On the potential of extratropical SST anomalies for improving climate predictions