Evaluation and Application of the ETA Adjoint Model

美国国家科学基金会和美国国家海洋和大气管理局（NOAA）建立了一个联合项目，以支持双方共同感兴趣的研究。一般来说，这项计划支持的研究包括一些有趣的研究工作，满足NOAA的需要，以及那些可以从NOAA国家环境预测中心的资源和计算机模型中受益的研究。该研究将在该联合项目的支持下进行，旨在评估和使用NOAA的Eta预测模型及其在NOAA NCEP开发的相关伴随模型。中尺度Eta模式的不断发展使得评估以前没有充分记录的中尺度天气现象成为可能，特别是在美国西部，那里的中尺度大气运动与底层复杂地形紧密耦合。首席研究员将建立和扩展目前正在进行的研究，并继续评估美国西部中尺度Eta模态的性能。Eta模型的伴随已在NCEP开发，用于使用称为四维变分（4DVAR）数据同化的高级数据同化方法。几个案例研究表明，4DVAR同化可以提高Eta模式预测的预测技能。随着NCEP开始在常规基础上测试这种数据同化程序，将向模型开发人员提供额外的信息，并最终向业务人员提供关于模型预测对初始观测误差和其他因素的敏感性的信息。本研究的主要目标是参与Eta伴随模型和相关数据同化方法的开发和应用。具体地说，作为数据同化方法的一部分，根据初始条件和伴随灵敏度获得的解将作为不同天气情况的函数进行评估。＊＊＊

9626380 Horel The National Science Foundation and the National Oceanic and Atmospheric Administration (NOAA) have established a joint program to support research of mutual interest. Generally, research supported under this program involves activities that are interesting research efforts, address a need of NOAA, and ones that can benefit from the resources and computer models of the NOAA National Center for Environmental Predictions. This research will be supported under this joint program and is aimed at evaluating and using the NOAA Eta forecast model and its associated adjoint model that have been developed at the NOAA NCEP. The ongoing development of the mesoscale Eta model makes it possible to evaluate mesoscale weather phenomena that have not been adequately documented before, especially in the western United States, where mesoscale atmospheric motions are tightly coupled to the underlying complex terrain. The Principal Investigator will build and expand on research underway now and continue to evaluate the performance of the mesoscale Eta modal over the Western United States. The adjoint of the Eta model has been developed at NCEP for using an advanced data assimilation method know as 4-dimensional variational (4DVAR) data assimilation. 4DVAR assimilation has been shown in several case studies to lead to improved forecast skill for Eta model forecasts. As this data assimilation procedure begins to be tested on a routine basis at NCEP, additional information will be available to model developers and, eventually, operational personnel regarding the sensitivity of the model forecasts to initial observational error and other factors. The major goal of the proposed work is to participate in the development and application of the Eta adjoint model and associated data assimilation methods. Specifically, the solutions in terms of both the initial conditions and adjoint sensitivities obtained as part of the data assimilation method will be evalua ted as a function of differing synoptic situations. ***