-
1 3
DOI 10.1007/s00382-014-2227-0Clim Dyn (2015) 44:807825
State of the tropical Pacific Ocean and its enhanced impact on
precipitation over East Asia during marine isotopic stage 13
M. P. Karami N. Herold A. Berger Q. Z. Yin H. Muri
Received: 28 August 2013 / Accepted: 19 June 2014 / Published
online: 16 July 2014 Springer-Verlag Berlin Heidelberg 2014
associated teleconnections with the extra-tropics favored
increased precipitation over the EASM. As compared to PrI, it is
found that the summer (JuneJulyAugust) sea surface temperature
(SST) is warmer in the eastern tropi-cal Pacific Ocean and colder
to the west. In concert with previous studies, we show that colder
summer SSTs in the central tropical Pacific during MIS-13 promotes
an upper-level teleconnection between the tropical Pacific Ocean
and EASM. It also contributes to the strengthening of the north-ern
Pacific subtropical high and, therefore, the transport of more
moisture into the EASM. We suggest that the reduced eastwest SST
difference in the tropical Pacific in summer helps to maintain the
teleconnection between the tropical Pacific and EASM. The
correlation between tropical Pacific SSTs and the EASM was higher
in our MIS-13 simula-tions, further supporting the enhancement of
their relation-ship. It is found that the pure impact of El Nio
Southern Oscillation on EASM precipitation increases by up to 30 %
in MIS-13 for HadCM3 while it is minor for CCSM3. Bet-ter
constraining the spatio-temporal variability of tropical Pacific
SST during the interglacials may thus help explain the anomalously
strong EASM during MIS-13 which has been observed from geological
records.
Keywords Paleoclimate modeling MIS-13 ENSO Teleconnection East
Asian summer monsoon
1 Introduction
Given the Earths current interglacial state, past intergla-cial
periods provide good candidates for studying potential climate
change. Quaternary interglacials exhibited changes in atmospheric
and oceanic circulations and are subject to high spatio-temporal
resolution data archives, making them
Abstract Multiple terrestrial records suggest that marine
isotopic stage 13 (MIS-13), an interglacial period approxi-mately
0.5 million years ago, had the strongest East Asian summer monsoon
(EASM) of the last one million years. This is unexpected given
that, compared to other intergla-cials, MIS-13 was globally cooler
with a lower CO2 con-centration. We use two coupled atmosphereocean
general circulation models, the Hadley Centre Coupled Model,
version 3 (HadCM3) and Community Climate System Model, version 3.0
(CCSM3), to simulate the climate of MIS-13 forced with different
insolation and greenhouse gas concentrations relative to the
pre-industrial (PrI) situ-ation. Both models confirm a stronger
EASM during MIS-13 compared to PrI. Here we specially focus on
analyzing the impact of the tropical Pacific Ocean on the EASM. Our
simulations suggest that the mean climatic state in the tropical
Pacific during MIS-13 was La Nia-like and that
M. P. Karami (*) N. Herold A. Berger Q. Z. Yin H. Muri Georges
Lematre Centre for Earth and Climate Research (TECLIM), Earth and
Life Institute (ELI), Universit Catholique de Louvain, Place Louise
Pasteur 3, Box L4.03.08, 1348 Louvain-La-Neuve, Belgiume-mail:
[email protected]
Present Address: M. P. Karami Geotop, Universit du Qubec Montral
(UQAM), Montreal, Canada
N. Herold Institute for the Study of Earth, Oceans and Space,
University of New Hampshire, Durham, NH 03824, USA
H. Muri Department of Geosciences, Meteorology and Oceanography
Section, University of Oslo, Blindern, Postboks 1022, 0315 Oslo,
Norway
-
808 M. P. Karami et al.
1 3
ideal candidates for investigating the sensitivity of climate to
forcings.
The interglacial stage of ~0.5 million years ago, identi-fied as
marine isotopic stage 13 (MIS-13), was character-ized by a
relatively low CO2 concentration (Luthi et al. 2008), cool
Antarctic temperatures (Jouzel et al. 2007) and high benthic 18O
values (e.g., Lisiecki and Raymo 2005) related to higher global ice
volume and/or colder deep-ocean temperatures. However, it featured
the strongest East Asian summer monsoon (EASM) of the last one
million years, an insight derived mainly from loess records from
northern China (e.g. Kukla et al. 1990; Guo et al. 1998) and
paleosols from southern China (Yin and Guo 2006). The strongest
EASM in MIS-13 was proposed as a paradox (Yin and Guo 2008) given
that CO2 and CH4 concentra-tions were relatively low during MIS-13
and its insolation is not abnormal as compared to other
interglacials. Strong African and Indian monsoons during MIS-13
were also suggested based on marine sediments from the equatorial
Indian Ocean (Bassinot et al. 1994) and the Mediterranean Sea
(Rossignol-Strick et al. 1998).
This study investigates the role of the tropical Pacific Ocean
in enhancing the EASM during MIS-13. In previous model studies of
MIS-13, the impacts of insolation, green-house gas (GHG)
concentrations and ice sheets were inves-tigated (Yin et al. 2008,
2009; Sundaram et al. 2012; Muri et al. 2012, 2013). The first
model simulations of MIS-13 climate were performed by Yin et al.
(2008, 2009) using an Earth system model of intermediate
complexity, LOVE-CLIM. They found that strong summer insolation in
the Northern Hemisphere (NH) is chiefly responsible for the more
intense EASM of MIS-13 compared to pre-industrial (PrI). However,
model comparison between MIS-13 and all the other interglacials (in
particular MIS-5e) showed that insolation alone did not make the
MIS-13 EASM excep-tionally strong (Yin and Berger 2012). Therefore,
the role of other factors was also needed to be investigated. The
presence of a Eurasian ice sheet further increased (by 5 %) the
precipitation in the EASM region through a topograph-ically-induced
atmospheric wave train (Yin et al. 2008, 2009). Whether there were
in fact additional ice sheets in the NH during MIS-13 is still
unclear from geological evidence (Guo et al. 2009). The results of
Yin et al. (2008, 2009) have since been confirmed by general
circulation models (Sundaram et al. 2012; Muri et al. 2012,
2013).
Based on hydrographic reconstructions from the South China Sea,
Yu and Chen (2011) conjectured the importance of the tropical
dynamics and the eastwest sea surface tem-perature (SST) gradient
of the Equatorial Pacific in MIS-13 climate. In the modern climate,
the tropical Pacific Ocean is an important component of climate
variability. The El Nio-Southern Oscillation (ENSO) is the dominant
source of interannual variability in the tropical Pacific Ocean
and
strongly influences global precipitation (Ropelewski and Halpert
1987) and the EASM (e.g., Chang et al. 2000) through changing
atmospheric circulation and teleconnec-tions (e.g., Hoskins and
Karoly 1981). El Nio (La Nia), which is the positive (negative)
phase of ENSO, is driven by significantly warm (cold) SST in the
equatorial eastern Pacific Ocean in winter. Mean summer
precipitation over China usually increases after onset of the El
Nio (Shen and Lau 1995; Chang et al. 2000). Moreover, the mean
climatic state of the tropical Pacific varies on interdecadal to
millennial time scales, which modifies the atmospheric
teleconnections accompanying ENSO events (Mller and Roeckner 2008)
and influences EASM precipitation. For instance, changes in the
ENSO-EASM relation on interdec-adal time scales (e.g., Wu and Wang
2002; Lee et al. 2008), was linked to the interdecadal changes in
the background state of the tropical Pacific and Indian Oceans
(Chang et al. 2000). At precessional frequencies, El Nio- or La
Nia- like configurations are found in the mean-state of the
tropi-cal Pacific depending on the time of perihelion (Clement et
al. 1999).
The atmospheric-oceanic interaction in the tropical Pacific
which was very likely to contribute to a stronger EASM in MIS-13
has not been explored. We use two fully coupled general circulation
models, Hadley Centre Cou-pled Model, version 3 (HadCM3) and
Community Cli-mate System Model, version 3.0 (CCSM3), to explore
the ocean and atmosphere interactions in the tropical Pacific and
their teleconnection with the EASM during MIS-13. This work goes
beyond previous studies of MIS-13 (e.g., Muri et al. 2013) by
showing that the changes in the mean state as well as in the
interannual variability of the tropi-cal Pacific increased the
ENSOEASM relationship and contributed to the higher EASM
precipitation during this period. In addition to unraveling the
causes for the climate of MIS-13, this study also contributes to
our understand-ing of the relationship between the tropical Pacific
Ocean and East Asian monsoon under different CO2 concentra-tion and
insolation.
This paper is organized as follows: Sect. 2 briefly describes
the models and boundary conditions. In Sect. 3, the difference in
the mean climate between MIS-13 and PrI, and the subsequent
difference in the monsoons are investigated. This will be followed
by our proposed mecha-nism concerning the enhanced relationship
between the tropical Pacific and EASM in MIS-13. In Sect. 4, we
will analyze the interannual variability around the mean state of
each model experiment and will quantify the contribution of ENSO
events to the enhancement of the EASM. Lastly, Sect. 5 draws
conclusions concerning the role of tropical Pacific Ocean in
enhancement of the EASM and the often disregarded relationship
between mean climate and interan-nual variability.
-
809State of the tropical Pacific Ocean
1 3
2 Model descriptions
Full descriptions of each models configuration is presented in
Muri et al. (2012) for HadCM3 and Herold et al. (2012) for CCSM3.
Here, we summarize the main information of the models and the
boundary conditions. The HadCM3 is a fully coupled atmosphereocean
general circulation model (AOGCM). Its atmospheric component has 19
verti-cal levels and horizontal resolution of 3.75 2.5 (lon-gitude
latitude). The oceanic component has 20 vertical layers with 1.25
1.25 horizontal resolution. Documen-tation of the model are
explained in e.g. Pope et al. (2000) and Gordon et al. (2000).
HadCM3 produces ENSO with a period in 34 year band (Collins et al.
2001), and has a realistic representation of the monsoons (Turner
and Slingo 2009). The CCSM3 (Collins et al. 2006) is an AOGCM of a
similar generation to HadCM3, both of which were a part of the
Fourth Assessment Report by the IPCC. The atmospheric component of
the CCSM3 represents 26 vertical levels with a horizontal T31
spectral resolution (~3.75 3.75). The ocean component represents 25
ver-tical levels with a nominal ~31 horizontal resolution. The
version of the CCSM3 used in this study simulates ENSO variability
with as much skill as higher resolution versions of the model
(Yeager et al. 2006), though exhibits too high a frequency compared
to observations (Deser et al. 2006).
Two simulations were performed for each model; a PrI and a
MIS-13 simulation. Only the climate response to insolation and GHGs
is analyzed while ice sheets are kept at their present-day states.
The largest difference between the PrI and MIS-13 simulations are
their astronomical con-figurations (Table 1). MIS-13 has a larger
eccentricity and its NH summer occurred at perihelion, which leads
to a higher summer insolation in the NH compared to our PrI
simulations [the summer solstice daily insolation at 65N is 50 W m2
(10 %) larger during MIS-13 than PrI]. This dif-ference in
insolation forcing induces significant changes in the climate
system as shown in previous studies (Yin et al. 2008; Muri et al.
2013) as well as in this study.
For HadCM3, the PrI control run is for the year 1850, and to
simulate the MIS-13 climate the astronomical parameters and GHG
concentrations at 506 ka BP are used (Table 1; Muri et al. 2013).
Both HadCM3 experiments
have the same insolation and GHG as Yin et al. (2008) where the
justification of the forcing was explained. The HadCM3 experiments
were run for 800 years. The PrI CCSM3 simulation is that of Herold
et al. (2012) and the MIS-13 CCSM3 simulation was set up in an
identical fashion to the interglacial simulations in that study.
The PrI simulation of CCSM3 was run for 1,300 years (thus a total
length of 1,300 years) and the MIS-13 one was ini-tiated from the
PrI simulation at year 500 (of 1,300) and run for 1,000 years. The
CCSM3 MIS-13 experiment has the same astronomical parameters as the
HadCM3 one, but has slightly different GHG concentrations (Table
1). This is because the HadCM3 MIS-13 experiment has followed the
strategy of Yin et al. (2008) where the average of GHG
concentrations over MIS-13 was used, whereas the CCSM3 MIS-13
experiment has followed the strategy of Yin and Berger (2012) where
interglacial climates were simulated under peak forcings. This
slight difference in GHG concen-trations between the two models is
equivalent to ~0.13 W/m2 radiative forcing (Myhre et al. 1998) and
thus would not likely change our conclusions.
Given that our focus is on the potential role of the tropi-cal
Pacific in enhancing EASM, a comprehensive inter-model comparison
is not done. However, recent studies provide comprehensive
overviews of various models skill in reproducing interglacial
climate (e.g., Lunt et al. 2013). The HadCM3 and CCSM3
underestimate high latitude warmth and thus also misrepresent the
equator to pole tem-perature gradient based on Lunt et al. (2013).
HadCM3 however does a better job than CCSM3 in reproducing
interglacial temperatures.
3 Climatology of MIS13
3.1 Annual/summer atmosphere and ocean MIS-13 climatologies
In both HadCM3 and CCSM3, the annually-averaged global SST of
MIS-13 is generally found to be lower than PrI. This cooling is
mainly due to its lower CO2 concentration (Yin and Berger 2012).
The strongest cooling particularly occurs in the tropical Pacific
and the North Atlantic Ocean for HadCM3, and in the western
Table 1 Astronomical parameters (Berger 1978) and greenhouse gas
concentrations used for simulating the PrI and MIS-13 climate in
HadCM3 and in CCSM3
Experiment Obliquity () Eccentricity Longitude of perihelion ()
CH4 (ppb) N2O (ppb) CO2 (ppm)
PrI 23.446 0.016724 102.04 760 270 280HadCM3 MIS-13 23.377
0.034046 274.05 510 280 240CCSM3 MIS-13 23.377 0.034046 274.05 508
258 247
-
810 M. P. Karami et al.
1 3
North Pacific and North Atlantic Ocean for CCSM3 (Fig. 1). In
the tropical Pacific, both models show cool-ing, though this is
stronger in HadCM3. The eastern tropical Pacific Ocean (250280E)
shows a smaller decrease in temperature relative to the central and
west-ern tropical Pacific. One possible explanation for this is the
westward shift of the Pacific cold tongue (where the SST is minimum
along the tropical Pacific) which is further discussed in Sect.
3.2. Also, the presence of strong ocean dynamics (e.g., upwelling)
in the eastern tropical Pacific Ocean has been known to cause
different responses to climate change relative to the western
tropi-cal Pacific (Clement et al. 1996).
Figure 2 shows the modeled precipitation for NH sum-mer
(JuneJulyAugust; JJA) as our particular interest is the summer
monsoon. The PrI values of the two models (Fig. 2a, b) are
qualitatively in agreement although there are differences in the
tropics and East Asia. For the differ-ence in precipitation between
MIS-13 and PrI, we focus on eastern China where most EASM proxy
data were collected, i.e., approximately 2040N and 100120E (e.g.,
Yin and Guo 2008; their Fig. 2). The MIS-13 JJA precipitation
increases in both models in MIS-13 except for central-eastern China
in the CCSM3. The increase in EASM precipitation is larger in the
HadCM3 and cov-ers eastern China in agreement with proxy data.
Overall, HadCM3 seems to perform better than CCSM3 in captur-ing
increased precipitation throughout the EASM region when compared to
the data of Yin and Guo (2008). For the modern simulations of EASM,
it was also found that the HadCM3 reproduced the EASM well (Lei et
al. 2013) while CCSM3 had deficient rainfall over EASM (Meehl et
al. 2006).
The position and strength of the western Pacific subtrop-ical
high is known to play a dominant role in the variability and
distribution of EASM precipitation (Zhou et al. 2008 and references
therein). In our MIS-13 simulations, JJA sea level pressure (SLP)
values for both models exhibit west-ward extension and
strengthening of the western Pacific subtropical high and deepening
of Asian low (Fig. 3). Such a change is in favor of increasing the
monsoon intensity by bringing more moisture from the Indian and
Pacific Oceans to East Asia (Fig. 3). The total moisture
transported to East Asia in our PrI simulations mainly originates
from the Indian Ocean and South China Sea (Fig. 3a, c). In CCSM3
MIS-13, it can be seen that the moisture coming from the northern
Pacific Ocean into EASM is increased compared to CCSM3 PrI (Fig.
3b). In the case of HadCM3 MIS-13, the additional moisture comes
not only from the northern Pacific Ocean, but also from the Indian
Ocean and South China Sea. More moisture from the Pacific Ocean
dur-ing MIS-13 suggests an increase in its contribution to the EASM
during MIS-13.
3.2 La Nia-type mean climate of the tropical Pacific
All the fields in this section are presented as annual mean
values averaged over 5S5N.
The cooling of the tropical Pacific SSTs (as shown in Sect. 3.1)
is not uniform along the equator and is largest in the
eastern-central equatorial Pacific Ocean around 240E (Fig. 4a). The
Pacific cold tongue has shifted westward in our MIS-13 simulations
by ~10 compared to PrI ones. The difference in SST (annual mean)
between the cold tongue and the western tropical Pacific shows an
increase in MIS-13 for both models. However, the difference in SST
between the easternmost and western tropical Pacific Ocean is
decreased in MIS-13. This is important to con-sider when
interpreting proxy SST records from the tropi-cal Pacific (cf.,
Mohtadi et al. 2006).
The magnitude of the zonal wind stress during MIS-13 has
decreased in the eastern tropical Pacific (230280E) compared to
PrI, while it has increased to the west (150230E; Fig. 4b).
The mean thermocline depth, where the largest vertical oceanic
temperature gradient occurs, is shown in Fig. 4c. The thermocline
along the equator is typically approximated by the depth of the 20
C isotherm (e.g., Merkel et al. 2010), which was found to be a good
approximation for our MIS-13 simu-lations and thus adopted here.
Compared to the PrI, MIS-13 in CCSM3 shows an overall shoaling of
the thermocline with a reduced eastwest slope. In HadCM3, the
MIS-13 thermo-cline is shallower in the east, deeper in the west
and, therefore, has steeper tilt. The enhanced vertical velocity in
the eastern tropical Pacific (not shown) also indicates a stronger
upwelling consistent with the shoaling of the thermocline in both
mod-els. When the eastwest tilt of the thermocline is compared
between MIS-13 and PrI, its relation with the corresponding zonal
wind stress is non-linear. Between 230E and 280E, the shallower
thermocline does not follow the weakened zonal wind stress. This is
likely to be related to the stronger meridi-onal wind (Fig. 12 in
Appendix) in the eastern equatorial Pacific (mainly south of the
equator) which causes stronger upwelling through Ekman pumping and,
therefore, a shal-lower thermocline. The equatorial Pacific
thermocline com-puted in both PrI and MIS-13, is deeper in HadCM3
than in CCSM3 although HadCM3 exhibits stronger zonal and
merid-ional winds. The difference in the thermocline of the two
mod-els might be an ocean-related process (e.g., Timmermann et al.
2005) and/or due to the difference in the vertical resolution of
the ocean component of the models.
The sea-level pressure differences between our MIS-13 and PrI
simulations (in both models) show a positive and negative anomaly
in the eastern and western tropical Pacific, respectively (Fig. 13
in Appendix). Regarding the Walker circulation, the atmospheric
vertical velocity at the equator versus longitude shows the
upwelling and
-
811State of the tropical Pacific Ocean
1 3
downwelling branches (Fig. 4d; note that negative velocity is
upward). HadCM3 MIS-13 shows westward shift (west-ward extension)
of the convection center in the western tropical Pacific and
smaller vertical velocity in the west-ern Pacific (150170E). CCSM3
MIS-13, however, shows increased convection in the western Pacific
(150180E) and no westward extension of the Walker circulation.
The shoaling of the thermocline in the eastern tropical Pacific,
the presence of the cold SST anomaly in the cen-tral tropical
Pacific, the westward shift of the cold tongue and Walker
circulation, and stronger zonal wind stress in the central
equatorial Pacific all indicate that MIS-13 was subject to a La
Nia-like mean climatic state, relative to PrI conditions.
3.3 Strengthened relationship between tropical Pacific SST and
East Asian monsoon
Higher summer insolation and the subsequent increase of
landocean thermal contrast (i.e., between East Asia and Western
Pacific Ocean) in MIS-13 was found as an impor-tant factor for the
enhancement of EASM precipitation (Yin et al. 2008). We find that
MIS-13 landocean thermal con-trast reaches its maximum in June, and
starts to decrease in July and August (not shown). This is while,
EASM precipi-tation of MIS-13, continues to increase in July and
August, and has its maximum in July for CCSM3 and in August for
HadCM3 (Fig. 5d). This suggests that other factors, next to the
landocean thermal contrast, could have also played a role in MIS-13
to explain the EASM precipitation evolu-tion in June, July and
August. The changes in the tropical Pacific SST, is seen a good
candidate.
The SST difference between MIS-13 and PrI shows a warming
anomaly in the eastern equatorial Pacific and a cooling one in the
central equatorial Pacific during summer (JJA) which coincides with
the increase in EASM precipi-tation (Fig. 5a, b). These SST
anomalies affect the atmos-pheric circulation by modifying the
tropical convection. Summer-time cooling in the central tropical
Pacific was suggested to have a critical role in maintaining the
western Pacific subtropical high and modifying the EASM
precipi-tation (Fan et al. 2013). The SST difference between MIS-13
and PrI has larger seasonal variability in the eastern equatorial
Pacific (260280E) than in the western equato-rial Pacific (140160E;
Fig. 5a, b) which also affects the tropical convection. Such
behavior is normally linked to the higher asymmetry in the
distribution of SSTs on the north-ern and the southern sides of the
eastern equatorial Pacific (Li and Philander 1996; Thuburn and
Sutton 2000).
Changes in the eastwest SST gradient in the tropical Pacific has
been suggested to have a larger influence on the tropical
convection than a uniform zonal cooling/warm-ing (Yin and Battisti
2001; Chiang 2009). The eastwest SST difference in MIS-13 (Fig. 5c)
is smaller than in PrI
Fig. 1 Annually-averaged sea surface temperature (SST; C). Left
column is CCSM3 and right one is HadCM3. a and b are PrI val-ues, c
and d are the difference between MIS-13 and PrI. The colored-
shaded areas in c and d are significant at the 95 % confidence
level based on the Students t test and the white-shaded areas are
non-sig-nificant values
-
812 M. P. Karami et al.
1 3
from May to October in CCSM3 (the maximum decrease is reaching 1
C), and from June to December in HadCM3 (the maximum decrease is
reaching 3 C). At around the same period, the EASM precipitation
(averaged over 2040N and 100120E) is found to be higher in MIS-13,
especially in HadCM3 (Fig. 5d). Accordingly, less winter
precipitation coincides with the enhanced eastwest SST difference
in MIS-13. Hence, the eastwest tropical Pacific SST gradient and
EASM precipitation are found to be anti-correlated (around 0.8 with
95 % significance) in MIS-13 which is not the case in PrI. Thus the
teleconnection between the tropical Pacific and EASM was present
dur-ing MIS-13 though not in PrI. We propose that the reduced
eastwest SST difference and the cooling in the central tropical
Pacific Ocean in MIS-13 are contributing factors in providing more
rainfall over EASM through maintaining the teleconnection between
the tropical Pacific and EASM as will be discussed in Sect.
3.4.
3.4 Teleconnection between tropical Pacific and East Asian
monsoon
Changes in the tropical Pacific convection affect the
atmos-pheric circulation in the extra-tropical regions through
atmospheric teleconnections (e.g., Seager et al. 2010) which can
vary with the seasonal cycle (Alexander et al. 2002). Atmospheric
teleconnections during modern ENSO (El Nio and La Nia) events are
normally attributed to the Ross by wave train generated due to
changes in tropical Hadley circulation (Trenberth et al. 1998) or
tropical dia-batic heating (De Weaver and Nigam 2004). These
wave
(a)
(c)
(e) (f)
(d)
(b)
Fig. 2 Summer (June-JulyAugust) precipitation (mm/day). Left
column is CCSM3 and right one is HadCM3. a and b are PrI values, c
and d are the difference between MIS-13 and PrI, e and f are also
the difference between MIS-13 and PrI but zoomed into the
East-Asian region
Fig. 3 Summer (June-JulyAugust) sea-level pressure superim-posed
on the vertically-integrated moisture transport. Left column is
CCSM3 and right one is HadCM3. a and b are PrI values, c and d are
the difference between MIS-13 and PrI. Please note the different
color legends for the isobars in c and d
-
813State of the tropical Pacific Ocean
1 3
-
814 M. P. Karami et al.
1 3
trains which are excited towards higher latitudes modify the
westerlies and are an important component of the tel-econnection
between the tropics and sub-tropics.
To assess the change in teleconnection between the tropi-cal
Pacific and EASM during MIS-13, we analyze the zonal eddy wind
field anomalies (differences between MIS-13 and PrI) at 250 mb. The
eddy wind represents the stationary waves (Mller and Roeckner
2008), and its anomaly shows for instance changes in the wave
amplitude and ENSO-related wave trains (e.g., De Weaver and Nigam
2004). We will mainly focus on the summer (JJA) teleconnection as
both of our models show excess precipitation during these
months.
The positive summer SST anomaly in the eastern tropi-cal Pacific
and the negative anomaly in the western tropi-cal Pacific (Fig. 5a,
b) slows down the Walker circulation by causing anomalous ascending
in the east and anomalous descending in the west, respectively
(Fig. 5 e, f; note that the negative velocity is upward). It should
be noted that the maxi-mum weakening of the Walker circulation in
MIS-13 relative to PrI occurs mainly around the summer season
consistent with the reduced eastwest SST difference. The subsequent
change in the vertical velocity and convection is transferred to
the middle and upper troposphere through a Gill-type effect (Gill
1980). In the Gill response, two anticyclones (cyclones) form to
the west of a warm (cold) anomaly in the middle-upper troposphere,
straddling the equator. Although our mod-els are more complex than
Gills simple model, some of these theoretically-predicted features
can be seen by comparing the
tropical Pacific wind anomaly (Fig. 6) with the correspond-ing
SST anomalies (Fig. 5). We highlight the cyclones (C) straddling
the equator by drawing schematic circles. The Gill-like cyclones
which are positioned on both sides of the equa-tor and extend to 20
north and south (Fig. 6) correspond to the cold SST anomaly in the
eastern-central tropical Pacific (Fig. 5). The anomalous easterlies
on the northern flank of the cyclone C1 interact with the westerly
jet and its meander which in return reinforces the teleconnection
between the tropical Pacific and other regions of the globe. This
interac-tion promotes the upper-level anomalous cyclone C2 in the
western north Pacific that connects the tropical Pacific to East
Asia. We therefore suggest that the tropical Pacific in MIS-13 had
a larger share or at least played a modulating role in providing
moisture over East Asia through a summer telecon-nection between
the EASM and the tropical Pacific. Due to the westward propagation
of the SST anomalies from June to August (cf., Thuburn and Sutton
2000), the corresponding cyclones (C1) also move westward (Figs. 5,
6). This might cause an additional interaction between C1 and the
westerly jet, and hence a stronger teleconnection.
Similar characteristics and patterns were also found for the
wind field anomaly in the observations of the modern La Nia (Yuan
and Yan 2013). Moreover, the SST and wind anomaly shown in Figs. 5
and 6, have similar characteris-tics as the La Nia Modoki (Ashok et
al. 2007) which has been suggested to promote more precipitation
over the monsoon front (Fan et al. 2013). The cooling anomaly
in
Fig. 4 Annual mean average over 5S5N in the tropical Pacific for
a sea surface temper-ature (C), b zonal wind stress (N/m2), c depth
of thermocline (m), d vertical wind velocity at 700 mb (Pa/s) note
that the negative velocity is upward. CCSM3 experiments are in Red
and the ones of HadCM3 are in blue
-
815State of the tropical Pacific Ocean
1 3
the central tropical Pacific and the development of the
asso-ciated upper-level anomalous cyclone in the western north
Pacific as discussed above also resembles the Pacific-East Asian
teleconnection of Wang et al. (2000). Although this teleconnection
was suggested to operate between the East Asian winter monsoon and
ENSO in the modern climate, it could also operate in summer as
found in our results.
4 MIS13 ENSO characteristics
The interannual variability of the tropical Pacific SST and EASM
precipitation around their climatological means (Sect. 3) will be
discussed in this section.
4.1 Was ENSO persistent in MIS-13?
As ENSO is the main mode of interannual variability in the
modern tropical Pacific, it is of interest to know
whether ENSO-related events were present in MIS-13 and if so, to
what extent they influenced EASM. To find the dominant modes of
interannual variability in the tropical Pacific SST, the Empirical
Orthogonal Function (EOF) was calculated from our 1,200 months of
SST data (taken from the last 100 years of our simulations) in the
region between 15S15N and 145E280E, which cover NINO regions 14.
EOFs are mathematical tool and known to show the spatial mode
(pattern) of vari-ability and correlations. Before calculating the
EOFs the monthly mean climatology was subtracted from the origi-nal
monthly time-series to produce a time-series of 1,200 monthly
anomalies. Our focus is on the first two EOFs that together, as
will be shown, explain more than 50 % of the total variance.
In the EOF plots, the regions with the same sign vary in phase
with each other, and the larger values correspond to the higher
amplitudes of the variability. The first EOF (EOF1) as the dominant
mode shows the canonical ENSO
Fig. 5 Annual cycle of the SST difference between MIS-13 and PrI
in the tropical Pacific versus longitude a CCSM3 b HadCM3. c annual
cycle of SST difference between western (147E) and east-ern (276E)
tropical Pacific, d annual cycle of the EASM precipita-
tion (averaged over 2040N and 100120E). Annual cycle of the
vertical velocity at 700 mb in the tropical Pacific versus
longitude e CCSM3 f HadCM3 (note that the negative velocity is
upward). The panels related to the tropical Pacific are first
averaged over 5S5N
-
816 M. P. Karami et al.
1 3
pattern in both MIS-13 and PrI (Fig. 7ad). The principal
components corresponding to the EOF1 vary at time scales of 35
years as would be expected for an ENSO mode (not shown). The
regions of maximum SST variability in EOF1 plots are more to the
west compared to the observations, particularly in the CCSM3 runs
(cf., Deser et al. 2006). The EOF1 of MIS-13 and PrI in CCSM3 are
almost spa-tially identical. HadCM3 MIS-13 has its maximum EOF1
variability in the central equatorial Pacific (220240E). The
percentage of the first EOFs relative to the total vari-ance for
the MIS-13 experiments (41 % in CCSM3 and 39 % in HadCM3) are
smaller than those computed for PrI (51 % in CCSM3 and 53 % in
HadCM3). This indi-cates a smaller contribution of the first mode
in the total variability of tropical Pacific SSTs in MIS-13. The
second EOF (EOF2) exhibits a dipole pattern on either side of the
equatorial Pacific (Fig. 7eh) and has a larger percent-age in
MIS-13 (11 % in CCSM3 and 13 % in HadCM3) than in PrI (9 % in CCSM3
and 5 % in HadCM3). As with EOF1, the CCSM3 MIS-13 and PrI
simulations have simi-lar EOF2 patterns. This is not the case in
HadCM3 where MIS-13 and PrI have different EOF2 patterns,
particularly in the northern and southern parts of the region
considered.
The EOF2 varies on the time scales of 23 years, slightly shorter
than EOF1.
To investigate the frequency of variability in the tropical
Pacific SST, the spectrum of SST in the NINO3.4 region were
computed. For this, the monthly mean cycles of SSTs were first
resolved and a 12-month moving average filter was applied following
Douglass (2011). The spectrum (Fig. 8a, b) shows a dominant
variability at the interannual time scale which is associated with
ENSO. The CCSM3 PrI run shows a broad spectral peak (above the 99 %
con-fidence level) in the 1.54 year band, with the highest peak
occurring at 4 years. Peaks in the CCSM3 MIS-13 are lower in
magnitude and are limited to a narrower band between 1.5 and 2.5
years. Similar results are found for the power spectrum of NINO3.4
SSTs in HadCM3 where the significant peaks (above the 99 %
confidence level) in MIS-13 have smaller amplitudes and are
concentrated in a nar-rower band (25.5 year band) than in PrI (27
year band).
The EOF of the tropical Pacific SST and spectral analy-sis of
NINO3.4 SST show, therefore, that ENSO-type inter-annual
variability remained present in MIS-13 and was occurring more
frequently than in PrI. The probability den-sity function (not
shown) for NINO3.4 SST also confirms
Fig. 6 Eddy wind field anomaly (difference between MIS-13 and
PrI) at 250 mb calculated after removing the zonal mean is shown
for June, July and August from top to down. Left column is CCSM3
and
right is HadCM3. We show the cyclones by schematic circles.
Vectors smaller than 0.5 m/s are not shown
-
817State of the tropical Pacific Ocean
1 3
these findings and shows that MIS-13 could have been
characterized by more La Nia events than El Nio ones.
To determine the amplitude of ENSO, the standard deviation of
SST in the NINO3.4 region (averaged over 5S5N; 190E240E) is used
(e.g., Collins et al. 2001). In the HadCM3, the ENSO amplitude in
MIS-13 (0.82 C) is found 15 % smaller than that in PrI (0.96 C).
Similarly for CCSM3, we also find that the ENSO ampli-tude in
MIS-13 (0.92 C) is 19 % smaller than that in PrI (1.14 C). By
applying the F-test, the difference in the amplitude of ENSO
between MIS-13 and PrI (in
both CCSM3 and HadCM3) is found statistically sig-nificant (to
95 % significance). The ENSO amplitudes in HadCM3 experiments are
at least 11 % smaller than those in CCSM3 with significance of 95
%. This can be due to the fact that the tropical Pacific
thermocline in HadCM3 is deeper than in CCSM3.
4.2 Enhanced relation between ENSO and the EASM
The spectral analysis of the modeled EASM precipitation
(averaged over 2040N and 100120E) is shown in Fig. 8
Fig. 7 The first two empirical orthogonal functions (addressed
as EOF1 and EOF2) of SST in the Pacific Ocean as a function of
latitude and longitude. Right column is for MIS-13 simulations and
the left column is for PrI
-
818 M. P. Karami et al.
1 3
(panels c and d). It demonstrates that interannual variability
of the EASM in both MIS-13 and PrI has a dominant period between 2
and 4 years, and thus appears largely synchronous with ENSO. The
EOF1 of precipitation calculated over the region between 20 and 50N
and 90E140E also varies at the same time scale as ENSO (i.e., 23
years) in both mod-els suggesting a possible impact of ENSO on the
interan-nual variability of EASM. EOF1 explains 12 % of the total
variance in both the PrI and MIS-13 simulations of CCSM3. In
HadCM3, however, MIS-13 has a larger variance for the EOF1 of
precipitation (22 %) than PrI (10 %) which indi-cates a larger
contribution of the ENSO in the total variability of EASM
precipitation in MIS-13. The pattern of EOF1 in both PrI runs (Fig.
14 Appendix) have opposite sign over southeast (2535N and 100120E)
and northeast China (3545N and 100120E). This suggests that ENSO
has opposite impacts on those regions as is also found for the
modern climate that the ENSOEASM relation may differ between the
northern and the southern parts of EASM (e.g., Wu and Wang 2002;
Lee et al. 2008). The CCSM3 MIS-13 simulation has a similar pattern
as the PrI one, but with a shift to the north. The EOF1 of HadCM3
MIS-13 shows negative values over large parts of China which
indicates that the impact of ENSO would be distributed more
uniformly.
To check if the EASM variability is correlated to the tropical
Pacific SST and whether this relation differed
during MIS-13, the Pearsons coefficient of correlation between
1,200 months of modelled NINO3.4 SST and the modelled precipitation
was computed (Fig. 9). Both models show an increase in correlation
towards more positive val-ues in the EASM region in MIS-13 compared
to PrI. A pos-itive correlation implies that higher (lower)
precipitation is associated with warmer (colder) NINO3.4 SSTs
relative to their respective means. The increase in correlation is
con-sistent with the enhanced teleconnection between the tropi-cal
Pacific and East Asia shown in Sects. 3.3 and 3.4. It can also be
seen that the HadCM3 MIS-13 simulation has a larger correlation
coefficient in the EASM region compared to the CCSM3 MIS-13
simulation. The positively increased correlation between NINO3.4
SST and precipitation in our MIS-13 simulations can be the result
of the ENSO (as the dominant mode of the tropical Pacific SST
variability) and/or the seasonal change in insolation, that force
the tropical temperature and the EASM precipitation in a similar
fash-ion. We will calculate the contribution of ENSO in Sect. 4.3.
The correlation coefficients between the NINO3.4 SST anomalies
(difference between MIS-13 and PrI) and pre-cipitation anomalies
were also calculated to find the pos-sible relation between the
changes in SST and precipitation of MIS-13 relative to PrI. The
positive correlation shows that NINO3.4 SST and precipitation
anomalies have a tendency to be greater or less than their
respective means
(a) (b)
(c) (d)
Fig. 8 Fast Fourier Transform spectrum after resolving the
monthly mean cycles by a 12-month moving average filter for: a
NINO3.4 SST in CCSM3 b NINO3.4 SST in HadCM3 c EASM precipita-tion
in CCSM3 d EASM precipitation in HadCM3. Continuous
lines are for MIS-13, dashed lines are for PrI, black lines are
99 % confidence lines. NINO3.4 is the region averaged over 5S5N and
190E240E. EASM precipitation was first averaged over 2040N and
100120E
-
819State of the tropical Pacific Ocean
1 3
simultaneously. CCSM3 has both positive and negative correlation
coefficients over EASM while HadCM3 has mainly positive values
(Fig. 15 Appendix).
The same correlation analysis as above was computed between
precipitation and the North Atlantic Oscillation (NAO) index (not
shown), but did not show a significant change over the EASM region.
However, the mean-cli-mate of MIS-13 has positive NAO-like features
(Muri et al. 2013) which strengthens the relationship between the
tropical Pacific Ocean and EASM (Wu et al. 2012) consist-ent with
our results. The correlation of the Indian Ocean Dipole Index with
EASM also did not show a difference between MIS-13 and PrI (Muri et
al. 2013). Thus ENSO took on great importance in controlling EASM
variability during MIS-13 than during PrI.
4.3 Isolating the impact of ENSO
Observations show that ENSO events (both El Nio and La Nia
phases) in the modern climate strongly influence global
precipitation (Ropelewski and Halpert 1987). In our model runs, we
also find that the precipitation changes dur-ing the modeled ENSO
events (e.g., Figure 8). Therefore, it is of interest to compute
the pure contribution of ENSO events to the total change in
precipitation of MIS-13. In other words, we are interested to
calculate how much of
the increased EASM precipitation in MIS-13 is related to changes
in ENSO. First, NINO3 SST anomalies were cal-culated with respect
to the mean climatology. ENSO events were defined to be those where
the December NINO3 SST anomaly exceeds one standard deviation (El
Nio) or falls below minus one standard deviation (La Nia) of NINO3
SST (e.g., Merkel et al. 2010). Those model years in which the
NINO3 SST anomaly fell between minus one and plus one standard
deviation were defined as non-ENSO or nor-mal years.
We define the pure impact of ENSO on precipitation as the
corresponding change in precipitation compared to the background
precipitation (mean precipitation of non-ENSO years). To compute
this, the background precipitation was subtracted from the
precipitation of the ENSO years and the average of the resulting
values for both MIS-13 and PrI were calculated. Then, the
precipitation difference between MIS-13 and PrI were separated into
their difference in the impact of ENSO and in background
precipitation (Fig. 10). As can be seen, the difference in the pure
impact of ENSO on precipitation between MIS-13 and PrI does not
change in CCSM3 but increases in HadCM3 (i.e., more rainfall) for
most of the regions throughout the globe. It can be shown that the
pure ENSO-driven precipitation over east China reaches up to 30 %
of the total precipitation differ-ence between MIS-13 and PrI. This
is consistent with the
(a) (b)
(c) (d)
Fig. 9 Pearson correlation map between NINO3.4 SST and
precipi-tation. The color contour/bar shows the value of
correlation coef-ficients. Only the correlation coefficients with a
significance level
larger than 95 % are shown. Top panels are for CCSM3 runs and
lower panels are for HadCM3
-
820 M. P. Karami et al.
1 3
increased EOF1 variance of EASM precipitation in MIS-13 compared
to PrI for HadCM3 (Sect. 4.2). The differ-ence in the pure impact
of ENSO on precipitation between
HadCM3 and CCSM3 illustrates one of the mechanisms contributing
towards HadCM3s better representation of MIS-13 EASM.
Fig. 10 The difference in precipitation between MIS-13 and PrI
(Fig. 2) separated into two parts: ENSO-driven and non-ENSO driven.
Left panels are the difference in precipitation between MIS-13
and PrI after taking into account only the pure impact of ENSO,
right panels are the difference in precipitation between MIS-13 and
PrI for their non-ENSO years
(a) (b)
(d)(c)
Fig. 11 Summer (JJA) precipitation change driven by ENSO
teleconnection shown as the difference between El Nio and La Nia
precipitation composites for: a CCSM3 MIS-13 b CCSM3 PrI c HadCM3
MIS-13 d HadCM3 PrI
-
821State of the tropical Pacific Ocean
1 3
4.4 ENSO teleconnection
To indentify the difference in ENSO teleconnections between
MIS-13 and PrI, a composite analysis of summer (JJA) precipitation
was performed. The average effect of El Nio and La Nia on
precipitation can be determined by these composite maps. By
averaging precipitation for the El Nio years, La Nia years and the
non-ENSO years, their corresponding so-called composite anomalies
were obtained. The difference between the El Nio and La Nia
precipitation composites shows changes in their corre-sponding
pattern of precipitation which is associated with the difference in
their teleconnections (Fig. 11). The El Nio and La Nia composite
difference in precipitation in PrI has a similar pattern to that of
the modern-day obser-vation (e.g., Figure 8 in Deser et al. 2006)
over southern and central east-China but with smaller magnitudes.
In both models a double ITCZ-like structure and the rainfall
distri-bution over the equatorial Pacific has disappeared in
MIS-13, likely due to the cooling of the central tropical Pacific.
Over east China, larger precipitation differences between El Nio
and La Nia composites can be seen in MIS-13 suggesting increased
ENSO teleconnection during MIS-13.
5 Summary and conclusions
With the aim of better understanding the strong EASM dur-ing the
relatively cool MIS-13 interglacial, we investigated the role of
the tropical Pacific Ocean. Two coupled general circulation models,
HadCM3 and CCSM3, were used to study the climate of MIS-13 with
different insolation and GHGs than the present day. Results from
both models con-firm the increased EASM precipitation during MIS-13
com-pared to PrI. Overall, we leave more confidence to the results
from HadCM3 given its better performance than CCSM3 in capturing
EASM in MIS-13 and at present-day (Lei et al. 2013). It was also
shown that the western Pacific subtropi-cal high was strengthened
and extended westward in MIS-13 providing more moisture to the
EASM. The additional mois-ture in MIS-13 came from the northern
Pacific Ocean for CCSM3, and from the Indian Ocean and South China
Sea as well as the northern Pacific Ocean for HadCM3.
We suggested that MIS-13 had a La Nia-like mean climate in the
tropical Pacific and the associated telecon-nection with the
extra-tropics acted to increase precipita-tion over EASM. In
MIS-13, the increasing trend of the EASM precipitation during
summer was not found for the landocean thermal contrast.
Summer-time cooling in the central tropical Pacific was suggested
to promote more rainfall over the EASM through maintaining the
summer western Pacific subtropical high (cf., Fan et al. 2013). In
MIS-13, the eastwest SST gradient in the tropical Pacific
was reduced during NH summer (reaching 3 C for HadCM3) and was
anticorrelated (around 0.8 with 95 % significance) with EASM
precipitation. The reduced eastwest SST gradient promoted an
upper-level anomalous cyclone in the western north Pacific which
was the main system connecting the tropical Pacific to the EASM.
Thus, the changes in the tropical Pacific SST contributed to the
intense EASM of MIS-13 next to the larger land-sea ther-mal
contrast driven by the higher summer insolation (e.g., Yin et al.
2008).
Based on our modelling analysis we conclude that ENSO
variability was present in MIS-13 with a smaller amplitude but
higher frequency than in PrI. On the other hand, the precipitation
rate in the EASM-region showed larger correlation with ENSO in
MIS-13. This means that although ENSO variability had smaller
amplitude in MIS-13 compared to PrI, it had a larger influence on
the EASM. This we relate to the enhanced teleconnection between the
tropical Pacific and East Asia during MIS-13. Moreover, the pure
impact of ENSO on increasing EASM precipita-tion during MIS-13 was
investigated. It was shown that in HadCM3, the ENSO-related
precipitation was stronger in MIS-13 than in PrI, and accounted for
up to 30 % of the total precipitation difference between MIS-13 and
PrI. This was not the case in CCSM3 in which the ENSO-related
precipitation did not significantly differ between MIS-13 and PrI.
This could also be one reason explaining better representation of
EASM in HadCM3 compared to CCSM3.
Our results suggest that the state of the tropical Pacific
during the past interglacials could be quite different from today,
which in turn could have changed the relationship between the
tropical Pacific and the EASM. Moreover, it could also affect the
ENSO properties and teleconnections as was shown in our results.
Future research constraining the state of the tropical Pacific
could serve the dual purpose of resolving the oceanic response of
this vitally important region to interglacial forcing as well as
constraining the enigmatic EASM during MIS-13.
Acknowledgments This work and M. P. Karami were supported by the
European Research Council Advanced Grant EMIS (No 227348 of the
Programme Ideas). Q. Z. Yin is supported by the Bel-gian National
Fund for Scientific Research (F.R.S.-FNRS). H. Muri is supported by
the Research Council of Norway (Grant agreement 229760). We are
grateful to the reviewers for their constructive com-ments and
suggestions. We thank Dr. Fred Kucharski, Dr. Carlos Almeida,
Gauillame Lenoir and Dr. Tobias Bayr for helpful discus-sions.
Access to computer facilities was facilitated through sponsor-ship
from S. A. Electrabel, Belgium. We are also grateful to CISM staff
at Universit catholique de Louvain for their technical support.
Appendix
See Figs. 12, 13, 14, 15.
-
822 M. P. Karami et al.
1 3
Fig. 13 Annual of average of sea-level pressure. First row is
PrI, second row is MIS-13 and the third row is the difference
between MIS-13 and PrI. Left and right columns are related to CCSM3
HadCM3 experiments, respectively
Fig. 12 Annual mean average over 5S5N in the tropical Pacific
for the meridional wind stress (Nm2)
-
823State of the tropical Pacific Ocean
1 3
References
Alexander MA, Blad I, Newman M, Lanzante JR, Lau NC, Scott JD
(2002) The atmospheric bridge: the influence of ENSO
telecon-nections on air-sea interaction over the global oceans. J
Clim 15:22052231. doi:10.1175/1520-0442(2002)0152.0.CO;2
Ashok K, Behera SK, Rao SA, Weng H, Yamagata T (2007) El Nio
Modoki and its possible teleconnection. J Geophys Res 112:C11007.
doi:10.1029/2006JC003798
Bassinot FC, Labeyrie LD, Vincent E, Quidelleur X, Shackleton
NJ, Lancelot Y (1994) The astronomical theory of climate and the
age of the Brunhes-Matuyama magnetic reversal. Earth Planet Sci
Lett 126:91108
Berger A (1978) Long-term variations of daily insolation and
quater-nary climatic changes. J Atmos Sci 35(12):23622367
Chang CP, Zhang YS, Li T (2000) Interannual and interdecadal
vari-ations of the East Asian summer monsoon and tropical Pacific
SSTs. Part I: roles of the subtropical ridge. J Clim 13:43104325.
doi:10.1175/1520-0442(2000)0132.0.CO;2
Chiang JCH (2009) The tropics in paleoclimate. Annu Rev Earth
Planet Sci 37:263297. doi:10.1146/annurev.earth.031208.100217
Clement AC, Seager R, Cane MA, Zebiak SE (1996) An ocean
dynamical thermostat. J Clim 9:21902196. doi:10.1175/1520-
0442(1996)0092.0.CO;2
Clement AC, Seager R, Cane MA (1999) Orbital controls on the El
Nio/Southern oscillation and the tropical climate.
Paleoceanog-raphy 14(4):441456. doi:10.1029/1999PA900013
Fig. 14 The first two empirical orthogonal functions (addressed
as EOF1 and EOF2) of summer precipitation. Right column is for
MIS-13 simulations and the left column is for PrI
0 50 100 150 200 250 300 350
50
0
50
Latit
ude
( N)
(a) Correlation of NINO 3.4 SST anomaly with precipitation
anomaly;CCSM3
0 50 100 150 200 250 300 350
50
0
50
Longitude ( E)Longitude ( E)
Correlation of NINO 3.4 SST anomaly with precipitation
anomaly;HadCM3
0.80.60.40.2
0.0.20.40.60.8
(b)
Latit
ude
( N)
Fig. 15 Pearson correlation map between NINO3.4 SST anomaly and
precipitation anomaly for a CCSM3 b HadCM3. By anomaly, we mean the
difference in a given variable between MIS-13 and PrI. Only the
correlation coefficients with a significance level larger than 95 %
are shown
-
824 M. P. Karami et al.
1 3
Collins M, Tett SFB, Cooper C (2001) The internal climate
variability of HadCM3, a version of the Hadley centre coupled model
with-out flux adjustments. Clim Dyn 17:6181
Collins WD, Bitz CM, Blackmon ML, Bonan GB, Bretherton CS,
Carton JA, Chang P, Doney SC, Hack JJ, Henderson TB, Kiehl JT,
Large WG DS, McKenna SB, Smith RD (2006) The commu-nity climate
system model version 3 (CCSM3). J Clim 19:21222143.
doi:10.1175/JCLI3761.1
De Weaver E, Nigam S (2004) On the forcing of ENSO
teleconnec-tions by anomalous heating and cooling. J Clim
17:32253235. doi:10.1175/1520-0442(2004)0172.0.CO;2
Deser C, Capotondi A, Saravanan R, Phillips AS (2006) Tropi-cal
pacific and Atlantic climate variability in ccsm3. J. Clim
19:24512481. doi:10.1175/JCLI3759.1
Douglas DH (2011) Separation of a signal of interest from a
seasonal effect in geophysical data: el Nio/La Nia phenomenon. Int
J Geosci 2:414419. doi:10.4236/ijg.2011.24045
Fan L, Shin SI, Liu Q, Liu Z (2013) Relative importance of
tropical SST anomalies in forcing East Asian summer monsoon
circula-tion. Geophys Res Lett 40:24712477.
doi:10.1002/grl.50494
Gill AE (1980) Some simple solutions for heat-induced tropi-cal
circulation. QJR Meteorol Soc 106:447462.
doi:10.1002/qj.49710644905
Gordon C, Cooper C, Senior CA, Banks H, Gregory JM, Johns TC et
al (2000) The simulation of SST, sea ice extent and ocean heat
transport in a version of the Hadley Centre coupled model with-out
flux adjustment. Clim Dyn 147168
Guo ZT, Liu TS, Fedoroff N, Wei LY, Ding ZL, Wu NQ, L HY, Jiang
WY, An ZS (1998) Climate extremes in loess of China coupled with
the strength of deep-water formation in the North Atlantic. Global
Planet Change 18:113128
Guo ZT, Berger A, Yin QZ, Qin L (2009) Strong asymmetry of
hemi-spheric climates during MIS-13 inferred from correlating China
loess and Antarctica ice records. Clim Past 5:2131
Herold N, Yin QZ, Karami MP, Berger A (2012) Modelling the
cli-matic diversity of the warm interglacials. Quaternary Science
Review 56:126141. doi:10.1016/j.quascirev.2012.08.020
Hoskins BJ, Karoly DJ (1981) The steady linear response of a
spheri-cal atmosphere to thermal and orographic forcing. J Atmos
Sci 38:11791196
Jouzel J, Masson-Delmotte V, Cattani O, Dreyfus G, Falourd S,
Hoffmann G, Minster B, Nouet J, Barnola JM, Chappellaz J, Fischer
H, Gallet JC, Johnsen S, Leuenberger M, Loulergue L, Luethi D,
Oerter H, Parrenin F, Raisbeck G, Raynaud D, Schilt A, Schwander J,
Selmo E, Souchez R, Spahni R, Stauffer B, Steffensen JP, Stenni B,
Stocker TF, Tison JL, Werner M, Wolff EW (2007) Orbital and
millennial antarctic climate variability over the past 800,000
years. Science 317:793796. doi:10.1126/science.1141038
Kukla G, An ZS, Melice JL, Gavin J, Xiao JL (1990) Magnetic
sus-ceptibility record of Chinese loess. Trans R Soc Edinb Earth
Sci 81:263288
Lee E, Chase TN, Rajagopalan B (2008) Seasonal forecasting of
East Asian summer monsoon based on oceanic heat sources. Int J
Cli-matol 28:667678. doi:10.1002/joc.1551
Lei Y, Hoskins B, Slingo J (2013) Natural variability of sum-mer
rainfall over China in HadCM3. Clim Dyn.
doi:10.1007/s00382-013-1726-8
Li T, Philander SGH (1996) On the annual cycle of the eastern
equa-torial pacific. J Clim 9:29862998.
doi:10.1175/15200442(1996)0092.0.CO;2
Lisiecki LE, Raymo ME (2005) A Pliocene-Pleistocene stack of 57
globally distributed benthic 18O records. Paleoceanography 20:
PA1003. doi:10.1029/2004PA001071
Lunt D, Abe-Ouchi A, Bakker P, Berger A, Braconnot P, Charbit S,
Fischer N, Herold N, Jungclaus J, Kohn V, Krebs-Kanzow U,
Lohmann G, Otto-Bliesner B, Park W, Pfeiffer M, Prange M,
Rachmayani R, Renssen H, Rosenbloom N, Schneider B, Stone E,
Takahashi K, Wei W, Yin Q (2013) A multi-model assess-ment of last
interglacial temperatures. Clim Past, COPERNICUS GESELLSCHAFT MBH
9:699717. doi:10.5194/cp-9-699-2013
Lthi D, Le Floch M, Bereiter B, Blunier T, Barnola JM,
Siegent-haler U, Raynaud D, Jouzel J, Fischer H, Kawamura K,
Stocker TF (2008) High-resolution carbon dioxide concentration
record 650,000800,000 years before present. Nature 453:379382.
doi:10.1038/nature06949
Meehl GA, Arblaster JM, Lawrence DM, Seth A, Schneider EK,
Kirt-man BP, Min D (2006) Monsoon regimes in the CCSM3. J Clim
19:24822495. doi:10.1175/JCLI3745.1
Merkel U, Prange M, Schultz M (2010) ENSO variability and
tele-connections during glacial climates. Quatern Sci Rev 29:86100.
doi:10.1016/j.quascirev.2009.11.006
Mohtadi M, Hebbeln D, Ricardo SN, Lange CB (2006) El Nino-like
pattern in the Pacific during marine isotope stages (MIS) 13 and
11? Paleoceanography 21: PA1015. doi:10.1029/2005PA001190
Mller W, Roeckner E (2008) ENSO teleconnections in projections
of future climate in ECHAM5/MPI-OM. Clim Dyn 31:533549.
doi:10.1007/s00382-007-0357-3
Muri H, Berger A, Yin QZ, Voldoire A, Salas D, Sundaram S (2012)
SST and ice sheet impacts on the MIS13 climate. Clim Dyn
39:17391761. doi:10.1007/s00382-011-1216-9
Muri H, Berger A, Yin QZ, Karami MP, Barriat PY (2013) The
cli-mate of the MIS-13 interglacial according to HadCM3. J Climate
26:96969712. doi:10.1175/JCLI-D-12-00520.1
Myhre G, Highwood EJ, Shine KP, Stordal F (1998) New estimates
of radiative forcing due to well mixed greenhouse gases. Geophys
Res Lett 25:27152718
Pope VD, Gallani ML, Rowntree PR, Stratton RA (2000) The impact
of new physical parametrizations in the Hadley Centre climate model
- HadAM3. Clim Dyn, 123146
Ropelewski CF, Halpert MS (1987) Global and regional scale
precipi-tation patterns associated with the El Nino/Southern
Oscillation. Mon Wea Rev 115:16061626
Rossignol-Strick M, Paterne M, Bassinot FC, Emeis KC, de Lange
GJ (1998) An unusual mid-Pleistocene monsoon period over Africa and
Asia. Nature 392:269272
Seager R, Naik N, Ting M, Cane MA, Harnik N, Kushnir Y (2010)
Adjustment of the atmospheric circulation to tropical Pacific SST
anomalies: variability of transient eddy propagation in the
Pacific-North America sector. Q J R Meteorol Soc 136:277296.
doi:10.1002/qj.588
Shen SH, Lau KM (1995) Biennial oscillation associated with the
East Asian summer monsoon and tropical sea surface tempera-ture. J
Meteor Soc Japan 73:105124
Sundaram S, Yin QZ, Berger A, Muri H (2012) Ice sheet induced
North Atlantic oscillation mode during an interglacial 500,000
years ago and its impact on East Asian summer mon-soon. Clim Dyn
39:10931105. doi:10.1007/s00382-011-1213-z
Thuburn J, Sutton RT (2000) The seasonal cycle of tropical
Pacific sea surface temperature in a coupled GCM. Clim Dyn
16(12):935947
Timmermann A, An SI, Krebs U, Goosse H (2005) Enso suppression
due to 30 weakening of the north Atlantic thermohaline
circula-tion. J Clim 18:31223139. doi:10.1175/JCLI3495.1
Trenberth KE, Branstator GW, Karoly D, Kumar A, Lau NC,
Ropelewski C (1998) Progress during TOGA in understanding and
modeling global teleconnections associated with tropical sea
surface temperatures. J Geophys Res 103(C7):1429114324.
doi:10.1029/97JC01444
Turner AG, Slingo JM (2009) Subseasonal extremes of
precipita-tion and active-break cycles of the Indian summer monsoon
in a climate-change scenario. QJR Meteorol Soc 135:549567.
doi:10.1002/qj.401
-
825State of the tropical Pacific Ocean
1 3
Wang B, Wu R, Fu X (2000) Pacific-East Asian teleconnection: how
does ENSO affect East Asian climate? J Clim 13:15171536.
doi:10.1175/1520-0442(2000)0132.0.CO;2
Wu R, Wang B (2002) A contrast of the East Asian summer
monsoonENSO relationship between 19621977 and 19781993. J Clim
15:32663279. doi:10.1175/1520-0442(2002)0152.0.CO;2
Wu Z, Li J, Jiang Z, He J, Zhu X (2012) Possible effects of the
North Atlantic oscillation on the strengthening relationship
between the East Asian summer monsoon and ENSO. Int J Climatol
32:794800. doi:10.1002/joc.2309
Yeager SG, Shields CA, Large WG, Hack JJ (2006) The
low-resolu-tion CCSM3. J Clim 19:25452566.
doi:10.1175/JCLI3744.1
Yin JH, Battisti DS (2001) The Importance of tropical sea
surface tem-perature patterns in simulations of last glacial
maximum climate. J Clim 14:565581.
doi:10.1175/1520-0442(2001)0142.0.CO;2
Yin QZ, Berger A (2012) Individual contribution of insolation
and CO2 to the the interglacial climates of the past 800,000 years.
Clim Dyn 38:709724. doi:10.1007/s00382-011-1013-5
Yin QZ, Guo ZT (2006) Mid-Pleistocene vermiculated red soils in
southern China as an indication of unusually strengthened East
Asian monsoon. Chin Sci Bull 51(2):213220
Yin QZ, Guo ZT (2008) Strong summer monsoon during the cool
MIS-13. Clim Past 4:2934
Yin QZ, Berger A, Driesschaert E, Goosse H, Loutre MF, Crucifix
M (2008) The Eurasian ice sheet reinforces the East Asian summer
monsoon during the interglacial 500,000 years ago. Clim Past 4:
7990. http://www.clim-past.net/4/79/2008/cp-4-79-2008.html
Yu PS, Chen MT (2011) A prolonged warm and humid interval during
marine isotope stage 1315 as revealed by hydrographic
recon-structions from the South China Sea (IMAGES MD972142). J
Asian Earth Sci 40(6):12301237
Yuan Y, Yan HM (2013) Different types of La Nia events and
differ-ent responses of the tropical atmosphere. Chin Sci Bull
58:406415. doi:10.1007/s11434-012-5423-5
Zhou T, Yu R, Zhang J, Drange H, Cassou C, Deser C, Hodson D,
Sanchez-Gomez E, Li J, Keenlyside N, Xin X, Okumura Y (2008) Why
the Western Pacific subtropical high has extended westward since
the late 1970s. J Clim 22:21992215. doi:10.1175/2008JCLI2527.1
State ofthe tropical Pacific Ocean andits enhanced impact
onprecipitation overEast Asia duringmarine isotopic stage
13Abstract 1 Introduction2 Model descriptions3 Climatology
ofMIS-133.1 Annualsummer atmosphere andocean MIS-13
climatologies3.2 La Nia-type mean climate ofthe tropical Pacific3.3
Strengthened relationship betweentropical Pacific SST andEast Asian
monsoon3.4 Teleconnection betweentropical Pacific andEast Asian
monsoon
4 MIS-13 ENSO characteristics4.1 Was ENSO persistent
inMIS-13?4.2 Enhanced relation betweenENSO andthe EASM4.3 Isolating
the impact ofENSO4.4 ENSO teleconnection
5 Summary andconclusionsAcknowledgments References