Implementation of a Multibeam Echosounder and Terrestrial LiDAR in Support of a Dam Spillway Construction Project

Nestled Northeast of the city of Folsom, California, sits the Folsom Lake—formed by an existing dam on the North Fork and South Fork of the American River. It is responsible for providing irrigation, drinking water, and electricity to parts of California, as well as flood prevention for the areas situated south of the lake. Despite the low risk of a damaging occurrence, it was decided (under the Reclamation Safety of Dams Act) that preventive measures needed to be taken to protect those living downstream. Construction of a new auxiliary spillway is in progress, which will provide a means for a steady water outflow during an increased water flow event. Carried out in phases, this project requires surveys from land and water to ensure successful completion. The $900-million cooperative project between the U.S. Army Corps of Engineers, U.S. Department of the Interior, and Bureau of Reclamation will assist the Sacramento region to achieve the level of flood protection.Completion of the auxiliary spillway is scheduled for late 2017.

To support the project, R.E.Y. Engineers, Inc. (Folsom, California) has conducted bathymetric surveys and back scatter imaging. Seafloor Systems (El Dorado Hills, California) provided the equipment and on-site operational expertise for the hydrographic survey. Specifically, R.E.Y. Engineers, Inc. was tasked to provide bathymetry and ground control for the new spillway construction.The hydrographic equipment supplied and operated by Seafloor remained the same throughout the project; however, three vessels were used, each with unique challenges to overcome:

The first vessel was a  Fisher pontoon boat, equipped withan Applied Marine System Hydro-Mount—which made sonar installation and vessel launching seamless with its mounting diversity and rotating arm. The stable platform of the pontoon and spacious work area seemed made for a very pleasant day.

The second vessel was a  Valco. Due to the limited space, an existing sonar mount was modified and strengthened to accommodate the heavier multibeam.Althoughit was challenging to work in such a small area with the various equipment topsides, the main issue was vessel vibration. At a certain engine RPM, the aluminum boat shook with the force of a small tremor and introduced noise into the bathymetry data; this made for a very slow moving and long day.

The third vessel was a  Bennington pontoon, custom outfitted specifically forhydrographic surveying. Moving back to the Applied Marine Hydro-Mount, it features a dedicated workstation, stable antenna frame, mobile LiDAR mount, and even a head...truly the Cadillac of the project so far.

The ability to mobilize quickly without sacrificing data integrity (due to +/- 0.5 foot accuracy specification) was crucial for the completion of these on-going surveys. With this in mind, the survey platform consisted of the following:

• Applanix POS-MV™Wavemaster Inertial Navigation System

http://www.seafloorsystems.com/posmvwavemaster.html

Used to provide roll stabilization to the multibeam in addition to positioning, heading, roll, pitch and heave of the vessel’s reference point. Due to the steep slope of the spillway embankment, the POS-MV was chosen in case of radio shadow or GPS dropout.

• Reson SeaBat T20-P Multibeam Sonar

http://www.seafloorsystems.com/multi_echosounder.html

Gathered both bathymetry and backscatter imagery, operating at its full capacity, with a reduced ping-rate of 20Hz to keep data processing time down without creating holes in the final surface. The T20-P was ideal for this project due to its small size, ability to be rotated, and high-resolution capability.

• ValeportMiniSVS

http://www.seafloorsystems.com/ctd.html

The ValeportMiniSVS was mounted on the pole, next to the sonar head, allowing it to provide surface speed of sound measurements used in the multibeam’s beam forming.

• Odom DigiBarS

http://www.seafloorsystems.com/ctd.html

The Odom DigiBarS acquired sound velocity profiles at a  increment throughout the water column.

• Trimble R7 GNSS receiver set as a RTK base station

http://www.seafloorsystems.com/satellite.html

Finally, a Trimble R7 GNSS base receiver occupied a control point at the edge of the spillway embankment. It made use of a Trimble TDL-450H UHF radio to send CMR+ corrections to the POS-MV, and a 0dB gain omnidirectional antenna, radiating in a spherical pattern better suited to overcome the vertical relief between the top of the embankment and the survey area below.

Once the equipment was installed, configured, and offsets measured with a total station, it was time to begin the survey. The control for the survey had already been established, based upon the primary control point for the damspillway construction project. An auxiliary control point was set at the edge of the spillway embankment, providing unobstructed line-of-sight to the bathymetric survey area below. The horizontal position for this point was established via two three-minute RTK observations from the dam spillway. The elevation of this point was established via digital barcode leveling, also from dam spillway.While on the water, the Reson Seabat Multibeam was busy calculating 10240 points per second—deciding whether each point passed both the brightness and collinearity tests—the Applanix provided data at 100Hz, logging the raw trajectory information for use in post-processing while the Trimble was on the bank logging GPS (L1, L2, L5) and GLONASS (L1, L2) internally at one second, to be used in the post-processing of the trajectory from the Applanix. (It is important to note that in order to successfully post-process the GNSS data, a sufficient duration of satellite data must be acquired to enable initialization of the solution. In general, the duration of data required is directly dependent upon the distance to the base station—in this case, the distance was less than 2km. Data logging on the POS-MV was initiated 5 minutes prior to the start of sonar acquisition, and continued for an additional 5 minutes after completion. A general rule of thumb is 30 minutes of data for every 20km of distance.)The small area was contained within a silt screen, and the quick rise of the lake bottom, coupled with large boulders made for a few nail-biting moments, but luckily, no damage was done to the sonar head. Finding a suitable patch test area was a challenge for the surveys, due to the restrictiveness of the silt screen and its habit to move around between surveys. In the end, each one of the surveys provided 400% bottom coverage and acceptable patch tests.

With the field work completed, processing the data began. First, the raw data from the base station and the POS-MV was post-processed in POSPac MMS version 7.1 SP1, producing a smoothed best estimate trajectory (SBET). During processing, the survey was processed forward in time, then backward in time, and then combined to produce the final trajectory. The final tightlycoupled trajectory was the result of a blended solution, incorporating all observables (GNSS and attitude), such that the trajectory maintains the positional accuracy of the GNSS data while incorporating the dynamic accuracy of the IMU and heading information. The resulting SBET file was then imported in to HYPACK 2014 for post-processing of the sonar data.Within HYSWEEP’s 64bit editor, the SBET file was applied, along with the sound velocity profiles, first to the patch test lines, and then to the survey area, once the patch test values were found.

The data was cleaned using a combination of methods. Only one base filter was applied to the bathymetry, which flagged any point that did not pass the Reson QC tests. After checking the impact on the data, those points were removed. Next, erroneous points (multipath & specular reflection) were removed in HYPACK’s Sweep Editor. The Sweep Editor will show a user-defined amount of pings (400 sweeps was used, which is 20 seconds of data), which allows trends in the data set to be observed. To keep the final project manageable, the data was reduced to a  grid, and the median value for all points within a cell was chosen. After viewing the gridded color model within HYSWEEP, areas of question were opened in HYSWEEP’s Cell Editor. The cell editor shows all points within a single cell, and gives an option to view the points in surrounding cells as well. With repeatability (accuracy) in mind, the overlapping data from separate lines in the questionable cells was examined, and points that strayed considerably from the repeatable bottom were removed.

The XYZ, which used actual XY data where possible, was exported from HYPACK and brought intoAutoCAD Civil 3D 2013 to create the digital terrain model (DTM). The bathymetry was then compared to a terrestrial LiDAR scan that was also performed by R.E.Y.Engineers. Prior to the first hydrographic survey, due to the ongoing drought in California, the water level of Folsom Lake dropped enough for construction vehicles to work in the (now underwater) survey area, which in turn, warranted a terrestrial LiDAR scan to be conducted.

Shortly before the first survey, a moderate amount of rainfall occurred, bringing the lake water level to a point where hydrographic survey was possible, and needed. In the comparison of the two data sets, it was revealed that, roughly, only 30% of the allowable error budget was used.

Thanks to the different types of hydrographic survey equipmentthat was used, and successful surveying that was conducted, the project was completed to the satisfaction of the client and on schedule.