Development of a new parameterization of tropical convection

A convective cloud is one which grows vertically. Examples include fair weather cumulus clouds, and cumulonimbus clouds which grow into thunderstorms. These clouds generate almost all of the rainfall in the tropics, and most of the summer rainfall in mid-latitudes. The small size of these clouds makes them difficult to represent in global weather forecast models. These models divide the atmosphere into grid boxes with a typical horizontal length of 100 km. Not surprisingly, the rainfall generated by convective clouds is also very difficult to forecast. In fact, current weather forecast models are unable to forecast tropical rainfall more accurately than very primitive statistical methods. Much of the rainfall in the tropics occurs in association with the movement of large coherent groups of thunderstorms, of which hurricanes are one example. One of these coherent groups, known as the MJO, has been shown to have a significant influence on mid-latitude 1 - 2 week forecasts. This research is designed to improve the way global models currently represent the effects of convective clouds. These improvements should lead to better rainfall forecasts. Convective clouds also have a strong influence on the concentration of water vapor in the atmosphere. They inject moist, saturated air into the upper troposphere, and generate rainfall which evaporates and moistens the atmosphere as it falls. Water vapor is also the most important greenhouse gas in the atmosphere. Its concentration is likely to change in response to current and future increases in carbon dioxide. Uncertainty over the magnitude of this change is perhaps the dominant source of uncertainty in the projections of global climate models. Improvements in the way models represent convective clouds should therefore also improve the climate predictions of these models.