A Physics Coupler for Climate Models

该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。气候在很大程度上取决于与云有关的过程。尽管如此，在当前的气候模式中，与云相关的过程通常只通过网格箱相互作用，而不是直接在小尺度上相互作用。例如，最好让小尺度云结构的全部可变性直接驱动降水过程，并让降水反过来直接改变小尺度云结构。缺乏这样的接口会导致各种云相关过程的模拟降级。为了满足这一需求，该项目将开发一种基于假设概率密度函数（PDF）方法的物理耦合器。假设的PDF方法通过提高局部过程速率来产生降水率的网格箱平均值。在过去，小尺度变异性的平均值是通过两种方法中的任一种进行的：（1）解析积分，或（2）拉丁超立方体采样，这是一种蒙特-卡罗积分。这两种方法都有缺点。研究人员将开发一种新的方法来减少噪音，同时允许使用复杂的降水方案。该技术将在气候模型中实施，并用于模拟云，并可能导致，在未来的工作中，一个广义的物理耦合器，将改善气候模拟的许多方面，这项工作的更广泛的影响包括与现实世界的软件工程本科院校的伙伴关系，广泛的研究生培训和教学，并在社区传播研究成果，造福社会。虽然这项研究旨在立即应用于气候模型中的降水过程，但这些技术将来可能应用于其他模型类型，如数值天气预报模型和其他物理现象，如大气化学。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). Climate depends strongly on processes related to clouds. Despite this, in current generation climate models, cloud-related processes typically interact with each other only through grid box means and not directly with each other on the small scales. For instance, it would be desirable to let the full variability in small-scale cloud structure directly drive precipitation processes and let precipitation in turn directly alter the small-scale cloud structure. The lack of such an interface leads to degraded simulations of a variety of cloud related processes. To address this need, the project will develop a physics coupler based on the assumed probability density function (PDF) method. The assumed PDF method produces grid box averages of precipitation rates by upscaling local process rates. In the past, the averaging over the small-scale variability has been performed by either of two methods: (1) analytic integration, or (2) Latin hypercube sampling, which is a type of Monte-Carlo integration. Both methods suffer drawbacks. The investigator will develop a new approach to reduce noise while permitting the use of sophisticated precipitation schemes. The technique will be implemented in a climate model and used to simulate clouds, and may lead, in future work, to a generalized physics coupler that will improve many facets of climate simulations.Broader impacts of this work include partnership with an undergraduate institution in real-world software engineering, broad training of graduate students and teaching, and disseminating research in the community and benefiting society. Although the research is intended to have immediate application to precipitation processes in climate models, the techniques may be applied in the future to other model types, such as numerical weather prediction models, and other physical phenomena, such as atmospheric chemistry.