Monitoring Tides with GNSS Buoys in Open Sea Areas

Weikang Sun

1,2

, Xinghua Zhou

1,2,*

, Yanguang Fu

1,2

, Zhaoyang Wang

and Dongxu Zhou

College of Geomatics, Shandong University of Science and Technology, Qingdao, 266590, China;

First Institute of Oceanography, State Oceanic Administration, Qingdao, 266061, China.

National Marine Information Center,Tianjin, 300171, China

Email: xhzhou@fio.org.cn

Keywords: GNSS buoys, open sea, Kalman filters, prior information, tide model

Abstract: Nowadays, the coastal tide observation technology has developed very well. However, there still lacks a

reliable and quick method of acquiring tidal level in open sea areas. To overcome this problem, a GNSS

buoy has been developed, which is equipped with data acquisition system and shore station system. Also,

some tidal level extraction algorithms are also proposed. Based on Precision Point Positioning, the GNSS

data is calculated. Kalman filtering, attitude correction, smoothing, and sampling are performed on the high-

frequency GNSS solution results. And the above results are corrected using the prior periodic information.

The tidal level derived from the GNSS buoy without and with prior periodic information are compared with

the tidal level derived from the regional tide model (NAO.99Jb). The RMSE are 7.6cm and 4.8cm,

respectively. Also, the tidal level derived from the GNSS buoy with prior periodic information in spring tide

period are less accurate than it in neap tide period. The RMSE are 6.4cm and 3.7cm, respectively. The

results show that GNSS buoy can provide high accuracy tidal level.

1 INTRODUCTION

At present, marine observation buoys have

developed rapidly and maturely, including wave

buoys, tsunami buoys (Kato et al., 2008) and so on.

GNSS buoys are a new type of marine surveying

instrument that has emerged with the rapid

development of GNSS satellite positioning

technology since the early 1990s. It can obtain high

precision tidal level and carry out the correction of

the sounding in the offshore areas. It is of great

significance for the construction of the channel and

the safety of the navigation. The measurement

principle of GNSS buoys is to obtain the precise

spatial geocentric coordinates of the phase center of

the GNSS antenna through GNSS Real Time

Kinematics(RTK), post-processing kinematics(PPK)

or precise point positioning (PPP). Then combined

with the measured height of the center of the

antenna's phase relative to the water surface, a time-

sampling sequence of the height from the sea surface

to the reference ellipsoid can be obtained. The

GNSS buoy measurement principle is shown in

Figure 1 (Bisnath, 2004):

Figure 1: The principle of GNSS buoy measurement.

When using the GNSS buoy to observe the water

level, the GNSS buoy elevation observation data has

a lot of noise due to the influence of the GNSS

observation error, the sea wave, the change of the

buoy attitude and so on (Bisnath, 2004). At present,

the GNSS observation error can be effectively

eliminated by using a perfect model and method

Sun, W., Zhou, X., Fu, Y., Wang, Z. and Zhou, D.

Monitoring Tides with GNSS Buoys in Open Sea Areas.

In Proceedings of the International Workshop on Environment and Geoscience (IWEG 2018), pages 105-109

ISBN: 978-989-758-342-1

105

(Hein et al., 1992; Kelecy et al., 2011), and the error

of the buoy attitude change can be eliminated by the

attitude correction. Therefore, eliminating the

influence of ocean waves in GNSS buoy height data

is particularly important. In distant sea areas,

changes in the attitude of the carrier caused by wind

waves will bring greater errors to the tide measuring

(Yang and Zhao, 2003; Zhao et al., 2006). The

attitude correction technology can reduce the error

effect caused by severe shaking of the carrier to the

tide inspection. The PPK or PPP mode is not limited

by radio signal transmission and can further expand

the tide measuring range. Therefore, PPK or PPP

mode can be considered for tide measuring

combined with attitude parameters in the offshore

areas.

Aiming at these problems, a combined GNSS

buoy was developed and a new algorithm for

extracting tidal information was proposed.

2 BUOY DEVELOPMENT

The GNSS buoy is mainly composed of two parts:

data acquisition system and shore station system.

The data acquisition system includes an

embedded data collector, a high-precision attitude

sensor, a high-accuracy GNSS module, a Beidou

transmit signal module, a power supply voltage

regulator module, and a corresponding software

system. The system can be operated independently

to collect raw observations of GNSS tidal level

measurements. The GNSS receiver is an

independently developed multimode and multi

frequency data acquisition device with the

NOV703GGG.R2 antenna. The TCM2.5 attitude

instrument is integrated inside the system and is

used to measure the instantaneous roll, pitch and

heave, providing attitude correction parameters for

the GNSS tidal data. The data collector of the buoy

is installed in the buoy cabin, and the GNSS antenna

and Beidou satellite terminal are installed on the

bracket at the buoy platform.

The shore station system includes the Beidou

reception signal module, data monitoring module,

and data preprocessing module. The system is used

to monitor the operational status of the buoy, such as

buoy position information, attitude information and

the working status of each module.

3 INTRODUCTION OF OPEN SEA

EXPERIMENT

Before the GNSS buoy was placed to the measuring

areas, the elevation of the antenna should be

calculated first.

A dynamic calibration method of measuring GPS

buoy antenna elevation based on GPS PPK and

synchronous tide observation data was adopted

(Zhou et al., 2015). After installing and debugging

the GNSS buoy system, the base station was set up

at the wharf, which was about 20m from the GNSS

buoy. At the same time, a pressure tide gauge was

installed at the same site with the base station to

observe the tidal level. Combined with three CORS

stations of Shipu, Dongshan and Kanmen in the

coastal GNSS observation network, the three-

dimensional coordinates of the base station under

the framework of ITRF2008 were obtained. Using

the synchronous observation data from the base

station, the dynamic coordinates of the GNSS buoy

antenna phase center are solved one epoch by one

epoch with the using of the Differential method. The

tidal level recorded by the tide gauge was converted

to geodetic height. The GNSS buoy's dynamic

antenna elevation was processed by a moving

average method, and the tidal level of geodetic

height recorded by the tide gauge was subtracted.

Then the GNSS buoy dynamic antenna elevation can

be obtained.

After completing the buoy antenna elevation

measurement, the buoy was placed in the East China

Sea for continuous observation of tidal level (shown

in Figure.2).

Figure 2: The distribution of reference stations and GNSS

buoy.

IWEG 2018 - International Workshop on Environment and Geoscience

106

4 DATA PROCESSING

4.1 GNSS Data Processing Based on

PPP Technology

The PPP location technology does not require a

reference station. Single epoch-precise single point

positioning is performed after initialization with the

using of the non-differential dual-frequency carrier

phase observations. Using post-processing mode, the

precise ephemeris and clock difference files

provided by IGS were selected, and then the phase

or pseudo-range observations collected by a single

receiver were subjected to non-differential

positioning processing.

4.2 Tidal Level Extraction

We developed a three steps procedure to extract the

actual tidal level from the GNSS data and to

subsample the data from 1sec intervals to 1min

interval. The data processing of the GNSS buoy is

performed by PPP technology, and the series of the

antenna phase center elevation is obtained.

4.2.1 Kalman Filtering

The best estimation criterion of the Kalman filtering

is the minimum mean square error, which requires

that the optimal estimation of the signal or state

should have the smallest error variance from the

corresponding true value. The basic idea is to use

the state space model of signal and noise, update the

estimation of the state variables with the estimates

of the previous time and the observations at the

current moment, and find the estimated value at the

moment of occurrence. Kalman filtering is an

efficient recursive filter that can estimate the state of

a dynamic system from a series of measurements

that do not completely contain noise.

4.2.2 Intermediate Data Subsampling and

Smoothing

In this step, the GPS raw data is subsampled at

1minute intervals. From this subsampled series,

records are smoothed by a 60minute average

window (Apel et al., 2012). The smoothing is

performed bi-directionally over the time series in

order to prevent phase shifts.

4.2.3 GNSS Buoy Data Processing with

Prior Periodic Information

With the rapid development of modern geodetic

technology, measuring methods are becoming more

and more advanced, and geodetic data has become

more and more abundant. Our understanding of the

geometric, physical and mechanical properties of

any observing target or object is also becoming

more and more sufficient, and the possibility of

establishing periodic based on prior information is

increasing. In the technique of extracting tidal

information from GNSS buoys, we know the period

information of each constituent, instantaneous tidal

level can be expressed as (Huang and Huang, 2005):

( ) ( cos sin )+ ( )

j j j j

t A t B t x t

  







(1)

Where,

()t



represent GNSS buoy-derived time

series;

is the number of constituents;



is the

angular rate of constituents;

is time;

()xt

is non-

astronomical tidal level and

AB、

are coefficient

matrixes waiting to be solved.

Taking into account the time length of

()t



, the

appropriate constituents should be selected and the

coefficient matrix of

AB、

can be solved.

()xt

can be obtained by polynomial fitting. Through the

obtained coefficient matrix

AB、

, non-

astronomical tidal level

()xt

, and known

constituents periodic information, the new

()t



can

be obtained, which is the tidal level extracted from

the GNSS buoy after corrected by the prior periodic

information. The tidal level terms are long-period

items, the period is generally several hours to ten

hours; the surge-wave items are short-wave items,

and the period is generally tens of seconds to tens of

minutes. The influence of short period surge-wave

can be eliminated by selecting the appropriate tide.

5 RESULTS AND DISCUSSION

The PPP calculation results for a single day are

shown in Figure 3. 7 April was chosen, as it was a

representative sample of the entire dataset. The data

set is referenced to WGS84 ellipsoid.

Monitoring Tides with GNSS Buoys in Open Sea Areas

107

As suggested in Figure 3, the processing filter

noise was a high frequency signal riding upon the

lower frequency water level changes. After Kalman

filtering, most of the gross errors were eliminated, as

shown in Figure 4.

Figure 3: GNSS buoy height solutions for 7 April 2016.

Figure 4: GNSS buoy height solutions and filtered data for

7 April 2016.

Figure 5: Subsampled and smoothed GNSS buoy height

solutions for 7 April 2016.

Figure 6: GNSS buoy height solutions verses NAO.99Jb

for the entire month of April.

The data in April was processed day-by-day

according to the above steps, and the overall results

for April were obtained. Then, the prior information

was used to correct the overall results in April. The

reference datum was transformed from WGS84

ellipsoid to monthly mean sea level. Compare the

two results with NAO.99Jb, as shown in Figure 6.

Since the GNSS buoy was far away fro m the

mainland and there was no tide gauge station in the

surrounding area, the NAO.99Jb tidal model was

selected to verify the accuracy of GNSS buoys for

measuring tidal level. The NAO.99Jb tidal model is

a regional tidal model constructed by the Japanese

National Observatory based on two-dimensional

nonlinear equations using the Blending assimilation

method (Matsumoto et al., 2000). Compared with

other global models, NAO.99Jb has the highest

accuracy in the layout areas of the buoy (Fu et al.,

2017) .

Figure 6 shows that the tide data without prior

information periodic are less accurate than the tide

data with prior information periodic compared with

the NAO.99Jb. The RMSE is 7.6cm and 4.8cm,

respectively. The RMSE is defined as (Lin et al.,

2017):

 

RMSE y y







(2)

where,

and

represent the tides obtained

from the regional tide model NAO.99Jb and the

GNSS buoy, respectively, and N is the number of

tide data.

It is proved that the tidal level correction based

on priori periodic information has a great

contribution to improving the accuracy of GNSS

buoys for extracting tidal information. Notably, the

amplitude of the GNSS buoy height solutions is

smaller than it of NAO.99Jb. A possible reason is

the inclination of the GNSS buoy (Lin et al., 2017).

Another possible reason is that the bathymetry

model that the NAO.99Jb used are not accuracy

enough, which affect the tidal level calculation.

Figure 7: GNSS buoy height solutions with prior periodic

information verses NAO.99Jb for the spring tide.

IWEG 2018 - International Workshop on Environment and Geoscience

108

Figure 8: GNSS buoy height solutions with prior periodic

information verses NAO.99Jb for the neap tide.

As can be seen from Figure 7 and Figure 8, the

tide period measured by the GNSS buoy is in good

agreement with the model. The accuracy in spring

tide period is lower than it in neap tide period. The

RMSE is 6.4cm and 3.7cm respectively.

6 CONCLUSIONS

Using GNSS technology to measure tidal level is a

new measurement method. This article introduces in

detail the measurement of tidal level in coastal areas

and open sea with the using of the modular GNSS

buoy.

By making full use of the precise three-

dimensional dynamic measurement technology of

precise point positioning, it is possible to measure

tidal level in areas where traditional methods are

difficult to perform and in open sea areas. Through a

series of processing to the elevation of the antenna,

such as Kalman filtering, attitude correction,

smoothing, subsampling and adding prior periodic

information correction, the time series of tidal level

with the accuracy of cm level can be obtained.

The open sea GNSS buoys have high

observation accuracy, and the deployment is not

limited by sea and land conditions. If long-term

effective observations are carried out, reliable tidal

level information can be obtained. On the one hand,

it is possible to supplement the tidal data with the

construction of a global tidal model to improve its

construction accuracy. On the other hand, the GNSS

buoy can be used to calibrate the altimeter directly at

the sub-satellite point. This method is not affected

by tidal models and geoid and it and is widely used

in the study of altimeter calibration (Bonnefond et

al., 2003).

The prospect of the application of buoys is to

realize the extraction of real-time tidal level, further

investigation is required to resolve the discrepancies

between the real-time and post-processed solutions.

We should give full play to the applicability and

expandability of GNSS buoy technology in marine

surveillance in the near future.

REFERENCES

Apel H, Hung N G, Thoss H, et al 2012 GPS buoys for

stage monitoring of large rivers Journal of Hydrology

412 182-192

Bisnath S 2004 Development of an Operational RTK

GPS-equipped Buoy for Tidal Datum Determination

International Hydrographic Review 54-64

Bonnefond P, Exertier P, Laurain O, et al 2003 Leveling

the Sea Surface Using a GPS-Catamaran Special Issue

Jason-1 Calibration/Validation Marine Geodesy 26

319-334

Fu Y G, Zhou X H, Zhou D X, et al 2017 Accuracy

analysis of ocean tidal model over China seas based on

the gauge data Science of Surveying and Mapping 42

28-32

Hein G W, Blomenhofer H, Landau H, et al 1992

Measuring Sea Level Changes Using Gps in Buoys Sea

Level Changes Determination and Effects American

Geophysical Union 101-105

Huang Z K, Huang L 2005 Tidal Principles and

Calculations

Kato T, Terada Y, Nagai T, et al 2008 Development of a

new Tsunami Monitoring System Using a GPS Buoy

AGU Fall Meeting

Kelecy T M, Born G H, Parke M E, et al 2011 Precise

mean sea level measurements using the Global

Positioning System Journal of Geophysical Research

Oceans 99 7951-7959

Lin Y P, Huang C J, Chen S H, et al 2017 Development of

a GNSS Buoy for Monitoring Water Surface

Elevations in Estuaries and Coastal Areas Sensors 17

1-19

Matsumoto K, Takanezawa T, Ooe M 2000 Ocean Tide

Models Developed by Assimilating

TOPEX/POSEIDON Altimeter Data into

Hydrodynamical Model A Global Model and a

Regional Model around Japan Journal of

Oceanography 56 567-581

Yang F L, Zhao J H 2003 Analyzing and Eliminating the

Effect of Wave in GPS Tide Observing Hydrographic

Surveying and Charting 23 1-4

Zhao J H, Zhang H M, Hughes-Clarke J E 2006

Determination of Precise Instantaneous Tidal Level at

Vessel Geomatics and Information Science of Wuhan

University 31 1067-1070

Zhou D X, Zhou X H, Liang G H, et al 2015 Research on

dynamic calibration method of GPS buoy Science of

Surveying and Mapping 40 121-124

Monitoring Tides with GNSS Buoys in Open Sea Areas

109