Components of 21 years (1995–2015) of absolute sea level trends in the Arctic

Abstract

The Arctic Ocean is at the frontier of the fast-changing climate in the northern latitudes, and sea level trends are a bulk measure of ongoing processes related to climate change. Observations of sea level in the Arctic Ocean are nonetheless difficult to validate with independent measurements, and this is globally the region where the sea level trend (SLT) is most uncertain. The aim of this study is to create a satellite-independent reconstruction of Arctic SLT, as it is observed by altimetry and tide gauges (TGs). Previous studies use Gravity Recovery and Climate Experiment (GRACE) observations to estimate the manometric (mass component of) SLT. GRACE estimates, however, are challenged by large mass changes on land, which are difficult to separate from much smaller ocean mass changes. Furthermore, GRACE is not available before 2003, which significantly limits the period and makes the trend more vulnerable to short-term changes. As an alternative approach, this study estimates the climate-change-driven Arctic manometric SLT from the Arctic sea level fingerprints of glaciers, Greenland, Antarctica and glacial isostatic adjustment (GIA) with the addition of the long-term inverse barometer (IB) effect. The halosteric and thermosteric components complete the reconstructed Arctic SLT and are estimated by interpolating 300 000 temperature (T) and salinity (S) in situ observations. The SLT from 1995–2015 is compared to the observed SLT from altimetry and 12 selected tide gauges (TGs) corrected for vertical land movement (VLM). The reconstructed estimate manifests the salinity-driven halosteric component as dominating the spatial SLT pattern with variations between −7 and 10 mm yr−1. The manometric SLT in comparison is estimated to be 1–2 mm yr−1 for most of the Arctic Ocean. The reconstructed SLT shows a larger sea level rise in the Beaufort Sea compared to altimetry, an issue that is also identified by previous studies. There is a TG-observed sea level rise in the Siberian Arctic in contrast to the sea level fall from the reconstructed and altimetric estimate. From 1995–2015 the reconstructed SLT agrees within the 68 % confidence interval with the SLT from observed altimetry in 87 % of the Arctic between 65∘ N and 82∘ N (R=0.50) and with 5 of 12 TG-derived (VLM-corrected) SLT estimates. The residuals are seemingly smaller than results from previous studies using GRACE estimates and modeled T–S data. The spatial correlation of the reconstructed SLT to altimetric SLT during the GRACE period (2003–2015) is R=0.38 and R=0.34/R=0.37 if GRACE estimates are used instead of the constructed manometric component. Thus, the reconstructed manometric component is suggested as a legitimate alternative to GRACE that can be projected into the past and future.