Journal of Tropical Oceanography >
Tropical ocean-atmosphere coupling modes and their relationship with ENSO during spring*
Copy editor: YIN Bo
Received date: 2022-05-10
Revised date: 2022-07-14
Online published: 2022-08-02
Supported by
National Natural Science Foundation of China(41976024)
National Natural Science Foundation of China(41830538)
National Natural Science Foundation of China(42090042)
Southern Marine Science, Engineering Guangdong Laboratory (Guangzhou)(2019BT02H594)
Independent Research Project Program of State Key Laboratory of Tropical Oceanography(LTOZZ2101)
The tropical ocean-atmosphere system in spring may simultaneously respond to the El Niño-Southern Oscillation (ENSO) in the Pacific ocean. At the same time, it can affect the ENSO development through coupled regional ocean-atmosphere interactions. Based on the joint empirical orthogonal function and open-source datasets, we identify two major global climate modes. The first EOF mode presents the ENSO pattern along with the spring meridional mode in the Atlantic and asymmetric mode in the Indian Ocean, in which the sea surface temperature warms up and precipitation increases in the tropical central and eastern Pacific ocean, accompanied by the equator-asymmetric pattern of precipitation in the tropical Atlantic and Indian Oceans as well as anomalous sea surface temperature gradient in the trans-equatorial. Further analyses suggest that the ENSO influences the intertropical convergence zone by adjusting atmospheric circulation during its mature phase and then induces regional ocean-atmosphere feedback resulting in the spring meridional modes. The differences in spring asymmetric modes of precipitation in the tropical Atlantic and the Indian Ocean are determined by the different positions of the intertropical convergence zone in winter and spring. The second mode shows a meridional sea surface temperature and precipitation anomalies in the tropical Pacific, i.e., the Pacific meridional mode. The warm pole of the spring Pacific meridional mode extends over the equator, causing westerly wind anomalies that favor the El Niño development. This study reveals the relationship between the Pacific ENSO and the global spring meridional mode, contributing to a better understanding of the seasonal 'footprint' of tropical climate modes.
ZHANG Yuhong , ZHANG Lianyi , DU Yan . Tropical ocean-atmosphere coupling modes and their relationship with ENSO during spring*[J]. Journal of Tropical Oceanography, 2023 , 42(2) : 34 -44 . DOI: 10.11978/2022105
图1 热带海洋北半球春季(3—5月)平均降水率(填色, 单位: mm·月-1)和海表温度(等值线, 单位: ℃)联合经验正交函数的前两模态空间分布和时间序列a. 第一模态空间分布; b. 第二模态空间分布; c. 第一和第二模态对应的时间序列。图c中短虚线标注了1倍标准差的值, 点线为0值; 灰色柱状图为ENSO盛期11月至次年1月平均的Niño3.4指数 Fig. 1 The first two modes of the Joint empirical orthogonal functions (EOF) of boreal spring (March—April—May) mean precipitation rate (shaded, mm·month-1) and sea surface temperature (contours, ℃) in the global tropical ocean and the corresponding time series. (a) the first mode; (b) the second mode; (c) the corresponding time series (PC1 and PC2). Short dashed lines indicate the value of one standard deviation, the dotted line indicates the zero. The gray bars in c indicate the three-month mean Nino3.4 index in November—next January (NDJ) |
图2 海表温度(单位: ℃)和降水率(单位: mm·月-1)联合EOF分析前两模态时间序列分别与Niño3.4指数之间的超前滞后相关a. PC1与Niño3.4指数; b. PC2与Niño3.4指数。虚线为t检验95%置信度的阈值, 灰色矩形标注了第一、二模态发生的时间, 超前和滞后表示Niño3.4指数超前和滞后第一、二模态 Fig. 2 Lead and lag correlation coefficient between the Noni 3.4 index and PC1, and PC2, respectively. (a) PC1 and Niño3.4 index; (b) PC2 and Niño3.4 index. Dashed lines indicate the threshold value of 95% confidence level of Student’s t-test, the grey rectangles mark the mature phase of the first and second modes, the lead and lag indicate the Nino3.4 index leads and lags the first and second modes |
图3 热带大西洋降水率、海表温度、海表面压强和风场异常回归到海表温度和降水联合EOF分析第一模态时间序列上的超前滞后回归系数a, c, e: 降水(填色)和海表温度(等值线, 单位: ℃); b, d, f: 海表面压强(填色)和风场(矢量)。(a, b: 上一年12月到当年1月; c, d: 2月到3月; e, f: 4月到5月。图中只展示了超过t检验95%置信度的结果 Fig. 3 Anomalies of precipitation (Precip, mm·month-1), sea surface temperature (SST, ℃), sea level pressure (Slp, hPa), and 10 m wind (m·s-1) in the tropical Atlantic Ocean obtained by regressing to the PC1, respectively. (a, c, e) precipitation (shaded) and SST (contours); (b, d, f) Slp (shaded) and winds (vectors); (a, b) the last December to January; (c, d) February—March; (e, f) April—May. The regressed results exceeding 95% confidence level of Student’s t-test are shown |
图4 热带印度洋降水、海表温度、海表面压强和风场异常回归到海表温度和降水联合EOF分析第一模态时间序列上的超前滞后回归系数a, c, e, g: 降水(填色)和海表温度(等值线, 单位: ℃); b, d, f, h: 海表面压强(填色)和风场(矢量)。a, b: 上一年10月到11月; c, d: 上一年12月到当年1月; e, f: 2月到3月; g, h: 4月到5月。图中只展示了超过t检验95%置信度的结果 Fig. 4 Anomalies of precipitation (Precip, mm·month-1), sea surface temperature (SST, ℃), sea level pressure (Slp, hPa), and 10 m wind (m·s-1) in the tropical Indian Ocean obtained by regressing to the PC1, respectively. (a, c, e, g) precipitation (shaded) and SST (contours); (b, d, f) Slp (shaded) and winds (vectors); (a, b) the last October to November; (c, d) December to January; (e, f) February—March; (g, h) April—May. The regressed results exceeding 95% confidence level of Student’s t-test are shown. |
图5 热带南印度洋12°—8°S海表面高度(填色, 单位: cm)和10m大气风场旋度(等值线, 单位: ×10-6s-1)回归到海表温度和降水联合EOF分析第一模态时间序列上的超前滞后回归系数经向平均的经度-时间变化图图中只展示了超过t检验95%置信度的结果 Fig. 5 Longitude-time diagram of the meridional averaged anomalies of sea surface height (SSH, shaded, cm), and 10 m wind curl (contours, s-1) in the south tropical Indian Ocean between 12°—8°S obtained by regressing to the PC1, respectively. The regressed results exceeding 95% confidence level of Student’s t-test are shown |
图6 热带印度洋、大西洋和太平洋降水率、海表温度、海表面压强(hPa)和风场异常回归到海表温度和降水联合EOF分析第一和第二模态时间序列上的超前滞后回归系数和气候态降水和风场的纬向平均的纬度-时间变化图a, d, g: 印度洋(50°—90°E); b, e, h: 大西洋(60°W—20°E); c, f, i: 太平洋(140°E—80°W)。a, b, c: 降水(填色)和海表温度(等值线, 单位: ℃)的回归系数; d, e, f: 海表面压强(填色)和风场(矢量)的回归系数; g, h, i: 气候态降水(填色)和风场(矢量), 图中白色粗实线标注了气候态降水量为130 mm·月-1 的等值线(强降水中心)。印度洋和大西洋的各变量回归到第一模态; 太平洋的变量回归到第二模态。图中回归系数只展示了超过t检验95%置信度的结果 Fig. 6 Latitude-time diagram of the zonal averaged anomalies of precipitation (mm·month-1), sea surface temperature (℃), sea level pressure (hPa), and 10 m wind (m·s-1) in the tropical Indian Ocean, the Atlantic Ocean and the Pacific Oceans obtained by regressing to the PC1 and PC2, respectively, and the climatological mean precipitation and wind. The regressed results exceeding 95% confidence level of Student’s t-test are shown. (a, d, g) Indian Ocean (50°—90°E); (b, e, h) Atlantic Ocean (60°W—20°E); (c, f, i) Pacific (140°E—80°W); (a, b, c) regressions of precipitation (shading) and SST (contours); (d, e, f) regressions of sea level pressure (shaded), and 10 m wind (vectors); (g, h, i) climatological mean precipitation (shaded) and wind (vectors). White bold contours in e and f labels the heavy precipitation centers with precipitation rate larger than 130 mm·month-1. The anomalies in the tropical Indian Ocean and the Atlantic Ocean are regressed to the PC1; the anomalies in the Pacific ocean are regressed to the PC2 |
图7 热带太平洋降水率(填色)、海表温度(等值线, 单位: ℃)和风场(箭头)异常回归到海表温度和降水联合EOF分析第二模态时间序列上的超前滞后回归系数的季节平均值a. 春季(3—5月); b. 夏季(6—8月); c. 秋季(9—11月); d. 冬季(12月—次年2月)。图中只展示了超过t检验95%置信度的结果 Fig. 7 Anomalies of precipitation (mm·month-1) and sea surface temperature (℃) in the tropical Pacific Ocean obtained by regressing to the PC2, respectively. The regressed results exceeding 95% confidence level of Student’s t-test are shown. (a) spring (March—April—May, MAM); (b) summer (June—July—August, JJA); (c) fall (September—October—November, SON); (d) winter (December—January—February, DJF) |
[1] |
|
[2] |
|
[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] |
|
/
〈 | 〉 |