ITR/AP: Collaborative Research: Diversifying Ensembles with Stochastic Convection

使用日益强大的计算机系统来改进数值天气预报的一个有效方法是通过集合预报。经验表明，集合预报的偏差不够大，因此大气的真实状态常常位于集合的扩散范围之外。在这个项目中，PI 试图通过对流过程的扰动来探索模型气候的可变性，从而扩大传播范围。该方法是在积云参数化方案中引入随机噪声来解释子网格尺度效应。波动的设计将基于来自云解析模型、雷达、飞机和无线电探空仪的高分辨率数据。这些想法将在具有不同类型复杂性的模型中实现。研究结果有可能为改进主要气象中心的集合预报系统提供新方法，并提高对现有积云参数化的理解。该项目是由博士之间合作完成的。科罗拉多大学 CIRES 的 Brian Mapes 和 Thomas Hamill 以及亚利桑那大学的 Steven Mullen 博士。

An effective way of using increasing powerful computer systems in improving numerical weather prediction is through ensemble forecasting. Experience shows that the ensemble forecasts do not diverge enough so that often the true state of the atmosphere lies outside the spread of the ensemble. In this project, the PIs seek to broaden the spread by exploring the variability of model's climate through the perturbation of the convective processes. The approach is to introduce stochastic noise in cumulus parameterization schemes to account for the sub-grid scale effects. The design of the fluctuations will be based on high-resolution data from cloud-resolving models, radars, aircraft, and radiosondes. These ideas will be implemented in models with different types of complexity. The results have the potential of providing new methods for improving ensemble forecast systems at major meteorological centers and also leading to improved understanding of existing cumulus parameterization. This project is being done collaboratively between Drs. Brian Mapes and Thomas Hamill, CIRES, University of Colorado and Dr. Steven Mullen, University of Arizona.