In this chapter we review what is known about ENSO as recorded in paleoclimatic proxies, and what is expected for ENSO as we enter a climate state altered by anthropogenic greenhouse gases. In neither case can we confidently construct a reliable picture from instrumental data; in one case we draw inferences from proxies to reconstruct what the climate was, and in the other we rely on imperfect models to foretell the future.

Our knowledge of ENSO in the paleoclimate record has expanded rapidly from the late 1990s. The ENSO cycle is present in all relevant records, going back 130 kyr (kilo-years) to the previous interglacial period (Hughen et al., 1999). It was systematically weaker during the early and middle Holocene (the last 10 000 years), and, as we shall see, model studies indicate that this results from reduced amplification in the late summer and early fall, a consequence of the altered mean climate in response to boreal summer perihelion. Data from corals show substantial decadal and longer variations in the strength of the ENSO cycle within the past 1000 years; it is suggested that this may be due to solar and volcanic variations in solar insolation, amplified by the Bjerknes feedback. There is some evidence that this feedback has operated in the twentieth century.

All of us now anticipate a change in climate brought about by human activity. Among other things, we will have to adjust to a change in the year-to-year variations in climate.