Cloud Parametrization
Cloud Parametrization
复制标题
云参数化
DOI：
--
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
A. Tompkins
中科院分区：
文献类型：
--
作者：
A. Tompkins
让我们通过研究 ECMWF 预报系统在代表大气层顶部 (TO A) 长波辐射收支方面的进展来介绍亚网格尺度参数化研讨会系列。图 1 比较了使用 ECMWF 集成预报系统根据观测到的海面温度进行的长达一年的集成。左组面板显示使用模型周期 23r4 的集成，该模型于 2000 年投入使用，并用于进行 ERA-40 再分析。右组面板显示与 2008 年投入运行的 CY33R1 循环的集成1。在八个中间年份中，误差已大大减少，全球平均 RMS 误差减少了近一半。虽然这种误差的减少无疑源于大规模模型动力学和平流的改进，但很大程度上是由于更好的亚网格尺度方案来表示对流、云和云-辐射相互作用过程的结果。如果对流方案在热带美洲和非洲大陆上产生的活动太少，就像在早期模型版本中那样，那么调整云方案调整参数以消除红外误差是极其困难的。同样，如果云方案的微观物理允许过多的冰从热带卷云中沉积出来，调整冰晶的辐射特性可能会导致其他地方的误差增加，例如两极上空的低冰云。如果假设的辐射特性或这些云的垂直重叠很差，那么云量和凝结水量的完美云特性仍然会导致辐射误差。显然，TOA IR 预算和其他云相关领域中的小错误只能通过所有子电网规模流程的协调改进来实现，2008 年 EC MWF 研讨会系列旨在介绍最近发生的一些进展。
Let us introduce the seminar series on subgrid-scale parame trization by examining the progress in the ECMWF forecast system at representing the top-of-atmosphere (TO A) longwave radiation budget. Figure 1 compares an ensemble of year-long integrations using the ECMWF integ ra d forecasting system forced by observed sea surface temperatures. The left set of panels shows an integr ation using model cycle 23r4, which was operational in 2000, and was used conduct the ERA-40 reanalysis. The righ t set of panels shows an integration with cycle CY33R1, operational in 2008 1. Over the eight intermediate years the errors have been grea tly duced, with the global mean RMS error reduced by almost one half. While so me f this error reduction undoubtedly derives from improvements in the large-scale model dynamic s and advection, much of it is a result of better subgrid-scale schemes to represent the processes of convec ti , clouds and cloud-radiation interaction. If the convection scheme produces too little activity over the Tro pical American and African continents, as it did in earlier model versions, then it is extremely difficult to adj ust cloud scheme tuning parameters to eradicate the IR error. Likewise, if the cloud scheme microphysics allow t oo much ice to sediment out of tropical cirrus clouds, adjusting the radiative properties of the ice cryst als will likely lead to increased errors elsewhere, such as low ice clouds over the poles. Perfect cloud properties of cloud cover and condensate amount will still result in radiative errors if the assumed radiative propert ies or vertical overlap of those clouds are poor. It is clear that small errors in the TOA IR budgets and other cloud r elated fields can only result from a harmonious improvement of all subgrid-scale processes, and the 2008 EC MWF seminar series aims to introduce some of the advances that have occurred recently.