The influence of the lower boundary SST conditions, in both the Pacific and the Atlantic sectors, is found also to be distinct and highly non-linear. While a local warm SST anomaly promotes TC activity, a local cool anomaly does not play a role in inhibiting it. By the same token, while a remote Pacific warm SST anomaly produces unfavorable atmospheric conditions, a cool La Nina event does not play the opposite role of promoting the Atlantic TC activity. An implication of these findings is that the conventional linear correlation analysis, or other linear tool such as the empirical orthogonal function analysis, is most likely to be inadequate to address and depict this highly non-linear phenomenon.
The long-lead predicative capability of TC's interannual variability by employing SST's antecedent modes of variability is carefully compared with that of the other variables, such as sea level pressure, low- and upper-level flows, and the vertical wind shear. The former variable appears to be a better contender. In general, the SST precursor is already in place in April or May, way ahead of the ensuing peak season of the TC activity, which falls in August to October.