Home | Contact ST  

Feature Articles

Freshwater Runoff Effects On Shallow-Water Multibeam Surveys
Using Multibeam Data Processing To Characterize Submarine Freshwater Springs

By Gabriella Di Martino
Renato Tonielli
Coastal Marine Environment Institute
Italian National Research Council
Naples, Italy

Incorrect sound-speed values have a significant impact on multibeam survey accuracy. When the sound velocity profile (SVP) is overestimated (called “smiling”) or undervalued (“frowning”), “ripples” typically appear in the logged data. In such cases, it is possible to recompute bathymetric data in post-processing by generating an equivalent SVP that can be applied to the raw data.

Over the last 10 years, the Coastal Marine Environment Institute of Italy’s National Research Council (IAMC-CNR) has developed national and regional research programs on morphobathymetric mapping.

In order to study the Campanian continental shelf from three to 200 meters of depth, the institute carried out the Regional Geological Cartography (CARG) project.

Multibeam data collected in shallow water during the CARG project often suffered from smiling or frowning because of freshwater runoff or upwelling. This paper describes the method used to correct these data and the use of these fixes for tracking unreported underwater springs.

In June 2006, high-resolution bathymetric data were collected in a depth range of 10 to 70 meters along the southern Campanian coasts, between Punta Licosa and Sapri, Italy. The survey was carried out from the RV Thetis, a 31-meter-long ship purchased by the CNR in 2001 and equipped for geophysical surveys.

The vessel was equipped with a Navtech (Springfield, Virginia) 12-channel Landstar differential global positioning system (GPS). A Trimble (Sunnyvale, California) GPS was also employed; it produced the same precision through the Landstar differential corrections.

The movements of the ship were recorded by the DMS02-05 motion sensor and by the Meridian Surveyor gyrocompass, both made by Teledyne TSS (Watford, England).

The bathymetric data were collected from a hull-mounted multibeam RESON (Slangerup, Denmark) SeaBat 8111R. The system operates at a frequency of 100 kilohertz with 101 horizontal beams centered 1.5° apart (150° across-track beam width and 1.5° along-track beam width). It provides a 150° swath coverage to a maximum depth of 1,000 meters. Phase values of central beams and amplitude values of external beams provide the bottom detection.

A RESON SVP-C velocity sound probe was installed near the transducers to provide real-time sound speed for beam-forming. RESON PDS2000 software was used for positioning and recording bathymetric data and for successive data processing.

During acquisition, the PDS2000 received data from the GPS and the motion sensor and applied the corrections to the bathymetric data. Sound velocity values from the SVP-C and the time string from the acquisition computer for data synchronization were transmitted to the central unit (CU). The CU transmits bathymetric data to the acquisition system through a local area network connection.

A sound velocity profiler, RESON SVP20, was lowered through the water column every six to eight hours. It provided a direct measurement of the sound velocity every 50 centimeters. PDS2000 uses the SVP to calculate depth.

Click To Enlarge.

Data Acquisition
The instrumental offsets and calibration values were computed in the shipyard when the Seabat 8111R was mounted on the vessel. At the beginning of the survey, a calibration was carried out to check the calibration values as required by International Hydrographic Organization (IHO) standards. Calibration was performed in Naples Bay, where a 20-meter high object was located at 70 meters’ depth.

The survey lines ran parallel to the coast, because the best definition is obtained when the slope is within the port or starboard swath coverage. The survey speed was 5.5 knots.

Since swath width is a function of water depth, the line spacing cannot be constant. High-resolution data were required for this survey, so the lines were planned to obtain 80 to 100 percent overlap between adjacent swaths. Low survey speed and high swath overlap allowed for a grid cell size of five by five meters during both data acquisition and data processing.

Some filters were applied in the PDS2000 acquisition window to remove spikes during data logging. The quality filter rejects the beams that do not meet the quality settings based on colinearity and brightness levels. The nadir filter rejects the beams outside a defined port or starboard angle and the nadir. The intersect filter checks the enclosed angle of three data points against the given minimum angle; a smaller angle than the intersection angle will reject the middle data point of the three data points. This filter can be used on flat seafloor.

Data logged in such a way meet Order 1b of the IHO standards for hydrographic surveys.

Data Processing
In order to produce fine resolution data, the PDS2000 Editing Module was used to process data from ship positioning, beam cleaning and filter proceeding.

The PositionEditView window displays the logged vessel position while the TimeBasedEditView displays some correlated parameters, such as number of satellites, horizontal dilution of precision values and GPS mode. Incorrect data can be deleted or moved in order to place each swath in the correct position.

The operator can delete or modify the acquisition filters or create new ones to recover missing data. It is also possible to change the SVP to be applied.

The MultibeamEditView, or Swath-Editor, displays a 2D and 3D view of the swaths. In this window the operator can manually remove bathymetric spikes.

During the final step of data processing for the CARG project, the research-ers generated a bathymetric grid in order to verify data quality.
Some ripples were evident on the draft grid that had one-meter amplitude and variable wavelength.

It was determined that these bottom shapes were not valid because of their features: The orientation was the same as the survey lines, and the wavelength was variable but always consistent with the swath coverage at different depths.

The undervalued sound velocity in the water column was considered to be the cause of the problem. The PDS2000 3D View confirmed that the swaths were affected by frowning. The problem was evident in the bathymetric range of 10 to 50 meters within an eight-square-kilometer area between Capo Palinuro and Marina di Camerota, Italy.

The survey in this area was carried out in few hours, so a single SVP logged at 70 meters’ depth 2.5 miles from the coast was used. It displays a 1,529 meters per second sound velocity from zero to five meters’ depth. A thermocline is evident between five and 35 meters’ depth: Along it, the sound velocity decreases to 1,508 meters per second. Finally, the SVP displays a rather constant value of 1,508 meters per second from 35 to 70 meters’ depth.

This sound velocity profile was not suitable for shallow-water data, even though it was consistent with documented Mediterranean Sea SVPs and thus produced the artifacts on the draft grid.

Numerous tests were carried out to modify the logged sound velocities in different bathymetric ranges in order to verify the effects on the swaths. The best results were obtained by decreasing the sound speed values between zero and 10 meters’ depth.

Four SVPs were generated from the logged one: The first SVP was created by reducing the values by 10 meters per second. It was used to correct the survey lines logged within a mile from the coast, where the outer beams were 1.5 meters lower than the actual bottom.

The second profile was generated by reducing the values by seven meters per second: It was used for data logged between one and 1.6 miles from the coast, where the outer parts of the swaths were one meter lower than the actual bottom.

The third SVP was applied to data logged between 1.6 miles and 2.2 miles from the coast, where the difference between the real bottom and the logged data was less than one meter—the new profile was generated by reducing the values by six meters per second.

Finally, frowning was noticeably reduced at distances greater than 2.2 miles from the coast: In this area, the values were reduced by three meters per second.
The method described in this paper to correct the bathymetric data for sound velocity errors produced satisfactory results. It allowed the team to reduce errors and to recover data so that there was no need to use statistical processing methods that remove errors but reduce data resolution.

Two grids were generated to verify the results. The first one, generated with a grid cell size of five by five meters, presents artifacts that are hardly visible. The second grid was generated with a cell size of 10 by 10 meters, and the artifacts are not visible.

Processed data were finally used to generate a high-resolution bathymetric map with a contour interval of one meter.

Once the problem was solved, the cause of the artifacts was investigated. The theory was that there was a fresh-water body in the area producing a strong temperature and salinity gradient, decreasing from the coast toward the open sea due to areal spread and water mixing. Therefore, it was not necessary to correct the bathymetric data logged far from the coast.

At first it was thought that the problem was that the mouths of two rivers (Lambro and Mingardo) were located in the surveyed area and spaced 500 meters from each other. However, areas with high or low freshwater rates were identified by considering the places where the SVP needed to be modified, and there were no strong SVP changes near the river mouths. Moreover, since they are both torrential rivers, their flow in June is negligible.

In the same area on dry land, there is a 112-square-kilometer carbonaceous massif called Mount Bulgheria. Because of their composition, the carbonaceous terranes hold main water supplies. Because they are prone to seepage events, most of the water runoff consists of underground circulation (85 to 95 percent). Mount Bulgheria’s water mass is about 47 million cubic meters per year, and the underground water flow southward produces springs along the coast in the surveyed area.

Studies on submarine freshwater springs located along the Campanian coast are currently in progress. Some sea-bottom structures, consistent with this hypothesis, are located in shallow water 13 miles south of the studied area. This information suggests that some freshwater springs may be located in the shallow water regions of the surveyed area. Freshwater runoff along the coast south of Capo Palinuro may be consistent with the shape of the areas located by the position of the modified SVP.

The authors would like to acknowledge Capt. Aimone Patané and the crew of the Thetis for their efforts. They also thank Dr. Fabrizio Lirer and Patricia Sclafani for their revisions of this article.

For a complete list of references, please contact author Gabriella Di Martino at gabriella.dimartino@iamc.cnr.it.

Gabriella Di Martino received her degree in environmental sciences from Parthenope University in Naples, Italy, in 2003. She has been involved in geophysical data acquisition and processing since 2001.

Renato Tonielli is a technologist at the Coastal Marine Environment Institute of the National Research Council in Naples, Italy. He has more than 15 years of experience in geophysical data acquisition and processing, and he earned his Ph.D. from the Sapienza University of Rome.

-back to top-

-back to to Features Index-

Sea Technology is read worldwide in more than 110 countries by management, engineers, scientists and technical personnel working in industry, government and educational research institutions. Readers are involved with oceanographic research, fisheries management, offshore oil and gas exploration and production, undersea defense including antisubmarine warfare, ocean mining and commercial diving.