Bathymetry grid and GPS data from Yellowfin UMS operation at Marconi Beach, CACO, Wellfleet, MA on February 26, 2025

The single-beam bathymetry data were collected by launching Yellowfin into the water from the surf zone. Yellowfin is equipped with a Cerulean s500 echosounder and collects singe-beam data and has an EMLID Reach M2 GNSs reciever for positioning data. The boat was controlled near the shore by Peter Traykovski and put on auto-pilot using predetermined tracklines for longer transects offshore. A base station was set-up in the Nauset Lighthouse parking lot to gather GNSS data. Step completed by Peter Traykovski with help from Athina Lange and Mark Carson.

The raw bathymetry data was processed with the following steps: 1.Parsed the echosounder data (NMEA GPS Puck, Time, and Echosounder range) into MATLAB (v. 2024a) using MATLAB Scripts developed by Peter Traykovski at WHOI - see contact below. 2. Post Processed Kinematic (PPK) GNSS data from Yellowfin was corrected using the base station (Java Triumph-1) data in Emlid Studio (Version 1.9) using the datum NAD83(2011) and then NAVD88 with GEOID 18 applied. A projection for UTM Zone 19N was also specified to get the Easting and Northing information. 3. Refined time alignment to account for any small-time delays between GPS data and echosounder data collection. This is accomplished using the time lagged cross-correlation of echosounder range and PPK GNSS altitude in locations with a relatively flat bottom. The resulting time shift was 18 seconds. The GPS measures vertical fluctuation of the boat due to waves and these fluctuations are also visible in echosounder data. Outliers in the echosounder data were removed using a despiking function. The equation in Step 5 removes the vertical fluctuation due to waves from the echosounder data leaving only true bathymetry if GPS and the echosounder are well synced. The lagged cross-correlation processing ensures they are synced optimally. 4. The PPK GNSS elevation (GPS_z) positions (collected at a higher frequency than the echosounder) were adjusted for the Antenna_Z_offset (0.15 m) and then interpolated to the time of echosounder samples (Echo_z) with the MATLAB function interp1 (linear interpolation). 5. Calculated sea-floor elevation with reference to the NAVD88 datum in meters using the following equation: Seafloor_z = Echo_z + GPS_z – Sonar_waterline_offset. These values data are available in 2025005FA_Marconi_Feb_Bathymetry_trackline.csv. 6. Gridded Seafloor_z values onto 1 m resolution UTM eastings, northings using MATLAB file exchange script regularizedata3d by Jamal (2020), a cubic interpolation with the optimum smoothness coefficient determined by a Monte Carlo optimization procedure for the data regularization. The function Regularizedata3d (Jamal, 2020) entailed removing one cross-shore line of data from the processing and adjusting the smoothness parameter, so the overall surface fit best fits the removed data. The process was then repeated for all the lines of data. The optimized smoothing parameter was used with all the data for the final processing. 7. The MATLAB function 'roipoly' returned the mask as a binary image, which sets pixels inside the region of interest (ROI) to 1 and pixels outside the ROI to 0. The boundary was developed so that there were no extrapolated bathymetry data outside the tracklines. The masks extent was the convex hull of the tracklines positions (x,y) and a bounding z coordinate of 0.3 represented the surface in reference to NAVD88 in meters. 8. Exported gridded 1 m data in NAD83(2011)/UTM Zone 19N in NAVD88 meters as a GeoTIFF: 2025005FA_Marconi_Feb_Yellowfin_Bathymetry_1m.tif. Note that using NAVD88 as the elevation datum, means that the depths are not true depths but the elevation of the seafloor above or below the datum and GEOID.
Jamal (2020). RegularizeData3D (https://www.mathworks.com/matlabcentral/fileexchange/46223-regularizedata3d), MATLAB Central File Exchange. Retrieved September 23, 2020. Step processed by P. Traykovski.