Abstract

This study investigates the impact of western central Pacific sea surface temperatures (SSTs) on the temperature variability over the Korea Peninsula during the early boreal summer season. The authors found a significant positive correlation between June temperature anomalies across the Korea Peninsula and a horseshoe pattern of ENSO-related SST anomalies in the western central Pacific during May. While this SST pattern is relatively stationary throughout the boreal summer, the SST-driven atmospheric teleconnection exhibits dramatic subseasonal changes. During May, positive precipitation anomalies caused by SST warming in the northern central Pacific (NCP; 20°–30°N, 160°E–160°W) induce a wave train extending in a northwest–southeast direction. During June, while this wave train pattern is weakened, a Pacific–Japan (PJ) pattern is formed by positive precipitation anomalies over the off-equatorial western Pacific (off-WP). Even though dramatic subseasonal changes exist in the teleconnection patterns, both patterns contribute to temperature warming over Korea. It is shown that the subseasonal change in the atmospheric teleconnection is caused by changes in climatological precipitation. During May, significant climatological precipitation related to the pre-mei-yu and pre-baiu fronts occurs over the NCP. These wet background conditions provide a favorable environment for anomalous convection over the NCP but not over the off-equatorial western Pacific. On the other hand, during June, convective activities over the off-WP result in the formation of a PJ pattern, enhanced by an increase in climatological precipitation associated with the activation of the western North Pacific (WNP) monsoon. These results indicate that climatological conditions play an important role in the formation of atmospheric teleconnection linking Pacific and East Asian climates.

1. Introduction

It has been widely recognized that the intensity of tropical convective activity can influence midlatitude climate through various atmospheric teleconnection patterns. One of the most well-documented tropics–midlatitude teleconnection patterns during the boreal summer is the western North Pacific anticyclone (WNPAC), excited by western Pacific convective anomalies during the El Niño decaying summer (Wang et al. 2000; Li and Wang 2005; Wu et al. 2010). This WNPAC is also considered part of a meridional wave train pattern along the western Pacific, also called the Pacific–Japan (PJ) pattern (Nitta 1987; Kosaka and Nakamura 2006, 2010).

The interannual variability of the WNPAC or PJ pattern is a major cause for monsoonal variation over East Asia (Huang and Sun 1992; Kwon et al. 2005). Kwon et al. (2005) found that the intensity of the East Asian summer monsoon, indicated as the precipitation anomalies associated with the Chinese mei-yu, the Korean changma, and the Japanese baiu events, is negatively correlated with that of the western North Pacific. Similarly, Kosaka and Nakamura (2006) showed that the typical PJ pattern associated with the suppressed convection over the off-equatorial western Pacific exhibits a dipole pattern between upper-level anticyclonic vorticity anomalies over the off-equatorial western Pacific around 20°N and those centered east of Japan. The positive meridional vorticity advection due to the climatological northerly between them can induce midlevel upward vertical motions over southeast Japan.

The suppressed convective activity over the off-equatorial western Pacific during the boreal summer season can be modulated by the sea surface temperature (SST) variability over the western Pacific and Indian Ocean (Wu et al. 2010; Kosaka et al. 2013). It can be expected that the suppressed convective activity over the off-equatorial western Pacific can be induced by negative in situ SST anomalies (Su et al. 2001). A possible influence of Indian Ocean SST on the WNPAC was proposed by Xie et al. (2009) and Wu et al. (2009), postulating that Indian Ocean SST warming can generate a boundary layer frictional convergence over the tropics. This may also lead to frictional divergence and the associated downward motion over the off-equatorial western Pacific. Wu et al. (2010) concluded that the western Pacific SST effectively excites the PJ pattern during June, while the Indian Ocean SST is crucial for the excitation of the PJ pattern during July and August.

While the off-equatorial SST in the western Pacific and Indian Ocean and the associated convective activities are keys to link climate variability over the tropics to that in East Asia, dominant SST variability during the early boreal summer season also exhibits a significant SST signal extending to midlatitudes, which can influence climate variation over East Asia through an independent teleconnection pattern. Figure 1a shows the first empirical orthogonal function (EOF) of SST averaged for May from 1982 to 2013. A positive SST anomaly in the off-equatorial western Pacific, which exhibits northeast–southwest-tilting structure, is clearly shown. In addition, negative SST anomalies south of Japan and in the eastern equatorial Pacific are shown. The correlation coefficient between the first principal component (PC1) of the EOF and the Niño-3.4 index during May is −0.9, denoting that this pattern is related to ENSO. It is interesting to note that the positive SST signal in the north-central Pacific (20–30°N, 160°E–160°W; blue box in Fig. 1a) is as strong as that over the off-equatorial western Pacific. Therefore, it is worthwhile to investigate whether this SST forcing has an impact on East Asian climate variability.

Fig. 1.

(a) First EOF of May SST anomalies from 1982 to 2013. (b) Lag correlation between the 5-pentad moving average of station-based temperature anomalies over the Korea Peninsula and SST anomalies in the off-equatorial western Pacific (5°–15°N, 120°–170°E; red box) from May to August. (c) As in (b), but for northern central Pacific SST anomalies (20°–30°N, 160°E–160°W; blue box).

Fig. 1.

(a) First EOF of May SST anomalies from 1982 to 2013. (b) Lag correlation between the 5-pentad moving average of station-based temperature anomalies over the Korea Peninsula and SST anomalies in the off-equatorial western Pacific (5°–15°N, 120°–170°E; red box) from May to August. (c) As in (b), but for northern central Pacific SST anomalies (20°–30°N, 160°E–160°W; blue box).

Even though the SST signal in the northern central Pacific (blue box in Fig. 1a) is stronger than over the off-equatorial western Pacific, a teleconnection with northern central Pacific SST has received less attention because the North Pacific is often considered an area where oceanic variability is forced by the atmosphere as a result of strong transient atmospheric activities independent from SST forcing (Cayan et al. 1995). However, Lau et al. (2004) emphasized the role of the North Pacific SST variability as a regulator of the boreal summer climate over Eurasia and North America. The authors showed that the zonally extended warm SST from east of the Korea Peninsula and the north-central Pacific around 40°N can induce a midlevel anticyclonic circulation, which can increase midlatitude temperature anomalies including those affecting the Korea Peninsula. This implies that it is worthwhile to investigate the influence of north-central Pacific SST variability associated with El Niño on East Asian climate.

Even though there are various atmospheric teleconnection patterns between ENSO and East Asian climate, a statistical relationship between ENSO and climate variability over the Korea Peninsula is not clear (Ahn et al. 1997; Kang 1998; Cha et al. 1999; Kug et al. 2010; Son et al. 2014). For example, Cha et al. (1999) showed that the statistical relationship between ENSO and precipitation anomalies during the boreal summer [i.e., the June–August (JJA) average] is marginal. It is understandable that most previous studies that analyzed the relationship between ENSO and the climate over the Korea Peninsula used seasonal mean data, as ENSO SST forcing is quasi-steady within a season. However, Son et al. (2014) recently found that the influence of ENSO on Korean climate variability exhibits a strong subseasonal variation by comparing the relationship between seasonal mean data and 5-pentad moving-averaged data. This suggests the possibility that atmospheric teleconnection patterns strongly vary within a season even though the SST forcing is quite steady, implying that it is worthwhile to adapt Son et al.’s (2014) approach to analyze the role of ENSO-related western central Pacific SST on Korean climate variability during the boreal summer season.

Therefore, in this study, we will investigate the teleconnection patterns induced by western central Pacific SST anomalies extending in a northeast–southwest direction (i.e., the positive SST anomalies in Fig. 1a) on a monthly time scale (i.e., 5-pentad moving average). We will divide the positive SST anomalies into two areas, the central North Pacific (blue box in Fig. 1a) and the off-equatorial western Pacific (red box in Fig. 1a), and demonstrate that there are two distinct teleconnection patterns caused by SSTs, which result in temperature variability over the Korea Peninsula. Section 2 describes the observational data and linearized model used in this study. In section 3, we will show two distinct teleconnection patterns induced by the western central Pacific SST, which result in climate variation over the Korea Peninsula during the early boreal summer. Section 4 will discuss a linearized model result to support the observational findings. The summary and conclusions are presented in section 5.

2. Description of observational data and the linear baroclinic model

a. Observational data

The observed daily mean SST data were derived from the National Oceanic and Atmospheric Administration’s (NOAA) Optimum Interpolation Sea Surface Temperature version 2 dataset (OISSTv2; Reynolds et al. 2002). Daily mean surface temperatures, atmospheric zonal and meridional wind, geopotential height, and temperature data are from ERA-Interim (Dee et al. 2011). For precipitation, we used the CPC Merged Analysis of Precipitation (CMAP) pentad mean precipitation data from the National Centers for Environmental Prediction (NCEP; Xie and Arkin 1997). To obtain temperature variability over the Korea Peninsula, daily temperature data from 86 stations on the Korea Peninsula, including South and North Korea, were provided by the Korea Meteorological Administration (Son et al. 2014). The analyzed period for all observations is from 1982 to 2013, and all data were detrended before analysis by removing any linear trend.

b. The LBM

We used the linear baroclinic model (LBM), which is based on primitive equations linearized about a basic state (Watanabe and Jin 2003). The model variables, including vorticity, divergence, temperature, and surface pressure, are horizontally represented by spherical harmonics with a spectral T21 horizontal resolution and 20 vertical levels. The basic state is obtained using the NCEP–NCAR reanalysis output (Kalnay et al. 1996). Biharmonic horizontal diffusion was set to have a damping time scale of 5 days for the smallest wave and a damping time scale of 1000 days for vertical diffusion. A linear model, solved with externally imposed diabatic heating (referred to as dry LBM), is utilized in this study. The LBM provides a steady response in time, which is useful to examine the steady atmospheric response to the given diabatic heating.

3. Observational results

Figure 1a shows the first EOF using SST anomalies for May from 1982 to 2013. As mentioned in the introduction, a positive SST anomaly in the western central Pacific is shown which exhibits a northeast–southwest-trending structure. To examine the detailed relationship between Korean temperatures and the SST anomalies in the western central Pacific, the off-equatorial western Pacific SST index and the northern central Pacific index are defined as area-averaged SST anomalies for 5°–15°N, 120°–170°E and 20°–30°N, 160°E–160°W, respectively. Then, the lag correlation between 5-pentad moving-average SST indices and 5-pentad moving-average Korean temperature anomalies during 1982–2013 were calculated and are shown in Figs. 1b,c. The x axis denotes the date at the center of the 5-pentad moving average, and the y axis denotes the lead–lag times in days. For example, the correlation coefficient on 10 June with a lag of −30 days denotes the correlation between the 5-pentad averaged Korean temperature from 29 May to 22 June (i.e., ±12 days centered at 10 June) and the SST index averaged from 29 April to 23 May (i.e., −30 days from 29 May and 22 June). We will denote the off-equatorial western Pacific SST index as off-WP SST index and the northern central Pacific SST index as NCP SST index. Note that the general conclusion is not changed with the slight change of averaged area for the SST indices. For example, the highest lag correlation between the NCP SST and the Korean temperature is still shown during early June when the NCP SST is defined over 25°–35°N, 180°–140°W where the midlatitude SST signal is maximum, even though the amplitude of the correlation is slightly decreased (not shown).

The off-WP SST exhibits the highest correlation coefficients with Korean temperature anomalies during early June with a 1-month lag, indicating a significant relationship between the off-WP SST in May and the Korean climate in June. In more detail, the maximum correlation appears on 12 June with a lag of 25 days. This significant positive relationship exists only for early June, implying a strong subseasonal variation in the relationship. For NCP SST, even though the general relationship is similar to that of the off-WP SST, lag days for the highest correlation are slightly different; that is, the highest correlation appears on 5 June with a lag of −15 days, which implies the existence of two different SST forcing mechanisms for Korean temperature variability during the early boreal summer.

Focusing on the teleconnection mechanism between SST forcing and Korean temperature variability during June, we analyzed the lead–lag relationship between 5-pentad moving-average Korean temperature anomalies centered at 10 June (i.e., averaged from 29 May to 22 June) and atmospheric/oceanic fields. Figure 2a shows the autocorrelation of Korean temperatures for June (averaged from 29 May to 22 June), demonstrating that the Korean temperature anomaly continuously evolves from −6 pentads and reaches the >95% confidence level from −4 to +3 pentads. This means that the Korean temperature variability during the early boreal summer season has a time scale of about one month (i.e., 5 pentads). Figures 2b–g show the lag correlation between Korean temperatures and skin temperatures from ERA-Interim data to examine the temporal evolution of Korean temperature anomalies and data consistencies between the ERA-Interim and in situ temperature data. A weak positive correlation coefficient exists starting at −3-pentad lag and becoming significant from −2-pentad lag forward. The maximum correlation coefficient between station-based Korean temperatures and ERA-Interim skin temperatures is above 0.7 at simultaneous lag, which means that analyses based on reanalysis data (i.e., ERA-Interim) are consistent with station-based data.

Fig. 2.

(a) Autocorrelation of station-based temperature anomalies over the Korea Peninsula. Note that the period for 0-pentad lag is from 29 May to 22 Jun (i.e., ±12 days centered at 10 Jun). (b)–(g) Lag correlation of station-based temperature anomalies over the Korea Peninsula and the skin temperature of ERA-Interim data from 1982 to 2013.

Fig. 2.

(a) Autocorrelation of station-based temperature anomalies over the Korea Peninsula. Note that the period for 0-pentad lag is from 29 May to 22 Jun (i.e., ±12 days centered at 10 Jun). (b)–(g) Lag correlation of station-based temperature anomalies over the Korea Peninsula and the skin temperature of ERA-Interim data from 1982 to 2013.

To examine the spatial pattern of SST forcing associated with Korean temperatures in more detail, Fig. 3 shows lag regressions of SST anomalies onto Korean temperatures. At −5-pentad lag, the spatial distribution of SST anomalies exhibits a pattern similar to the first EOF of May SST shown in Fig. 1a, even though the La Niña signal over the equator is weakened. As the time lag reduces, the positive SST signal in the off-equatorial western Pacific is weakened, which is consistent with Wu et al. (2010), who found that the impact of the off-equatorial western Pacific SST on WNPAC becomes weak after June. In addition, the negative equatorial eastern Pacific SST anomaly is gradually weakened, indicating a decay of the La Niña events (Wu et al. 2010). On the other hand, positive SST anomalies east of the Korea Peninsula–Japan become stronger as the lead lag is increased. This zonally elongated pattern around 40°N is similar to the dominant EOF of SST anomalies during the JJA season (Lau et al. 2004), which can be linked directly to the increased temperatures over the Korea Peninsula.

Fig. 3.

Lag regression of SST anomalies onto station-based Korean temperature anomalies. The period for 0-pentad lag is from 29 May to 22 Jun (i.e.,±12 days centered at 10 Jun). Note that the regression coefficient above the 95% confidence level is denoted as red and blue dots.

Fig. 3.

Lag regression of SST anomalies onto station-based Korean temperature anomalies. The period for 0-pentad lag is from 29 May to 22 Jun (i.e.,±12 days centered at 10 Jun). Note that the regression coefficient above the 95% confidence level is denoted as red and blue dots.

Figure 4 shows lag-regression precipitation anomalies onto Korean temperatures. At −5-pentad lag, positive precipitation anomalies are clearly shown for 30°N, 150°E–180°, where the western part of the positive SST anomaly in the northern central Pacific is located. Over Japan, negative precipitation anomalies are clearly shown. The dipole pattern of precipitation anomalies at −5-pentad lag is gradually weakened, and the positive precipitation anomalies over the off-equatorial western Pacific are significantly enhanced. At the same time, negative precipitation anomalies over the western Pacific at 20°–30°N are generated, which is also evident in the PJ-related precipitation pattern (Kosaka and Nakamura 2006). It is interesting to note that the precipitation signal over the off-equatorial western Pacific gradually intensifies as the time lag is reduced (Fig. 4), while SST forcing in the off-equatorial western Pacific becomes weaker with time (Fig. 3).

Fig. 4.

As in Fig. 3, but for precipitation anomalies.

Fig. 4.

As in Fig. 3, but for precipitation anomalies.

To examine the temporal evolution of SST and precipitation anomalies over the off-equatorial western Pacific and northern central Pacific in more detail, Fig. 5 shows the lag correlation between the off-WP SST and NCP SST indices and Korean temperature over May 29 and June 22 (5-pentad moving average centered on June 10; Fig. 5a) and local precipitation anomalies (Fig. 5b), respectively. Based on Fig. 4, local precipitation associated with off-WP SST and NCP SST is defined as the area-averaged precipitation anomalies for 0°–15°N, 105°–130°E and 30°–35°N, 155°–175°E, respectively. Note that the precipitation index is shifted to the west compared to the SST indices in both regions. This is associated with the fact that the climatological precipitation is greatest over the western Pacific. As the SST anomaly over the climatological dry regions has difficulty generating the precipitation response (Ham and Kug 2012), the precipitation index is shifted to the west where the climatological precipitation is stronger. (This point will be discussed in detail in Fig. 9.)

Fig. 5.

(a) Lag correlation between station-based Korean temperatures and the off-WP (5°–15°N, 120°–170°E; red) and NCP (20°–30°N, 160°E–160°W; blue) SST indices. (b) As in (a), but for off-WP (0°–15°N, 105°–130°E; red) and NCP (30°–35°N, 155°–175°E; blue) precipitation indices.

Fig. 5.

(a) Lag correlation between station-based Korean temperatures and the off-WP (5°–15°N, 120°–170°E; red) and NCP (20°–30°N, 160°E–160°W; blue) SST indices. (b) As in (a), but for off-WP (0°–15°N, 105°–130°E; red) and NCP (30°–35°N, 155°–175°E; blue) precipitation indices.

The maximum correlation between off-WP SST and Korean temperatures is slightly above 0.53 and that between NCP SST and Korean temperatures is 0.46. Both the off-WP SST and NCP SST have the strongest positive relationship with Korean temperature at −5-pentad lag. However, the temporal evolution of the precipitation signal related to Korean temperatures is quite different for the two regions. Over the off-equatorial western Pacific, the precipitation signal exhibits its strongest relationship with Korean temperatures at −2-pentad lag, which is 3 pentads after the local SST peak. This implies that it takes 2–3 pentads for SST forcing to lead to a local precipitation response over the off-equatorial western Pacific. Then, the off-WP precipitation would influence Korean temperatures through atmospheric teleconnections after another 2–3 pentads.

On the other hand, the positive precipitation anomalies over the northern central Pacific show a peak at −7 pentads, implying that the precipitation signal leads the SST signal in time. One may think that this denotes that the atmospheric signal can lead the oceanic response in the NCP region (Frankignoul and Reynolds 1983; Alexander et al. 2002). However, in the case that the atmospheric anomaly induces the SST anomaly, the negative relationship should be appear, as the positive precipitation anomalies act to decrease the SST anomalies by reducing the incoming solar radiation (Wang et al. 2005). Therefore, the positive precipitation signal leading the positive SST signal does not indicate that the NCP SST anomaly is induced by the atmosphere. Then why does the NCP precipitation anomaly show a peak earlier than that of the NCP SST anomaly? We found that the climatological precipitation provides favorable conditions for robust precipitation response at −7-pentad lag, even though this is the period when the SST forcing amplitude is not strongest. (Note that the climatological precipitation will be provided in Fig. 9.) This implies that the amplitude of the precipitation response over the northern central Pacific is mainly determined by the climatological condition. This leads an independent teleconnection pattern by NCP SST prior to that induced by the off-WP SST.

To investigate teleconnection patterns related to convective activity over the northern central Pacific and the off-equatorial western Pacific, Fig. 6 shows precipitation, zonal and meridional wind, geopotential height at 500 hPa, and temperature anomalies at 925 hPa lag regressed onto NCP and off-WP precipitation. Note that, based on Fig. 5, NCP precipitation at −7-pentad lag (i.e., from 24 April to 18 May) from the maximum Korean temperature warming was used for the regression, and 3 pentads after the NCP precipitation is chosen for circulation anomalies. Note that the 3-pentad lag for the circulation anomalies is chosen because the circulation anomalies induced by the NCP precipitation forcing are mixed with those induced by the off-WP precipitation after 3-pentad lag. For off-WP precipitation at −2-pentad lag (i.e., from 19 May to 12 June) from the maximum Korean temperature warming, the circulation pattern at 2 pentads after the off-WP precipitation forcing is shown. Note that the 2-pentad lag for the circulation anomalies is chosen because it is when the correlation between the off-WP precipitation and the Korean temperature is strongest.

Fig. 6.

(left) Regression between NCP (30°–35°N, 155°–175°E) precipitation index and (a) precipitation anomalies, (b) zonal and meridional wind (vectors; m s−1) and geopotential height anomalies at 500 hPa, and (c) temperature anomalies at 925 hPa. The period used for precipitation is from 24 Apr to 18 May and that for other variables is 3 pentads after the period for precipitation. (right) Regression between off-WP (0°–15°N, 105°–130°E) precipitation index and (d)0 precipitation anomalies, (e) zonal and meridional wind (vectors; m s−1) and geopotential height anomalies at 500 hPa, and (f) temperature anomalies at 925 hPa. The period used for precipitation is from 19 May to 12 Jun, and that for other variables is 2 pentads after the period for precipitation. Note that the regression coefficient above the 95% confidence level is denoted as red and blue dots.

Fig. 6.

(left) Regression between NCP (30°–35°N, 155°–175°E) precipitation index and (a) precipitation anomalies, (b) zonal and meridional wind (vectors; m s−1) and geopotential height anomalies at 500 hPa, and (c) temperature anomalies at 925 hPa. The period used for precipitation is from 24 Apr to 18 May and that for other variables is 3 pentads after the period for precipitation. (right) Regression between off-WP (0°–15°N, 105°–130°E) precipitation index and (d)0 precipitation anomalies, (e) zonal and meridional wind (vectors; m s−1) and geopotential height anomalies at 500 hPa, and (f) temperature anomalies at 925 hPa. The period used for precipitation is from 19 May to 12 Jun, and that for other variables is 2 pentads after the period for precipitation. Note that the regression coefficient above the 95% confidence level is denoted as red and blue dots.

The precipitation pattern regressed onto the NCP precipitation in Fig. 6a clearly exhibits local positive precipitation anomalies and, to the northwest, a negative precipitation anomaly over Japan. This negative precipitation persists up to late May, as shown in the pattern regressed onto Korean temperatures shown in Fig. 4, implying that there might be a physical connection between NCP precipitation and the negative precipitation anomaly over Japan. The midlevel circulation pattern after 3 pentads of NCP precipitation forcing exhibits cyclonic flow and associated low geopotential height anomalies northwest of the positive NCP precipitation anomaly and anticyclonic flow and associated positive geopotential height anomalies southeast of the NCP precipitation anomaly. In addition, there is an area of anticyclonic flow over the Korea Peninsula. As the positive geopotential height anomalies are linked to increased thickness, they can cause warm temperature anomalies over the Korea Peninsula (Fig. 6c). The relationship between vorticity and low-level temperature anomalies was also discussed by Hoskins and Karoly (1981). According to the thermal wind relationship, negative vorticity that increases with height is linked to low-level positive temperature anomalies. Because the circulation pattern excited by NCP precipitation exhibits an equivalent barotropic structure whose amplitude increases with height (not shown), the anticyclonic circulation over the Korea Peninsula is also dynamically consistent with the positive temperature anomalies.

Kosaka and Nakamura (2006) used a linearized vorticity equation to examine the structure of the PJ pattern as follows:

 
formula

where S denotes the linearized Rossby wave source (RWS), which is linked to the horizontal convergence/divergence, given by

 
formula

Note that the positive RWS is associated with horizontal convergence. The variables uψ and υψ represent the rotational zonal and meridional wind components, and uχ and υχ are the divergent zonal and meridional wind components, respectively. Figure 7 shows the NCP precipitation-related linearized vorticity advection terms at 200 hPa and the precipitation averaged for 30°–35°N. Over the NCP region at roughly 160°E, where the positive precipitation signal is shown, a positive zonal advection of anomalous vorticity caused by the climatological jet is dominant. As the positive (negative) vorticity anomalies are located over the west (east) of the NCP region, the strong westerly jet stream leads to strong positive anomalous vorticity advection. This positive zonal advection is balanced by the negative Rossby wave source term, which implies upper-level divergence (i.e., enhanced convection). The amplitude of other terms (i.e., meridional advection and the beta term) is systematically smaller than the zonal advection term at the upper level. On the other hand, at a low level, a cyclonic Rossby wave source (i.e., low-level convergence) is dominant over the NCP region as a result of a negative meridional advection of the anomalous vorticity caused by a climatological southerly wind (not shown), which is balanced by a low-level convergence. This means that the positive precipitation anomaly over the NCP region is associated with a low-level convergence, which is balanced by a negative meridional vorticity advection, while the upper-level divergence is balanced by a positive zonal vorticity advection. In short, the barotropic circulation structure can be balanced by the baroclinic RWS term over the NCP region.

Fig. 7.

(a) Precipitation anomalies averaged for 30°–35°N lag regressed onto NCP precipitation. (b) Lag regression of zonal advection of the anomalous vorticity caused by climatological wind (i.e., ) onto NCP precipitation. (c) As in (b), but for meridional vorticity advection (i.e., ). (d) As in (b), but for meridional advection of climatological absolute vorticity caused by anomalous wind [i.e., ].

Fig. 7.

(a) Precipitation anomalies averaged for 30°–35°N lag regressed onto NCP precipitation. (b) Lag regression of zonal advection of the anomalous vorticity caused by climatological wind (i.e., ) onto NCP precipitation. (c) As in (b), but for meridional vorticity advection (i.e., ). (d) As in (b), but for meridional advection of climatological absolute vorticity caused by anomalous wind [i.e., ].

In addition to the positive precipitation anomaly, there is a negative precipitation signal west of the positive anomaly. It is found that the negative precipitation signal and the related upper-level convergence at 120°–140°E are also balanced by negative zonal vorticity advection. This negative zonal vorticity advection is located between the negative vorticity over the Korea Peninsula, and the positive vorticity east of Japan, which indicates that a strong climatological westerly jet stream leads to negative zonal vorticity advection. This in turn means that the wave train extending in a northwest–southeast direction is responsible for the negative precipitation anomaly south of the Korea Peninsula.

The regression pattern related to the off-WP precipitation shown in Figs. 6d–f clearly shows the northward-propagating wave train (Fig. 6e), which is quite similar to the PJ pattern (Kosaka and Nakamura 2006). The precipitation anomalies also show a weak tripolar pattern over the western Pacific. The large geopotential height anomalies over the Korea Peninsula are associated with positive temperature anomalies, which are zonally elongated from the Korea Peninsula to the central Pacific. This shows that both the NCP and off-WP precipitation anomalies can jointly affect warming over the Korea Peninsula.

To compare the atmospheric response related to the NCP and off-WP precipitation and related warming over the Korea Peninsula during the early boreal summer season, Fig. 8 shows a lag regression of geopotential height anomalies at 500 hPa related to the warming over the Korea Peninsula in June (i.e., average from 29 May to 22 June). For a pentad lag between −5 and −4, a wave train extending in a northwest–southeast direction is clear. Negative geopotential height anomalies are shown at the west of the NCP precipitation anomaly (Figs. 4a,b), and there are positive geopotential height anomalies over the Korea Peninsula. This wave train pattern is quite similar to the atmospheric response associated with the NCP precipitation anomaly as shown in Fig. 6. The NCP-related positive geopotential height over the Korea Peninsula during −5- and −4-pentad lag may be responsible for the dramatic temperature increase in time as shown in Fig. 2a by increasing incoming shortwave flux. Above a −3-pentad lag, a zonally elongated positive geopotential height anomaly becomes clear. In addition, a weak negative geopotential height anomaly is also shown over the off-equatorial western Pacific. This pattern is similar to the off-WP precipitation pattern shown in Fig. 6e, which means that off-WP precipitation plays a role in the warming of the Korea Peninsula with a −3-pentad lag.

Fig. 8.

As in Fig. 3, but for geopotential height anomalies at 500 hPa.

Fig. 8.

As in Fig. 3, but for geopotential height anomalies at 500 hPa.

So far, two pathways are demonstrated that a horseshoe warming pattern over the western central Pacific can result in warming of the Korea Peninsula during the early boreal summer season. One is a teleconnection pattern induced by SST anomalies in the northern central Pacific (NCP). These positive NCP SST anomalies result in local positive precipitation during May, which induces anomalous cyclonic flow over the southeast of Japan–Korea Peninsula. Then, the negative vorticity advection at the north-eastern part of this cyclonic flow leads negative RWS, which excites the positive geopotential anomalies over the Korea Peninsula, which eventually leads the temperature increase. During June, SST anomalies in the off-equatorial western Pacific (off-WP) result in local precipitation anomalies, which induce a northward-propagating wave train. This leads to zonally elongated positive geopotential height and related positive temperature anomalies over the Korea Peninsula. In short, the positive temperature anomalies over the western central Pacific act to warm the Korea Peninsula by a teleconnection pattern induced by NCP precipitation during May and off-WP precipitation during June.

The remaining question is that the precipitation signal over the NCP leads the local SST signal in time as shown in Fig. 6, and this may imply that the atmospheric precipitation signal is not a result of local SST forcing. In addition, one can question why the teleconnection patterns linked to the Korean temperature are dramatically changed within one season (i.e., teleconnection related to NCP precipitation in May but to off-WP precipitation in June), even though the spatial distribution of SST anomalies is almost stationary. To address this question, Fig. 9 shows the climatological precipitation at −7- and −2-pentad lags from the maximum warming of the Korea Peninsula (i.e., 24 April–18 May and 19 May–12 June, respectively). During late April–mid-May (i.e., 24 April–18 May), there are regions where climatological precipitation exceeds 10 mm day−1 over the northeastern Indian Ocean and the intertropical convergence zone (ITCZ) located at around 5°N. In addition, climatological precipitation along the east coast of China and south of the Korea Peninsula–Japan is also clearly shown, which is related to the pre-mei-yu and pre-baiu signals (Kobashi et al. 2008). One month later, during late May to mid-June (i.e., 19 May to 12 June), climatological precipitation over the off-equatorial western Pacific is intensified, and the center of the ITCZ is clearly shifted to the north. At the same time, the mei-yu and baiu fronts over the South China Sea and south of the Korea Peninsula–Japan are shifted to the northwest and intensified in amplitude. The difference map (Fig. 9c) reflects those changes, showing a clear precipitation increase over the off-equatorial western Pacific, south of the Korea Peninsula, and over India, while precipitation is decreased southeast of Japan at around 15°–30°N, 130°–160°E.

Fig. 9.

Climatological precipitation (mm day−1) from (a) 24 Apr to 18 May and (b) 19 May to 12 Jun averaged for 1982–2013. (c) Difference in climatological precipitation between the two periods [i.e., (b) minus (a)].

Fig. 9.

Climatological precipitation (mm day−1) from (a) 24 Apr to 18 May and (b) 19 May to 12 Jun averaged for 1982–2013. (c) Difference in climatological precipitation between the two periods [i.e., (b) minus (a)].

The amplitude of the climatological precipitation can determine the sensitivity of the anomalous precipitation to the given SST anomalies. Ham and Kug (2012) argued that a strong climatological dryness prevents anomalous convective activities because the anomalous ascending motion has to overcome the climatological descending motion to generate anomalous precipitation. Therefore, climatological dryness tends to reduce the amplitude of precipitation response to SST forcing (Ham and Kug 2014). On the other hand, wet climatological conditions provide a favorable environment for anomalous convective activity. To examine the evolution of the precipitation sensitivity to the SST in NCP and off-WP regions in time, Fig. 10 shows the precipitation anomalies regressed onto the local SST anomalies over the NCP and off-WP regions. The unit of the regression is millimeters per day per degree Celsius, and the ratio of the standard deviation of NCP SST to that of off-WP SST is multiplied to the NCP regression to compare the precipitation response between two regions. Also, note that the time lag between the precipitation anomalies and the SST anomalies is 2 pentads, which takes into consideration the time for the SST to induce the precipitation response. For the NCP region, it is clear that the most sensitive precipitation response to the given SST anomalies occurred during early May, which is similar to when the precipitation anomalies over the NCP region are greatest (Fig. 4a). Also, this is when the climatological precipitation related to pre-mei-yu and baiu is clear over the NCP region. Except for early May, the amplitude of the precipitation anomalies resulting from the local SST forcing becomes quite weak. This means that the dramatic decrease in the climatological precipitation over the NCP region after May is responsible for the weak precipitation response to the NCP SST forcing, even though the NCP SST forcing is strongest during June. Therefore, it supports our arguments that the climatological precipitation plays a role in determining the atmospheric response to the local SST anomalies. On the other hand, the sensitivity of precipitation anomalies resulting from the off-WP SST is weak during May, and it systematically increases from June. This means that the precipitation response to the given off-WP SST can be robust during June when the amount of the climatological precipitation starts to increase. This implies that a subseasonal change in climatological precipitation is key to understanding a dramatic change in the precipitation response to the given quasi-stationary SST anomalies in the western central Pacific.

Fig. 10.

The regressed precipitation anomalies onto the SST anomalies (mm day−1 °C−1) over the NCP (25°–35°N, 160°E–160°W) and off-WP regions (5°–20°N, 105°–160°E). Note that the ratio of the standard deviation of NCP SST to that of off-WP SST is multiplied by the NCP regression to compare the precipitation response between the two regions.

Fig. 10.

The regressed precipitation anomalies onto the SST anomalies (mm day−1 °C−1) over the NCP (25°–35°N, 160°E–160°W) and off-WP regions (5°–20°N, 105°–160°E). Note that the ratio of the standard deviation of NCP SST to that of off-WP SST is multiplied by the NCP regression to compare the precipitation response between the two regions.

4. LBM experiments

The observed teleconnection patterns induced by NCP and off-WP precipitation were examined in more detail by model experiments using the dry version of the LBM (Watanabe and Jin 2003). First, we obtained precipitation anomalies averaged from −7- to −5-pentad lag (i.e., from 26 April to 30 May) from the maximum Korean temperature warming to simulate the atmospheric response to NCP precipitation, and precipitation anomalies averaged from −3 to 0 lag (i.e., from 14 May to 22 June) were obtained to simulate the atmospheric response to off-WP precipitation. Note that precipitation anomalies for 20°–40°N, 120°E–120°W were used for NCP forcing, and the anomalies for 10°S–40°N, 120°E–120°W were used for off-WP forcing. Assuming that the spatial distribution of precipitation is quite similar to that of column-integrated diabatic heating, a three-dimensional diabatic heating pattern was obtained by multiplying bell-type vertical profiles with maximum values at 600 hPa. The period used to obtain the basic state corresponds to the period used to calculate the regression pattern (i.e., the 26 April–30 May average for NCP forcing and the 14 May–22 June average for off-WP forcing).

Figure 11 shows temperature anomalies at 925 hPa and geopotential anomalies at 500 hPa with the previously described atmospheric forcing for −7- to −5-pentad lag from the Korean temperature warming. The observed patterns for the same period are shown for comparison. As mentioned earlier, the observed temperature (at 925 hPa) and geopotential anomalies (at 500 hPa) exhibit a wave train pattern elongated in a southeast–northwest direction. The LBM experiment with precipitation forcing over the western central Pacific shows a similar, yet slightly different, wave train pattern. For example, the positive geopotential height anomalies and related warming over the Korea Peninsula are simulated well, and the negative geopotential anomalies over 30°–40°N, 150°E–180° are also shown at west of the positive precipitation anomaly. In addition, the barotropic structure is simulated well in the LBM experiments (not shown). When the total diabatic heating is divided into positive (Figs. 11e,f) and negative diabatic heating anomalies (Figs. 11g,h), it becomes clear that the negative geopotential height anomalies are mostly induced by positive diabatic heating over the NCP region (i.e., 30°–35°N, 155°–175°E), while the positive geopotential height anomaly over the Korea Peninsula is caused by negative diabatic heating over Japan.

Fig. 11.

Averaged lag regression fields from −7- to −5- pentad lag of (a) temperature at 925 hPa, (b) geopotential height anomalies at 500 hPa onto the Korean temperatures that occur from 29 May to 22 Jun. LBM results for (c) temperature anomalies at 925 hPa and (d) geopotential height anomalies at 500 hPa with total diabatic heating (contours; K s−1) over 20°–40°N, 120°E–120°W. LBM results for (e) temperature at 925 hPa and (f) geopotential height at 500 hPa with positive diabatic heating (contours; K s−1) and (g) temperature at 925 hPa and (h) geopotential height at 500 hPa with negative diabatic heating (contours; K s−1) are also shown.

Fig. 11.

Averaged lag regression fields from −7- to −5- pentad lag of (a) temperature at 925 hPa, (b) geopotential height anomalies at 500 hPa onto the Korean temperatures that occur from 29 May to 22 Jun. LBM results for (c) temperature anomalies at 925 hPa and (d) geopotential height anomalies at 500 hPa with total diabatic heating (contours; K s−1) over 20°–40°N, 120°E–120°W. LBM results for (e) temperature at 925 hPa and (f) geopotential height at 500 hPa with positive diabatic heating (contours; K s−1) and (g) temperature at 925 hPa and (h) geopotential height at 500 hPa with negative diabatic heating (contours; K s−1) are also shown.

Figure 12 shows the result of the LBM experiments with diabatic heating for −3- to 0-pentad lag (i.e., 14 May–22 June) from the maximum warming over the Korea Peninsula to examine the role of off-WP forcing. Positive temperature anomalies, zonally elongated from the Korea Peninsula to the central Pacific at 40°N, can clearly be recognized. In addition, there are significant positive geopotential height anomalies in the same region. The LBM result shows that the positive temperature anomalies along with the positive geopotential height anomalies can be simulated well using diabatic heating. We also performed the LBM experiment with positive (Figs. 12e,f) and negative diabatic heating (Figs. 12g,h) only. The LBM experiment with positive diabatic heating simulates a PJ pattern, which exhibits northward propagation over the western Pacific. Negative diabatic heating south of the Korea Peninsula simulates a dipole pattern, which exhibits positive geopotential height anomalies over the Korea Peninsula and negative geopotential height anomalies over the subtropical regions (0°–25°N). It is clear that both positive diabatic heating over the off-equatorial western Pacific and negative diabatic heating south of the Korea Peninsula contribute to the warming over the Korea Peninsula. According to Kosaka and Nakamura (2006), the negative diabatic heating (i.e., upper-level convergence related to a descending motion) is balanced by dipole-like vorticity anomalies over the Korea Peninsula and South China Sea. There are negative upper-level vorticity anomalies over the Korea Peninsula and positive upper-level vorticity anomalies over the South China Sea. Between them, there is the climatological northerly during boreal summer, which leads negative vorticity advection, which in turn is linked to the negative diabatic heating. This means that the negative precipitation signal south of the Korea Peninsula can be induced by a PJ-type wave train, implying that both the positive precipitation anomaly over the off-equatorial western Pacific and the negative precipitation anomaly south of Japan, which result in warming over the Korea Peninsula during June, can be induced by the PJ wave train.

Fig. 12.

As in Fig. 11, but for the period of −3- to 0-pentad lag from maximum temperature warming (i.e., from 14 May to 22 Jun).

Fig. 12.

As in Fig. 11, but for the period of −3- to 0-pentad lag from maximum temperature warming (i.e., from 14 May to 22 Jun).

5. Summary and conclusions

This study investigated the impact of SST anomalies in the western central Pacific on temperature variability over the Korea Peninsula during the early boreal summer. We found a significant positive lag relationship between temperature anomalies over the Korea Peninsula during June and western central Pacific SST anomalies during May. In detail, the positive SST anomaly over the western central Pacific, which is similar to the horseshoe pattern produced during La Niña events, induces two distinct atmospheric teleconnection patterns, which commonly result in warming over the Korea Peninsula. During May, positive precipitation anomalies related to SST warming of the northern central Pacific (NCP) induces an atmospheric teleconnection extending in a northwest–southeast direction. As part of this atmospheric teleconnection, a positive geopotential height anomaly over the Korea Peninsula is generated, which can result in warming over the Korea Peninsula. During June, a wave train pattern induced by NCP precipitation is dramatically weakened, while a northward-propagating wave train pattern is strengthened by positive precipitation anomalies over the off-equatorial western Pacific (off-WP). This northward-propagating wave train pattern is quite similar to the Pacific–Japan (PJ) pattern (Kosaka and Nakamura 2006). It is found that this PJ pattern may also cause the positive temperature anomalies over the Korea Peninsula. This means that the combined effects of two distinct atmospheric teleconnection patterns are responsible for the strong relationship between positive SST anomalies in the western central Pacific and temperature anomalies over the Korea Peninsula.

While the atmospheric teleconnection patterns exhibit dramatic changes from May to June, the spatial distribution of the SST anomalies is nearly stationary within one season, as those SST anomalies are associated with the ENSO. We found that the dramatic change in the atmospheric teleconnection from May to June is caused by changes in climatological precipitation. During May, climatological precipitation related to the pre-mei-yu and pre-baiu fronts is over 8 mm day−1 over the northern central Pacific. In contrast, climatological precipitation over the off-equatorial western Pacific is relatively low (i.e., about 6 mm day−1). As wet climatological conditions provide a favorable environment for anomalous convective activities, the precipitation anomaly over the NCP is relatively stronger than that over the off-equatorial western Pacific during May.

During June, climatological precipitation over the off-equatorial western Pacific is dramatically increased along with the activation of the western North Pacific (WNP) monsoon. With the aid of increased climatological precipitation, the positive SST anomalies in the off-equatorial western Pacific efficiently induce local convective activities. On the other hand, as the precipitation band related to the mei-yu and baiu fronts is shifted to the northwest, the climatological precipitation over the northern central Pacific is decreased. Therefore, the precipitation anomaly over the northern central Pacific is suppressed even though the SST signal remains quite similar from May to June. This implies that the response of precipitation anomalies to SST forcing is highly dependent on the atmospheric climatology, so the atmospheric teleconnection exhibits dramatic subseasonal changes even though the SST anomaly remains quasi-stationary throughout the season.

One can wonder whether the relationship between the Korean temperature and the western central Pacific SST found in this study can link the Korean temperature variability to the ENSO. It is found that there is a significant Korean temperature anomaly during the ENSO decaying phase with the aid of relatively strong NCP and off-WP SST anomalies compared to those in the ENSO developing phase (not shown). This is consistent with the result from Wu et al. (2010), which examined the role of Indian Ocean and western Pacific SST on the western North Pacific anticyclone during ENSO decaying phase.

The linkage between the intensity of the atmospheric teleconnection patterns related to ENSO and the climatological precipitation shown in this study implies that the quality of the simulation of midlatitude circulation patterns by atmosphere–ocean general circulation models (AOGCMs) can be determined by the amplitude of the climatological bias. For example, during May, the climatological precipitation related to the mei-yu and baiu bands in most of the AOGCMs that participated in phase 5 of CMIP (CMIP5) (Ham and Kug 2015) was systematically weaker than the observations (Fig. 13). This implies that the atmospheric teleconnection induced by northern central Pacific SSTs might be systematically underestimated in AOGCMs. In addition, several studies pointed out excessive precipitation over the off-equatorial western Pacific (Lin et al. 2006; Ham and Kug 2014; see also Fig. 13). Therefore, it is possible that dynamical prediction systems commonly fail to simulate realistic atmospheric teleconnection patterns, which may cause East Asian climate variability during the boreal summer season. Additional research to examine this issue may help improve climate prediction skills for East Asia.

Fig. 13.

Multimodel ensemble (MME) bias of climatological precipitation during May. Note that the historical run of the 34 climate models that participated in CMIP5 was analyzed. The model list is same as the one used in Ham and Kug (2015).

Fig. 13.

Multimodel ensemble (MME) bias of climatological precipitation during May. Note that the historical run of the 34 climate models that participated in CMIP5 was analyzed. The model list is same as the one used in Ham and Kug (2015).

Most previous studies of teleconnection patterns stimulated by SST anomalies have focused on the role of tropical SST anomalies. As the climatological precipitation is lower over nontropical regions, it was thought that it would be difficult for SST anomalies to induce anomalous convective activities and trigger an atmospheric response in these regions. However, this study has shown that SST anomalies in subtropical regions near 20°–30°N may also excite systematic atmospheric teleconnections. This finding may contribute to improvements of seasonal prediction skills by discovering additional predictors of temperature variability over East Asia. However, because the atmospheric teleconnection induced by subtropical SST anomalies is highly sensitive to the analyzed period and climatological conditions, further detailed analysis is required for the successful development of nontropical SST predictors to improve seasonal temperature predictions.

Acknowledgments

This work was funded by the Korea Meteorological Administration Research and Development Program under Grant KMIPA2015-6170.

REFERENCES

REFERENCES
Ahn
,
J.
,
J.
Ryu
,
E.
Cho
,
J.
Park
, and
S.
Ryoo
,
1997
:
A study of correlations between air-temperature and precipitation in Korea and SST over the tropical Pacific
.
J. Korean Meteor. Soc.
,
33
,
487
495
.
Alexander
,
M. A.
,
I.
Bladé
,
M.
Newman
,
J. R.
Lanzante
,
N.-C.
Lau
, and
J. D.
Scott
,
2002
:
The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans
.
J. Climate
,
15
,
2205
2231
, doi:.
Cayan
,
D. R.
,
A. J.
Miller
,
T. P.
Barnett
,
N. E.
Graham
,
J. N.
Ritchie
, and
J. M.
Oberhuber
,
1995
: Seasonal-to-interannual fluctuations in surface temperature over the Pacific: Effects of monthly winds and heat fluxes. Natural Climate Variability on Decade-to-Century Time Scales, National Academy Press,
133
150
.
Cha
,
E.
,
J.
Jhun
, and
H.
Chung
,
1999
:
A study on characteristics of climate in South Korea for El Niño/La Niña years
.
J. Korean Meteor. Soc.
,
35
,
98
117
.
Dee
,
D.
, and Coauthors
,
2011
:
The ERA‐Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Frankignoul
,
C.
, and
R. W.
Reynolds
,
1983
:
Testing a dynamical model for mid-latitude sea surface temperature anomalies
.
J. Phys. Oceanogr.
,
13
,
1131
1145
, doi:.
Ham
,
Y.-G.
, and
J.-S.
Kug
,
2012
:
How well do current climate models simulate two types of El Nino?
Climate Dyn.
,
39
,
383
398
, doi:.
Ham
,
Y.-G.
, and
J.-S.
Kug
,
2014
:
Effects of Pacific intertropical convergence zone precipitation bias on ENSO phase transition
.
Environ. Res. Lett.
,
9
, 064008, doi:.
Ham
,
Y.-G.
, and
J.-S.
Kug
,
2015
:
Improvement of ENSO simulation based on intermodel diversity
.
J. Climate
,
28
,
998
1015
, doi:.
Hoskins
,
B. J.
, and
D. J.
Karoly
,
1981
:
The steady linear response of a spherical atmosphere to thermal and orographic forcing
.
J. Atmos. Sci.
,
38
,
1179
1196
, doi:.
Huang
,
R.
, and
F.
Sun
,
1992
:
Impacts of the tropical western Pacific on the East Asian summer monsoon
.
J. Meteor. Soc. Japan
,
70
,
243
256
.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
Kang
,
I.
,
1998
:
Relationship between El Niño and Korean climate variability
.
J. Korean Meteor. Soc.
,
34
,
390
396
.
Kobashi
,
F.
,
S.-P.
Xie
,
N.
Iwasaka
, and
T. T.
Sakamoto
,
2008
:
Deep atmospheric response to the North Pacific oceanic subtropical front in spring
.
J. Climate
,
21
,
5960
5975
, doi:.
Kosaka
,
Y.
, and
H.
Nakamura
,
2006
:
Structure and dynamics of the summertime Pacific–Japan teleconnection pattern
.
Quart. J. Roy. Meteor. Soc.
,
132
,
2009
2030
, doi:.
Kosaka
,
Y.
, and
H.
Nakamura
,
2010
:
Mechanisms of meridional teleconnection observed between a summer monsoon system and a subtropical anticyclone. Part I: The Pacific–Japan pattern
.
J. Climate
,
23
,
5085
5108
, doi:.
Kosaka
,
Y.
,
S.-P.
Xie
,
N.-C.
Lau
, and
G. A.
Vecchi
,
2013
:
Origin of seasonal predictability for summer climate over the northwestern Pacific
.
Proc. Natl. Acad. Sci. USA
,
110
,
7574
7579
, doi:.
Kug
,
J.-S.
,
M.-S.
Ahn
,
M.-K.
Sung
,
S.-W.
Yeh
,
H.-S.
Min
, and
Y.-H.
Kim
,
2010
:
Statistical relationship between two types of El Niño events and climate variation over the Korean Peninsula
.
Asia-Pac. J. Atmos. Sci.
,
46
,
467
474
, doi:.
Kwon
,
M.
,
J. G.
Jhun
,
B.
Wang
,
S. I.
An
, and
J. S.
Kug
,
2005
:
Decadal change in relationship between East Asian and WNP summer monsoons
.
Geophys. Res. Lett.
,
32
, L16709, doi:.
Lau
,
K.
,
J.
Lee
,
K.
Kim
, and
I.
Kang
,
2004
:
The North Pacific as a regulator of summertime climate over Eurasia and North America
.
J. Climate
,
17
,
819
833
, doi:.
Li
,
T.
, and
B.
Wang
,
2005
:
A review on the western North Pacific monsoon: Synoptic-to-interannual variabilities
.
Terr. Atmos. Oceanic Sci.
,
16
,
285
314
.
Lin
,
J.-L.
, and Coauthors
,
2006
:
Tropical intraseasonal variability in 14 IPCC AR4 climate models. Part I: Convective signals
.
J. Climate
,
19
,
2665
2690
, doi:.
Nitta
,
T.
,
1987
:
Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation
.
J. Meteor. Soc. Japan
,
65
,
373
390
.
Reynolds
,
R. W.
,
N. A.
Rayner
,
T. M.
Smith
,
D. C.
Stokes
, and
W.
Wang
,
2002
:
An improved in situ and satellite SST analysis for climate
.
J. Climate
,
15
,
1609
1625
, doi:.
Son
,
H.-Y.
,
J.-Y.
Park
,
J.-S.
Kug
,
J.
Yoo
, and
C.-H.
Kim
,
2014
:
Winter precipitation variability over Korean Peninsula associated with ENSO
.
Climate Dyn.
,
42
,
3171
3186
, doi:.
Su
,
H.
,
J. D.
Neelin
, and
C.
Chou
,
2001
: Tropical teleconnection and local response to SST anomalies during the 1997–1998 El Niño. J. Geophys. Res., 106, 20 025–20 043, doi:.
Wang
,
B.
,
R.
Wu
, and
X.
Fu
,
2000
:
Pacific–East Asian teleconnection: How does ENSO affect East Asian climate?
J. Climate
,
13
,
1517
1536
, doi:.
Wang
,
B.
,
Q.
Ding
,
X.
Fu
,
I. S.
Kang
,
K.
Jin
,
J.
Shukla
, and
F.
Doblas‐Reyes
,
2005
:
Fundamental challenge in simulation and prediction of summer monsoon rainfall
.
Geophys. Res. Lett.
,
32
, L15711, doi:.
Watanabe
,
M.
, and
F.-F.
Jin
,
2003
:
A moist linear baroclinic model: Coupled dynamical–convective response to El Niño
.
J. Climate
,
16
,
1121
1139
, doi:.
Wu
,
B.
,
T.
Zhou
, and
T.
Li
,
2009
:
Contrast of rainfall–SST relationships in the western North Pacific between the ENSO-developing and ENSO-decaying summers
.
J. Climate
,
22
,
4398
4405
, doi:.
Wu
,
B.
,
T.
Li
, and
T.
Zhou
,
2010
:
Relative contributions of the Indian Ocean and local SST anomalies to the maintenance of the western North Pacific anomalous anticyclone during the El Niño decaying summer
.
J. Climate
,
23
,
2974
2986
, doi:.
Xie
,
P.
, and
P. A.
Arkin
,
1997
:
Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs
.
Bull. Amer. Meteor. Soc.
,
78
,
2539
2558
, doi:.
Xie
,
S.-P.
,
K.
Hu
,
J.
Hafner
,
H.
Tokinaga
,
Y.
Du
,
G.
Huang
, and
T.
Sampe
,
2009
:
Indian Ocean capacitor effect on Indo–western Pacific climate during the summer following El Niño
.
J. Climate
,
22
,
730
747
, doi:.