It’s often discussed by climate scientists (who in many instances need to lay off the coffee or switch to decaf), that the subsurface temperatures ACTUALLY MEAN SOMETHING. However, the evidence (and calling it evidence is drawing a long bow) is rather rubbish. When the length of time of which the two records (subsurface verses surface) correlate with a useful 6 month lead, is similar to the length of time of which the two records (subsurface verses surface) correlate with a un-useful 0 month lead, the inescapable conclusion is that using sub surface data is no better than using analog years. How much of the forecast of slightly favoring el nino is due to a possible correlation fallacy?