熱帶太平洋環流的動力: 演化、調節與解釋

全文

(1)國立台灣師範大學地球科學研究所博士論文. 熱帶太平洋環流的動力: 演化、調節與解釋 Dynamics of Tropical Pacific Ocean circulation: evolution, modulation, and interpretation. 王儷樵 Li-Ch iao Wang. 指導教授：吳朝榮博士 Advisor : Dr. Chau - Ron Wu. 中華民國一百零三年十月.

(2) 謝誌 (Acknowledgements). 本論文得以順利完成，首先要感謝我的指導老師-吳朝榮教授。謝謝老師的悉心指導，即便公務繁忙也不時關心我的進度，充分利用時間進行討論，讓這份工作盡善盡美。老師巧妙地扮演了我研究生涯中所有重要的角色：嚴父，慈母，甚至是一個朋友。研究之路偶有瓶頸，老師更引用了王鼎鈞的《人生三書》來勉勵我：「人生就是為了要解決問題。」使我如夢初醒。研究是人生的一部分，有問題本是常態，一路順遂才是稀有。如何在解決問題的過程中使自己成長、精進，才是最重要的。除了做研究，老師更教給了我寶貴的事：與其祈求萬事一帆風順，不如挺身做好面對問題的準備。另外一定要感謝的，是我的口試委員們：師大地科系的陳正達老師，台大大氣系的林依依老師，海大海洋環境資訊系的郭南榮老師以及中研院環境變遷研究所的許晃雄老師。謝謝委員們細心閱讀我的論文，在口試時給了我許多建議，使這篇論文朝著更完善的方向邁進。當然不能忘記，行雲流水研究室的各位學長學姊、學弟學妹以及三位助理們。你們的存在為我的研究生涯注入許多甘甜，即使是跌倒了因為有同袍在，再痛都能笑著爬起來。謝謝你們一直以來溫暖的照應和陪伴。最後還是要感謝最疼愛包容我的父母，做了這麼多年讀書人，您們還是開口只有勉勵，嘮叨放在心裡。在我疑惑的時刻，對我永遠都有滿滿的信心，是我的最強後盾，也是我努力的意義。謝謝您們。最後的最後，想把這篇論文獻給我在天上的外婆。親愛的外婆您在我口試前兩天住進醫院裡，謝謝您還等著我口試完，趕去醫院向您報喜。您開心地握著我的手：這樣啊！不久便陷入了迷離。那表情令我欣慰又難過。也許因為沒有牽絆，您離開得是那麼快。衷心希望您滿足地啟程，一路平安，到新世界去，展開一段嶄新的人生。.

(3) Abstract Outputs from ocean models are adopted to analyze the ocean circulations in the low-latitude Pacific and explain the responsible mechanisms behind them.. Due to. the prominent role that El Niño-Southern Oscillation (ENSO) plays in the tropical climate system, the analyses focused on ocean circulation patterns under the influence of ENSO in equatorial Pacific and the interaction between ENSO and other decadal variability in the extra-tropical region. Simulation outputs show that results for the central Pacific El Niño (CP-El Niño) corresponded well with previous studies which suggested that thermocline variations in the equatorial Pacific contain an east-west oscillation. The eastern Pacific El Niño (EP-El Niño) experienced an additional north-south seesaw oscillation between approximately 15N° and 15°S.. Moreover, the same outputs show that the flow. patterns in the CP-El Niño are consistent with previous El Niño studies that both the eastward Equatorial Undercurrent and westward South Equatorial Current (SEC) weaken.. On the other hand, the EP-El Niño displays a significantly distinct. circulation pattern.. The North Equatorial Counter-Current strengthens in the. developing phase and persists into the peak of the warm event, while the northern branch of the SEC also intensifies during the mature phase and lasts for about half a year.. The South Equatorial Counter-Current strengthens during the decaying phase. of EP-El Niño.. It is found that the shifting of the wind stress curl associated with the. thermocline variability is chiefly responsible for the unique performance of EP-El Niño. Simulated current velocity shows that wind stress curl anomaly (WSCA) in the region of 10°N–15°N and 160°E–170°E generates Rossby waves and affects the North Equatorial Current (NEC) bifurcation along the Philippine coast.. From 1976. to 1992, following a regime shift to the positive Pacific Decadal Oscillation (PDO) I.

(4) phase, PDO and ENSO match each other in strength and have a neutralized effect on the WSCA.. From 1993 to 2009, WSCA matches PDO well, and its correlation with. ENSO is lower.. A linear regression model demonstrates that the influence of PDO. has nearly thirteen times weight over that of ENSO.. Prior to the 1976 regime shift,. WSCA is closely related to ENSO from 1961 to 1975, and it does not correlate significantly with PDO.. It is suggested that Rossby waves are preferentially. generated in either the negative PDO phase when the ENSO signal dominates, or in the positive PDO phase when the ENSO signal is overshadowed.. In the phase when. the positive PDO counteracts with the ENSO signal, neither ENSO nor PDO has a significant influence on Rossby wave generations through the WSCA.. Key Words: El Niño-Southern Oscillation, eastern Pacific El Niño, central Pacific El Niño, wind stress curl, Rossby waves. II.

(5) 摘要 (Abstract in Chinese). 本研究使用海洋模式資料，分析太平洋低緯度區域洋流的演變過程，並探究其背後機制。由於聖嬰–南方振盪現象(El Niño-Southern Oscillation，ENSO)在太平洋熱帶區域扮演了極其重要的角色，本研究主要著重於 ENSO 在赤道區域對洋流的影響，以及 ENSO 和其他氣候因子交互作用之下，熱帶區域洋流所產生的變化。模式資料顯示，中太平洋聖嬰發生時，赤道地區的溫躍層產生了東西向振盪的變化，且南赤道洋流和赤道潛流明顯減弱，這和前人研究的結果是一致的；東太平洋聖嬰的情況完全不同：赤道區域的溫躍層產生了東西向及南北向的振盪，且自聖嬰的發展期開始，北赤道反流不斷增強，到了成熟期，南赤道洋流不僅未減弱，還持續增強了六個月。本研究進一步發現，風應力旋度透過艾克曼作用而帶動了溫躍層的變化，是造成兩種聖嬰現象之下太平洋環流有如此差異的關鍵。另外，模式資料顯示，中太平洋位於 10°N–15°N , 160°E–170°E 的風應力旋度場，激發了羅士培波(Rossby waves)，間接影響了北赤道洋流在菲律賓沿岸分支點的變化。1976~1992 年，剛從負相位轉為正相位的太平洋十年期振盪(Pacific Decadal Oscillation，PDO)和 ENSO 勢均力敵，效力中和，使得中太平洋風場失去了激發 Rossby waves 的能力，間接導致北太平洋分支點緯度偏南；1993~2009 年，PDO 主導了中太平洋的風場並激發了 Rossby waves，使得北太平洋分支點緯度偏北，其效力約為來自 ENSO 效力的 13 倍。而在 1976 年以前，當 PDO 仍為負相位時，中太平洋的風場由 ENSO 所主導。本研究發現，唯有 PDO 和 ENSO 兩者效力相差甚遠時，方能激發 Rossby waves 並使得北赤道洋流分支點往北偏移；當 PDO 初經歷相位轉換，與 ENSO 效力互相削弱，中太平洋的風場就失去了遠端影響北太平洋分支點的能力。關鍵字：聖嬰-南方振盪東太平洋聖嬰中太平洋聖嬰 III. 風應力旋度. 羅士培波.

(6) List of Contents Abstract. I. Abstract in Chinese. III. List of Contents. IV. List of Tables. VI. List of Figures. VII. Chapter 1: Introduction. 1. 1.1 Background information. 1. 1.2 Structure of the thesis. 3. Chapter 2: Contrasting the evolution between two types of El Niño. 6. 2.0 Abstract. 6. 2.1 Introduction. 7. 2.2 Data. 8. 2.3 Result and Discussion. 9. 2.3.1 Simulated flow patterns associated with two types of El Niño. 9. 2.3.2 The pattern of evolution for the two types of El Niño. 11. 2.3.3 Forcing mechanism. 15. 2.3.4 The distinct decay phase of the EP El Niño. 17. 2.4 Concluding Remarks. 19. Chapter 3: Contrasting the flow patterns in the equatorial Pacific between two types of El Niño. 29. 3.0 Abstract. 29. 3.1 Introduction. 30. 3.2 Data. 32 IV.

(7) 3.3 Results and Discussion. 33. 3.3.1 The equatorial currents. 33. 3.3.2 Spatial variations associated with CP- and EP-El Niño. 35. 3.3.3 Temporal variations of the equatorial currents. 37. 3.3.4 Forcing Mechanism. 40. 3.4 Concluding Remarks. 42. Chapter 4: Modulation of Rossby Waves on the Pacific North Equatorial Current Bifurcation Associated with the 1976 Climate Regime Shift. 56. 4.0 Abstract. 56. 4.1 Introduction. 57. 4.2 Data. 59. 4.3 Results. 60. 4.4 Mechanisms. 62. 4.5 Discussion. 65. 4.6 Concluding Remarks. 67. Chapter 5: Conclusions. 80. List of References. 82. V.

(8) List of Tables Table 1.1. Classification of the El Niño events……………………………………..5. Table 4.1. Correlation and linear regression coefficients………………………….70. VI.

(9) List of Figures Figure 2.1. (a) Mean surface circulation (averaged from 0 to 50 m) in the equatorial. Pacific based on GODAS model assimilation (Units: m/s).. Velocity. composites during the mature phase of (b) CP-El Niño events and (c) EP-El Niño events.. Shading indicates the current intensity. Contour interval is. 0.05 m/s. ……………….…………………………………………………21 Figure 2.2 The modeled 20°C isotherm depth anomalies together with the graphs for (a) September 1994, (b) December 1994, (c) March 1995, and (d) June 1995.. Contour interval for the D20 isotherm depth is 10 m. (e) Schematic. diagram for Figure 2.2b and (f) schematic diagram for Figure 2.2c. …….22 Figure 2.3 Sea surface height anomalies from AVISO in (a) September 1994, (b) December 1994, (c) March 1995, and (d) June 1995. Contour interval for sea surface height anomalies is 5 cm. …………………………………….23 Figure 2.4 Same as Figure 2.2 except in (a) September 1997, (b) December 1997, (c) March 1998, and (d) June 1998, with respective schematic diagrams in (e) to (h). ……………….……………………………………………………..24 Figure 2.5 Same as Figure 2.3 except in (a) September 1997, (b) December 1997, (c) March 1998, and (d) June 1998. ………………………………………….25 Figure 2.6 (a) Wind stress curl averaged over the southwestern Pacific (150~180°E, 5~7°S) during the mature phase of an El Niño between 1980 and 2010. Wind stress curl pattern composited for the mature phase of (b) the EP-El Niño events, and (c) the CP-El Niño events.. Contour interval for wind. stress curl is 0.5 × 10-7 Nm-3. ……………………………………………..26 Figure 2.7 The modeled 20°C isotherm depth anomalies in (a) March 1983 and (b) June 1983. Contour interval is 10 m. …………………………………...27 Figure 2.8 Wind stress curl (shading) and wind stress (vector) averaged from May to VII.

(10) July in (a) 1983 and (b) 1998. Negative contours are shaded. Vector scale is 0.1 m/s. The red rectangle encloses the region from 1°N–6°S, 170°E–130°W in the central equatorial Pacific. ………………………….28 Figure 3.1 (a) Mean surface circulation (averaged from 0 to 50 m) in the equatorial Pacific based on GODAS model assimilation (Units: m/s).. The shading. indicates the current intensity. Contour interval of is 0.05 m/s. (b) Vertical velocity profile along the equator averaged from 2°N ~ 2°S from the assimilated annual mean (Units: m/s). Contour interval is 0.1 m/s. and. blue. shading. indicate. the. eastward. and. westward. Red. current,. respectively. ……………….………………………………………………45 Figure 3.2 The annual cycle of zonal currents on the equator from the model assimilation at (a) 165°E, (b) 140°W, and (c) 110°W, respectively (Units: m/s).. (d)-(f) are the same as (a)-(c), but based on TAO array data (Figure 2. of Keenlyside and Kleeman, 2002).. Contour interval is 0.05 m/s. ……...46. Figure 3.3 Composited meridional velocity profiles from GODAS along 140°W in winter of (a) normal years, (b) CP-El Niño, (c) EP-El Niño.. Red and blue. shading indicate the eastward and westward current (Units: m/s).. Contour. interval is 0.1 m/s. . ………………………………………………………..47 Figure 3.4 Same as Figure 3.1a, but in winter of (a) normal years, (b) CP-El Niño, (c) EP-El Niño ……………….………………………………………………..48 Figure 3.5 Composited velocity anomalies of the SECn during the evolution of El Niño.. In (a) EP-El Niño and (b) CP-El Niño.. Positive values (red. shading) indicate westward anomalies, while negative values (blue shading) indicate eastward anomalies. and depth from 0 to 300 m.. The data are averaged over the region 0~7°N, Contour interval is 0.02 m/s. ………………49. Figure 3.6 Same as Figure 3.5 except for the NECC. VIII. Positive values (red shading).

(11) indicate eastward anomalies, while negative values (blue shading) indicate westward anomalies. The data are averaged over the region 3.5°N ~ 10°N, and depth from 0 to 500 m.. Contour interval is 0.01 m/s. ………………50. Figure 3.7 Velocity intensity pattern of (a) the SECn during the mature phase together with (b) the NECC during the developing phase of the two EP El Niño events.. Black line indicates the 82/83 event, while red line indicates the. 97/98 event.. (Units: m/s). (c) and (d) show the t value of the SECn and. NECC during the composited EP-El Niño (black line). the 95% confidence level.. Red line denotes. Light-blue color highlights the significant. zone. ……………….………………………………………………………51 Figure 3.8 Statistical t value (bars) for (a) the SECn during the mature phase and (b) the NECC during the developing phase between EP-El Niño and CP-El Niño. Red curve denotes the 90% confidence level. ……………………52 Figure 3.9 (a) Wind stress curl pattern and (b) 20°C isotherm depth anomalies in September 1997. Contour interval for wind stress curl is 0.5×10-7 Nm-3, and for the isotherm depth is 10 m. …………………………………….…53 Figure 3.10 Same as Figure 3.9 except in June 1998. ………………………………..54 Figure 3.11 Schematic diagram illustrating the flow patterns between the eastward NECC and westward SECn during El Niño evolutional phases.. The. colored arrows indicate the currents and the solid lines display the variation of the thermoline around the equator. ……………………………………..55 Figure 4.1 Wind stress curl anomalies generated from NCEP/NCAR (shading) together with modeled velocity field (vectors) where the magnitude larger than 0.2 m/s is presented. (averaged from 0 to 50 m) in (a) July 1984 and (b) July 2003. Vector scale is 0.7 m/s. The black rectangle identifies the region from 10–15°N and 160–170°E in the central equatorial Pacific. Contour IX.

(12) interval for wind stress curl is 10-8 Nm-3. ………………………………...71 Figure 4.2 Meridional velocity anomalies in the region of 130–170°E (a) from 1982 to 1992 and (b) from 1993 to 2005. Data are averaged over the 10–15°N region. Contour interval is 0.05 m/s. Rossby waves are highlighted with yellow lines….…………………………………………………………….72 Figure 4.3 Wind stress curl anomalies in the region of 160–170°E (a) from 1982 to 1992 and (b) from 1993 to 2005.. Data are averaged over the 10–15°N. region. Contour interval for wind stress curl anomaly is 10-8 Nm-3. ……..73 Figure 4.4 (a) Time series of wind stress curl anomaly averaged in the region of 10–15°N and 160–170°E (gray bars, units: 10-7 Nm-3), together with Niño-3.4 (red curve, units: °C), PDO (blue curve, units: °C) and (b) the correlation coefficient between wind stress curl anomaly in the region of 10–15°N and 160–170°E and those in the North Pacific from 1976 to 1992. Correlation coefficients in that are larger than or equal to 0.4 are shown in colored shading and those larger than 0 but less than 0.4 are shown by contours. Contour interval is 0.1. …………………………………………74 Figure 4.5 Same as Figure 4.4 except from 1993 to 2009. …………………………..75 Figure 4.6 The EKE pattern in (a) October 1985 and (b) October 1997 (Units: m2/s2). Contour interval is 0.005. ………………………………………………...76 Figure 4.7 Time series of wind stress curl anomalies (blue curve, units: Nm-3) in C-BOX and the EKE (red curve, units: m2/s2) averaged in (10–20°N, 130–150°E) from 1982 to 2005. ………………………………………….77 Figure 4.8 Same as Figure 4.4 except from 1961 to 1975. …………………………..78 Figure 4.9 Time series of wind stress curl anomaly averaged in the region of 10–15°N and 160–170°E (gray bars, units: 10-7 Nm-3), together with Niño-3.4 (red curve, units: °C), PDO (blue curve, units: °C) from 1961 to 2009. ……….79 X.

(13) Chapter 1 Introduction 1.1 Background information El Niño-Southern Oscillation (ENSO) is the most notable interannual variability in the tropical Pacific climate system.. The observed temporal evolution, spatial. features and interactions between oceanic and atmospheric variables have been broadly explored and profoundly discussed.. Over the past two decades, a thorough. understanding about the effects of ENSO on variations in either global climate or regional weather phenomena have been achieved through a series of studies based on observations and model simulations. The analyses of ENSO impacts on ocean activities are much more limited. Due to the sensitivity of atmospheric variability to fluctuations of sea surface temperature (SST), the description of ENSO in previous studies is often based on the SST variability, which is studied extendedly. However, the equatorial currents and the processes that modulate their circulation pattern are less discussed due to the lack of measurements in the region.. On interannual or longer timescale how the equatorial. currents vary is not yet well understood and is of interest to further investigate the dynamics involved.. The current fluctuation in the tropical domain is largely. influenced by the atmospheric condition, which in turn is modulated by the ENSO phenomenon. Although ENSO has been recognized as important variability which involves in a substantial portion of many climate phenomena, it is not appropriate to take it as an invariable “background state”.. Complex ENSO evolution feature has been studied. recently. These studies suggest that there are two different types of El Niño (Ashok et al., 2007; Kug et al., 2009; Kao and Yu, 2009; Yeh et al., 2009).. The 1982/83 and. 1997/98 events were representative of the canonical eastern Pacific El Niño (EP-El 1.

(14) Niño, Kao and Yu 2009), which is also referred to as a cold tongue (CT) El Niño (Kug et al., 2009). Table 1.1 summarizes the characteristic and classification of El Niño events.. In Table 1.1, El Niño Modoki is defined according to the locations where. SST anomalies occur (Ashok et al., 2007). During El Niño Modoki, SST anomalies appear associated with a horse-shoe pattern in the central Pacific. Kau and Yu (2009) pointed out that during the eastern Pacific El Niño, SST anomalies emerge from the coast of the South America, propagating westward to the central Pacific, and decay off the equator.. In the central Pacific El Niño, SST anomalies first appear around the. date line, developing and mature in a V-shaped anomaly structure extending toward the subtropics in both hemispheres, and then decay in the equatorial central Pacific. Kug et al. (2009) found and defined CT El Niño (canonical type) with the pattern that large SST anomalies develop in the eastern Pacific during the developing summer. The maximum value of the warm SST is located in the coastal region of the eastern boundary. SST anomalies of warm pool (WP) El Niño are mostly confined in the central Pacific during boreal summer.. One objective of this study is to contrast the. temporal evolution in subsurface-surface structure and the circulation patterns in the tropical Pacific during different flavors of El Niño. Furthermore, how longer timescale variation interacts with ENSO, for instance, the ocean current variations under the influences of both ENSO and extra-tropical decadal variability, is not yet fairly understood.. The tropical and extra-tropical. interaction can occur via meridional ocean circulations, and in turn it brings about irregularity behavior of ocean flow patterns. From tropical to extra-tropical domain in the northwestern Pacific, modulation of Rossby waves associated with ENSO and Pacific Decadal Oscillation (PDO) on the North Equatorial Current (NEC) bifurcation is examined in this study.. In the northwestern Pacific, the westward-flowing NEC 2.

(15) splits as it encounters the Philippine coast.. The poleward-flowing branch forms the. root of the Kuroshio, while the equatorward-flowing branch becomes the Mindanao Current (MC). Both of these western boundary currents ultimately recirculate eastward as the Kuroshio Extension and North Equatorial Countercurrent, indicating that the NEC plays a noticeable role in regulating the North Pacific subtropical and tropical gyres.. Specifically, the bifurcated NEC provides important pathways for. heat, mass, and salt transport exchanges between the Kuroshio and MC, which approximately represent the mid-latitude and low-latitude western North Pacific, respectively (Lukas et al., 1991).. Therefore, variation in the NEC bifurcation. latitude, accompanied by a quantitative change in the partitioning of the NEC transport between the Kuroshio and MC (e.g., Qiu and Lukas, 1996; Kim et al., 2004), plays an essential role in mid-latitude and low-latitude circulation changes.. 1.2 Structure of the thesis The thesis includes three basic topics and is organized as follows.. For the first. topic, based on outputs of a data assimilation model, the distinct evolution patterns between two types of El Niño are examined and the responsible mechanism is explained in Chapter 2.. For the second topic, due to the close relationship between. ocean and atmosphere, it is worthwhile to understand the ocean circulation behavior under different flavors of El Niño.. The flow patterns in the equatorial Pacific. between two types of El Niño are contrasted in Chapter 3.. For the third topic, in. addition to the air-sea interaction taking place among equatorial region, I am also interested in the low-latitude ocean current variations in the presence of ENSO and other long timescale variability.. The modulation of the NEC bifurcation by ENSO. and PDO via generations of Rossby waves is investigated in Chapter 4. 3. The.

(16) conclusion is given in Chapter 5.. 4.

(17) Table 1.1 Classification of the El Niño events. El Niño Modoki Ashok et al. (2007). Canonical type. New type. Canonical El Niño SSTA: warm in the Eastern Pacific; With eastward Kelvin waves. El Niño Modoki SSTA: warm in the Central Pacific; Without eastward Kelvin waves. Years:. Years: 1979/80、1986/87、 1990/91、1991/92、 1992/93、1994/95、 2002/03、2004/05. The other events.. EP and CP El Niño Kao and Yu (2009). EP El Niño SSTA: warm in the Eastern Pacific associated with thermocline. The third type (cannot be classified yet):. CP El Niño SSTA: warm in the Central Pacific influenced by the atmospheric forcing. variations, surface winds Years: Years: 1972/73、1976/77、 1991/92、1994/95、 1982/83、1997/98 2002/03、2004/05、 2006/07 CT and WP El Niño Kug et al. (2009). CT El Niño SSTA: warm in the Niño-3 (5°S~5°N, 90°W~150°W). WP El Niño SSTA: warm in the Niño-4 (5°S~5°N, 150°W~160°E). The third type SSTA: warm in Niño-3.4 (5°S~5°N, 120°W~170°E). Years: Years: Years: 1972/73、1976/77、 1977/78、1990/91、 1986/87、1987/88、 1982/83、1997/98 1994/95、2002/03、 1991/92 2004/05. 5.

(18) Chapter 2 Contrasting the evolution between two types of El Niño An edited version of this paper was available online: Wu, C.-R. and L.-C. Wang (2013), Contrasting the evolution between two types of El Niño. in. a. data. assimilation. model. Ocean. Dynamics,. 63,. 577-587,. doi:10.1007/s10236-013-0610-8.. 2.0 Abstract Simulation outputs were used to contrast the distinct evolution patterns between two types of El Niño.. The modeled isotherm depth anomalies closely matched. satellite sea surface height anomalies.. Results for the El Niño Modoki (central. Pacific El Niño) corresponded well with previous studies which suggested that thermocline variations in the equatorial Pacific contain an east-west oscillation. The eastern Pacific El Niño experienced an additional north-south seesaw oscillation between approximately 15N° and 15°S.. The wind stress curl pattern over the. west-central Pacific was responsible for the unusual manifestation of the eastern Pacific El Niño. The reason why the 1982/83 El Niño was followed by a normal state whereas a La Niña phase developed from the 1997/98 El Niño is also discussed. In 1997/98, the Intertropical Convergence Zone (ITCZ) retreated faster and easterly trade winds appeared immediately after the mature El Niño, cooling the sea surface temperature in the equatorial Pacific and generating the La Niña event.. The slow. retreat of the ITCZ in 1982/83 terminated the warm event at a much slower rate and ultimately resulted in a normal phase.. 6.

(19) 2.1. Introduction The recharge-discharge oscillator theory (e.g., Jin, 1997a, b; Meinen and McPhaden, 2000) describes the El Niño Southern–Oscillation (ENSO) phenomenon as the zonal transport of warm water along the equator associated with El Niño and La Niña. The recharge and discharge can be estimated by variations in the volume of water warmer than 20°C, also known as the warm water volume (WWV). The WWV builds up at the equator prior to El Niño and is transported to off-equatorial Pacific locations during the mature warm phase.. The recharge of the WWV in the. equatorial Pacific is a necessary condition for the development of El Niño (e.g., Cane et al., 1986). Recent studies have shown that there are two types of El Niño (e.g., Ashok et al., 2007; Yu and Kao, 2007; Kug et al., 2009). The 1982/83 and 1997/98 events were representative of the canonical eastern Pacific El Niño (EP-El Niño, Kao and Yu, 2009), which is also referred to as a cold tongue El Niño (Kug et al., 2009). The more recently identified El Niño type has a distinct difference in the location of maximum sea surface temperature (SST) anomalies from the canonical El Niño. This type has been termed the El Niño Modoki (Ashok et al., 2007), as well as a central Pacific El Niño (CP-El Niño, Kao and Yu, 2009), Date Line El Niño (Larkin and Harrison, 2005), or Warm Pool El Niño (Kug et al., 2009).. The. recharge-discharge oscillator theory has been applied to these two different types of El Niño.. It has been suggested that the EP-El Niño is produced by basin-wide. thermocline variations, whereas the CP-El Niño involves only local air-sea interactions (e.g., Kao and Yu, 2009; Kug et al., 2009; Yu and Kim, 2010). These studies concluded that the EP-El Niño acts as a mechanism to remove excess heat content to off-equatorial Pacific locations and operates as proposed by the recharge-discharge theory. 7.

(20) The canonical ENSO recharge-discharge theory does not emphasize the important role of the wind stress curl on the recharge-discharge process (e.g., Jin, 1997a, b). According to Jin’s parameters (Jin, 1997b), the zonal wind stress has a meridional scale that is so large that the stress remains almost constant near the equator and its curl is negligible. However, thermocline depth variations are highly related to surface wind variations, and the essential role of wind stress curl in controlling the thermocline displacement cannot be ignored. Clarke et al. (2007) also emphasized that the wind stress curl over the west-central Pacific is closely linked to equatorial thermocline variations. The wind stress curl variation and its intensity near the equatorial region trigger both zonal and meridional transports, which can explain the recharge and discharge of the WWV. The present report is structured as follows. Section 2 describes the datasets used in this study.. In section 3 we reveal two distinct patterns of El Niño evolution. and their forcing mechanisms, and explore why two different decay patterns exist for the EP-El Niño. Section 4 presents our conclusions.. 2.2 Data The 20°C isotherm depth anomaly and current velocity data used in this study were based on the National Centers for Environmental Prediction (NCEP) Global Ocean Data Assimilation System (GODAS) product (Behringer and Xue, 2004). The GODAS domain covers 65°N to 75°S with a resolution of 1/3° × 1/3°. The temporal range extends from January 1980 to September 2010. The model has 40 levels with 10-m resolution near the sea surface and is forced by momentum, heat, and freshwater fluxes from the NCEP-Department of Energy (NCEP-DOE) reanalysis 2 (R2) (Kanamitsu et al., 2002), which is an improved version of the NCEP Reanalysis 1 (R1) model with updated parameterizations of physical processes and 8.

(21) several errors fixed. The temperature profiles assimilated in GODAS include those from the Tropical Ocean and Global Atmosphere-Tropical Atmosphere Ocean (TOGA-TAO), Pilot Research Array in the Tropical Atlantic (PIRATA), Triangle Trans-Ocean Buoy Network (TRITON) moorings (McPhaden et al., 2001), bathythermographs (XBTs), and Argo profiling floats.. In addition to the. temperature, a synthetic salinity profile was computed for each temperature profile using a local temperature-salinity climatology based on the annual mean fields of temperature and salinity from the National Oceanographic Data Center (NODC) World Ocean Database (Conkright et al., 1999). Surface wind stress data from 1948 were provided by the NCEP/National Center for Atmospheric Research reanalysis project (R1) (Kistler et al., 2001) and were also used to calculate the wind stress curl. Monthly averages on global grids (2.5° × 2.5°) were available from the Climate Diagnostics Center of the National Oceanic and Atmospheric Administration’s (NOAA) (http://www.cdc.noaa.gov/).. Earth System Research Laboratory. Sea surface height anomalies were obtained from. AVISO (Archiving, Validation and Interpretation of Satellite Oceanographic data; http://www.aviso.oceanobs.com) and were merged from altimeter observations by multiple satellites, including TOPEX/Poseidon, European Remote Sensing satellites 1 and 2 (ERS-1&2), Jason-1 and 2, Envisat, and Geosat Follow On (GFO).. The. product was gridded on a Mercator grid of 1/3° × 1/3° and a time interval of seven days, and is available from October 1992 onward.. 2.3 Result and Discussion 2.3.1 Simulated flow patterns associated with two types of El Niño The general circulation pattern of the equatorial Pacific was quite realistic in the GODAS model. Figure 2.1a shows the simulated mean circulation pattern of the 9.

(22) Pacific equatorial current based on GODAS products and is consistent with the observed mean flow pattern in the region (e.g., Lukas, 2001). The equatorial zonal currents consist of a westward-flowing wide-range surface current, the South Equatorial Current (SEC), which is located between about 8°S and 3°N.. Following. the definition of Wyrtki (1974), the SEC is further split into a northern branch (SECn) and a southern branch (SECs) at the equator.. To the north, the eastward-flowing. current is the North Equatorial Counter Current (NECC, between about 5°N and 10°N).. North of the NECC, there is an intense westward current, the North. Equatorial Current (NEC, between about 10°N and 20°N). The NEC is the southern limb of the North Pacific subtropical gyre and the upstream of the Kuroshio.. The. South Equatorial Counter Current (SECC) extends eastward from the western boundary region, but it only intermittently reaches the central and eastern Pacific. Wang and Wu (2012) showed that the vertical profile along the equator from GODAS is quite similar to that plotted by Lukas (2001) based on the NOAA/NCEP ocean assimilation/reanalysis product.. Furthermore, the annual cycle of zonal currents on. the equator from GODAS is also consistent with that based on TAO array data (Wang and Wu, 2012). In addition to the mean circulation, the GODAS model reveals distinct circulation patterns between El Niño Modoki (CP-El Niño) and EP-El Niño. Figures 2.1b-c show the spatial circulation patterns in the equatorial Pacific composited by the mature phases of El Niño Modoki and EP-El Niño years, respectively.. El Niño. Modoki years include 1991/92, 1994/95, 2002/03, 2004/05, and 2006/07. Typical EP-El Niño events occurred in 1982/83 and 1997/98.. The phases of El Niño. evolution have been defined by Kug et al. (2010). The period of the developing phase is from March to November, the mature phase is from December to February of the following year, and the decay phase is from February to October. 10. The mean.

(23) circulation in Figure 2.1a shows that the SECn and SECs are clearly separated from each other.. In the mature phase of El Niño Modoki, the eastward NECC strengthens. while the westward SECn and SECs weaken compared with the velocity composite of normal years (Figure 2.1b). The SECn becomes much weaker and is confined to the east of 125°W.. The SECs is probably blocked by the shoaling Equatorial. Undercurrent (EUC) and is seldom able to reach the western boundary. The circulation during the mature phase of EP-El Niño has a distinctive pattern. The SECn and SECs merge together and strengthen. This unusual pattern appears only during the EP-El Niño and is very different from the events considered in the earlier El Niño studies. Moreover, it is often observed that the NEC tends to migrate northward rather than flowing westward during an El Niño event.. Figure 2.1c shows. that the latitude of the NEC is ~16°N as it reaches the western Pacific boundary at the Philippine coast.. The bifurcation point of the NEC is much farther north during an. EP-El Niño than during an El Niño Modoki (CP-El Niño).. 2.3.2 The pattern of evolution for the two types of El Niño The pattern of evolution for the El Niño can be followed through depth changes in the equatorial Pacific thermocline.. Two distinct patterns of El Niño evolution. have been documented using the tropical 20°C isotherm depth (D20) of the GODAS product.. The D20 is located in the middle of the main thermocline throughout the. region (Kessler, 1990; Ji and Leetmaa, 1997; Braganza, 2008). Figure 2.2 shows the typical evolution of an El Niño Modoki (CP-El Niño) in 1994/95.. CP type D20. anomalies are relatively weaker and found at depths less than 30 m.. In September. 1994, during the onset of a warm El Niño, a positive D20 anomaly occurred in the central and eastern Pacific (Figure 2.2a). The positive anomaly expanded zonally and generated a basin-wide warming.. In December 1994, when the warm event 11.

(24) reached its mature phase, the positive D20 anomaly began to concentrate in the eastern Pacific at 5°N~5°S. A negative D20 anomaly occurred in the western Pacific, but was not as obvious (Figure 2.2b).. The schematic diagram in Figure 2.2e. demonstrates that the thermocline depth was elevated in the western Pacific but deepened in the eastern Pacific. During the decaying phase in March 1995, the positive D20 anomaly was replaced by a weak basin-wide negative D20 anomaly (Figure 2.2c), tilting the thermocline toward the opposite side (Figure 2.2f). When the La Niña was developing in June 1995, a negative anomaly was observed in the eastern Pacific, although it was still weak (Figure 2.2d).. Similar patterns of. evolution have been observed in the other El Niño Modoki events (e.g., 1991/92, 2002/03, 2004/05, 2006/07, and 2009/10). To relate the D20 anomalies to the sea surface height pattern in the tropic Pacific Ocean, sea surface height anomalies based on AVISO satellite altimeter observations in 1994/95 are shown in Figure 2.3. The development and decay of the 1994/95 El Niño are evident in the gridded sea surface height anomalies. A high sea surface first appeared in the east-central equatorial Pacific in September 1994 prior to the El Niño. The large and positive sea surface height anomalies extended to the eastern Pacific during the mature phase in December 1994. During the decay phase in March 1995, negative sea surface height anomalies developed in the eastern Pacific, before becoming weaker in June 1995 (Figures 2.3c-d). These patterns of evolution of satellite sea surface height anomalies, which show the zonal seesaw oscillation during an El Niño Modoki, corresponded well with modeled thermocline depth variations, further validating the GODAS product. Relatively strong thermocline depth variations were observed during the EP-El Niño event (Figure 2.4).. In September 1997 during the onset of an EP-El Niño, a. large and negative D20 anomaly (> 30 m) was observed over the region 5°N~15°N, 12.

(25) 140°E~170°E (Figure 2.4a). Unlike the El Niño Modoki pattern, the D20 anomaly occurred not only around the equator, but also extended to the extratropics. This resulted in a thermocline that tended to tilt in a peculiar way. in a schematic diagram.. Figure 2.4e shows this. The thermocline in the northwestern Pacific was higher. than the gradient between the eastern and western Pacific, which was observed during an El Niño Modoki.. In December 1997 during the mature phase, the negative D20. anomaly had extended to the east and had become larger (> 50 m) (Figure 2.4b). The symmetric anomaly around the equator generated a thermocline pattern similar to that in the mature phase of El Niño Modoki (Figure 2.2e or Figure 2.4f), except for a larger D20 anomaly gradient between the east and west Pacific.. In March 1998,. during the decay phase, the negative anomaly separated into two portions.. One. portion kept moving southeastward and extended over the region 5°S~15°S, 160E°~150°W (Figure 2.4c). Figure 2.4g shows that the thermocline tilted to the other side and became elevated in the southwestern Pacific (Figure 2.4g). Another portion was trapped at the equator, becoming eastward-propagating Kelvin waves. The D20 fluctuations are the manifestation of Kelvin waves. The occurrence of upwelling Kelvin waves agrees with findings of previous studies (e.g., Wang et al., 1999; Luo and Yamagata, 2001).. In June 1998, negative D20 anomalies in the form. of Kelvin waves shifted eastward to about 120°W~150°W at the equator (Figure 2.4d). The negative D20 anomalies moved eastwards and raised the thermocline depth in the eastern Pacific.. The thermocline depth gradient in the east-west direction was. weakened or even reversed (Figure 2.4h).. Another recent EP type of El Niño. (1982/83) displayed a similar evolution pattern except during the decay phase. Such differences might affect SST variations and the evolution of the following La Niña, as is discussed further in Section 3.4. The satellite-observed sea surface height anomaly patterns are also comparable 13.

(26) to the simulated thermocline depth variations observed during the 1997/98 EP-El Niño. Figure 2.5 shows a distinctly different pattern from the sea surface height anomalies observed in 1994/95 (Figure 2.3). Compared with 1994/95, the monthly sea surface height anomaly patterns in 1997/98 fluctuated to a larger degree. A strong and positive sea surface height developed in the central and eastern Pacific Ocean in September 1997 (Figure 2.5a).. It further intensified and the maximum sea. surface height anomaly shifted to the eastern boundary in the mature phase of December 1997 (Figure 2.5b). After the peak of the El Niño event, the positive sea surface height anomaly gradually declined (Figures 2.5c-d) as previously described using satellite SST data.. During its evolution, the negative sea surface height. anomaly displayed a meridional oscillation in the western Pacific during EP-El Niño, similar to the development of the negative D20 anomaly. The lower sea surface height anomaly occupied the western Pacific (Figure 2.5a), bifurcating into the two components shown in Figure 2.5b. The southeastward component appeared to be stationary in the southwestern Pacific, whereas the eastward component extended to the eastern Pacific roughly following the equator (Figures 2.5c-d). In summary, the equatorial thermocline depth variation of an El Niño Modoki (CP-El Niño) demonstrated an east-west seesaw oscillation.. However, in addition to. the east-west oscillation, the EP-El Niño displayed unusual behavior in the way the D20 tilted, with an additional north-south seesaw oscillation between approximately 15N° and 15°S.. From September to December when the EP-El Niño was. developing, the D20 in the eastern Pacific was almost 100 m deeper than the corresponding D20 in the west-central Pacific.. The D20 in the northwestern Pacific. was about 50 m shallower than in the southwestern Pacific (Figure 2.4b).. After the. peak of the EP-El Niño, the sharp east-west D20 gradient gradually diminished or even reversed. The meridional seesaw oscillation also reversed and the D20 in the 14.

(27) northwestern Pacific became deeper than that in the southwestern Pacific (Figure 2.4d). The meridional seesaw oscillation could be interpreted as the movements of the elevated D20. The elevated D20 in the northwestern Pacific migrated toward the southeast.. Part of it was trapped at the equator, becoming eastward-propagating. Kelvin waves. For EP-El Niño events, the ENSO oscillator theory should take the north-south oscillation within 15N°~15°S into consideration.. 2.3.3 Forcing mechanism The meridional oscillation of the thermocline depth could be explained by the effect of the tropical wind stress curl pattern. Previous studies have shown that the Intertropical Convergence Zone (ITCZ) tends to move southward during El Niño events (e.g., Rasmusson and Carpenter, 1982; Philander, 1985; Wolter and Timlin, 1993), and so does the zero wind stress curl line (the nodal line separates into positive curls to the north and negative curls to the south).. However, the meridional. migration of the zero wind stress curl line differs among El Niño events.. Figure 2.6a. shows the wind stress curl averaged over the southwestern Pacific (150°E~180°E, 5°S~7°S) during the mature phase of El Niño events between 1980 and 2010. We focused on the west-central Pacific because the tropical wind anomaly associated with the ENSO is large over the western to central Pacific Ocean (Deser and Wallace, 1990; Jin, 1997a), indicating that the main ocean-atmosphere coupling occurs in the west-central equatorial Pacific rather than in the eastern equatorial Pacific (as demonstrated by Clarke et al., 2007). A large and positive curl was observed during the mature phases of the 1982/83 and 1997/98 El Niño (EP type) events (Figure 2.6a). Wind stress curl patterns composited by the mature phase of the two EP-El Niño events and the other El Niño Modoki (CP-El Niño) events are shown in Figures 2.6b 15.

(28) and 2.6c, respectively. During the mature phase of the EP-El Niño events, the zero wind stress curl line shifted southward and large positive curls occupied the western equatorial region.. Different wind stress curl patterns will trigger distinct thermocline. variations within the tropical Pacific, which will in turn affect the SST evolution between the two types of El Niño. During the onset of El Niño Modoki (CP-El Niño) events, a positive wind stress curl occurred in the northwestern Pacific and a negative curl was located to the southwest, which agrees well with the preconditions of the warm water discharge process.. However, during the onset of EP-El Niño events, large positive curls in the. northwestern Pacific induced an upwelling through Ekman pumping and raised the thermocline depth (figures not shown). A raised D20 was clearly observed in the northwestern Pacific (Figure 2.4a). During mature EP-El Niño events, the zero wind stress curl line migrated southward toward the equator. A strong positive wind stress curl occurred in the northwestern Pacific and a rather strong negative curl was located to the southwest, both elevating the thermocline depth and inducing the WWV discharge. As a result, the D20 in the western Pacific was raised (Figure 2.4b). This phenomenon prevailed in the western Pacific, which agrees with several recent studies which concluded that the main region of ENSO air-sea interaction occurs in the west-central Pacific (e.g., Clarke et al., 2007). When evolving into the decay phase, the zero wind stress curl line began to shift northwards, while a negative wind stress curl occurred near the equatorial region.. The negative curl in the northwestern. Pacific deepened the thermocline and resulted in a recharge of the WWV, while the negative curl in the southwestern Pacific elevated the thermocline depth and then induced a discharge of the WWV. The elevated thermocline depth appeared to occur in the northwestern Pacific during the event onset, extending to the equator during the mature phase. During the decay phase, it separated into two portions extending to 16.

(29) the southern Pacific Ocean and propagating eastward along the equator, respectively. Furthermore, in addition to the meridional migration of the zero wind stress curl line, the intensity of wind stress curl was much stronger in an EP-El Niño event than an El Niño Modoki (CP-El Niño) event as shown in Figures 2.6b and 2.6c. The stronger wind stress curl resulted in stronger thermocline depth variations in an EP-El Niño. The negative D20 anomalies of an EP-El Niño extended to 70 m and were much larger than the D20 anomalies of a CP-El Niño (20 m), resulting in SST anomalies for the EP type being stronger than those in the CP type.. Both the large. magnitude and the meridional seesaw pattern of the oscillation were characteristics of EP-El Niño, and the meridional seesaw oscillation of the equatorial thermocline was not only a necessary condition, but also a sufficient precondition of an EP-El Niño.. 2.3.4 The distinct decay phase of the EP El Niño The 1982/83 EP-El Niño displayed a similar evolution pattern to the 1997/98 EP-El Niño except during the decay phase.. In March 1983, after the peak of the El. Niño, a negative D20 anomaly in the southern Pacific reached 50 m, which was much deeper than its counterpart trapped at the equator (~10 m) (Figure 2.7a).. The pattern. deviated from that in March 1998, when the D20 anomalies were almost equal (~40 m) between the southern Pacific and equator (see Figure 2.4c). Such differences affect the evolution of the cold event following the El Niño. For example, the 1982/83 warm event was brought to a normal state, whereas the 1997/98 El Niño was followed by three years of La Niña.. In 1982/83, because the negative D20 anomaly in the. equator was relatively weak, it almost disappeared during its eastward propagation as Kelvin waves (Figure 2.7b). Consequently, the negative SST anomaly in 1983 could not be defined as a La Niña using the Nino 3.4 index (5°S~5°N, 120°W~170°W). However, in 1997/98 the negative D20 anomaly at the equator was large and became 17.

(30) stronger as it propagated eastward during the decay phase, while the anomaly in the southern Pacific became weaker (Figure 2.4d).. From August to September 1998,. negative D20 anomalies at the equator occupied the region of 120°W~150°W and thereafter spread westward.. This induced a particularly strong negative SST. anomaly, followed by three years of cold events. Although thermocline variations could lead to distinct decay patterns, the direct effects are likely to be wind stress curl patterns in the equatorial Pacific.. Wind stress. curl shifted northward far away from the equator during both EP-El Niños, but their pattern in the southern equatorial Pacific was different.. During the decay phase in. 1982/83, a negative curl occurred south of the equatorial region (red rectangle in Figure 2.8a).. However, in the 1997/98 decay phase, positive wind stress curls. together with strong easterly trade winds occurred in the southern equatorial Pacific Ocean (Figure 2.8b).. The wind stress curl variation was connected to the ITCZ. migration. During the decay phase of an El Niño, the ITCZ generally starts to retreat northward. This movement of the ITCZ was faster during 1997/98 than in 1982/83 (figure not shown).. Recent research has demonstrated how the retreat of an. equatorial ITCZ abruptly terminates El Niño events, bringing back easterly winds accompanied by a cooling at the equator (Lengaigne and Vecchi, 2009). Thus, in 1997/98, easterly trade winds appeared immediately after the mature phase, due in part to the faster retreat of the ITCZ.. The trade winds cooled the SST in the. equatorial Pacific and immediately initiated the transition phase into a La Niña event. However, in 1982/83, the ITCZ retreated relatively slowly and the trade winds did not appear quickly enough to cool down the SST. As a result, the transition phase returned to normal conditions.. Specifically, the weak negative thermocline. anomalies under the prolonged warming conditions were not powerful enough to instantly convert the 1982/83 El Niño into a La Niña. 18.

(31) 2.4 Concluding Remarks Two distinct patterns of El Niño evolution were identified using the tropical 20°C isotherm depth of the GODAS product.. Both patterns of evolution were. confirmed by satellite altimeter sea surface height anomaly data. The thermocline depth variation of El Niño Modoki (CP-El Niño) events displayed an east-west seesaw oscillation in the equatorial Pacific.. In addition to the east-west oscillation,. the EP-El Niño also displayed an extra north-south seesaw oscillation between approximately 15N° and 15°S.. This unique feature only appeared during an EP-El. Niño and was not recorded in earlier El Niño studies.. We demonstrated that the. wind stress curl over the western and central Pacific was closely linked to the north-south seesaw oscillation of the equatorial thermocline depth during an EP-El Niño. This argument does not contradict the standard recharge-discharge paradigm (Jin, 1997a, b). Because the definition of an ENSO event is based on the zonal (east-west) redistribution of warm water along the equator, the recharge and discharge of warm water are not necessarily associated with the local wind stress curl variability. The wind stress curl fluctuation triggers distinct thermocline depth variations within the tropical Pacific Ocean, modulating the evolution of the SST evolution in the two types of El Niño. Thus, EP-El Niño events, with both east-west and north-south seesaw oscillations are not only produced by basin-wide thermocline variations, but are also influenced by the wind stress curl in the equatorial Pacific. EP-El Niño also involves strong air-sea interactions.. Hence, the. The meridional seesaw. oscillation of thermocline variation is a major factor in inducing the distinct evolution patterns of the two types of El Niño. Thermocline depth anomalies in the equatorial Pacific were stronger in 1997/98 19.

(32) than 1982/83, which resulted in distinct differences during the decay phases between the two EP-El Niño events, and subsequently affected the SST variations and the evolution of the following La Niña event.. The state of the thermocline in the. equatorial Pacific Ocean can be attributed to the rate of the ITCZ migration. The 1997/98 El Niño with a rapidly retreating ITCZ abruptly terminated the warm event and immediately became a La Niña event, whereas the slow retreat of the ITCZ in 1982/83 terminated the warm event at a much slower rate resulting in a return to a normal phase.. 20.

(33) Figure 2.1 (a) Mean surface circulation (averaged from 0 to 50 m) in the equatorial Pacific based on GODAS model assimilation (Units: m/s).. Velocity composites. during the mature phase of (b) CP-El Niño events and (c) EP-El Niño events. Shading indicates the current intensity. Contour interval is 0.05 m/s.. 21.

(34) Figure 2.2 Modeled 20°C isotherm depth anomalies together with the graphs for (a) September 1994, (b) December 1994, (c) March 1995, and (d) June 1995.. Contour. interval for the D20 isotherm depth is 10 m. (e) Schematic diagram for Figure 2.2b and (f) schematic diagram for Figure 2.2c.. 22.

(35) Figure 2.3 Sea surface height anomalies from AVISO in (a) September 1994, (b) December 1994, (c) March 1995, and (d) June 1995. surface height anomalies is 5 cm.. 23. Contour interval for sea.

(36) Figure 2.4 Same as Figure 2.2 except in (a) September 1997, (b) December 1997, (c) March 1998, and (d) June 1998, with respective schematic diagrams in (e) to (h).. 24.

(37) Figure 2.5 Same as Figure 2.3 except in (a) September 1997, (b) December 1997, (c) March 1998, and (d) June 1998.. 25.

(38) Figure 2.6 (a) Wind stress curl averaged over the southwestern Pacific (150~180°E, 5~7°S) during the mature phase of an El Niño between 1980 and 2010. Wind stress curl pattern composited for the mature phase of (b) the EP-El Niño events, and (c) the CP-El Niño events.. Contour interval for wind stress curl is 0.5 × 10-7 Nm-3.. 26.

(39) Figure 2.7 The modeled 20°C isotherm depth anomalies in (a) March 1983 and (b) June 1983. Contour interval is 10 m.. 27.

(40) Figure 2.8 Wind stress curl (shading) and wind stress (vector) averaged from May to July in (a) 1983 and (b) 1998. Negative contours are shaded. Vector scale is 0.1 m/s. The red rectangle encloses the region from 1°N–6°S, 170°E–130°W in the central equatorial Pacific.. 28.

(41) Chapter 3 Contrasting the flow patterns in the equatorial Pacific between two types of El Niño An edited version of this paper was available online: Wang, L.-C. and C.-R. Wu (2012), Contrasting the flow patterns in the equatorial Pacific. between. two. types. of. El. Niño. Atmosphere-Ocean,. doi:10.1080/07055900.2012.744294.. 3.0 Abstract Outputs based on NCEP Global Ocean Data Assimilation System (GODAS) are adopted to contrast the current variations in the equatorial Pacific between two types of El Niño. The model fully resolves the equatorial currents.. We found that CP-El. Niño (the central Pacific El Niño) corresponds well with the previous El Niño studies that both the eastward Equatorial Undercurrent (EUC) and westward South Equatorial Current (SEC) weaken.. On the other hand, EP-El Niño (the eastern Pacific El Niño). displays a significantly distinct circulation pattern.. The North Equatorial. Counter-Current (NECC) strengthens in the developing phase and persists into the peak of the warm event, while the northern branch of the SEC (SECn) also intensifies during the mature phase and lasts for about half a year.. The South Equatorial. Counter-Current (SECC) strengthens during the decaying phase of EP-El Niño. The shifting of the wind stress curl associated with the thermocline variability is chiefly responsible for the unique current performance of EP-El Niño.. It is worth noting. that the air-sea interaction plays an important role in the current variability not only during CP-El Niño, but also during EP-El Niño.. 29.

(42) 3.1. Introduction The significance of ocean–atmosphere interaction in the tropical Pacific Ocean has been recognized and documented (e.g. Bjerknes, 1969; Philander et al., 1984). Understanding dynamics of the equatorial Pacific is essential to the global climate. The seasonal cycle is one of the conspicuous fluctuations in the region (Yu and McPhaden, 1999), resulting ultimately from the solar radiation, and coupled air-sea-land interaction (Li and Philander, 1996).. Beyond the seasonal timescale, the. climate variability demonstrates an interannual variation related to El Niño/Southern Oscillation (ENSO) (e.g. Philander, 1990).. Since the atmospheric anomalies. variability is sensitive to the variations of sea surface temperature (SST), the description of ENSO in the previous studies is often based on the SST variability, which is studied extendedly.. However, the equatorial currents and the processes that. modulate their circulation pattern are less discussed due to the lack of measurements in the region.. On interannual or longer timescales how the equatorial currents vary. is still not well understood and is worthy to further investigate the dynamics involved. The current fluctuation in the region is largely influenced by the atmospheric condition, which in turn is modulated by the ENSO phenomenon. Among the studies of the tropical Pacific circulation, there are only a few capable of providing sufficient data to describe the mean flow pattern from contemporaneous observations. For example, Yu and McPhaden (1999) described the annual cycle of zonal currents along the equator based on (Tropical Ocean-Global Atmosphere (TOGA) experiment Tropical Atmosphere-Ocean (TAO) array).. In the. western Pacific, hydrographic sections and Acoustic Doppler Current Profiler (ADCP) observations collected between 1984 and 1991 were used in a dynamical interpretation for the mean circulation (Gouriou and Toole, 1993). The termination of the Equatorial Undercurrent (EUC) has been examined with historical 30.

(43) hydrographic data in the eastern Pacific (Lukas, 1986).. Although these. measurements have shed some light on the mean circulation in the region, much less is known about the interannual variability related to ENSO. A couple of pioneering studies found out that the currents in the equatorial Pacific are considerably altered during ENSO events.. Using hydrographic sections and ADCP data, Delcroix et al.. (1992) documented current variations during 1986/87 El Niño and the subsequent La Niña. The same sections have been further examined by Johnson et al. (2000) to investigate the equatorial flow fluctuation during 1997/98 El Niño. Complex ENSO evolution feature has been recognized and studied just recently. These studies suggest that there are two different types of El Niño (Ashok et al., 2007; Kug et al., 2009; Kao and Yu, 2009; Yeh et al., 2009). Table 1.1 summarizes the characteristic and classification of the El Niño events.. In Table 1.1, El Niño Modoki. is defined according to the locations where SST anomalies occur (Ashok et al., 2007). During El Niño Modoki, SST anomalies appear associated with a horse-shoe pattern in the central Pacific.. The monthly indices and anomaly fields are seasonally. averaged over the period from June to September (December to February) as the boreal summer (winter) values. Kau and Yu (2009) pointed out that during the eastern Pacific El Niño, SST anomalies emerge from the coast of the South America, propagating westward to the central Pacific, and decay off the equator.. In the central. Pacific El Niño, SST anomalies first appear around the date line, developing and mature in a V-shaped anomaly structure extending toward the subtropics in both hemispheres, and then decay in the equatorial central Pacific.. Kug et al. (2009). found and defined CT El Niño (canonical type) with the pattern that large SST anomalies develop in the eastern Pacific during the developing summer.. The. maximum value of the warm SST is located in the coastal region of the eastern boundary. SST anomalies of WP El Niño are mostly confined in the central Pacific 31.

(44) during boreal summer.. A broad seasonal mean from September to the following. February [SOND(0)JF(1)] is taken to classify the El Niño events. In this study, we follow the definition and methodology of Kau and Yu (2009). The canonical El Niño is significant in the eastern Pacific (hereafter EP-El Niño). The new type of El Niño, which is different from the canonical El Niño in both the location of maximum SST anomalies and tropical-midlatitude teleconnections, is termed central Pacific El Niño (hereafter CP-El Niño). The two types of El Niño have their own evolution process and thermocline structure, bringing significantly distinct modulation on the ocean currents. To have a better description of the equatorial current pattern on the ENSO timescale, the analyses have to contrast current patterns of the two types of El Niño. The limited measurements in the equatorial Pacific are not capable for the longer timescale study.. In this study, we use an ocean model product to investigate the. equatorial current variability and how they are associated with the different types of El Niño.. In section 2 we describe the dataset used in this study.. In section 3,. spatial and temporal variations of assimilated equatorial currents from GODAS are presented. The distinct evolution patterns of the currents associated with CP-El Niño and EP-El Niño and their forcing mechanisms are also described.. Section 4. concludes this work.. 3.2 Data Current velocity and 20°C isotherm depth anomaly data used in this study are based on the National Centers for Environmental Prediction (NCEP) Global Ocean Data Assimilation System (GODAS) product (Behringer and Xue, 2004).. The. GODAS domain extends from 75°S to 65°N with a horizontal resolution of 1/3° × 1/3°. The model has 40 levels with 10 m resolution near the sea surface, and is 32.

(45) forced by momentum, heat, and fresh water flux from NCEP atmospheric Reanalysis 2 (R2) (Kanamitsu et al., 2002). The temperature profiles assimilated in GODAS include TOGA experiment TAO, PIRATA, TRITON moorings, XBTs, and Argo profiling floats.. In addition to the temperature, a synthetic salinity profile is. computed for each temperature profile using a local T-S climatology based on annual mean fields of temperature and salinity from the NODC World Ocean Database. Surface wind stress data are provided by the NCEP/NCAR reanalysis project since 1948.. Monthly averages on global grids (2.5°×2.5°) are available at the. Climate Diagnostics Center of NOAA’s Earth System Research Laboratory (http://www.cdc.noaa.gov/).. 3.3 Result and Discussion 3.3.1 The equatorial currents Assimilated mean circulation pattern of the Pacific equatorial current is plotted in Figure 3.1a, which is consistent with the observational mean flow pattern in the region (e.g. Lukas, 2001).. The equatorial zonal currents consist of a. westward-flowing wide-range surface current, the South Equatorial Current (SEC), which locates between about 8°S and 3°N.. Following the definition of Wyrtki. (1974), the SEC is further split into a northern and a southern branch according to the equator, hereafter the SECn and SECs.. To the north, the eastward-flowing current is. the North Equatorial Counter Current (NECC, between about 5°N and 10°N). North of the NECC, there is an intense westward current, the North Equatorial Current (NEC, between about 10°N and 20°N). The NEC is the southern limb of the North Pacific subtropical gyre and the upstream of the Kuroshio. The South Equatorial Countercurrent (SECC) extends eastward from the western boundary region, but it 33.

(46) only intermittently reaches the central and eastern Pacific. A vertical profile along the equator averaged from 2°N ~ 2°S from GODAS is shown in Figure 3.1b.. The most prominent currents here are the subsurface eastward. EUC (red shading) and the westward SEC (blue shading). equatorial current, which travels across the Pacific.. The EUC is the strongest. It originates at about 140°E,. getting stronger and reaches its maximum speed between 155°W and 125°W, weakening considerably east of 90°W.. The assimilated results from GODAS agree. well with the observational data (Yu and McPhaden, 1999; Johnson et al., 2002). Furthermore, the core of the EUC gets closer to the sea surface as flowing eastward, which feeds the equatorial upwelling. This shoaling process largely influences the SST in the eastern Pacific, where the air-sea interaction is important.. Through the. influences, the EUC plays an important role in both the seasonal and interannual fluctuations around the equatorial region (Gu and Philander, 1997) and the climate response to the global warming (Cai and Whetton, 2000).. The strong EUC prevails. over the depth up to 400 m, while the relative weaker SEC is limited in the upper 50 m of depth.. The GODAS result is quite similar to the velocity profile plotted by. Lukas (2001) based on the ocean assimilation/reanalysis product from the NOAA/NCEP. Figures 3.2a-c show the annual cycle of zonal currents on the equator at 165°E, 140°W, and 110°W, respectively.. The sections have been used previously to. describe the annual cycle of the equatorial currents (Yu and McPhaden, 1999). Following Keenlyside and Kleeman (2002), the annual cycle is described in terms of velocity anomalies (i.e., with climatology removed). The structures of the annual cycle at all three locations look quite similar.. There is an eastward intensification of. the currents, occurring between March and July, which extends from the surface to the depth of the EUC.. At 165°E, the EUC is the strongest in July, and the depth of the 34.

(47) core locates about 60~150 m (Figure 3.2a).. At 140°W, the EUC core falls upon 60. m and the fastest flow shows in May (Figure 3.2b).. At 110°W, the EUC is the. strongest during April and May and the core shoals to ~25 m (Figure 3.2c).. These. annual cycle patterns agree well with those shown in the TAO array data, except slightly discrepancy in the magnitude of the current velocity (Figures 3.2d-f, after Keenlyside and Kleeman 2002).. 3.3.2 Spatial variations associated with CP- and EP-El Niño The modulation on the equatorial currents by the two types of El Niño is investigated.. Figure 3.3 shows meridional velocity profiles along 140°W in boreal. winter, including composited normal year, CP-El Niño, and EP-El Niño.. We. averaged from December to February of the following year as the boreal winter, identified as the mature phase of El Niño by Kug et al. (2009; 2010).. In boreal. winter of normal years, eastward currents include the EUC and NECC (Figure 3.3a). The EUC locates in the thermocline from 20 m to 300 m and its core is near 140 m. It is trapped in the equatorial region between 2°S to 2°N. 7°N, has a maximum velocity near a depth of 50 m.. The NECC, centered at. Westward currents include the. SECn, SECs, and NEC. The SECn is present at ~3°N, which is the strongest in the central Pacific between 90°W and 140°W. The SECs, centered ~5°S, is relatively weaker at 140°W, but has a wider range than that of the SECn.. The NEC locates. between 10°N and 20°N, and it is very weak at 140°W. During CP-El Niño winter, the EUC is significantly weakened near its core (Figure 3.3b).. The SECn becomes weaker and narrower, probably due to the. weakening easterly trade winds on the equator (Johnson et al., 2002).. On the other. hand, the SECs becomes a little stronger than the normal year at 140°W. is also stronger during CP-El Niño than the normal year. 35. The NEC. The flow patterns during.

(48) CP-El Niño agree well with the previous El Niño studies (e.g. Yu and McPhaden, 1999; Lukas, 2001; Johnson et al., 2002; Keenlyside and Kleeman, 2002). The velocity profile during the EP-El Niño mature phase is significantly different from that during CP-El Niño (Figure 3.3c).. The westward SECn, which weakens. during CP-El Niño compared with those in the normal year, merges with SECs and strengthens at 140°W during EP-El Niño.. Probably due to the depression from the. strong SEC, the eastward EUC weakens and descends to a deeper depth of more than 600 m.. The feature is not consistent with the earlier finding that the EUC. disappeared during 1997/98 El Niño (e.g. Johnson et al., 2002).. We suggest that the. EUC exists still, but it is depressed and descends to a deeper depth due to the intensification of the SEC during EP-El Niño.. The similar pattern is seen in the. other ocean data assimilation models, such as Simple Ocean Data Assimilation (SODA) Parallel Ocean Program (POP) version 1.4.2, and Geophysical Fluid Dynamical Laboratory (GFDL) Couple Model Assimilation (figures not shown). The spatial circulation patterns in the equatorial Pacific composited by the normal, CP-El Niño, and EP-El Niño years are plotted in Figure 3.4, respectively. During the normal winter, the SECn and SECs are separated from each other obviously.. The SECn lasts until the dateline and extends westward intermittently. thereafter, while the SECs seems to be capable of reaching the western Pacific basin (Figure 3.4a).. The eastward NECC locates between 4°N and 9°N, and the. subsurface current EUC is not obvious at the depth of 50 m. However, in the mature phase of CP-El Niño, the NECC strengthens and the EUC shoals up to the sea surface west of 170°E (Figure 3.4b).. The SECn gets much weaker and is confined east of. 120°W. The SECs is blocked by the shoaling EUC and not able to reach the western boundary. The circulation during EP-El Niño has a distinctive pattern. 36. Most of the.

(49) previous studies indicate that the EUC weakens obviously as flowing eastward during El Niño.. In this study, we found that these previous findings are valid only during. CP type of El Niño. The EUC strengthens from the onset till the mature phase of EP-El Niño (Figure not shown), and is found to merge with the NECC, surging eastward. This unusual performance of the eastward currents is supposed to be a great disturbance to the westward currents. merge together and strengthen.. However, the SECn and SECs also. These two merged current systems with opposite. direction match each other in strength, and get deadlocked around 170°W.. This. unusual pattern appears only during EP-El Niño and is far from what was discussed in the earlier El Niño studies. The strengthening region of the eastward EUC during the mature phase of EP-El Niño is confined to west of the dateline, while it weakens significantly east of the dateline, where the merged SECn and SECs is exceptionally vigorous. The intense westward currents suppress the eastward EUC and NECC around the equator, which produces an unusual flow pattern during EP-El Niño (Figure 3.4c).. Moreover, it is often observed that the NEC tends to migrate. northward as flowing westward during El Niño.. Figure 3.4c shows that the latitude. of the NEC is at ~16°N when it reaches the western Pacific boundary.. The. bifurcation point of the NEC is much farther north during EP-El Niño than CP-El Niño.. 3.3.3 Temporal variations of the equatorial currents To place emphasis on the interannual variability, we remove the annual cycle signal and deal with the velocity anomalies in the subsequent analysis.. We identified. the phases of El Niño evolution according to Kug et al. (2009; 2010). The period of developing phase is from March to November, the mature phase is from December to February of the following year, and the decay phase is from February to October. 37.