Advances in NE-Atlantic coastal sea level change monitoring by Delay Doppler altimetry

Abstract

A significant part of the World population lives in the coastal zone, which is affected by sea level rise and extreme events. Consistent and precise new measurements are needed to assess and predict these changes. New altimeter missions equipped with Synthetic Aperture Radar (SAR) mode provide more accurate sea level heights. In this work we analyse their impacts on the estimation of the coastal sea level variability in the last 10 km. First, by analysing various standard and improved SAR altimetric products it is found that SAR altimetry will reduce the minimum usable distance from five to three kilometres. The best performance is achieved with the SAR coastal retrackers SAMOSA+ and SAMOSA++. A comparable performance is obtained in low resolution Reduced SAR (RDSAR) mode with the Spatio-Temporal Altimeter sub-waveform Retracker (STAR). In both cases, sea level heights are recovered with 4 cm accuracy up to 3 km from the coast at the Helgoland station, and the accuracy does not decrease with the distance to coast. In estuaries and coastal zones with hight tidal regimes, the discrepancy between altimetry and in-situ observations remains large (40 cm standard deviation difference (stdd) with SAR-SAM+). CryoSat-2 and Sentinel-3A have similar accuracy, despite their different repeat cycle and sampling, which require a different post-processing. The monthly coastal variability from the SAR-SAM+ product agrees most favourably with the NEMO-WAM model, with stdd 3.9 and correlation 0.90 for Sentinel-3A. Instead, the maximum departure between altimeter data products is between SAR-SAM+ and RDSAR-TALES CryoSat-2 data (stdd 2.3 and correlation 0.96). An average trend of about 3 1.3 mm/yr has been detected using low resolution altimetry at ninetheen in-situ stations along the German coast over time intervals longer than 15 years. The difference of the trends of altimetry and tide gauge co-located time-series (al-tg) agrees with the GNSS rates within 1.5 mm/yr at half of the co-located locations, with error of al-tg larger than the error of the GPS rate by a factor bigger than two. The larger departure between monthly tide gauge and LRM altimetric time-series as between tide gauge and SAR altimetric time-series confirms the higher accuracy of SAR data compared to LRM. Finally, the sea level trends of the merged LRM and SAR altimetry time-series are consistent with the LRM trends.