Multidecadal Variability in the Rate of Global Sea-level Rise Since 1900

Observed global mean sea-level (GMSL) rise since 1900 is superimposed on pronounced multidecadal variability that modulates short-term rates and accelerations, increasing uncertainty in coastal adaptation planning. However, concerns remain about whether this variability reflects a physical signal or arises from uneven tide-gauge sampling and early-century data gaps. Here, we use synthetic sea-level experiments to assess the robustness of reconstructed multidecadal variability. The experiments show that this variability is not an artifact of early-century data gaps or spatial sampling bias. Applied to tide-gauge observations over 1900–2022, the GMSL reconstruction yields a trend of 1.6 ± 0.2 mm yr⁻¹, consistent within 90% confidence level with the sum of contributions from glaciers, ice sheets, terrestrial water storage, and thermosteric expansion (1.5 ± 0.2 mm yr⁻¹). Superimposed on this trend, 30-year running GMSL rise rates exhibit multidecadal variability, with a rapid rise to the late 1930s, a slowdown toward the 1970s, and sustained acceleration thereafter. This GMSL multidecadal variability reflects time-varying barystatic and thermosteric contributions over different periods, with the barystatic component dominating. By contrast, multidecadal regional sea-level patterns are governed by sterodynamic sea level (SDSL), exhibiting ENSO-like spatial patterns. This regional pattern is consistent with independent steric sea-level estimates from in situ temperature and salinity observations, largely confined to thermosteric sea level in the upper 700 m. These findings highlight the roles of wind-driven circulation and upper-ocean heat accumulation in shaping multidecadal regional departures from GMSL and emphasize the need to improve ocean dynamics in future regional sea-level projections.