聖嬰在全球暖化下之改變
No Thumbnail Available
Date
2013
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
全球暖化對聖嬰的影響看法仍相當分歧，目前尚無一致結論。本研究透過比較CMIP3模式中GFDL-CM2.1模式及UKMO-HADCM3模式對不同氣候情境模擬，探討暖化對聖嬰的影響，旨在歸納影響(1)兩類型聖嬰(即，中太平洋聖嬰與東太平洋聖嬰)出現頻率與(2)極端強聖嬰振幅的關鍵因子。研究中選擇GFDL-CM2.1模式及UKMO-HADCM3模式的原因是，兩個模式在現今氣候模擬(20C3M)對赤道東太平洋海溫的統計特徵(如平均態、變異度與偏度)皆與觀測非常相似，但在未來氣候模擬(A1B)則出現完全相反的結果。
在工業革命前控制實驗(pre-industrial control experiment，PICTL)長期模擬顯示，GFDL-CM2.1模式與UKMO-HADCM3模式皆顯示或中太平洋聖嬰(Central Pacific El Niño，簡稱CP-EN)偏多時期，赤道東太平洋平均場之斜溫層深度(或海溫、海水高度)變淺(小)，東太平洋聖嬰(Eastern-Pacific El Niño，簡稱EP-EN)偏多時期則反之。此外，研究發現平均場的計算若移除聖嬰與反聖嬰年，與不移除聖嬰年與反聖嬰年的結果相反(即，CP-EN偏多時期赤道東太平洋斜溫層深度變深)。此結果表示，聖嬰對平均場的調整作用不容忽略。
A1B模擬與PICTL實驗比較發現，平均場在赤道西－東太洋斜溫層厚度的斜率變化對兩類型聖嬰出現頻率具關鍵性影響。相較於PICTL實驗，UKMO-HADCM3模式在A1B模擬，平均場之赤道西－東太平洋斜溫層厚度之斜率明顯變小，由於赤道中太平洋斜溫層厚度變淺，有利於斜溫層－海溫回饋，CP-EN出現頻率明顯增多。然而，上述平均場變化在GFDL-CM2.1模式並不明顯，CP-EN出現頻率，因此未明顯增多。
至於暖化對極端強聖嬰振幅的影響，GFDL-CM2.1在A1B模擬顯著減弱，但UKMO-HADGEM1模式反而增強。GFDL-CM2.1極端強聖嬰振幅減弱，歸納原因為：(1)風-海溫回饋機制，(2)斜溫層-海溫回饋機制與(3)IOD-ENSO回饋機制的減弱。UKMO-HADGEM1模式，極端強聖嬰振幅在A1B模擬增加則歸納與赤道大西洋海溫變冷有關。
The change of ENSO (El Niño and Southern Oscillation) in future climate projection is still an open question. The change of the frequency of two types El Niños, the central Pacific El Nino (CP-EN), and eastern Pacific El Nino (EP-EN) and the amplitude of El Niño in future climate under the global warming was investigated by diagnosing the CMIP3 simulations. The GFDL-CM2.1 and UKMO-HADCM3 simulations are particularly selected for diagnosis because that both models successfully capture the behaviors of El Niño (e.g., SST variance and skewness) in present simulation (20C3M run), but shows a widely distinction in the future climate (A1B run). In the pre-industry control experiment (PICTL), both models showing consistent results, reveals that that the occurrence of two type El Niños tightly related with the mean state of the depth of thermocline (i.e., D20, depth at 20oC, which is substantially modified by the ENSO events, in the equatorial eastern Pacific (EEP). The frequency of CP-EN increases substantially during the period when the mean state of D20 in EEP deepens, and vice versa. The change of the frequency of the two type El Niños in future climate (A1B) exhibits a characteristic of model dependent: the frequency of CP-EN increases significantly in UKMO-HADCM3, but does not have essential change in GFDL-CM2.1, which the D20 in EEP does not shows remarkable deepening in A1B. As the amplitude of El Nino, the GFDL-CM2.1 shows significant decrease in A1B, but the El Niño amplitude remains unchanged and even shows a bit stronger in A1B than that in 20C3M. It shows that the change of the El Niño amplitude is primarily dominated by the following factors: 1) wind-evaporation-SST feedback, 2) thermocline-SST feedback, 3) IOD (Indian Ocean Diploe)-ENSO feedback, and 4) TA(tropical Atlantic)-ENSO feedback. The decrease of El Niño amplitude in GFDL-CM2.1 is due to the weakening of Bjerkness feedback (factor 1 and 2), and the weakening of IOD-ENSO feedback. Conversely, the enactment of El Niño amplitude in UKMO-HADCM3 is due to the enhancement of TA-ENSO feedback.
The change of ENSO (El Niño and Southern Oscillation) in future climate projection is still an open question. The change of the frequency of two types El Niños, the central Pacific El Nino (CP-EN), and eastern Pacific El Nino (EP-EN) and the amplitude of El Niño in future climate under the global warming was investigated by diagnosing the CMIP3 simulations. The GFDL-CM2.1 and UKMO-HADCM3 simulations are particularly selected for diagnosis because that both models successfully capture the behaviors of El Niño (e.g., SST variance and skewness) in present simulation (20C3M run), but shows a widely distinction in the future climate (A1B run). In the pre-industry control experiment (PICTL), both models showing consistent results, reveals that that the occurrence of two type El Niños tightly related with the mean state of the depth of thermocline (i.e., D20, depth at 20oC, which is substantially modified by the ENSO events, in the equatorial eastern Pacific (EEP). The frequency of CP-EN increases substantially during the period when the mean state of D20 in EEP deepens, and vice versa. The change of the frequency of the two type El Niños in future climate (A1B) exhibits a characteristic of model dependent: the frequency of CP-EN increases significantly in UKMO-HADCM3, but does not have essential change in GFDL-CM2.1, which the D20 in EEP does not shows remarkable deepening in A1B. As the amplitude of El Nino, the GFDL-CM2.1 shows significant decrease in A1B, but the El Niño amplitude remains unchanged and even shows a bit stronger in A1B than that in 20C3M. It shows that the change of the El Niño amplitude is primarily dominated by the following factors: 1) wind-evaporation-SST feedback, 2) thermocline-SST feedback, 3) IOD (Indian Ocean Diploe)-ENSO feedback, and 4) TA(tropical Atlantic)-ENSO feedback. The decrease of El Niño amplitude in GFDL-CM2.1 is due to the weakening of Bjerkness feedback (factor 1 and 2), and the weakening of IOD-ENSO feedback. Conversely, the enactment of El Niño amplitude in UKMO-HADCM3 is due to the enhancement of TA-ENSO feedback.
Description
Keywords
平均場, 兩類型聖嬰, 全球暖化, 極端強聖嬰振幅, Two types El Niño, Bjerkness feedback, IOD-ENSO feedback, TA-ENSO feedback