MEAN SEA LEVEL VALIDATION AND ESTIMATION OF ERRORS

When calculating and analyzing the evolution of mean Sea Level (MSL), the question legitimately arises as to whether altimeters are reliable enough to measure a rise in MSL of a few millimeters per year over a period of more than 20 years. There are potentially many sources of errors that could affect the long-term stability of the GMSL record. These errors come from the ageing of intruments for instance, from the uncertainty of their calibration, from the ageing and incidents/maneuvers of the satellite platform, from the orbit calculations, as well as from all the uncertainties associated to the geophysical corrections used to derive the SSH (tides, troposphere, ionosphere, ect). The estimation of the uncertainties of the MSL altimetry data is thus key to be able to validate the accuracy of such climate data record. 

UNCERTAINTY ON THE GMSL RISE

Following the approach of Ablain et al. (2019) who consider all potential sources of errors affecting the altimetric system, an updated error budget has been established for the GMSL timeseries up to 2021. Using an Ordinary Least Square (OLS) fit that considers this error budget, the GMSL trend uncertainty (i.e., the mean rise uncertainty) is estimated at 0.4 mm/yr over the whole altimetry era (1993-2021) at the 90% confidence level.

 

Focus on TOPEX-A instrumental drift and its impact on the GMSL curve

The GMSL uncertainty is significantly larger during the first decade (1993-2002) of the record, due to larger errors affecting the T/P sea level measurements on climatic time scales, compared to subsequent missions. For instance, the T/P orbit solutions are more uncertain due to lower precision of the gravity field solutions estimated without GRACE data. Furthermore, the switch from TOPEX-A to TOPEX-B in February 1999 leads to a 2mm bias uncertainty (at 1-sigma) in the GMSL time series (Ablain et al. 2019). 

However, the most significant error affecting the first 6 years (January-1993 to February 1999) of the T/P GMSL measurements comes from an instrumental drift of TOPEX-A altimeter. This effect on the GMSL time series was recently highlighted via comparisons with tide gauges (Valladeau et al. 2012; Watson et al. 2015; Chen et al. 2017; Ablain et al. 2017), via a sea level budget approach (i.e., comparison with the sum of mass and steric components; Dieng et al., 2017) and via comparisons to Poseidon-1 measurements (Zawadzki, personal communication). In a recent study, Beckley et al. (2017) suggested that the corresponding error on the GMSL during the 1993-1998 time-span, results from incorrect onboard calibration parameters. 

All three approaches conclude that during the period January 1993 to February 1999, the altimetry-based GMSL slope was overestimated by ~1.5 mm/yr, with an uncertainty in the range of 0.5 to 1.0 mm/yr (Watson et al. 2015; Chen et al. 2017; Dieng et al. 2017). Beckley et al. (2017) proposed to not apply the suspicious onboard calibration correction on TOPEX-A measurements. The latter approach is similar to apply the TOPEX-A drift correction estimated by Dieng et al. (2017) and Ablain et al. (2017b). In the latter study, accurate comparisons between TOPEX A-based GMSL and tide gauge measurements lead to a drift correction of about -1.0 mm/yr between January 1993 and July 1995, and +3.0 mm/yr between August 1995 and February 1999, with an uncertainty of 1.0 mm/yr (68% confidence level). 

 

Evolution of the ensemble mean GMSL time series (average of the 6 GMSL products from AVISO/CNES, SL_cci/ESA, University of Colorado, CSIRO, NASA/GSFC, and NOAA). On the black, red and green curves, the TOPEX-A drift correction is applied, using the correction proposed respectively by Ablain et al.(2017b), Watson et al. (2015), Dieng et al. (2017) and Beckley at al. (2017). On the figure, the annual signal is removed, and a 6-month smoothing is applied; The GIA correction is also applied. Uncertainties (90% confidence interval) of correlated errors over a 1-year period are superimposed for each individual measurement (shaded area).

Comparison with in-situ measurements

Another approach for describing the GMSL error consists in comparing the sea level data with in situ data such as tide gauges or ARGO data (temperature and salinity profiles) (Ablain et al., 2009 (a), (b)). The stability of the altimetry system can thus be assessed by detecting drifts or leaps in the GMSL (SALP annual GMSL report 2017).

Comparison to tide gauges

The method to compare altimeter and tide gauges data (noted Alti-TG hereafter) is based on the main following steps: 

- The along-track altimeter data are averaged on 360x180 global grids (1°x1° cells). One grid is computed per mission cycle, except for comparisons to the PSMSL tide gauge network, where monthly grids are used.

- In a 300 km radius around each tide gauge, the best correlated grid element is chosen and a GMSL difference time series (Altimeter GMSL - Tide gauge GMSL) is computed.

- A mean global Alti-TG time series is computed from Alti-TG time series. The trend difference is computed from this global time series, which represents the estimated altimeter GMSL drift.

- Vertical Land Motion (VLM) correction: each trend difference is corrected for Glaciar Isostatic Ajustment (GIA) induced VLM, using Pr. Peltier’s ICE-5G(VM2) model (see Peltier 2004). Due to insufficient GPS data at the tide gauge locations over the whole GLOSS-CLIVAR or PSMSL networks, no correction is applied for the other VLM contributions.

The altimeter GMSL drift is estimated within a confidence level depending on different source of errors: vertical land motion at the tide gauges, high-frequency noise from altimeter and tide gauge data, collocation errors from both datasets, and the global alti-TG averaging scheme. Describing the temporal correlation of these errors and their evolution with time allows to provide the uncertainty of the GMSL altimeter drift over the entire altimetry period. It is estimated to 0.4 mm/yr between 1993 and 2017 at 90% confidence level (Ablain et al., 2018). Furthermore, the uncertainties have been also provided for any altimeter periods between 1993 and 2017 (see figure below).

Altimeter GMSL drift uncertainties (mm/yr) estimated from alti-TG comparison using the GLOSS/CLIVAR tide gauge network for any altimeter periods between 1993 and 2017. The confidence level is 90 % (1.65-sigma). On the Y-axis is represented the length of the window (in years) and on the X-axis the central date of the window used (in years).

The GMSL drifts and associated uncertainties (at the 90% confidence level), based on comparisons with GLOSS-CLIVAR and PSMSL data, are shown for each altimeter on the figure below. For most missions, no significant drift can be detected. Most notably, the drift for the reference GMSL time series provided on the AVISO website is estimated between -0.35 mm/yr and +0.45 mm/yr at the 90% confidence level, thus is not considered as significant.

It is worth noting that a significant drift is detected on the TOPEX-A –based GMSL. This drift is evaluated at 1.6 +/- 1.2 mm/yr (90% confidence level) when compared to data from the GLOSS-CLIVAR tide gauge network. Such a drift is well-known by the scientific community, and several corrections have been proposed. A significant drift, of -1.3 +/- 0.8 mm/yr, is also detected in the ERS-2 GMSL time series. Data from this altimeter are not used to compute the AVISO reference GMSL (based on only TOPEX and Jason missions). The ERS-1 altimeter may have been affected by a significant drift but the uncertainty from the comparison to tide gauges for this specific mission is too high to accurately estimate a drift. This high uncertainty is due to the short duration of the mission (only 3.5 years, with many data gaps) and very high altimeter noise levels.

At the moment, the Jason-3 and Sentinel-3A records are too short for a comparison with tide gauges and the detection of a drift.

GMSL drift on each altimeter estimated by comparison to tide gauges from the GLOSS-CLIVAR and PSMSL networks
Global MSL difference between the AVISO reference missions (Topex + Jason missions) and tide gauges from the GLOSS-CLIVAR network. This is obtained by the Alti-TG comparison process described previously

Local errors in MSL slopes

With respect to local errors of MSL slopes, the same statistical approach was applied in order to describe locally the different sources of uncertainty in altimetry data [Ablain et al., 2010]. This work revealed the strong local impact of inter-annual variability since it accounts for 80% of the formal adjustment error, which demonstrates that the periods of altimetry data are still too short to estimate slopes locally. The residual errors (20%) due to errors in the altimetry system vary essentially according to latitude between 1 and 2 mm/year due to uncertainties in orbit determination and correction for the wet troposphere respectively at high latitudes and in the tropical band. In this case as well, the description of local errors should be refined to take into account in particular the impact of variable gravity fields for the orbit calculation [Cerri et al., 2010] or again uncertainty in zones close to coasts or to ice due to the degradation of the altimetry measurement.

Recent work has been published by Prandi et al. 2021 that described the state-of-the-art regional MSL uncertainties.

References

  • (a) Ablain, M., Cazenave, A., Valladeau, G., Guinehut, S., 2009. A new assessment of the error budget of global mean sea level rate estimated by satellite altimetry over 1993-2008. Ocean Science, 5(2), 193-201. DOI: 10.5194/os-5-193-2009
  • (b) Ablain M., 2009. Quality assessment of tide gauge and altimeter measurements through SSH comparisons. Oral presentation in OSTST, Seattle 2009.
  • Ablain M., 2010. Quality assessment of tide gauge and altimeter measurements through SSH comparisons. Oral presentation in ESA Symposium, Bergen 2010.
  • Ablain, M., Philipps, S., Urvoy, M., Tran, N., Picot, N., 2012. Detection of long-term instabilities on altimeter backscatter coefficient thanks to wind speed data comparisons from altimeters and models. Marine Geodesy, 35(sup1), 258-275. DOI: 10.1080/01490419.2012.718675
  • Ablain, M., Legeais, J. F., Prandi, P., Marcos, M., Fenoglio-Marc, L., Dieng, H. B., Benveniste, J., Cazenave, A., 2017. Satellite altimetry-based sea level at global and regional scales. In: Cazenave A., Champollion N., Paul F., Benveniste J. (eds) Integrative Study of the Mean Sea Level and Its Components. Space Sciences Series of ISSI, vol 58. Springer, Cham. DOI: 10.1007/978-3-319-56490-6_2
  • Ablain M., R. Jugier, N. Picot., 2018. Estimation of any Altimeter Mean Seal Level Drifts between 1993 and 2017 by Comparison with Tide-gauges Measurements. Presented at International Review Workshop on Satellite Altimetry Cal/Val Activities and Applications 2018.
  • Beckley, B. D., Callahan, P. S., Hancock, D. W., Mitchum, G. T., Ray, R. D., 2017. On the "Cal-Mode" Correction to TOPEX Satellite Altimetry and Its Effect on the Global Mean Sea Level Time Series. Journal of Geophysical Research: Oceans, 122(11), 8371-8384. DOI: 10.1002/2017JC013090
  • Cerri L., Berthias J.P., Bertiger W., Haines B., Lemoine F.G., Mercier F., Ries J.C., Willis P., Zelensky N.P., Ziebart M., 2010. Precision Orbit Determination Standards for the Jason Series of Altimeter Missions. Marine Geodesy 2010.
  • Chen, X., Zhang, X., Church, J. A., Watson, C. S., King, M. A., Monselesan, D., Legresy, B., Harig, C., 2017. The increasing rate of global mean sea-level rise during 1993–2014. Nature Climate Change, 7, 492-495. DOI: 10.1038/nclimate3325
  • Couhert, A., Cerri, L., Legeais, J. F., Ablain, M., Zelensky, N. P., Haines, B. J., Lemoine, F.G., Bertiger, W.I., Desai, S.D., Otten, M., 2015. Towards the 1 mm/y stability of the radial orbit error at regional scales. Advances in Space Research, 55(1), 2-23. DOI: 10.1016/j.asr.2014.06.041
  • Dieng, H. B., Cazenave, A., Meyssignac, B., Ablain, M., 2017. New estimate of the current rate of sea level rise from a sea level budget approach. Geophysical Research Letters, 44(8), 3744-3751. DOI: 10.1002/2017GL073308
  • Escudier, P., Couhert, A., Mercier, F., Mallet, A., Thibaut, P., Tran, N., Amarouche, L., Picard, B., Carrère, L., Dibarboure, G., Ablain, M., Richard, J., Steunou, N., Dubois, P., Rio, M. H., and Dorandeu, J., 2017. Satellite radar altimetry: principle, geophysical correction and orbit, accuracy and precision, in: Satellite Altimetry Over Oceans and Land Surfaces, edited by: Stammer, D. and Cazenave, A., CRC Press, Taylor Francis, Boca Raton
  • Legeais, J.-F., Ablain, M., Zawadzki, L., Zuo, H., Johannessen, J. A., Scharffenberg, M. G., Fenoglio-Marc, L., Fernandes, M. J., Andersen, O. B., Rudenko, S., Cipollini, P., Quartly, G. D., Passaro, M., Cazenave, A., and Benveniste, J., 2018. An improved and homogeneous altimeter sea level record from the ESA Climate Change Initiative. Earth System Science Data, 10, 281-301. DOI: 10.5194/essd-10-281-2018
  • SALP annual GMSL report 2017. Validation of altimeter data by comparison with tide gauges measurements and with in-situ T/S ARGO profiles
  • Valladeau, G., Legeais, J. F., Ablain, M., Guinehut, S., Picot, N., 2012. Comparing altimetry with tide gauges and Argo profiling floats for data quality assessment and mean sea level studies. Marine Geodesy, 35(sup1), 42-60. DOI: 10.1080/01490419.2012.718226
  • Watson, C. S., White, N. J., Church, J. A., King, M. A., Burgette, R. J., Legresy, B., 2015. Unabated global mean sea-level rise over the satellite altimeter era. Nature Climate Change, 5(6), 565. DOI: 10.1038/nclimate2635
  • Zawadzki, L., Ablain, M., 2016. Accuracy of the mean sea level continuous record with future altimetric missions: Jason-3 vs. Sentinel-3a. Ocean Science, 12(1), 9-18. DOI: 10.5194/osd-12-1511-2015