Gauging the Improvement of Recent Mean Sea Surface Models: A New Approach for Identifying and Quantifying Their Errors

Abstract

AbstractThis paper presents an assessment and comparison of recent mean sea surface (MSS) models. Using a new approach and independent altimeter data sets, we quantify the major improvement of the CNES_CLS15 and the DTU15 models. We observe a reduction in the amplitude of omission errors thanks to the use of new geodetic altimeter data sets (i.e., Cryosat‐2 and Jason‐1 geodetic): they are reduced by a factor of 2 compared with previous generations (CNES_CLS11). We also quantify commission errors resulting from the leakage of residual ocean variability and altimeter noise into the MSS models. For wavelengths shorter than 250 km, the error is of the order of 1–2 cm2, i.e., ∼10% to 20% of the sea level anomaly (SLA) variance. The global error of both 2015 models has similar orders of magnitude and spectral power densities, although the commission errors of the CNES_CLS15 model are about half as large as those of the DTU15 model. Its absolute error is also slightly smaller than for the DTU15 model in coastal regions and at high latitudes. Conversely, the DTU15 model produces smaller omission errors, especially in the open ocean over strong bathymetric features. More importantly, the MSS errors still have a substantial impact on altimetry products for wavelengths ranging from 30 to 100 km: the error explains ∼30% of the global SLA variance, and the error can be 2.5 times higher on uncharted ground tracks (e.g., Sentinel‐3) over rugged bathymetry.