The rate of global sea level rise doubled during the past three decades

With the launch of Sentinel-6 Michael Freilich in 2020, the satellite radar altimeter record of sea level surpassed 30 years, enabling definitive measurement of both the net increase in global mean sea level (GMSL) and changes in its rate. GMSL rise is driven primarily by ocean thermal expansion as the ocean absorbs ~90% of excess heat and by mass addition from land ice loss, processes that are now well measured by Argo, GRACE/GRACE-FO, and targeted missions like OMG. On shorter timescales, interannual to decadal variability, including ENSO-driven exchanges of water between land and ocean and intrinsic ocean heat content variability, can temporarily modulate GMSL, but their influence diminishes as the record lengthens. The commonly cited long-term average rate of 3.3 mm/year, computed as a linear fit to the full record, has become less representative due to a recent acceleration in GMSL rise beginning around 2017/2018. Understanding the evolving trajectory—represented here by a quadratic fit—provides actionable insight for coastal adaptation by addressing not only how much sea level will rise but also how fast, and offers an observation-based near-term pathway to compare against projection frameworks such as those used by IPCC AR6.

Previous work established significant acceleration in GMSL during the altimeter era and explored methods to separate forced signals from natural variability (e.g., ENSO). Earlier estimates often removed climate variability signals before estimating acceleration, but more recent studies demonstrated that statistically significant acceleration can be detected without such removals. Reported accelerations of ~0.08–0.12 mm/year^2 are broadly consistent across analyses, with regional accelerations sometimes higher but more uncertain due to ocean dynamics that do not affect GMSL. The literature also emphasizes uncertainties in satellite altimetry (instrumental and geophysical corrections), the role of ENSO in short-term fluctuations, and the importance of aligning observation-based estimates with model-based projections such as IPCC AR6’s SSP scenarios for planning and risk assessment.

Data: The GMSL time series is from the Integrated Multi-Mission Ocean Altimeter Data for Climate Research, combining TOPEX/Poseidon, Jason-1, OSTM/Jason-2, Jason-3, and Sentinel-6 Michael Freilich. A glacial isostatic adjustment (GIA) correction is applied and the seasonal cycle removed to ensure comparability with IPCC AR6 projections and to reflect true ocean volume/mass changes. Estimation: A quadratic fit is applied to monthly GMSL from January 1993 to December 2023, referenced to the midpoint (2008.5). The GMSL is approximated as ½·a·t^2 + b·t + c, where acceleration = 2a and the instantaneous rate at time t is a·t + b. The analysis follows established approaches (e.g., Nerem et al.). Uncertainty assessment: Three error sources are quantified: (i) serially correlated residuals (accounted for using lag-1 correlation methods following Maul & Martin), (ii) uncertainties in the applied GIA correction (based on a Bayesian inversion across ice history and Earth structure ensembles), and (iii) mission-dependent measurement errors assumed spatially uniform, including altimeter noise, geophysical corrections, orbit determination, wet troposphere correction, precision orbit determination, inter-mission biases, large-scale reference frame drifts, and TOPEX/Poseidon instabilities. A Monte Carlo framework (10,000 realizations) perturbs rate and acceleration according to their 1-sigma uncertainties from these sources to derive 90% confidence intervals for the trajectory and its extrapolation (GIA excluded from acceleration perturbations). Robustness test: The start year is fixed at 1993, while the end year varies from 2017.99 to 2024 to examine sensitivity to record length and recent ENSO events (including the strong 2023 El Niño).

- Total rise: 111 mm increase in GMSL from 1993 to the end of 2023; the article also notes 111 mm from 1993 to 2024 in a summary statement. - Average rate and acceleration (1993–2023): 3.3 ± 0.3 mm/year (referenced to the time series midpoint) with acceleration 0.08 ± 0.06 mm/year^2 (90% CI), consistent with prior estimates. - Doubling of rate: Instantaneous rate rose from 2.1 ± 1.0 mm/year at the beginning of 1993 to 4.5 ± 1.0 mm/year at the end of 2023 (more than doubling over 31 years). - Robustness to ENSO and record length: Varying the end year (2017.99–2024) yields stable rates (~3.3 mm/year) and accelerations (0.08–0.09 mm/year^2) with decreasing uncertainty as the record lengthens; the 2023 El Niño has only minor impact on the estimated acceleration. - Observation-driven near-term projection: Extrapolating the quadratic trajectory to 2050 gives 205 ± 54 mm of rise from 2008 to 2050. The observation-based estimate aligns closely with the median and likely range of the IPCC AR6 SSP2-4.5 scenario, with substantial overlap across scenarios. - 2020–2050 rise from varying record lengths (Table 1): 158–174 mm; for the current full record (to 2024), 169 ± 52 mm (90% CI). - Projected instantaneous rates if current trajectory continues: 5.0 ± 1.4 mm/year by 2030, 5.8 ± 2.0 mm/year by 2040, and 6.5 ± 2.6 mm/year by 2050.

The findings demonstrate an unequivocal acceleration in GMSL and a more than doubling of the rate since 1993, directly addressing the need to quantify not only total rise but the evolving pace of change. By using an observation-based quadratic trajectory that does not explicitly remove natural variability such as ENSO, the study shows that the forced signal is now strong enough to yield statistically significant acceleration. The stability of rate and acceleration estimates across different end dates, including during a strong El Niño, supports the robustness of the inferred trajectory. Importantly, the extrapolated near-term rise to 2050 tracks the central tendency and likely range of IPCC AR6 SSP2-4.5 projections, providing an independent observational line of evidence that can inform and constrain planning scenarios. For coastal risk management, the increasing rates imply shrinking adaptation lead times and the potential need to accelerate or adjust adaptation plans as the trajectory evolves. Monitoring GMSL thus remains critical both as a global integrator of cryospheric mass loss and ocean warming and as a contextual benchmark for regional and local sea-level planning.

Over the satellite altimetry era, the rate of global sea level rise has more than doubled, with a total rise of about 111 mm since 1993 and an acceleration of ~0.08 mm/year^2. Extrapolation of the observed trajectory suggests about 205 ± 54 mm of rise from 2008 to 2050 and approximately 169 mm from 2020 to 2050, aligning closely with mid-range model projections (e.g., IPCC AR6 SSP2-4.5). These observation-based estimates provide a valuable, independent complement to model-based projections and a practical basis for near-term adaptation planning. Continued satellite altimetry with upcoming missions (Sentinel-6B in 2025 and potentially Sentinel-6C in 2030) is essential to track the trajectory, detect shifts that could warrant adaptation strategy adjustments, and refine projections. Future work should maintain and enhance observing systems, reduce uncertainties (e.g., measurement and GIA), and further integrate observations with models to improve near-term, regionalized sea-level guidance for decision-making.

- The analysis relies on a quadratic fit to observations through 2023 and extrapolates to 2050; it is not intended for longer-term projections where poorly constrained processes (e.g., ice-sheet instabilities) could dominate. - Short-term natural variability (e.g., ENSO) can modulate year-to-year rates, though its impact on long-record estimates is small; nevertheless, interannual anomalies may transiently affect inferred rates near record endpoints. - Uncertainties remain in GIA corrections, mission-dependent measurement errors, and treatment of serially correlated residuals, although these are quantified and incorporated into confidence intervals. - The study focuses on global mean sea level; regional sea-level dynamics and local factors are not directly resolved but can differ substantially from GMSL. - Observation-based extrapolations do not directly assimilate process-based model constraints and thus represent an empirical, near-term trajectory rather than a fully mechanistic projection.