EOT20: A global ocean tide model from multi-mission satellite altimetry

EOT20 is the latest in a series of empirical ocean tide (EOT) models derived using residual tidal analysis of multi-mission satellite altimetry at DGFI-TUM. The amplitudes and phases of seventeen tidal constituents are provided on a global 0.125-degree grid based on empirical analysis of seven satellite altimetry missions and four extended missions. The EOT20 model shows significant improvements compared to the previous iteration of the global model (EOT11a) throughout the ocean, particularly in the coastal and shelf regions, due to the inclusion of more recent satellite altimetry data as well as more missions, the use of the updated FES2014 tidal model as a reference to estimated residual signals, the inclusion of the ALES retracker and improved coastal representation. In the validation of EOT20 using tide gauges and ocean bottom pressure data, these improvements in the model compared to EOT11a are highlighted with the root-square sum (RSS) of the eight major tidal constituents improving by ~3 cm for the entire global ocean with the major improvement in RSS (~3.5 cm) occurring in the coastal region. Concerning the other global ocean tidal models, EOT20 shows an improvement of ~0.2 cm in RSS compared to the closest model (FES2014) in the global ocean. Variance reduction analysis was conducted comparing the results of EOT20 with FES2014 and EOT11a using the Jason-2, Jason-3 and SARAL satellite altimetry missions. From this analysis, EOT20 showed a variance reduction for all three satellite altimetry missions with the biggest improvement in variance occurring in the coastal region. These significant improvements, particularly in the coastal region, provides encouragement for the use of the EOT20 model as a tidal correction for satellite altimetry in sea-level research. All ocean and load tide data from the model can be freely accessed at https://doi.org/10.17882/79489 (Hart-Davis et al., 2021).

Polar motion resonance in the prograde diurnal band

Summary Until now, the polar motion resonance (PMR) complex frequency has been determined in the seasonal and retrograde diurnal band of the polar motion. In this study this resonance is studied in the prograde diurnal band, where polar motion is mainly composed of periodic terms caused by the diurnal oceanic tide. The resonance parameters (period and quality factor) are encompassed in the frequency transfer function between generating tidal potential and polar motion, and can be estimated accordingly. To this aim, we gather three published sets of prograde diurnal terms determined from GNSS and VLBI, to which we append our own estimates based upon a processing of the VLBI delays over the period 1990-2020. Then, by fitting the PMR parameters so that the prograde diurnal terms match the corresponding components of the tide generating potential, we obtained a resonance period of about 401 days and an equivalent quality factor of −22, differing from the ones reigning in the seasonal band (PPMR ≈ 431 days; QPMR ≈ 56 − 255) and in the retrograde diurnal band (PPMR ≈ 380 days; QPMR ≈ −10). Our estimates confirm strikingly the theoretical prediction derived from the tidal ocean angular momentum derived from the FES 2014 ocean tide model.

Impact of the stochastic model of the ocean tides on GRACE monthly gravity field solutions

<p><span>I</span><span>n </span><span>GRACE data </span><span>processing</span><span> </span><span>t</span><span>he geophysical </span><span>background </span><span>models, which are needed to compute </span><span>the </span><span>monthly gravity field solutions, </span><span>usually </span><span>e</span><span>nter as</span><span> error-free. </span><span>This</span><span> </span><span>means that model errors could influence and distort the gravity field solution</span><span>.</span></p><p><span>The geophysical models </span><span>which influence the solution the most</span><span> a</span><span>re</span><span> the </span><span>atmosphere and ocean dealiasing product (AOD1B) and the ocean tide model. </span><span>In this presentation we focus on the </span><span>ocean tide model and on incorporati</span><span>ng</span><span> </span><span>its </span><span>stochastic information </span><span>in data processing</span><span>. </span></p><p><span>We use </span><span>the FES2014 ocean tide model presented as a spherical harmonic expansion till degree and order 180. The information about its uncertainties and the correlations between different spherical harmonics is provided by the research unit NEROGRAV (New Refined Observations of Climate Change from Spaceborne Gravity Missions). In a first step, the stochastic properties of the tide model are considered to be static and are expressed as variance-covariance matrices (VCM) of the spherical harmonics of the 8 main tidal waves till degree and order 30. The incorporation of this stochastic information is done by setting up the respective ocean tide harmonics as parameters to be solved for. Since ocean tides cannot be freely estimated within monthly GRACE solutions, the provided VCMs for the 8 tidal waves are used for constraining the tidal parameters.</span></p><p><span>T</span><span>his procedure was used to compute monthly gravity field solutions for the year 2007. For a comparison, we computed also monthly gravity fields without taking into account the stochastic information on ocean tides. In this contibution we present and discuss the first results of this comparison.</span></p>

A new high resolution Mean Sea Surface (DTU21MSS) for improved sea level monitoring

<p>A new Mean Sea Surface (DTU21MSS) for referencing sea level anomalies from satellite altimetry is presented. The major new advance leading up to the release of this MSS the use of 5 years of Sentinel-3A and an improved 10 years Cryosat-2 LRM+SAR+SARin record including retracked altimetry in Polar regions using the SAMOSA+ physical retracker via the ESA GPOD facility.</p><p>A new processing chain with updated editing and data filtering has been implemented. The filtering implies, that the 20Hz sea surface height data are filtered using the Parks-McClellan filter to derive 1Hz. This has a clear advantage over the 1 Hz boxcar filter in not introducing sidelobes degrading the MSS in the 10-40 km wavelength band. Similarly, the use of consistent ocean tide model for the Mean sea surface improves the usage of sun-syncronous satellites in high latitudes.</p><p>The presentation will also focus on the difficult issues to consolidating Cryosat-2 and Sentinel-3 onto a past 20 year mean sea surface. This is implemented using simultaneous estimation of the mean, sea level trend and annual and semi-annual variations in sea level.  </p>

EOT20: A new global empirical ocean tide model derived from multi-mission satellite altimetry.

<p>EOT20 is the latest in a series of empirical ocean tide (EOT) models derived using residual tidal analysis of multi-mission satellite altimetry at DGFI-TUM. The amplitudes and phases of seventeen tidal constituents are provided on a global 0.125-degree grid based on empirical analysis of eleven satellite altimetry missions. The EOT20 model shows significant improvements compared to the previous iteration of the global model (EOT11a) throughout the ocean, particularly in the coastal and shelf regions, due to the inclusion of more recent satellite altimetry data as well as more missions, the use of the updated FES2014 tidal model as a reference to estimated residual signals, the inclusion of the ALES retracker and improved coastal representation. In the validation of EOT20 using tide gauges and ocean bottom pressure data, these improvements in the model compared to EOT11a are highlighted with the root-square sum (RSS) of the eight major tidal constituents improving by ~3 cm for the entire global ocean with the major improvement in RSS (~3.5 cm) occurring in coastal regions (<1 km to the coast). Compared to the other global ocean tidal models, EOT20 shows a clear improvement of ~0.4 cm in RSS compared to the closest model (FES2014) in the global ocean. Compared to the FES2014 model, the RSS improvement in EOT20 is mainly seen in the coastal region (~0.45 cm) while in the shelf and open ocean regions these two models only vary in terms of RSS by ~0.005 cm. The significant improvement of EOT20, particularly in the coastal region, provides encouragement for the use of the EOT20 model as a tidal correction of satellite altimetry in coastal sea level research. </p>

Evaluation of global ocean tide models based on tidal gravity observations in China

<p>Solid Earth is affected by tidal cycles triggered by the gravity attraction of the celestial bodies. However, about 70% the Earth is covered with seawater which is also affected by the tidal forces. In the coastal areas, the ocean tide loading (OTL) can reach up to 10% of the earth tide, 90% for tilt, and 25% for strain (Farrell, 1972). Since 2007, a high-precision continuous gravity observation network in China has been established with 78 stations. The long-term high-precision tidal data of the network can be used to validate, verifying and even improve the ocean tide model (OTM).</p><p>In this paper, tidal parameters of each station were extracted using the harmonic analysis method after a careful editing of the data. 8 OTMs were used for calculating the OTL. The results show that the Root-Mean-Square of the tidal residuals (M<sub>0</sub>) vary between 0.078-1.77 μgal, and the average errors as function of the distance from the sea for near(0-60km), middle(60-1000km) and far(>1000km) stations are 0.76, 0.30 and 0.21 μgal. The total final gravity residuals (Tx) of the 8 major constituents (M<sub>2</sub>, S<sub>2</sub>, N<sub>2</sub>, K<sub>2</sub>, K<sub>1</sub>, O<sub>1</sub>, P<sub>1</sub>, Q<sub>1</sub>) for the best OTM has amplitude ranging from 0.14 to 3.45 μgal. The average efficiency for O<sub>1</sub> is 77.0%, while 73.1%, 59.6% and 62.6% for K<sub>1</sub>, M<sub>2</sub> and Tx. FES2014b provides the best corrections for O<sub>1</sub> at 12 stations, while SCHW provides the best for K<sub>1 </sub><sub>,</sub>M<sub>2</sub>and Tx at 12,8and 9 stations. For the 11 costal stations, there is not an obvious best OTM. The models of DTU10, EOT11a and TPXO8 look a litter better than FES2014b, HAMTIDE and SCHW. For the 17 middle distance stations, SCHW is the best OTM obviously. For the 7 far distance stations, FES2014b and SCHW model are the best models. But the correction efficiency is worse than the near and middle stations’.</p><p>The outcome is mixed: none of the recent OTMs performs the best for all tidal waves at all stations. Surprisingly, the Schwiderski’s model although is 40 years old with a coarse resolution of 1° x 1° is performing relative well with respect to the more recent OTM. Similar results are obtained in Southeast Asia (Francis and van Dam, 2014). It could be due to systematic errors in the surroundings seas affecting all the ocean tides models. It's difficult to detect, but invert the gravity attraction and loading effect to map the ocean tides in the vicinity of China would be one way.</p>

Determination of Coastal Sea Level Heights around Taiwan by improved GNSS Reflectometry

<p>Rapid sea level rise, a severe consequence of global warming, could significantly damage the lives and properties of numerous human beings living in low-lying coastal areas. Therefore, realizing and monitoring coastal sea level variations are of great importance for human society. Conventionally, sea level heights are measured by using tide gauges; however, the records are contaminated by vertical land motions which are difficult to be separated. Recently, Global Navigation Satellite System Reflectometry (GNSS-R) technology has been proved to effectively monitor the coastal sea level changes from GNSS signal-to-noise ratio (SNR) data. However, the generation of detrended SNR ( SNR) depending on different satellite elevation angle intervals via a quadratic fitting, considerably influences the accuracy of sea level retrievals. Moreover, the quadratic fitting cannot perfectly describe the trend of SNR data. Therefore, we proposed a method combining ensemble empirical mode decomposition (EEMD) and ocean tide model to compute SLHs. EEMD can decompose the original SNR data into several intrinsic mode functions (IMFs) corresponding to specific frequencies. Then, Lomb-Scargle Periodogram (LSP) is applied to calculate the dominant frequency of the IMF with maximum spectral power. EEMD is not only suitable for dealing with nonlinear and nonstationary data but also eliminates the mode mixing problem of empirical mode decomposition (EMD) by adding white noises. In addition, we set an empirical SLH interval from ocean tide model as a quality control. In this study, the existing GNSS stations at the coasts of Taiwan are used to examine the proposed approach and then compare the results with those from the traditional quadratic fitting. Finally, the measurements from co-located or nearby traditional tide gauges are served as ground truth to evaluate the accuracy and stability of the mentioned methods.</p>

Asthenospheric anelasticity effects on ocean tide loading in the East China Sea region observed with GPS

Anelasticity may decrease the shear modulus of the asthenosphere by 8–10 % at semi-diurnal tidal periods compared with the reference 1 s period of seismological Earth models. We show that such anelastic effects are likely to be significant for ocean tide loading displacement at the M2 tidal period around the East China Sea. By comparison with tide gauge observations, we establish that NAO99Jb is the most accurate numerical ocean tide model in this region, and that related errors in the predicted M2 vertical ocean tide loading displacements will be 0.2–0.5 mm. In contrast, GPS observations on the Ryukyu Islands (Japan), with uncertainty 0.2–0.3 mm, show discrepancies of over 1.5 mm with respect to ocean tide loading displacements predicted using the purely elastic radial Preliminary Reference Earth Model. We show that the use of an anelastic PREM-based Earth model reduces these discrepancies to no more than 0.8 mm, which is of the same order as the sum of the remaining errors due to uncertainties in the ocean tide model and the GPS observations. Use of a regional Earth model based on the laterally-varying S362ANI, with or without further empirical tuning, results in minor additional improvements in fit.