the bottom of the X axis.