INVESTIGATING AND REDUCING THE DIFFERENCES BETWEEN THE SATELLITE ALTIMETRY-BASED GLOBAL MEAN SEA LEVEL TIME SERIES PROVIDED BY THE DIFFERENT PROCESSING GROUPS

摘要

Satellite altimetry-based global mean sea level (GMSL)rntime series provided by the different processing groupsrnagree well in terms of trend over the whole altimetry erarn(1993-2010) but show significant differences atrninterannual time scale. In a recent paper, Masters et alrn(2012) investigated the effect of the different geophysicalrncorrections and processing methodology adopted by therndifferent groups, and found that the method of datarnaveraging and minimum depth criteria seemed to be thernmain causes of the observed discrepancies.rnHere we go one step further and attempt to discriminaternthe best approach for averaging the satellite data tornproduct the GMSL time series. We also address the effectrnof considering or not the shelf areas. For that purpose, wernuse the versions GLORYS2V1of the high-resolutionrnMERCATOR ocean circulation model, with datarnassimilation (temporal and spatial resolutions of 1-dayrnand 0.25°). We produce a set of synthetic sea surfacernheight (SSH) data by interpolating the model data at therntime and location of the 'true' along-track satelliternaltimetry measurements. We focus on the Jason-1rnoperating period (i.e., 2002-2009). These synthetic SSHrndata are then treated as “true” altimetric measurements.rnWith this synthetic data set, we test the differentrnaveraging methods classically used by the processingrngroups: along-track averaging, simple gridding (on 1°×1°,rn2°×2° or 3°×3° grids) and more sophisticated griddingrnprocedures. We also test the effect of considering or notrnshallow depths (<120 m). Finally we discuss the effect ofrnthe wet troposphere correction replacement in JASON-1rndata.