数值模式内物理参数化.PDFVIP

数值模式内物理参数化 个人的看法 潘华陆 The purpose of these lectures • Give you a flavor of how complicated the phenomena we try to parameterize are and that the only way to make the forecasts better is to build physically based schemes • Give you a personalized view of the problems and the solutions. What is the purpose of a parameterization package? • Because of the limited computer resource, we can not hope to simulate the atmosphere to the molecular level. Some important physics (e.g. heating) are needed to make the model simulation closer to the real atmosphere. • Convection and turbulence are two mechanism that have been considered necessary for the maintenance of the atmospheric circulation. Over the years we learned that complications of the earth surface and the complexities of the land also are important. In addition, cloud and radiation also make important contributions. While the model resolution increase seemed to give us better handle of the circulation of the large scale, errors made in the parameterization schemes continue to hamper our ability to simulate the atmosphere. Parameterization for weather and climate models • Traditionally, climate modelers worry about time mean of the simulation. The idea is that, if we can get the time mean right, we don’t need to know how it gets there. Over time, we are learning that how the model gets there is still important. In that, we mean the weather. Climate is still the average of the weather. It is important to get the weather statistics right for climate models as well. • Weather modelers, on the other hand, worry more about getting the timing and intensity of the weather events and less about the time mean. In that sense, weather modelers traditionally worry less about cloud-radiation feedbacks and more about the