On the consistency of coastal sea-level measurements in the Mediterranean Sea from tide gauges and satellite radar altimetry

We assess the consistency of sea-level variability derived from tide-gauge (TG) and satellite radar altimeter (ALT) data acquired along the coasts of the Mediterranean Sea. For a coherent comparison between these techniques, we use GNSS observations to characterize the local vertical land movement embedded in TG records, but not affecting ALT data. We first investigate the performance of CMEMS, a gridded altimeter product covering the period 1993-2019. TG and GNSS series are not required to cover the whole altimeter period. The inter-technique comparison reveals good agreement at annual and semi-annual scales, but also the occasional occurrence of nonlinear discrepancies impacting trend estimation. Large-scale patterns of variability are observed in the Ionian region and along the continental shores from the Alboran to the Adriatic Sea. The estimates of linear trends based on TG + GNSS or CMEMS observations are found consistent within 1 sigma at 27/45 sites, with the best agreement in the Western Mediterranean and Adriatic Seas. We also consider the X-TRACK/ALES altimeter dataset, provided along the tracks of the Jason missions (2002-2018) and optimized for coastal applications. In this case, we identify only 12 sites suitable for the comparison. The results show that inter-technique consistency is impacted by the length of the series used in the comparison. Optimum agreement between X-TRACK/ALES and TG + GNSS trends is reached at the two sites closer to a satellite track. However, we find sites where X-TRACK/ALES-derived sea-level trends present suspicious along-track variations at < 5 km from the coast, where the corresponding 20 Hz series are incomplete and noisy.

Automated summary

This study assesses the consistency of sea-level variability derived from tide-gauge and satellite radar altimeter data along the coasts of the Mediterranean Sea. The results show good agreement at annual and semi-annual scales, with occasional nonlinear discrepancies impacting trend estimation. Large-scale patterns of variability are observed in certain regions. The length of the data series used in the comparison affects the inter-technique consistency.