Simulated sea levels during 1948–2009 in a global ocean-sea ice model for OMIP

Abstract

This paper assesses the ability of the State Key Laboratory of Atmospheric Science and Geophysical Fluid Dynamics of the Institute of Atmospheric Physics (LASG/IAP) Climate System Ocean Model version 3 (LICOM3) to simulate sea levels. For the time series of global changes, most of the global mean steric sea level (SSL) changes were dominated by thermosteric sea level (TSSL) changes in the upper 700 m, with a relatively small contribution from halosteric sea level (HSSL). LICOM3 can reproduce the observed dynamic sea level (DSL) in most regions to reflect the spatial patterns of regional changes. Deviations between LICOM3 and observations were mainly concentrated in the western boundary current and the tropical eastern Pacific. The departure in the western boundary current was primarily related to the model resolution. LICOM3 could not depict DSL changes in the high latitudes where the mesoscale eddies were active. The excessive negative trend in the tropical eastern Pacific was mainly caused by the trend error in the CORE-II forcing products. For the sea level variability, the analysis results of different DSL periods showed that the LICOM3 model could simulate interannual and decadal variability and the satellite altimeter period trend. At least 50 years of data were used to separate the decadal variability and long-term trend. The decadal variability had no significant effect on the 50-year long-term trend change in the DSL in the tropical western Pacific.Analysis of sea level component contributions showed that the changes in global time series, spatial patterns, and sea level variability mainly resulted from the change in TSSL. The TSSL changes in the upper 700 m could roughly reflect the TSSL changes for the whole depth, except for the North Atlantic and the Southern Ocean. In the Atlantic, the contribution of HSSL was comparable to and compensated by TSSL. When compared with reanalysis, the simulation generated differences in the Southern Ocean and North Atlantic. The linear trend of TSSL simulated by LICOM3 was much stronger than that simulated by reanalysis. The deviation of the Southern Ocean may be related to the CORE-II forcing product. The lack of observations in the Southern Ocean led to the uncertainty of the CORE-II forcing product.