Obtaining optimal fingerprints of the AMOC from sea surface observations

Records of Atlantic sea surface temperature (SST) and sea surface salinity (SSS) integrated over different areas have been proposed as fingerprints of the AMOC over the period of observations. Boers (2021) analyses eight different time series of fingerprints as AMOC indices and finds for all of them early warming signals of an approach to a tipping point. However, if two different AMOC indices are not strongly correlated, they can obviously not both be trustworthy representations of the AMOC. Thus, if significant EWSs are found in both of such records, at least one is spurious. In that case, it is questionable if EWSs for a forthcoming collapse can be trusted at all, since it is observed for unrelated reasons in (at least) one record. Here we propose a consistent fingerprint based on a combination of all the fingerprints.