GRAPES全球模式静力平衡奇异向量改进及应用试验

The improved computation scheme for singular vectors based on hydrostatic equilibrium and application experiments using the GRAPES global model

中文关键词:  奇异向量  静力平衡关系  GRAPES模式  集合预报

英文关键词:Singular vectors  Hydrostatic equilibrium  GRAPES model  Ensemble forecasts

中文摘要:

      采用线性化物理过程方案的GRAPES全球模式奇异向量在进行非线性模式积分时会有部分奇异向量出现崩溃问题，这说明奇异向量结构可能存在扰动变量之间不协调之处，需要对奇异向量扰动的计算方法优化，进而改进基于奇异向量的集合预报初值扰动，提高GRAPES全球集合预报效果。基于原有的GRAEPS全球奇异向量计算方法，在求解奇异向量时，对气压扰动的处理进行改进，将初始时刻的气压扰动分量通过位温扰动根据静力平衡关系导出获得，其他保持一致，发展了静力平衡奇异向量改进方法。基于有两个台风过程的个例（2019年8月8日12时（世界时）），分别采用原奇异向量方法和静力平衡奇异向量改进方法进行热带气旋目标区奇异向量的计算求解，并进行相应奇异向量的非线性模式积分，对比分析奇异向量非线性积分的稳定性。进而，对比分析奇异向量求解方法改进前、后热带气旋奇异向量的结构特征和初值扰动特征，开展了集合预报试验，评估改进后的奇异向量求解方法对GRAPES全球集合预报系统预报性能的影响。试验结果表明，静力平衡奇异向量改进方法通过产生协调的气压扰动和位温扰动场，解决了奇异向量非线性积分崩溃的问题，消除了原来不利于积分稳定性的气压扰动过于局地化的小尺度结构。静力平衡奇异向量改进方法对奇异向量中位温扰动分量和纬向风扰动分量结构影响较小，使得气压扰动分量的大值区位于台风附近，更好地描述热带气旋初值不确定性，与位温扰动分量的分布更加协调。采用静力平衡奇异向量改进方法，可以提高GRAPES全球集合预报在北半球和南半球等压面要素集合预报技巧和中国地区24 h累计降水概率预报技巧，增大台风路径集合离散度。

英文摘要:

      The singular vectors (SV) based on linearized physical processes have been applied in the Global and Regional Assimilation PrEdiction System (GRAPES) global ensemble forecast system to generate initial perturbations. However, the nonlinear integration of some singular vectors may occasionally break down. This means that there exist discordances between different perturbed variables and the method for SV computation should be improved, which could further improve the SV-based initial perturbations and the GRAPES global ensemble forecasts. Based on the original computation scheme for the GRAPES global SV, when computing SV, the pressure perturbation component at the initial time is obtained through the potential temperature perturbation component according to the hydrostatic equilibrium relation, and others remain the same. Since the tropical cyclone SV (TCSV) are sensitive to the linearized physical processes, one single case that consisted of two typhoon processes is selected for the present study (the initial time is 12:00 UTC 8 August 2019). TCSV are respectively computed with the original SV computation scheme and the improved computation scheme of SV based on the hydrostatic equilibrium, and the nonlinear integrations of the TCSV are given to analyze the stability of the nonlinear integrations for the TCSV. The structure characters of the TCSV and the initial perturbations are further analyzed, the ensemble forecasting experiments are conducted, and the influences of the improved computation method of SV on the GRAPES global ensemble prediction skills are investigated. The experiments show that by generating coordinated pressure perturbation and potential temperature perturbation, the improved computation method of SV based on the hydrostatic equilibrium solves the breakdown problem of the nonlinear integrations of SV and eliminates the original small-scale structure that is too local and may result in nonlinear integration instability. The improved computation method of SV based on the hydrostatic equilibrium has slight effects on the structures of potential perturbation and zonal wind perturbation. It makes the maximum value of pressure perturbation locate near the tropical cyclone, which can better describe the initial uncertainty of the tropical cyclone and make the pressure perturbation and the potential temperature perturbation more coordinating. Using the improved computation method of SV based on the hydrostatic equilibrium in the GRAPES global ensemble forecasts, the Northern and Southern isobaric variables ensemble forecasts and the probability forecast skill of 24 h accumulated precipitation in China can be improved, and the tropical cyclone track ensemble spread is also increased.