We still don't know what causes the SST behavior. Predictions based on statistics is not true science. I recall things like volcanic vent activity at the ocean floor had been considered as a cause, without success. (It had been proposed that historical records of concomitant earthquakes might serve as a predictor, but a reliable correlation was not found.) I guess that still leaves us with statistics until a scientific predictor can be found. The SST variations have an economic impact, demanding further scientific research. Speaking from a science background: BA Physics, (1972, Knox College); MS Atmospheric Science (1976, under Bill Gray at CSU); NCAR data "steward" 1976-2011 (under Roy Jenne at NCAR/CISL/DSS).