Recent observational studies show that these effects almost balance, but that the cooling effect is somewhat more important. From the point of view of global change, however, it is crucial to note that this small difference is about five times larger than the radiative effect anticipated from a doubling of atmospheric carbon dioxide (CO2), and that the individual components of the difference are orders of magnitude larger. In existing climate models about one third of the predicted warming due to increasing CO2 arises because of the predicted cloud changes. These predictions, however, are highly speculative because none of the models include interactive cloud physics.
Convection also plays an important role in the surface heat budget. The main mechanism whereby the surface of the Earth cools is not radiation, but evaporation. The heat lost by evaporation is carried deep into the atmosphere by convection and is realized as heat by the atmosphere when water vapor condenses into rain and snow. Much of this heat is released above 6 km. In existing models, another one third or so of the anticipated warming due to increasing CO2 arises because the amount of water vapor (the major greenhouse gas) increases due to increased evaporation with increasing temperature. Moreover, increased evaporation is likely to lead to increased convection.
But convection has two competing effects: increased convection forces increased subsidence in the environment of clouds, which is a strong drying effect; but increased convection also increases the rate at which water vapor from near the Earth’s surface is transported to higher altitudes. However the bulk of this water vapor condenses as it rises and falls out as precipitation leaving open how much is actually available to moisturize the atmosphere. Current representations of cumulus convection handle these processes crudely and the inclusion of more realistic representations of cumulus convection in climate models could alter significantly what is currently thought to be a major positive feedback. A major challenge is the observational validation of representation of cumulus convection.
Clearly, without a proper treatment of both layer clouds and convection, model predictions of climate are uncertain. Cloud effects are so much larger than the anticipated effects of added greenhouse gases, that small changes in the cloud picture can easily alter predictions of global warming. In addition, existing methods of representing convection and clouds are crude, and, in some cases, can be shown even to be qualitatively incorrect.
By utilizing theories and observations of cumulus clouds, CGCS researchers have constructed a new and significantly improved representation of cumulus convection. A comprehensive and systematic series of tests of this new scheme are now being performed to optimize the parameters of the scheme and to assess the magnitude of water vapor errors produced by this and other convective schemes. Sensitivity of atmospheric water vapor content to the parameters of this new representation of convection is also being evaluated through use of the so-called adjoint of the scheme.
The treatment of layer clouds is a more difficult problem at the moment, and demands a combination of expertise in large scale dynamics, cloud dynamics, and cloud physics. Comparable cloud covers exist under a wide variety of climate conditions, and thus, no obvious parameterization suggests itself. Careful studies of the dynamic and thermodynamic conditions that lead to the formation of layer clouds are crucial. We note that better treatment of layer clouds will, among other things, require better prediction of atmospheric water vapor content, which requires better representations of convection.
Observations in this area are also critical. CGCS researchers have extensive experience in satellite passive microwave observations of humidity, precipitation and temperature profiles, and in active radar observations of severe storms. The research emphasis of the CGCS is on understanding, since without it one is unable to foresee the likely feedbacks. Observations and theory together will provide this understanding. Efforts at MIT include cooperative projects with various modeling groups to assess both the behavior of convection and the processes that determine water vapor in these models.