Journal of Tropical Oceanography >
Deficiency of CMIP5 models in simulating changes of Pacific Walker circulation in recent three decades: the role of Sea Surface Temperature *
Copy editor: YIN Bo
Received date: 2019-01-01
Request revised date: 2019-02-27
Online published: 2019-10-09
Supported by
National Key Research and Development Plan Project(2018YFC1506903)
National Natural Science Key Fund Project(41830538)
Independent project of State Key Laboratory of Tropical Oceanography(South China Sea Institute of Oceanology, Chinese Academy of Sciences)(LTOZZ1802)
Copyright
The Pacific Walker circulation (PWC) is the most important atmospheric system over the tropical Pacific Ocean, and the cause of the long-term change of the PWC in response to global warming still remains debatable. The observations consistently indicate that the PWC has significantly strengthened in the past three decades. We examine the changes of the PWC in 18 climate models participated in phase 5 of the Coupled Model Intercomparison Project (CMIP5). Most CMIP5 models have produced successful simulations for the climatological spatial distribution features of the PWC, but no one can simulate the trend of significant enhancement of the PWC as observed. The deficiency of the models to simulate the trend of the PWC depends mainly on the capability of the models to simulate changes in sea surface temperature (SST). The trend pattern of SST is similar to that of the observation (i.e., La Niña-like) in the coupled models that simulate a strengthening PWC, but there are still some differences between the two. However, the distribution of SST shows an El Niño-like trend pattern in the coupled models that simulate a weakening PWC, which does not match that of the observation. For the latter models, if the observed SST is used to drive its corresponding atmospheric models, it can simulate the enhancement of the PWC, which fully demonstrates the leading role of SST change in the long-term change of the PWC. Therefore, to reasonably predict the change of the PWC in the context of global warming, the CMIP5 models need to improve their simulation capability for SST changes in the tropical Pacific.
Key words: CMIP5; Walker circulation; AMIP simulation; SST trend pattern
LIN Shuheng , GUAN Yuping , ZHANG Banglin . Deficiency of CMIP5 models in simulating changes of Pacific Walker circulation in recent three decades: the role of Sea Surface Temperature *[J]. Journal of Tropical Oceanography, 2019 , 38(5) : 52 -67 . DOI: 10.11978/2019002
表1 本文所使用的再分析资料及其详细信息Table 1 Reanalysis datasets used in this study and their detailed information |
数据 | 来源 | 同化方法 | 参考文献 | 时间跨度 | 空间分辨率 | |
---|---|---|---|---|---|---|
水平 | 垂直 | |||||
20CR | NOAA-CIRES | 3D-Var | Compo等(2011) | 1871—现在 | 2°×2° | 24层 |
ERA-Interim | ECMWF | 4D-Var | Dee等(2011) | 1979—现在 | 1.5°×1.5° | 37层 |
JRA-55 | JMA | 4D-Var | Ebita等(2011) | 1958—现在 | 1.25°×1.25° | 37层 |
MERRA | NASA | 3D-Var | Rienecker等(2011) | 1979—现在 | 2/3°×1/2° | 42层 |
NCEP-1 | NCEP-NCAR | 3D-Var | Kalnay等(1996) | 1948—现在 | 2.5°×2.5° | 17层 |
NCEP-2 | NCEP-DOE | 3D-Var | Kanamitsu等(2002) | 1979—现在 | 2.5°×2.5° | 17层 |
表2 本文中所使用的18个CMIP5耦合模式(水平分辨率用纬度×经度格点的数目表示)Tab. 2 Details of 18 CMIP5 models used in this study (Resolutions are given in terms of latitude×longitude grid points) |
序号 | 模式 | 分辨率/(纬度×经度) | 所属机构 |
---|---|---|---|
1 | ACCESS1.0 | 144×192 | 联邦科学与工业研究组织——气象局(澳大利亚) |
2 | ACCESS1.3 | 144×192 | 联邦科学与工业研究组织——气象局(澳大利亚) |
3 | BCC-CSM1.1 | 64×128 | 北京气候中心——中国气象局(中国) |
4 | BCC-CSM1.1(m) | 160×320 | 北京气候中心——中国气象局(中国) |
5 | CCSM4 | 192×288 | 美国国家大气研究中心(美国) |
6 | CMCC-CM | 240×480 | 意大利欧洲-梅迪特拉内奥气候研究中心(意大利) |
7 | CNRM-CM5 | 128×156 | 法国气象局气候变化研究中心(法国) |
8 | CSIRO Mk3.6.0 | 96×192 | 澳大利亚联邦科学和工业研究, 昆士兰气候变化卓越中心(澳大利亚) |
9 | GFDL CM3 | 90×144 | 美国国家海洋和大气管理局地球流体实验室(美国) |
10 | INM-CM4.0 | 120×180 | 俄罗斯数值数学研究所(俄罗斯) |
11 | IPSL-CM5A-LR | 96×96 | 皮埃尔-西蒙-拉普拉斯研究所(法国) |
12 | IPSL-CM5A-MR | 143×144 | 皮埃尔-西蒙-拉普拉斯研究所(法国) |
13 | IPSL-CM5B-LR | 96×96 | 皮埃尔-西蒙-拉普拉斯研究所(法国) |
14 | MIROC5 | 128×256 | 日本气候系统研究中心, 国家环境研究所和前沿气候变化研究中心(日本) |
15 | MPI-ESM-LR | 96×192 | 马克斯普朗克气象研究所(德国) |
16 | MPI-ESM-MR | 96×192 | 马克斯普朗克气象研究所(德国) |
17 | MRI-CGCM3 | 160×320 | 日本气象研究所(日本) |
18 | NorESM1-M | 96×144 | 挪威气候中心(挪威) |
图1 PWC强度在1979—2008年的线性趋势a. AMIP试验与再分析资料; b. CMIP5耦合模式与再分析资料。横坐标为各再分析资料(对应 Fig. 1 Linear trends of PWC intensity over the period from 1979 to 2008. (a) AMIP stimulations and six reanalysis datasets; (b) CMIP5 and six reanalysis datasets. Error bars denote standard errors of linear trends. The labels of x-axis are the names of reanalysis data (corresponding to |
图2 赤道太平洋区域(5ºS—5ºN)平均的纬向质量流函数的气候态(1979—2008年) (a~c, 等值线和填色, 单位: ×109kg·s-1)和热带太平洋海表面温度的气候态(1979—2008年) (d~f, 等值线及填色, 单位: ℃)a. 6个再分析集合平均; b. AMIP试验集合平均; c. CMIP5耦合模式集合平均; d. ERSST; e. HadISST; f. CMIP5耦合模式集合平均 Fig. 2 Long-term mean (1979-2008) of zonal mass stream function ψ along the equatorial Pacific (5ºS—5ºN) (a~c, shading and contour; units: ×109 kg·s-1) and long-term mean (1979-2008) of SST in the tropical Pacific Ocean (d~f, shading and contour; units: ºC). (a) Ensemble mean of six reanalysis datasets; (b) ensemble mean of AMIP simulations; (c) ensemble mean of CMIP5 simulations; (d) ERSST; (e) HadISST; (f) ensemble mean CMIP5 simulations |
图3 18个CMIP5耦合模式模拟的热带太平洋沃克环流(a)和海表面温度(b)的气候态相对于观测的泰勒诊断图a中参考点REF为6个再分析资料的集合平均; b中为NOAA的扩展重建海表温度数据(ERSST.v4b)与英国气象局哈德莱中心海面温度数据集(HadISST)的平均。极坐标表示模拟场到原点的距离, 代表模拟场相对于参考点的标准差; 角坐标表示模拟场的方位角的余弦, 代表模拟场与观测场的相关系数。角坐标和极坐标值越接近于1 (即点REF), 表示模拟场的空间分布越接近于观测 Fig. 3 Taylor diagrams of the climatology of tropical Pacific Walker circulation and SST in 18 CMIP5 models. (a) Tropical Pacific Walker circulation; (b) SST. The REF in (a) is ensemble mean of the six reanalysis datasets and that in (b) is mean of ERSST and HadISST datasets. The radial coordinate is the standard deviation of model results divided by standard deviation of the observations. The angular coordinate is weighted pattern correlation coefficient between model results and observations. The closer the angular and polar values are to 1 (i.e., REF), the closer the spatial distribution of the simulated field is to the observation |
图4 1979—2008年纬向质量流函数的气候态(等值线, 单位: ×109kg·s-1)和该时段的线性趋势(填色, 单位: ×109kg·s-1·decade-1)a. 6个再分析资料集合平均(ENS); b. 18个AMIP试验集合平均; c. CMIP5 (+PWCI)模式; d. CMIP5 (-PWCI)模式; e. AMIP (+CMIP5)试验; f. AMIP (-CMIP5)试验。黑色打点表示趋势通过95%的置信检验。位于坐标底下的黑色加粗线段分别表示海洋性大陆以及南美洲大陆的地形范围 Fig. 4 Climatology (contour, units: ×109 kg·s-1) and linear trend (shading, units: ×109 kg·s-1·decade-1) of zonal mass stream function ψ over 1979-2008. (a) Ensemble mean of six reanalysis datasets (ENS); (b) ensemble mean of 18 AMIP stimulations; (c) CMIP5 (+PWCI) models; (d) CMIP5 (-PWCI) models; (e) AMIP (+CMIP5) stimulations; (f) AMIP (-CMIP5) stimulations. Stippling indicates the trend is statistically significant at the level 95%. Black thick lines at the bottom indicate the Maritime and South American Continent |
图5 海表面气压(填色, 单位: hPa·decade-1)、1000hPa等压面上的风(箭头, 单位m·s-1·decade-1)和降水速率(等值线, 单位: mm·d-1·decade-1)在1979—2008年的线性趋势a. 观测; b. 18个AMIP试验集合平均; c. CMIP5 (+PWCI)模式; d. CMIP5 (-PWCI)模式; e. AMIP (+CMIP5)试验; f. AMIP (-CMIP5)试验。a中观测的风来自6个再分析资料的集合平均, SLP为HadSLP2数据, 降水资料来自GPCP数据。图中绿色实线表示降水速率的趋势为正值, 紫色虚线表示降水速率的趋势为负值, 黑色打点表示SLP趋势通过95%的置信检验。图中只给出纬向风趋势通过95%置信检验的风场 Fig. 5 Linear trends of SLP (shading, units: hPa·decade-1), wind at 1000 hPa (vector, units: m·s-1·decade-1) and precipitation (contour, units: mm·d-1·decade-1) during 1979-2008. (a) observation; (b) ensemble mean of 18 AMIP stimulations; (c) CMIP5 (+PWCI) models; (d) CMIP5 (-PWCI) models; (e) AMIP (+CMIP5) stimulations; (f) AMIP (-CMIP5) stimulations. Wind and SLP data are from the ENS and precipitation is from the GPCP in (a). Green solid contour denotes the trend of precipitation is positive, and purple denotes the negative trend. Vectors are plotted only for regions with surface zonal wind trends that are statistically significant at the 95% level. Stippling indicates the trend of SLP is statistically significant at the 95% level |
图6 CMIP5耦合模式模拟的赤道太平洋纬向SST梯度[$\Delta \text{SST}$, (80º—160ºW, 5ºS—5ºN)平均SST减去(80º—160ºE, 5ºS—5ºN)平均SST]与PWC的强度在1979—2008年的线性趋势的散点图黑色直线表示18个模式之间的最小二乘的拟合直线, 黑点表示CMIP5耦合模式的集合平均结果, 红点表示再分析资料得到的PWCI的趋势与观测得到的SST趋势 Fig. 6 Scatterplot of linear trends of SST gradient [$\Delta \text{SST}$, regional mean over (80º—160ºW, 5ºS—5ºN) minus that over (80º—160ºE, 5ºS—5ºN)] and PWC intensity during 1979-2008. The black line denotes the least-squares linear fit of the trend based on 18 CMIP5 models. Black dot denotes the trend of PWC intensity and $\Delta \text{SST}$ in ensemble mean of CMIP5 models, and red dot denotes the trend in the observations |
图7 海表面温度在1979—2008年的线性趋势a. ERSST; b. HadISST; c. CMIP5 (+PWCI)模式; d. CMIP5 (-PWCI)模式。黑色打点表示趋势通过95%的置信检验 Fig. 7 Linear trend of annual-mean SST (shading, units: K·decade-1) during 1979-2008. (a) ERSST; (b) HadISST; (c) CMIP5 (+PWCI) models; (d) CMIP5 (-PWCI) models. Stippling indicates the trend is statistically significant at the 95% level |
图8 {Invalid MML}海表温度梯度(a、b)和沃克环流强度(c、d)的异常值年变化的时间序列a. CMIP5 (+PWCI)模式与观测; b. CMIP5 (-PWCI)模式与观测; c. CMIP5 (+PWCI)模式、AMIP (+CMIP5)与观测; d. CMIP5 (-PWCI)模式、AMIP (-CMIP5)与观测。CMIP5 (+PWCI)模式包括GFDL-CM3、IPSL-CM5A-MR、INMCM4、CSIRO-Mk3-6-0模式, CMIP5 (-PWCI)模式包括MPI-ESM-MR和MIROC5。模式具体细节见 Fig. 8 Time series of the annual anomalies of SST gradient and PWCI. (a) CMIP5 (+PWCI) models and observation; (b) CMIP5 (-PWCI) models and observation; (c) CMIP5 (+PWCI) models, AMIP (+CMIP5) stimulations and observations; (d) CMIP5 (-PWCI) models, AMIP (-CMIP5) stimulations and observation. The CMIP5 (+PWCI) models include GFDL-CM3, IPSL-CM5A-MR, INMCM4, CSIRO-Mk3-6-0 mode, and CMIP5 (-PWCI) model include MPI-ESM-MR and MIROC5. See |
图9 海表温度梯度(a、b)和沃克环流强度(c、d)的异常值9年滑动平均的年变化时间序列a. CMIP5 (+PWCI)模式与观测; b. CMIP5 (-PWCI)模式与观测; c. CMIP5 (+PWCI)模式、AMIP (+CMIP5)与观测; d. CMIP5 (-PWCI)模式、AMIP(-CMIP5)与观测。CMIP5 (+PWCI)模式包括GFDL-CM3、IPSL-CM5A-MR、INMCM4、CSIRO-Mk3-6-0模式, CMIP5(-PWCI)模式包括MPI-ESM-MR和MIROC5。模式具体细节见 Fig. 9 Time series of the nine-year smooth anomalies of SST gradient and PWCI. (a) CMIP5 (+PWCI) models and observation; (b) CMIP5 (-PWCI) models and observation; (c) CMIP5 (+PWCI) models, AMIP (+CMIP5) stimulations and observations; (d) CMIP5 (-PWCI) models, AMIP (-CMIP5) stimulations and observation. The CMIP5 (+PWCI) models include GFDL-CM3, IPSL-CM5A-MR, INMCM4, CSIRO-Mk3-6-0 mode, and CMIP5 (-PWCI) model include MPI-ESM-MR and MIROC5. See |
感谢中国科学院南海海洋研究所曙光高性能计算集群系统的帮助。
[1] |
马双梅, 周天军 , 2014. 热带太平洋沃克环流变化的数值模拟[J]. 中国科学: 地球科学, 44(11):2576-2592.
|
[2] |
孙稚权, 项杰, 管玉平 , 2016. 太平洋沃克环流近几十年来的加强[J]. 热带海洋学报, 35(2):19-29.
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
[29] |
|
[30] |
|
[31] |
|
[32] |
|
[33] |
|
[34] |
|
[35] |
|
[36] |
|
[37] |
|
[38] |
|
[39] |
|
[40] |
|
[41] |
|
[42] |
|
[43] |
|
[44] |
|
[45] |
|
[46] |
|
[47] |
|
[48] |
|
[49] |
|
[50] |
|
[51] |
|
[52] |
|
[53] |
|
[54] |
|
[55] |
|
[56] |
|
[57] |
|
[58] |
|
[59] |
|
[60] |
|
[61] |
|
[62] |
|
[63] |
|
[64] |
|
[65] |
|
[66] |
|
/
〈 | 〉 |