The Earth’s climate is strongly affected by water vapor, which is a greenhouse gas and which also controls the activity of clouds, which reflect sunlight to space and which also act as greenhouse agents, absorbing infrared radiation and re-emitting some of it back to the surface. We know from observations that the distribution of water vapor is very complex, and also varies greatly in time.

               The source of water vapor is the earth’s surface:  oceans, lakes, and wet land surfaces. Turbulence in the atmosphere carries water vapor up a few thousand feet. But water vapor is transported much higher up into the atmosphere by deep cumulus clouds, such as thunderstorms, and the physics of those clouds controls how much water is put into the atmosphere versus coming back down as rain or snow.

               Since the real earth is very complicated, we try first to understand idealized climate states, one of which is called “radiative-convective equilibrium”, or “RCE”. This is just a hypothetical climate over an ocean of infinite horizontal extent, above which only radiation (both solar and terrestrial) and convection (cumulus clouds) are the only processes that transport energy through the atmosphere, and convection is the only process that transports water. In recent decades, we have been able to simulate such states over small chunks of the ocean using very powerful computers. When we first started doing this, we saw pretty much what we expected to see…  random cumulus clouds developing and decaying, a bit like watching boiling water.

               But about ten years ago, during one such simulation, all the convection grouped together into one massive clump, which stayed around for as long as the simulation was run. Since then, other researchers have replicated this process, which has become known as “self-aggregation”. My colleague Marat Khairoutdinov and I discovered that whether this happens or not depends, among other things, on how warm the underlying sea surface is.

               This NSF grant enabled us finally to unlock the key physical processes that lead to self-aggregation, and we explained clearly why the process depends on the sea surface temperature. We think that aggregation of convection also occurs in nature, and because it dries out the atmosphere profoundly, it might act as a tropical climate regulator. A drier atmosphere has a weaker greenhouse effect, so it cools.

               Along the way, we decided to see what happens if we repeated the experiments but this time on a rotating planet. (The earlier work was done on a nonrotating planet.) The aggregation, in this case, takes the form of hurricanes, and we were able to discover what precisely controls their size, their spacing, and their peak wind speeds. This may yield clues to how and why hurricanes form in nature, and how they might depend on the state of the climate.

Last Modified: 11/25/2014
Modified by: Kerry A Emanuel