Retrospective monitoring of salinity in coastal waters with mussel shells.

Sea surface salinity (SSS) is a key parameter to understand and predict many physical, chemical and biological processes in dynamic coastal environments. Yet, in many regions, instrumental measurements are spatially sparse and insufficiently long, hindering our ability to document changes, causes, and consequences of SSS across different time scales. Therefore, there is an need to develop a robust proxy to extend SSS records back in time. Here, we test whether SSS can be reconstructed reliably and quantitatively from shell oxygen isotopic ratios (δO) of the mussel Mytilus galloprovincialis (Lamarck, 1819) in Otsuchi Bay, Northern Japan. δO ratios vary spatially and temporally and exhibit strong linear correlations with both sea surface temperature (SST) and SSS measurements, indicating that the composite signal recorded by δO measurably responds to variations in both parameters. By combining contemporaneous variations of SST and δO, SSS records encoded into mussel shells are deconvolved that significantly correlate with in situ SSS values. To further validate the robustness of δO as a quantitative SSS proxy, high-resolution and temporally aligned time-series of δO-derived SSS are reconstructed that are highly synchronous with the instrumental records. In particular, two lowered SSS scenarios occur concomitantly with periods of the summer monsoon and typhoon events. δO-derived SSS time-series are also comparable to those from numerical modeling. In conclusion, our findings demonstrate that mussel δO signatures can be used as a useful tool to construct high-resolution records of SSS in the coastal regions.