An Enhanced Moisture Convergence–Evaporation Feedback Mechanism for MJO Air–Sea Interaction

Andrew G. Marshall Bureau of Meteorology Research Centre, Melbourne, Australia

Search for other papers by Andrew G. Marshall in
Current site
Google Scholar
PubMed
Close
,
Oscar Alves Bureau of Meteorology Research Centre, Melbourne, Australia

Search for other papers by Oscar Alves in
Current site
Google Scholar
PubMed
Close
, and
Harry H. Hendon Bureau of Meteorology Research Centre, Melbourne, Australia

Search for other papers by Harry H. Hendon in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Simulations using an atmospheric model forced with observed SST climatology and the same atmospheric model coupled to a slab-ocean model are used to investigate the role of air–sea interaction on the dynamics of the MJO. Slab-ocean coupling improved the MJO in Australia’s Bureau of Meteorology atmospheric model over the Indo-Pacific warm pool by reducing its period from 70–100 to 45–70 days, thereby showing better agreement with the 30–80-day observed oscillation.

Air–sea coupling improves the MJO by increasing the moisture flux in the lower troposphere prior to the passage of active convection, which acts to promote convection and precipitation on the eastern flank of the main convective center. This process is triggered by an increase in surface evaporation over positive SST anomalies ahead of the MJO convection, which are driven by the enhanced shortwave radiation in the region of suppressed convection. This in turn generates enhanced convergence into the region, which supports evaporation–wind feedback in the presence of weak background westerly winds. A subsequent increase in low-level moisture convergence acts to further moisten the lower troposphere in advance of large-scale convection in a region of reduced atmospheric pressure. This destabilizing mechanism is referred to as enhanced moisture convergence–evaporation feedback (EMCEF) and is utilized to understand the role of air–sea coupling on the observed MJO. The EMCEF mechanism also reconciles traditionally opposing ideas on the roles of frictional wave–conditional instability of the second kind (CISK) and wind–evaporation feedback. These results support the idea that the MJO is primarily an atmospheric phenomenon, with air–sea interaction improving upon, but not critical for, its existence in the model.

* Current affiliation: Met Office, Hadley Centre, Exeter, United Kingdom

Corresponding author address: Andrew G. Marshall, Hadley Centre, Met Office, FitzRoy Rd., Exeter EX1 3PB, United Kingdom. Email: andrew.marshall@metoffice.gov.uk

Abstract

Simulations using an atmospheric model forced with observed SST climatology and the same atmospheric model coupled to a slab-ocean model are used to investigate the role of air–sea interaction on the dynamics of the MJO. Slab-ocean coupling improved the MJO in Australia’s Bureau of Meteorology atmospheric model over the Indo-Pacific warm pool by reducing its period from 70–100 to 45–70 days, thereby showing better agreement with the 30–80-day observed oscillation.

Air–sea coupling improves the MJO by increasing the moisture flux in the lower troposphere prior to the passage of active convection, which acts to promote convection and precipitation on the eastern flank of the main convective center. This process is triggered by an increase in surface evaporation over positive SST anomalies ahead of the MJO convection, which are driven by the enhanced shortwave radiation in the region of suppressed convection. This in turn generates enhanced convergence into the region, which supports evaporation–wind feedback in the presence of weak background westerly winds. A subsequent increase in low-level moisture convergence acts to further moisten the lower troposphere in advance of large-scale convection in a region of reduced atmospheric pressure. This destabilizing mechanism is referred to as enhanced moisture convergence–evaporation feedback (EMCEF) and is utilized to understand the role of air–sea coupling on the observed MJO. The EMCEF mechanism also reconciles traditionally opposing ideas on the roles of frictional wave–conditional instability of the second kind (CISK) and wind–evaporation feedback. These results support the idea that the MJO is primarily an atmospheric phenomenon, with air–sea interaction improving upon, but not critical for, its existence in the model.

* Current affiliation: Met Office, Hadley Centre, Exeter, United Kingdom

Corresponding author address: Andrew G. Marshall, Hadley Centre, Met Office, FitzRoy Rd., Exeter EX1 3PB, United Kingdom. Email: andrew.marshall@metoffice.gov.uk

1. Introduction

The impact of the air–sea interaction on the behavior of the Madden–Julian oscillation (MJO; Madden and Julian 1971) has been a major area of research interest particularly since observations from the Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) showed that sea surface temperature (SST) in the Indo-Pacific warm pool is modulated by the passage of the MJO (e.g., Weller and Anderson 1996; Hendon and Glick 1997). Intraseasonally varying SSTs associated with the MJO contribute to the intraseasonal variation of heat and moisture flux variations produced by the MJO (e.g., Krishnamurti et al. 1988; Shinoda et al. 1998), giving rise to the hypothesis that the strength, frequency, propagation, and maintenance of the MJO may critically depend on the interaction–feedback with an active upper ocean (e.g., Flatau et al. 1997; Sperber et al. 1997; Wang and Xie 1998).

Recent modeling studies have reinforced the notion that the intraseasonal SST anomalies produced by the MJO can feed back onto the atmospheric circulation associated with the MJO (e.g., Schubert and Wu 2001; Woolnough et al. 2001; Wu et al. 2002). These studies show that an “MJO-like” atmospheric disturbance is generated in response to an imposed intraseasonally varying SST anomaly that is representative of that produced by the observed MJO. The implication is that MJO-induced SST anomalies have sufficient amplitude and appropriate structure to generate a response in the atmosphere that is relevant to the MJO. The notion that two-way coupling is important has been further demonstrated in simulations whereby incorrect phasing between an imposed intraseasonally varying SST and the resultant surface flux anomalies has been remedied with coupling to an interactive ocean (e.g., Matthews 2004; Fu et al. 2003; Fu and Wang 2004; Zheng et al. 2004).

Various mechanisms by which an active upper ocean feeds back onto and improves the simulation of the MJO have been suggested. For example, intraseasonally varying SSTs have been found to destabilize the atmosphere to the east of the MJO-induced convection through an increase in moist static energy (Flatau et al. 1997), an increase in low-level convergence (Wang and Xie 1998; Kemball-Cook et al. 2002), and an enhancement of low-level moisture (Waliser et al. 1999), thereby promoting coherent eastward propagation at the MJO time and space scales. Such proposed interactions critically depend on the cooperative phasing of the surface fluxes of heat (primarily latent), radiation (primarily short wave), and momentum, such that the SST warms in advance of enhanced convection and cools in advance of suppressed convection (Shinoda et al. 1998). The lack of any significant impact of coupling on the MJO in some model simulations has been attributed to improper phasing of the component surface fluxes, which partly stems from biases in the simulated mean state (Hendon 2000; Kemball-Cook et al. 2002).

The motivation for the present study is to explore further the impact and mechanism of ocean–atmosphere coupling associated with the MJO. Realistic simulation of the MJO in both uncoupled and coupled climate models remains an elusive goal (e.g., Lin et al. 2006; Zhang et al. 2006). An emerging consensus, however, is that coupling acts to subtly modify a preexisting MJO-like disturbance in the uncoupled simulations (e.g., Waliser et al. 1999; Watterson 2002; Zhang et al. 2006): no general circulation model (GCM) to date has simulated the MJO in the coupled version when it was completely absent from the uncoupled version. However, there are also clear model dependencies for simulation of the MJO, especially associated with the parameterization of moist convection (e.g., Wang and Schlesinger 1999; Maloney 2002), and different models suggest different mechanisms for coupled atmosphere–ocean feedback (e.g., Flatau et al. 1997; Wang and Xie 1998; Waliser et al. 1999).

We will build on the work of Waliser et al. (1999) and others to address the impact and mechanism of coupling on the simulation of the MJO in an atmospheric GCM coupled to a slab mixed layer ocean. The uncoupled GCM used here simulates an MJO with realistic structure, but it is much less spectrally peaked at 40–50 days than observed. The focus here will be on the mechanism by which coupling acts to sharpen the spectral peak associated with the MJO. Model experiments will be of sufficient length (∼70 yr) so as to be able to detect systematic yet subtle impacts of coupling in the presence of large year-to-year variations in MJO activity. Use of a slab mixed layer ocean is justified by the fact that the relevant intraseasonal SST anomalies driven by the MJO in the warm pool of the Indian and Pacific Oceans result primarily from variations of surface heat flux and are adequately simulated by one-dimensional mixed layer ocean models (e.g., Shinoda and Hendon 1998). Recent research—making use of new satellite SST estimates that are not influenced by the extensive cloud cover and rainfall pervading the equatorial Indian Ocean where the MJO has large amplitude—suggests large and important SST–atmosphere interactions associated with the MJO that might not be fully described using a slab-ocean mixed layer (e.g., Harrison and Vecchi 2001; Duvel et al. 2004). Nonetheless, use of a slab-ocean mixed layer is an appropriate initial simplification.

A description of the control and slab-ocean experiments is presented in section 2. Section 3 describes the general characteristics of the MJO in the uncoupled and coupled simulations in summer and winter seasons. Section 4 contains a detailed analysis of the coupling mechanism during southern summer over the Indo-Pacific warm pool. The summary and conclusions are presented in section 5.

2. Experiments

a. Atmospheric model

We use the Bureau of Meteorology Research Centre’s atmospheric GCM version 3.0 (BAM3; Colman et al. 2005), which is a unified climate–weather prediction model with a spectral truncation of T47 and 17 vertical levels (Zhong et al. 2004). Details of the physical parameterization are provided in Colman et al. (2005), but here we note that moist convection is parameterized with a mass flux scheme closed on the time rate of change of convective available potential energy (CAPE; Tiedtke 1989; Nordeng 1994).

b. Uncoupled integration

A 70-yr control simulation is conducted by imposing daily varying climatological SSTs that are calculated from the monthly phase 2 of the Atmospheric Model Intercomparison Project (AMIP2; Gates 1992) data spanning the years 1979–97. This uncoupled experiment will hereafter be referred to as the control (CTL) experiment.

c. Slab model and coupled integration

To investigate the impact of ocean–atmosphere coupling while retaining a realistic mean SST, we introduce a simple slab-ocean anomaly model to compute prognostic SST anomalies. Based on Waliser et al. (1999), the simple slab-ocean SST anomaly is governed by
i1520-0469-65-3-970-e1
where T ′ is the SST anomaly, F′ is the net surface heat flux anomaly, H is a fixed mixed layer depth, γ is a damping factor, ρ is the density of seawater, and Cp is the specific heat capacity of water. In the coupled configuration, the atmosphere feels the SST anomaly (T ′) plus the climatological SST from the CTL simulation. To compute the heat flux anomaly (F′), the daily climatological heat flux is first calculated from the CTL. This climatology is then subtracted from the total heat flux in the coupled simulation prior to passing to (1). In this fashion, the same climatological SST as in the uncoupled integration is maintained while allowing for intraseasonal variability. The slab-ocean model is fully active equatorward of 30° latitude; in the range of 30° to 50° latitude, the SST anomalies are linearly weighted from 1.0 to 0.0, respectively. The damping term −γT ′ is a simplification of neglected processes including ocean internal dynamics. The γ is set to 60 (day)−1, which is similar to the value of 50 (day)−1 used by Waliser et al. (1999).

Three slab-ocean model experiments were run using fixed mixed layer depths of 10, 30, and 50 m. The 50-m slab model produced intraseasonally varying SST anomalies of the smallest magnitude (up to around 0.2°C for individual MJO events), while the 10-m slab model generated the largest SST anomalies, with maximum perturbations of >1°C for individual MJO events. The 30-m slab model produced SST anomalies similar to those typically observed, around 0.5°C for individual MJO events (e.g., Krishnamurti et al. 1988; Hendon and Glick 1997). The shallower slab produces a more drastic impact on the MJO (e.g., Watterson 2002), but we chose to explore the coupled behavior in the experiment with a 30-m slab, as this produces the most realistic SST anomalies. A 70-yr coupled integration is performed, which we refer to as SLAB.

d. Observations

Simulated MJO behavior is also compared to observed behavior as depicted in the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis 1 dataset (NNR1; Kalnay et al. 1996). Here we utilize 24 yr of analyzed surface zonal wind as obtained from the National Oceanic and Atmospheric Administration (NOAA)/Earth System Research Laboratory and Cooperative Institute for Research in Environmental Sciences (CIRES) Climate Diagnostics Center in Boulder, Colorado. Variations of deep tropical convection are inferred with the daily mean estimates of outgoing longwave radiation (OLR) described by Liebmann and Smith (1996).

3. Results

a. MJO behavior in an uncoupled simulation

Wavenumber–frequency power spectra are used to assess the large-scale characteristics of the simulated MJO. Spectra are computed in 6-month seasons that are centered on Southern Hemisphere summer (SHS: November–April) and Northern Hemisphere summer (NHS: May–October). Considering these seasons separately may be important for assessing the dynamics of the MJO, since mechanisms driving the MJO may differ between NHS and SHS because of the observed climatological asymmetries about the equator (e.g., Kemball-Cook et al. 2002) and because of each hemisphere’s different proportions of land–ocean coverage.

Spectra are computed for surface zonal wind anomalies (seasonal cycle removed) that are averaged between 20°S and 0° for the SHS period, and between 0° and 20°N for the NHS period. The time-mean and linear trends were removed from each 6-month segment using a least squares fit, and then each segment was padded with zeros to 256 days. Power was then computed for each 256-day segment using Fourier transforms (nominal bandwidth 1/256 cycles day−1), and then averaged over all 70 segments. We also compare simulated behavior to spectra computed in a similar fashion using the surface zonal wind from NNR1, however we use only 24 yr of data (as opposed to 70 yr from the simulations).

The observed MJO is characterized by a single but broad spectral peak centered on eastward wavenumber 1 and periods of 35–80 days during both SHS and NHS (Figs. 1a,b; see also Salby and Hendon 1994). MJO power is weaker in NHS than in SHS, but its spectral character is otherwise similar in the two seasons. Similar results are obtained with OLR (not shown), however power is spread into wavenumbers 2–4 because of the localization of the MJO convective signal to the Indian and western Pacific Oceans (Salby and Hendon 1994).

The CTL experiment exhibits a similar preponderance of eastward power at low zonal wavenumbers for periods of 30–80 days, but the spectrum is broader than that observed (Fig. 1c). Furthermore, there appears to be two dominant intraseasonal spectral peaks in SHS: a “fast” mode with periods ∼30–40 days and a “slow” mode with periods 70–100 days. A fast mode is generic to many GCM simulations of the MJO (e.g., Slingo et al. 1996). A slow mode has been reported by Watterson (2002) in an uncoupled simulation of the Commonwealth Scientific and Industrial Research Organisation (CSIRO) climate model. In NHS, the CTL exhibits only one intraseasonal spectral peak at eastward wavenumber 1, but it is at a lower frequency (i.e., a 70–100-day period as compared to a 50-day period) and weaker than observed.

Geographical distributions of variance at MJO time and space scales were calculated to assess the horizontal characteristics of MJO variability during southern summer and northern summer (Fig. 2). Prior to computing variance, surface zonal wind was spectrally filtered to eastward wavenumbers 1–4 for periods of 20–120 days. The maximum MJO variance in observations is centered at about 10°S, 120°E in SHS and at about 10°N, 120°E in NHS. In SHS the region of maximum variance extends across much of the Indian Ocean and the western Pacific Ocean. In NHS it is mainly concentrated in the Bay of Bengal, the South China Sea, and the far western Pacific Ocean. The maximum variance is about a factor of 2 stronger in SHS than in NHS. The CTL simulates a similar distribution of MJO variance as observed (Figs. 2c,d). However, in SHS, two distinct regions of maximum variance are evident in the Indian Ocean and western Pacific Ocean, and there is a spurious local maximum north of the equator at about 130°E. In NHS, the CTL simulates a maximum MJO variance in about the right location, but it is approximately 50% stronger than observed.

We apply further spectral filtering to explore the geographical distribution of the slow and fast intraseasonal variations identified in Fig. 1. We repeat our variance calculations, but now using data filtered for two period ranges: 20–45 days and 45–120 days (Fig. 3). During SHS the slow intraseasonal variance in CTL predominantly occurs west of Indonesia over the south equatorial Indian Ocean (Fig. 3a), while the fast intraseasonal variance occurs over the south equatorial western–central Pacific Ocean (Fig. 3b). Hence, the peak in variance in the Indian Ocean in Fig. 2c is predominantly due to the slower component of MJO variability, while that in the western Pacific is due to the faster component of MJO variability. In NHS there is no strong indication of separate fast and slow modes by the spectra of Fig. 1. Similarly, there is no geographical separation of the fast and slow intraseasonal variance (not shown), with both the slow and fast variances showing maxima around 15°–20°N in the western Pacific Ocean similar to Fig. 2d.

The distinction between the slow mode in the Indian Ocean and the fast mode in the western Pacific Ocean in SHS is demonstrated further in Fig. 4a, which shows a sample 8-month time–longitude section of zonal wind anomalies averaged between 20°S and the equator. The thin shaded contours show raw zonal wind anomalies; the solid thick contours show westerly wind anomalies associated with the slow intraseasonal mode (45–120 days); and the thick dashed contours show westerly wind anomalies associated with the fast intraseasonal mode (20–45 days). These plots confirm that in the CTL simulation, the MJO propagates slowly across the Indian Ocean and faster over the western Pacific Ocean, prior to decaying near the date line. This differing behavior may partly be explained by the decoupling of the circulation anomaly from the convection anomaly as the MJO moves over colder waters of the central Pacific (Hendon and Salby 1994). For instance, Fig. 4b shows a convective (negative OLR) anomaly propagating in conjunction with the slow-moving westerly anomaly over the Indian Ocean and far western Pacific Ocean; however, east of about 150°E, where the eastward propagation of the westerly anomaly speeds up (Fig. 4a), the convective anomaly appears to decay. Analysis of the CTL suggests that this decoupling occurs farther west than observed.

b. Coupled simulation

Wavenumber–frequency spectra of surface zonal wind anomalies from the SLAB experiment are shown in Fig. 1e for SHS and in Fig. 1f for NHS. Coupling to the slab ocean sharpens the spectral peak associated with the MJO, with now a single maximum occurring at a period of about 70 days for both SHS and NHS. This appears to result from power at the higher frequencies being reduced and power associated with the slower mode shifting to a slightly higher frequency. The strength of the MJO in the SLAB simulation, however, remains similar to that in the CTL experiment. Some studies have found that slab-ocean coupling acts to increase variability at MJO scales (e.g., Watterson 2002), but this does not appear to be the case in our model.

Inspection of time–longitude sections of OLR and surface zonal wind from the SLAB simulation (not shown) suggests that during SHS the MJO is characterized by a 50–75-day variability across the Indian and far western Pacific Oceans and a 30–40-day variability in the central Pacific. This suggests that air–sea coupling acts to speed up the slow mode of propagation across the Indian and far western Pacific Oceans while not significantly affecting the faster mode of propagation over the central Pacific. The spatial distribution of MJO variance in the SLAB experiment during SHS (not shown) closely resembles that of observations, now with a single maximum centered on about 10°S, 130°E.

4. Mechanism of coupling

a. Composite analysis for SHS

To explore the mechanism for the impact of air–sea coupling associated with the MJO we focus on the behavior during SHS when coupling appears to have its greatest impact. We form composites of the MJO based on its convective signal. The MJO events are identified when MJO-filtered OLR anomalies (eastward wavenumbers 1–4 with periods of 20–120 days), averaged from the equator to 20°S along 100°E, are less than −10 W m−2 (this is approximately one standard deviation of the MJO-filtered OLR anomaly averaged along this longitude). Each individual event is then combined into a composite by first determining when the maximum in convective activity (defined as the minimum in anomalous OLR) passed over 100°E for each event. This is referred to as day 0 of the composite. The number of individual MJO events in each composite is summarized in Table 1. The 70-yr integrations provide ∼150 individual MJO events, which allow for the detection of relatively small but significant differences in the composites from the two experiments.

The composite MJO in both the CTL and the SLAB experiments exhibits eastward propagation similar to that in the observed MJO; however, minimum OLR at day 0 is slightly weaker in CTL (−20 W m−2) than in the SLAB (–23 W m−2) experiment. At a lag of −10 days, the SLAB OLR anomaly is centered about 5° longitude farther west than that of the CTL simulation (Fig. 5), implying faster propagation in the SLAB simulation. The estimated eastward phase speeds are ∼3–4 and 4–5 m s−1 for the CTL and SLAB simulations, respectively. Such differences in phase speed are consistent with the spectral analysis in Fig. 1. The faster phase speed in the SLAB experiment is closer to reality (e.g., Hendon and Salby 1994). Both the CTL and SLAB also produce an increase in phase speed when the anomalous convection enters the Pacific Ocean (not shown), consistent with the observed behavior in this region (Hendon and Salby 1994).

b. Air–sea interaction

The increased phase speed of the MJO in the SLAB experiment over that in the CTL experiment is attributed to air–sea interaction. An indication of this air–sea interaction is the resultant intraseasonal SST anomalies that propagate eastward across the warm pool ahead of the convective anomaly in the SLAB experiment (Fig. 6). Positive SST anomalies are centered about 30° of longitude to the east of enhanced convection in the vicinity of clear skies, and negative SST anomalies are centered about 30° longitude to the west. The magnitude and spatial scale of these SST anomalies and their approximate quadrature relationship with the convection anomaly are similar to those observed (e.g., Hendon and Glick 1997).

To assess the role of these SST anomalies for the eastward propagation of the MJO, we examine in detail the phasing and evolution of the surface heat fluxes and their relationship to the convective anomaly in the SLAB experiment, and how they differ from that in the CTL experiment. The SLAB anomalies are also compared with those from the observed composite. This analysis is presented in Fig. 7; note that the sign convention of the heat fluxes is such that latent heat flux anomalies are positive into the atmosphere, while shortwave and total heat flux anomalies are positive into the ocean. We first examine the longitudinal variation along the equator at day 0. In the CTL experiment, enhanced convection (low OLR) centered on 100°E is accompanied by anomalous surface westerlies that are shifted west by about 20° longitude and weaker anomalous easterlies extending to the east (Figs. 7a,b). The basic state in these regions is predominantly westerly, but weak in magnitude (<1 m s−1; Fig. 8). Similar to observed behavior (Fig. 7g), positive anomalous latent heat flux into the atmosphere occurs to the west of the convective anomaly in conjunction with anomalous westerly flow. Easterly anomalies to the east of the convective anomaly are associated with decreased latent heat flux and suppressed convection that spans the region ∼120°E–180°. Note, however, that the negative latent heat flux anomaly extends westward to about 90°E, which means that reduced latent heat flux underlies the eastern flank of the convective anomaly. Such a phasing is not conducive to eastward propagation of convection. Presumably, atmospheric moisture convergence on the eastern flank of the convective anomaly must be overcoming this negative latent heat flux to support eastward propagation of enhanced convection. Additionally, note that the region of enhanced convection is also a region of low surface pressure (low geopotential height) and high surface humidity, while the region of suppressed convection to the east is associated with weak negative humidity anomaly and near-zero surface pressure anomaly.

The impact of coupling is revealed through a similar analysis of the SLAB simulation (Figs. 7c,d). On the western flank of the region of enhanced convection (westward of ∼90°E), the composites from the SLAB and the CTL are similar, with enhanced latent heat flux being driven primarily by anomalous surface westerlies that act to increase the surface wind speed. However, on the eastern flank of the region of enhanced convection, significant differences between the two simulations are evident. Whereas in the CTL the OLR anomaly is positive east of about 120°E (Fig. 7b), the OLR anomaly in the SLAB simulation (Fig. 7d) does not become positive until about 140°E, as is also seen in the observations (Fig. 7h), and is near zero in the region ∼120°–150°E. Furthermore, and more strikingly, the latent heat flux anomaly is negative east of about 90°E in the CTL experiment (Fig. 7a), while it is near zero or even positive in the region ∼90°–140°E in the SLAB experiment (Fig. 7c) as it is in NNR1 (Fig. 7g). In other words, the latent heat flux is negative on the eastern flank of the convective anomaly in the CTL simulation, but it is near zero or even positive there in the SLAB simulation, which agrees with observations.

These and other differences are highlighted in Figs. 7e,f, which show the SLAB minus the CTL composite differences. The two-dimensional structures of some of these differences are displayed in Fig. 9. Relative to the CTL experiment, the SLAB composite exhibits enhanced convection on the eastern flank of the main convective anomaly (roughly 90°–140°E), accompanied by anomalously westerly surface winds (Figs. 7e, 9c), enhanced latent heat flux into the atmosphere (Figs. 7e, 9a), increased specific humidity (Figs. 7f, 8e), and reduced surface pressure farther east (Figs. 7f, 9e). The relative increase in the westerly wind anomaly is up to ∼0.5 m s−1, and the increased latent heat flux is approximately 5–10 W m−2. Recall that in an absolute sense, the winds are anomalously easterly to the east of the convection in both the CTL and the SLAB simulations. Coupling thus acts to reduce these easterly anomalies. This westerly anomaly in the SLAB relative to the CTL therefore acts to relatively increase the wind speed (Fig. 9b) and surface evaporation (Fig. 9a) ahead of the convective core, which is in the region of the positive SST anomaly. This implies that a type of evaporation–wind feedback mechanism is playing a role in favoring convection on the eastern flank of the MJO convective anomaly in the SLAB simulation. Indeed, calculations reveal that the contribution of wind speed to the change in total heat flux ahead of the MJO is around 50% (not shown). Note also that a relative decrease in evaporation occurs to the west of the main convective core, which is the region of the negative SST anomaly. Slab-ocean coupling results in a relative easterly anomaly there that in turn results in reduced wind speed because of the mean westerly basic state (Fig. 8). Hence, coupling reduces the positive latent heat flux anomaly to the west of convection.

At this point we should note that Fig. 7e shows peak anomalous westerlies at 120°E and peak easterlies at 170°E, with near-zero values between 130° and 165°E that are consistent with the weak gradient in geopotential height seen in this region (Fig. 7f). The near-zero nature of these zonal wind anomalies is due to the meridional averaging (5°–10°S) used to produce the plot. Figure 9c shows that weak westerlies extend to only 140°E equatorward of 5°S, but to 165°E at 10°S; Fig. 7e captures this feature.

Low-level specific humidity increases (decreases) by up to 0.2 g kg−1 around 15°–30°E (W) of the reference longitude in the SLAB experiment relative to the CTL (Figs. 7f, 9d). This supports the premise that the increased (decreased) latent heating over positive (negative) SST anomalies contributes to the moisture flux on intraseasonal scales (e.g., Kemball-Cook and Weare 2001). An associated relative decrease (increase) in low-level atmospheric pressure also occurs ahead of (behind) the MJO convection, as seen in the SLAB–CTL 1000-mb geopotential height anomaly (Figs. 7f, 9e).

Relative moistening of the lower troposphere in the SLAB simulation compared to the CTL to the east of the convective center is sustained/amplified through increased low-level convergence, which acts to further enhance the incoming moisture flux at low levels in a region of reduced atmospheric pressure. An increase in anomalous moisture convergence begins about two weeks ahead of the convective center (Fig. 10a) and continues to accumulate at low levels as the MJO convection intensifies. An associated increase in anomalous specific humidity extends from low levels into the upper troposphere, concurrent with the increase in moisture convergence and reaching a maximum of 0.3 g kg−1 near the 700-mb pressure level around a lag of −5 days (Fig. 10b). (Note that the negative moisture change seen near the surface in Fig. 10b is an artifact of meridional averaging over the tropical Maritime Continent and corresponds to the weak negative SLAB–CTL surface specific humidity change seen east of the reference longitude over the Malaysian island of Sumatra in Fig. 9d.) After day 0, a reduction in anomalous specific humidity extends from the surface to upper levels in the SLAB simulation relative to the CTL, which is associated with an increase in low-level moisture divergence. A calculation of the moist static energy difference throughout the atmosphere (not presented here) shows the same evolutionary behavior as the specific humidity profile of Fig. 10b.

c. Proposed coupling mechanism

In the absence of coupling, the model’s MJO exhibits reduced latent heat flux into the atmosphere and increased shortwave radiation into the ocean ahead of (east of) the convective center. When coupled to the mixed layer ocean, these surface fluxes act together to warm the SST ahead of the convective center. This warm SST then acts to mitigate the decreased evaporation ahead of the convective center. The warm SST also acts to hydrostatically decrease pressure, thereby increasing surface convergence. Surface humidity and moisture convergence are thus increased ahead of the convective center in the SLAB experiment relative to the CTL. Shinoda et al. (1998) estimate that the observed SST anomaly of about 0.25°C associated with the MJO acts to mitigate the decreased latent heat flux ahead of the convective center by about 5 W m−2 (climatological wind speed was assumed), which is similar to the relative increase in the SLAB experiment over that in the CTL. However, in the experiments here, the negative wind speed anomaly ahead of the convective center is also mitigated in the SLAB experiment. But close inspection of Fig. 7c reveals that just east of the convective center (110°–140°E) the latent heat flux anomaly is actually positive, even though the zonal wind anomaly is slightly negative (hence resulting in a slightly negative wind speed anomaly because this is a region of mean westerlies; Fig. 8). Thus, it appears that the increased evaporation east of the convective center in the SLAB experiment relative to the CTL stems from the increase of SST that acts to in turn increase the air–sea humidity difference. The buildup of low-level moisture east of the convective center over warm SSTs in the SLAB simulation compared to the CTL subsequently leads to an increase in convection (reduction in OLR) and an increase in precipitation of up to 2 mm day−1 (not shown) centered around 120°–130°E, while equivalent reductions in precipitation and increases in OLR are found west of the reference longitude.

Air–sea coupling results in a feedback between evaporation, wind, and moisture over positive SST anomalies that develop ahead of the MJO-convective anomaly, thereby promoting eastward propagation of the MJO. We refer to this mechanism as enhanced moisture convergence–evaporation feedback (EMCEF), which is shown schematically in Fig. 11. It involves wind–evaporation feedback and enhanced low-level moisture convergence in the region of low surface pressure and warm SST to the east of the convective anomaly. The results consolidate the findings of Flatau et al. (1997), Wang and Xie (1998), Kemball-Cook et al. (2002), and Waliser et al. (1999) in one comprehensive mechanism. The novel interpretation here is that the SST anomaly acts to mitigate the deleterious impact of the MJO-induced decreased latent heat flux ahead of the convective center in the absence of coupling.

Also portrayed in Fig. 11 (cf. Figs. 7, 9) is a relative reduction in surface evaporation over negative SST anomalies to the west of the MJO convection in the SLAB simulation compared to the CTL. The negative SST anomaly acts to decrease the air–sea humidity difference. Coincident westerly anomalies to the west of convection in the SLAB experiment are also weaker than in the CTL. Thus, positive latent heat flux is not as strong in the SLAB simulation compared to the CTL in the wake of convection, which also acts to promote eastward movement of the convective center in the SLAB integration. Overall, then, coupling appears to reduce the magnitude of the latent heat flux variation driven by the MJO, as speculated by Shinoda et al. (1998), thereby leading to less inhibition for the eastward propagation of the convective anomaly.

5. Conclusions

Air–sea coupling acts to modestly speed up and strengthen the MJO in the Bureau of Meteorology Research Centre (BMRC) atmospheric GCM over the Indo-Pacific warm pool by acting to reduce the surface moisture flux suppression prior to the passage of active convection. As a result, precipitation is increased around 20°–30° east of the convective center in the coupled experiment. This process is triggered by an increase in surface evaporation over the induced positive SST ahead of the MJO convection in the coupled simulation. The positive SST to the east of the convection is a result of a combination of reduced evaporation, associated with anomalous easterlies, and increased surface shortwave radiation, associated with reduced cloudiness. These flux anomalies occur in both the coupled and uncoupled integrations. But in the coupled simulation, the induced SST anomaly acts to mitigate the latent heat flux suppression just to the east of the convective center, resulting in a weakly positive latent heat flux anomaly in this region that matches observed behavior. Compared to the uncoupled experiment, there is an increase in precipitation and surface westerlies to the east of the convective center. Atmospheric pressure is also lowered over the positive SST anomaly. Together, these processes are indicative of an evaporation–wind feedback to the east of the large-scale convection in the presence of weak background westerly winds. The relative increase in low-level moisture convergence also acts to moisten the lower troposphere ahead of the convective center. As mentioned, this destabilizing mechanism is referred to as enhanced moisture convergence–evaporation feedback (EMCEF) and is depicted in the schematic of Fig. 11. It consolidates previous mechanisms into one general mechanism involving both wind–evaporation feedback and enhanced low-level moisture convergence in the region of low pressure to the east of the convective anomaly, but it operates only in a coupled sense. However, the current results suggest that the MJO is primarily an atmospheric phenomenon, with air–sea interaction improving upon, but not critical for, the existence of the MJO in the model.

Our study focused on the Indo-Pacific warm pool in SHS, where the EMCEF mechanism was shown to be responsible for a speeding up of the MJO propagation and for a modest increase in amplitude. The overall effect is to sharpen the spectral peak around the 50–60-day period associated with the MJO in the coupled simulation as compared to the uncoupled simulation. There is little impact of air–sea coupling on the MJO in the relatively cooler waters of the central and eastern Pacific Ocean. There are a few possible explanations for this. First, the faster propagation over the central Pacific Ocean may not allow the buildup of low-level moisture to take place before being advanced upon by the eastward-moving convective anomaly associated with the MJO. Second, the convective response to an SST anomaly is relatively small over colder water (e.g., Woolnough et al. 2000). Third, the easterly mean state in the central and eastern Pacific Ocean prevents the form of wind–evaporation feedback discussed above.

The power spectra for NHS showed a similar impact of air–sea coupling as did SHS. However, because of the larger proportion of landmass in the Northern Hemisphere, analysis of air/sea coupling during NHS is difficult. This is left as a topic for future work.

Coupling the atmosphere GCM used in this study to an ocean GCM, instead of a slab mixed layer ocean model, also results in a modest strengthening of the MJO, together with an overall increase of eastward propagation at intraseasonal frequencies (Fig. 1 from Zhang et al. 2006). It is not obvious, however, that this can be interpreted as stemming from an increase in phase speed of the existing MJO in the uncoupled run. Differences in experimental design hinder such an interpretation. In our SLAB experiments, interannual variability of the coupled mean state is minimal (i.e., SST anomalies are damped back to the climatological seasonal cycle with an e-folding time of 60 days). In the fully coupled run examined by Zhang et al. (2006), El Niño and other low-frequency coupled variations naturally develop. Therefore, the sample population is much more homogeneous in the present study, which allows for the detection of relatively subtle differences.

Intraseasonal variations of the mixed layer depth over the life cycle of the MJO (e.g., Hendon and Glick 1997) can have important ramifications for the thermodynamics of the mixed layer. For example, Shinoda and Hendon (1998) showed that a shallow mixed layer during the calm, clear warm phase of the MJO results in enhanced sensitivity of the mixed layer temperature to surface heat flux forcing. Conversely, a deep mixed layer during the windy, cloudy cool phase of the MJO results in reduced sensitivity. This implies that in the present study where the slab-ocean thickness is fixed, the sensitivity of the mixed layer to surface heat flux variations ahead of the MJO convection may be somewhat underestimated. The resultant increase in low-level moisture prior to the passage of active convection may be more pronounced with the use of a more realistic mixed layer ocean model whose depth varies intraseasonally. This is left as a topic for future work.

With the exception of Hendon (2000), in which deficiencies in the atmospheric GCM prevented the formation of coherent SST anomalies, the air–sea interaction in previous studies has generally resulted in an improved periodicity of the MJO to around 40–60 days. This result has been produced in studies in which the uncoupled GCM has produced MJO-like activity with a shorter-than-observed period of 25–30 days (e.g., Flatau et al. 1997) and in studies in which the simulated MJO propagates with a longer-than-observed period of 80–120 days (e.g., see also Watterson 2002). This tendency for coupling to result in an MJO with a period of about 50 days in models that have both too slow and too fast MJOs when uncoupled could be coincidental, indicative of poor consistency among a variety of current GCMs, or it may suggest that the MJO has a preferred period of propagation when coupled to an interactive ocean. If the latter, the relative phasing between SST and convection introduced in the presence of coupling may be the key. For example, Woolnough et al. (2001) suggest that the feedback between convection and surface conditions could mediate the eastward propagation speed of the SST anomalies, thus providing a characteristic speed for the MJO. That is, slow-moving SST anomalies will lead to enhanced convection and strengthened surface flux anomalies, which will in turn encourage faster growth and decay of the SST anomalies. Conversely, for fast-moving SST anomalies, the generation of surface fluxes will be inhibited, thus encouraging slower growth and decay of the SST anomalies. A model intercomparison study involving the coupling of a common mixed layer ocean to a range of present-day atmospheric GCMs may shed more light on this matter.

Finally, these results raise the issue of whether wind–evaporation feedback acts as a negative feedback in the absence of interactive SSTs. Although this suggestion is counter to the earlier work of Neelin et al. (1987), it is now apparent that their model was simulating convectively coupled Kelvin waves in easterly basic states rather than in the MJO (e.g., after Wheeler and Kiladis 1999). Hence, since the conclusions of Neelin et al. (1987) are not necessarily applicable to the MJO, the role of wind–evaporation feedback should perhaps be revisited in present-day GCMs that support MJO-like behavior in an uncoupled version.

Acknowledgments

The authors thank the Bureau of Meteorology Research Centre and Monash University for supporting Andrew Marshall’s Ph.D., during which this work was conducted. The authors also thank the three anonymous reviewers of this manuscript, whose comments and suggestions improved the overall quality of the paper. We would like to extend our thanks to Dr. Matthew Wheeler for his insights and contributions to the direction of this research and to Dr. Robert Colman, Dr. Lilia Lemus-Deschamps, Dr. Scott Power, Dr. Guomin Wang, Ms. Aihong Zhong at BMRC, and Mr. Mark Williams at the Victorian Regional Office (VRO). The NCEP–NCAR reanalyses were provided by the NOAA/CIRES Climate Diagnostics Centre in Boulder, Colorado.

REFERENCES

  • Colman, R., and Coauthors, 2005: BMRC Atmospheric Model (BAM) version 3.0: Comparison with mean climatology. BMRC Research Rep. 108, Bureau of Meteorology, 66 pp.

  • Duvel, J. P., R. Roca, and J. Vialard, 2004: Ocean mixed layer temperature variations induced by intraseasonal convective perturbations over the Indian Ocean. J. Atmos. Sci., 61 , 10041023.

    • Search Google Scholar
    • Export Citation
  • Flatau, M., P. J. Flatau, P. Phoebus, and P. P. Niiler, 1997: The feedback between equatorial convection and local radiative and evaporative processes: The implications for intraseasonal oscillations. J. Atmos. Sci., 54 , 23732386.

    • Search Google Scholar
    • Export Citation
  • Fu, X. H., and B. Wang, 2004: Differences of boreal summer intraseasonal oscillations simulated in an atmosphere–ocean coupled model and an atmosphere-only model. J. Climate, 17 , 12631271.

    • Search Google Scholar
    • Export Citation
  • Fu, X. H., B. Wang, T. Li, and J. P. McCreary, 2003: Coupling between northward-propagating, intraseasonal oscillations and sea surface temperature in the Indian Ocean. J. Atmos. Sci., 60 , 17331753.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73 , 19621970.

  • Harrison, D. E., and G. A. Vecchi, 2001: January 1999 Indian Ocean cooling event. Geophys. Res. Lett., 28 , 37173720.

  • Hendon, H. H., 2000: Impact of air–sea coupling on the Madden–Julian oscillation in a general circulation model. J. Atmos. Sci., 57 , 39393952.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., and M. L. Salby, 1994: The life cycle of the Madden-Julian oscillation. J. Atmos. Sci., 51 , 22252237.

  • Hendon, H. H., and J. Glick, 1997: Intraseasonal air–sea interaction in the tropical Indian and Pacific Oceans. J. Climate, 10 , 647661.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kemball-Cook, S., and B. C. Weare, 2001: The onset of convection in the Madden–Julian oscillation. J. Climate, 14 , 780793.

  • Kemball-Cook, S., B. Wang, and X. Fu, 2002: Simulation of the intraseasonal oscillation in the ECHAM-4 model: The impact of coupling with an ocean model. J. Atmos. Sci., 59 , 14331453.

    • Search Google Scholar
    • Export Citation
  • Krishnamurti, T. N., D. K. Oosterhof, and A. V. Mehta, 1988: Air–sea interaction on the time scale of 30 to 50 days. J. Atmos. Sci., 45 , 13041322.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77 , 12751277.

    • Search Google Scholar
    • Export Citation
  • Lin, J-L., and Coauthors, 2006: Tropical intraseasonal variability in 14 IPCC AR4 climate models. Part I: Convective signals. J. Climate, 19 , 26652690.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., 2002: An intraseasonal oscillation composite life cycle in the NCAR CCM3.6 with modified convection. J. Climate, 15 , 964982.

    • Search Google Scholar
    • Export Citation
  • Matthews, A. J., 2004: Atmospheric response to observed intraseasonal tropical sea surface temperature anomalies. Geophys. Res. Lett., 31 .L14107, doi:10.1029/2004GL020474.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., I. M. Held, and K. H. Cook, 1987: Evaporation–wind feedback and low-frequency variability in the tropical atmosphere. J. Atmos. Sci., 44 , 23412348.

    • Search Google Scholar
    • Export Citation
  • Nordeng, T. E., 1994: Extended versions of the convective parameterization scheme at ECMWF and their impact on the mean and transient activity of the model in the tropics. ECMWF Research Department Tech. Memo. 206, European Centre for Medium-Range Weather Forecasts, 41 pp. [Available from ECMWF, Shinfield Park, Reading, Berkshire RG2 9AX, United Kingdom.].

  • Salby, M. L., and H. H. Hendon, 1994: Intraseasonal behavior of clouds, temperature, and motion in the tropics. J. Atmos. Sci., 51 , 22072224.

    • Search Google Scholar
    • Export Citation
  • Schubert, S. D., and M. L. Wu, 2001: Predictability of the 1997 and 1998 South Asian summer monsoon low-level winds. J. Climate, 14 , 31733191.

    • Search Google Scholar
    • Export Citation
  • Shinoda, T., and H. H. Hendon, 1998: Mixed layer modeling of intraseasonal variability in the tropical western Pacific and Indian Oceans. J. Climate, 11 , 26682685.

    • Search Google Scholar
    • Export Citation
  • Shinoda, T., H. H. Hendon, and J. D. Glick, 1998: Intraseasonal variability of surface fluxes and sea surface temperature in the tropical western Pacific and Indian Oceans. J. Climate, 11 , 16851702.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., and Coauthors, 1996: Intraseasonal oscillations in 15 atmospheric general circulation models: Results from an AMIP diagnostic subproject. Climate Dyn., 12 , 325357.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., J. M. Slingo, P. M. Inness, and W. K-M. Lau, 1997: On the maintenance and initiation of the intraseasonal oscillation in the NCEP/NCAR Reanalysis and the GLA and UKMO AMIP simulations. Climate Dyn., 13 , 769795.

    • Search Google Scholar
    • Export Citation
  • Spiegel, M. R., 1988: Schaum’s Outline of. Theory and Problems of Statistics. 2nd ed. McGraw-Hill, 504 pp.

  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev., 117 , 17791800.

    • Search Google Scholar
    • Export Citation
  • Waliser, D. E., K. M. Lau, and J-H. Kim, 1999: The influence of coupled sea surface temperatures on the Madden–Julian oscillation: A model perturbation experiment. J. Atmos. Sci., 56 , 333358.

    • Search Google Scholar
    • Export Citation
  • Wang, B., and X. Xie, 1998: Coupled modes of the warm pool climate system. Part I: The role of air–sea interaction in maintaining Madden–Julian oscillation. J. Climate, 11 , 21162135.

    • Search Google Scholar
    • Export Citation
  • Wang, W., and M. E. Schlesinger, 1999: The dependence on convection parameterization of the tropical intraseasonal oscillation simulated by the UIUC 11-layer atmospheric GCM. J. Climate, 12 , 14231457.

    • Search Google Scholar
    • Export Citation
  • Watterson, I. G., 2002: The sensitivity of subannual and intraseasonal tropical variability to model ocean mixed layer depth. J. Geophys. Res., 107 .4020, doi:10.1029/2001JD000671.

    • Search Google Scholar
    • Export Citation
  • Weller, R. A., and S. P. Anderson, 1996: Surface meteorology and air–sea fluxes in the western equatorial Pacific warm pool during the TOGA Coupled Ocean–Atmosphere Response Experiment. J. Climate, 9 , 19591990.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., and G. N. Kiladis, 1999: Convectively coupled equatorial waves: Analysis of clouds and temperature in the wavenumber–frequency domain. J. Atmos. Sci., 56 , 374399.

    • Search Google Scholar
    • Export Citation
  • Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2000: The relationship between convection and sea surface temperature on intraseasonal timescales. J. Climate, 13 , 20862104.

    • Search Google Scholar
    • Export Citation
  • Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2001: The organization of tropical convection by intraseasonal sea surface temperature anomalies. Quart. J. Roy. Meteor. Soc., 127 , 887907.

    • Search Google Scholar
    • Export Citation
  • Wu, M. L., S. Schubert, I-S. Kang, and D. Waliser, 2002: Forced and free intraseasonal variability over the South Asian monsoon region simulated by 10 AGCMs. J. Climate, 15 , 28622880.

    • Search Google Scholar
    • Export Citation
  • Zhang, C. D., M. Dong, S. Gualdi, H. H. Hendon, E. D. Maloney, A. Marshall, K. R. Sperber, and W. Q. Wang, 2006: Simulations of the Madden-Julian oscillation in four pairs of coupled and uncoupled global models. Climate Dyn., 27 , 573592.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., D. E. Waliser, W. F. Stern, and C. Jones, 2004: The role of coupled sea surface temperatures in the simulation of the tropical intraseasonal oscillation. J. Climate, 17 , 41094134.

    • Search Google Scholar
    • Export Citation
  • Zhong, A., O. Alves, A. Schiller, F. Tseitkin, and N. Smith, 2004: Results from a preliminary version of the ACOM2/BAM coupled seasonal forecast model. BMRC Research Rep. 95, Bureau of Meteorology, Australia, 45 pp.

Fig. 1.
Fig. 1.

Wavenumber–frequency spectra of surface zonal wind anomalies in (a) SHS for NNR1, (b) NHS for NNR1, (c) SHS for CTL, (d) NHS for CTL, (e) SHS for SLAB, and (f) NHS for SLAB. Contour interval is 0.001 m2 day s−2, with gray areas indicating spectral density values greater than 0.01 m2 day s−2.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 2.
Fig. 2.

Geographical distributions of variance at MJO scales (eastward wavenumbers 1–4 for periods of 20–120 days) for surface zonal wind anomalies in (a) SHS for NNR1, (b) NHS for NNR1, (c) SHS for CTL, and (d) NHS for CTL. Contour interval (CI) is 0.3 m2 s−2, with gray areas indicating variances greater than 0.9 m2 s−2.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 3.
Fig. 3.

Geographical distributions of variance at MJO scales for surface zonal wind anomalies in CTL for SHS. Spectral filtering is for eastward wavenumbers 1–4 for periods (a) 45–120 days and (b) 20–45 days. The CI is 0.3 m2 s−2, with gray areas indicating variances greater than 0.9 m2 s−2.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 4.
Fig. 4.

Time–lon section of 5-day mean (a) surface zonal wind and (b) OLR anomalies averaged 20°S–0° for an arbitrary 8-month period for the CTL in SHS. The thin shaded contours show raw zonal wind anomalies, with a CI of (a) 2 m s−1 and (b) 20 W m−2, and gray areas indicate negative anomalies. Solid thick contours indicate slow eastward-propagating anomalies (spectrally filtered for positive wavenumbers 1–4 and periods of 45–120 days), while dashed bold contours indicate fast eastward-propagating anomalies (spectrally filtered for positive wavenumbers 1–4 and periods of 20–45 days).

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 5.
Fig. 5.

Geographical locations of anomalous OLR for the SLAB (solid contours) and the CTL (dashed contours) simulations in southern summer, superimposed for various lag composites. Black contours represent −22 W m−2 OLR anomalies for day 0 composites based at 100°E, while gray contours represent −12 W m−2 OLR anomalies for a lag of −10 days.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 6.
Fig. 6.

Composite SST anomalies for the SLAB experiment at lags of −10, 0, and 10 days. Base longitude for the composites is 100°E. Contour interval is 0.05 K, with gray areas indicating negative values. A land–sea mask is applied such that skin temperatures over land are omitted. Bold contours are negative OLR anomalies, with a contour interval of 5 W m−2 and a minimum value (lowest bold contour) of −22 W m−2. Positive OLR anomalies are not shown.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 7.
Fig. 7.

Composite anomalies of latent heat and shortwave fluxes, surface zonal wind speed, and SST for (a) CTL, (c) SLAB, and (e) SLAB–CTL. Composite anomalies of total heat flux, OLR, surface specific humidity, and 1000-mb geopotential height for (b) CTL, (d) SLAB, and (f) SLAB–CTL. Data are shown as a function of longitude at day 0, averaged 5°–10°S. Base longitude for the composites is 100°E. Latent heat flux anomalies are positive into the atmosphere, while shortwave and total heat flux anomalies are positive into the ocean.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 8.
Fig. 8.

Composite absolute zonal wind speeds at day 0 for the CTL run. Base lon for the composite is 100°E. The CI is 1 m s−1, with gray (white) areas indicating mean westerlies (easterlies).

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 9.
Fig. 9.

Difference between the SLAB and the CTL composites at lag 0 of anomalous (a) surface latent heat flux, (b) surface wind speed, (c) zonal wind speed, (d) specific humidity, (e) geopotential height, and (f) total heat flux. Base lon for the composites is 100°E. CIs are (a) 5 W m−2, (b) 0.2 m s−1, (c) 0.2 m s−1, (d) 0.1 g kg−1, (e) 1 m, and (f) 5 W m−2. Gray areas indicate negative values. Bold contours incorporate regions of statistical significance at the 95% confidence level using a two-tailed significance test for the difference between two means (Spiegel 1988).

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 10.
Fig. 10.

SLAB–CTL difference composites (vertical sections) of anomalous (a) moisture convergence and (b) specific humidity as a function of lag time. Base longitude for the composites is 100°E. The CI is 2.5 × 10−9 s−1 and 0.1 g kg−1, respectively, with gray shaded areas indicating negative values.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Fig. 11.
Fig. 11.

Schematic illustrating the impact of air–sea coupling on the dynamics of the MJO in BAM.

Citation: Journal of the Atmospheric Sciences 65, 3; 10.1175/2007JAS2313.1

Table 1.

Number of individual MJO events selected for southern summer composites at 100°E for NNR1, the CTL simulation, and the SLAB simulation.

Table 1.
Save
  • Colman, R., and Coauthors, 2005: BMRC Atmospheric Model (BAM) version 3.0: Comparison with mean climatology. BMRC Research Rep. 108, Bureau of Meteorology, 66 pp.

  • Duvel, J. P., R. Roca, and J. Vialard, 2004: Ocean mixed layer temperature variations induced by intraseasonal convective perturbations over the Indian Ocean. J. Atmos. Sci., 61 , 10041023.

    • Search Google Scholar
    • Export Citation
  • Flatau, M., P. J. Flatau, P. Phoebus, and P. P. Niiler, 1997: The feedback between equatorial convection and local radiative and evaporative processes: The implications for intraseasonal oscillations. J. Atmos. Sci., 54 , 23732386.

    • Search Google Scholar
    • Export Citation
  • Fu, X. H., and B. Wang, 2004: Differences of boreal summer intraseasonal oscillations simulated in an atmosphere–ocean coupled model and an atmosphere-only model. J. Climate, 17 , 12631271.

    • Search Google Scholar
    • Export Citation
  • Fu, X. H., B. Wang, T. Li, and J. P. McCreary, 2003: Coupling between northward-propagating, intraseasonal oscillations and sea surface temperature in the Indian Ocean. J. Atmos. Sci., 60 , 17331753.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., 1992: AMIP: The Atmospheric Model Intercomparison Project. Bull. Amer. Meteor. Soc., 73 , 19621970.

  • Harrison, D. E., and G. A. Vecchi, 2001: January 1999 Indian Ocean cooling event. Geophys. Res. Lett., 28 , 37173720.

  • Hendon, H. H., 2000: Impact of air–sea coupling on the Madden–Julian oscillation in a general circulation model. J. Atmos. Sci., 57 , 39393952.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., and M. L. Salby, 1994: The life cycle of the Madden-Julian oscillation. J. Atmos. Sci., 51 , 22252237.

  • Hendon, H. H., and J. Glick, 1997: Intraseasonal air–sea interaction in the tropical Indian and Pacific Oceans. J. Climate, 10 , 647661.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kemball-Cook, S., and B. C. Weare, 2001: The onset of convection in the Madden–Julian oscillation. J. Climate, 14 , 780793.

  • Kemball-Cook, S., B. Wang, and X. Fu, 2002: Simulation of the intraseasonal oscillation in the ECHAM-4 model: The impact of coupling with an ocean model. J. Atmos. Sci., 59 , 14331453.

    • Search Google Scholar
    • Export Citation
  • Krishnamurti, T. N., D. K. Oosterhof, and A. V. Mehta, 1988: Air–sea interaction on the time scale of 30 to 50 days. J. Atmos. Sci., 45 , 13041322.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77 , 12751277.

    • Search Google Scholar
    • Export Citation
  • Lin, J-L., and Coauthors, 2006: Tropical intraseasonal variability in 14 IPCC AR4 climate models. Part I: Convective signals. J. Climate, 19 , 26652690.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., 2002: An intraseasonal oscillation composite life cycle in the NCAR CCM3.6 with modified convection. J. Climate, 15 , 964982.

    • Search Google Scholar
    • Export Citation
  • Matthews, A. J., 2004: Atmospheric response to observed intraseasonal tropical sea surface temperature anomalies. Geophys. Res. Lett., 31 .L14107, doi:10.1029/2004GL020474.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., I. M. Held, and K. H. Cook, 1987: Evaporation–wind feedback and low-frequency variability in the tropical atmosphere. J. Atmos. Sci., 44 , 23412348.

    • Search Google Scholar
    • Export Citation
  • Nordeng, T. E., 1994: Extended versions of the convective parameterization scheme at ECMWF and their impact on the mean and transient activity of the model in the tropics. ECMWF Research Department Tech. Memo. 206, European Centre for Medium-Range Weather Forecasts, 41 pp. [Available from ECMWF, Shinfield Park, Reading, Berkshire RG2 9AX, United Kingdom.].

  • Salby, M. L., and H. H. Hendon, 1994: Intraseasonal behavior of clouds, temperature, and motion in the tropics. J. Atmos. Sci., 51 , 22072224.

    • Search Google Scholar
    • Export Citation
  • Schubert, S. D., and M. L. Wu, 2001: Predictability of the 1997 and 1998 South Asian summer monsoon low-level winds. J. Climate, 14 , 31733191.

    • Search Google Scholar
    • Export Citation
  • Shinoda, T., and H. H. Hendon, 1998: Mixed layer modeling of intraseasonal variability in the tropical western Pacific and Indian Oceans. J. Climate, 11 , 26682685.

    • Search Google Scholar
    • Export Citation
  • Shinoda, T., H. H. Hendon, and J. D. Glick, 1998: Intraseasonal variability of surface fluxes and sea surface temperature in the tropical western Pacific and Indian Oceans. J. Climate, 11 , 16851702.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., and Coauthors, 1996: Intraseasonal oscillations in 15 atmospheric general circulation models: Results from an AMIP diagnostic subproject. Climate Dyn., 12 , 325357.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., J. M. Slingo, P. M. Inness, and W. K-M. Lau, 1997: On the maintenance and initiation of the intraseasonal oscillation in the NCEP/NCAR Reanalysis and the GLA and UKMO AMIP simulations. Climate Dyn., 13 , 769795.

    • Search Google Scholar
    • Export Citation
  • Spiegel, M. R., 1988: Schaum’s Outline of. Theory and Problems of Statistics. 2nd ed. McGraw-Hill, 504 pp.

  • Tiedtke, M., 1989: A comprehensive mass flux scheme for cumulus parameterization in large-scale models. Mon. Wea. Rev., 117 , 17791800.

    • Search Google Scholar
    • Export Citation
  • Waliser, D. E., K. M. Lau, and J-H. Kim, 1999: The influence of coupled sea surface temperatures on the Madden–Julian oscillation: A model perturbation experiment. J. Atmos. Sci., 56 , 333358.

    • Search Google Scholar
    • Export Citation
  • Wang, B., and X. Xie, 1998: Coupled modes of the warm pool climate system. Part I: The role of air–sea interaction in maintaining Madden–Julian oscillation. J. Climate, 11 , 21162135.

    • Search Google Scholar
    • Export Citation
  • Wang, W., and M. E. Schlesinger, 1999: The dependence on convection parameterization of the tropical intraseasonal oscillation simulated by the UIUC 11-layer atmospheric GCM. J. Climate, 12 , 14231457.

    • Search Google Scholar
    • Export Citation
  • Watterson, I. G., 2002: The sensitivity of subannual and intraseasonal tropical variability to model ocean mixed layer depth. J. Geophys. Res., 107 .4020, doi:10.1029/2001JD000671.

    • Search Google Scholar
    • Export Citation
  • Weller, R. A., and S. P. Anderson, 1996: Surface meteorology and air–sea fluxes in the western equatorial Pacific warm pool during the TOGA Coupled Ocean–Atmosphere Response Experiment. J. Climate, 9 , 19591990.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., and G. N. Kiladis, 1999: Convectively coupled equatorial waves: Analysis of clouds and temperature in the wavenumber–frequency domain. J. Atmos. Sci., 56 , 374399.

    • Search Google Scholar
    • Export Citation
  • Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2000: The relationship between convection and sea surface temperature on intraseasonal timescales. J. Climate, 13 , 20862104.

    • Search Google Scholar
    • Export Citation
  • Woolnough, S. J., J. M. Slingo, and B. J. Hoskins, 2001: The organization of tropical convection by intraseasonal sea surface temperature anomalies. Quart. J. Roy. Meteor. Soc., 127 , 887907.

    • Search Google Scholar
    • Export Citation
  • Wu, M. L., S. Schubert, I-S. Kang, and D. Waliser, 2002: Forced and free intraseasonal variability over the South Asian monsoon region simulated by 10 AGCMs. J. Climate, 15 , 28622880.

    • Search Google Scholar
    • Export Citation
  • Zhang, C. D., M. Dong, S. Gualdi, H. H. Hendon, E. D. Maloney, A. Marshall, K. R. Sperber, and W. Q. Wang, 2006: Simulations of the Madden-Julian oscillation in four pairs of coupled and uncoupled global models. Climate Dyn., 27 , 573592.

    • Search Google Scholar
    • Export Citation
  • Zheng, Y., D. E. Waliser, W. F. Stern, and C. Jones, 2004: The role of coupled sea surface temperatures in the simulation of the tropical intraseasonal oscillation. J. Climate, 17 , 41094134.

    • Search Google Scholar
    • Export Citation
  • Zhong, A., O. Alves, A. Schiller, F. Tseitkin, and N. Smith, 2004: Results from a preliminary version of the ACOM2/BAM coupled seasonal forecast model. BMRC Research Rep. 95, Bureau of Meteorology, Australia, 45 pp.

  • Fig. 1.

    Wavenumber–frequency spectra of surface zonal wind anomalies in (a) SHS for NNR1, (b) NHS for NNR1, (c) SHS for CTL, (d) NHS for CTL, (e) SHS for SLAB, and (f) NHS for SLAB. Contour interval is 0.001 m2 day s−2, with gray areas indicating spectral density values greater than 0.01 m2 day s−2.

  • Fig. 2.

    Geographical distributions of variance at MJO scales (eastward wavenumbers 1–4 for periods of 20–120 days) for surface zonal wind anomalies in (a) SHS for NNR1, (b) NHS for NNR1, (c) SHS for CTL, and (d) NHS for CTL. Contour interval (CI) is 0.3 m2 s−2, with gray areas indicating variances greater than 0.9 m2 s−2.

  • Fig. 3.

    Geographical distributions of variance at MJO scales for surface zonal wind anomalies in CTL for SHS. Spectral filtering is for eastward wavenumbers 1–4 for periods (a) 45–120 days and (b) 20–45 days. The CI is 0.3 m2 s−2, with gray areas indicating variances greater than 0.9 m2 s−2.

  • Fig. 4.

    Time–lon section of 5-day mean (a) surface zonal wind and (b) OLR anomalies averaged 20°S–0° for an arbitrary 8-month period for the CTL in SHS. The thin shaded contours show raw zonal wind anomalies, with a CI of (a) 2 m s−1 and (b) 20 W m−2, and gray areas indicate negative anomalies. Solid thick contours indicate slow eastward-propagating anomalies (spectrally filtered for positive wavenumbers 1–4 and periods of 45–120 days), while dashed bold contours indicate fast eastward-propagating anomalies (spectrally filtered for positive wavenumbers 1–4 and periods of 20–45 days).

  • Fig. 5.

    Geographical locations of anomalous OLR for the SLAB (solid contours) and the CTL (dashed contours) simulations in southern summer, superimposed for various lag composites. Black contours represent −22 W m−2 OLR anomalies for day 0 composites based at 100°E, while gray contours represent −12 W m−2 OLR anomalies for a lag of −10 days.

  • Fig. 6.

    Composite SST anomalies for the SLAB experiment at lags of −10, 0, and 10 days. Base longitude for the composites is 100°E. Contour interval is 0.05 K, with gray areas indicating negative values. A land–sea mask is applied such that skin temperatures over land are omitted. Bold contours are negative OLR anomalies, with a contour interval of 5 W m−2 and a minimum value (lowest bold contour) of −22 W m−2. Positive OLR anomalies are not shown.

  • Fig. 7.

    Composite anomalies of latent heat and shortwave fluxes, surface zonal wind speed, and SST for (a) CTL, (c) SLAB, and (e) SLAB–CTL. Composite anomalies of total heat flux, OLR, surface specific humidity, and 1000-mb geopotential height for (b) CTL, (d) SLAB, and (f) SLAB–CTL. Data are shown as a function of longitude at day 0, averaged 5°–10°S. Base longitude for the composites is 100°E. Latent heat flux anomalies are positive into the atmosphere, while shortwave and total heat flux anomalies are positive into the ocean.

  • Fig. 8.

    Composite absolute zonal wind speeds at day 0 for the CTL run. Base lon for the composite is 100°E. The CI is 1 m s−1, with gray (white) areas indicating mean westerlies (easterlies).

  • Fig. 9.

    Difference between the SLAB and the CTL composites at lag 0 of anomalous (a) surface latent heat flux, (b) surface wind speed, (c) zonal wind speed, (d) specific humidity, (e) geopotential height, and (f) total heat flux. Base lon for the composites is 100°E. CIs are (a) 5 W m−2, (b) 0.2 m s−1, (c) 0.2 m s−1, (d) 0.1 g kg−1, (e) 1 m, and (f) 5 W m−2. Gray areas indicate negative values. Bold contours incorporate regions of statistical significance at the 95% confidence level using a two-tailed significance test for the difference between two means (Spiegel 1988).

  • Fig. 10.

    SLAB–CTL difference composites (vertical sections) of anomalous (a) moisture convergence and (b) specific humidity as a function of lag time. Base longitude for the composites is 100°E. The CI is 2.5 × 10−9 s−1 and 0.1 g kg−1, respectively, with gray shaded areas indicating negative values.

  • Fig. 11.

    Schematic illustrating the impact of air–sea coupling on the dynamics of the MJO in BAM.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 335 112 8
PDF Downloads 223 67 0