Poleward propagation of boreal summer intraseasonal oscillations in a coupled model: role of internal processes

Poleward propagation of boreal summer intraseasonal oscillationsin a coupled model: role of internal processes

R. S. Ajayamohan • H. Annamalai •

Jing-Jia Luo • Jan Hafner • Toshio Yamagata

Received: 3 February 2010 / Accepted: 4 May 2010 / Published online: 22 May 2010

� Springer-Verlag 2010

Abstract The study compares the simulated poleward

migration characteristics of boreal summer intraseasonal

oscillations (BSISO) in a suite of coupled ocean–atmo-

spheric model sensitivity integrations. The sensitivity

experiments are designed in such a manner to allow full

coupling in specific ocean basins but forced by temporally

varying monthly climatological sea surface temperature

(SST) adopted from the fully coupled model control runs

(ES10). While the local air–sea interaction is suppressed in

the tropical Indian Ocean and allowed in the other oceans

in the ESdI run, it is suppressed in the tropical Pacific and

allowed in the other oceans in the ESdP run. Our diag-

nostics show that the basic mean state in precipitation and

easterly vertical shear as well as the BSISO properties

remain unchanged due to either inclusion or exclusion of

local air–sea interaction. In the presence of realistic east-

erly vertical shear, the continuous emanation of Rossby

waves from the equatorial convection is trapped over the

monsoon region that enables the poleward propagation of

BSISO anomalies in all the model sensitivity experiments.

To explore the internal processes that maintain the tropo-

spheric moisture anomalies ahead of BSISO precipitation

anomalies, moisture and moist static energy budgets are

performed. In all model experiments, advection of ano-

malous moisture by climatological winds anchors the

moisture anomalies that in turn promote the northward

migration of BSISO precipitation. While the results indi-

cate the need for realistic simulation of all aspects of the

basic state, our model results need to be taken with caution

because in the ECHAM family of coupled models the

internal variance at intraseasonal timescales is indeed very

high, and therefore local air–sea interactions may not play

a pivotal role.

1 Introduction

The seasonal mean rainfall associated with the Asian

summer monsoon (ASM) dictates the livelihoods of mil-

lions of people across India, southeast and east Asia. The

economic and industrial developments of these countries

are dependent on the rain-fed agriculture. The ASM rainfall

variability occurs from synoptic to decadal and longer time

scales. Of interest here is assessing the relative roles of

local air–sea interactions and internal processes on the

propagation characteristics of the variability associated

with intraseasonal time scales (30–60 days).

All-year round, the tropics experience strong and

coherent variations in winds and precipitation at intrasea-

sonal time scales, known as Madden-Julian oscillation

(MJO; Madden and Julian 1994). While the MJO is pre-

dominantly confined to the equatorial region in boreal

winter and spring, rainfall and wind anomalies in boreal

summer propagate northward from the equatorial Indian

R. S. Ajayamohan (&)

Canadian Centre for Climate Modelling and Analysis,

University of Victoria, P.O. Box 3065, STN CSC, Victoria,

BC V8W 3V6, Canada

e-mail: [email protected]

H. Annamalai � J. Hafner

International Pacific Research Center,

University of Hawaii, Honolulu, USA

J.-J. Luo � T. Yamagata

Frontier Research Centre for Global Change,

Japan Agency for Marine-Earth Science and Technology,

Yokohama, Japan

T. Yamagata

Department of Earth and Planetary Science,

The University of Tokyo, Tokyo, Japan

123

Clim Dyn (2011) 37:851–867

DOI 10.1007/s00382-010-0839-6

Ocean (IO) and western Pacific to the land masses of India

and Southeast Asia, pacing the active and break cycles of

the ASM. While the changes in the mean conditions

between boreal winter and summer are attributed to the

changes in the propagation characteristics, during boreal

summer observed seasonal mean rainfall show that there

are three regional rainfall zones over the ASM domain. Of

them, the primary one is noted over the Indian subcontinent

and the Bay of Bengal between 15�N and 25�N. The sec-

ondary zones are observed over the equatorial IO between

the equator and 10�S, and over the warm waters of the

tropical western Pacific (Goswami and Ajayamohan 2001;

Annamalai and Slingo 2001). The boreal summer intra-

seasonal oscillation (BSISO) originates over the equatorial

IO and through eastward and north-northwestward propa-

gations influences other regional heat sources (Annamalai

and Sperber 2005).

Understanding the processes responsible, and modeling

and predicting the complex space–time propagation

characteristics of the BSISO have received wide attention

(see the reviews in Goswami 2005; Wang 2005; Waliser

2006). Many models of increasing complexity (simple toy

models to fully coupled ocean–atmosphere general cir-

culation models) have been employed to identify the

BSISO mechanisms (e.g., Yamagata and Hayashi 1984).

The large zonal scale (*15,000 km) compared to the

meridional scale (*3,000 km) of the BSISO motivated

the application of zonally symmetric dynamics for an

explanation of its genesis, poleward propagation and

temporal scale selection (Webster 1983; Goswami and

Shukla 1984). A notable result from these studies is the

existence of meridional gradient in sensible heat flux,

moist static energy and mean specific humidity (northern

side being more unstable than the southern side) that

favors the northward movement of the BSISO. Employing

three-dimensional models and invoking wave dynamics

Lau and Peng (1990) suggested that convective feedback

between monsoon flow and the equatorial MJO could

trigger westward propagating Rossby waves over the

Indian monsoon region. The model results of Wang and

Xie (1997) show that BSISO is strongly influenced by the

background circulation and low-level moisture. They

described the BSISO as a convection front formed by

continuous emanation of equatorial Rossby waves and the

northward propagation is anchored due to the eastward

movement of the northwest tilted convection front. The

GCM results of Wu et al. (2006) support the findings of

Wang and Xie (1997). Krishnan et al. (2000) also sug-

gested the importance of Rossby waves in triggering

monsoon break conditions. While the role of Rossby

waves in promoting poleward migration requires eastward

propagating component of the BSISO, Jiang et al. (2004)

proposed the importance of barotropic vorticity for

independent poleward migrating events (i.e., without

eastward component) over the Indian longitudes.

Apart from the internal processes, local air–sea inter-

action has also been identified as one of the factors

responsible for pre-conditioning the boundary layer mois-

ture ahead of convection. Idealized modeling studies

(Woolnough et al. 2001), and observational diagnostics

with high temporal resolution satellite and in-situ data

highlighted the temporal phase relationship among SST,

surface fluxes, precipitation and other related parameters

(Bhat et al. 2004; Vecchi and Harrison 2002; Sengupta

et al. 2001). Adopting a series of perturbation experiments

with ECHAM4 Atmospheric General Circulation Model

(AGCM) coupled with an intermediate ocean model, Fu

et al. (2003); Fu and Wang (2004) showed that the coupled

model simulates a more realistic phase structure of BSISO

compared to an atmosphere-only version of the model. The

central hypothesis arrived from these studies is that warm

(cold) SST anomalies lead the northward-propagating wet

(dry) phase of convection by about 7–10 days, and this

positive SST anomaly promotes enhanced moisture per-

turbation through increase in evaporation (Shinoda et al.

1998; Woolnough et al. 2001; Annamalai and Sperber

2005).

Waliser et al. (2004) examined a suite of AGCMs and

their coupled versions and noted that the BSISO is better

represented in latter. In a comprehensive study, Sperber

and Annamalai (2008) examined 18 coupled model simu-

lations and noted that all models capture the equatorial

component of the BSISO but not necessarily the poleward

component over the monsoon region. The authors con-

cluded that while the basic state is a necessary but not a

sufficient condition for realistic simulation of the BSISO.

In summary, it is well recognized that MJO or BSISO owes

its existence to internal atmospheric variability (Madden

and Julian 1994; Slingo et al. 2005; Wang 2005) but their

spatial structure and statistical properties such as frequency

and amplitude may be influenced by factors such as soil

moisture (e.g., Ferranti et al. 1999), and air–sea interaction

(e.g., Waliser et al. 2004).

Based on observational and modeling studies (Kemball-

Cook and Wang 2001; Lawrence and Webster 2001), the

sequence of events of internal origin leading to poleward

migration of BSISO can be summarized as follows: (1) In

the presence of background mean flow with easterly ver-

tical shear, the anomalous circulation driven by the diabatic

heating associated with the BSISO produces cyclonic

vorticity to the north of BSISO-related cloud band; (2) The

cyclonic vorticity then drives frictional convergence in the

planetary boundary layer that leads to higher moisture

convergence north of the cloud band; and (3) In addition,

the mean flow and the associated meridional gradient in

boundary layer moisture favors moisture convergence in

852 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

the north. It appears that moistening of the planetary

boundary layer either by internal processes and/or by local

air–sea interactions ahead of convection plays a role in the

poleward propagation of the BSISO. Three-dimensional

specific humidity observations from satellite support this

particular view (Tian et al. 2006).

During boreal summer, it is clear that maximum intra-

seasonal variability is observed over the ASM region. A

question that is not examined in detail is: what is the rela-

tive role of local air–sea interactions versus internal pro-

cesses in the life-cycle of the BSISO, in particular for

moistening of the boundary layer ahead of convection? The

present study examines these particular aspects by diag-

nosing output from a suite of coupled model integrations.

The model experiments are designed by permitting free

coupling in certain ocean basins while forced by tempo-

rally varying monthly climatological SST taken from

control integrations in other ocean sites (details in Sect. 2).

To motivate our interest, Fig. 1 shows the latitude–time

precipitation regression maps (data processing is discussed

in Sect. 2) from observations (Fig. 1a), the control simu-

lation (Fig. 1b) and sensitivity experiment where air–sea

interaction is not allowed only over the tropical Indian

Ocean (Fig. 1c). Compared to observations, both model

runs simulate the northward migration of precipitation

anomalies over the Indian monsoon region, but there are

differences in details. This finding allows us to identify

other processes apart from local air–sea interactions on the

BSISOs poleward characteristics in this model. We use a

suite of diagnostic tools on selected model variables, and

apply moisture and moist static energy budget analyses to

identify the processes that promote tropospheric moisture

ahead of the convection.

The remainder of the paper is organized as follows. The

model details and experimental designs are described in

Sect. 2. The model’s basic state and BSISO characteristics

are presented in Sect. 3. Budget diagnostics and mecha-

nisms identified for the poleward migration are discussed

in Sect. 4. Major findings of the present study are sum-

marized in Sect. 5.

2 Model and experimental design

The global ocean–atmosphere coupled general circulation

model used in this study, SINTEX-F1 was developed from

the original European SINTEX model (Gualdi et al. 2003;

Guilyardi et al. 2003) at the Frontier Research Center for

Global Change under the European Union-Japan collabo-

ration (more details in Luo et al. 2005). The atmospheric

component (ECHAM4) has a resolution of 1.1� 9 1.1�(T106) with 19 vertical levels (Roeckner et al. 1996). The

oceanic component (OPA8.2) has a resolution of a 2�Mercator mesh (increased to 0.5� in the latitudinal direc-

tion near the equator) with 31 vertical levels (Madec et al.

1998). To avoid the singularity in the coordinate system,

the finite mesh is designed in a way such that the North

Pole is replaced by two nodal points located over North

America and Eurasia. Those two components are directly

coupled every 2 h without any flux corrections using a

standardized model coupler (Valcke et al. 2000). No sea

ice model is incorporated in this version. The SINTEX-F1

(a) (b) (c)

Fig. 1 Lag-latitude diagrams of regressed precipitation anomalies

(mm day-1) averaged over 80�E–90�E illustrating poleward propa-

gation in a observations (CMAP), b ES10 simulations and c ESdI

simulations. First solid (dashed) contour of precipitation is 0.2

(-0.2) mm day-1 and contour interval is 0.8 (-0.8) mm day-1

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 853

123

model realistically simulates the El Nino and Southern

Oscillation (ENSO) and Indian Ocean Dipole (IOD)

(Gualdi et al. 2003; Yamagata et al. 2004; Luo et al. 2007)

and shows skill in experimental seasonal predictions (Luo

et al. 2008; Jin et al. 2008). The fully coupled control runs

are referred to as ES10, and pentad-mean data from the last

50 years of the total 70-year integration are analyzed here.

For the present study, to assess the regional air–sea

interactions on BSISO characteristics, coupled model

experiments have been conducted by allowing full cou-

pling in specific ocean basins but forced by temporally

varying monthly climatological SST adopted from ES10

runs in other ocean basins. This experimental set-up allows

one to suppress the air–sea interaction at all time scales in

specific ocean sites by decoupling the ocean and atmo-

sphere in that region. A ‘sponge layer’ of 10� width in

longitude and 5� in latitude, where weights of the free air–

sea coupling increase linearly toward the outside ocean,

was applied in order to avoid unrealistic instability near the

boundary. We briefly explain the experiments here, and

more details are available in Luo et al. (2007, 2009).

• Air–sea interactions suppressed over the tropical Indian

Ocean:

Monthly mean values of SST derived from the control

run (ES10 run) are specified in the tropical IO (see

Fig. 2a) to force the atmosphere, and this experiment is

referred to as ESdI runs. By comparing results from

ES10 and ESdI runs, the possible role of local air–sea

interactions on BSISO characteristics in the Indian

monsoon region can be assessed.

• Air–sea interactions suppressed over the tropical

Pacific Ocean:

In this experiment, coupling is suppressed over the

tropical Pacific (25�S–25�N, Fig. 2b) and is referred to

as ESdP runs. By comparing results from ES10 and

ESdP, we can assess the importance of local air–sea

interactions on BSISO characteristics over the tropical

west Pacific.

In both ESdI and ESdP simulations pentad-mean data

from the last 50 years of the total 70-year integration are

diagnosed here. First, pentad anomalies are calculated by

removing the respective pentad climatology. Then,

Lanczos bandpass filter with 200 weights and retaining

periods of 20–100 days are applied to the full data set

covering all the months. Finally, pentads corresponding to

the boreal summer season (June through September) are

extracted for analysis. For validating the model results,

we use pentad precipitation data from the Climate Pre-

diction Center Merged Analysis (CMAP) and winds from

NCEP/NCAR reanalyses, respectively for the period

1979–2006.

3 Simulation of mean state and BSISO characteristics

In this section, we present the ability of the model in

simulating the basic state and BSISO characteristics, first in

ES10 (Sect. 3.1), and then in the sensitivity experiments

(Sect. 3.2). We close the section by discussing space–time

evolution in BSISO characteristics (Sect. 3.3).

3.1 Basic state and BSISO in ES10 runs

The description of the SINTEX mean state has been dis-

cussed in detail elsewhere (Gualdi et al. 2003; Cherchi

et al. 2007; Ajayamohan et al. 2009). Here, we discuss the

simulated boreal summer precipitation and SST climatol-

ogies over the ASM region in ES10 runs (Fig. 3a, b).

Compared to observations, the spatial extent of the Indo-

Pacific warm pool (area covered by SST [28�C) is rea-

sonably simulated. In the model, the lack of a minimum

SST along the Somali–Oman coast implies a weaker cross-

equatorial flow and hence reduced upwelling (not shown).

As regards to precipitation, while the model captures the

location of the major regional centers the simulated

intensities are indeed weak. The excess rainfall over the

equatorial western IO may be attributed to high mean SST

simulated there. The pattern correlation between the

observed and model simulated precipitation climatology

over the ASM region (40�E–120�E, 20�S–30�N) is 0.85.

This model simulated precipitation climatology is compa-

rable to other state-of-art coupled models that showed

(a) (b)Fig. 2 SINTEX-F decoupling

experiments. The shaded region

represents decoupling area in

a ESdI, b ESdP experiments. A

‘sponge layer’ where weights of

the free air–sea coupling

increase linearly toward the

outside ocean was applied near

the boundary

854 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

realistic monsoon precipitation climatology (Sperber and

Annamalai 2008).

The integrated diabatic heating (or precipitation) forces

westerlies in the lower and easterlies in the upper tropo-

sphere, respectively, resulting in strong easterly vertical

shear (Webster and Yang 1992, Fig. 3c). Theoretical and

modeling studies have shown that Rossby waves that are

forced by equatorial MJO are trapped over the ASM

domain during boreal summer by this shear (Wang and Xie

1997; Krishnan et al. 2000; Annamalai and Sperber 2005).

Consistent with simulated precipitation or diabatic heating

pattern, the model simulates the spatial structure of the

easterly shear but its intensity is weaker than observed

(Fig. 3d). The meridional extent of the shear to about 25�N,

and its eastward extension till 160�E are expected to favour

poleward propagation over both Indian and west Pacific

monsoon regions. The simulated intraseasonal variance

pattern and propagation characteristics are presented next.

In general, the model simulated intraseasonal variance

pattern (Fig. 4b) agrees with that of the observation

(Fig. 4a) but the simulated magnitude is very high. Despite

the weakened mean monsoon precipitation, the ECHAM

family of coupled and uncoupled models has large intra-

seasonal variance (e.g., Gualdi et al. 1999; Sperber et al.

2005; Sperber and Annamalai 2008). To discern both the

eastward propagation along the equator and northward

migration over the Indian monsoon region, lead/lag

regression analysis is performed. Statistical significance of

regression analysis is estimated using a t test and only

anomalies greater than 0.1 significance level is used for the

entire analysis. Both in model and observations (Fig. 4a–c)

a local maximum in variance is noted over the region

(80�E–100�E, 5�S–5�N) of the equatorial IO. Time series

of precipitation anomalies averaged over this region is

considered as the ‘base’ time series against which regres-

sions of precipitation and other variables are estimated.

The observed equatorial component of the BSISO

(Fig. 4d) constructed from the regression analysis confirms

the genesis over the equatorial western IO, coherent east-

ward propagation and local maximum in precipitation

anomalies over the eastern equatorial IO. The simulated

equatorial component (Fig. 4e, f) depicts similar features.

Note that a realistic simulation of the equatorial component

is a necessary condition for simulating subsequent pole-

ward migration over the Indian monsoon region (Sperber

and Annamalai 2008). Substantiating that, the regression

analysis shows a realistic simulation of meridional pro-

gression (Fig. 1b, c) compared to observations (Fig. 1a).

Notably, at all lead/lags, a north–south dipole pattern in

precipitation anomalies between the equatorial IO and

Indian subcontinent region around 10�N–15�N is observed

(Fig. 1a) as well as simulated (Fig. 1b, c). Combined

interpretation of equatorial and poleward components from

both observations (Figs. 4d, 1a) and model simulations

(Figs. 4e, 1b) suggest that this model is capable of cap-

turing the basic elements of the BSISO over the Indian

(a) (b)

(d)(c)

Fig. 3 JJAS mean precipitation (shaded, mm day-1) and SST (contours, �C) in a observations and b ES10 simulations. Only SST contours

above 27�C are shown. Mean easterly zonal wind shear (U850 - U200, m s-1) in c observations and d ES10 simulations

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 855

123

monsoon region. Few earlier studies also noted that the

SINTEX model performs well in simulating the boreal

winter MJO (Sperber et al. 2005) and BSISO (Ajayamohan

et al. 2009). In addition, Sperber and Annamalai (2008)

note that coupled models whose atmospheric component is

ECHAM4, out-performed other coupled models in simu-

lating the BSISO.

3.2 Basic state and BSISO in ESdI and ESdP runs

Does the suppression of local air–sea interaction influence

the simulation of the basic state? Figure 5 shows boreal

summer climatology difference in precipitation and vertical

shear between ESdI and ES10 runs. The seasonal mean

precipitation in ESdI run shows enhanced values of the

order of 2 mm day-1 over the Bay of Bengal and central IO

while reduced precipitation is noticeable over the eastern

equatorial IO and parts of the Maritime Continent sector

(Fig. 5a). The net changes in precipitation or diabatic

heating over the ASM appears small. Thus, change in the

easterly zonal wind shear is also subtle with little increase

(&1.5 m s-1) over the Arabian Sea and north-west India

and decrease over the tropical west Pacific. Over the ASM

region similar minor differences in the basic state are also

noted between ESdP and ES10 runs (not shown). There-

fore, in this model, the basic state in diabatic heating and in

the first internal baroclinic mode (shear) over the ASM

region remain largely unchanged due to either inclusion or

exclusion of local air–sea interaction over the tropical

Indian or Pacific Oceans.

In ESdI runs, the regression analysis also reveals the

realistic simulation of meridional progression (Fig. 1c) in

agreement with observations (Fig. 1a) and ES10 runs

(Fig. 1b). The equatorial component shown in Fig. 4f is

also well captured in ESdI runs. Similar results in ESdP

runs are also obtained (not shown). Hence forth, most of

(a) (b) (c)

(f)(e)(d)

Fig. 4 Mean variance of 20–100 day bandpass filtered June–Sep-

tember precipitation anomalies (shaded contours; mm2 day-2) from aCMAP , b ES10 and c ESdI. Lag-latitude diagrams of regressed

precipitation anomalies averaged over 5�S–5�N illustrating eastward

propagation in d observations, e ES10 simulations and f ESdI

simulations. First solid (dashed) contour of precipitation in the lowerpanels is 0.2 (-0.2) mm day-1 and contour interval is 0.8 (-0.8) mm

day-1

856 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

the discussions are confined to ES10 and ESdI runs, as we

are interested in identifying the possible processes related

to poleward migration over the Indian monsoon region.

Towards the end, however, we highlight the results from

ESdP runs also.

3.3 Space–time evolution of BSISO

As mentioned in Sect. 1, previous studies have shown that

systematic lead/lag relationship in SST and boundary layer

moisture with respect to precipitation as the one of the

reasons for the coherent poleward propagation of BSISO

(Woolnough et al. 2000; Sengupta et al. 2001; Fu et al.

2003). Following a similar approach, the precipitation base

time series is regressed against precipitation and SST at

various lead times over the ASM region. Here lag = 0 is

referred to as simultaneous correlation with the base time

series and lag = 10 (lead = 10) indicates regressed pre-

cipitation anomalies 2 pentads before (after) lag = 0.

Figure 6 shows precipitation (shaded) and SST (con-

tours) regressions for a 30-day period in ES10 integrations,

and each panel is 5-days apart. In this model, typical

BSISO characteristics include (a) initiation over the wes-

tern IO (Fig. 6a); (b) amplification over the central/eastern

IO (Fig. 6b, c); (c) bifurcation into northern and southern

latitudes over the eastern IO (Fig. 6d); (d) pronounced

poleward migration over the northern latitudes (Fig. 6e, f),

and (e) eastward migration over the Maritime Continent

(Fig. 6b–d). These features in the simulated space–time

evolution in precipitation are in good agreement with

observations (Lau and Chen 1986; Wang and Rui 1990;

Annamalai and Slingo 2001; Ajayamohan and Goswami

2007). Consistent with previous studies that examined

coupled models or observations, both in the equatorial IO

and over the northern IO the warm (cold) SST anomalies

lead positive (negative) precipitation anomalies by about

5–10 days.

Figure 7 shows similar spatial regressions in precipita-

tion (shaded) from ESdI runs. Like in observations and

ES10 runs, both the equatorial and poleward propagating

components are realistically captured in ESdI runs. Spe-

cifically, without coherent SST signals at intraseasonal

time scales, this particular model is able to simulate space–

time evolution in BSISO precipitation anomalies over the

ASM domain. The prevailing notion that the leading warm

SST anomalies precondition the boundary layer for initia-

tion of new convection center does not appear to be borne

out here.

4 Mechanisms for BSISO

In this section, first we present and discuss the lead/lag

relationship among different variables from ES10 and ESdI

runs (Sect. 4.1), followed by diagnostics that identify the

moist processes leading precipitation anomalies (Sect. 4.2),

and finally, we summarize the present findings with exist-

ing hypotheses (Sect. 4.3). Since the results from ES10 and

ESdI experiments are similar some diagnostics are pre-

sented from only one of the experiments.

4.1 Poleward propagation in sensitivity experiments

Figures 6, 7, 8 and 9 show the spatial as well as latitude–

time regression maps (averaged over the longitudes 80�E–

90�E) of precipitation and other-related variables from both

runs. We repeated our regression analysis for a longitudinal

average over 65�E–95�E and obtained similar results (not

shown). In the control run (ES10), poleward propagation is

expected to be influenced by both air–sea interaction and

internal processes whereas in ESdI run, the latter alone

influences poleward migration. We examine these possi-

bilities now.

In ES10 integrations, both in the near-equatorial region

and over the Bay of Bengal (Figs. 6, 8a), the signatures in

SST lead precipitation anomalies by about 1–2 pentads.

Intriguingly, enhanced latent heat flux anomalies (negative

contours in Fig. 8b) lead precipitation variations in the

near-equatorial IO (10�S–0�), but not so clearly in the

northern IO between 10�N and 20�N. In ESdI run, as

expected, no coherent lead/lag relationship between SST

and precipitation anomalies are evident (Fig. 8d). The

(a)

(b)

Fig. 5 Difference (ESdI-ES10) in JJAS seasonal mean a precipita-

tion (mm day-1) and b easterly wind shear (U850–U200, m s-1)

between ESdI and ES10 experiments

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 857

123

signatures in latent heat flux anomalies (Fig. 8e) are,

however, consistent with those noted in ES10 experiment.

Based on the experimental designs, in ES10 integration

changes in both SST and wind can lead to changes in

evaporation, while in ESdI run only changes in wind lead

to evaporation anomalies. In ESdI run, a combined exami-

nation of low-level wind anomalies (Fig. 7) and latent heat

flux anomalies (Fig. 8e) suggest that reduced (enhanced)

flux anomalies occur during easterly (westerly) wind

anomalies over the northern IO. In addition, Fig. 7 suggests

that low-level westerly (easterly) wind anomalies, either

along the equatorial region or over the northern IO, are a

response to enhanced (reduced) precipitation anomalies.

Alternatively, consistent with observational findings in

Kemball-Cook and Wang (2001), in both runs, anomalous

low-level convergence does not lead positive precipitation

anomalies in the northern IO (not shown). In summary,

Figs. 6, 7 and 8 imply that in both runs variations in latent

heat flux are largely due to wind anomalies. Our interpre-

tation is consistent with the fact that in both runs, wind

variations cause identical changes in mixed-layer depth

(Fig. 8c, f); shallowest during the calm–clear phase and

deeper during cloudy–windy phase). Note that the changes

in mixed-layer depth are not permitted to make changes in

SST in ESdI runs. Therefore, irrespective of whether or not

local air–sea interaction is allowed the robustness in the

results allows us to examine internal processes role in the

BSISO characteristics.

4.2 Internal processes

Figure 7 shows space–time evolution of precipitation and

low-level wind anomalies from ESdI runs. Beginning from

lag 10 days the development and eastward migration of

positive rainfall anomalies along the equatorial IO and its

further extension into the Maritime Continent till lag 0

(Fig. 7c) are accompanied with development and eastward

extension of negative rainfall anomalies over the northern

IO extending into the tropical west Pacific. In response to

these dry conditions in the northern IO, low-level anticy-

clonic circulation anomalies are simulated. On the other

hand, in response to moist conditions in the equatorial IO,

beginning at lag 0, low-level cyclonic circulation anomalies

are noticed. These wind anomalies over the ASM region are

likely Rossby-wave response to the respective heating

anomalies. The concurrent intensification of positive

(a) (b) (c)

(f)(e)(d)

Fig. 6 Regressed precipitation (shaded, mm day-1) and SST (con-tour, �C) anomalies at different lead/lags from the ES10 experiment.

Regression is calculated with respect to a base time series of

precipitation averaged over 80�E–100�E, 10�S–5�N and lag = 0

represents simultaneous correlation with the base time series

858 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

rainfall anomalies along the equatorial IO and negative

rainfall anomalies poleward of it (Fig. 7b, c) suggest the

possibility of mutual interaction among the regional heat

sources (Annamalai and Sperber 2005). The largest easterly

wind anomalies over the Arabian Sea region occurs at lag 0,

possibly due to constructive interference of anti-cyclonic

circulation anomalies forced by dry conditions over India-

Bay of Bengal, and the cyclonic circulation anomalies

forced by moist conditions along the eastern equatorial IO.

Over the Arabian Sea, for instance, the easterly wind

anomalies acting against the climatological westerlies tend

to warm the SST there (Fig. 6) but not necessarily promote

an increase in latent heat flux anomalies.

Figure 9 shows the space–time evolution of lower-tro-

pospheric specific humidity from ESdI runs. A comparison

of Fig. 7 and Fig. 9 suggest that enhanced moisture

anomalies lead the increase in rainfall anomalies. For

example, at lag 5 days while the increased rainfall anom-

alies along the equatorial IO cover the region (50�E–

120�E; 10�S–5�N) the moisture anomalies penetrate further

polewards in both hemispheres as well eastwards beyond

130�E. This poleward intrusion of moisture anomalies in

the Northern Hemisphere ahead of enhanced precipitation

anomalies is clearly visible at lag 0 (Fig. 9c), and thereafter

over the Arabian Seas and the Indian subcontinent reaching

as far as 30�N (Fig. 9d–f). The north-northwestward

migration of the moisture anomalies from the central-

eastern equatorial IO ahead of similar propagation of

rainfall anomalies illustrate the possible effect of Rossby

waves emanated from the equatorial IO. The continuous

emanation of moist Rossby waves due to equatorial east-

ward extension of positive rainfall anomalies into the

Maritime Continent (Fig. 7c, d) likely promotes the pole-

ward migration over the Bay of Bengal and South China

Sea regions (Fig. 7e, f). Similar north-northwestward

propagation of moisture anomalies (Fig. 9c–f) leading the

precipitation anomalies (Fig. 7d–f) are also noticeable over

the tropical western Pacific. In both runs, another view of

moisture anomalies leading precipitation anomalies is

shown in Fig. 10a, b. Both in the equatorial and off-

equatorial regions, there is a clear indication for specific

humidity at 600-hPa to lead precipitation anomalies. While

the role of Rossby waves for poleward migration over the

Indian monsoon region is consistent with other studies

(e.g., Wang and Xie 1997; Lawrence and Webster 2002;

Wu. et al. 2006), in the present study, the lead/lag rela-

tionship between lower-tropospheric moisture and preci-

pitation variations deserves attention.

Next, we perform moisture and moist static energy

(MSE) budgets to identify the dominant term that promotes

(a) (b) (c)

(f)(e)(d)

Fig. 7 Same as Fig. 6, but for precipitation and 850 hPa wind anomalies (m s-1) for the ESdI experiment

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 859

123

enhanced moisture anomalies ahead of precipitation

anomalies. The equations for moisture and MSE budgets

are described and discussed in Neelin and Su (2005) and

Annamalai (2009) and hence not repeated here. The lead-

ing terms of the moisture budget are precipitation, moisture

convergence, latent heat flux and moisture advection. It

should be noted here that in Fig. 11a, positive sign means

net moisture convergence into the column and vice versa.

Similarly, moisture advection (Fig. 12a) into the column is

positive and dry advection is negative. Figure 11 shows

regressions of total moisture convergence, and sensible

heat flux from both ES10 and ESdI runs. Note that latent

heat flux anomalies are already shown (Fig. 8b, e) and

moisture advection results are shown in Fig. 12a. While

precipitation and moisture convergence are the dominant

terms in the moisture budget, there is little evidence for the

latter to lead the former (Fig. 11a, b). In fact, enhanced

precipitation anomalies lead or appear in situ to moisture

convergence anomalies. Observational (Kemball-Cook and

Wang 2001) and modeling studies (Annamalai and Sperber

2005) note that BSISO-related low-level convergence does

not lead convection over the northern IO. Similarly, we

already noted that latent heat flux anomalies do not lead

precipitation anomalies over the northern IO (Fig. 8b, e). In

both experiments, sensible heat flux anomalies lead the

precipitation anomalies but their contribution is very small

(Fig. 11c, d). Of the moisture budget terms, therefore, it is

the moisture advection that clearly shows a lead with

respect to precipitation anomalies (Fig. 12a).

A further examination of the individual terms of the

moisture advection (Fig. 12c–f) indicates that it is the

advection of anomalous moisture by the climatological

(a) (b) (c)

(f)(e)(d)

Fig. 8 Lag-latitude diagrams of regressed precipitation anomalies

(shaded, mm day-1) averaged over 80�E–90�E a, b, c ES10

simulations and d, e, f ESdI simulations. The abscissa is time in

pentads relative to day 0 of the composite. Contours represent a, d

SST (�C), b, e Latent heat flux (W m-2) and c, f mixed-layer depth

(m) regressed anomalies respectively. Stippled lines in c, f indicate

propagation speed discussed in Sect. 4.3

860 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

winds (Fig. 12e) that dominates the moist advection

(shown as color shading) as well as the term that leads

positive precipitation anomalies (shown as contours).

Similarly, dry advection leads negative precipitation

anomalies. For example, at about 15�N, enhanced moist

advection (shaded in blue) is noted 2 pentads ahead of local

precipitation (solid contours) maximum. The contribution

from the advection of anomalous moisture by the

(a) (b) (c)

(f)(d) (e)

Fig. 9 Regressed anomalies of lower tropospheric specific humidity

(kg hPa kg-1) at different lead/lags from the ESdI experiment. Lower

tropospheric humidity is the integrated specific humidity from the

boundary to 600 hPa. Regression is calculated with respect to a base

time series mentioned in the text

(a) (b)

Fig. 10 Lag-latitude diagram of regressed 600 hPa specific humidity

(contours; kg kg-1 9 104) and precipitation (shaded; mm day-1)

anomalies averaged over 60�E–80�E for a ES10 b ESdI experiments.

First solid (dashed) contour of specific humidity is 0.5 (-0.5)

kg kg-1 and contour interval is 0.5 (-0.5) kg kg-1

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 861

123

climatological winds can be understood as follows:

Beginning at lag 10 days onwards to lag 0 days (Figs. 6, 7)

the north–south gradient in precipitation anomalies

extending over 60�E–100�E are associated with similar

signatures in tropospheric specific humidity (Fig. 9a–c).

The climatological cross-equatorial flow covers the entire

equatorial IO, turning into strong westerlies in the northern

IO (not shown). This, in conjunction with the pronounced

north–south gradient in anomalous moisture (Fig. 9a–c)

results in the dominance of the term, advection of ano-

malous moisture by the climatological winds (Fig. 12e). In

sharp contrast, the anomalous winds acting on the clima-

tological moisture lags precipitation anomalies (Fig. 12c)

primarily due to the fact that the anomalous winds are

result (response) of anomalous precipitation (Fig. 7). The

contribution from anomalous winds acting on anomalous

moisture gradient is small, and hence not shown.

Observational (Kemball-Cook and Weare 2001) and

modeling studies (e.g., Maloney 2009) suggest that MSE is

a useful diagnostic tool to examine the recharge–discharge

paradigm for tropical intraseasonal oscillations. In other

words, a buildup of column MSE occurs before the initi-

ation of MJO deep convection. Figure 13 clearly shows

that that is the case for the BSISO in ES10 run, as well as in

ESdI run (not shown). In addition, in deep tropics, while

moisture convergence dominates the moisture budget

(Fig. 11), an analysis of MSE budget allows to assess the

contribution of smaller terms in contributing to the mois-

ture convergence itself (Neelin and Su 2005; Annamalai

2009). The terms that make up MSE budget includes

(a) (b)

(c) (d)

Fig. 11 Lag-latitude plot of the sum of leading terms of regressed

total moisture convergence (W m-2) and surface sensible heat flux

(W m-2) anomalies. Shaded contours in top panels represents total

moisture convergence anomalies for a ES10 and b ESdI experiments.

Shaded contours in the bottom panels indicate sensible heat flux for c

ES10 and d ESdI experiments. Overlaid solid and dashed contoursrepresents regressed precipitation anomalies for the corresponding

experiments. First solid (dashed) contour of precipitation is 0.2 (-

0.2) mm day-1 and contour interval is 0.4 (-0.4) mm day-1

862 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

(a) (b)

(d)(c)

(e) (f)

Fig. 12 Lag-latitude diagram of regressed precipitation (contours;

mm day-1) and moisture advection terms (shaded; W m-2) averaged

over 80�E–90�E in a, c, e ES10 b, d, f ESdI experiments. Shadedcontours represents a, b total moisture advection b, c anomalous wind

on climatological moisture gradient and c, d climatological wind on

anomalous moisture gradient respectively. First solid (dashed)

contour of precipitation is 0.2 (-0.2) mm day-1 and contour interval

is 0.4 (-0.4) mm day-1

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 863

123

moisture advection, sensible and latent heat fluxes, and

temperature advection (cf. Neelin and Su 2005). While

neglecting the small contributions from temperature

advection (not shown), it is noted that the moisture

advection term dominates the MSE budget itself (Figs. 8,

11, 12).

4.3 Discussion

Both in observations (e.g., Goswami and Ajayamohan

2001; Annamalai and Slingo 2001; Lawrence and Webster

2001) and in coupled models (e.g., Sperber and Annamalai

2008) the BSISO is dominated by both eastward and north-

northwestward propagating components. In this scenario,

the continuous emanation of moist Rossby waves (Figs. 7,

9), and their subsequent trapping in the monsoon region by

the basic state easterly shear (e.g., Wang and Xie 1997)

enable the poleward migration of precipitation anomalies.

This hypothesis is at work in both ES10 and ESdI runs. The

present study finds that, the column integrated MSE as well

as specific humidity (Figs. 9, 13) tend to destabilize the

atmosphere poleward and create new centers of precipita-

tion anomalies, and thus promote poleward migration. The

finding that anomalous mid-tropospheric moisture leads

precipitation anomalies (Fig. 10) suggests the proper rep-

resentation of mid-level congestus clouds in the model

(Tian et al. 2006).

Fu et al. (2003) coupled ECHAM4 AGCM to a mixed-

layer ocean model and noted improvement in the basic

state as well as in capturing the equatorial and poleward

components of the BSISO, and an active role of air–sea

interactions is attributed for this improvement. Jiang et al.

(2004) examined the independent northward moving

component (without the existence of an eastward compo-

nent) in an ECHAM4 integration forced with climatologi-

cal SST. They found that barotropic vorticity of the BSISO

in the presence of mean easterly shear enables poleward

migration of precipitation anomalies. Apart from this

internal process, Jiang et al. (2004) also speculated the role

of moisture advection. Sperber and Annamalai (2008)

noted that irrespective of the ocean model component,

ECHAM4 family of coupled models captured the BSISO

more realistically compared to any other coupled model. It

should be mentioned here that Sperber and Annamalai

(2008), however, did not examine the SINTEX coupled

model. In summary, compared to observations the

ECHAM4 family of coupled models, irrespective of the

ocean model component, has very high BSISO variance

(a) (b) (c)

(f)(e)(d)

Fig. 13 Regressed anomalies of moist static energy (shaded; W m-2) and precipitation (contours; mm day-1) at different lead/lags from the

ES10 experiment. First solid (dashed) contour of precipitation is 0.4 (-0.4) mm day-1 and contour interval is 0.4 (-0.4) mm day-1

864 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

over the ASM region (e.g., Fig. 4). Thus, the model

internal variability at intraseasonal time scales dominates.

The comparable variance in precipitation and other asso-

ciated variables (Figs. 6, 7, 8, 9) in both ES10 and ESdI

simulations further indicate that the propagating compo-

nent of the BSISO is not necessarily sensitive to local

air–sea interaction. Only notable difference in BSISO

characteristics between ES10 and ESdI integrations is that

the rate of poleward migration of precipitation anomalies is

faster by about &1 pentad in the latter (see stippled lines in

Fig. 8).

As mentioned earlier, the mean BSISO characteristics

and the lead/lag relationships of various parameters with

respect to BSISO precipitation in the ESdP simulations are

similar to the ES10 runs (figures not shown for brevity). An

example is shown in Fig. 14, where the regressed anoma-

lies of precipitation, averaged over latitudes 15�N–20�N

are plotted. The westward migration of precipitation

anomalies over the tropical west Pacific is realistically

simulated in the ES10 integrations (Fig. 14a), indicative of

Rossby waves. This feature is simulated reasonable well in

the ESdP runs compared to the control runs (Fig. 14b), but

with a higher phase speed. Irrespective of the suppression

of air–sea interaction in the tropical Indian and Pacific

Oceans, this model maintains the steady northward pro-

pagation of BSISO anomalies in the IO and westward

propagation of BSISO anomalies over the tropical west

Pacific. The realistic simulation of the large spatial extent

of BSISO precipitation anomalies covering from 50�E to

beyond 130�E in all the model experiments is a noteworthy

feature (Figs. 6, 7).

5 Summary

In the the present study, using coupled ocean–atmospheric

model sensitivity experiments, the relative role of internal

processes versus local air–sea interaction in influencing the

poleward propagation characteristics of BSISO are exam-

ined. The experiments are designed in such a way to allow

full coupling in specific ocean basins but forced by tem-

porally varying monthly climatological SST adopted from

the control runs in other ocean basins. Accordingly, we

have two sets of experiments apart from the fully coupled

control (ES10) run. While the local air–sea interaction is

suppressed in the tropical IO and allowed in the other

oceans in the ESdI run, it is suppressed in the tropical

Pacific and allowed in the other oceans in the ESdP run.

Detailed diagnostics are carried out to study the BSISO

characteristics in these three experiments with the objective

of assessing the influence of regional air–sea interaction on

BSISO propagation.

The simulation of mean state and BSISO properties of

the model is realistic compared to observations. It is found

that the basic state in diabatic heating and BSISO charac-

teristics like northward and eastward propagation remain

unchanged due to either inclusion or exclusion of local air–

sea interaction over the tropical Indian or Pacific Oceans. It

is also found that the BSISO variance of precipitation and

other variables in all model sensitivity experiments are

comparable, suggesting that local air–sea interaction is not

necessarily sensitive in the realistic simulation of BSISO

properties. Our model results do not support the prevailing

notion that the leading warm SST anomalies precondition

the boundary layer for initiation of new convection center

leading to steady propagation. Due to the fact that ECHAM

family of coupled or uncoupled models possess large

internal variance at intraseasonal timescales, the results

presented here may be model dependent. In this scenario,

the possible internal mechanisms responsible for realistic

simulation of BSISO characteristics are examined.

The basic hypothesis that the continuous emanation of

Rossby waves from the warm equatorial IO and their sub-

sequent trapping in the ASM region by the basic easterly

(b)(a)

Fig. 14 Lag-longitude plot of regressed anomalies of precipitation (mm day-1) averaged over 15�N–20�N in a ES10 and b ESdP experiment

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 865

123

wind shear enabling the poleward propagation of BSISO

anomalies is working well in all the model sensitivity

experiments. It is found that mid-tropospheric moisture and

MSE anomalies lead precipitation anomalies in the ESdI

run like the ES10 runs, thereby destabilizing the atmosphere

poleward to create new centers of convection. By analyzing

the MSE budget, it is further shown that the moisture

advection of anomalous moisture gradient by climatological

winds plays a dominant role in BSISO propagation.

The results presented here confirm earlier suggestions

that a realistic basic state, in particular the zonal and

meridional extent of the vertical easterly shear (e.g., Wang

and Xie 1997), and a proper representation of the equato-

rial component with sufficient intensity to force equatorial

Rossby waves (e.g., Sperber and Annamalai 2008) are

basic ingredients that models need to possess for a rea-

sonable representation of the BSISO. In the present study,

the identification of the role of climatological winds

advecting anomalous moisture gradient in the life-cycle of

the BSISO further emphasizes the need for realistic simu-

lation of the basic elements.

Acknowledgements RSA is supported by the Global Atmosphere–

Ocean Prediction and Predictability research network, funded by the

Canadian Foundation for Climate and Atmospheric Sciences.

H. Annamalai is supported by the IPRC through its institutional grants

from NOAA, NASA and JAMSTEC.

References

Ajayamohan RS, Goswami BN (2007) Dependence of simulation of

boreal summer tropical intraseasonal oscillations on the simu-

lation of seasonal mean. J Atmos Sci 64:460–478. doi:

10.1175/JAS3844.1

Ajayamohan RS, Rao SA, Luo JJ, Yamagata T (2009) Influence of

Indian Ocean dipole on boreal summer intraseasonal oscillations

in a coupled general circulation model. J Geophys Res

114:D06119. doi:10.1029/2008JD011096

Annamalai H (2009) Moist dynamical linkage between the equatorial

Indian Ocean and the south Asian monsoon trough. J Atmos Sci.

doi:10.1175/2009JAS2991.1

Annamalai H, Slingo JM (2001) Active/break cycles: diagnosis of the

intraseasonal variability of the Asian summer monsoon. Clim

Dyn 18:85–102

Annamalai H, Sperber KR (2005) Regional heat sources and the

active and break phases of boreal summer intraseasonal (30–

50 day) variability. J Atmos Sci 62:2726–2748

Bhat GS, Vecchi GA, Gadgil S (2004) Sea surface temperature of the

Bay of Bengal derived from the TRMM microwave imager.

J Atmos Ocean Technol 21:1283–1290

Cherchi A, Gualdi S, Behera S, Luo JJ, Masson S, Yamagata T,

Navarra A (2007) The influence of tropical Indian Ocean SST on

the Indian summer monsoon. J Clim 20:3083–3105

Ferranti L, Slingo JM, Palmer TN, Hoskins BJ (1999) The effect of

land-surface feedbacks on monsoon circulation. Q J R Meteor

Soc 125:1527–1550

Fu X, Wang B (2004) Differences of boreal summer intraseasonal

oscillations in an atmosphere–ocean coupled model and atmo-

sphere-only model. J Clim 17:1263–1271

Fu X, Wang B, Li T, McCreary J (2003) Coupling between northward

propagating intraseasonal oscillations and sea-surface tempera-

ture in the Indian Ocean. J Atmos Sci 60(15):1733–1753

Goswami BN (2005) South Asian monsoon. In: Lau WKM, Waliser

DE (ed) Intraseasonal variability in the Atmosphere–Ocean

climate system, chap 2. Praxis Springer, Berlin, pp 19–61

Goswami BN, Ajayamohan RS (2001) Intraseasonal oscillations and

interannual variability of the Indian summer monsoon. J Clim

14:1180–1198

Goswami BN, Shukla J (1984) Quasi-periodic oscillations in a

symmetric general circulation model. J Atmos Sci 41:20–37

Gualdi S, Navarra A, Tinarelli G (1999) The interannual variability of

the MaddenJulian oscillation in an ensemble of GCM simula-

tions. Clim Dyn 15:643–658

Gualdi S, Navarra A, Guilyardi E, Delecluse P (2003) Assessment of

the tropical Indo-Pacific climate in the sintex CGCM. Ann

Geophys 46:1–26

Guilyardi E, Delecluse P, Gualdi S, Navarra A (2003) Mechanisms

for ENSO phase change in a coupled GCM. J Clim 16:1141–

1158

Jiang X, Li T, Wang B (2004) Structures and mechanisms of the

northward propagating boreal summer intraseasonal oscillation.

J Clim 17:1022–1039

Jin EK, Kinter JL, Wang B, Park CK, Kang IS, Kirtman BP, Kug JS,

Kumar A, Luo JJ, Schemm J, Shukla J, Yamagata T (2008)

Current status of ENSO prediction skill in coupled ocean–

atmosphere model. Clim Dyn. doi:10.1007/s00382-008-0397-3

Kemball-Cook SR, Wang B (2001) Equatorial waves and air–sea

interaction in the boreal summer intraseasonal oscillation. J Clim

14:2923–2942

Kemball-Cook SR, Weare BC (2001) The onset of convection in the

Madden-Julian oscillation. J Clim 14:780–793

Krishnan R, Zhang C, Sugi M (2000) Dynamics of breaks in the

Indian summer monsoon. J Atmos Sci 57:1354–1372

Lau KM, Chen PH (1986) Aspects of 30–50 day oscillation during

summer as inferred from outgoing longwave radiation. Mon Wea

Rev 114:1354–1369

Lau KM, Peng L (1990) Orgin of low-frequency (intraseasonal)

oscillations in the tropical atmosphere. Part III: monsoon

dynamics. J Atmos Sci 47:1443–1462

Lawrence DM, Webster PJ (2001) Interannual variations of the

intraseasonal oscillation in the south Asian summer monsoon

region. J Clim 14:2910–2922

Lawrence DM, Webster PJ (2002) The boreal summer intraseasonal

oscillation: relationship between northward and eastward move-

ment of convection. J Atmos Sci 59:1593–1606

Luo JJ, Masson S, Roeckner E, Madec G, Yamagata T (2005)

Reducing climatology bias in an ocean–atmosphere CGCM with

improved coupling physics. J Clim 18:2344–2360

Luo JJ, Masson S, Behera S, Yamagata T (2007) Experimental

forecasts of the Indian Ocean Dipole using a coupled OAGCM.

J Clim 20:2178–2190

Luo JJ, Masson S, Behera S, Yamagata T (2008) Extended ENSO

predictions using a fully coupled ocean–atmosphere model.

J Clim 21:84–93

Luo JJ, Zhang R, Behera S, Masumoto Y, Jin FF, Lukas R, Yamagata

T (2009) Interaction between El Nino and extreme Indian ocean

dipole. J Clim. doi:10.1175/2009JCLI3104.1

Madden RA, Julian PR (1994) Observations of the 40–50 day tropical

oscillation: a review. Mon Wea Rev 122:813–837

Madec G, Delecluse P, Imbard M, Lvy C (1998) Opa 8.1 Ocean

general circulation model reference manual. Tech. Rep. Note 11,

LODYC/IPSL, Paris, France

Maloney ED (2009) The moist static energy budget of a composite

tropical intraseasonal oscillation in a climate model. J Clim

22:711–729

866 R. S. Ajayamohan et al.: BSISO propagation: role of internal processes

123

Neelin JD, Su H (2005) Moist teleconnection mechanisms for the

tropical South America and Atlantic sector. J Clim 18:3928–

3950

Roeckner E, Arpe K, Bengtsson L, Christoph M, Claussen M, Dmenil

L, Esch M, Giorgetta M, Schlese U, Schulzweida U (1996) The

atmospheric general circulation model ECHAM-4: model

description and simulation of present-day climate. The Max

Planck Institute for Meteorology, Hamburg, Germany, report

no:218, 90 pp

Sengupta D, Goswami BN, Senan R (2001) Coherent intraseasonal

oscillations of ocean and atmosphere during the Asian summer

monsoon. Geophys Res Lett 28:4127–4130

Shinoda T, Hendon HH, Glick J (1998) Intraseasonal variability of

surface fluxes and sea surface temperature in the tropical western

Pacific and Indian Oceans. J Clim 11:2668–2685

Slingo JM, Inness PM, Sperber KR (2005) Modelling the MJO. In:

Lau WKM, Waliser DE (eds) Intraseasonal variability in the

atmosphere–ocean climate system, chap 11. Praxis Springer,

Berlin, pp 361–388

Sperber KR, Annamalai H (2008) Coupled model simulations of

boreal summer intraseasonal (3050 day) variability, part 1:

Systematic errors and caution on use of metrics. Clim Dyn

31:345–372. doi:10.1007/s00382-008-0367-9

Sperber KR, Gualdi S, Legutke S, Gayler V (2005) The Madden-

Julian oscillation in ECHAM4 coupled and uncoupled general

circulation models. Clim Dyn 25:117–140

Tian B, Waliser DE, Fetzer EJ, Lambrigtsen BH, Yung TL, Wang B

(2006) Vertical moist thermodynamical structure and spatio-

temporal evolution of the MJO in AIRS observations. J Atmos

Sci 63:2462–2485

Valcke S, Terray L, Piacentini A (2000) Oasis 2.4 ocean atmosphere

soil user’s guide. Tech. Rep. TR/CGMC/00-10, CERFACS, 42

Avenue Coriolis, 31057 Toulouse, 85 pp. http://citeseer.ist.psu.

edu/valcke00oasis.html

Vecchi GA, Harrison DE (2002) Monsoon breaks and subseasonal sea

surface temperature variability in the Bay of Bengal. J Clim

15:1485–1493

Waliser DE (2006) Intraseasonal variability. In: Wang B (ed) The

Asian monsoon, chap 5. Praxis Springer, Chichester, pp 203–258

Waliser DE, Murtugudde R, Lucas L (2004) Indo-Pacific response to

atmospheric intraseasonal variability. Part ii: boreal summer and

the intraseasonal oscillation. J Geophys Res 109:C03030. doi:

10.1029/2003JC002002

Wang B (2005) Theory. In: Lau WKM, Waliser DE (eds) Intrasea-

sonal variability in the atmosphere–ocean climate system, chap

10. Praxis Springer, Berlin, pp 307–360

Wang B, Rui H (1990) Synoptic climatology of transient tropical

intraseasonal convection anomalies: 1975–1985. Met Atmos

Phys 44:43–61

Wang B, Xie X (1997) A model for the boreal summer intraseasonal

oscillations. J Atmos Sci 54:72–86

Webster PJ (1983) Mechanism of monsoon low-frequency variability:

surface hydrological effects. J Atmos Sci 40:2110–2124

Webster PJ, Yang S (1992) Monsoon and ENSO: selectively

interactive systems. Q J Meteor Soc 118:877–920

Woolnough SJ, Slingo JM, Hoskins BJ (2000) The relationship

between convection and sea surface temperature anomalies on

intraseasonal time scales. J Clim 13:2086–2104

Woolnough SJ, Slingo JM, Hoskins BJ (2001) The organization of

tropical convection by intraseasonal sea surface temperature

anomalies. Q J R Meteor Soc 127:888–907

Wu MLC, Schubert SD, Suarez MJ, Pegion PJ, Waliser DE (2006)

Seasonality and meridional propagation of the MJO. J Clim

19:1901–1921

Yamagata T, and Y. Hayashi (1984) A simple diagnostic model for

the 30–50 day oscillation in the tropics. J Meteor Soc Jpn

62(5):709–7117

Yamagata T, Behera SK, Luo JJ, Masson S, Jury MR, Rao SA (2004)

Coupled ocean–atmosphere variability in the tropical Indian

Ocean. In: Wang C, Xie SP, Carton JA (eds) In earth climate: the

ocean–atmosphere interaction. Geophys. Monogr. Ser., vol 147.

AGU, Washington, DC, pp 189–212

R. S. Ajayamohan et al.: BSISO propagation: role of internal processes 867

123