Climate Dynamics (2021) 56:2651–2665

https://doi.org/10.1007/s00382-020-05609-4

Inter‑annual variability of spring precipitation over the Indo‑China

Peninsula and its asymmetric relationship with El Niño‑Southern
Oscillation
Gen Li1,2 · Chujie Gao1 · Bo Lu3 · Haishan Chen4

Received: 25 July 2020 / Accepted: 26 December 2020 / Published online: 18 January 2021
© The Author(s), under exclusive licence to Springer-Verlag GmbH, DE part of Springer Nature 2021

Abstract
Previous studies suggested that the dry–wet surface state over the Indo-China Peninsula (ICP), closely associated with the
local spring precipitation, is an important seasonal predictor for the East Asian summer monsoon and extreme climate.
Hence, this work investigates the inter-annual variability of spring precipitation over the ICP and its relationship with El
Niño-Southern Oscillation (ENSO) during 1958–2019. The results show that the spring precipitation anomalies over the ICP
are highly linked to the ENSO-induced atmospheric circulation anomalies. In particular, there are large asymmetries in the
precipitation anomalies for the spring following ENSO. During the decaying spring of the El Niño events, the precipitation
decrease mainly occurs over the Western ICP associated with an anomalous low-level anticyclone over the western North
Pacific. In contrast, during the decaying spring of the La Niña events, a stronger precipitation increase broadly extends into
the Southeastern ICP. This is owing to a nonlinear effect of ENSO on the atmospheric circulation. Compared to El Niño,
the abnormal center of La Niña extends too far westwards, inducing a westward movement of the anomalous atmospheric
circulation, which results in a stronger effect on the spring ICP precipitation. Our findings emphasize the nonlinear responses
of the spring ICP precipitation to ENSO. This has important implications for the seasonal climate prediction over the ICP,
especially for the Southeastern ICP countries/regions.

Keywords Spring precipitation · Inter-annual variability · Indo-China Peninsula · El Niño-Southern Oscillation ·

Asymmetry · Atmospheric circulation

1 Introduction

The Indo-China Peninsula (ICP) mainly includes Myan-

mar, Thailand, Laos, Vietnam, and Cambodia (Fig. 1a),
* Gen Li with dense population and high vegetation coverage (Taka-
[email protected] hashi and Yasunari 2006; Ge et al. 2017). The inter-annual
1 variation of spring precipitation over the ICP can have great
College of Oceanography/Key Laboratory of Marine
Hazards Forecasting, Ministry of Natural Resources/Key influences on the local agriculture, ecology, economy, and
Laboratory of Ministry of Education for Coastal Disaster people’s livelihood (Kanae et al. 2001; Tanaka et al. 2008;
and Protection, Hohai University, Nanjing, China De Bruyn et al. 2014; Mandapaka et al. 2017).
2
Southern Marine Science and Engineering Laboratory The ICP is one of the typical regions that are affected
(Zhuhai), Zhuhai, China by the Asian monsoon, which is a complex coupled
3
Laboratory for Climate Studies and CMA‑NJU Joint ocean–atmosphere-land system (Huang and Yan 1999;
Laboratory for Climate Prediction Studies, National Climate Wen et al. 2004, 2006; Ding and Chan 2005). Meanwhile,
Center, China Meteorological Administration, Beijing, China the land surface processes over the ICP also have great
4
Key Laboratory of Meteorological Disaster, Ministry influences on the Asian monsoon system (Zhang and
of Education (KLME) and Collaborative Innovation Center Qian 2002a; Chen and Zhang 2013; Chen et al. 2015).
on Forecast and Evaluation of Meteorological Disasters
(CIC‑FEMD), Nanjing University of Information Science In particular, previous studies suggested that the ICP
and Technology, Nanjing, China land surface thermal condition plays an important role

13
Vol.:(0123456789)
2652 G. Li et al.

and extreme climate. For instance, Ma et al. (2018) have

shown that the abnormal sensible heating associated with
the spring soil moisture anomaly over the ICP is closely
linked to the onset and development of the East Asian
summer monsoon. Moreover, Yang et al. (2019) pointed
out that the spring ICP soil moisture can be used as a sea-
sonal predictor for the extreme high temperature events
over the Yangtze River basin. Our recent study (Gao et al.
2019) further clarified that the effects of the soil moisture
over the ICP on the monsoonal system are partly due to
the local dry–wet transitional climate condition in spring.
As shown in Fig. 1c, the spring precipitation over the ICP
is medial compared with the dry India and Mongolia and
the wet Southern China. Gao et al. (2020a, b) also indi-
cated that the spring soil moisture anomalies over the ICP
are primarily related to the inter-annual variation of local
precipitation. However, the features and causes of inter-
annual variation of the spring precipitation over the ICP
remain unclear.
El Niño-Southern Oscillation (ENSO) is one of the most
important climatic events over the whole world and is the
strongest inter-annual signal in the ocean–atmosphere
coupling system (Philander 1983; Rasmusson and Wal-
lace 1983; Wang et al. 2017). It not only has great impacts
on the marine environment and weather/climate states
over the tropical areas, but also exerts large influences on
the global climate, and even leads to the occurrences of
extreme weather/climate events worldwide (Neelin et al.
1998; Wallace et al. 1998; Lyon and Barnston 2005; Lin
et al. 2018; Luo and Lau 2019). It has been known that
the ENSO-related sea surface temperature (SST) anoma-
lies can induce complex precipitation anomaly patterns,
especially over the tropical areas (Dai and Wigley 2000;
Wang et al. 2000; Juneng and Tangang 2005; Hu and
Huang 2010; Ge et al. 2017). For the El Niño events, the
abnormally higher SST over the equatorial eastern Pacific
intensifies convection and precipitation over the tropical
Fig. 1  a Topography (units: m) in Asia. b Mean water vapor flux central and eastern Pacific. Meanwhile, over the maritime
­ −1 s−1) integrated from 300 to 1000 hPa and its
(arrows; units: kg m continent and the western Pacific, a precipitation deficit
divergence (colors; units: 1­ e5 kg ­m−2 s−1) in spring during 1958– usually occurs (Ropelewski and Halpert 1987; Hendon
2019. c Total precipitation (units: mm) in spring for the period of
1958–2019. The red boxes denote the Indo-China Peninsula (ICP; 2003; Lau and Nath 2003; Chang et al. 2004). ENSO
95°–110° E, 8°–22° N) affects the remote precipitation mainly by its influence on
the atmospheric circulation (Ropelewski and Halpert 1987;
Ju and Slingo 1995; Ng et al. 2019). For example, there is
in the East Asian summer monsoon system (Zhang and usually an anomalous anticyclone at the lower troposphere
Qian 2002b; Jin et al. 2006), since the ICP is located in over the western North Pacific from the El Niño developing
the upper region of the southeasterly monsoon wind flow autumn to the decaying summer, transporting excessive
(Fig. 1b). The work of Jin et al. (2006) suggested that moisture to the Southern China and increasing local pre-
the thermal contrast between the ICP and the Western cipitation (Wang et al. 2000; Wang and Zhang 2002; Jiang
Pacific is of great importance for the East Asian summer et al. 2019). Such an anomalous anticyclone is induced
monsoon formation. By affecting the local thermal state, by the local SST cooling as a Rossby wave response to El
the surface dry–wet condition in spring over the ICP can Niño (Wang et al. 2000). In addition, the Walker circula-
exert profound effects on the East Asian summer monsoon tion adjustment is closely linked to ENSO (Ropelewski

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2653

and Halpert 1989; Power and Smith 2007; Sasaki et al. atmospheric data, including monthly wind, humidity, and
2014). During the El Niño and La Niña years, the Walker vertical velocity, are collected from the Japanese 55-year
circulation is weakened and intensified, respectively (Phi- reanalysis dataset (JRA-55; Kobayashi et al. 2015) pro-
lander 1990), affecting the precipitation anomalies around vided by the Japan Meteorological Agency. The monthly
the world. SST data with a resolution of 1° × 1° are from the Hadley
The ENSO-related climate responses are sensitive to the Center Sea Ice and SST dataset (HadISST; Rayner et al.
location and intensity of the SST anomaly center. However, 2003). The present study period is set to be 1958–2019,
due to the existence of nonlinearity in the climate system, which covers the common time span of these datasets. In
the warm and cold events of ENSO are not entirely coun- addition, a surface (0–10 cm) soil moisture dataset (1° × 1°)
tered, but with significant asymmetries (Timmermann et al. from the Global Land Data Assimilation System (GLDAS)
2018). For the intensity, previous studies indicated that the V2.0 during 1958–2012 is used in our discussion (Rodell
El Niño events tend to be stronger than the La Niña events et al. 2004; unavailable for the data after 2012).
(An and Jin 2004). Moreover, the SST anomaly pattern of The water vapor flux and its divergence are computed
the La Niña events exhibits a westward shift compared to the by the following equations:
El Niño events (Hoerling et al. 1997). Even the durations of
the warm and cold ENSO events exhibit asymmetries (Oku-
mura and Deser 2010). Some studies have suggested that
Table 1  List of the El Niño and La Niña decaying years during 1958–
asymmetries also consist in the climate effects of ENSO (An 2019 used in the composite analysis
et al. 2005; Chou and Lo 2007; Ng et al. 2019). For exam-
ple, Hoerling et al. (1997) investigated the tropical precipita- El Niño 1958 1964 1966 1973 1983
tion anomalies associated with ENSO, and pointed out that 1987 1992 1998 2010 2016
the maximum of precipitation anomaly over the equatorial La Niña 1971 1974 1976 1985 1989
Pacific during mature winter of El Niño is farther east than 1996 1999 2000 2008 2011
that of La Niña. Thus, both the linearity and nonlinearity
exist in the ENSO effects, causing a great challenge in the
seasonal prediction.
In this study, the primary objective is to explore the inter-
annual variability of spring precipitation over the ICP and
its relationship with ENSO. The results show that the spring
precipitation anomalies over the ICP are largely affected by
ENSO, with distinct asymmetrical responses to El Niño and
La Niña. Our findings are important for the seasonal climate
prediction over the ICP, especially for the Southeastern ICP
countries/regions. The rest of this work is arranged as fol-
lows. Section 2 introduces the data and methods used in this
study. Section 3 explores the connection between the spring
ICP precipitation and ENSO events. Section 4 demonstrates
the nonlinearity in the spring ICP precipitation responses to
ENSO. Section 5 investigates how the asymmetric ENSO-
induced circulation anomalies affect the spring precipita-
tion anomalies over the ICP. Summary and discussions are
presented in Sect. 6.

2 Data and methods

In this study, the topographic data are from the Global 30

Arc-Second Elevation (GTOPO30 10 min) dataset (https​ Fig. 2  a The first EOF mode of inter-annual variability of spring pre-
://lta.cr.usgs.gov/GTOPO​30). The ground-based monthly cipitation anomalies over the ICP during 1958–2019. b Standardized
precipitation data with a high horizontal resolution of time series of the principal component of the first EOF mode (PC1),
0.5° × 0.5° are gathered from the Climatic Research Unit regional mean spring precipitation (RMP) over the ICP, and preced-
ing winter Oceanic Niño Index (ONI). The r1, r2, and r3 are the cor-
(CRU) at the University of East Anglia (https​: //cruda​ relation coefficients between PC1 and RMP, between PC1 and ONI,
ta.uea.ac.uk/cru/data/hrg/cru_ts_4.03/). The 1.25° × 1.25° and between RMP and ONI, respectively

13
2654 G. Li et al.

Ps largest negative ONI values are selected, and their decay-

g ∫Pt
1 ��⃗
Q= ∇ ⋅ (qV)dP (1) ing years are listed in Table 1. The long-term trends of
all datasets are removed to avoid the possible influences
where g is the acceleration of the gravity, q is the specific of climate change (such as global warming). The com-
humidity, V⃗ is the horizontal wind velocity, P is the air pres- posite anomalies for the variables are calculated by the
sure, and Pt and Ps denote the top and surface pressure lay- deviations apart from the climatology of 62-year mean
ers, respectively. As little moisture exists above 300 hPa, states. As this study is to focus on the asymmetry in the
Eq. (1) is integrated from 300 to 1000 hPa. In this study, climate responses associated with the ENSO, a summa-
the positive (negative) value of Q denotes the divergence tion of the two composite anomalies in the El Niño and La
(convergence) of the tropospheric moisture transport, unless Niña years is used to highlight the nonlinear components
otherwise specified. (Hoerling et al. 1997).
We also employ the Oceanic Niño Index (ONI) pro- In addition, some ordinary methods, such as the correla-
vided by the US Climate Prediction Center (CPC) in the tion, regression, and empirical orthogonal function (EOF)
mature winter (December, January, and February; DJF) analyses, are adopted in this study. Specifically, a regression
to denote the strength of ENSO events, which is based approach is used to remove the SST anomaly signals:
on SST anomalies in the NINO3.4 region (120°–170°
W, 5° S–5° N). This index and its detailed information
V = V ∗ − S × cov(V ∗ , S)∕var(S) (2)
can be found at the following website: http://origi​n.cpc. where V* and S are the given variable and regional averaged
ncep.noaa.gov/produ​c ts/analy​s is_monit​o ring​/ ensos​t uff/ SST anomalies, cov(V*, S) is the covariance between the
ONI_v5.php. The selection of the El Niño and La Niña variable and the SST anomalies, var(S) is the variance of the
events is based on the ONI. The ten years having the SST anomaly, and V is the variable anomaly after removing
largest positive ONI values and the ten years having the the SST anomaly signal.

Fig. 3  Correlation coefficients

of the a autumn, b winter, and c
spring sea surface temperature
(SST) anomalies with the spring
precipitation anomaly averaged
over the ICP for the period of
1958–2019. Numeral “− 1” in
the bracket denotes the develop-
ing year of ENSO events. The
dotted areas are significant with
p < 0.1. All data are linearly
detrended

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2655

3 Inter‑annual variability of spring (Fig. 2a). The largest precipitation anomalies are in the
precipitation over the ICP and its Western coast and the Northern region, and the anomalies
relationship with ENSO over the Eastern part are relatively tender. This is possibly
related to the topography and spring water vapor transport
The EOF analysis is applied for the inter-annual anom- over the ICP. There are high mountains along the coastal
alies of the spring (March–May) precipitation over the region of the ICP (Fig. 1a). Thus, when the moisture is
ICP. The first EOF mode demonstrates that the positive transporting from the Indian Ocean to the Southern China
anomaly dominants the whole ICP area, which indicates a (Fig. 1b), the air is forced to rise ahead of the mountains,
regional consistency of the abnormal spring precipitation causing relatively larger precipitation amount (Fig. 1c) and

Fig. 4  a Correlation coefficients of spring stream function (con- p < 0.1) and divergent wind (arrows; only arrows significant with
tours; contour interval (CI): 0.2; the shaded areas are significant with p < 0.1 are shown) anomalies at 200 hPa. The dashed and continu-
p < 0.1) and rotational wind (arrows; only arrows significant with ous lines denote the negative and positive values, respectively. The
p < 0.1 are shown) anomalies at 850 hPa with the preceding winter zero contours are not shown. All data are linearly detrended. The red
ONI for the period of 1958–2019. b Same as in (a), but for the veloc- boxes denote the ICP region
ity potential (contours; CI 0.2; the shaded areas are significant with

Fig. 5  Correlation coefficients between spring vertical velocity (con- tive values, respectively. The zero contours are not shown. A negative
tours; CI 0.2; the shaded areas are significant with p < 0.1) averaged (positive) value indicates an upward (downward) movement. All data
over 0°–20° N and the preceding winter ONI for the period of 1958– are linearly detrended. The red box is above the ICP region
2019. The dashed and continuous lines denote the negative and posi-

13
2656 G. Li et al.

variability along the Western coastal region and the North- in the air column, which is also an unfavorable condition for
eastern ICP. The first EOF mode contributes to about 54% the ICP precipitation.
of inter-annual variability of the ICP spring precipitation. To further examine the circulation responses at the upper
By contrast, the second EOF mode, which only contrib- layer of the troposphere, the correlation distributions of the
utes to about 9% of the spring ICP precipitation variation, 200 hPa velocity potential and divergent wind in the decaying
shows an opposite anomaly pattern between the northeast- spring with the preceding winter ONI are shown in Fig. 4b.
ern and southwestern parts of the ICP (figure not shown). There are evident divergence and convergence regions at a
Besides, the correlation coefficient between the principal higher level over the eastern and western Pacific during the
component time series of the first EOF mode (PC1) and El Niño events, respectively. It also indicates that there is an
the regional mean of spring precipitation anomaly over the abnormal ascending (descending) motion over the eastern
ICP (95°–110° E, 8°–22° N) even reaches 0.99 (p < 0.001;
Fig. 2b). Thus, the first EOF mode can well capture the
inter-annual variability of area-averaged spring precipita-
tion over the ICP.
Being one of the most important climatic driving factors
on the inter-annual timescale around the world, ENSO plays
a crucial role in regulating the tropical precipitation in the
decaying spring (Ropelewski and Halpert 1987). The abnor-
mal spring ICP precipitation is highly in association with the
tropical SST anomalies. As shown in Fig. 3, the area-aver-
aged spring precipitation over the ICP exhibits significant
negative correlations with the equatorial eastern Pacific SST
anomalies for the preceding autumn (September–Novem-
ber), preceding winter (December–February), and spring. To
check this relationship between the precipitation and ENSO-
related SST anomalies, we further utilize the preceding win-
ter ONI from the CPC to denote the intensity of ENSO.
The correlation coefficient between the area-averaged spring
precipitation anomaly over the ICP and the ONI is − 0.47
during the past six decades (Fig. 2b), which is statistically
significant with p < 0.001. This indicates that there is usually
an enhanced precipitation in spring over the ICP when the
tropical SST over the eastern Pacific is abnormally lower,
i.e., La Niña events, and vice versa. The abnormal ENSO
SST would motivate the atmospheric circulation anomalies,
and further induce anomalous precipitation over Asia (Ju
and Slingo 1995). Particularly, an anomalous anticyclone
is usually developed over the western North Pacific dur-
ing the decaying spring as a response to the El Niño events
(Wang et al. 2000; Wang and Zhang 2002). Figure 4a shows
the abnormal spring stream function and rotational wind
at 850 hPa correlated with the preceding winter ONI. An
abnormal anticyclone circulation pattern exists with a center
to the east of the ICP. This indicates that when the SST
is abnormally higher over the tropical eastern Pacific, i.e.,
an El Niño event, the ICP would be under the effects of
an abnormal low-level anticyclone circulation system. This
abnormal circulation would guide the water vapor transport-
ing to the Southern China excessively. In turn, this implies
less moisture convergence over the ICP, lowering the water Fig. 6  a Correlation coefficients between the spring precipitation
anomalies and the preceding winter ONI for the period of 1958–
source for local spring precipitation. On the other hand, an
2019. The dotted areas are significant with p < 0.1. All data are lin-
anticyclone normally associated with a downward movement early detrended. b and c are the same as in (a), but for the years with
ONI > 0 and ONI < 0, respectively

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2657

Fig. 7  Composite anomalies of

spring precipitation (units: mm)
in the a El Niño cases and c La
Niña cases. The dotted areas are
significant with p < 0.1. e The
summation of (a) and (c). The
red boxes denote the Western
ICP (96°–104° E, 12°–21° N)
and Southeastern ICP (104°–
109° E, 11°–17° N) regions.
b, d, and f are the same as in
a, c, and e, but for the anoma-
lies divided by the preceding
winter ONI of each El Niño/La
Niña year. All data are linearly
detrended

(western) Pacific accompanied by a warmer equatorial east- the ICP is in the ascending branch of the Walker circulation,
ern Pacific. This could be attributed to the responses of the this further confirms that the ICP is under the evident effects
Walker circulation to the abnormal tropical SST. Theoretically, of ENSO-induced atmospheric circulation anomalies. Dur-
the abnormal higher SST over the tropical eastern Pacific can ing the decaying spring of the El Niño events, the weakened
weaken the Walker circulation during El Niño years (Philan- Walker circulation leads to an abnormal descent above the
der 1990). As shown in Fig. 5, the strong negative correla- ICP (Fig. 5), further hindering the local precipitation. On the
tions of vertical velocity anomalies with the preceding winter contrary, during the decaying spring of the La Niña events, an
ONI indicate an anomalous upward motion over the tropical abnormal low-level cyclone with an upward movement above
eastern Pacific in the El Niño decaying spring. On the other the ICP would favor the local precipitation.
hand, the significant positive correlations over the tropical
western Pacific denote an anomalous downward motion. As

13
2658 G. Li et al.

4 The nonlinear responses of the spring ICP Niño and 10 strongest La Niña events based on the preced-
precipitation to ENSO ing winter ONI for the composite analysis (see Table 1 and
Sect. 2 for details). During the decaying spring of the El
However, we know that the positive (El Niño) and negative Niño events, the Western ICP exhibit an evident precipita-
(La Niña) phases of ENSO are not simply the reverse of tion deficit, while there is no statistically significant pre-
each other, but with significant asymmetries in magnitude, cipitation anomaly over the Southeastern ICP (Fig. 7a). In
duration, and spatial structure (An and Jin 2004; Timmer- contrast, the precipitation anomaly is relatively larger during
mann et al. 2018). For example, compared to the El Niño the decaying spring of the La Niña events and most regions
events, the cool SST anomalies on the equator for the La of the ICP display a precipitation surplus, especially for the
Niña events often extends too westwards into the central Southeastern ICP (Fig. 7c).
and western Pacific (An and Jin 2004; Guan et al. 2019). To show the linear and nonlinear parts of the spring
Such an asymmetry in the ENSO spatial structure could precipitation responses more intuitively, the summation of
potentially exert complex influences on the global climate the anomalies in the above two cases is further illustrated
(Wallace et al. 1998; Alexander et al. 2002; An and Jin in Fig. 7e. The lower values indicate that the precipitation
2004; Timmermann et al. 2018). anomalies are usually opposite with similar magnitudes
To explore the linear and nonlinear responses of spring between the decaying springs of the El Niño and La Niña
precipitation over the ICP to ENSO, Fig. 6a–c demonstrate events. Correspondingly, the higher values highlight the
the correlations between the spring precipitation anomalies nonlinear component of the anomalies. In general, the spring
and the preceding winter ONI under different conditions. precipitation over the Western ICP is linearly related to the
Firstly, for the whole period of 1958–2019, the spring precip- ENSO SST anomalies from the regional average perspec-
itation over the ICP bears significantly negative correlations tive. However, the Southeastern ICP (including the Eastern
with the preceding winter ONI (Fig. 6a). However, when we Thailand, most of Cambodia, and the Southern Laos and
further investigate this relationship separately in the warm Vietnam) exhibits a distinct nonlinear response of the spring
and cold ENSO phases, the negative correlations exhibit evi- precipitation to ENSO. Different from El Niño, La Niña can
dent asymmetries (Fig. 6b and c). In the years with positive cause a strong precipitation anomaly in the decaying spring
SST anomalies over the NINO3.4 region (i.e., ONI > 0), the over the Southeastern ICP, suggesting a unique regional cli-
negative correlations mainly dominate the Western ICP. In mate effect. In addition, we also identify the El Niño (La
the years when the ENSO SST anomalies are lower than nor- Niña) events when the NINO3.4 index is greater (less) than
mal (i.e., ONI < 0), the domain with the negative correlation 0.5 °C (− 0.5 °C) for at least 5 months during 1958–2019 for
largely extends into the Southeastern ICP. This implies that comparison, the results are in agreement with the analyses of
the impact of ENSO-related SST anomalies on the spring the 20 strongest ENSO years (figure not shown).
ICP precipitation in the La Niña decaying years is relatively As an effort to preliminarily explore the ENSO asym-
larger than in the El Niño decaying years. metries in amplitude and duration, the composite evolutions
To verify the nonlinear components of spring precipita- of the NINO3.4 SST anomalies in the El Niño and La Niña
tion responses to ENSO events, we select 10 strongest El events are shown in Fig. 8. The El Niño SST anomalies are

Fig. 8  Composite evolutions

of SST anomaly (SSTA; units:
°C) averaged over the NINO3.4
region (170° W–120° W, 5°
N–5° S) since the develop-
ing summer to the decaying
summer for the El Niño and La
Niña cases. All data are linearly
detrended. The solid bars are
significant with p < 0.1

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2659

slightly stronger from the developing summer to the mature precipitation over the ICP, especially for the Southeastern
winter than the La Niña ones. For their duration, both the El ICP. Now we turn our attention to investigate the nonlin-
Niño and La Niña SST anomalies could sustain to the decay- earity in ENSO and its asymmetric effects on the atmos-
ing spring, and the La Niña events exhibit a longer duration pheric circulation and spring precipitation over the ICP. In
until the decaying summer. For April and May of the decay- this section, we find that the western shift of the abnormal
ing spring, the El Niño SST anomalies are slightly weaker SST center in the La Niña events can lead to a westward
than the La Niña ones. In order to exclude the influence of
the amplitude and duration asymmetries of ENSO, we have
also normalized the composite spring precipitation anomalies
by scaling the absolute value of the preceding winter (Fig. 7;
or decaying spring, the results are similar, figure not shown)
ONI of the ENSO years. The results still display a strong
asymmetry in the precipitation anomalies over the ICP for
the spring following ENSO (Fig. 7b, d, and f). This indicates
that the amplitude and duration of the ENSO events have lit-
tle influence on our major findings in this study.
We further select two sub-regions of the ICP: the West-
ern ICP (96°–104° E, 12°–21° N) and the Southeastern ICP
(104°–109° E, 11°–17° N). Figure 9 shows the detailed rela-
tionships between the preceding winter ONI and spring pre-
cipitation anomalies averaged over different regions. For the
whole ICP region, the correlation coefficients between the
preceding winter ONI and spring precipitation anomalies are
− 0.38 (p < 0.05) and -0.40 (p < 0.05) during the warm and
cold phases of ENSO, respectively (Fig. 9a). This implies
that the La Niña events could result in a relatively stronger
effect on the ICP precipitation in the decaying spring than the
El Niño events. For the Western ICP region, a linear response
is dominated between the spring precipitation anomalies and
the ENSO. The negative relationships are both robust for
the years with ONI < 0 and ONI > 0 (Fig. 9b), with a cor-
relation coefficient of − 0.41 (p < 0.05). In particular, for the
Southeastern ICP region, the negative relationship is strong
under the condition of ONI < 0, but it nearly vanishes when
ONI > 0 (Fig. 9c). This is suggestive of an asymmetric rela-
tionship between the spring precipitation over the ICP and
the ENSO. During the decaying phase of ENSO, the Indian
Ocean basin-wide SST anomalies, i.e., the Indian Ocean
basin mode (IOBM), can usually affect the Asian climate by
regulating the atmospheric circulation (Xie et al. 2016; Tao
et al. 2016). We further examine the relationship between
the preceding winter ONI and the spring ICP precipitation
anomalies after removing the IOBM signals in each condition
(figure not shown). The results also support our conclusion
that the effect of La Niña on the following spring precipita-
tion over the ICP is larger than El Niño.

Fig. 9  Relationships of spring precipitation anomalies over the a ICP,

5 The asymmetric effects of ENSO b Western ICP, and c Southeastern ICP with the preceding winter
ONI for the period of 1958–2019. The blue and red dots are for the
years with ONI < 0 (28 years) and ONI > 0 (31 years), respectively.
So far, our results indicate that the ICP precipitation
The r is the correlation coefficient between the spring ICP precipita-
responses to ENSO exhibit an obvious asymmetry. Spe- tion anomalies and the preceding winter ONI. Precipitation data are
cifically, La Niña has a stronger effect on the spring linearly detrended and standardized

13
2660 G. Li et al.

Fig. 10  Composite anomalies

of the decaying spring SST
(contours and colors; units: °C;
CI 0.2 °C; the zero contours are
not shown) for the a El Niño
cases and b La Niña cases.
The dashed and continuous
lines denote the negative and
positive values, respectively.
The dotted areas are significant
with p < 0.1. c The summation
of a and b. All data are linearly
detrended

movement of the atmospheric circulation response. The (Hoerling et al. 1997; An and Jin 2004; Timmermann et al.
anomalous low-level cyclone associated with La Niña shifts 2018) on the ENSO nonlinearities for their mature phases
westwards covering the ICP, compared with the anomalous (boreal winter).
low-level anticyclone in the El Niño events. This induces a Figure 11 demonstrates the evolution of SST anomalies
more effective influence on the precipitation over the ICP along the equatorial Pacific from the developing summer to
during the decaying spring of La Niña. decaying summer of ENSO events. For the El Niño events,
To analyze the causes of the nonlinear responses of the SST over the equatorial eastern Pacific is abnormally
atmospheric circulation and precipitation, we further higher during the developing summer to decaying spring,
look into the asymmetries of the ENSO patterns during with its maximum in the mature winter (Fig. 11a). For the
the decaying spring. Due to the huge thermal capacity La Niña events, the situation is basically on the contrary
of the ocean and the strong memory of SST anomalies, (Fig. 11b): such SST is abnormally lower during the devel-
the SST anomalies generally maintain the pattern in their oping summer to decaying spring. The nonlinear pattern
mature phases for the ENSO decaying spring (Fig. 10a shows a negative center and a positive center over the central
and b). In the El Niño cases, the abnormally higher SST and eastern Pacific since developing summer, respectively
anomalies are mainly located in the tropical eastern Pacific (Fig. 11c). This indicates that the westward movement of
and Indian Oceans. In the La Niña cases, the equatorial the SST anomaly center exists in the whole process of the
cool SST anomalies extend westwards. Figure 10c high- La Niña events, which also implies that the response of
lights the nonlinear part of the tropical SST anomalies, the abnormal atmospheric circulation system might move
and indicates that the SST anomaly center for La Niña westwards.
moves westwards. This indicates that the equatorial Pacific To verify our hypothesis, Fig. 12 shows the anomalous
SST anomalies during the decaying spring for the La circulations at 850 hPa in the two ENSO phases. During the
Niña events extend too far westwards in comparison to decaying spring of the El Niño events, it is evident that an
the El Niño events, which is similar with previous studies anomalous anticyclone exists over the western North Pacific,

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2661

Fig. 11  Composite evolutions of SST anomalies (contours and colors; Niña cases. The dashed and continuous lines denote the negative and
units: °C; CI 0.2 °C; the zero contours are not shown; only areas sig- positive values, respectively. c The summation of a and b. Numeral
nificant with p < 0.1 are shaded with colors) averaged over 2° S–2° N “− 1” in the bracket denote the developing year of ENSO events. All
from June of the previous year to July in the a El Niño cases and b La data are linearly detrended

and the anomalous center is to the east of the ICP (Fig. 12a). west side of the ICP. On the contrary, the abnormal cyclone
Normally, this would guide the moisture transporting to the over the ICP in the La Niña events prevents the water vapor
Southern China, and hamper the local precipitation over the transporting northeastward (Fig. 13b), resulting in an abnor-
ICP. For the La Niña events, an anomalous cyclone is located mal convergence of air moisture over the ICP. This would
over the western North Pacific (Fig. 12b). In addition, the benefit local spring precipitation. The nonlinear part of
anomalous center moves westwards and covers the whole water vapor flux anomaly in the ENSO events indicates that
ICP compare with the El Niño events. Figure 12c shows a the Na Niña events exert a stronger effect on the moisture
La Niña-like pattern of nonlinear response anomalies, con- flux over the ICP (Fig. 13c), owing to a westward shift of the
firming that the abnormal low-level circulation system in low-level abnormal cyclone in the La Niña years (Fig. 12c).
response to the La Niña events shifts westwards obviously. For the Walker circulation anomalies in response to
This hints that the atmospheric circulation and precipitation ENSO in the decaying spring, there also exist distinct dif-
over the ICP can be affected more effectively by the La Niña ferences reflected by the vertical motion anomalies above the
in the decaying spring. ICP. In the El Niño years, there is an abnormal descending
Generally, the lower tropospheric circulation anomaly movement at the middle troposphere over the ICP (Fig. 14a).
would lead to an abnormal water vapor transport. In the Particularly, the west side of the ICP shows most evident
decaying spring of the El Niño events, the abnormal south- positive (downward motion) velocity anomaly, which is in
westerly wind along the west rim of the abnormal anticy- corresponding to the spring precipitation deficit over the
clone brings excessive water vapor from the ICP and South Western ICP. During the decaying spring of the La Niña
China Sea to the Southern China (Fig. 13a). Meanwhile, events, the anomaly area with abnormal ascending move-
there exists an evident divergence of moisture flux over the ments extends eastward, and the vertical velocity anomalies

13
2662 G. Li et al.

Fig. 12  Composite anomalies

of the decaying spring stream
function (contours; units: ­m2
­s−1; CI 0.4 ­m2 s−1; the zero
contours are not shown and the
shaded areas are significant
with p < 0.1) and rotational
wind (arrows; units: m ­s−1; only
arrows significant with p < 0.1
are shown) at 850 hPa for the
a El Niño cases and b La Niña
cases. The dashed and continu-
ous lines denote the negative
and positive values, respec-
tively. c The summation of a
and b, but the stream function
is denoted by colors instead of
contours. All data are linearly
detrended. The red boxes denote
the ICP region

are evidently enhanced compared to the El Niño events ICP precipitation in the decaying spring when the tropical
(Fig. 14b). This would benefit the spring precipitation for eastern Pacific SST is abnormally colder (i.e., a La Niña
the entire ICP to a great extent. The nonlinear responses event), and vice versa.
of vertical motion to ENSO also confirm that the La Niña Further analyses show that the responses of precipita-
events exert a stronger effect over the ICP than the El Niño tion anomalies over the ICP exhibit large asymmetries in the
events (Fig. 14c). This suggests a more favorable condition decaying springs of the El Niño and La Niña events. This
for the ICP precipitation anomaly during the La Niña decay- is owing to the nonlinear effects of the ENSO events. Dur-
ing spring. ing El Niño years, the abnormally warmer tropical eastern
Pacific SST motivates an anomalous anticyclone at the lower
troposphere over the western North Pacific near the ICP.
6 Summary and discussions This would result in water vapor transporting from the ICP
to the Southern China excessively. Besides, the abnormal
The agrarian-based socioeconomic livelihood of densely downward movement responding to the ENSO SST anom-
populated ICP countries/regions is vulnerable to modest aly only covers the western part of the ICP. Therefore, the
changes in spring precipitation. Thus, this study further abnormally lower precipitation usually occurs over the West-
explores the inter-annual variability of spring precipitation ern ICP during the El Niño decaying spring. Compare to El
over the ICP and its asymmetric relationship with ENSO. On Niño, La Niña SST anomalies tend to extend too far west-
the inter-annual timescale, the ENSO strongly affect the pre- wards. As a result, the anomalous low-level cyclone in the
cipitation anomaly over the ICP during the decaying spring decaying spring induced by La Niña shifts westwards with
through inducing the atmospheric circulation variation. The an abnormal center covering the whole ICP. This impedes
spring precipitation averaged over the ICP and the preced- the water vapor transporting to the Southern China and leads
ing winter ONI are closely correlated with a statistically to an abnormal moisture convergence over the ICP. Mean-
significant (p < 0.001) correlation coefficient of -0.47 during while, the resultant abnormal upward movement coverage
1958–2019. This indicates that there is usually excessive extends eastward and covers the whole ICP (including the

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2663

Fig. 14  Composite anomalies of spring vertical velocity averaged

over 12° N–17° N (contours; units: ­1e−2 Pa s−1; CI 0.2 e­ −2 Pa s−1; the
zero contours are not shown and the shaded areas are significant with
Fig. 13  Composite anomalies of spring water vapor flux (arrows;
p < 0.1) in the a El Niño cases and b La Niña cases. c The summation
­ −1 s−1; only arrows significant with p < 0.1 are shown)
units: kg m
of a and b. The dashed and continuous lines denote the negative and
integrated from 300 to 1000 hPa and its divergence (colors; units: ­1e5
positive values, respectively. The arrows denote the vertical motion
kg ­m−2 s−1; the areas within white lines are significant with p < 0.1)
directions. All data are linearly detrended
in the a El Niño cases and b La Niña cases. c The summation of a
and b. All data are linearly detrended. The red boxes denote the ICP
region
the decaying spring (Fig. 7c). Such a fundamental regional
Southeastern ICP). Thus, the positive spring precipitation difference in precipitation between the springs following El
anomaly in response to La Niña exhibits a relatively stronger Niño and La Niña is vitally important for the seasonal cli-
intensity and extends into the Southeastern ICP. mate prediction over the Southeastern ICP. The benefits of
In this study, we emphasize the asymmetric responses of improved seasonal climate prediction are enormous for more
the ICP spring precipitation to ENSO. In particular, while than one hundred million agricultural people living in the
there is no statistically significant anomaly of precipitation Southeastern ICP countries/regions (including the Eastern
over the Southeastern ICP during the decaying spring of Thailand, most of Cambodia, and the Southern Laos and
El Niño (Fig. 7a), La Niña often results in a strong posi- Vietnam).
tive precipitation anomaly over the Southeastern ICP for Owing to the strong persistence (memory), the spring soil
moisture anomalies over the ICP can exert great influences

13
2664 G. Li et al.

on both the mean (Gao et al. 2019, 2020a) and extreme (Gao An SI, Jin F-F (2004) Nonlinearity and asymmetry of ENSO. J Clim
et al. 2020c) climates over the Yangtze River basin for the 17:2399–2412
An S-I, Ham Y-G, Kug J-S, Jin F-F, Kang I-S (2005) El Niño–La Niña
following summer by influencing the local surface heating asymmetry in the coupled model intercomparison project simula-
and remote atmospheric circulation. On the other hand, the tions. J Clim 18:2617–2627
spring ICP precipitation variation is always considered to be Ashok K, Yamagata T (2009) Climate change: the El Nio with a
the primary driving factor for the local soil moisture anoma- difference. Nature 461:481–484
Chang CP, Wang Z, Ju J, Li T (2004) On the relationship between
lies (Gao et al. 2020a, b). Using the surface (0–10 cm) soil western maritime continent monsoon rainfall and ENSO during
moisture data from GLDAS V2.0 (Rodell et al. 2004), we northern winter. J Clim 17:665–672
further calculate the regional averaged soil moisture anoma- Chen H, Zhang Y (2013) Sensitivity experiments of impacts of large-
lies in spring over the ICP for the period of 1958–2012 (una- scale urbanization in East China on East Asian winter monsoon.
Chin Sci Bull 58:809–815
vailable for the data after 2012). The correlation coefficient Chen H, Zhang Y, Yu M, Hua W, Sun S, Li X, Gao C (2015) Large-
reaches 0.84 (p < 0.001) between the spring precipitation and scale urbanization effects on eastern Asian summer monsoon
soil moisture anomalies averaged over the ICP. This indi- circulation and climate. Clim Dyn 47:1–20
cates that about 70% of the changes in the surface spring soil Chou C, Lo M-H (2007) Asymmetric responses of tropical precipita-
tion during ENSO. J Clim 20:3411–3433
moisture over the ICP can be attributed to the local spring Dai A, Wigley TML (2000) Global patterns of ENSO-induced pre-
precipitation. Thus, the spring ICP precipitation can pro- cipitation. Geophys Res Lett 27:1283–1286
vide an important source of seasonal predictability of the De Bruyn M, Stelbrink B, Morley RJ, Hall R, Carvalho GR, Can-
East Asian summer monsoon and extreme climate events. non CH, van den Bergh G, Meijaard E, Metcalfe I, Boitani L
(2014) Borneo and Indochina are major evolutionary hotspots
The present findings suggest that the two phases of ENSO for Southeast Asian biodiversity. Syst Biol 63:879–901
have asymmetric effects on the ICP precipitation during the Ding Y, Chan JCL (2005) The East Asian summer monsoon: an
decaying spring. This would potentially increase the com- overview. Meteorol Atmos Phys 89:117–142
plexity of seasonal predictability of the East Asian summer Gao C, Chen H, Li G, Ma H, Li X, Long S, Xu B, Li X, Zeng X, Yan
H, Wang Z, Yang S (2019) Land-atmosphere interaction over
monsoon and extreme climate. Therefore, the present results the Indo-China Peninsula during spring and its effect on the
are also conducive to a better understanding of seasonal pre- following summer climate over the Yangtze River basin. Clim
dictions for the summer mean and extreme climates over Dyn 53:6181–6198
East Asia. Gao C, Li G, Chen H, Yan H (2020a) Inter-decadal change in the
effect of spring soil moisture over the Indo-China Peninsula
El Niño exists multiple SST anomaly patterns. For exam- on the following summer precipitation over the Yangtze River
ple, there are eastern Pacific (EP) and central Pacific (CP) basin. J Clim 33:7063–7082
El Niño events, with abnormal SST centers in the equatorial Gao C, Li G, Xu B (2020b) Weakening influence of spring soil mois-
EP and CP, respectively (Ashok and Yamagata 2009; Kao ture over the Indo-China Peninsula on the following summer
Meiyu front and precipitation extremes over the Yangtze River
and Yu 2009). Such a difference in SST anomaly patterns basin. J Clim 33:10055–10072
can exert different effects on the Asian climate during their Gao C, Li G, Xu B, Li X (2020c) Effect of spring soil moisture over
decaying phases, which has attracted increasing attention the Indo-China Peninsula on the following summer extreme
(Gao et al. 2020d; Luo and Lau 2020). While the present precipitation events over the Yangtze River basin. Clim Dyn
54:3845–3861
study focuses on the conventional EP El Niño, it is also Gao T, Luo M, Lau NC, Chan TO (2020d) Spatially distinct effects of
important to check whether/how the CP and EP El Niño two El Niño types on summer heat extremes in China. Geophys
events can cause the different effects on the spring ICP pre- Res Lett 47:e2020GL086982
cipitation, since our findings highlight the effects of different Ge F, Zhi X, Babar ZA, Tang W, Chen P (2017) Interannual variability
of summer monsoon precipitation over the Indochina Peninsula
patterns of the El Niño and La Niña events. This could be in association with ENSO. Theor Appl Climatol 128:523–531
thoroughly investigated in the future. Guan C, McPhaden MJ, Wang F, Hu S (2019) Quantifying the role
of oceanic feedbacks on ENSO asymmetry. Geophys Res Lett
Acknowledgements This work was supported by the Fundamen- 46:2140–2148
tal Research Funds for the Central Universities (B210201015), Hendon HH (2003) Indonesian rainfall variability: impacts of ENSO
the National Key Research and Development Program of China and local air–sea interaction. J Clim 16:1775–1790
(2018YFC1506002), the Natural Science Foundation of China Hoerling MP, Kumar A, Min Z (1997) El Niño, La Niña, and the non-
(41831175, 41905054, 41861144013, and 41975097), and the China linearity of their teleconnections. J Clim 10:1769–1786
Postdoctoral Science Foundation (2020T130168 and 2019M651665). Huang G, Yan Z (1999) The East Asian summer monsoon circula-
tion anomaly index and its interannual variations. Chin Sci Bull
44:1325–1329
Hu K, Huang G (2010) The formation of precipitation anomaly patterns
References during the developing and decaying phases of ENSO. Atmos Ocean
Sci Lett 3:25–30
Alexander MA, Bladé I, Newman M, Lanzante JR, Lau NC, Scott JD Jiang F, Zhang W, Geng X, Stuecker MF, Liu C (2019) Impacts of central
(2002) The atmospheric bridge: the influence of ENSO telecon- Pacific El Niño on southern China spring precipitation controlled by
nections on air–sea interaction over the global oceans. J Clim its longitudinal position. J Clim 32:7823–7836
15:2205–2231

13
Inter-annual variability of spring precipitation over the Indo-China Peninsula and its… 2665

Jin Q, He J, Chen L, Zhu C (2006) Impact of ocean-continent distribution Sasaki W, Doi T, Richards KJ, Masumoto Y (2014) The influence of
over Southern Asia on the formation of summer monsoon. Acta ENSO on the equatorial Atlantic precipitation through the Walker
Meteorol Sin 20:95–108 circulation in a CGCM. Clim Dyn 44:191–202
Juneng L, Tangang FT (2005) Evolution of ENSO-related rainfall anoma- Takahashi HG, Yasunari T (2006) A climatological monsoon break in
lies in Southeast Asia region and its relationship with atmosphere– rainfall over Indochina—a singularity in the seasonal march of the
ocean variations in Indo-Pacific sector. Clim Dyn 25:337–350 Asian summer monsoon. J Clim 19:1545–1556
Ju J, Slingo J (1995) The Asian summer monsoon and ENSO. Q J R Tanaka N, Kume T, Yoshifuji N, Tanaka K, Takizawa H, Shiraki K, Tan-
Meteorol Soc 121:1133–1168 tasirin C, Tangtham N, Suzuki M (2008) A review of evapotran-
Kanae S, Oki T, Musiake K (2001) Impact of deforestation on regional spiration estimates from tropical forests in Thailand and adjacent
precipitation over the Indochina Peninsula. J Hydrometeorol regions. Agric Forest Meteorol 148:819
2:51–70 Tao W, Huang G, Hu K, Gong H, Wen G, Liu L (2016) A study of biases
Kao H-Y, Yu J-Y (2009) Contrasting Eastern-Pacific and Central-Pacific in simulation of the Indian Ocean basin mode and its capacitor effect
types of ENSO. J Clim 22:615–632 in CMIP3/CMIP5 models. Clim Dyn 46:205–226
Kobayashi S, Ota Y, Harada Y, Ebita A, Moriya M, Onoda H, Onogi Timmermann A, An S-I, Kug J-S, Jin F-F, Cai W, Capotondi A, Cobb
K, Kamahori H, Kobayashi C, Endo H, Miyaoka K, Takahashi K KM, Lengaigne M, McPhaden MJ, Stuecker MF, Stein K, Witten-
(2015) The JRA-55 reanalysis: general specifications and basic char- berg AT, Yun K-S, Bayr T, Chen H-C, Chikamoto Y, Dewitte B,
acteristics. J Meteorol Soc Jpn Ser II 93:5–48 Dommenget D, Grothe P, Guilyardi E, Ham Y-G, Hayashi M, Ineson
Lau N-C, Nath MJ (2003) Atmosphere–ocean variations in the Indo- S, Kang D, Kim S, Kim W, Lee J-Y, Li T, Luo J-J, McGregor S,
Pacific sector during ENSO episodes. J Clim 16:3–20 Planton Y, Power S, Rashid H, Ren H-L, Santoso A, Takahashi K,
Lin L, Chen C, Luo M (2018) Impacts of El Niño-Southern Oscillation Todd A, Wang G, Wang G, Xie R, Yang W-H, Yeh S-W, Yoon J,
on heat waves in the Indochina peninsula. Atmos Sci Lett 19:e856 Zeller E, Zhang X (2018) El Niño-Southern Oscillation complexity.
Luo M, Lau N-C (2019) Amplifying effect of ENSO on heat waves in Nature 559:535–545
China. Clim Dyn 52:3277–3289 Wallace JM, Rasmusson EM, Mitchell TP, Kousky VE, Sarachik ES,
Luo M, Lau N-C (2020) Summer heat extremes in northern continents Storch HV (1998) On the structure and evolution of ENSO-related
linked to developing ENSO events. Environ Res Lett 15:074042 climate variability in the tropical Pacific: lessons from TOGA. J
Lyon B, Barnston AG (2005) ENSO and the spatial extent of interannual Geophys Res Oceans 103:14241–14259
precipitation extremes in tropical land areas. J Clim 18:5095–5109 Wang B, Zhang Q (2002) Pacific-East Asian teleconnection. Part II: how
Mandapaka PV, Qin X, Lo EYM (2017) Analysis of spatial patterns of the Philippine Sea anomalous anticyclone is established during El
daily precipitation and wet spell extremes in Southeast Asia. Int J Niño development. J Clim 15:3252–3265
Climatol 37:1161–1179 Wang B, Wu R, Fu X (2000) Pacific-East Asian teleconnection: how does
Ma J, Yang S, Wang Z (2018) Influence of spring soil moisture anomaly ENSO affect East Asian climate? J Clim 13:1517–1536
in the Indo-China peninsula on the establishment and development Wang C, Deser C, Yu J-Y, DiNezio P, Clement A (2017) El Niño
of Asian tropical summer monsoon. Meteorol Environ Sci 41:19–30 and Southern Oscillation (ENSO): a review. Coral reefs of
(in Chinese) the world. Springer, Netherlands, pp 85–106. https​ : //doi.
Neelin JD, Battisti DS, Hirst AC, Jin F-F, Wakata Y, Yamagata T, Zebiak org/10.1007/978-94-017-7499-4_4
SE (1998) ENSO theory. J Geophys Res Oceans 103:14261 Wen Z, He H, Huang R (2004) The influence of 30–60 days oscillation
Ng CHJ, Vecchi GA, Muñoz ÁG, Murakami H (2019) An asymmetric on the development of the South China Sea summer monsoon. Acta
rainfall response to ENSO in East Asia. Clim Dyn 52:2303–2318 Oceanol Sin 23:569–579
Okumura YM, Deser C (2010) Asymmetry in the duration of El Niño and Wen Z, Liang Z, Wu L (2006) The relationship between the Indian Ocean
La Niña. J Clim 23:5826–5843 sea surface temperature anomaly and the onset of the South China
Philander SGH (1983) El Niño/southern oscillation phenomena. Nature Sea summer monsoon II: analyses of mechanisms. Chin J Atmos Sci
302:295–301 30:1138–1146 (in Chinese)
Philander SG (1990) El Niño, La Niña, and the Southern Oscillation. Xie S-P, Kosaka Y, Du Y, Hu K, Chowdary JS, Huang G (2016) Indo-
Academic, New York Western Pacific Ocean capacitor and coherent climate anomalies in
Power SB, Smith IN (2007) Weakening of the Walker Circulation and post-ENSO summer: a review. Adv Atmos Sci 33:411–432
apparent dominance of El Niño both reach record levels, but has Yang K, Zhang J, Wu L, Wei J (2019) Prediction of summer hot extremes
ENSO really changed? Geophys Res Lett 34:L18702 over the middle and lower reaches of the Yangtze River valley. Clim
Rasmusson EM, Wallace JM (1983) Meterological aspects of the El Niño/ Dyn 52:2943–2957
Southern Oscillation. Science 222:1195–1202 Zhang Y, Qian Y (2002a) Characteristics of surface temperature varia-
Rayner NA, Parker DE, Horton EB, Folland CK, Alexander LV, Row- tions in the SCS monsoon region and connections with the monsoon
ell DP (2003) Global analyses of sea surface temperature, sea ice, onset. J Nanjing Inst Meteorol 25:192–198 (in Chinese)
and night marine air temperature since the late nineteenth century. Zhang Y, Qian Y (2002b) Mechanism of thermal features over the Indo-
J Geophys Res 108:D144407 China Peninsula and possible effects on the onset of the South China
Rodell M, Houser PR, Jambor U, Gottschalck J, Mitchell K, Meng CJ, Sea monsoon. Adv Atmos Sci 19:885–900
Arsenault K, Cosgrove B, Radakovich J, Bosilovich M, Entin JK,
Walker JP, Lohmann D, Toll D (2004) The global land data assimila- Publisher’s Note Springer Nature remains neutral with regard to
tion system. Bull Am Meteorol Soc 85:381–394 jurisdictional claims in published maps and institutional affiliations.
Ropelewski CF, Halpert MS (1987) Global and regional scale precipita-
tion patterns associated with the El Niño/Southern Oscillation. Mon
Weather Rev 115:1606–1626
Ropelewski CF, Halpert MS (1989) Precipitation patterns associated with
the high index phase of the Southern Oscillation. J Clim 2:268–284

13