Ensemble Forecasting of Convective Weather Events Using a Mesoscale Model

小行星9501532 由于大气物理的非线性，大气的确定性预测有其局限性。 一种被称为集合预报的技术被认为提供了一种改进大气预报的机制。 包围预报涉及通过合理地改变模型初始条件或模型物理条件来运行一系列数值预报。 其目的是提供预测，确定可能的解决方案的范围。 这些技术已显示出改善预测的前景。 中尺度气象学研究合作研究所和科罗拉多大学提出了一个联合项目，目的是开发和测试中尺度模式的集合预报技术。 拟议的研究将集中在准确预报中尺度对流天气事件所涉及的困难。 研究的三个具体目标是：1）发展一种技术来修改中尺度模式初始条件，以定义集合组，该集合组使用最少数量的模拟来最大化可预报相空间上模式解的跨度，同时要求集合初始条件保持约束于可用观测的误差特性; 2）探索这种技术在强和弱强迫大尺度环境下的结果; 3）确定在集合预报方法中使用变化模式物理的适用性。 从这项研究中获得的知识将提供必要的信息，如何有效地产生集合预报使用中尺度模式。 预计模式的结果也将提供一些指导预测模式的技能，为给定的对流天气事件。 ***

9501532 Warner Due to nonlinearities in the physics of the atmosphere, deterministic predictions of the atmosphere have their limitations. A technique known as ensemble forecasting is believed to provide a mechanism for improving forecasts of the atmosphere. Ensemble forecasting involves running a series of numerical predictions by changing, within reason, model initial conditions or model physics. The objective is to provide forecasts that define the range of possible solutions. These techniques have shown promise for forecast improvement. A joint project between the Cooperative Institute for Mesoscale Meteorological Studies and the University of Colorado is proposed with the objective to develop and test ensemble forecasting techniques for mesoscale models. The proposed research will focus upon the difficulties involved in accurately predicting mesoscale convective weather events. Three specific objectives of the research are: 1) to develop a technique to modify the mesoscale model initial conditions to define the ensemble group that maximizes the span of model solutions across the predictability phase space using a minimum number of simulations, while requiring that the ensemble initial conditions remain constrained to the error characteristics of the available observations; 2) to explore the results of this technique for both strongly and weakly forced large-scale environments; 3) to determine the applicability of using varying model physics in ensemble forecasting methods.. Knowledge gained from this study will provide necessary information on how to efficiently produce an ensemble forecast using a mesoscale model. It is anticipated that the model results also will provide some guidance on predicting the model skill for given convective weather events. ***