On the basis of analyses of spatial patterns of Atlantic sea surface temperatures (SSTs) and salinity concentrations, together with four SST- and four salinity-based fingerprints of the Atlantic Meridional Overturning Circulation (AMOC) introduced in previous studies, Boers concluded that the stability of the AMOC may have declined during the past century. Chen and Tung raise three concerns in their comment: (1) the two AMOC fingerprints that are based on subpolar North Atlantic salinity fields have the wrong sign, (2) the positive trends of the restoring rate λ inferred in Boers1 and interpreted as a stability decline are caused by non-stationary coverage of observational data, and (3) a general criticism of SST-based AMOC fingerprints. Here I respond to each of their concerns, arguing that the fingerprints’ sign does not matter for the stability analysis and that their way to test the effect of non-stationary data coverage is flawed, and demonstrate that the autocorrelation and restoring rate λ are robust against non-stationary observational coverage. Note that the stability decline inferred by Boers remains statistically significant when accounting for the effects of non-stationary observational coverage and that it has since been confirmed on longer timescales using independent data. |
