地球科学进展 ›› 2017, Vol. 32 ›› Issue (4): 396 -408. doi: 10.11867/j. issn. 1001-8166.2017.04.0396

上一篇    下一篇

BCC_CSM1.1气候模式对全球海表温度年代际变化的回报能力评估
韩振宇 1( ), 吴波 2, 辛晓歌 1   
  1. 1.中国气象局 国家气候中心,北京 100081
    2.中国科学院大气物理研究所大气科学和地球流体力学数值模拟国家重点实验室,北京 100029
  • 收稿日期:2016-10-17 修回日期:2016-12-29 出版日期:2017-04-20
  • 基金资助:
    公益性行业(气象)科研专项项目“基于FGOALS-s、CMA和CESM气候系统模式的年代际集合预测系统的建立与研究”(编号:GYHY201506012);中国气象局LCS气候研究开放课题青年基金2016年度资助项目“BCC_CSM1.1模式对中国近期气候的预测能力研究”资助

Decadal Prediction Skill of the Global Sea Surface Temperature in the BCC_CSM1.1 Climate Model

Zhenyu Han 1( ), Bo Wu 2, Xiaoge Xin 1   

  1. 1. National Climate Center, China Meteorological Administration, Beijing 100081, China
    2. State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics,Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China
  • Received:2016-10-17 Revised:2016-12-29 Online:2017-04-20 Published:2017-04-20
  • About author:

    First author:Han Zhenyu(1987-),male,Yuanping City,Shanxi Province,Associate professor. Research areas include climate modelling, East Asian climate change and climate variability.E-mail:hanzy@cma.gov.cn

  • Supported by:
    Foundation item:Project supported by the R&D Special Fund for Public Welfare Industry (Meteorology) “Development and research of ensemble decadal climate prediction system based on global climate models FGOALS-s, CAMS and CESM”(No.GYHY201506012);The LCS Open Funds for Young Scholars (2016) “Study on near-term climate prediction in China by BCC_CSM1.1 model”

通过与观测资料和20世纪历史气候模拟试验结果(NoINT)的对比分析,评估了BCC_CSM1.1年代际预测试验(INT)中对全球海表温度(SST)年代际变化的回报能力。分析结果显示:①INT试验模拟的全球平均SST增暖趋势比NoINT的更加接近观测;②其在热带大西洋、热带西太平洋和热带印度洋有较高的预测技巧;③对于太平洋年代际振荡2个关键区——北太平洋和热带中东太平洋,模式的预测技巧较低,且海洋初始化的作用也很小;④在热带南印度洋,INT的预测技巧普遍高于NoINT,在提前3~6年和4~7年时技巧最高。这些结论与基于其他模式得到的已有研究结果类似,但是BCC模式对北大西洋,特别是其副极地区域的预测技巧明显低于其他模式。BCC模式无法合理模拟出北大西洋SST与热盐环流间的交替变化规律,可能是其预测技巧偏低的原因。

This study assesses retrospective decadal prediction skill of Sea Surface Temperature (SST) variability in initialized climate prediction experiments (INT) with the Beijing Climate Center Climate System Model (BCC_CSM1.1). Ensemble forecasts were evaluated using observations, and compared to an ensemble of uninitialized simulations (NoINT). The results show as follows: ①The warming trend of global mean SST simulated by the INT runs is closer to the observation than that in the NoINT runs.②The INT runs show high SST prediction skills over broad regions of tropical Atlantic, western tropical Pacific and tropical Indian Oceans. ③ In the North Pacific and the east-central tropical Pacific Ocean, the prediction skills are very weak, and there are few improvements coming from the initialization in the INT runs. ④ In the southern Indian Ocean, the prediction skills of the INT runs are significantly larger than that of the NoINT runs, with the maximum skill at the 3~6 and 4~7 years lead time. The above-mentioned conclusions are similar to the results of other climate models. However, the prediction skill in the North Atlantic Ocean is much lower than that of other models, especially in the subpolar region. The low skills in the Atlantic Ocean may be attributed to the misrepresentation of the lead-lag relationship between the Atlantic meridional heat transport and the SST in the BCC_CSM1.1.

中图分类号: 

图1 全球平均SST的模拟误差
(a)INT(红色)和NoINT(蓝色)试验中各起报时间合成序列的平均误差,以及INT试验的误差统计分布的盒须图;(a)中虚线表示单样本试验结果,实线表示样本集合结果;(b)INT试验中各起报时间的合成序列(如lead 1a表示提前1年起报)
Fig.1 Biases on global averaged SST
(a) Mean biases and boxplots along the forecast lead time from the INT (red) and NoINT (blue) runs, the dashed (solid) lines denote ensemble members (means); (b) Series of the biases from the INT runs over different forecast lead time (“lead 1 a” means that the lead time is one year)
图2 SST年代际时间尺度变率与总变率的比值
比值计算采用1870—2005年的数据,NoINT的比值是3组样本分别计算比值后的平均,而不是基于样本集合后的SST;方框所示区域分别是副极地大西洋、热带大西洋、北太平洋、热带中太平洋、热带南印度洋
Fig.2 Ratio of the decadal time-scale variance to the total variance in SST
The calculation is based on the time series of 1870-2005; The ratio in the NoINT SST is the average of three variance ratios,not the variance ratio of the average of the ensemble members. Boxes correspond to the subpolar Atlantic, the tropical Atlantic, the North Pacific, the central tropical Pacific and the southern tropical Indian Ocean
图3 INT试验对历史SST回报技巧的空间分布
(a)INT回报与观测相关系数;(b)INT回报的RMSSS(乘以100)评分;(c)相对于观测,INT试验的RMSE与NoINT的RMSE的比值。回报技巧全部基于提前6~9年的回报序列计算,相关系数0.32表示通过95%显著性检验
Fig.3 Spatial patterns of the hindcast skills of the SST predicted by the INT runs
(a) Correlation skills and (b) RMSSS (multiplied by 100) of the ensemble mean of the INT runs for predictions averaged over the hindcast years 6~9;(c) Ratio of RMSE between the ensemble mean of the INT runs and that of the NoINT runs for predictions averaged over the hindcast years 6~9. Correlation coefficient of 0.32 represents passing the 95% significance level
图4 INT试验对全球平均SST的回报技巧
(a)提前2~5年回报的全球平均SST距平随时间的演变(℃);(b)同(a),但基于提前6~9年回报结果;(c)INT和NoINT与观测的时间相关系数随预测时效的变化;(d)相对于观测,INT和NoINT的RMSE随预测时效的变化;红线为INT试验结果,蓝线为NoINT试验结果,黑线为观测,图中虚线为单样本试验结果,实线为样本集合结果。(a),(b)中右下角数字表示各序列的线性趋势(℃/decade),箭头指示4次火山喷发的年份
Fig.4 Hindcast skills of the global averaged SST predicted by the INT runs
(a) Time series of the global averaged SST anomalies predicted by the ensemble mean of the INT (NoINT) runs for predictions averaged over the hindcast years 2~5 and corresponding observational reference (℃); (b) As in (a) but for hindcast years 6~9;(c) Correlations between the model runs and the observational references along the forecast time for 4-year averages;(d) RMSE of the model runs along the forecast time for 4-year averages. The red (blue) lines denote INT (NoINT) runs. The black lines denote the observation. The dashed (solid) lines denote ensemble members (means). In (a),(b), the numbers at the bottom-right corner denote the linear trend of the time series (℃/decade), and arrows indicate the volcanic eruption years
图5 INT试验对副极地大西洋和热带大西洋平均SST的回报技巧
(a)提前2~5年回报的副极地大西洋地区平均SST距平随时间的演变(℃);(b)同(a),但基于提前6~9年回报结果;(c)INT和NoINT与观测的时间相关系数随预测时效的变化;(d)相对于观测,INT和NoINT的RMSE随预测时效的变化。红线为INT试验结果,蓝线为NoINT试验结果,黑线为观测,图中虚线为单样本试验结果,实线为样本集合结果;(e)~(h)同图(a)~(d),但是热带大西洋;(i),(j)同图(a),(b),但是AMO指数
Fig.5 Hindcast skills of the SST averaged over the subpolar Atlantic and the tropical Atlantic Ocean predicted by the INT runs
(a) Time series of the averaged SST anomalies over the subpolar Atlantic Ocean predicted by the ensemble mean of the INT (NoINT) runs for predictions averaged over the hindcast years 2~5 and corresponding observational reference (℃); (b) As in (a) but for hindcast years 6~9;(c) Correlations between the model runs and the observational references along the forecast time for 4-year averages;(d) RMSE of the model runs along the forecast time for 4-year averages. The red (blue) lines denote INT (NoINT) runs. The black lines denote the observation. The dashed (solid) lines denote ensemble members (means); (e)~(h) same as (a)~(d), but for the tropical Atlantic Ocean; (i),(j) same as (a), (b), but for the AMO index
图6 大西洋经向热输送与副极地大西洋SST的超前滞后相关
(a)SODA;(b)NoINT试验结果;(c)INT试验提前2~5年预测结果;(d)INT试验提前1年预测结果,相关系数0.32和0.42分别表示通过95%和99%显著性检验
Fig.6 Lead-lag correlation between the Atlantic MHT and the SST averaged over the subpolar Atlantic at each latitude
(a) SODA; (b) NoINT;(c) INT runs over the hindcast years 2~5;(d) INT runs over the hindcast year 1,correlation coefficient of 0.32 and 0.42 represent passing the 95% and 99% significance level respectively
图7 INT试验对北太平洋和热带中东太平洋平均SST的回报技巧
(a)提前2~5年回报的北太平洋地区平均SST距平随时间的演变(℃);(b)同(a),但基于提前6~9年回报结果;(c)INT和NoINT与观测的时间相关系数随预测时效的变化;(d)相对于观测,INT和NoINT的RMSE随预测时效的变化;红线为INT试验结果,蓝线为NoINT试验结果,黑线为观测,图中虚线为单样本试验结果,实线为样本集合结果;(e)~(h)同图(a)~(d),但是热带中东太平洋
Fig.7 Hindcast skills of the SST averaged over the North Pacific and the east-central tropical Pacific Ocean predicted by the INT runs
(a) Time series of the averaged SST anomalies over the North Pacific Ocean predicted by the ensemble mean of the INT (NoINT) runs for predictions averaged over the hindcast years 2~5 and corresponding observational reference (℃); (b) as in (a) but for hindcast years 6~9; (c) Correlations between the model runs and the observational references along the forecast time for 4-year averages; (d) RMSE of the model runs along the forecast time for 4-year averages; The red (blue) lines denote INT (NoINT) runs; The black lines denote the observation; The dashed (solid) lines denote ensemble members (means); (e)~(h) same as (a)~(d), but for the east-central tropical Pacific Ocean
图8 INT试验对热带南印度洋平均SST的回报技巧
(a)提前2~5年回报的热带南印度洋地区平均SST距平随时间的演变(℃);(b)同(a),但基于提前6~9年回报结果;(c)INT和NoINT与 观测的时间相关系数随预测时效的变化;(d)相对于观测,INT和NoINT的RMSE随预测时效的变化;红线为INT试验结果,蓝线为NoINT试验结果,黑线为观测,图中虚线为单样本试验结果,实线为样本集合结果
Fig.8 Hindcast skills of the SST averaged over the southern tropical Indian Ocean predicted by the INT runs
(a) Time series of the averaged SST anomalies over the southern tropical Indian Ocean predicted by the ensemble mean of the INT (NoINT) runs for predictions averaged over the hindcast years 2~5 and corresponding observational reference (℃); (b) as in (a) but for hindcast years 6~9;(c) Correlations between the model runs and the observational references along the forecast time for 4-year averages; (d) RMSE of the model runs along the forecast time for 4-year averages; The red (blue) lines denote INT (NoINT) runs; The black lines denote the observation; The dashed (solid) lines denote ensemble members (means)
[1] Xiong Wei, Feng Lingzhi, Ju Hui, et al.Possible impacts of high temperatures on China’s rice yield under climate change[J] . Advances in Earth Science, 2016, 31(5): 515-528.
[熊伟, 冯灵芝, 居辉, 等. 未来气候变化背景下高温热害对中国水稻产量的可能影响分析[J]. 地球科学进展, 2016, 31(5): 515-528.]
[2] Shi Peijun, Wang Aihui, Sun Fubao, et al.A study of global change population and economic system risk forming mechanism and assessment[J]. Advances in Earth Science, 2016, 31(8): 775-781.
[史培军, 王爱慧, 孙福宝, 等. 全球变化人口与经济系统风险形成机制及评估研究[J]. 地球科学进展, 2016, 31(8): 775-781.]
[3] Taylor K E, Stouffer R J, Meehl G A.An overview of CMIP5 and the experiment design[J]. Bulletin of the American Meteorological Society, 2012, 93(4): 485-498.
[4] Gao Feng, Xin Xiaoge, Wu Tongwen.A study of the prediction of regional and global temperature on decadal time scale with BCC_CSM1. 1 model[J]. Chinese Journal of Atmospheric Sciences, 2012, 36(6): 1 165-1 179.
[高峰, 辛晓歌, 吴统文. BCC_CSM1.1对10年尺度全球及区域温度的预测研究[J]. 大气科学, 2012, 36(6): 1 165-1 179.]
[5] Wang Bin, Liu Mimi, Yu Yongqiang, et al.Preliminary evaluations of FGOALS-g2 for decadal predictions[J]. Advances in Atmospheric Sciences, 2013, 30(3): 674-683.
[6] Wu Bo, Zhou Tianjun.Prediction of decadal variability of sea surface temperature by a coupled global climate model FGOALS_gl developed in LASG/IAP[J]. Chinese Science Bulletin, 2012, 57(19): 2 453-2 459.
[7] Wu Bo, Chen Xiaolong, Song Fengfei, et al.Initialized decadal predictions by LASG/IAP climate system model FGOALS-s2: Evaluations of strengths and weaknesses[J]. Advances in Meteorology, 2015, doi:10.1155/2015/904826.
[8] Kim H, Webster P J, Curry J A.Evaluation of short-term climate change prediction in multi-model CMIP5 decadal hindcasts[J]. Geophysical Research Letters, 2012, 39(10): L10701, doi:10.1029/2012GL051644.
[9] Doblas-Reyes F J, Andreu-Burillo I, Chikamoto Y, et al. Initialized near-term regional climate change prediction[J]. Nature communications, 2013, 4: 1 715, doi: 10.1038/ncomms2704.
[10] Guemas V, Corti S, García-Serrano J, et al.The Indian Ocean: The region of highest skill worldwide in decadal climate prediction[J]. Journal of Climate, 2013, 26(3): 726-739.
[11] Karspeck A, Yeager S, Danabasoglu G, et al.An evaluation of experimental decadal predictions using CCSM4[J]. Climate Dynamics, 2015, 44(3/4): 907-923.
[12] Mignot J, García-Serrano J, Swingedouw D, et al.Decadal prediction skill in the ocean with surface nudging in the IPSL-CM5A-LR climate model[J]. Climate Dynamics, 2016, 47(3):1 225-1 246.
[13] Shaffrey L C, Hodson D, Robson J, et al.Decadal predictions with the HiGEM high resolution global coupled climate model: Description and basic evaluation[J]. Climate Dynamics, 2016,doi: 10.1007/s00382-016-3075-x.
[14] Chen Weilin, Jiang Zhihong.Evaluation of the decadal prediction skill over China based on four global atmosphere-ocean coupled climate models[J]. Climatic and Environmental Research, 2012, 17(1): 81-91.
[陈威霖, 江志红. 全球海气耦合模式对中国区域年代际气候变化预测能力的评估[J]. 气候与环境研究, 2012, 17(1): 81-91.]
[15] Zhou Xin, Li Qingquan, Sun Xiubo, et al.Simulation and projection of temperature in China with BCC_CSM1. 1 model[J]. Journal of Applied Meterological Science, 2014, 25(1): 95-106.
[周鑫,李清泉,孙秀博,等. BCC_CSM1.1模式对我国气温的模拟和预估[J].应用气象学报, 2014, 25(1): 95-106.]
[16] Wei Linxiao, Xin Xiaoge, Cheng Bingyan, et al.Hindcast of China climate with decadal experiment by BCC_CSM1.1 climate model[J]. Advances in Climate Change Research, 2016, 12(4): 294-302.
[魏麟骁,辛晓歌,程炳岩,等. BCC_CSM1.1气候模式年代际试验对中国气候的回报能力评估[J]. 气候变化研究进展, 2016, 12(4): 294-302.]
[17] Luo Feifei, Li Shuanglin.Joint statistical-dynamical approach to decadal prediction of East Asian surface air temperature[J]. Science in China(Series D), 2014, 57(12): 3 062-3 072, doi: 10.1007/s11430-014-4984-3.
[罗菲菲, 李双林. 动力统计相结合的未来30年东亚气温年代际预测[J]. 中国科学: D辑, 2015, 45(4):402-413.]
[18] Xin Xiaoge, Wu Tongwen, Zhang Jie.Introduction of CMIP5 experiments carried out with the climate system models of Beijing Climate Center[J]. Advances in Climate Change Research, 2013, 4(1): 41-49.
[辛晓歌, 吴统文, 张洁. BCC气候系统模式开展的CMIP5试验介绍[J]. 气候变化研究进展, 2012, 8(5): 378-382.]
[19] Carton J A, Giese B S.A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA)[J]. Monthly Weather Review, 2008, 136(8): 2 999-3 017.
[20] Rayner N A, Parker D E, Horton E B, et al.Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century[J]. Journal of Geophysical Research: Atmospheres, 2003, 108(D14): 4 407, doi:10.1029/2002JD002670.
[21] lato G, Marotzke J, Abiodun B, et al. Evaluation of climate models[M]∥Stocker T F, et al, eds. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press, 2013: 741-866.
[22] Goddard L, Kumar A, Solomon A, et al.A verification framework for interannual-to-decadal predictions experiments[J]. Climate Dynamics, 2013, 40(1/2): 245-272.
[23] ICPO (International CLIVAR Project Office).Decadal and Bias Correction for Decadal Climate Predictions[R/OL]. International CLIVAR Project Office, 2011, CLIVAR Publication Series No.150.[2016-12-20]..
URL    
[24] Knight J R, Allan R J, Folland C K, et al.A signature of persistent natural thermohaline circulation cycles in observed climate[J]. Geophysical Research Letters, 2005, 32(20): L20708, doi:10.1029/2005GL024233.
[25] Msadek R, Frankignoul C.Atlantic multidecadal oceanic variability and its influence on the atmosphere in a climate model[J]. Climate Dynamics, 2009, 33(1): 45-62.
[26] Cheng Wei, Chiang J C H, Zhang Dongxiao. Atlantic Meridional Overturning Circulation (AMOC) in CMIP5 models: RCP and historical simulations[J]. Journal of Climate, 2013, 26(18): 7 187-7 197.
[27] Zhang Liping, Wang Chunzai.Multidecadal North Atlantic sea surface temperature and Atlantic meridional overturning circulation variability in CMIP5 historical simulations[J]. Journal of Geophysical Research: Oceans, 2013,118(10): 5 772-5 791.
[28] Zhang Rong.Latitudinal dependence of Atlantic Meridional Overturning Circulation (AMOC) variations[J]. Geophysical Research Letters, 2010, 37(16): L16703, doi:10.1029/2010GL044474.
[29] Zhou Tianjun, Chen Xiaolong, Dong Lu, et al.Chinese contribution to CMIP5: An overview of five Chinese models’ performances[J]. Journal of Meteorological Research, 2014, 28(4): 481-509.
[30] Han W, Vialard J, McPhaden M J, et al. Indian Ocean decadal variability: A review[J]. Bulletin of the American Meteorological Society, 2014, 95(11): 1 679-1 703.
[31] Dong Lu, Zhou Tianjun, Dai Aiguo, et al.The footprint of the inter-decadal pacific oscillation in Indian Ocean sea surface temperatures[J]. Scientific Reports, 2016, 6: 21 251, doi:10.1038/srep21251.
[1] 汪芋君, 任宏利, 王琳. 第三极地区气温和积雪的季节—年际气候预测研究[J]. 地球科学进展, 2021, 36(2): 198-210.
[2] 吴波, 周天军, 孙倩. 海洋模式初始化同化方案对IAP近期气候预测系统回报试验技巧的影响[J]. 地球科学进展, 2017, 32(4): 342-352.
[3] 满文敏, 周天军. IAP年代际预测试验中火山活动对太平洋海温预测技巧的影响[J]. 地球科学进展, 2017, 32(4): 353-361.
[4] 容新尧, 刘征宇, 段晚锁. 耦合模式中北太平洋和北大西洋海表面温度年代际可预报性和预报技巧的季节依赖性[J]. 地球科学进展, 2017, 32(4): 382-395.
[5] 张丽霞, 张文霞, 周天军, 吴波. ENSEMBLES耦合模式对全球陆地季风区夏季降水的年代际预测能力评估[J]. 地球科学进展, 2017, 32(4): 409-419.
[6] 马浩,李春. 南大洋淡水通量的气候效应研究进展[J]. 地球科学进展, 2010, 25(2): 140-146.
[7] 周天军,张学洪,刘海龙. 大洋环流模式的温盐表面边界条件处理及其影响研究评述[J]. 地球科学进展, 2009, 24(2): 111-122.
[8] 刘宇,管玉平,林一骅. 大洋热盐环流研究的一个焦点:北太平洋是否有深水形成[J]. 地球科学进展, 2006, 21(11): 1185-1192.
阅读次数
全文


摘要