基于CMA-GEPS的延伸期预报能力评估

齐倩倩; 朱跃建; 陈静; 李晓莉; 田华; 汪叶

doi:10.16032/j.issn.1004-4965.2024.050

基于CMA-GEPS的延伸期预报能力评估

doi: 10.16032/j.issn.1004-4965.2024.050

齐倩倩^{1, 2, 3},
朱跃建⁴,
陈静^{1, 2, 3, ,},
李晓莉^{1, 2, 3},
田华^{1, 2, 3},
汪叶⁵

1.
中国气象局地球系统数值预报中心，北京 100081
2.
灾害天气国家重点实验室，北京 100081
3.
中国气象局地球系统数值预报重点开放实验室，北京 100081
4.
美国国家海洋和大气管理局/国家气象局/国家环境预报中心/环境模式中心，美国马里兰州 20742
5.
河南大学数学与统计学院，河南开封 475004

基金项目:

国家自然科学基金面上项目 42375152

国家自然科学基金青年基金项目 41906022

国家自然科学基金青年基金项目 42105057

中国气象局地球系统数值预报中心青年基金 CEMC-QNJJ-2021005

详细信息

通讯作者:
陈静，女，研究员，四川省人，主要从事集合预报技术研发。E-mail：chenj@cma.gov.cn

中图分类号: P435
计量
- 文章访问数: 105
- HTML全文浏览量: 20
- PDF下载量: 15
- 被引次数: 0
出版历程
- 收稿日期: 2023-05-08
- 修回日期: 2024-07-16
- 网络出版日期: 2024-10-15
- 刊出日期: 2024-08-20

Assessment of Extended-Range Prediction Capability Based on CMA-GEPS

QI Qianqian^{1, 2, 3},
ZHU Yuejian⁴,
CHEN Jing^{1, 2, 3
, ,},
LI Xiaoli^{1, 2, 3},
TIAN Hua^{1, 2, 3},
WANG Ye⁵

1.
CMA Earth System Modeling and Prediction Center (CEMC), Beijing 100081, China
2.
State Key Laboratory of Severe Weather (LaSW), Beijing 100081, China
3.
Key Laboratory of Earth System Modeling and Prediction of China Meteorological Administration, Beijing 100081, China
4.
National Oceanic and Atmospheric Administration (NOAA)/National Weather Service (NWS)/ National Centers for Environmental Prediction (NCEP)/Environmental Modeling Center (EMC), College Park, Maryland 20742, USA
5.
School of Mathematics and Statistics, Henan University, Kaifeng, Henan 475004, China

摘要

摘要: 基于CMA-GEPS系统开展1~35天的延伸期集合预报，并对该系统的延伸期尺度天气进行预报能力评估。结果表明：关于500 hPa位势高度，以距平相关系数（ACC）为表征的集合预报有效天数在北半球和南半球分别为9天和8.7天，且在北半球呈现季节循环特征，即冬（夏）季值高（低），为大气内在性质的表现；定量分析离散度-均方根误差关系表明，集合预报系统比确定性预报在延伸期尺度上可预报性更高，且北半球及南半球的潜在可预报天数分别为18天和16天。关于2 m温度，CMA-GEPS在延伸期尺度上可较好地描述温度场的空间分布特征，其较大的系统偏差主要位于热力强迫显著的高原或沙漠地区。关于MJO，CMA-GEPS对MJO的有效预报技巧达到15天，优于一般的大气模式，说明CMA-GEPS有潜力进一步发展延伸期天气预报。进一步诊断分析表明：CMA-GEPS对MJO预报的强度偏弱，这与CMA-GEPS描述的热带对流系统偏弱有关；传播速度前8天略偏快，8天之后偏慢；CMA-GEPS可较好地预报出MJO东传及北传运动；比较发现，CMA-GEPS对环流信号传播特征的预报优于对流信号，且描述的MJO东传优于北传特征。
- CMA-GEPS /
- 延伸期天气 /
- 集合预报 /
- MJO /
- 预报能力评估
Abstract: This study assessed the extended-range prediction capability of the CMA-Global Ensemble Prediction System (CMA-GEPS). Using this system, we conducted 35-day ensemble prediction experiments and evaluated the forecast capability. Results showed that, at the geopotential height of 500hPa, the ACC skills of the extended-range ensemble prediction system were 9 days and 8.7 days in the northern and southern hemispheres, respectively. The ACC skills in the northern hemisphere exhibited a seasonal cycle with higher prediction skills in winter and lower skills in summer, reflecting the inherent atmospheric properties. Quantitative analysis of the spread-root mean square error (RMSE) skill relations revealed that the ensemble forecasting system was more predictive than deterministic prediction on the extended scale. The potential forecast days in the northern and southern hemispheres were 18 and 16 days, respectively. In terms of 2m temperature, CMA-GEPS effectively captured the spatial distribution characteristics of the temperature field on the extended period scale. Further analysis showed that prediction errors for 2m temperature mainly occurred in desert or plateau areas with significant thermal forcing effects. Additionally, CMA-GEPS demonstrated skillful forecasting of the Madden-Julian Oscillation (MJO) with a lead time of 15 days, outperforming other general circulation models. This suggested that CMA-GEPS has the potential to develop extended-range ensemble predictions further. Moreover, analyses revealed that the predicted intensity of MJO was weaker than the analyses, possibly due to a weak tropical convective system. The predicted propagation speed for the first 1-8 days was slightly faster, while the predicted propagation speed of the 9-35 days was slower compared to the analyses. CMA-GEPS effectively predicted the eastward and northward signals of MJO. Furthermore, the prediction of circulation signal propagation characteristics was better than that of convective signal, and MJO eastward propagation characteristics were better predicted than northward propagation.
- CMA-GEPS /
- extended-range weather /
- ensemble prediction /
- MJO /
- assessment of prediction capability

HTML全文

图 1 北半球(a)与南半球（b）500 hPa位势高度的集合平均预报及控制预报的ACC技巧评分

下载: 全尺寸图片幻灯片

图 2 CMA-GEPS对北半球500 hPa位势高度集合平均预报及控制预报的ACC技巧

超前预报1周(a)、超前预报2周（b）、超前预报3&4周（c）、超前预报1~4周（d）；其中，蓝色和红色实线分别代表控制预报和集合平均预报的ACC技巧，蓝色和红色虚线分别代表控制预报和集合平均预报ACC技巧的平均值。

下载: 全尺寸图片幻灯片

图 3 基于CMA-GEPS系统的1年平均（2021年1月—2021年12月）的北半球（a）和南半球（b）500 hPa位势高度集合平均预报（红色实线）、控制预报（黑色实线）及集合离散度（蓝色实线）的35天预报

分析场和ERA5气候平均值的均方根误差以黑色虚线表示。

下载: 全尺寸图片幻灯片

图 4 CMA-GEPS系统集合平均预报2 m温度的分析场误差（a），参考态为ERA5资料；超前1周（a₁）、2周（a₂）、3&4周（a₃）及28天（a₄）的预报误差，参考态为CMA-GEPS自身分析场；超前1周（b₁）、2周（b₂）、3&4周（b₃）及28天（b₄）的预报误差，参考态为ERA5资料

下载: 全尺寸图片幻灯片

图 5 三组试验（集合预报CMA-GEPS V1.2, 控制预报CMA-GFS V3.1, CMA-GFS V2.4）2 m温度平均均方根误差（RMSE）对比

下载: 全尺寸图片幻灯片

图 6 基于CMA-GEPS分析场计算及基于观测计算的MJO振幅时间序列

下载: 全尺寸图片幻灯片

图 7 赤道（15 °S~15 °N）CMA-GEPS集合平均预报及控制预报的MJO要素（850 hPa纬向风、200 hPa纬向风和OLR）距平场的相关系数

水平虚线为相关系数0.5。

下载: 全尺寸图片幻灯片

图 8 2021年1—12月平均的MJO预报技巧

水平虚线为相关系数0.5。

下载: 全尺寸图片幻灯片

图 9 CMA-GEPS集合平均预报及观测的MJO振幅（虚线，左y坐标轴）和振幅偏差

（实线，右y坐标轴）（a），相位角度偏差（b）

下载: 全尺寸图片幻灯片

图 10 CMA-GEPS集合平均OLR延伸期天气预报的相对误差（单位：W·m^-2）

a₁.超前预报1周的相对误差；a₂.超前预报2周的相对误差；a₃.超前预报3&4周的相对误差；a₄.超前预报1~4周平均的相对误差；

下载: 全尺寸图片幻灯片

图 11 15 °S~15 °N平均的CMA-GEPS分析场（a）、7天预报(b)、14天预报(c)和21天预报(d)的时滞-经度传播图

阴影为OLR异常相关，等值线为850 hPa纬向风异常相关。

下载: 全尺寸图片幻灯片

图 12 30~150 °E平均的CMA-GEPS分析场(a)、7天预报(b)、14天预报(c)和21天预报(d)的时滞-纬度传播图

阴影为OLR异常相关，等值线为850 hPa纬向风异常相关。

下载: 全尺寸图片幻灯片

表 1 CMA-GEPS不同预报时效下，MJO东传及北传的预报场与分析场的相似系数

传播方向	7天预报	14天预报	21天预报
东传	R_U₈₅₀ = 0.95	R_U₈₅₀ = 0.79	R_U₈₅₀ = 0.54
东传	R_OLR = 0.80	R_OLR = 0.61	R_OLR = 0.58
北传	R_U₈₅₀ = 0.93	R_U₈₅₀ = 0.76	R_U₈₅₀ = 0.53
北传	R_OLR = 0.78	R_OLR = 0.59	R_OLR = 0.54

下载: 导出CSV

参考文献(43)

[1]	HOSKINS B. The potential for skill across the range of the seamless weather-climate prediction problem: A stimulus for our science[J]. Quart J Roy Meteor Soc, 2013, 139(672): 573-584.
[2]	MA J, JIN R H, ZONG Z P, et al. Analysis of predictability in extended range forecast[J]. Desert and Oasis Meteorology, 2019, 13(3): 137-143.
[3]	金荣花, 马杰, 任宏昌, 等. 我国10~30天延伸期预报技术进展与发展对策[J]. 地球科学进展, 2019, 34(8): 814-825.
[4]	章大全, 郑志海, 陈丽娟, 等. 10~30 d延伸期可预报性与预报方法研究进展[J]. 应用气象学报, 2019, 30(4): 416-430.
[5]	BAUER P, THOROE A, BRUNET G. The quiet revolution of numerical prediction[J]. Nature, 2015, 525(7 567): 47-55.
[6]	段晚锁, 汪叶, 霍振华, 等. 数值天气预报和气候预测的集合预报方法: 思考与展望[J]. 气候与环境研究, 2019, 24(3): 396-406.
[7]	KANG I S, JANG P H, ALMAZROUI M. Examination of multi-perturbation methods for ensemble prediction of the MJO during boreal summer[J]. Climate Dyn, 2014, 42(9-10): 2 627-2 637.
[8]	梁萍, 丁一汇. 基于季节内振荡的延伸预报试验[J]. 大气科学, 2012, 36(1): 102-116.
[9]	MADDEN R A, JULIAN P R. Detection of a 40-50 day oscillation in the zonal wind in the tropical Pacific[J]. J Atmos Sci, 1971, 28(5): 702-708.
[10]	MADDEN R A, JULIAN P R. Description of global－scale circulation cells in the tropics with a 40-50 day period[J]. J Atmos Sci, 1972, 29 (6): 1 109-1 123.
[11]	JONES C, WALISER D E, LAU K M, et al. The Madden-Julian Oscillation and its impact on Northern Hemisphere weather predictability [J]. Mon Wea Rev, 2004, 132(6): 1 462-1 471.
[12]	任宏利, 沈雨旸. MJO对我国天气气候影响的新事实[J]. 气象科技进展, 2016, 6(3): 97-105.
[13]	李崇银, 凌健, 袁媛, 等. 大气MJO研究的几个前沿问题[J]. 热带气象学报, 2016, 32(6): 797-816.
[14]	VITART F. Madden-Julian Oscillation prediction and teleconnections in the S2S database[J]. Quart J Roy Meteor Soc, 2017, 143(706): 2 210-2 220.
[15]	WANG W, HUNG M P, WEAVER S J, et al. MJO prediction in the NCEP climate forecast system version 2[J]. Climate Dyn, 2014, 42(9-10): 2 509-2 520.
[16]	KIM H M, WEBSTER PJ, TOMA V E, et al. Predictability and prediction skill of the MJO in two operational forecasting systems[J]. J Climate, 2014, 27(14): 5 364-5 378.
[17]	RASHID H A, HENDON H H, WHEELER MC, et al. Prediction of the Madden-Julian oscillation with the POAMA dynamical prediction system[J]. Climate Dyn, 2011, 36: 649-661.
[18]	LIU X W, WU T W, YANG S, et al. MJO prediction using the sub-seasonal to seasonal forecast model of Beijing Climate Center[J]. Climate Dyn, 2017, 48: 3 283-3 307.
[19]	LIU X W, LI W J, WU T W, et al. Validity of parameter optimization in improving MJO simulation and prediction using the sub-seasonal to seasonal forecast model of Beijing Climate Center[J]. Climate Dyn, 2019, 52: 3 823-3 843.
[20]	VITART F. Madden-Julian Oscillation prediction and teleconnections in the S2S database[J]. Quart J Roy Meteor Soc, 2017, 143(706): 2 210-2 220.
[21]	ZHU Y, ZHOU X, PENA M, et al. Impact of sea surface temperature forcing on weeks 3 and 4 forecast skill in the NCEP global forecasting system[J]. Wea Forecasting, 2017, 32(6): 2 159-2 174.
[22]	ZHU Y J, ZHOU X, LI W, et al. Towards the improvement of sub-seasonal prediction in the NCEP Global Ensemble Forecast System (GEFS)[J]. J Geophys Res, 2018, 123(13): 6 732-6 745.
[23]	ZHOU X Q, ZHU Y J, HOU D, et al. The development of the NCEP Global Ensemble Forecast System Version 12[J]. Wea Forecasting, 2022, 37(6): 1 069-1 084.
[24]	GUAN H, ZHU Y J, SINSKY E, et al. GEFSv12 reforecast dataset for supporting sub-seasonal and hydrometeorological applications[J]. Mon Wea Rev, 2022, 150(3): 647-665.
[25]	朱跃建. 基于最新全球集合预报系统的可预报性评估[J]. 大气科学学报, 2020, 43(1): 193-200.
[26]	陈静, 李晓莉. GRAPES全球/区域集合预报系统10年发展回顾及展望[J]. 气象科技进展, 2020, 10(2): 9-18.
[27]	李晓莉, 陈静, 刘永柱, 等. GRAPES全球集合预报初始条件及模式物理过程不确定性方法研究[J]. 大气科学学报, 2019, 42(3): 348-359.
[28]	龚建东, 张林, 王金成. 背景误差水平相关结构对四维变分资料同化的影响研究[J]. 气象学报, 2020, 78(6): 988-1001.
[29]	FENG J, WANG J C, DAI G K, et al. Spatiotemporal estimation of analysis errors in the operational global data assimilation system at the China Meteorological Administration using a modified SAFE method[J]. Quart J Roy Meteor Soc, 2023, 149(75): 2 049-2 653.
[30]	祁春娟, 潘留杰. CMA-GFS模式对东北半球环流形势预报能力的检验评估[J]. 高原气象, 2023, 42(2): 483-494.
[31]	齐倩倩, 朱跃建, 陈静, 等. 基于GRAPES-GFS次季节预报的误差诊断和预报能力分析[J]. 大气科学, 2022, 46(2): 327-345.
[32]	刘永柱, 龚建东, 张林, 等. 线性化物理过程对GRAPES 4DVar同化的影响[J]. 气象学报, 2019, 77(2): 196-209.
[33]	瞿安祥, 麻素红, 张进, 等. CMA-GFS全球预报系统中的台风初始化[J]. 气象学报, 2022, 80(2): 269-279.
[34]	WHEELER M C, HENDON H H. An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction [J]. Mon Wea Rev, 2004, 132(8): 1 917-1 932.
[35]	GOTTSCHALCK J, WHEELER M, WEICKMANN K, et al. A frame work for assessing operational model MJO forecasts: a project of CLIVAR Madden-Julian oscillation working group[J]. Bull Amer Meteor Soc, 2010, 91(9): 1 247-1 258.
[36]	LIN H, BRUNET G, DEROME J. Forecast skill of the Madden-Julian oscillation in two Canadian atmospheric models[J]. Mon Wea Rev, 2008, 136(11): 4 130-4 149.
[37]	HENDON H H, LIEBMANN B, NEWMAN M E, et al. Medium range forecasts errors associated with active episodes of the MaddenJulian oscillation[J]. Mon Wea Rev, 2000, 128(1): 69-85.
[38]	JONES C. Occurrence of extreme precipitation events in California and relationships with the Madden-Julian oscillation[J]. J Climate, 2000, 13(20): 3 576-3 587.
[39]	SPERBER K R. Madden-Julian variability in NCAR CAM 2.0 and CCSM2. 0[J]. Climate Dyn., 2004, 23: 259-278.
[40]	KIM H, JANIGA M A, PEGION K. MJO Propagation Processes and Mean Biases in the Subx and S2S Reforecasts[J]. J Geo Res, 2019, 124 (16): 9 314-9 331.
[41]	汪栩加, 郑志海, 封国林, 等. 延伸期预报中大气初值与海温边值的相对作用[J]. 气象学报, 2017, 75(1): 111-122.
[42]	PEGION K, KIRTMAN B P. The impact of air-sea interactions on the predictability of the tropical intra-seasonal oscillation[J]. J Climate, 2008, 21(22): 5 870-5 886.
[43]	WOOLNOUGH S J, VITART F, BALMASEDA M A. The role of the ocean in the Madden-Julian oscillation: Implications for MJO prediction[J]. Quart J Roy Meteor Soc, 2007, 133(622): 117-128.