Bathymetric survey
Near-shore bathymetric sounding points were created from high-resolution dGPS and single beam sonar. Surveys were conducted using a post-process kinematic (PPK) surveying method with a one second sampling rate for dGPS. Surveys were conducted by Craig Tweedie and Gabriela Tarin in July/August of 2015.
Bathymetric surveys were conducted using a Trimble 5700 dGPS roving unit, Trimble Zephyr antenna, Trimble TSC2 survey controller, Seafloor SonarMite single beam echo sounder, Dell Rugged laptop and inflatable Achilles boat. After launching the boat, surveyors began logging both dGPS and sonar data. A photo was taken that included the GPS time (time displayed on survey controller) and computer time (the time displayed on the computer logging sonar data). This photo would later be used to correct computer time to GPS time and allow for accurate time syncing of dGPS and sonar data. Survey points follow the boat tracks on any given survey. Boat surveying speed was ~6-7mph.
Processing and QAQC
Overview: Sounding points were collected and logged at 2hz with Google Terminal version 1.9b. Associated dGPS points at 1 second resolution were post processed in Trimble Business Center version 2.70, using UTM Zone 4 North coordinate system, NAD 1983 (Alaska) datum transformation and a Geoid99 (Alaska) Geoid model. Sonar data and dGPS points were then joined and QA/QCed in ArcGIS Desktop version 10.2.2. Depth correction was performed before time syncing sonar and dGPS data using the UNESCO equation in the calculator available here: http://keisan.casio.com/exec/system/1258122391. The sea floor depths, elevations and ellipsoid heights were collected and post processed at high resolution. Survey lines may, however, contain gaps in data where dGPS data logging stopped due to various reasons such as poor satellite coverage or power issues. The average horizontal precision for any one sounding point is 2.3cm. Vertical precision of depth values is a combination of the vertical precision of dGPS surveying (4.6cm on average for any one sounding point) plus 0.1% of the total depth (sonar manufacturers stated precision).
Processed dGPS surveys were brought into ArcMap 10.2.2 for manual QA/QC. Uncorrected survey points were removed. There were areas on a survey where the dGPS unit lost satellite coverage. Data point in these areas could also not be corrected and were cleaned and discarded. Other quality measures taken into account were survey quality (“Survey”, “Mapping” or “Unknown” from best to worst) and horizontal and vertical precessions. The bathymetric surveys here very rarely lost satellite coverage as there were no obstructions on the open water. Furthermore, dGPS data points are of high precision. Then, the photos displaying GPS time and computer time were used to determine the time offset between the two datasets. This offset was applied to the sonar dataset. The two data sets were joined via their respected time fields in ArcMap keeping only matching records (sonar was logged at 2hz while dGPS at 1 second).
To account for the effects of salinity and temperature on sonar velocity and thus depth, a series of calculations were made: calculate pulse time for each point with manufacturers stated sonar velocity; calculate new velocity; and calculate new depth with (new velocity * pulse time) / 2. New velocity was calculated here using the “velocity of sound in sea-water calculator” (http://keisan.casio.com/exec/system/1258122391) using the “UNESCO” formula. Salinity and temperature were logged aboard the survey boat at varying logging rates. Median values for each were used in the UNESCO formula to calculate new sonar velocity speed. Corrected depth values (NewDepth) were subtracted from Ellipsoid height and Elevations values output from Trimble Business Center to create ocean bottom ellipsoid height (BottomElli) and ocean bottom elevation (BottomElev).
The attribute table for this dataset includes: FID, an internal feature number, Shape, the feature geometry, Easting, the X coordinate in meters for the sounding point, Northing, the Y coordinate for the sounding point, Lat_G, the latitude for the sounding point, Lon_G, the longitude for the sounding point, H_Precsn, the horizontal precision, in meters, of the sounding point (generated from dGPS), V_Precsn, the vertical precision, in meters, of the sounding point (generated from dGPS), Local_Time, or “GPS time”, of each survey point that was used to sync sonar data to dGPS data, Date_, the date that the survey a particular sound point came from was conducted, NewDepth, the water depth at any given sounding after correction for salinity and temperature, BottomElev, elevation of the ocean bottom at any given sounding point, BottomElli, ellipsoid height of the ocean bottom at any given sounding point, Survey, an internal survey name pertaining to the location (“CH” for Chukchi Sea and “EL” for Elson Lagoon) and date of the particular survey, UID, a unique number given to each sounding point in the dataset in no particular order.
Surface interpolation
The 10 meter resolution interpolated surface was generated from bathymetric sounding points.Processed and finalized sounding points were brought into ArcMap 10.3.1 for subsetting. Of the original 231,266 sounding points, 1,638 were chosen to train the interpolated surface. Manual cleaning was done on points outside of the study area extent (in North Salt Lagoon). A polygon shapefile representing the interpolation processing extent was created using a 2014 digitized shoreline shapefile that includes the barrier islands and a boundary connecting the easternmost island back to the mainland shore. Another polygon shapefile was created to represent the shoreline, or zero line. Raster values beyond the boundary of this shoreline would be considered as “no data” in the interpolation process. This shapefile was similar to the extent shapefile, however, its boundaries extend beyond the extent in areas where lines were created to connect barrier islands to each other or back to the mainland.
The interpolation method used was the Topo to Raster tool within the 3D Analysis license in ArcMap 10.2.2. This method was chosen for its ability to utilize polygon features to delineate processing extents and shorelines (zero lines). Three shapefile were used as inputs for processing: the extent polygon for “Boundary”; the sounding point subset for “Point” with “Depth” chosen as field; the shoreline polygon for “Coast”; “Drainage enforcement” turned off; “Primary type of input data” as spot;
20 iterations; 0.5 for “Roughness penalty”; “Tolerance” at 200; 0.25 as the upper “Zlimit”; and 10m cell size. The final surface was then converted from GRID to TIFF format.
