Teleconnections within the tropical Pacific began to be understood thanks to the idealized calculations of A.E. Gill and later through more complex models. Building upon the "proto-model", much of the early theory of teleconnections dealt with barotropic, linearized model of atmospheric flow about a constant mean state. However, the model was soon invalidated when it was discovered that actual teleconnection patterns were nearly insensitive to the location of the forcing, in direct contradiction with the predictions offered by this simple picture. Simmons and collaborators showed that if a more realistic background state was prescribed, it would become unstable, leading to a similar pattern regardless of the location of the forcing, in accordance to observations. This "modal" property turned out to be an artifact of the barotropicity of the model, though it has appeared for more subtle reasons in more realistic models. More recent work has shown that most teleconnections from the tropics to the extratropics can be understood to surprising accuracy by the propagation of linear, planetary waves upon a 3-dimensional seasonally-varying basic state. Because the patterns are persistent over time and somewhat "locked" to geographical features such as mountain ranges, these waves are called stationary. Another mechanism of teleconnection between tropical oceans and midlatitude regions is symmetric along latitude circles and between hemispheres, unlike the stationary wave mechanism. It relies on interactions between transient eddies and the mean atmospheric flow that are mutually reinforcing. It has been shown to explain some aspects of ENSO teleconnections in temperature and rainfall. Other authors suggested, as well, a correlation between many teleconnection patterns and local climate change factors.
Applications
Since tropical sea surface temperatures are predictable up to two years ahead of time, knowledge of teleconnection patterns gives some amount of predictability in remote locations with an outlook sometimes as long as a few seasons. For instance, predicting El Niño enables prediction of North American rainfall, snowfall, droughts or temperature patterns with a few weeks to months lead time. In Sir Gilbert Walker's time, A strong El Niño usually meant a weaker Indian monsoon, but this anticorrelation has weakened in the 1980s and 1990s, for controversial reasons.