Vertical Control via GPS vs. Tide Gauges: A Pilot Study

Vertical Control via GPS vs. Tide Gauges: A Pilot Study Rebecca Quintal, John Shannon Byrne, John Kiernan, Evan Robertson, Walter Simmons, Gary Davis, Paul Donaldson, Deborah Smith, Jason Infantino and Jeff Parker Science Applications International Corporation 221 Third Street Newport, RI 02840 USA Abstract Under contract to the National Oceanic and Atmospheric Administration (NOAA) Office of Coast Survey, the Marine Science and Technology Division of Science Applications International Corporation (SAIC) is performing a debris mapping survey using 200% side-scan sonar coverage with resulting vertical beam echo sounder coverage, south of Terrebonne Bay Louisiana. Water depths in the survey area range from approximately seven feet (near shore) to 25 feet (off shore, three-mile limit). The primary purpose of the survey is to provide survey data suitable for item detection and debris mapping in the wake of Hurricanes Katrina and Rita, and subsequent to funding, Hurricanes Ike and Gustav, which moved through the area. A secondary purpose of the survey is providing modern, accurate hydrographic survey data for updating the nautical charts. SAIC has deployed two shallow draft boats equipped with SAIC s ISS-2000 data acquisition and navigation system, a vertical beam echo sounder and a side-scan sonar. One vessel will also employ a multibeam echo sounder for item investigations. Both survey platforms are equipped with inertial navigation systems. Vertical control is established using an SAIC-installed tide station at Caillou Bay (8763535), and NOAA tide stations at Grand Isle, La. (8761724); Port Fourchon, La. (8762075); and LAWMA, La. (8764227). SAIC s ISS-2000 system as well as the Position and Orientation System for Marine Vessels (POS/MV) are recording raw L1 and L2 Global Positioning System observables within portions of the survey area. These data will be used to compute echo sounder ellipsoidal height, which is then used to correct soundings to the specified mean lower low water (MLLW) vertical datum for comparison with the traditional method of correcting soundings using zoning of water level measurements from shore-based tide stations. The raw L1 and L2 GPS observables acquired on the survey platforms are post-processed using precise point positioning (PPP) techniques. PPP post-processing operates independently of any user-established shore base stations. Rather than relying on user-established shore stations for kinematic corrections, PPP techniques make use of GPS satellite clock corrections and satellite orbit corrections that are freely available via the Internet within 24 hours. Use of GPS-based techniques for vertical control requires analysis to properly define the separation (SEP) from the ellipsoid to the desired vertical datum. The SEP value can change significantly over the spatial extents of the survey area. An integrated survey system, configured with current state-of-the-art equipment for the echo sounder, motion sensor, and sound speed profiling sensor, can meet International Hydrographic Organization (IHO) Order 1 survey requirements using a PPP-based GPS trajectory when the separation uncertainty is suitably known. This paper provides a description of the methods adopted for comparison between the vertical control established from the tide gauges and the vertical control established using GPS PPP techniques. Based on the preliminary results from this pilot study, the potential merits of using a 1

GPS-based vertical control solution, that operates independently of user-established shore base stations, are presented and discussed. 1 INTRODUCTION On August 29, 2005, Hurricane Katrina struck the Louisiana and Mississippi gulf coasts, causing catastrophic damage and depositing tons of debris into the surrounding water bodies, creating hazards to vessel navigation and causing adverse effects to the local commercial fishing industry. In September of 2006, the National Oceanic and Atmospheric Administration (NOAA) started surveying the coastal waters to locate marine debris. These data are being used to assist in the removal of the debris to re-establish safe navigation and commercial fishing and to update the nautical charts throughout the affected area 1. Under contract to NOAA, Science Applications International Corporation s (SAIC s) Marine Science and Technology Division was tasked to survey 161 square nautical miles offshore of Terrebonne Bay, Louisiana (Figure 1), with the primary purpose to provide data suitable for item detection and debris-mapping in the wake of Hurricanes Katrina and Rita. A secondary purpose of the survey is providing modern, accurate hydrographic survey data for updating the nautical charts. The survey specifications required 200% side-scan coverage and single-beam echo sounder data of the three sheets to the inshore limit of eight feet. In addition, a Reson (Reson A/S) 8101 multibeam echo sounder will be employed for item investigations on each sheet at the end of the survey. As part of the survey, a tide station was to be installed to aid in vertical control. SAIC chose to deploy two vessels to complete the Terrebonne Bay surveys, but planned to stagger the vessels starting with one in operation in October 2008 and the second coming online in January 2009. SAIC planned for a field office at the Louisiana Universities Marine Consortium s (LUMCON) DeFelice Marine Center in Cocodrie, La. Figure 1. SAIC Survey Area of Terrebonne Bay, La., Showing Sheets A, B and C. 2

Fortunately, the SAIC team was able to begin survey operations on schedule with the installation of the tide gauge occurring in early October and survey operations beginning on October 21, 2008, despite two hurricanes impacting the Cocodrie, La., area. Hurricane Gustav made landfall on September 1, 2008, near Cocodrie, La., causing severe wind damage and flooding (Figure 2). Gustav was followed by Hurricane Ike, which made landfall on September 13, 2008, in Galveston, Texas, and caused storm surge in the Cocodrie, La., which was still recovering from Gustav. Perhaps it was fortuitous that the debris mapping survey did not begin until after these hurricanes potentially deposited more debris in the Terrebonne Bay area. Figure 2. Photograph of Cocodrie, La., Around September 17, 2008. (Photo courtesy Tim Osborn, NOAA) Recent software enhancements to SAIC s data acquisition software, ISS-2000, added the capability to both collect raw L1 and L2 Global Positioning System (GPS) observable data. At the same time, SAIC s post-processing software, SABER, was enhanced to apply the corrected position and ellipsoidal height data to the Generic Sensor Format (GSF) bathymetry data files. The ellipsoidal height data serves as vertical control as opposed to traditional tidal zoning vertical control. Other correctors that are required with tide vertical controlled data are heave, draft, settlement and squat. The raw L1 and L2 GPS observables acquired on the survey platforms are post-processed using GPS precise point positioning (PPP) techniques. PPP post-processing operates independently of any user-established shore base stations. Rather than relying on user-established shore stations for kinematic corrections, PPP techniques make use of GPS satellite clock corrections and satellite orbit corrections that are freely available via the Internet within 24 hours of realtime. Positioning accuracies with PPP techniques of better than 20 cm in the horizontal and 30 cm in the vertical have been demonstrated using commercial GPS processing software packages 2. The Terrebonne Bay area surveys provided a logical scenario to perform a pilot study of the PPP technology for hydrographic surveys. The survey area is remote and not easily accessible for establishing and maintaining shore stations if real-time kinematic (RTK) positioning were to be used. Also, since a tide station was installed for this survey, a cost/benefit analysis of the two methods for vertical control may be performed once the survey is complete and final comparison results have been analyzed. 3

2 SYSTEMS Mobilization for the survey commenced in early October 2008, with the installation of a tide station at Caillou Bay, La. (8763535). The station was installed on a Texas Gas Transmission, LLC platform (Figure 3) and comprises two Design Analysis H350XL digital bubbler gauges, a tide staff, and Geosynchronous Operational Environmental Satellite (GOES) satellite telemetry link provided by John Oswald & Associates. Figure 3. Tide Station at Caillou Bay, La. (8763535). SAIC Photograph SAIC is employing two vessels to conduct the survey. The Sea Beneath is a 32-foot aluminum work boat with twin 500hp outboard motors (Figure 4). The vessel is owned and operated by EMC, Inc., in Greenwood, Miss. The Lacey Marie is a 42-foot Lafitte Skiff work boat with a two-foot draft and is diesel powered (Figure 5). The vessel is owned and operated by Campo Marine in Shell Beach, La., and under charter to SAIC via a sub-contract with Lowe Engineers, LLC, of Atlanta, Ga. Figure 4. Sea Beneath. SAIC Photograph 4

Figure 5. Lacey Marie. SAIC Photograph Both vessels are configured with SAIC s ISS-2000 real-time survey system, which consists of a dual processor computer with Microsoft Corporation s Windows XP operating system running SAIC s ISS-2000 software. This software provides survey planning and control in addition to data acquisition and logging for single beam, multibeam and navigation data. The system includes: Applanix Corporation POS/MV 320 Position and Orientation System with a Trimble ProBeacon differential receiver Trimble 4000 GPS receiver with a differential receiver Sea-Bird Electronics, Inc., Model SBE 19-01 conductivity, temperature, depth (CTD) profiler L-3 Communications/Klein Associates Inc. (Klein), 3000 side-scan sonar (bow-mounted) Odom Echotrac CV vertical-beam echo sounder with a pole-mounted 200-kHz transducer 3 DATA ACQUISITION The task order with NOAA provided preliminary water-level zoning based on NOAA tide stations in Port Fourchon, La. (8762075), and LAWMA, La. (8764227). In addition, SAIC was required to install a tide station at Caillou Bay, La. (8763535), and operate it for the duration of the survey. Subsequent to the start of survey operations, SAIC s ISS-2000 system and the POS/MV system were configured to record raw L1 and L2 GPS observables within portions of the survey area to test the use of GPS PPP techniques for determining water-level correctors. The data presented in this paper were collected on H11784 (Sheet B) over six days (Figure 6). The data fall within one modified NOAA-provided tide zone for the Port Fourchon station (zone CGM749) and include four cross lines. 5

Figure 6. Track Lines of Data Used for This Paper (yellow lines), with Port Fourchon, La. (8762075), and Caillou Bay, La. (8763535), Tide Zoning (red polygons). SAIC graphic 4 DATA PROCESSING The following sections provide details on the specific processing steps performed on the raw GPS-observable data and the bathymetry data to support GPS-based vertical control and traditional tidal control. Figure 7 is a processing flow chart that aids in the understanding of the processes detailed below. The steps colored in black are steps performed on both the tidal vertical control dataset and the GPS-based vertical control dataset. The green steps are for the GPS-based vertical control processing pipeline, while the blue steps are for the tidal verticalcontrol processing pipeline. 6

GSF Bathymetry Files Navigation Processing Apply Delayed Heave POS/MV Data File Clock Orbit Corrections Edit and QC Separation Definition Water Level Measurements Merge XYZ Navigation Tide Benchmark Data EGM2008 Model Apply Tidal Corrections GPS Based Vertical Control Uncertainty Attribution Uncertainty Attribution Surface Generation Surface Generation Surface Differencing Figure 7. Flow Chart for Processing Steps. 4.1 VERTICAL CONTROL PROCESSING As part of the task order, SAIC is to compute the datum at the SAIC-installed Caillou Bay station by comparison with the NOAA control station Grand Isle, La. (8761724). The determination of final water-level correctors is also an SAIC responsibility. SAIC modified the NOAA-provided preliminary tide zones to cover the receding shoreline on the north because the contracted inshore limit crossed into the zones intended for surveys north of the islands. Once the survey began, it quickly became apparent that the zoning from the LAWMA station was not adequate and would not provide the data needed to determine whether we had reached the inshore limit of the survey. Figure 8 illustrates this problem. At 18:30 UTC on October 27, 2008, there is a 0.565-meter difference between the observed data and the WGM413 zoned data. This two-foot (0.565 meter) uncertainty in the water level could equate to a displacement of soundings horizontally by 2000 feet (610 meters) or more in this very flat bottom. 7

Caillou Bay WGM413 0.8 0.6 Tidal Height - Meters 0.4 0.2 0-0.2-0.4 0:00 1:54 3:48 5:42 7:36 9:30 11:24 13:18 15:12 17:06 19:00 20:54 22:48 0:42 2:36 4:30 6:24 8:18 10:12 12:06 14:00 15:54 17:48 19:42 21:36 23:30 Time October 26-27, 2008 Figure 8. Caillou Bay Record Compared to Zone WGM413 on LAWMA. SAIC graphic Because of the numerous shape and height differences between the LAWMA-zoned water levels and those observed at Caillou Bay, SAIC elected to prepare preliminary zones based on the observed water levels at Caillou Bay. At the same time, SAIC modified the western bounds of zone CGM749 (Port Fourchon) to match the new CAI### (Caillou Bay) zones. Where acquisition of the L1 and L2 data is accomplished for the Terrebonne Bay surveys, SAIC may use the GPS PPP solution to aid in development and improvement of the final water level zoning from the tide stations. For purposes of this paper, the observed tides based on the modified CGM749 zone on the NOAA Port Fourchon station are being used for tidal corrections. Future comparisons will be made against the Caillou Bay tide data, where applicable. 4.2 NAVIGATION PROCESSING SAIC chose to use the NovAtel, Inc., Waypoint Products Group GrafNav GPS post-processing software for the PPP navigation processing. GrafNav operates on the raw GPS observables and utilizes the precise clock and orbit corrections to develop the PPP trajectory. It is necessary to ensure a sufficiently high sampling rate of the PPP trajectory to resolve all of the motion of the platform. The required sampling rate of the trajectory has a one-for-one mapping to the sampling rate of the raw GPS observables. This is in part due to the fact that GrafNav does not make use of the high rate POS/MV inertial measurement unit (IMU) data. For the small boats utilized on this project, a 10 Hz sampling rate of the raw GPS observables is used. The following are the high-level processing steps: 8

1. Extract the primary L1 and L2 GPS data stream records from the ISS-2000 or POS/MV raw data files using SAIC s SABER. This creates a GrafNav compatible Global Navigation Satellite System (GNSS) file. 2. Using GrafNav, convert the raw GNSS file to a binary GPB (*.gpb) file and an ASCII EPP (*.epp) file. The GPB has the GPS measurement data and the EPP stores the ephemeris data. Both are GrafNav s custom formats. 3. Load the GPB file into the current GrafNav project. 4. Download the GPS raw service data, the Precise Ephemeris File (*.sp3) and Precise Clock (*.clk) data files from the Internet. The Download Service Data automatically retrieves the data from the appropriate ftp servers. 5. Using the four files above, create a PPP solution using Multi-Pass (forward, reverse, and forward again) process to converge on the PPP solution. 6. Export the time, date, latitude, longitude, ellipsoidal height, and the standard deviations to an ASCII file. 4.3 SEPARATION DEFINITION The PPP data provide the ellipsoidal height of the vessel s reference point; however, the distance (or separation) between the ellipsoid and the chart datum must be determined and applied to use the GPS (PPP) data for vertical control. The ellipsoidal height of the primary bench mark at the Caillou Bay station was determined by GPS observations during SAIC s station installation. Preliminary datum determination established the height of the benchmark above the mean lower low water (MLLW) datum. Combining these two values yielded the separation (SEP) from the ellipsoid to MLLW at the Caillou Bay station. NOAA published data for the Grand Isle station provided the ellipsoidal height of the primary benchmark, and that mark s height above the MLLW datum at the station. Combining these two values yielded the SEP from the ellipsoid to MLLW at the Grand Isle station. With Grand Isle beyond the east end of the project and Caillou Bay at the west end, SAIC had the means to calculate the SEP values across the survey area. A worldwide Earth Gravitational Model (EGM2008), which provides the WGS-84 to geoid undulation has been published by the National Geospatial-Intelligence Agency 3. This model is available at two resolutions: 1 by 1-minute and 2.5 by 2.5-minute resolution. At both the Caillou Bay and Grand Isle tide station the 2.5 x 2.5-minute EGM2008 model undulation height was subtracted from the SEP to MLLW for that position. These difference values were then interpolated across the survey area and added to the EGM2008 height at specific points to determine the SEP to MLLW as shown in Table 1 and Figure 9. The EGM2008 model was used as a tool for interpolating the SEP to MLLW across the survey area. 9

Latitude (N) Longitude (W) Table 1. Preliminary SEP Values Used SEP SEP Uncertainty Latitude (N) Longitude (W) SEP SEP Uncertainty 29.17533 90.97676-25.992 0.15 29.05840 90.73884-25.643 0.15 29.17649 91.03423-26.032 0.15 29.08295 90.65671-25.648 0.15 29.19390 90.90063-26.022 0.15 29.03149 90.66651-25.508 0.15 29.13297 90.88838-25.892 0.15 28.99081 90.66743-25.398 0.15 29.13350 90.94234-25.897 0.15 29.00625 90.59557-25.369 0.15 29.12243 91.03423-25.922 0.15 29.04948 90.60148-25.499 0.15 29.09084 91.03413-25.852 0.15 29.10307 90.59194-25.629 0.15 29.10093 90.95717-25.842 0.15 29.10154 90.49238-25.529 0.15 29.11010 90.88263-25.842 0.15 29.04397 90.49896-25.399 0.15 29.07847 90.89918-25.782 0.15 28.98860 90.50997-25.259 0.15 29.06607 90.98205-25.782 0.15 28.98909 90.43915-25.229 0.15 29.05629 91.03928-25.782 0.15 29.02659 90.43097-25.319 0.15 29.01899 91.02271-25.672 0.15 29.06289 90.42385-25.399 0.15 29.03283 90.97337-25.683 0.15 29.07019 90.34624-25.355 0.15 29.06406 90.89083-25.728 0.15 29.03565 90.35058-25.285 0.15 29.02207 90.91518-25.633 0.15 28.99360 90.35821-25.185 0.15 28.98221 90.91045-25.513 0.15 29.01245 90.25341-25.175 0.15 28.98115 90.84381-25.482 0.15 29.05975 90.25008-25.275 0.15 29.02034 90.84403-25.592 0.15 29.09951 90.24937-25.355 0.15 29.05428 90.83952-25.682 0.15 29.11393 90.19781-25.355 0.15 29.05712 90.79003-25.668 0.15 29.06898 90.19730-25.275 0.15 29.01856 90.79085-25.558 0.15 29.03468 90.19788-25.195 0.15 28.97974 90.79375-25.458 0.15 29.05478 90.12892-25.196 0.15 28.98020 90.74020-25.433 0.15 29.09948 90.14634-25.296 0.15 29.01831 90.73937-25.533 0.15 29.12348 90.16124-25.346 0.15 Figure 9 shows the color-coded EGM2008 WGS-84 to geoid separation with overlays of preliminary tide zoning (red), survey bounds (blue) and SEP to MLLW values (blue) as determined by SAIC. Flags indicate the three tide stations in the area. 10

Caillou Bay SAIC Grand Isle NOAA Port Fourchon NOAA Figure 9. EGM2008 Predicted WGS-84 to GEOID Undulation in Survey Area, Values Are SEP to MLLW. SAIC graphic 4.4 MERGING PROCESSED NAVIGATION The PPP navigation files contain updated latitude, longitude and height values. As the assumption that the PPP latitude and longitude are more accurate than the Differential Global Positioning System positions collected during acquisition, the position for each ping is updated along with adding in the height value when the PPP navigation is merged with the GSF bathymetry files. In order to make a direct comparison between the vertical control of the PPP processed files and the traditionally processed tidal files, the positions need to be the same in both datasets. Once all other corrections and processing have been performed, the PPP data are merged into to the bathymetry files using SAIC s SABER version 4.3. The positions for each ping are updated and a new field within the GSF files is populated with the GPS height corrector (the height above the ellipsoid). At this point, the depths for each beam have not been recalculated. That process comes after the next step in the process flow. After merging the navigation data, copies of the files are made so that they may be processed in parallel pipelines for GPS and tidal vertical control. As mentioned above, there are several updates to the GSF format to accommodate GPS vertical control. A new version of GSF will be issued in the April/May 2009 timeframe. Two of these modifications are the addition of new ping record fields and a new ping flag. The new ping record fields in the GSF file are SEP: Separation value the distance between the ellipsoid and the chart datum Height: Ellipsoidal height the height above the ellipsoid GPS Tide Corrector: GPS vertical correction value applied to each ping. This value also includes the correction for heave. 11

A new ping flag has been defined for GPS-based vertical control. This ping flag is mutually exclusive with conventional tide flags for any given ping, but may change from one ping to the next within a file. Post-Processing Kinematic (PPK) and PPP are defined as positioning modes for the HV Nav Error record, and Mean Sea Level is added as a vertical datum intended for deep water (greater than 200 meters) work where a low water datum is not available. 4.5 APPLYING TIDES/GPSZ DATA For the copy of the GSF files that are to be processed using tidal data, the water level files are applied to the bathymetry files. Any predicted tidal corrections that may have been applied during acquisition are removed and then the new observed tidal corrections are applied to the files. The SAIC SABER software automatically sets a flag for each ping to indicate the type of tide correction that has been applied (predicted, observed preliminary or observed verified). Based on this ping flag, the SAIC SABER software will ignore any fields that are populated for GPS vertical control (such as the new ellipsoidal height corrector field) and will instead use the fields such as heave, draft, tide corrector, depth corrector (settlement and squat) to calculate the depth value for each beam. For the copy of the GSF files that are to be processed using the PPP data, the SEP value is required. The user has the choice of applying a single SEP value, using the EGM2008 modeled values or providing a text file that contains the SEP values determined by the user. When the user chooses to apply PPP data instead of tidal data, the SAIC SABER software will set a flag for each ping indicating that GPS vertical control is being used for this ping. Therefore, when the depth for each beam is calculated, the software will remove the corrections for tides, heave, draft, and settlement and squat, and will then add in the ellipsoidal height corrector and SEP value to calculate the final depth value for each beam. For both processing pipelines, uncertainty attribution is applied to the data next. 4.6 UNCERTAINTY ATTRIBUTION The Total Propagated Uncertainty (TPU) model incorporated within SAIC SABER has been updated to support GPS-based vertical control. The uncertainty model uses the ping flag that is applied to determine the uncertainty parameters that should be used in the horizontal and vertical uncertainty attribution for each beam. If the ping flag indicates that tidal corrections are applied to the ping, then the program will ignore the uncertainty fields pertaining to GPS corrections. If the ping flag indicates that GPS-based vertical control corrections are applied to the ping, then the program will ignore the uncertainty fields pertaining to tidal corrections (zoning uncertainty and gauge measurement uncertainty). The SAIC SABER TPU model includes two new fields for GPS vertical control. A user-defined SEP uncertainty value and a new vertical uncertainty value for the positioning system. When vertical control is established based on GPS, the dynamically updated vertical uncertainty from the positioning system is combined with the estimated SEP uncertainty and this value is combined with the sonar contributions to the model to provide the final uncertainty. It is important to note that whenever a new navigation source is applied, the dynamically changing 12

horizontal and vertical uncertainty values are correspondingly updated. When using GPS-based vertical control, the component uncertainties associated with conventional tides, settlement and squat, static draft, and loading draft are all bypassed. If the vertical uncertainty from the positioning system is stored in the GSF file, then the values stored in the GSF file are used by the TPU model; if they do not exist in the GSF file, then the model uses the fixed value defined by the user. Once the uncertainty attribution is complete, combined uncertainty bathymetry estimator (CUBE) grids are generated for each dataset (PPP and tides) and the resulting surfaces can then be compared. 5 RESULTS CUBE grids were generated at one-meter resolution for both the GPS-based vertical control dataset and the tidal vertical control dataset. The minimum and maximum CUBE depths were 1.20 meters and 8.92 meters, respectively for the GPS vertical controlled data and 1.20 meters and 8.67 meters, respectively, for the tide vertical controlled data. Various comparisons were made between gridded crossings data. The first junction analysis was performed by subtracting the entire one-meter GPS CUBE surface from the one-meter tide CUBE surface. Comparisons of these crossing data showed that 97.25% of comparisons were within 15 centimeters, and 99.46% of comparisons were within 20 centimeters (Table 2). A comparison was performed on the tidal mainscheme lines versus the four tidal cross lines by subtracting the cross line grid from the mainscheme grid. Comparison of these crossing data showed that 87.61% of comparisons were within 15 centimeters, and 97.01% of comparisons were within 20 centimeters (Table 3). Almost all the differences are positive, which indicates a possible issue with the water level corrector determination. A similar comparison was performed on the GPS mainscheme lines versus the four GPS cross lines. These data showed that 98.29% of comparisons were within 15 centimeters, and 99.15% of comparisons were within 20 centimeters (Table 4). While there are limited crossings in these datasets, the results show an appreciable improvement with the GPS vertical control data. Table 2. GPS Vertical Control Versus Tidal Vertical Control Junction Analysis Results Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 322044 56.73% 139189 57.07% 145488 50.80% 37367 100.00% 5 10 181374 88.69% 86219 92.42% 95155 84.03% 10 15 48642 97.25% 15048 98.59% 33594 95.76% 15 20 12494 99.46% 2486 99.60% 10008 99.26% 20 25 2723 99.94% 964 100.00% 1759 99.87% 25 30 140 99.96% 1 100.00% 139 99.92% 30 35 91 99.98% 0 100.00% 91 99.95% 35 40 133 99.99% 0 100.00% 133 99.99% 40 45 2 100.00% 0 100.00% 2 100.00% Totals 567643 100.00% 243907 42.97% 286369 50.45% 37367 6.58% 13

Table 3. Tidal Vertical Control Main to Cross Junction Analysis Results Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 28 11.97% 23 10.04% 3 100.00% 2 100.00% 5 10 64 39.32% 64 37.99% 0 100.00% 10 15 113 87.61% 113 87.34% 0 100.00% 15 20 22 97.01% 22 96.94% 0 100.00% 20 25 7 100.00% 7 100.00% 0 100.00% Totals 234 100.00% 229 97.86% 3 1.28% 2 0.85% Table 4. GPS Vertical Control Main to Cross Junction Analysis Results Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 123 52.56% 68 46.58% 40 54.79% 15 100.00% 5 10 93 92.31% 68 93.15% 25 89.04% 10 15 14 98.29% 10 100.00% 4 94.52% 15 20 2 99.15% 0 100.00% 2 97.26% 20 25 2 100.00% 0 100.00% 2 100.00% Totals 234 100.00% 146 62.39% 73 31.20% 15 6.41% All of the data used for this paper falls within one Port Fourchon tide zone; therefore, we were not able to analyze tide zone boundaries. However, we did evaluate the cross line data on the eastern (Figure 10, Figure 11 and Table 5) and western (Figure 12, Figure 13 and Table 6) ends of tide zone CGM749. These comparisons indicate that the single CGM749 zone may not be adequate to define the water level correctors for this area. For the eastern cross line there is a larger difference between tide-corrected main and cross lines than at the western cross line. In addition, the GPS height-controlled crossings show much better agreement at both the east and west cross lines. 14

Figure 10. Tidal Eastern Cross Line and Northern Three Mainscheme Lines Figure 11. GPS Eastern Cross Line and Northern Three Mainscheme Lines Table 5. Water Level Values Applied at Three Crossings for Eastern Cross Line Line File Day Tide Time Ping Eastern Cross Line 088.d11 29-Mar 0.089 22:53:09-22:53:36 8200-8700 Northernmost Mainscheme Line 080.d03 21-Mar 0.168 14:38:57 4550 Next Mainscheme South 079.d09 20-Mar 0.251 16:10:37 5650 Second Mainscheme South 079.d08 20-Mar 0.233 15:56:16 120910 15

Figure 12. Tidal Western Cross Line and Northern Three Mainscheme Lines Figure 13. GPS Western Cross Line and Northern Three Mainscheme Lines Table 6. Water Level Values Applied at Three Crossings for Western Cross Line Line File Day Tide Time Ping Western Cross Line 088.d08 29-Mar 0.318 20:27:03-20:27:19 9500-9800 Northernmost Mainscheme Line 079.d08 20-Mar 0.177 14:32:40 27111 Next Mainscheme South 078.d12 19-Mar 0.460 20:10:13 21558 Second Mainscheme South 088.d09 29-Mar 0.221 21:22:04 34726 Tables 7, 8, and 9 show that there is good agreement between the GPS vertical control and the tidal vertical control near the center of the CGM749 zone, but that there appears to be an offset at the east and west ends of the test area. In the middle of the tide zone, the percentage of positive to negative comparisons is more equally distributed. However, there is a larger percentage of negative comparisons on either end of the zone. Because the GPS CUBE surface was subtracted from the tide CUBE surface, this means that there are more cells that have deeper depths in the GPS dataset. An example of this is shown in Figure 14. This figure shows the final depths of the tide and GPS-based vertical control from file 088.d09, which is the rightmost mainscheme 16

line shown in Figure 10 and Figure 11 (eastern side of the tide zone). The GPS PPP trajectory data will prove useful for updating the tide zoning in this survey area for the zones based on both the Port Fourchon and Caillou Bay tide gauges. Table 7. GPS Vertical Control Versus Tidal Vertical Control Junction Analysis East Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 11861 61.75% 3456 64.61% 7175 56.81% 1230 100.00% 5 10 6313 94.61% 1528 93.18% 4785 94.70% 10 15 427 96.83% 181 96.56% 246 96.64% 15 20 608 100.00% 184 100.00% 424 100.00% Totals 19029 100.00% 5349 27.85% 12630 65.75% 1230 6.40% Table 8. GPS Vertical Control Versus Tidal Vertical Control Junction Analysis Middle Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 10213 45.87% 4580 44.82% 4531 41.41% 1102 100.00% 5 10 9702 89.45% 4472 88.59% 5230 89.20% 10 15 2298 99.78% 1116 99.51% 1182 100.00% 15 20 50 100.00% 50 100.00% 0 100.00% Totals 22263 100.00% 10218 45.90% 10943 49.15 1102 4.95% Table 9. GPS Vertical Control Versus Tidal Vertical Control Junction Analysis West Depth All Positive Negative Zero Difference Range (cm) Count Percent Count Percent Count Percent Count Percent 0 5 17166 52.85% 6385 49.91% 9107 50.55% 1674 100.00% 5 10 12265 90.60% 4929 88.45% 7336 91.26% 10 15 3039 99.96% 1466 99.91% 1573 99.99% 15 20 13 100.00% 12 100.00% 1 100.00% Totals 32483 100.00% 12792 39.38% 18017 55.47% 1674 5.15% 17

File 088.d09 Depth Comparison Tidal Vertical Control GPS PPP Vertical Control 2 2.1 2.2 Depth - Meters 2.3 2.4 2.5 2.6 2.7 80450 80500 80550 80600 80650 80700 80750 80800 Time of Day - Seconds Figure 14. File 088.d09 Depth Comparison Between Tidal and GPS PPP Vertical Control. With water depths in this test dataset between one meter and nine meters, the allowable International Hydrographic Organization (IHO) Order 1 vertical uncertainty is between 0.50 and 0.51 meters. The junction analysis results of both the tidal vertical control and GPS-based vertical control datasets had 100% of the comparisons at 25 centimeters or less. There was a constant average total propagated vertical uncertainty for all of the soundings that contribute to a give node for the tide CUBE grid of 0.30 meters. The minimum and maximum average total propagated vertical uncertainty for all of the soundings that contribute to a given node for the GPS data was 0.14 and 0.23 meters, respectively. 6 SUMMARY At the time of this writing, it is not possible to draw strong conclusions about the benefits of the GPS-based approach to vertical control versus the traditional tidal approach; however, preliminary results show a significant improvement in junction analysis results by using GPS vertical control. An added benefit to the survey is that the GPS PPP trajectory data will be used to aid in updating the current tide zoning for the survey area based on the NOAA Port Fourchon gauge and the SAIC-installed Caillou Bay gauge. Analysis of the trajectory data will continue throughout the remainder of the survey for all three sheets. 7 REFERENCES 1 Gulf of Mexico Marine Debris Project, NOAA Website http://gulfofmexico.marinedebris.noaa.gov/ 2 Russell, D.C., Wong, R. and Toor, P., 2007. Real-time Tidal Observation Using High-accuracy GPS Positioning. Proc Hydro 2007 Conference, Cairns, 21-24 November 2007. 3 Earth Gravitational Model 2008 (EGM2008), NGA Website http://earthinfo.nga.mil/gandg/wgs84/gravitymod/egm2008/index.html 18