Ensemble Forecasting of Convective Weather Events Using a Mesoscale Model

9501532华纳由于大气物理的非线性，对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。集合预报包括通过合理地改变模型初始条件或模型物理来进行一系列数值预测。目标是提供确定可能解决方案范围的预测。这些技术已显示出改善预报的希望。提出了中尺度气象研究合作研究所和科罗拉多大学之间的一个联合项目，目的是开发和测试中尺度模式的集合预报技术。建议的研究将集中于准确预测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)开发一种修改中尺度模式初始条件的技术，以定义集合组，该集合组可以使用最少的模拟次数在可预测性相空间中最大化模型解的跨度，同时要求集合初始条件仍然受现有观测值的误差特征的约束；2)探索该技术在强强迫和弱强迫大尺度环境中的应用结果；3)确定变模式物理在集合预报方法中的适用性。从这项研究中获得的知识将为如何使用中尺度模式有效地产生集合预报提供必要的信息。预计模式结果还将为预测给定对流天气事件的模式技能提供一些指导。* * *

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. ***