The history of the ENSO cycle is known fairly well since 1879, the beginning of a sufficiently detailed and reliable instrument record. Indirect methods are methods are needed to deduce El Niño events before 1879. There are various methods, ranging from anecdotal evidence to isotope ratios in corals.
Tree ring chronologies
Perhaps the most detailed information arises from tree-ring chronologies from areas whose climate correlates strongly with the ENSO cycle. Tree growth rate in northwestern Mexico and southern Texas is mainly controlled by soil moisture, and especially winters tend to be wetter there in El Niño years. Therefore it is no surprise that the December-February average SOI values for 1879-1996 correlates well with tree growth there. The trees used here were Douglas fir, ponderosa pine and pinyon pine. A similar technique has been used with teak trees on Java, Indonesia, where an El Niño implies dry conditions. The negative correlation between tree ring widths in North America and Indonesia demonstrates the occurrence of ENSO’s since 1706 (1). Overall there appear to rhythms of about 4 and about 6 years in the inferred SOI values. The new chronology agrees with that constructed by Quinn, but not very well (2).
El Niño events before the 20th century
The list of El Niño events based on tree ring comparisons shows 22 cold events during 1706-1878, compared to 25 in the much shorter period 1879-1977 (1). There was a similar (but less dramatic) increase in the number of warm events after 1878. Also the variation of winter SOI seems to have increased notably since that date.
While the correlation between ENSO and monsoon activity in India is weak, every major drought in peninsular India between 1526 - 1900 appears to be followed, about half a year later, by an El Niño in the east Pacific (3). The linkage appears to be as follows: higher-than-normal sea-level pressure over inner Asia blocks the progression of the south-east Asian monsoon. That leads to a rise in sea-surface temperatures in the Indian and then the Pacific oceans, possibly initiating an El Niño in the east equatorial Pacific. The correlations which indicate this chain of events mean that southeast Asian and (especially) Indian monsoon failures accompany or foreshadow strong El Niño events (4). However during the 20th century these correlations broke down.
A particularly strong El Niño in 1789-91 affected newly colonised Australia; the enterprise was almost ended with the drying up of the Tank Stream, which provided water for the settlement at Sydney. At the same time in France there was an unusually cold winter then some of the worst harvests on record a year before the El Niño, in 1788, fomenting the French Revolution. Incidentally, El Niño events tend to occur after a weak phase of the North Atlantic Oscillation, i.e. more meridional winds, which bring colder weather in Europe. There was also a prolonged drought in southern Africa, and one in Andhra Pradesh in India which occurred more than a year before droughts in Australia and Mexico. Reduced rainfalls in Ethiopia reduced flows in the Nile to very low levels throughout 1790--97.
A similar El Niño in 1200 BC can be blamed for the sudden end of the Hittite empire and for the Trojan Wars.
(1) Stahle, D.W., R.D. D’Arrigo et al. 1998. Experimental dendroclimatic reconstruction of the Southern Oscillation. Bull. Amer. Meteor. Soc., 79, 2137-52.
(2) Quinn, W.H., V.T. Neal & S.E. A. de Mayola 1987. El Nino occurrences over the past four and a half centuries, J. Geophys. Res.,92, (C13), 14,449-61.
(3) Grove, R. 1998. Global impact of the 1789-93 El Nino. Nature, 393, 318-9.
(4) Normand, C. 1953. Monsoon seasonal forecasting. Quart. J. Roy. Meteor. Soc., 79, 463-73.