Time-Space Sampling-Related Uncertainties of Altimetric Global Mean Sea Level Estimates

Uncertainty measures for global mean sea level (GMSL) estimates are quantified, resulting from limited (in space and time) along-track altimetric sampling of the global sea level field by altimetric satellite missions. To estimate such sampling-related uncertainty, sea surface height (SSH) fields simulated by the high-resolution STORM/NCEP ocean circulation model were subsampled along altimeter tracks for the period 1993-2010 and subsequently processed into global SSH averages using similar techniques to those used by six groups worldwide. Results show that the underlying satellite space-time sampling has a substantial impact on the accuracy of GMSL estimates. This uncertainty originates primarily from data missing over sea ice-covered regions; omitted data from shallow seas also contributes. Uncertainties in GMSL estimates result both from interpolation techniques required to fill data gaps such as missing tracks, and the choice of the mean sea surface required to estimate SSH anomalies. Cumulative effects lead to errors in GMSL estimates from similar to 0.8 to similar to 3.2 mm (root-mean-square, RMS), depending on the underlying details of the estimation method. Results suggest that sampling limitations in meridional direction are a fundamental constraint on the accuracy level reachable for any altimetric GMSL estimate, resulting in a systematic Gaussian uncertainty of about 1.2 mm (RMS), 50% of which occurs on monthly time scales, while some fraction occurs on time scales of several years. In all cases, a significant fraction of the error results from a mass exchange between the global ocean and sea ice-covered polar regions. Contributions from data processing details are measurable but less significant.