An innovative geodetic tool called GNSS-IR is a more effective technique than tide gauge for monitoring sea surface height as the basis for modelling sea level variations, according to the study outlined in this article.

The utilization of remote sensing observations to monitor essential climate variables (ECVs) has become increasingly important in studying their regional and global impacts, as defined by the Global Climate Observing System (GCOS). Understanding the Earth’s surface conditions, including soil moisture runoff, snow, temperature, precipitation, water vapour, radiation, groundwater and sea surface height (SSH), can positively impact the environment and ecosystems. Here, the authors present an overview of how global navigation satellite systems (GNSS) can be employed for environmental monitoring, with a particular focus on sea surface height monitoring. This includes examination of the advantages and disadvantages of utilizing a network of permanent GNSS stations for monitoring sea level rise along shorelines.

Monitoring sea level rise is crucial to understanding and preparing for the potential impacts of climate change, such as flooding, erosion and saltwater intrusion in coastal areas. It can also affect global ocean circulation patterns and climate. Scientists can provide helpful information to policymakers and stakeholders by monitoring sea level change for informed decisions about land use, infrastructure development and emergency preparedness. Different techniques and sensors can be used individually or in combination to provide a complete picture of sea level changes, including tide gauge stations, satellite altimetry missions, satellite gravimetry techniques, GNSS stations, ocean buoys and acoustic sensors. One of the important applications of SSH monitoring is the study and analysis of tidal frequencies in the context of predicting tides, modelling sea currents, and harbour planning (e.g. breakwater design).

GNSS interferometric reflectometry

GNSS is primarily designed for precise positioning, timing and navigation applications. For precise positioning in geodesy, noises and systematic errors affecting the precise positioning are filtered out, e.g. for establishing reference frames and geodetic networks. Hence, the error sources from the atmosphere (due to the tropospheric and ionospheric delays), satellite and receiver clocks, satellite orbit data errors, and multipath error (Teunissen and Montenbruck 2017) are the main issues that should be solved for precise positioning applications. For example, the multipath error occurs when the transmitted signals from GNSS satellites are reflected from the receiver’s surrounding areas. These are unwelcome signals for precise positioning and are considered as noises in geodesy (See Figure 1). Therefore, the reflected (indirect) signals should be cancelled out, e.g. by increasing the cut-off-angle and using precise antenna models equipped with a chock ring plate (a metallic plate that can be mounted under the antenna). The reflected signals increase the uncertainty of the positioning. However, the reflected signals can be used for other applications, e.g. environmental monitoring. In fact, the noises contain an imprint of the Earth’s surface conditions. Since the mid-1990s, the reflected signals have been used by scholars in different fields. These signals can be used to detect snow depth, ice thickness, vegetation growth, soil moisture and sea level change by employing the GNSS reflectometry (GNSS-IR) technique.

Martin-Neira (1993) and Larsen et al (2008) are two pioneer scientists who studied oceanic surface and soil moisture, respectively, using the Global Positioning System (GPS) signals reflected from the surface for the first time. The reflected signals can be collected using terrestrial techniques (geodetic ground stations) or spaceborne techniques (using receivers mounted on satellites). For instance, the reflected signals can be recorded using onboard spaceborne platforms, e.g. UK-DMC and TechDemoSat-1, which were successfully launched in 2003 and 2014 respectively and used for GNSS-IR applications. In addition, eight microsatellites fully dedicated to GNSS-R for ocean surface and wind monitoring were launched in 2016 by NASA Cyclone GNSS (CYGNSS). And in 2021, a cube satellite (10x10x30cm³) was designed for GNSS-IR applications by the Passive REflecTomeTrY (PRETTY) project. This Austrian consortium, led by RUAG GmbH, relies on results from the former ESA OPS-SAT mission, conducted by TU Graz.

Sea surface height determination using GNSS-IR

Using GNSS-IR, one can estimate the absolute vertical distance from a GNSS antenna to the reflective surface (See Figure 1). This allows the sea level changes to be studied using the GNSS-IR technique. The main advantage of GNSS-IR concerning other methods, e.g. tide gauge stations, is that the GNSS can be used as a multipurpose system. The tide gauge stations relatively record the SSH with respect to the land adjacent to the station. Therefore, a precise vertical land motion model (land uplift or subsidence model) is required for correcting the tide gauge data. For example, the largest land uplift rate (due to the post-glacial rebound) can be seen in Fennoscandia and Laurentia. However, GNSS can be used for land motion monitoring in combination with GNSS-IR to monitor the SSH.

The other advantage of using GNSS reflectometry instead of tide gauge for sea level monitoring is that GNSS-IR can provide sea level information over a wider area. Moreover, the tide gauges are often installed in harsh environments, e.g. piers, making them vulnerable to damage from storms or extreme weather events. GNSS-IR is less vulnerable to such damage because the receivers are usually located on stable land. Furthermore, the maintenance of tide gauges can be expensive due to the need for regular calibration. In contrast, GNSS-IR needs less maintenance because the equipment is generally more robust. It is also worth mentioning that satellite altimetry accuracy around the coastlines considerably decreases due to the imperfect reflection of the radar waveforms in the shallow areas (i.e. radar signal reflections from land).

The height parameter of the GNSS antenna to the reflected surface (e.g. sea level) can be obtained using the interference between direct and reflected signals. There are two methods for this purpose: carrier phase, and signal-to-noise ratio (SNR). In the SNR method, the main observation is SNR data collected by the receiver. The SNR measures the signal strength received by a GNSS antenna (i.e. composite signal power and noise power ratio calculated in GNSS receivers). The SNR data is the main input in the GNSS-IR technique to extract the SSH using zenith-looking antenna. It is important to check and ensure the data shows up in the RINEX file (Larson et al).

Challenges associated with GNSS-IR

The main issue in the GNSS-IR method is determining the predominant reflected/multipath frequency in SNR data using different spectral analysis techniques (sophisticated signal processing algorithms) such as Wavelet decomposition, Lomb-Scargle Periodograms (LSP), least squares harmonic estimation, etc. These methods help to find the multipath frequencies in the frequency domain. The relation between the predominant reflected frequency (f_M) from the desired surface (e.g. sea level) and the height of the antenna (h) is expressed in the following formula (in which λ is the GNSS signal wavelength): 

Different GNSS carrier phases (signals) can be used for this purpose (see Figure 1). However, the main question is, which GNSS carrier phase signal provides better results for SSH? This question still needs to be investigated to obtain suitable uncertainty for SSH determination using GNSS. One of the challenges associated with using GNSS-IR is that the GNSS signal strength reflected off the ocean surface is typically weak, making detection and accurate measurement difficult. In addition, the signals from GNSS satellites can reflect off several surfaces before reaching the receiver, causing interference and measurement errors (called multipath interference). Moreover, the weather conditions such as rain, fog and cloud cover can affect the strength and quality of GNSS signals, making it more tricky to achieve precise measurements. Also, parameters such as satellite geometry, antenna patterns and internal receiver noises can affect the accuracy and quality of GNSS-IR results.

A toolbox for GNSS-IR

A free Matlab toolbox called GNSS-IR-UT has been developed for SSH studies using the GNSS-IR method (Farzaneh et al. 2021). The main advantages of this toolbox with respect to the other code packages are: The existing code packages are not user-friendly. For example, the code packages are Python-based without any user interface, making it difficult for new users, especially those unfamiliar with Python coding. To run the codes, SNR data should be extracted and prepared from RINEX and SP3 (satellite orbit data) files.

The GNSS-IR-UT toolbox involves the following three steps:

Step 1: The toolbox reads the GNSS data stored in the RINEX file and automatically extracts the SNR of various signals.

Step 2: GNSS-IR-UT estimates the dominant frequencies using spectral analysis methods.

Step 3: At the end of the processing stage, it determines the antenna height relative to the reflecting surface for each epoch (observation time). The user can fully control different parameters and tune them easily to obtain the desired result. For example, the user can define which GNSS system should be used for the processing. In addition, different spectral analysis techniques can be applied to determine the predominant reflected/multipath frequency.

Assessment using tide gauge stations

The obtained SSH time series can be assessed using other SSH monitoring methods e.g. tide gauge stations. In a comparison study, the hourly and daily data results were compared over three months using GNSS (station ID: MERS) and tide gauge stations in Erdemli, Turkey. The GNSS station’s distance from the tide gauge station was about 524m (at 36.563737 latitude, 34.255305 longitude). The GNSS receiver and antenna types were Leica GR50 and Leica AR10, respectively. The GNSS receiver could collect different GNSS carrier phases, i.e. GPS: L1/L2/L5, Galileo: L1/L5/L7/L8, GLONASS: L1/L2, BeiDou: L2/L7.

The results (see Figure 2) show that the high temporal and spatial resolution measurements of sea level altitude can be made using GNSS signals that are reflected from the water. So which GNSS signal provides better SSH results? The results showed that the SSH measured by the L1 signal of GPS, GLONASS and Galileo satellites provided a better and more accurate solution than other GNSS systems systems (Gholamrezaee et al. 2023). However, the BeiDou L2 signal provided a better outcome for the SSH. As can be seen in Figure 2, some of the GNSS signals showed poor SSH retrieval performance, so the root mean square errors (RMSEs) are larger than 1m. In other words, the large RMSEs are due to the signal structure and the antenna and receiver tracking method. For example, the GPS signals originated from the same oscillator; hence GPS L2 signal is contaminated by the GPS L1 signal. There are two tracking methods, i.e. Z-tracking and squaring methods, to track the signals (e.g. GPS L2P signal). The performance of the Z-tracking method depends on the precision of the W code obtained from L1. The squaring method also depends on squaring the L2 signal and bandpass filter, which can cause a loss of 20-30dBHz and weaken the SNR. Therefore, the SSH retrieval performance using the GPS L2 signal will be weak (cf. Wang et al. 2021). A similar comparison was performed using different stations, i.e. GTGU (Onsala Space Observatory, Sweden), MCHN (Michipicoten Harbor, Ontario, Canada), MARS (Marseille, France), TGDE (Tregde, Norway) and AT01 (St Michael, Alaska, USA), and obtained similar results.

Conclusion

In conclusion, while GNSS-IR and tide gauge are both useful techniques for monitoring sea level height, GNSS-IR offers several advantages. GNSS-IR can provide more accurate and continuous measurements of sea level height, even in areas with complex topography or limited access. Additionally, GNSS-IR provides complementary information on other geophysical parameters, such as soil moisture, vegetation and snow depth, particularly in areas where traditional monitoring methods are difficult or impractical. Therefore, considering the challenges and limitations of tide gauge measurements, it can be concluded that GNSS-IR is a more suitable technique for sea level height monitoring.

Further reading

Farzaneh, S., Parvazi, K., & Shali, H. H. (2021). GNSS-IR-UT: a MATLAB-based software for SNR-based GNSS interferometric reflectometry (GNSS-IR) analysis. Earth Science Informatics, 14(3), 1633-1645.

Gholamrezaee, S., Bagherbandi, M., Parvazi, K., & Farzaneh, S. (2023). A study on the quality of GNSS signals for extracting the sea level height and tidal frequencies utilizing the GNSS-IR approach. GPS Solutions, 27(2), 72.

Larson KM, Small EE, Gutmann E, Bilich A, Axelrad P, Braun J (2008) Using GPS multipath to measure soil moisture fluctuations: initial results. GPS Solut 12(3):173–177 Martin-Neira M (1993) A passive reflectometry and interferometry system (PARIS): Application to ocean altimetry. ESA J 17(4):331–355

Teunissen, P. J., & Montenbruck, O. (Eds.). (2017). Springer handbook of global navigation satellite systems (Vol. 10, pp. 978-3). Cham, Switzerland: Springer International Publishing.

Wang, X., He, X., Xiao, R., Song, M., & Jia, D. (2021). Millimeter to centimeter scale precision water-level monitoring using GNSS reflectometry: Application to the South-to-North Water Diversion Project, China. Remote Sensing of Environment, 265, 112645.