A revised acceleration rate from the altimetry-derived global mean sea level record

Marcel Kleinherenbrink,Remko Scharroo,Riccardo Riva

doi:10.1038/s41598-019-47340-z

Abstract

Satellite radar altimetry has been providing estimates of global mean sea level (GMSL) since 1992. The early TOPEX record originates from two identical instruments, which requires the estimation of an intermission bias and careful handling of the problematic first part of the record. Calibration of TOPEX is crucial to obtain a continuous and consistent record, which is needed to quantify any recent acceleration. We propose a novel approach to calibrate TOPEX altimeter data using sea surface height differences at crossovers of TOPEX and ERS. Tide gauges are only used to determine a drift in one of the two datasets. We provide a new and more accurate estimate of the intra-mission bias, which leads to a much reduced GMSL acceleration over the whole record. Hence, the conundrum of an uncertain GMSL acceleration from altimetry is still unsolved, in spite of recent opposite claims, and in contrast to the expected effect of ocean warming and continental freshwater fluxes.

Highlights

Data from the satellite radar altimetry missions TOPEX/Poseidon (1992–2002), Jason-1 (2002–2008) and Jason-2 (2008–2016) need to be merged in order to construct continuous time series of global mean sea level (GMSL)
Over the TOPEX-A period, the tide-gauge validation revealed a U-shaped drift in TOPEX sea surface heights[3], which has been related to the degradation of the Point Target Response (PTR) of the instrument[4]
The first 1.5 years of the timeseries had to be omitted due to data processing issues. 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

Summary

Crossover Analysis

With all four averaging methods TOPEX-A/B drifts are estimated that are statistically equivalent to each other and to those obtained with the crossovers, independent from the applied SSB correction. Model 1 is not the best fit with the crossover time series when the CSR or CLS SSB correction were applied and the F-score indicated that Gaspar time series might require the estimation of an intramission bias as well. According to the crossover analysis, no intramission bias has to be estimated when cal-1 is applied to TOPEX-B only in combinationn with the Gaspar SSB correction. These tide-gauge-derived drifts are equivalent to those estimated with the crossovers. When the Gaspar SSB correction is applied, while cal-1 is removed from TOPEX-A, only a drift should be removed

Accelerations in Global Mean Sea Level

Methods

Additional Information