Improving the Prediction of Tropical Precipitation using a new Convective Parameterization

点击翻译按钮获取中文摘要

Climate models are the main tools used to predict how the earth's atmosphere will respond to the current increases in carbon dioxide. Climate models are widely used in seasonal weather forecasting, such as in the prediction of droughts and cold winters. Although climate models have been very successful, they do have important weaknesses. These weaknesses originate from difficulties in trying to represent processes that are turbulent or involve clouds. Convective clouds are especially difficult to represent in climate models. They are the main cloud type in the tropics and are very common in mid-latitudes during spring and summer. They range in size from small fair weather cumulus to thunderstorms, and are often organized into larger patterns such as squall lines and hurricanes. With current computing power, it is impossible to represent the vertical circulations of individual convective clouds within a climate model. Instead, climate models resort to approximate methods called parameterizations. To make a successful convective rainfall forecast, a convective parameterization must accurately simulate the factors which determine the average growth rate of the convective clouds inside a model grid box. However, the growth rates of convective clouds are extremely sensitive to a large number factors in their environment. Convective clouds modify the background atmosphere in ways which suppress and enhance the growth rates of neighboring clouds. Because of these difficulties, climate and weather forecast models perform most poorly in regions where convective clouds are most frequent. The purpose of this research is to improve the representation of convective clouds in climate and weather forecast models. If successful, the research can be expected to improve the forecasting of convective precipitation, better predict the impacts of extreme weather events, generate more credible climate projections, improve the forecasting of seasonal precipitation, and by improving the way climate models couple to land surface models, improve the prediction of quantities important to agriculture such as soil moisture.