Item

Abstract

The subpolar North Atlantic is a hotspot with respect to predictability of sea surface temperature (SST) and upper-ocean heat content, as well as one of the regions worldwide, where initialisation with an observation-based state can further improve this predictability. In contrast to longer-term SST variability, high-frequency, impact-relevant extreme events are still challenging to predict by state-of-art decadal prediction systems. One such event in the eastern subpolar North Atlantic is the record-cold anomaly in summer 2015, often referred to as "Cold Blob". We analyse ensemble prediction experiments with the MPI-ESM1.2 "eddy-resolving" climate prediction system initialized in November 2013 and 2014 to demonstrate that individual ensemble members can reforecast strength and extent of the 2015 summer record "Cold Blob". About half of the members initialized in November 2014 and one third of the members initialized in November 2013 reforecast cold SST conditions in the entire eastern subpolar North Atlantic, with maximum anomalies reaching a similar magnitude as in observations. With respect to upper-ocean heat content, we expect an even higher skill. As in the real world, strong surface heat loss in the (two) winter preceding winters plays a key role in maintainig and deepening the subpolar cold anomaly in most ensemble members reforecasting the "Cold Blob". In some ensemble members, oceanic drivers such as heat transport divergence seem to be important in shaping subpolar SST anomalies. Individual ensemble members reforecasting the "Cold Blob" also reforecast the associated European summer heat wave.