AUSTRALIAN researchers are treating with caution new research that suggests a new central Pacific form of El Nino, called El Nino Modoki, is on the increase.
The joint Korean-United States research indicates that because of global warming, El Nino Modoki may become five times more common late this century, forcing more drought over Australia.
Occurrences of the Modoki phenomenon have become increasingly common over recent decades, the paper says.
The typical El Nino phenomenon occurs when waters in the eastern Pacific warm and change atmospheric circulation in ways that make Australia drier and the Americas wetter.
Dr Scott Power, principal research scientist with the Bureau of Meterology (BoM), believes that the El Nino Modoki theory needs further scrutiny before any conclusions can be drawn.
The hypothesis is based on analysis on only a small number of climate models, Dr Power said, and draws on data extending back before WWII, when reporting was less reliable.
"Another caveat is that while the climate models have come a long way, and are getting very good, it's still hard to reproduce an El Nino event," Dr Power said.
While he believes the research is "interesting and important", he says it should be regarded as a preliminary investigation that needs rigorous testing before the effect of El Nino Modoki can be properly assessed.
A joint CSIRO-BoM effort is underway to develop a new climate model that improves forecasting of El Nino events.
Currently forecasters are grappling with a very rare phenomenon in which El Nino conditions prevail at the same time as the waters of the Coral Sea, to Australia's north, are also warm.
Because very little statistical data exists, traditional statistically-based forecasts may be less relevant under such unusual conditions. The forecasting community has been saying for some years that due to global warming, unusual conditions are becoming more frequent, rendering the past an ineffective guide to the future.
The new dynamic model, part of an ongoing initiative to supplement statistical-based forecasting, is scheduled to go through its first runs in 2010.