North Atlantic sea level budget revisited

Abstract. Based on satellite altimetry, GRACE space gravimetry and Argo-based steric data down to 2000 m, recent studies have shown that the North Atlantic sea level budget (i.e., altimetry-based sea level minus sum of components) of the past two decades is not closed, with strong regional residuals in the North Atlantic. This was suggested to result from salinity errors reported since ~2015 in some Argo float measurements. In this study, we revisit the North Atlantic sea level budget, using satellite altimetry, GRACE and GRACE-FO data, different Argo products and two ocean reanalyses (CIGAR and ORAS5) over the 2004–2022 time span. The ocean reanalyses are used to estimate the manometric contribution, an alternative to using GRACE data, as well as the deep ocean contribution to the sea level budget, not yet fully sampled by Argo. Analyzing different data sets allows us to evaluate their impact on the previously reported non-closure of the North Atlantic sea level budget. We first find that using the CIGAR ocean reanalysis-based manometric component significantly reduces the residuals of the North Atlantic sea level budget compared to GRACE. We also find that accounting for the deep ocean (below 2000m) thermal expansion (using the CIGAR reanalysis) allows for 30 % reduction of the North Atlantic budget residuals when using GRACE for the manometric component, while the mean residual trend is reduced by a factor of 2 when using CIGAR for the manometric sea level. In the latter case, the budget is closed within data uncertainties. The North Atlantic halosteric component based on Argo and CIGAR in the upper 2000m displays a small decrease since the early 2010s. However, this negative trend becomes stronger after 2016. The 2010–2016 halosteric decrease may reflect a real salinity increase in the region, although salinity measurement errors may have impacted the halosteric component after that date.