Subpolar North Atlantic overturning: the 1990s versus the 2010s

The subpolar North Atlantic exhibits pronounced variability on decadal and longer time scales, which has implications for decadal predictability of climate and marine biogeochemical fields. Yet, its driving mechanisms are still under debate. It has been shown from both observations and modeling that this variability is associated with anomalous deep water formation in the Labrador Sea generated by the surface heat fluxes associated with the North Atlantic Oscillation (NAO) and the consequent adjustment of thermohaline circulation, most evident during the 1990s when the NAO was persistently positive. However, the direct observations of overturning circulation along the sections east and west of Greenland (Overturning in the Subpolar North Atlantic Program; OSNAP) do not support this view, as observed overturning in the Labrador Sea is very weak under positive NAO conditions during the observed period from 2014 onward. In this study, we use high-resolution (0.1°) forced ocean–sea-ice simulations, which reasonably reproduce the mean overturning in density coordinates observed at the OSNAP line, to elucidate this contrasting overturning between the two periods under similar positive NAO conditions. Simulated deep water formation in the Labrador Sea is indeed weak during the 2010s, while it is very active during the 1990s. These signals are meridionally coherent, suggesting coherent changes in Atlantic meridional overturning circulation (AMOC). We also find that this anomalous overturning in the Labrador Sea takes place over densities far heavier than the density where maximum overturning occurs, thus the maximum overturning time series cannot accurately capture these signals. We have conducted sensitivity experiments to identify whether the weak overturning during the 2010s is due to oceanic or atmospheric conditions. These experiments reveal that the weaker overturning is largely generated by weak surface heat release due to a warmer air temperature over the Labrador Sea. We have further performed composite analyses using the CESM2 pre-industrial and transient (historical plus SSP370) simulations to investigate how such warm air conditions come about over the Labrador Sea. The composite analyses suggest that the warm air temperature is likely due to a warm SST condition in the Labrador Sea, internally generated, rather than externally forced. Conversely, the strong overturning during the 1990s was likely because of cooler conditions in the Labrador Sea.