LIVERMORE, California — Using ocean observations and a large suite of climate models, Lawrence Livermore National Laboratory scientists have found that long-term salinity changes have a stronger influence on regional sea level changes than previously thought.
“By using long-term observed estimates of ocean salinity and temperature changes across the globe, and contrasting these with model simulations, we have uncovered the unexpectedly large influence of salinity changes on ocean basin-scale sea level patterns,” said LLNL oceanographer Paul Durack, lead author of a paper appearing in the November issue of the journal Environmental Research Letters (link is external).
Sea level changes are one of the most pronounced effects of climate change impacts on the Earth and are primarily driven by warming of the global ocean along with added water from melting land-based glaciers and ice sheets. In addition to these effects, changes in ocean salinity also can affect the height of the sea, by changing its density structure from the surface to the bottom of the ocean.
The team found that there was a long-term (1950-2008) pattern in halosteric (salinity-driven) sea level changes in the global ocean, with sea level increases occurring in the Pacific Ocean and sea level decreases in the Atlantic. These salinity-driven sea level changes have not been thoroughly investigated in previous long-term estimates of sea level change. When the scientists contrasted these results with models, the team found that models also simulated these basin-scale patterns, and that the magnitude of these changes was surprisingly large, making up about 25 percent of the total sea level change.
“By contrasting two long-term estimates of sea level change to simulations provided from a large suite of climate model simulations, our results suggest that salinity has a profound effect on regional sea level change,” Durack said. “This conclusion suggests that future sea level change assessments must consider the regional impacts of salinity-driven changes; this effect is too large to continue to ignore.”
Other collaborators include LLNL’s Peter Gleckler, along with Susan Wijffels, an oceanographer from Australia’s Commonwealth Scientific and Industrial Research Organization (CSIRO). The study was conducted as part of the Climate Research Program at Lawrence Livermore National Laboratory through the Program for Climate Model Diagnosis and Intercomparison, which is funded by the Department of Energy’s Regional and Global Climate Modeling Program.
Long-term sea-level change revisited: the role of salinity
Paul J Durack, Susan E Wijffels and Peter J Gleckler
Of the many processes contributing to long-term sea-level change, little attention has been paid to the large-scale contributions of salinity-driven halosteric changes. We evaluate observed and simulated estimates of long-term (1950-present) halosteric patterns and compare these to corresponding thermosteric changes. Spatially coherent halosteric patterns are visible in the historical record, and are consistent with estimates of long-term water cycle amplification. Our results suggest that long-term basin-scale halosteric changes in the Pacific and Atlantic are substantially larger than previously assumed, with observed estimates and coupled climate models suggesting magnitudes of ~25% of the corresponding thermosteric changes. In both observations and simulations, Pacific basin-scale freshening leads to a density reduction that augments coincident thermosteric expansion, whereas in the Atlantic halosteric changes partially compensate strong thermosteric expansion via a basin-scale enhanced salinity density increase. Although regional differences are apparent, at basin-scales consistency is found between the observed and simulated partitioning of halosteric and thermosteric changes, and suggests that models are simulating the processes driving observed long-term basin-scale steric changes. Further analysis demonstrates that the observed halosteric changes and their basin partitioning are consistent with CMIP5 simulations that include anthropogenic CO2 forcings (Historical), but are found to be inconsistent with simulations that exclude anthropogenic forcings (HistoricalNat).
Full PDF: http://iopscience.iop.org/1748-9326/9/11/114017/pdf/1748-9326_9_11_114017.pdf
Figure 1. Long-term trends in 0–2000 dbar total steric anomaly (left column; (A1)–(C1)), thermosteric anomaly (middle column; (A2)–(C2)) and halosteric anomaly (right column; (A3)–(C3)). Units are mm yr−1. Observational maps show results from (A) Ishii and Kimoto (2009; 1950–2008), (B) Durack and Wijffels (2010; 1950–2008) and (C) the CMIP5 Historical multi-model mean (MMM; 1950–2004). Stippling is used to mark regions where the two observational estimates do not agree in their sign ((A1)–(A3), (B1)–(B3)) and where less than 50% of the contributing models do not agree in sign with the averaged (MMM) map obtained from the ensemble ((C1)–(C3)). Results presented in columns 1–3 (above) relate to equations (1)–(3) presented in section 2