Radar in atmospheric physics and in forecasting

对风暴演变的正确短期预测在很大程度上取决于我们将观测，特别是雷达观测引入目前用于预报的数值模式的能力。为此，已开发出一种数据同化技术，允许将当前观测与过去发起的数值预报相结合，从而使预报被所有可用的雷达观测修正。要使这项技术成功，降水过程的模式描述必须与雷达观测相兼容。该建议的目标是优化代表降水粒子之间相互作用的方程(描述雨、雪、融雪、冰雹等性质的物理过程)。在云模型中称为微物理的参数化)，以便使雷达观测可用这些方程描述，并使数据同化对测量和物理过程描述中的不确定性尽可能不敏感。我们之前从雷达数据中对降水物理的研究将指导我们的探索。这项工作的出发点是用粒度分布的尺度描述来描述微物理，由此可以只使用分布的矩来进行完全的参数化，而不需要参考粒度分布本身。雷达观测也可以用分布的矩来描述，因此，将使用接近雷达观测矩的预报矩来进行参数化。我们将通过发展雷达观测的数学描述来推导微观物理方程，并从这些方程中推导出参数化。我们将在雷达观察到的简单情况下测试我们的开发，在这些情况下，每个过程都可以合理地隔离。

The correct short term prediction of the evolution of a storm is critically dependent on our ability to introduce observations, particularly radar observations, into the numerical models that are presently used for forecasting.  For this a data assimilation techniques have been developed that allow to combine current observations with numerical forecast initiated in the past so that the forecast is corrected by all available radar observations.  For this technique to be successful model description of the precipitation processes and radar observations must be compatible.  The objective of this proposal is to optimise the equations that represent the interactions between precipitating particles (the physical processes that describe properties of rain, snow, melting snow, hail, etc., called parameterization of microphysics in cloud models) in such a manner as to make radar observations describable by these equations and make the data assimilation as insensible as possible to uncertainties in measurements and in the description of the physical processes. Our previous studies of the physics of precipitation from radar data will be guiding us in our quest. The starting point of this work is a formulation of microphysics in terms of a scaling description of particle size distribution from which a full parameterization can be made using only the moments of the distributions without any reference to size distribution itself.  Radar observations are also described by the moments of the distribution, thus, the parameterization will be made using the prognostic moments close to the radar observed moments.  We will derive the microphysical equations by developing mathematical descriptions of radar observations and from these derive the parameterization. We will test our developments on simple situations observed by radar in which each process can be reasonably isolated.