I don't agree with this.. there is a +3 day-0 correlation, then it drops down to near 0, then it goes up to +10 at day+20 (20 days after 10mb warming). If the events saw a -NAO during initial warming didn't make a difference in the eventual lagged event. The correlation isn't perfect, but it's about 65-70% of time, indifferent to what happens at D+0.