Increasingly powerful thunderstorms may be largely responsible for greater rainfall in the tropics as the Earth warms, new research has shown.
Thunderstorms are known to play an important role in keeping the tropics wet, delivering almost half of all the rainfall in equatorial regions.
Climate models have predicted that global warming will lead to more precipitation, especially in parts of the world that are already wet and warm.
The new study, based on satellite observations, showed that the frequency of big and powerful thunderstorms increased in the tropics between 1998 and 2009 and was strongly correlated with observed rainfall changes.
Lead researcher Dr Jackson Tan, from the Australian Research Council Centre of Excellence for Climate System Science (ARCCSS) said: "The observations showed the increase in rainfall is directly caused by the change in the character of thunderstorms in the tropics rather than a change in the total number of thunderstorms.
" What we are seeing is more big and organised storms and fewer small and disorganised storms."
The trend indicated a "dynamic change" in the climate system.
Large thunderstorms emerge from "organised deep convention" systems and involve powerful currents of warm, moist air.
Dr Tan added: "If this rainfall change was caused simply by a warmer atmosphere holding more moisture, we would have expected an increase in the average rainfall when each system, organised or disorganised, occurs.
"Instead, the number of organised storms, which is largely controlled by the dynamics of the atmosphere, have increased in frequency, suggesting that the increase in rainfall is related to more than a simple warming of the atmosphere."
The findings, published in the journal Nature, could help explain why climate simulations have difficulty accurately representing the details of tropical rainfall.
Current computing power is insufficient to simulate the small-scale processes that give rise to thunderstorms.
Co-author Professor Christian Jakob, from Monash University in Australia, said: "This limitation, which is a well-known issue in global climate models, might well be a contributing factor to the precipitation errors and the bias towards light rain.
" Given how important these large storms are to rainfall in the tropics, it is vital that there is a renewed effort to represent convective organisation in global climate models if we are to fully understand precipitation changes in the future."