We challenge these solutions by means of a crossover analysis between TOPEX an ERS-1&2 and we derive a revised acceleration rate, whose point estimate is much reduced in magnitude with respect to those previous studies. The crossover analysis shows that not applying cal-1 is justified and removes the U-shaped drift in TOPEX-A. However, the removal of cal-1 from the whole TOPEX record leads to an intramission bias with an amplitude that depends on the applied SSB correction. Additionally, this approach reveals the presence of a drift of −1.1 ± 0.3 mm yr−1 between the TOPEX sea surface heights and those of ERS-1&2. An alternative approach consists in removing cal-1 from TOPEX-A alone, which yields a negligible intramission bias in case the Gaspar et al.12 SSB correction is applied. Also in this case, a drift between TOPEX and ERS of −1.3 ± 0.1 mm yr−1 is found, with is statistically equivalent to the previous value.

Based on four different weighting methods used in a tide-gauge comparison it is determined that TOPEX is drifting and not ERS. Therefore, we suggest to calibrate the TOPEX GMSL record with the crossover of ERS1&2 after the removal of cal-1. The calibration reduces the observed acceleration in GMSL, so that it becomes statistically equivalent to zero at the 95%-confidence level.