Contrasting Madden Julian Oscillation activity during various stages of EP and CP El Niños

Transcription

1 ATMOSPHERIC SCIENCE LETTERS Atmos. Sci. Let. 16: (2015) Published online 4 July 2014 in Wiley Online Library (wileyonlinelibrary.com) DOI: /asl2.516 Contrasting Madden Julian Oscillation activity during various stages of EP and CP El Niños Juan Feng, 1 * Ping Liu, 2 Wen Chen 1 and Xiaocong Wang 3 1 Center for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China 2 School of Marine and Atmospheric Sciences, Stony Brook University, NY, USA 3 LASG, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China *Correspondence to: Dr J. Feng, Institute of Atmospheric Physics, Chinese Academy of Sciences, P.O. Box 2718, Beijing , China. juanfeng@mail.iap.ac.cn Received: 13 June 2013 Accepted: 4 June 2014 Abstract This study revisits the relations between Madden Julian Oscillation (MJO) and El Niño that is classified as eastern Pacific (EP) and central Pacific (CP) types. During an EP-type event that produces lasting dry conditions in the Maritime Continent and western Pacific, the MJO activity is weak especially in its 4th 6th phases as a strong response. In contrast during CP-type episodes, the MJO exhibits notably enhanced and suppressed amplitudes in its different phases. The low-level westerly wind anomalies of enhanced MJO impact the CP-type El Niño more than vice versa. Keywords: MJO; EP El Niño; CP El Niño 1. Introduction The Madden Julian Oscillation (MJO) is a dominant mode of the tropical intraseasonal variability, with the coupling of atmospheric circulation and moist convection propagating eastward across the Indian Ocean, Maritime Continent and the western Pacific at a speed of 5 m s 1 (Hendon and Salby, 1994; Zhang, 2005). El Niño-Southern Oscillation (ENSO) is another significant large-scale air-sea coupled phenomenon in the tropical Pacific Ocean at the interannual timescales (McPhaden and Zhang, 2009; among others). Both ENSO and MJO are involved in the shifts of the large-scale convection and atmospheric circulation in the tropics. The connection between the two modes, however, is complicated and ambiguous (Hendon et al., 1999, 2007; Slingo et al., 1999; Tam and Lau, 2005). Slingo et al. (1999) and Hendon et al. (1999) reported that there is no obvious correlation between the year-to-year MJO amplitude and ENSO. Hendon et al. (2007) found out that the MJO ENSO relation has a seasonal dependent feature in which the spring enhanced MJO precedes autumn winter El Niño by several months. Nevertheless, interaction may exist between MJO and ENSO when the respective physical processes are considered. On the one hand, pronounced MJO events produce westerly wind passage in the western Pacific, a key factor for the onset and development of ENSO (Harrison and Schopf, 1984; Kessler et al., 1995; Belamari et al., 2003; Wang et al., 2013). On the other hand, ENSO may impact the evolution of the MJO in space and time through changing the background convection in the tropical Pacific (Hendon et al., 1998; Takayabu et al., 1999; Tam and Lau, 2005; Pohl and Matthews, 2007; Roundy and Kravitz, 2009). Such interactions may appear more intrinsic as El Niño demonstrates distinct shifts in spatial structure during past decades. Before the 1990s El Niño events showed a dominant dipole pattern (Yeh et al., 2009), with a warmer sea surface temperature anomalies (SSTA) pole in the eastern Pacific and a colder pole in the western Pacific. These traditional events are classified as eastern Pacific (EP) type. Some El Niño events have a tri-pole pattern with warmer SSTA in the central Pacific and colder ones in the eastern and western Pacific; these events are newly named as central Pacific (CP) El Niño (Ashok et al., 2007; Kao and Yu, 2009; Kug et al., 2009). The CP-type event tends to occur more frequently than the EP-type. McPhaden et al. (2011) disclosed that three of the four warm ENSO events are CP-type during the last decade ( ), whereas only two of the six belong to this type in Such pattern changes are reproduced partially by the coupled climate models (Yeh et al., 2009; Kim and Yu, 2012). The relationship between El Niño and MJO was discussed in many studies while the results are mixing (Hendon et al., 1999; Slingoet al., 1999; Hendon et al., 2007). It becomes intriguing whether the complicated relationship arises from the different flavors of El Niño or a more robust relation can be established when the El Niño is classified into EP and CP types. Gushchina and Dewitte (2012) documented the interaction of the MJO with the two types of El Niño. They found that the MJO is enhanced during the developing phase of the EP El Niño. In contrast the enhanced MJO occurs not only during the developing phase of CP El Niño but also during the mature and decaying phases. In this study, we focus on the event-to-event variability of MJO in the two types of El Niño years. Meanwhile, we further investigate the MJO characteristics during 2014 Royal Meteorological Society

2 MJO activity during EP and CP El Niño 33 Figure 1. Intensity of each MJO event (solid circles) and the evolution of the weekly mean Niño4 (left column) and Niño3 (right column) indices (black lines), respectively. The left (right) column is for the CP (EP) El Niño. Vertical lines indicate the dates of the onset and termination of El Niño. The x-axis denotes the date of the year. its various phases and possible relation with SSTA evolution associated with the two types of El Niño events. 2. Data and methods The weekly mean Optimum Interpolation (OI) sea surface temperature (SST) version 2, produced by the National Oceanic and Atmospheric Administration (NOAA), is used in this study (Reynolds et al., 2002). This SST dataset is blended by in situ and satellite sources, covering the periods from November 1981 to present as a horizontal resolution of 1.0 by 1.0 in latitude and longitude. The CP- and EP-type El Niños are distinguished using the Niño4 and Niño3 indices defined as the SSTA averaged over the regions of (5 N-5 S, 160 E-150 W) and (5 S-5 N, W), respectively (Kug et al., 2009; Yeh et al., 2009). A CP El Niño is identified when the winter-mean (December-January-February, DJF) Niño4 index is greater than Niño3 index and 0.5 C (Yeh et al., 2009), while an EP El Niño has the Niño3 index greater than Niño4 index and 0.5 C

3 34 J. Feng et al. (a) Event-to-event amp. (b) Event-to-event amp. Lead/lag days Lead/lag days Figure 2. (a) Scatterplots of each MJO amplitude with the lead/lag days relative to the respective reference day that is defined by the onset time of each EP El Niño. Sliding averaged event-to-event MJO amplitude with a 101-day window (blue line) and the composite Niño3 index (green line). (b) Same as (a) but for the CP El Niño. Blue, purple, red and pink circles indicate the MJO happens during the pre-developing, developing, mature and decaying stages of El Niño. The thick black lines denote the averaged MJO amplitude for one El Niño stages. Thick solid black lines indicate confidence level above 90% according to a two-tailed Student s t-test. (Yeh et al., 2009). This approach has identified six CP-type El Niño events in 1990/1991, 1994/1995, 2002/2003, 2004/2005, 2006/2007 and 2009/2010 and four EP-type events in 1982/1983, 1986/1987/1988, 1991/1992 and 1997/1998 (Figure 1). The onset dates for the CP (EP) El Niño are defined as when the Niño4 (Niño3) index changes from negative to positive values; and termination dates are similarly defined by which the Niño4 (Niño3) index changes from positive to negative values (Figure 1, vertical lines). However, for the CP-type events in 1990/1991, 2002/2003 and 2004/2005, there is no clear intersection with the zero line; the onset and termination dates are then defined by the lowest value of the Niño4 index (Figure 1, vertical lines). In addition, we examined the SSTA evolutions for these three CP-type events to accurately locate the transition point. We further divided the lifetime of CPand EP-type El Niños into four stages: a pre-developing stage as the days before the onset date, a developing stage as the days from the onset date to when the Niño 4 (Niño3) index is smaller than 1 C, mature stage as the days when the Niño4 or Niño3 index is greater than 1 C, and decaying stage as the days when the Niño4 or Niño3 index is smaller than 1 C. The threshold is increased to 2 C for the 1982/1983 and 1997/1998 EP El Niño, which is about half of the maximum Niño3 index. Sliding averaged amp. Sliding averaged amp. The day-to-day MJO activity is described by the Real-time Multivariate MJO (RMM) index (Wheeler and Hendon, 2004). This index is calculated from the leading two combined empirical orthogonal functions (EOFs) of OLR and zonal winds at 850 and 200-hPa. The leading pair of principal component (PC) time series forms the index as RMM1 and RMM2 which defines eight MJO phases (Figure 7 in Wheeler and Hendon, 2004). Using the RMM index, an MJO event is selected when it has an obvious and continuous eastward propagation from the 1st to the 8th phases (Pohl and Matthews, 2007) and the date of an event is represented by the middle day of an MJO event. The square root of (RMM12 + RMM22) represents the daily amplitude of an MJO event while the MJO intensity is the averaged amplitude during its lifetime (Figure 1, circles). An MJO is considered strong and active when its intensity is greater than 1.4. This value is obtained from the averaged intensity of all selected MJO events in the EP- and CP-type El Niño years. It is defined as 1.5 in Garfinkel et al. (2012) and Yoo et al. (2012), because these studies focus on the MJO only in the winter season. The slightly smaller threshold in our analysis is reasonable. 3. Characteristics of the MJO activity during the EP and CP El Niño events Figure 2 gives the scatterplots of the event-to-event MJO amplitudes with the lead/lag days relative to the onset date of the El Niño events. Also shown in Figure 2 are the composited Niño3 index for the four EP El Niño events and Niño4 index for the six CP El Niño events (green lines), which represents the SSTA evolutions of each type. The CP El Niño has slow evolutions from developing, mature to decaying phases, whereas the EP El Niño undergoes fast developing and decaying stages but a long-lasting mature episode. These different characteristics of EP and CP El Niños are associated with different event-to-event MJO activity and with distinct MJO-El Niño relationship. During the EP El Niño period, strong MJOs with intensity greater than 1.4 mainly occur in the predeveloping (about days) and developing stages (about days) (Figure 2(a), blue and purple circles). However, similar number of weak MJOs also occur during the pre-developing and developing stages. Meanwhile, the averaged amplitudes of the event-to-event MJOs for these two stages (thick black lines) are close to the threshold 1.4, hence the anomalous MJO activity is insignificant. As an EP El Niño reaches its mature stage (about days), most MJOs are weak with only a couple of exceptions (Figure 2(a), red circles). During the rapidly decaying stage of the EP El Niño, few MJOs occur (Figures 1 (left column) and 2(a)). The averaged amplitudes of MJOs for the mature and decaying stages are only close to the threshold 1.4, indicating the weak variability of MJO activity.

4 MJO activity during EP and CP El Niño 35 Pre-developing stage Pre-developing stage Developing stage Developing-early stage Developing-late stage Mature stage Mature stage Decaying stage Decaying stage Figure 3. Scatterplots of the event-to-event MJO amplitude with the different MJO phases during the (a, e) pre-developing, (b, f, g) developing, (c, h) mature and (d, i) decaying EP and CP El Niño periods. The solid black line is the averaged MJO amplitude in different MJO phases. Hollow and Solid circles indicate that the averaged MJO amplitude is above 80 and 90% confidence level according to the two-tailed Student s t-test, respectively. In contrast, strong MJOs with the intensity greater than 1.4 distribute in each stage of the CP El Niño (Figure 2(b), solid circles). This result is consistent with Gushchina and Dewitte (2012). Interestingly, during the developing phase (about days), MJOs show different intensity between the early and late portion (Figure 2(b), blue line). From 0 to 80 days, the MJO is as strong as that in the pre-developing period; while from 80 to 160 days, it becomes substantially suppressed. This feature is much more evident that in Gushchina and Dewitte (2012). During the mature and decaying stages, the MJO activities are enhanced. Statistically, the anomalous MJO activity is above 80% confident level for the developing-early, mature and decaying stages (Figure 2(b), thick dashed black lines) and above 90% confident level for the pre-developing and developing-late stages (thick solid black lines). These results demonstrate that the anomalous MJO is more active during the CP than during the EP El Niño. The MJO are notably different at near 0 days during the EP- and CP-type El Niños when the MJO may affect the onset and growth of warm SSTA in the tropical Pacific. Specifically, the MJO is basically inactive during the EP-type El Niño (Figure 2(a)), while it is

5 36 J. Feng et al. substantially enhanced during the CP-type episodes (Figure 2(b)). Such distinctive result suggests that the MJO activity impacts the onset of CP-type El Niño, but not of the EP El Niño. This is partially inconsistent with Gushchina and Dewitte (2012) in which the MJO is enhanced during the onset period of both the EP and CP El Niños. Such difference is possibly caused by the different data and methods employed. In addition, Hendon et al. (2007) found that there is a clear relationship between the MJO and ENSO, and suggested that enhanced MJO precedes mature El Niño by several months. Our results indicate that this relationship is more prominent during the CP than during the EP El Niño years. Furthermore, Tang and Yu (2008) pointed out that this relationship between MJO and ENSO is nonlinear in nature and experiences a decadal variation: it is stronger in the 1990s than in the 1980s, which may be contributed by the decadal variation of El Niño too. The CP El Nino occurs more frequently in the 1990s than in the 1980s (Yeh et al., 2009), likely in favor of the enhanced MJO-El Niño relationship as well. Figure 3 gives the MJO characteristics in its various phases during different El Niño stages. The MJO is slightly enhanced in the 6th 8th phases, especially in the 7th phase during the pre-developing stage of EP-type El Niño. These phases correspond to the low-level westerly wind anomalies propagating from the tropical western Pacific to the central Pacific, favoring the onset of the EP-type El Niño. During the developing, mature and decaying stages of the EP El Niño, the MJO activity is nearly suppressed in all eight phases and weakest in the 4th 6th phases (Figure 3(b) (d)). In contrast, the MJO demonstrates large variability in its eight phases during the pre-developing, developing-early, mature and decaying stages of the CP-type El Niño (Figure 3(e) (i)). Specifically, the MJO is strong in its 3rd, 4th and 5th phases during pre-developing stage of CP El Niño; these phases correspond to enhanced westerly wind anomalies propagating from the tropical Indian Ocean to the western Pacific. As such, the wind anomalies favor the formation of the CP El Niño. Later during the developing-early stage of the CP El Niño, the MJO amplitude becomes strong in the 3rd 8th phases, whereas during the developing-late stage of CP El Niño, the MJO amplitude shows an evident opposite features. During the mature stage of the CP El Niño, the MJO is significantly strong in the 3rd 5th phases. Notably during the decaying stage of the CP El Niño, the MJO is strong almost in its every phase although it is below a meaningful statistical significance level. We next explore why the MJO features are distinctly different during the two types of El Niños, recognizing that the convective activity is different over the Maritime Continent and western Pacific a key region to the MJO development and eastward propagation. Clearly the convective anomalies over the key region are significantly suppressed during the EP El Niño nearly from the developing to decaying stages (Figure 4). This long-lasting dry condition acts as a Figure 4. Composite daily OLR anomalies during the (a) EP and (b) CP El Niño events. The y-axis denotes the lead/lag day relative to the respective reference day that is defined by the onset time of each EP and CP El Niño. barrier to the MJO development and eastward propagation (Figure 4(a)), producing suppressed MJO in the 4th 6th phases (c.f. Figure 3) and similarly weak MJO activity (Figures 3(b) (d)). The results in Figures 3 and 4 indicate that the EP-type El Niño dictates the weak MJO activity and thus plays a dominant role in the MJO ENSO relationship. This relationship changes during the CP El Niño stages. The convective anomalies are also suppressed over the Maritime Continent and western Pacific but much weaker than those during the EP El Niño stages (Figure 4). This difference can be caused by the weaker intensity and more westward SSTA location of the CP El Niño (c.f. Figures 1 and 2). Such weakly suppressed convection allows the MJO to develop and propagate eastward (c.f. Figures 3(e) (i)), suggesting a weaker modulation of the CP El Niño to the MJO. Conversely, the large MJO variability may impact the CP El Niño in two ways: the suppressed MJO (Figure 3(g)) provides less low-level westerly wind anomalies to strengthen the SSTA in the central Pacific (Figures 1 and 2), resulting in overall weaker CP El Niño; while the strong MJO slows down the decaying speed of the warm SSTA in the tropical Pacific (c.f. Figure 2(b)). 4. Conclusions and discussions This study contrasts the MJO characteristics on an event-to-event basis during the CP and EP El Niño years using observational and reanalysis data. Results show that the MJO is remarkably different during the two El Niño types: it is generally suppressed especially in the 4th 6th phases by the EP-type El Niño during the developing to the decaying stages.

6 MJO activity during EP and CP El Niño 37 The descending branch of the EP-type Walker Circulation produces severely suppressed convection as a long-lasting background over the Maritime Continent and western Pacific, hindering the MJO development. Thus the EP-type El Niño plays a dominant role in the MJO ENSO relationship. Contrastingly, the MJO has a large variability during the CP-type El Niños: its phases can be either suppressed or enhanced significantly. Stronger MJOs are observed mainly during the pre-developing, mature and decaying phases of the CP El Niño. While during the developing phase of the CP-type, the MJO is strong in the earlier portion but extremely weak in the late stage. Specifically, the MJO is strong in the 3rd 5th phases during the pre-developing and mature stages of the CP El Niño, while it is strong in all eight phases during the developing-early and decaying stages. With its varying low-level westerly wind anomalies, the MJO appears to play a more important role in relation to the ENSO during these events. Seasonal cycle does not seem to affect the relationship disclosed above although it tends to be phase-locking to both ENSO and MJO. This is supported by a similar MJO ENSO relationship in the composites of monthly MJO amplitude anomalies after the climatological seasonal mean is removed during the EP and CP El Niño years (not shown). As most of the six CP-type El Niños occurred after year 2000 and all the EP-type events were before, the changes of the MJO ENSO relation may be originated from the decadal variation of MJO activity. This possibility is under exploration and can benefit from the larger samples from future climate models which can successfully produce both the MJO and two types of El Niños at the same time. Acknowledgements The authors are thankful to the constructive comments from the anonymous reviewers. This study is supported by the National Natural Science Foundation of China Grant ( and ). References Ashok K, Behera SK, Rao SA, Weng HY, Yamagata T El Niño Modoki and its possible teleconnection. Journal of Geophysical Research 112: C11019, DOI: /2006JC Belamari S, Redelsperger JL, Pontaud M Dynamic role of a westerly wind burst in triggering an equatorial Pacific warm event. Journal of Climate 16: Garfinkel CI, Feldstein SB, Waugh DW, Yoo C, Lee S Observed connection between stratospheric sudden warmings and the Madden-Julian Oscillation. Geophysical Research Letters 39: L18807, DOI: /2012GL Gushchina D, Dewitte B Intraseasonal tropical atmospheric variability associated to the two flavors of El Niño. Monthly Weather Review 140: Harrison D, Schopf PS Kelvin-wave-induced anomalous advection and the onset of surface warming in El Niño events. Monthly Weather Review 112: Hendon HH, Salby ML The life cycle of the Madden-Julian oscillation. Journal of the Atmospheric Sciences 51: Hendon HH, Liebmann B, Glick JD Oceanic Kelvin waves and the Madden-Julian oscillation. Journal of the Atmospheric Sciences 55: Hendon HH, Zhang C, Glick JD Interannual variation of the Madden-Julian oscillation during austral summer. Journal of Climate 12: Hendon HH, Wheeler MC, Zhang C Seasonal dependence of the MJO-ENSO relationship. Journal of Climate 20: Kao HY, Yu JY Contrasting eastern-pacific and central-pacific types of ENSO. Journal of Climate 22: Kessler WS, McPhaden MJ, Weickmann KM Forcing of intraseasonal Kelvin waves in the equatorial Pacific. Journal of Geophysical Research 100: Kim ST, Yu JY The two types of ENSO in CMIP5 models. Geophysical Research Letters 39: L Kug JS, Jin FF, An SI Two types of El Niño events: cold tongue El Niño and warm pool El Niño. Journal of Climate 22: McPhaden MJ, Zhang X Asymmetry in zonal phase propagation of ENSO sea surface temperature anomalies. Geophysical Research Letters 36: L13703, DOI: /2009GL McPhaden M, Lee T, McClurg D El Niño and its relationship to changing background conditions in the tropical Pacific Ocean. Geophysical Research Letters 38: L Pohl B, Matthews AJ Observed changes in the lifetime and amplitude of the Madden-Julian oscillation associated with interannual ENSO sea surface temperature anomalies. Journal of Climate 20: Reynolds RW, Rayner NA, Smith TM, Stokes DC, Wang W An improved in situ and satellite SST analysis for climate. Journal of Climate 15: Roundy PE, Kravitz JR The association of the evolution of intraseasonal oscillations to ENSO phase. Journal of Climate 22: Slingo JM, Rowel DP, Sperber KR On the predictability of the interannual behavior of the Madden-Julian Oscillation and its relationship with El Nino. Quarterly Journal of the Royal Meteorological Society 125: Takayabu YN, Iguchi T, Kachi M, Shibata A, Kanzawa H Abrupt termination of the El Nino in response to a Madden-Julian oscillation. Nature 402: Tam CY, Lau NC Modulation of the Madden-Julian oscillation by ENSO: inferences from observations and GCM simulations. Journal of the Meteorological Society of Japan 83: Tang Y, Yu B MJO and its relationship to ENSO. Journal of Geophysical Research 113: D14106, DOI: /2007JD Wang X, Wang CZ, Zhou W, Liu L, Wang DX Remote influence of North Atlantic SST on the equatorial westerly wind anomalies in the western Pacific for initiating an El Niño event: an Atmospheric General Circulation Model Study. Atmospheric Science Letters 14: Wheeler MC, Hendon HH An all-season real-time multivariate MJO index: development of an index for monitoring and prediction. Monthly Weather Review 132: Yeh SW, Kug JS, Dewitte B, Kwon MH, Kirtman BP, Jin FF El Niño in a changing climate. Nature 461: Yoo C, Lee S, Feldstein SB Mechanisms of arctic surface air temperature change in response to the Madden-Julian Oscillation. Journal of Climate 25: Zhang C Madden-Julian oscillation. Reviews of Geophysics 43: RG2003, DOI: /2004RG