2019-01-242024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/677893摘要：本計畫在中央氣象局高速電腦環境下建置Finite Volume Cubed-Sphere Dynamical Core (FV3)全球模式，以未來氣象局作業化上線為目標進行評估，發展具氣象局特色的全球預報系統。計畫第一年已對FV3程式碼及操作流程進行了解與分析，將FV3模式建置於氣象局HPC作業環境及完成其前置、後製處理程序，並完成氣象局分析資料場轉換。另外考量台灣地理位置測試三種網格配置，初步完成對FV3預報表現的評估。第二年將進一步使用NCEP版本的FV3模式，評估模式對一些重要歷史天氣事件的預報表現、探討動力過程，並評估準作業化(包含後處理及校驗過程)的工作流程、計算資源與計算時間。第三年將評估作業化流程之建置策略，開始測試串接資料同化系統後FV3的表現，且與目前CWB/GFS預報結果進行技術得分比較。本計畫預期將可連結國內外學術、作業及研究單位發展FV3模式之研究能量，培養國內自主發展、維護全球模式的能力，強化國內下一代全球模式預報系統研究人員的專業訓練，並協助氣象局發展全球預報系統。<br> Abstract: The Finite Volume Cubed-Sphere Dynamical Core (FV3) global model is plan to be operational in Central Weather Bureau’s (CWB) high-performance computing environments. The goal of this project is to evaluate and implement the FV3 model within the CWB operational environment. For the first year, the FV3 model has been studied and constructed in CWB. The format conversion process was also established for using CWB GFS analysis data in FV3 simulations. Three mesh configurations have been tested for the performance of 2017 August to 2017 September forecast experiments. In the second year, more typhoon cases in 2017-2018 will be selected to analyze the forecast performance and the dynamic processes of convection. The optimization for post processing, the computational demand, and verification system will be used in the assessment of the CWB’s high-performance computing environments. In the third year, this project will evaluate the forecast performance systematically through a data assimilation system in the base of CWB GSI data assimilation system,for the purpose of establish the operational process. Through the three years study and validate of FV3 model, we hope to enhance Taiwan’s own ability in researching and improving the numerical weather prediction for the next generation CWB GFS.中央氣象局全球模式預報系統FV3全球模式數值天氣預報CWB Global Forecast SystemFV3 global modelnumerical weather prediction