On-The-Fly GPS for Vertical Control of Hydrographic Surveys
Modern Measurement of Vessel Squat and Settlement using GPS

On-The-Fly GPS for Vertical Control of Hydrographic Surveys

Background

The National Ocean Service (NOS) of NOAA, the civilian hydrographic mapping and charting authority in the United States employs Mean Lower Low Water (MLLW) as the vertical reference datum for its nautical charts. Hydrographic surveys are regularly conducted such that the measured depths must be adjusted prior to presentation on charts. The adjustments to measured depths are necessary to account for the manner in which the survey platform reacts with the environment, as well as, known effects of the environment. Given a depth sounder on a surface ship, the former stems from the manner in which the height of the transducer above the bottom changes with the vessel's static draft, dynamic draft, and the vessel's response to surface waves. The latter stems from the manner in which the height of the free water surface above the bottom changes with astronomical, hydrological, and meteorological conditions.

The Galveston Bay Project relied heavily on the personnel and officers of the NOAA Ship HECK and the NOAA Launch 518 during the field work. Their dedication and professional efforts were essential to the overall success of this project. It is recognized that the collective contributions from NOS and USACE to the partnership on this project greatly exceed those specifically described herein.

Utilization of GPS

Starting in 1983, NOS has developed centimeter-level kinematic GPS which supports land, air, and marine applications which required initialization at a known geodetic location. Starting in 1989, NOS extended the previous successes by developing techniques, known as on-the-fly (OTF) GPS, for practical real-time centimeter-accuracy positioning which did not require static initialization (Remondi, 1992). The simplified essence of OTF is: compute the number of complete carrier cycles between two GPS receiving antennas and the transmitting antennas on five GPS satellites, then solve for the relative positions between the two GPS receiver antennas. OTF can be applied during real-time, going forward on the time line. GPS data can also be processed after the fact via OTF going either forward or reverse, or both, on the time line.

The Hydrographic Systems and Technology Programs (HSTP) in NOS has conducted investigations into the concept of utilizing OTF GPS to provide a temporary vertical reference for use during the conduct of hydrographic surveys. HSTP's initial reason for investigation into this application of GPS technology was provided by its participation in the 1993 campaign of the Hydrographic Ground Truth Experiment (HYGRO) (Mayer et al., 1993). Conducting hydrographic experiments on the CSS FREDERICK G. CREED, required that HSTP know the dynamic draft characteristics of this unique hydrographic platform (Grodin et al., 1992). HSTP employed OTF for estimating the CREED's dynamic draft. Additionally there was an impromptu extension to HYGRO'93 wherein HSTP conducted hydrographic measurements using the MV Mary-O operating above the Reversing Falls on the St. John River which was supported with OTF processing of GPS (Deloach et al., 1994). The following year, during HYGRO'94, HSTP utilized OTF to track the vertical movements of the depth sounder's transducer on the CREED and during low tide to ground truth the ellipsoidal heights of a transect along the bottom. Figure 1 presents the demeaned concurrent time histories of the OTF determined GPS height of the transducer and the depth measured closest in time to the 1-second GPS epochs.

Figure 1. Concurrent Demeaned GPS Heights and Acoustic Depths

This data sequence was obtained during a short period when the CREED, having stopped on station to take a sound velocity profile, was gently responding to low amplitude swell. The agreement between the demeaned parameters, OTF heights and acoustic depths, strongly suggests that proper combination of the actual parameters should yield a reasonably consistent value for the ellipsoidal bottom height at the location of the sound velocity cast. Figure 2 illustrates the use of OTF GPS as a temporary vertical reference for hydrographic surveys. When an OTF measured height of the launch-based GPS antenna is combined with the acoustic depth measurement and the physical distance of the GPS antenna above the acoustic transducer, the result is the height of the bottom relative to the ellipsoid. This is true regardless of the stage of the tide and regardless of either the static or dynamic draft of the launch. The field of ellipsoidal bottom heights would then be moved to chart datum by combination with the field of ellipsoidal MLLW heights.

Figure 2. Hydrograpic Surveys with Vertical Control by GPS

Galveston Bay Project

The HSTP, seeking to further investigate the utility of OTF GPS for a temporary vertical reference during hydrographic surveys, developed the Galveston Bay Project in partnership with three additional units within NOS and with the Topological Engineering Center (TEC) of USACE. The NOAA Ship HECK and the NOAA Launch 518 were conducting ongoing hydrographic surveys in and around Galveston Bay, Texas. Consequently several NOS Next Generation Water Level Measuring Systems (NGWLMS) (Scherer and Beaumariage, 1988) had been installed to support those surveys. The general activities of the Galveston Bay Project were conducted between May and August, 1995 and were divided among several principles: Drs. G. Mader, R. Schmalz, and L. Huff and Mr. D. Martin of NOAA and Mr. S. DeLoach of USACE. The respective activities were: (1) conduct static GPS surveys to accurately tie the 3- dimensional position of the NGWLMS and the NOAA fixed GPS continuous operating reference station into the International Terrestrial Reference Frame (ITRS-93),via McDonald Observatory (MDO1); (2) develop and run a Galveston Bay tide model to estimate the ellipsoidal height of MLLW throughout the bay; (3) acquire and process CTD data, depth data, and GPS data from NOAA's roving and fixed GPS receivers; (4) acquire and process NGWLMS data to provide local MLLW and stage of tide, relative to the ellipsoid; and (5) deploy and operate a moored GPS instrumented buoy and an independent fixed GPS reference station, OTF process the associated GPS data, and determine MLLW at the USACE buoy. This paper addresses the precision and accuracy with which OTF GPS can provide a temporary vertical reference for hydrographic surveys.

Figure 3. Physical Layout of Galveston Bay, Texas

Figure 3 illustrates the physical layout of Galveston Bay, Texas. Of particular note are: (1) NGWLMS locations at Eagle Point, Trinity River Platform, Smith Point, and Port Bolivar; and (2) the triangle of survey lines which were repeatedly run by the NOAA Launch 518. The reader's attention is also called the location of Pier 21 in the Galveston Ship Channel, just south of Pelican Island.

Station Name	Station ID	MLLW (m)	MLLW Data Interval
Port Bolivar	877-1328	-28.868	6/94--5/95
Smith Point	877-0931	-29.046	6/95--8/95
Trinity River	877-1021	-28.997	6/95--7/95
Eagle Point	877-1013	-28.949	1/94--12/94
Pier 21	877-1450	-28.778	1/60--1/78

Table 1. Vertical Position of Tidal References in ITRS-88

Partial results of the first and fourth activities, listed above, appear in Table 1, which presents the ellipsoidal heights for MLLW at specific NGWLMS locations. In Table 1, there are differences of up to 25 cm in the ellipsoidal MLLW heights at the locations of these five NGWLMS. This indicates that careful attention must be given to establishing the field of ellipsoidal MLLW heights which need to be combined with the field of ellipsoidal bottom heights to provide a field of depths relative to MLLW.

Figure 4. Data Collection Systems

Figure 4 illustrates the suites of standard and project equipment used for data collection. The survey lines were planned based on requirements of the Galveston Bay Project. The launch and the standard suite of equipment were operated by personnel of the NOS Atlantic Field Party in accordance with NOS field procedures for cross-track allowance, sounding interval, vessel speed, and weather conditions. The launch's survey lines were broken at nominal 30-minute intervals to conduct CTD casts. The GPS and depth measurements associated with the casts were processed as two special cases. One case was where the ellipsoidal height of the launch GPS antenna and the ellipsoidal bottom height were devoid of the dynamic draft effects associated with higher launch speeds. The OTF GPS determined heights obtained as the launch speed decreased and increased to stop and resume a survey line provided the second special case which measured the launch's dynamic draft.

Figure 5. NOAA Survey Launch 518 Dynamic Draft

Figure 5 illustrates the mean and rms of the launch's dynamic draft as a function of the launch speed. Also plotted in the figure is the launch's quantized dynamic draft curve employed in processing the depth soundings via Hydrographic Data Acquisition and Processing System (HDAPS), NOS's standard system for single beam depth sounder data. (Brown and Wallace, 1988).In addition to the triangle of survey lines which were surveyed on eight different days over a 21-day period, the launch was also loosely moored one or more times at five of the NGWLMS sites for periods varying from one-half hour to one hour. The Pier 21 site was occupied more times and for longer periods than the sites at Eagle Point, Trinity River Platform, Smith Point, and Port Bolivar were occupied. The data at Pier 21 were intended to provide an accurate calibration of the distance from the phase center of the GPS antenna on the launch down to the free water surface. The tidal variations at Pier 21 are well removed in both phase and amplitude from tidal variations at the four NGWLMS sites close to the triangle of survey lines. Determination of the antenna height above the water can be accurately realized through this calibration procedure (Goldan, 1991).

Figure 6. Data Processing Flow Chart

Figure 6 illustrates the data processing associated with estimating the accuracy and precision of the OTF GPS measurements and determining the population statistics. Temporal smoothing was applied to OTF heights to be comparable to the 361-second smoothing of water levels internally applied by NGWLMS. The Ashtech PNAV (Anon.,1994) software was set to OTF process the GPS data sets in both forward and reverse time. It was observed, in approximately twelve percent of the 1-second epochs, that the differences between the OTF forward and reverse solutions exceeded 15 cm. They were flagged and not used in subsequent analyses. Some unrealistic depth data were also flagged and not used. Depths sampled between half-a-second before and half-a -second after each GPS one-second epoch were averaged, corrected for the actual sound speed of the water column, and associated with the average of the OTF forward and reverse positions for that one-second epoch.

Figure 7. Distribution of Depths Measured on the Repeat Survey Lines

Figure 7 illustrates the distribution of averaged, sound velocity corrected acoustic depths observed on eight repeat surveys of the triangle lines presented in Figure 3.

The precision of the ellipsoidal bottom heights measured during the Galveston Bay Project was estimated by comparing repeat measurements of the ellipsoidal bottom height at various positions along the triangle of survey lines. The precision analysis was limited to depths of 6 m and less.

Figure 8. Observed Variability between Repeated Surveys

Figure 8 presents the standard deviation of repeat measurements in various sized sampling circles, with and without OTF GPS vertical control, which were taken on any given run (same day) and on different runs (different days).

Figure 9. Distribution of Sample Population with Sampling Area

Figure 9 presents the number of repeat measurements within sampling circles of various sizes. The character of standard deviations of the same day measurements, at small radius sampling circles, appears to mirror variations in the number of samples. The standard deviations of the same day repeat depth measurements, with and without OTF GPS vertical control, are low. At these temporal and spatial scales, potential sources of differences between measurements on different days are essentially common to measurements on the same day and therefore could not contribute to the differences between measurements on the same day.

The standard deviations of the different day repeated depth measurements, with and without OTF GPS vertical control, show considerable differences. The potential sources of differences in depth measurements made on different days has elevated the standard deviations of the different day measurements without OTF GPS vertical control considerably above those of the different day measurements with OTF GPS vertical control. The projection of the different day repeat measurements with OTF GPS vertical control intersects the vertical at approximately 4.2 cm which represents the precision of the temporary OTF GPS vertical reference. In Galveston Bay, there was no independent ground truthing of depths as there was for each of the HYGRO campaigns. However correcting the launch OTF GPS heights for vertical offset of the GPS antenna above the water line, static draft, and dynamic draft of the launch does provide an estimate for the heights of the free water surface. Such height estimates, obtained in the vicinity of a NGWLMS, could be compared with the accurate heights of free water surface measured by that NGWLMS. The four NGWLMS, which were close to the triangle of repeated survey lines, were selected for use in estimating the accuracy with which OTF GPS could provide a temporary vertical reference for the control of hydrographic surveys.

An analysis, designated accuracy test, was setup which was limited to those data meeting the following conditions: (1) the launch speed was less that 0.5 m/s, (2) there were a minimum of 300 nonflagged OTF GPS launch heights in the 361-second period centered on the 6-minute NGWLMS, (3) the average position of the launch was within 2500 m of the comparison NGWLMS, and (4) the NGWLMS reported rms of the 361 1-second samples, which comprised its 6-minute reported mean water level, was less than 6 cm. This accuracy test discriminated against the more obvious limitations of estimating the OTF GPS accuracy using two independent water level measurements. Previous to computing the 361-second average OTF GPS height centered on each independent 6-minute NGWLMS, the 1-second epoch OTF GPS heights were individually adjusted for launch dynamic draft, according to speed. The tight restriction on launch speed essentially eliminated the influence of dynamic draft on the accuracy test. The tight restriction on distance from the NGWLMS greatly decreased the influence of spatial and temporal variability of tides in Galveston Bay on the accuracy test. The mean height error in this accuracy test was 0.07 cm with a standard deviation of 4.7 cm. Due to the restrictive selection of the data sets, this mean error and standard deviation are attributed totally to the OTF GPS.

An intercomparison analysis was conducted that was similar to the accuracy test except for the following: launch speeds less than 15 m/s, 5000 m search radius and NGWLMS rms less than 30 cm. In an intercomparison such as this, the statistical results must be interpreted in light of the several potential independent contributors to the differences. In this intercomparison, potential contributors to the differences were: (1) the spatial and temporal variability of tides at increased distance from any particular NGWLMS location; (2) the speed dependent dynamic draft correctors; (3) the OTF GPS height measurements; and (4) the actual variations of tides at each of the NGWLMS. The rms of data from the individual NGWLMS used in the intercomparison was 9.1 cm. The precision of the OTF GPS heights, as determined from the repeatability of the ellipsoidal bottom height measurements, was 4.2 cm. The mean height difference in the intercomparison was 2.13 cm with a standard deviation of 9.01 cm. Given the potential sources of differences in the intercomparison, this mean and standard deviation are consistent with an accuracy on the order of 5 cm for the OTF GPS heights, as was determined via the accuracy test previously discussed.

Conclusions

The Galveston Bay Project has shown that OTF GPS can provide a temporary vertical reference for hydrographic surveys with an accuracy on the order of 5 cm and a precision on the order of 4 cm.

Future Activities

Additional understanding is required in three particular areas before this application of OTF GPS can be uniformly employed in practice. The first important issue is how to best establish the ellipsoidal MLLW heights throughout a survey area and it has already been addressed in a portion of the Galveston Bay Project not discussed in this brief paper. The second technical area requiring better understanding is the cause and remedy for the short-term discrepancies observed between the OTF GPS heights obtained from forward and reverse OTF processing the GPS data. The third of those areas is to determine how to correctly match spectral variations in OTF GPS heights measured at 1-second intervals with those in vessel heave measurements taken at 50-msec intervals with a motion reference unit. When that is achieved, it will truly be possible to provide a valid estimate of sounding transducer ellipsoidal heights from very long periods up to short periods, approaching the measurement intervals of shallow water acoustic depth sounders.

Acknowledgements

The Galveston Bay Project relied heavily on the personnel and officers of the NOAA Ship HECK and the NOAA Launch 518 during the field work. Their dedication and professional efforts were essential to the overall success of this project. It is recognized that the collective contributions from NOS and USACE to the partnership on this project greatly exceed those specifically described in this paper. Furthermore, the scope and significance of the total results from the Galveston Bay Project extends far beyond the scope of this paper.

References

Anon, (1994), "Precise Differential GPS Navigation - PNAV Trajectory," Software User's Manual, Ashtech, Inc., Sunnyvale, CA, p420.

Brown, M. B., J. L. Wallace, (1988), "Hydrographic Data Acquisition and Processing System (HDAPS), Proceedings of the Third Biennial National Ocean Service International Hydrographic Conference, Baltimore, MD, April 12-15, pp117-120.

DeLoach, S., et al., (1994), "Delineation of Tidal Datums and Water Surface Slopes with the GPS," Proceedings of the Sixth Biennial National Ocean Service International Hydrographic Conference, Norfolk, VA, April 18-23, pp214-221.

Godin, A. B., et al., (1992), "Simrad EM1000 and Swath Vessel Technology : The Perfect Match?" Proceedings of the International Hydrographic Conference HYDRO'92, 30 Nov-3 Dec., Copenhagen, Denmark.

Goldan, H. J., (1991), "Tidal Observations with Kinematic GPS on Ships," Proceedings from the First International Symposium Real Time Differential Applications of the Global Positioning System, Deutche Gesellschaft fur Ortung und Navigation, September 1991, Braunscheig, Germany, pp585-594.

Mayer, L. A., et al., (1993), "A Multi-faceted Acoustics Groundtruth Experiment in the Bay of Fundy," Acoustic Classification and Mapping of the Sea Floor, Proceedings of the Institute of Acoustics, Vol. 15, part 2, pp203-220.

Remondi, B. W., (1992), "Real-Time Centimeter-Accuracy GPS for Marine Applications," Proceedings of the Fifth Biennial National Ocean Service International Hydrographic Conference, Baltimore, MD, February 25-28, pp19-23.

Scherer, W. D., D. C. Beaumariage, (1988), "NGWLMS APPLICATIONS: From Hydrography to Absolute Global Sea Level Monitoring," Proceedings of the Third Biennial National Ocean Service International Hydrographic Conference, Baltimore, MD, April 12-15, pp117-120.

Top of page