Abstract

The initiation and developing mechanisms of four major central Pacific (CP) El Niño events in 1994, 2002, 2004, and 2009 were investigated by analyzing oceanic and atmospheric reanalysis data. A mixed layer heat budget analysis was conducted and the result shows that the initiation mechanism of the 1994 CP El Niño is very different from other CP El Niños in 2000s, while the developing mechanisms are similar among these events. The initial sea surface temperature (SST) warming of the 1994 El Niño was caused by enhanced solar radiation, which was related to atmospheric meridional overturning circulation in association with positive SST anomaly forcing in the subtropical Pacific. The subtropical SST anomalies also induced anticyclonic surface wind stress curl anomalies, which caused the formation of subsurface warmer waters in the off-equatorial regions. The off-equatorial subsurface warmer waters were transported farther equatorward by the mean subsurface ocean currents, leading to the subsurface warming in the central equatorial Pacific. The deepened thermocline anomaly at the equator further promoted a positive advective and thermocline feedback so that the SST anomaly grew. During the initiation phase of the 2000s El Niños, ocean dynamics played a dominant role, while the effect of surface heat flux anomalies was minor. Preexisting subsurface warmer waters appeared in the equatorial region during their initiation phases. Such subsurface anomalies can cause the SST warming in the central Pacific through induced anomalous eastward zonal currents that advect high mean SST eastward. This positive zonal advective feedback, along with a positive thermocline feedback, continued to warm the local SST throughout the developing phase of the 2000s El Niño events.

1. Introduction

The El Niño–Southern Oscillation (ENSO) is the most pronounced interannual variability in the tropics, and has far-reaching climatic impacts in many regions over the world. Based on the spatial pattern of the sea surface temperature anomalies (SSTAs), the El Niño events are classified into two types: one is the canonical eastern Pacific (EP) El Niño with maximum SSTA centered in the eastern equatorial Pacific and another is the central Pacific (CP) El Niño with maximum SSTA centered in the central equatorial Pacific (Fu et al. 1986; Trenberth and Stepaniak 2001; Ashok et al. 2007; Kao and Yu 2009; Kug et al. 2009). The atmospheric responses associated with the two types of El Niños exhibit significant differences. For example, the westerly anomalies during CP El Niño events have a smaller spatial scale and are located in the western Pacific, farther westward compared to those in the EP El Niño. Positive precipitation anomalies appear in the central-eastern equatorial Pacific during the EP El Niño, but are located mainly in the western Pacific during the CP El Niño (Kao and Yu 2009; Kug et al. 2009; Chung and Li 2013).

Compared with the generally accepted classic theories about the generation mechanism of the EP El Niño, there are debates concerning the initiation dynamics of the CP El Niño. The generation of the CP El Niño is attributed to wind-forced thermocline variations (Ashok et al. 2007), the zonal advection of mean SST by anomalous zonal currents (Kug et al. 2009), or subtropical SSTA in the northeastern Pacific (Yu et al. 2010). The decadal mean climate state tends to play an important role in favoring more frequent occurrence of the CP El Niño during the recent decade after 1999 (Xiang et al. 2013; Chung and Li 2013).

Furthermore, there are different developing features and climatic impacts among the CP El Niño events. For example, Wang and Wang (2013) classified the CP El Niño events into two groups with different SST evolution characteristics. The complex evolution features of the CP El Niño imply that more than one mechanism may operate during the CP El Niño initiation.

In the present study, we will investigate the different initiation mechanisms for the CP El Niño events. The rest of this paper is organized as follows. In section 2, the data and analysis methods are described. The initiation and developing mechanisms of four major CP El Niños since 1980 are investigated in section 3. A summary and discussions are given in section 4.

2. Data and methods

The ocean data used in this study are the National Centers for Environmental Prediction (NCEP) Global Ocean Data Assimilation System (GODAS; Saha et al. 2006), the European Centre for Medium-Range Weather Forecasts (ECMWF) Ocean Reanalysis System, version 4 (ORAS4; Balmaseda et al. 2013), and the Simple Ocean Data Assimilation (SODA) reanalysis, version 2.1.6 (Carton and Giese 2008). The GODAS has a constant zonal resolution of 1° and a variable meridional grid of 1° enhanced to ⅓° within 10° of the equator, and has 40 levels with a 10-m resolution in the upper 200 m. The ORAS4 has a horizontal resolution of 1° × 1° and 42 levels in the vertical. The SODA 2.1.6 has an average horizontal resolution of 0.4° (longitude) × 0.25° (latitude) and 40 levels of eddy-permitting resolution with 10-m spacing near the surface.

The atmospheric data are from the 40-yr ECMWF Re-Analysis (ERA-40) data (Uppala et al. 2005). The SST data are from the Met Office Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) with a resolution of 1.0° × 1.0° (Rayner et al. 2003). The precipitation data are Global Precipitation Climatology Project (GPCP), version 2.2 with a grid resolution of 2.5° × 2.5° (Adler et al. 2003). The outgoing longwave radiation (OLR) data are from the National Oceanic and Atmospheric Administration (NOAA) interpolated OLR with a resolution of 2.5° × 2.5° (Liebmann and Smith 1996). The surface heat flux data are from the Woods Hole Oceanographic Institution (WHOI) objectively analyzed air–sea fluxes (OAFlux; Yu et al. 2008), NCEP reanalysis version 2 (NCEP2; Kanamitsu et al. 2002), and the twentieth-century reanalysis, version 2 (20CRv2).

To understand the relative roles of ocean advection and surface heat flux terms in causing the SSTA tendencies, the oceanic mixed layer heat budget is diagnosed. The mixed layer temperature (MLT) tendency equation may be written as

 
formula

where represents the 3D ocean current, denotes the 3D gradient operator, a prime () represents the anomaly variables, a bar () represents the climatologic annual cycle variables, is the sum of linear advection terms, denotes 3D nonlinear temperature advection terms, represents the net heat flux at the ocean surface, represents the residual term, represents the density of water, represents the specific heat of water, and denotes the climatologic mixed layer depth. All of the mixed layer fields are calculated based on the layer average. The climatologic annual cycle is calculated based on the period of 1980–2000. The heat budget is calculated based on the ensemble mean of three ocean reanalysis datasets (GODAS, ORAS4, and SODA 2.1.6) and three heat flux datasets (OAFlux, NCEP2, and 20CRv2), and only statistically significant terms will be discussed in the following. To investigate the specific cause of ocean current anomalies, wind-induced Ekman currents and geostrophic currents are diagnosed respectively in an equatorial β-plane framework (see Su et al. 2010 for details). Note that the Ekman currents discussed here differ from traditional Ekman currents in midlatitudes, and strictly speaking they should be called “wind-driven ocean surface currents” in an equatorial β plane.

3. Initiation and developing dynamics of four major CP El Niños

a. Evolution features of CP El Niños

Four major CP El Niño events (1994, 2002, 2004, and 2009) after 1980 were identified (Fig. 1). A common feature of the four CP El Niño events is that the maximum SSTA were all confined in the central equatorial Pacific (around 180°–130°W) during their major developing period and they reached a mature phase in DJF (Fig. 1). Xiang et al. (2013) and Chung and Li (2013) have a detailed discussion on why the four events are classified as “pure” CP El Niño events and how they differ from other mixed events.

Fig. 1.

The evolution of SSTA along the equator (within 5°N–5°S) for four major CP El Niño events. Contour intervals for SSTA are 0.25°C. The SST data are from OISST. The initiation phases of the El Niño cases are indicated by gray shading.

Fig. 1.

The evolution of SSTA along the equator (within 5°N–5°S) for four major CP El Niño events. Contour intervals for SSTA are 0.25°C. The SST data are from OISST. The initiation phases of the El Niño cases are indicated by gray shading.

To investigate the SSTA evolution process during the four CP El Niño events, a CP region (5°N–5°S, 180°–130°W) is defined. Figure 2 illustrates the temporal evolution characteristics of the SSTA averaged in the CP region for the four CP El Niño cases. Because we are interested in how each of the CP El Niños was initiated and what contributed to their development, we separate the El Niño evolution processes into two periods: an initiation phase and a developing phase. The initiation phase is defined as the period when the SSTA in the CP region are close to zero but the SSTA tendency is positive. The reason to focus on this phase is that we intend to investigate what causes the initial warming tendency while the SST anomaly itself is nearly normal. The gray shaded period represents the initiation phase for each of the CP El Niño events (i.e., February–April 1994, November 2001–January 2002, March–May 2004, April–June 2009). The developing phase follows the initiation phase and ends when the SSTA reach a maximum. During this phase warm SSTAs have already set up, and there are positive feedbacks between the ocean and atmosphere. The periods of the developing phase for the four events are May–November 1994, February–November 2002, June–November 2004, and July–November 2009. Before selecting the initiation and developing phases, the SSTA time series are smoothed with a 3-month running mean to remove high-frequency signals. The initiation and developing phases in the observed SSTA are well captured by the MLT anomalies obtained from the GODAS, ORAS4, and SODA 2.1.6 ocean data, which adds confidence in using those ocean reanalysis data for the heat budget diagnosis. Because the SODA 2.1.6 data cover the period of 1958–2008, only the GODAS and ORAS4 data are used for the analysis of the 2009 case.

Fig. 2.

Time series of SSTA in the CP region (5°N–5°S, 180°–130°W). The SST data are from OISST. The temperature anomalies in the mixed layer of GODAS, ORAS4, and SODA 2.1.6 are also shown. The initiation phases for each of CP El Niño cases are indicated by gray shading.

Fig. 2.

Time series of SSTA in the CP region (5°N–5°S, 180°–130°W). The SST data are from OISST. The temperature anomalies in the mixed layer of GODAS, ORAS4, and SODA 2.1.6 are also shown. The initiation phases for each of CP El Niño cases are indicated by gray shading.

An examination of subsurface temperature anomalies reveals that there are apparent discrepancies among the four CP El Niño events (Fig. 3). During the initiation of the 2002, 2004, and 2009 El Niños, positive subsurface temperature anomalies appeared in the western–central equatorial Pacific. However, there was no subsurface warming during the initiation of the 1994 El Niño. In fact, during that time anomalous cold subsurface waters occupied the whole basin, expanding from the western Pacific warm pool to the eastern boundary. Positive subsurface temperature anomalies started to appear in the central equatorial Pacific around June 1994, when the local SSTA already become positive. The significant difference of the subsurface temperature anomalies indicates that the 1994 El Niño is a unique event compared with other CP El Niños. Hence, the initiation mechanism for the 1994 El Niño may be different from other CP El Niños. In the following, we separate the CP El Niño events into two groups: the 1994 El Niño and 2002, 2004, and 2009 El Niños (hereafter the 2000s El Niño), and investigate the potential different mechanisms of the two groups of CP El Niños.

Fig. 3.

Mean temperature anomalies along the equator (within 5°N–5°S) during the initiation phase for each of CP El Niño events. Contours intervals are 0.5°C. The temperature data are from GODAS.

Fig. 3.

Mean temperature anomalies along the equator (within 5°N–5°S) during the initiation phase for each of CP El Niño events. Contours intervals are 0.5°C. The temperature data are from GODAS.

b. Initiation mechanism of the 1994 El Niño

The SST anomalies in the central equatorial Pacific became positive around March 1994. However, the local subsurface temperature anomalies were still negative at that period (Fig. 3). This suggests that the sources of the initial SST warming of the 1994 El Niño should not arise from the local subsurface but from other factors. Our heat budget analysis indicates that the major cause of the SST warming during the initiation phase of the 1994 El Niño is the surface heat flux anomaly (0.10°C month−1), which is contributed mainly by the enhanced surface shortwave radiation (0.11°C month−1; see Table 1). The enhanced shortwave heat fluxes in the CP region during the initiation phase were associated with local anomalous descending motion (omega) at 850 hPa (0.01 Pa s−1), positive OLR anomalies (10 W m−2), and suppressed precipitation (15 mm day−1) in the CP region (Fig. 4).

Fig. 4.

The time series of downward shortwave heat flux anomalies (red line; unit: W m−2), downward latent heat flux anomalies (blue line; unit: W m−2), anomalous downward vertical motion at 850 hPa (green line; unit: 10−3 Pa s−1), anomalous OLR (magenta line; unit: W m−2), and precipitation anomalies (black line; unit: mm day−1) averaged in the CP region (5°N–5°S, 180°–130°W) for the 1994 El Niño event. The heat flux fields are derived from OAFlux. The OLR values are from NOAA data. The vertical motion anomalies are from ERA-40. The precipitation anomalies are from GPCP. The initiation phase (February–April) for the 1994 El Niño is indicated by gray shading.

Fig. 4.

The time series of downward shortwave heat flux anomalies (red line; unit: W m−2), downward latent heat flux anomalies (blue line; unit: W m−2), anomalous downward vertical motion at 850 hPa (green line; unit: 10−3 Pa s−1), anomalous OLR (magenta line; unit: W m−2), and precipitation anomalies (black line; unit: mm day−1) averaged in the CP region (5°N–5°S, 180°–130°W) for the 1994 El Niño event. The heat flux fields are derived from OAFlux. The OLR values are from NOAA data. The vertical motion anomalies are from ERA-40. The precipitation anomalies are from GPCP. The initiation phase (February–April) for the 1994 El Niño is indicated by gray shading.

Table 1.

The mixed layer temperature tendency terms averaged in the CP region (5°N–5°S, 180°–130°W) for (a) the initiation phase and (b) the developing phase of the 1994 El Niño and the 2000s El Niños. The units are °C month−1. The values are the ensemble mean of results from three ocean reanalysis datasets (GODAS, ORAS4, and SODA 2.1.6) and three surface heat flux datasets (OAFlux, NCEP2, and 20CRv2). The terms that have a ratio of mean to standard deviation greater than 1.3 are set bold.

The mixed layer temperature tendency terms averaged in the CP region (5°N–5°S, 180°–130°W) for (a) the initiation phase and (b) the developing phase of the 1994 El Niño and the 2000s El Niños. The units are °C month−1. The values are the ensemble mean of results from three ocean reanalysis datasets (GODAS, ORAS4, and SODA 2.1.6) and three surface heat flux datasets (OAFlux, NCEP2, and 20CRv2). The terms that have a ratio of mean to standard deviation greater than 1.3 are set bold.
The mixed layer temperature tendency terms averaged in the CP region (5°N–5°S, 180°–130°W) for (a) the initiation phase and (b) the developing phase of the 1994 El Niño and the 2000s El Niños. The units are °C month−1. The values are the ensemble mean of results from three ocean reanalysis datasets (GODAS, ORAS4, and SODA 2.1.6) and three surface heat flux datasets (OAFlux, NCEP2, and 20CRv2). The terms that have a ratio of mean to standard deviation greater than 1.3 are set bold.

The descending anomalies in the equatorial region were connected to anomalous ascending motions in the subtropical region, which can be seen both in the divergent wind field at 850 hPa (Fig. 5a) and in the meridional-vertical section zonally averaged in the central Pacific (Fig. 5b). In the northern subtropics, there were marked warm SSTAs, and the anomalous ascending motions located over the center of the warm SSTA in the northeastern subtropical Pacific. In the southern subtropical Pacific (about 10°–20°S, 180°–130°W), there were also anomalous ascending motions over the local warm SSTA there. The surface wind anomalies converged toward the center of the warm SSTA in the northern and southern subtropics. In fact, these positive SSTAs in the subtropical regions persisted from the previous winter (November 1993–January 1994), and so were the anomalous ascending (descending) motions (not shown). Thus, the observational analysis indicates that the warm SSTAs in the northern (southern) subtropical Pacific induced anomalous ascending motions locally, which further caused the descending anomalies near the equator through the change of atmospheric meridional vertical overturning circulation.

Fig. 5.

(a) The anomalous velocity potential (contours, with intervals of 0.25 × 10−6 m2 s−1) and divergent wind (vectors, in m s−1) at 850 hPa and (b) meridional and vertical velocity anomalies along the meridional section averaged over the region of 180°–130°W during the initiation phase (February–April) of the 1994 El Niño. The anomalous SSTA (shading, with intervals of 0.25°C) are shown in (a). The contour interval for vertical velocity anomalies is 0.25 × 10−2 Pa s−1 in (b).

Fig. 5.

(a) The anomalous velocity potential (contours, with intervals of 0.25 × 10−6 m2 s−1) and divergent wind (vectors, in m s−1) at 850 hPa and (b) meridional and vertical velocity anomalies along the meridional section averaged over the region of 180°–130°W during the initiation phase (February–April) of the 1994 El Niño. The anomalous SSTA (shading, with intervals of 0.25°C) are shown in (a). The contour interval for vertical velocity anomalies is 0.25 × 10−2 Pa s−1 in (b).

The potential role of the subtropical SSTA on the onset of theCP El Niño has been discussed in previous studies (e.g., Yu et al. 2010). Yu et al. (2010) emphasized that the initial warming in the CP region is a result of the southward propagation of warm SSTAs from the subtropics by a mechanism similar to the seasonal footprinting mechanism (Vimont et al. 2003). However, our analyses show that at least for the 1994 event, the positive SSTAs in the subtropics did not move, rather they made a remote impact on the SST warming in the central equatorial Pacific through the atmospheric overturning circulation change. Yu et al. (2010) mentioned that the southward SSTA movement depends on weakened latent heat fluxes on the equatorward side of the subtropical warm SSTA. However, the changes of latent heat fluxes in the CP region were minor during the initiation phase of the 1994 El Niño (Fig. 4). The major contribution of the initial SST warming in the central Pacific for the 1994 El Niño came from the enhanced solar radiation (Table 1).

c. The developing mechanism of the 1994 El Niño

Once the SSTAs were initiated, they can further grow through various positive air–sea feedback processes (Li 1997). The heat budget analysis results (Table 1) show that the major contributor for the SST warming during the developing phase of the 1994 El Niño was the zonal advection feedback term (; about 0.20°C month−1). The thermocline feedback term (; about 0.03°C month−1) also made a positive contribution for the developing SSTA. Note that during the developing phase there were pronounced eastward zonal ocean current anomalies () across the whole equatorial Pacific basin (Fig. 6b). Since the zonal climatologic SST gradient is negative (), the zonal advective feedback is positive () in the CP region. To investigate what causes the anomalous zonal ocean current, we diagnosed anomalous geostrophic and wind-induced Ekman currents in an equatorial β-plane framework, following Su et al. (2010). The result shows that the anomalous eastward zonal currents are largely caused by the geostrophic component, while the wind-induced Ekman currents are very weak (Fig. 7). Note that the geostrophic current is negatively correlated with the meridional second derivative of Z20 (20°C isothermal) anomalies . A local maximum of positive Z20 anomalies on the equator results in an anomalous eastward geostrophic current ().

Fig. 6.

Mean mixed layer (0–50 m) current anomalies (vectors), temperature anomalies (magenta contour, with intervals of 1°C), and 20°C isotherm depth anomalies (shading, with intervals of 2.5 m) during (top) the initiation phase and (bottom) the developing phase of (left) the 1994 El Niño and (right) the 2000s El Niños. The climatological mean temperatures are indicated by gray heavy contours with intervals of 3°C. The composite fields are the ensemble mean of GODAS, ORAS4, and SODA 2.1.6.

Fig. 6.

Mean mixed layer (0–50 m) current anomalies (vectors), temperature anomalies (magenta contour, with intervals of 1°C), and 20°C isotherm depth anomalies (shading, with intervals of 2.5 m) during (top) the initiation phase and (bottom) the developing phase of (left) the 1994 El Niño and (right) the 2000s El Niños. The climatological mean temperatures are indicated by gray heavy contours with intervals of 3°C. The composite fields are the ensemble mean of GODAS, ORAS4, and SODA 2.1.6.

Fig. 7.

Anomalies of zonal currents from the ocean reanalysis data (ensemble mean of GODAS and SODA 2.1.6, solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (averaged within 2°N–2°S) during the developing phase of each El Niño.

Fig. 7.

Anomalies of zonal currents from the ocean reanalysis data (ensemble mean of GODAS and SODA 2.1.6, solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (averaged within 2°N–2°S) during the developing phase of each El Niño.

The positive Z20 anomalies in the central Pacific can also lead a positive thermocline feedback term. This is because a positive Z20 anomaly causes a warmer subsurface temperature anomaly, which can further warm the surface water through anomalous temperature advection by the mean upwelling (). However, such a positive thermocline feedback () is weaker in the central Pacific than in the eastern Pacific.

As the zonal advective feedback is crucial for the development of the 1994 El Niño, we further examined the temporal evolution of the anomalous zonal ocean current in the CP region. The time evolution of the area-averaged anomalous zonal ocean current obtained from the aforementioned three ocean assimilation datasets, an independent ocean current reanalysis dataset of the Ocean Surface Current Analyses Real-time (OSCAR), and their ensemble mean were plotted in Fig. 8. The anomalous zonal current averaged in the CP region is negative (westward) at the beginning of 1994. After the initiation phase, the anomalous zonal current changes to positive (eastward), and reaches its maximum at the mature phase of the 1994 El Niño. The positive zonal current anomalies started to develop rapidly in May 1994, which coincided well with the occurrence of positive Z20 anomalies at the equator. Thus, a key point is what caused the development of positive equatorial Z20 anomalies during the initial developing phase (around May).

Fig. 8.

Time series of zonal velocity anomalies from GODAS (red line), ORAS4 (green line), SODA 2.1.6 (blue line), OSCAR reanalysis dataset (magenta line), and their ensemble mean (thick black dash line) in the CP region (5°N–5°S, 180°–130°W). The developing phase (May–November) of the 1994 El Niño is indicated by gray shading.

Fig. 8.

Time series of zonal velocity anomalies from GODAS (red line), ORAS4 (green line), SODA 2.1.6 (blue line), OSCAR reanalysis dataset (magenta line), and their ensemble mean (thick black dash line) in the CP region (5°N–5°S, 180°–130°W). The developing phase (May–November) of the 1994 El Niño is indicated by gray shading.

d. Origins of the equatorial subsurface warming of the 1994 El Niño

As mentioned previously, the temperature anomalies in the subsurface in the equatorial Pacific were negative at the beginning of 1994, and the subsurface waters became warmer after the initiation phase. A question that needs to be addressed is what are the origins of the later subsurface warming?

During the initiation phase, the subsurface temperature anomalies in the western equatorial Pacific were negative (Figs. 3 and 6a). Hence, the local Z20 deepening in the central equatorial Pacific was not originated from the western equatorial Pacific. It is noted that the local Z20 deepening at the central equatorial Pacific came from off-equatorial subsurface warming (around 5°–10°N and 5°–10°S; Fig. 6a). In the meridional-depth section averaged in the central Pacific (Figs. 9a–c), there is clear evidence that the subsurface temperature anomalies in the equatorial region changed gradually from negative values (February–April) to positive values (May–July). During the February–April period, the anomalous subsurface warming located mainly in the off-equatorial region (south of 6°S and north of 6°N). The off-equatorial subsurface warm waters gradually moved toward the equator in May–July due to the advection of the mean subsurface ocean currents (vectors in Figs. 9a–c). During August–October, the subsurface warming reached a maximum value (about 1.0°C) near the equator, while the off-equatorial warming tended to fade. From the time–latitude section of temperature anomalies along the layers between the base of mixed layer (50 m) and the isopycnal surface (Fig. 9d), the equatorward migration of temperature anomalies from 5°–10°N to 5°–10°S can be seen explicitly. The movement of the off-equatorial warm waters to the equatorial region was caused by the mean meridional ocean circulation associated with the subtropical cell (STC) in the tropical Pacific. The STC is characterized by poleward currents at the surface and equatorward currents at the subsurface (around 100-m depth; e.g., Capotondi et al. 2005). Because of the asymmetric structure of STC on both sides of the equator, the equatorward transportation of off-equatorial subsurface warm waters mainly came from the Southern Hemisphere. The mean equatorward velocity of the subsurface waters near the central equatorial Pacific region is about 0.04 m s−1, or about 100 km month−1. Hence, it took 3–5 months for the off-equatorial subsurface waters to be transported to the equatorial region.

Fig. 9.

(a)–(c) Composite ocean temperature anomalies (contour, with intervals of 0.2°C) and climatologic mean meridional ocean currents (vectors) averaged in the central Pacific (180°–130°W) during (a) February–April, (b) May–July, and (c) August–October for the 1994 El Niño. (d) The time–latitude section of temperature anomalies along the layers between the base of mixed layer (50 m) and the isopycnal surface averaged over (180°–130°W). The time is from January 1993 to December 1994. The annual mean depth of isopycnal surface is indicated by magenta lines in (a)–(c). The magenta arrows indicate the equatorward movement of the temperature anomalies. The composite fields are from the ensemble mean of GODAS, ORAS4, and SODA 2.1.6.

Fig. 9.

(a)–(c) Composite ocean temperature anomalies (contour, with intervals of 0.2°C) and climatologic mean meridional ocean currents (vectors) averaged in the central Pacific (180°–130°W) during (a) February–April, (b) May–July, and (c) August–October for the 1994 El Niño. (d) The time–latitude section of temperature anomalies along the layers between the base of mixed layer (50 m) and the isopycnal surface averaged over (180°–130°W). The time is from January 1993 to December 1994. The annual mean depth of isopycnal surface is indicated by magenta lines in (a)–(c). The magenta arrows indicate the equatorward movement of the temperature anomalies. The composite fields are from the ensemble mean of GODAS, ORAS4, and SODA 2.1.6.

The causes of the off-equatorial subsurface warming were attributed to anomalous surface wind stress forcing. In response to the subtropical warm SSTA in the North (South) Hemisphere during the initiation phase, the wind stress anomalies formed an anticyclonic wind stress curl at both side of the equator (blue shaded areas in Fig. 10). The anticyclonic wind curl tends to deepen the local thermocline depth, leading to a positive subsurface warming in the off-equatorial regions.

Fig. 10.

The surface wind stress anomalies (vectors, in 1 × 10−9 N m−2), wind stress curl (shading, with intervals of 1 × 10−8 Pa m−1), and SSTA (contours) during the initiation phase (February–April) of the 1994 El Niño. The positive (negative) SSTA are shown as red (blue) contours (at intervals of 0.25°C). Blue (red) shading denotes anticyclonic (cyclonic) wind stress curl anomalies.

Fig. 10.

The surface wind stress anomalies (vectors, in 1 × 10−9 N m−2), wind stress curl (shading, with intervals of 1 × 10−8 Pa m−1), and SSTA (contours) during the initiation phase (February–April) of the 1994 El Niño. The positive (negative) SSTA are shown as red (blue) contours (at intervals of 0.25°C). Blue (red) shading denotes anticyclonic (cyclonic) wind stress curl anomalies.

e. Initiation and developing mechanisms for the 2000s El Niños

The diagnosis of the mixed layer heat budget analysis for the 2000s El Niños shows that the SSTA warming during the initiation phase in the CP region was primarily caused by ocean dynamics, rather than surface heat flux anomalies. Table 1 shows that the ocean temperature advection term dominated during the initiation phase (0.20°C month−1). Among the advection terms, major contributions were the zonal advection feedback term (0.13°C month−1) and the thermocline feedback term (0.04°C month−1) (Table 1). The diagnosis results from individual CP El Niño events in the 2000s resemble the composite average shown in Table 1.

During the initiation phase of the 2000s El Niño, there were preexisting positive subsurface temperature anomalies in the central equatorial Pacific (Fig. 3), and the composite Z20 anomaly had a maximum near the equator (Fig. 6c). The deepened Z20 anomalies at the equator could cause anomalous eastward geostrophic currents (figure not shown) and promote a positive zonal ocean advection feedback. Meanwhile the warmer subsurface waters could be upwelled to the surface to warm the SST through a positive thermocline feedback.

What induced the initial subsurface warming in the equatorial Pacific prior to the initiation phase? It is noted that a few months prior to the initiation phases of the 2000s El Niños the ocean thermocline exhibited a positive anomaly in the western equatorial Pacific. Associated with the positive thermocline anomaly were the equatorial easterly anomalies and/or anticyclonic wind stress curl anomalies in the off-equatorial western Pacific (figure not shown). A further study is needed to understand the origin of these precursor wind anomalies.

During the developing phase of the 2000s El Niños, the zonal advection feedback and the thermocline feedback continued to make positive contributions to the SST warming in the CP region (Table 1). The diagnosis of the zonal current anomaly in an equatorial β-plane framework shows that the anomalous positive zonal ocean currents are largely contributed by the geostrophic current during the developing phase of the 2000s El Niños (Fig. 7), a similar mechanism operated during the developing phase of the 1994 El Niño.

After the positive SST anomalies in the central Pacific were induced, the equatorial subsurface warming was enhanced through the positive Bjerknes feedback during the developing phase of the 2000s El Niños. As a result, the anomalous positive zonal ocean currents increased gradually from the initiation phase to the developing phase (Figs. 6c,d). Hence, the positive zonal advective feedback became larger from the initiation phase to the developing phase. The results obtained here in general agree with previous studies, such as the wind-forced thermocline variations (Ashok et al. 2007) and the zonal advection of the mean SST by anomalous zonal currents (Kug et al. 2009). In fact, these two mechanisms operated together during the evolution of the 2000s CP El Niño events.

Hence, the observational analysis above indicates that the initiation mechanism of the 1994 El Niño is different from that of the 2000s El Niños. The reason for such a difference lies on the distinctive precursor patterns of subsurface temperature anomalies in the equatorial region and SSTA patterns in the subtropical regions. During the initiation phase of the 1994 El Niño, there were negative subsurface temperature anomalies at the equator, while there were positive subsurface temperature anomalies during the initiation phase of the 2000s El Niños (Figs. 3 and 6a,c). It is the subtropical warm SSTA that initiated the initial warming in the CP in 1994 through the change of anomalous atmospheric meridional overturning circulation and so-induced shortwave radiation anomalies and through the generation and advection of the anomalous subsurface warm waters.

4. Summary and discussion

By analyzing the oceanic and atmospheric reanalysis data, the initiation and developing mechanisms of the CP El Niños in 1994, 2002, 2004, and 2009 were investigated. It is found that the initiation mechanism of the 1994 El Niño is very different from the 2000s El Niños, while the developing mechanisms bear some similarity.

Figure 11 is a schematic diagram illustrating the major initiation/developing processes associated with the two CP El Niño groups. The initiation of the 1994 El Niño was preceded by positive SSTA in the subtropical Pacific, which triggered anomalous convection in the subtropics and led to anomalous descending motion on the equator through the atmospheric meridional overturning circulation change. Associated with the anomalous descending, the downward solar radiation was enhanced in the equatorial central Pacific, inducing the initial SSTA warming there. Meanwhile, the subtropical SSTA forced anomalous surface winds in such a way that they induced anticyclonic surface wind stress curl at both sides of the equator and led to warm subsurface temperature anomalies in the off-equatorial regions. The off-equatorial subsurface warm waters were further transported equatorward by the mean equatorward currents associated with the lower branch of the STC, causing the subsurface warming right on the equator.

Fig. 11.

Schematic for the initiation/developing mechanisms of two central Pacific El Niño groups.

Fig. 11.

Schematic for the initiation/developing mechanisms of two central Pacific El Niño groups.

During the initiation phase of the 2000s El Niños, the subsurface warm waters have already appeared in the equatorial region. The anomalous deepened Z20 brought out the anomalous eastward zonal geostrophic currents, warming the SST through a positive zonal advection anomaly. Meanwhile, the anomalous warm subsurface waters were upwelled to the surface to warm the local SST. Both the processes were responsible for the initial warming during the 2010 El Niño events.

Ramesh and Murtugudde (2013) claimed that all flavors of El Niños have similar early subsurface origins. However, the 1994 El Niño event is excluded in their study. On the other hand, many previous studies (see Table 1 of Xiang et al. 2013) classified the 1994 El Niño into the CP El Niño. The analysis here clearly shows that the 1994 El Niño is a unique CP El Niño due to the precursor negative subsurface temperature in the equatorial region during the initiation phase. From this respect, the 1994 El Niño can be viewed as a sensitive experiment in the nature climate system. During the initiation of this unique CP El Niño, there was no influence from previous subsurface warming signals, and the warming arose from shortwave radiation forcing in association with the subtropical SST anomalies.

Concerning the impacts of the subtropical SST anomalies on the formation of CP El Niño, Yu et al. (2010) emphasized the equatorward moving of the SST anomalies from the subtropics. Our analysis shows that such a subtropical SSTA forcing scenario did not operate in 1994. The impact of the SSTA in the subtropics was primarily through the change of the atmospheric meridional overturning circulation. It is worth mentioning that there were no significant positive SSTAs in the subtropical Pacific during the initiation phase of the 2000s CP El Niños, which raises a question about whether or not Yu et al.’s subtropical SSTA propagating mechanism is valid.

Classical El Niño theories suggested that the origin of the subsurface temperature anomalies at the equator arose from off-equatorial subsurface temperature anomalies through the westward propagation and reflection of the equatorial Rossby waves (e.g., Suarez and Schopf 1988) or zonal mean meridional Sverdrup upper-ocean mass transport (Li 1997; Jin 1997). This study suggests an alternative way, that is, the off-equatorial subsurface warm waters can be transported to the equatorial region by climatologic mean subsurface currents. This subsurface advection mechanism is consistent with a previous study by Zhang and Rothstein (2000), who noticed that the off-equatorial subsurface anomalies could be transported equatorward along isopycnal layers during the initiation of the 1991/92 El Niño.

One may wonder, given similar precursor thermocline signals between 2000s CP El Niños and strong EP El Niños such as those in 1982 and 1997, what is the essential cause of distinctive subsequent development between the two types of El Niños? We argue that it is primarily attributed to markedly different anomalous surface wind–SST spatial phase relationships between the two types of El Niños. For the CP El Niños, the zonal wind and precipitation anomalies were located to the west of a SSTA center, whereas for the 1982 and 1997 EP El Niños, they were approximately in phase with the SSTA center (Xiang et al. 2013). As a result, the former led to local growth of the SSTA in the CP because the maximum thermocline anomaly is approximately in phase with the SSTA, and the latter led to the eastward propagation of the maximum SSTA because a positive thermocline anomaly appeared to the east of the SSTA center. The cause of the distinctive zonal wind–SST phase relationships is, to a large extent, attributed to the decadal change of the mean state, in particular, the decadal change of the background zonal SST gradient, which was demonstrated by Chung and Li (2013) in idealized atmospheric and oceanic model experiments. The 1994 CP El Niños may be regarded as a special case in which subtropical SSTA forcing had a maximum vertical motion and shortwave radiation response in the central equatorial Pacific.

The observational analysis in this study suggests that there were different evolution features of the recorded El Niño events and that the initiation mechanisms for individual El Niños could be different. For a particular El Niño type, its formation may be due to a single process or the combination of several processes. Thus, it is necessary to investigate the initiation/developing mechanisms for each of the El Niños. The results derived from the current analysis have important implication for seasonal climate prediction. It has been shown by many previous studies (e.g., Ashok et al. 2007) that seasonal rainfall anomalies over the Asian monsoon region have distinctive characteristics between the CP and EP El Niños. Tropical cyclone activity is also markedly different in the western Pacific (Chen and Tam 2010; Hong et al. 2011; Chung and Li 2014). Thus, it is crucial to predict the type of El Niños a few months in advance, in order to accurately forecast regional climate variations.

Acknowledgments

This work was done when JS visited IPRC. JS and RZ were supported by the National Natural Science Foundation of China (under Grants 41221064 and 41376020), the International S&T Cooperation Project of the Ministry of Science and Technology of China under Grant 2009DFA21430, and the key program of 2012Z001 in the Chinese Academy of Meteorological Sciences. TL acknowledged the support of ONR Grant N00014-1210450 and NSF Grant AGS-1106536. GPCP Precipitation data were provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from website (http://www.esrl.noaa.gov/psd/).

REFERENCES

REFERENCES
Adler
,
R. F.
, and Coauthors
,
2003
:
The Version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present)
.
J. Hydrometeor.
,
4
,
1147
1167
, doi:.
Ashok
,
K.
,
S. K.
Behera
,
S. A.
Rao
,
H.
Weng
, and
T.
Yamagata
,
2007
:
El Niño Modoki and its possible teleconnection
.
J. Geophys. Res.
,
112
, C11007, doi:.
Balmaseda
,
M. A.
,
K.
Mogensen
, and
A. T.
Weaver
,
2013
:
Evaluation of the ECMWF ocean reanalysis system ORAS4
.
Quart. J. Roy. Meteor. Soc.
, 139, 1132–1161, doi:.
Capotondi
,
A.
,
M. A.
Alexander
,
C.
Deser
, and
M. J.
McPhaden
,
2005
:
Anatomy and decadal evolution of the Pacific subtropical–tropical cells (STCs)
.
J. Climate
,
18
,
3739
3758
, doi:.
Carton
,
J. A.
, and
B. S.
Giese
,
2008
:
A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA)
.
Mon. Wea. Rev.
,
136
,
2999
3017
, doi:.
Chen
,
G.
, and
C. Y.
Tam
,
2010
:
Different impacts of two kinds of Pacific Ocean warming on tropical cyclone frequency over the western North Pacific
.
Geophys. Res. Lett.
,
37
, L01803, doi:.
Chung
,
P.-H.
, and
T.
Li
,
2013
:
Interdecadal relationship between the mean state and El Niño types
.
J. Climate
,
26
,
361
379
, doi:.
Chung
,
P.-H.
, and
T.
Li
,
2014
:
Characteristics of tropical cyclone genesis in the western North Pacific during the developing and decaying phases of two types of El Niño
.
J. Trop. Meteor.
, in press.
Fu
,
C.
,
H. F.
Diaz
, and
J. O.
Fletcher
,
1986
:
Characteristics of the response of sea surface temperature in the central Pacific associated with warm episodes of the Southern Oscillation
.
Mon. Wea. Rev.
,
114
,
1716
1739
, doi:.
Hong
,
C.-C.
,
Y.-H.
Li
,
T.
Li
, and
M.-Y.
Lee
,
2011
:
Impacts of central Pacific and eastern Pacific El Niños on tropical cyclone tracks over the western North Pacific
.
Geophys. Res. Lett.
,
38
,
L16712
, doi:.
Jin
,
F.-F.
,
1997
:
An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model
.
J. Atmos. Sci.
,
54
,
811
829
, doi:.
Kanamitsu
,
M.
,
W.
Ebisuzaki
,
J.
Woollen
,
S.-K.
Yang
,
J. J.
Hnilo
,
M.
Fiorino
, and
G. L.
Potter
,
2002
:
NCEP–DOE AMIP-II Reanalysis (R-2)
.
Bull. Amer. Meteor. Soc.
,
83
,
1631
1643
, doi:.
Kao
,
H.-Y.
, and
J.-Y.
Yu
,
2009
:
Contrasting eastern Pacific and central Pacific types of ENSO
.
J. Climate
,
22
,
615
632
, doi:.
Kug
,
J.-S.
,
F.-F.
Jin
, and
S.-I.
An
,
2009
:
Two types of El Niño events: Cold tongue El Niño and warm pool El Niño
.
J. Climate
,
22
,
1499
1515
, doi:.
Li
,
T.
,
1997
:
Phase transition of the El Niño–Southern Oscillation: A stationary SST mode
.
J. Atmos. Sci.
,
54
,
2872
2887
, doi:.
Liebmann
,
B.
, and
C. A.
Smith
,
1996
:
Description of a complete (interpolated) outgoing longwave radiation dataset
.
Bull. Amer. Meteor. Soc.
,
77
,
1275
1277
.
Ramesh
,
N.
, and
R.
Murtugudde
,
2013
:
All flavours of El Niño have similar early subsurface origins
.
Nat. Climate Change
,
3
,
42
46
, doi:.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
, doi:.
Saha
,
S.
, and Coauthors
,
2006
:
The NCEP Climate Forecast System
.
J. Climate
,
19
,
3483
3517
, doi:.
Su
,
J.
,
R.
Zhang
,
T.
Li
,
X.
Rong
,
J.-S.
Kug
, and
C.-C.
Hong
,
2010
:
Causes of the El Niño and La Niña amplitude asymmetry in the equatorial eastern Pacific
.
J. Climate
,
23
,
605
617
, doi:.
Suarez
,
M. J.
, and
P. S.
Schopf
,
1988
:
A delayed action oscillator for ENSO
.
J. Atmos. Sci.
,
45
,
3283
3287
, doi:.
Trenberth
,
K. E.
, and
D. P.
Stepaniak
,
2001
:
Indices of El Niño evolution
.
J. Climate
,
14
,
1697
1701
, doi:.
Uppala
,
S. M.
, and Coauthors
,
2005
:
The ERA-40 Re-Analysis
.
Quart. J. Roy. Meteor. Soc.
,
131
,
2961
3012
, doi:.
Vimont
,
D. J.
,
J. M.
Wallace
, and
D. S.
Battisti
,
2003
:
The seasonal footprinting mechanism in the Pacific: Implications for ENSO
.
J. Climate
,
16
,
2668
2675
, doi:.
Wang
,
C.
, and
X.
Wang
,
2013
:
Classifying El Niño Modoki I and II by different impacts on rainfall in the southern China and typhoon tracks
.
J. Climate
,
26
,
1322
1338
, doi:.
Xiang
,
B.
,
B.
Wang
, and
T.
Li
,
2013
:
A new paradigm for the predominance of standing central Pacific warming after the late 1990s
.
Climate Dyn.
, 41, 327–340, doi:.
Yu
,
J.-Y.
,
H.-Y.
Kao
, and
T.
Lee
,
2010
:
Subtropics-related interannual sea surface temperature variability in the central equatorial Pacific
.
J. Climate
,
23
,
2869
2884
, doi:.
Yu
,
L.
,
X.
Jin
, and
R. A.
Weller
,
2008
: Multidecade global flux datasets from the Objectively Analyzed Air-sea Fluxes (OAFlux) Project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. WHOI Rep. OA-2008-1, Woods Hole Oceanographic Institution, 64 pp.
Zhang
,
R. H.
, and
L. M.
Rothstein
,
2000
:
Role of off-equatorial subsurface anomalies in initiating the 1991-1992 El Niño as revealed by the NCEP ocean reanalysis data
.
J. Geophys. Res.
,
105
,
6327
6339
, doi:.