• Adler, R. F., , G. J. Huffman, , D. T. Bolvin, , S. Curtis, , and E. J. Nelkin, 2000: Tropical rainfall distributions determined using TRMM combined with other satellite and rain gauge information. J. Appl. Meteor., 39 , 20072023.

    • Search Google Scholar
    • Export Citation
  • Annamalai, H., , and J. M. Slingo, 1999: The mean evolution and variability of the Asian summer monsoon: Comparison of ECMWF and NCEP–NCAR reanalyses. Mon. Wea. Rev., 127 , 11571186.

    • Search Google Scholar
    • Export Citation
  • Annamalai, H., , and K. R. Sperber, 2005: Regional heat sources and the active and break phases of boreal summer intaseasonal (30–50 day) variability. J. Atmos. Sci., 62 , 27262748.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., , and M. J. Miller, 1986: A new convective adjustment scheme. Part II: Single-column testing using GATE wave, BOMEX, ATEX, and arctic air-mass datasets. Quart. J. Roy. Meteor. Soc., 112 , 693709.

    • Search Google Scholar
    • Export Citation
  • Bladé, I., , and D. L. Hartmann, 1993: Tropical intraseasonal oscillation in a simple nonlinear model. J. Atmos. Sci., 50 , 29222939.

  • Chen, S. S., , B. E. Mapes, , and R. A. Houze Jr., 1996: Multiscale variability of deep convection in relation to large-scale circulation in TOGA COARE. J. Atmos. Sci., 53 , 13801409.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46 , 30773107.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1993: A nonhydrostatic version of the Penn State–NCAR Mesoscale Model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev., 121 , 14931513.

    • Search Google Scholar
    • Export Citation
  • Fasullo, J., , and P. J. Webster, 2000: Atmospheric and surface variations during westerly wind bursts in the tropical western Pacific. Quart. J. Roy. Meteor. Soc., 128 , 123.

    • Search Google Scholar
    • Export Citation
  • Frederiksen, J. S., , and C. S. Frederiksen, 1997: Mechanisms of the formation of intraseasonal oscillations and Australian monsoon disturbances: The roles of latent heat, barotropic and baroclinic instability. Contrib. Atmos. Phys., 70 , 3956.

    • Search Google Scholar
    • Export Citation
  • Ghil, M., , and K. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part I: Northern hemisphere and tropics. J. Atmos. Sci., 48 , 752779.

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., , J. Dudhia, , and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 117 pp.

    • Search Google Scholar
    • Export Citation
  • Gustafson, W. I., , and B. C. Weare, 2004a: MM5 modeling of the Madden–Julian oscillation in the Indian and west Pacific Oceans: Model description and control run results. J. Climate, 17 , 13201337.

    • Search Google Scholar
    • Export Citation
  • Gustafson, W. I., , and B. C. Weare, 2004b: MM5 modeling of the Madden–Julian oscillation in the Indian and west Pacific Oceans: Implications of 30–70-day boundary effects on MJO development. J. Climate, 17 , 13381351.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., 1988: A simple model of the 40–50-day oscillation. J. Atmos. Sci., 45 , 569584.

  • 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
  • Hoskins, B. J., , and G-Y. Yang, 2000: The equatorial response to higher-latitude forcing. J. Atmos. Sci., 57 , 11971213.

  • Hsu, H. H., , B. J. Hoskins, , and F-F. Jin, 1990: The 1985/86 intraseasonal oscillation and the role of the extratropics. J. Atmos. Sci., 47 , 823839.

    • Search Google Scholar
    • Export Citation
  • Hu, Q., , and D. A. Randall, 1994: Low-frequency oscillations in radiative-convective systems. J. Atmos. Sci., 51 , 10891099.

  • Hu, Q., , and D. A. Randall, 1995: Low-frequency oscillations in radiative-convective systems. Part II: An idealized model. J. Atmos. Sci., 52 , 478490.

    • Search Google Scholar
    • Export Citation
  • Inness, P. M., , J. M. Slingo, , E. Guilyardi, , and J. Cole, 2003: Simulation of the Madden–Julian oscillation in a coupled general circulation model. Part II: The role of the basic state. J. Climate, 16 , 365382.

    • Search Google Scholar
    • Export Citation
  • Janjic, Z. I., 1994: The step-mountain coordinate model: Further development of the convection, viscous sublayer, and turbulence closure schemes. Mon. Wea. Rev., 122 , 927945.

    • Search Google Scholar
    • Export Citation
  • Jones, C., , and J-K. Schemn, 2000: The influence of intraseasonal variations on medium-range weather forecasts over South America. Mon. Wea. Rev., 128 , 486494.

    • Search Google Scholar
    • Export Citation
  • Jones, C., , D. E. Waliser, , J-K. Schemn, , and W. K. M. Lau, 2000: Prediction skill of the Madden–Julian Oscillation in dynamical extended range forecasts. Climate Dyn., 16 , 273289.

    • 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. R., , and B. C. Weare, 2001: The onset of convection in Madden–Julian oscillation. J. Climate, 14 , 780793.

  • Kiladis, G. N., , K. H. Straub, , and P. T. Haertel, 2005: Zonal and vertical structure of the Madden–Julian oscillation. J. Atmos. Sci., 62 , 27902809.

    • Search Google Scholar
    • Export Citation
  • Knutson, R. R., , and K. M. Weickmann, 1987: 30–60 day atmospheric oscillations: Composite life cycles of convection and circulation anomalies. Mon. Wea. Rev., 115 , 14071436.

    • Search Google Scholar
    • Export Citation
  • Knutson, R. R., , K. M. Weickmann, , and J. E. Kutzbach, 1986: Global-scale intraseasonal oscillations of outgoing longwave radiation and 250-mb zonal wind during Northern Hemisphere summer. Mon. Wea. Rev., 114 , 605623.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., , and L. Peng, 1987: Origin of low-frequency (intraseasonal) oscillations in the tropical atmosphere. Part I: Basic theory. J. Atmos. Sci., 44 , 950972.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., , and D. E. Waliser, 2005: Intraseasonal Variability in the Atmosphere–Ocean Climate System. Praxis, 436 pp.

  • Lin, H., , G. Brunet, , and J. Derome, 2007: Intraseasonal variability in a dry atmospheric model. J. Atmos. Sci., 64 , 24222441.

  • 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
  • Madden, R. A., , and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29 , 11091123.

    • Search Google Scholar
    • Export Citation
  • Matthews, A., 2008: Primary and successive events in the Madden–Julian oscillation. Quart. J. Roy. Meteor. Soc., 134 , 439453.

  • Mechem, D. B., , S. S. Chen, , and R. A. Houze Jr., 2006: Momentum transport processes in the stratiform regions of mesoscale convective systems over the western Pacific warm pool. Quart. J. Roy. Meteor. Soc., 132A , 709736.

    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., , S. J. Taubman, , P. D. Brown, , M. J. Jacono, , and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102 , 1666316682.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , and J-Y. Yu, 1994: Modes of tropical variability under convective adjustment and the Madden–Julian oscillation. Part I. Analytical theory. J. Atmos. Sci., 51 , 18761894.

    • Search Google Scholar
    • Export Citation
  • Newman, M., , P. D. Sardeshmukh, , and J. W. Bergman, 2000: An assessment of the NCEP, NASA, and ECMWF reanalyses over the tropical west Pacific warm pool. Bull. Amer. Meteor. Soc., 81 , 4148.

    • Search Google Scholar
    • Export Citation
  • Roundy, P. E., 2008: Analysis of convectively coupled Kelvin waves in the Indian Ocean MJO. J. Atmos. Sci., 65 , 13421359.

  • Rui, H., , and B. Wang, 1990: Development characteristics and dynamic structure of tropical intraseasonal convection anomalies. J. Atmos. Sci., 47 , 357379.

    • Search Google Scholar
    • Export Citation
  • Salby, M. L., , and H. H. Hendon, 1994: Planetary-scale circulations in the presence of climatological and wave-induced heating. J. Atmos. Sci., 51 , 23442367.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., , and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45 , 12281251.

    • 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
  • Straub, K. H., , and G. N. Kiladis, 2002: Observations of a convectively coupled Kelvin wave in the eastern Pacific ITCZ. J. Atmos. Sci., 59 , 3053.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 reanalysis. Quart. J. Roy. Meteor. Soc., 131 , 29613012.

  • Vukicevic, T., , and R. M. Errico, 1990: The influence of artificial and physical factors upon predictability estimates using a complex limited-area model. Mon. Wea. Rev., 118 , 14601482.

    • 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
  • Waliser, D. E., , K. M. Lau, , W. Stern, , and C. Jones, 2003: Potential predictability of the Madden–Julian oscillation. Bull. Amer. Meteor. Soc., 84 , 3350.

    • Search Google Scholar
    • Export Citation
  • Wang, B., , and T. Li, 1994: Convective interaction with boundary layer dynamics in the development of the tropical intraseasonal signal. J. Atmos. Sci., 51 , 13861400.

    • Search Google Scholar
    • Export Citation
  • Warner, T. T., , L. E. Key, , and A. M. Lario, 1989: Sensitivity of mesoscale-model forecast skill to some initial data characteristics, data density, data position, analysis procedure, and measurement error. Mon. Wea. Rev., 117 , 12811310.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., , and H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132 , 19171932.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., , G. N. Kiladis, , and P. J. Webster, 2000: Large-scale dynamical fields associated with convectively coupled equatorial waves. J. Atmos. Sci., 57 , 613640.

    • Search Google Scholar
    • Export Citation
  • Wilson, J. D., , and M. Mak, 1984: Tropical response to lateral forcing with a latitudinally and zonally nonuniform basic state. J. Atmos. Sci., 41 , 11871201.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , and M-M. Lu, 1983: Equatorially trapped waves at 200 mb and their association with meridional convergence of wave energy flux. J. Atmos. Sci., 40 , 27852803.

    • Search Google Scholar
    • Export Citation
  • Yu, J-Y., , and J. D. Neelin, 1994: Modes of tropical variability under convective adjustment and the Madden–Julian oscillation. Part II: Numerical results. J. Atmos. Sci., 51 , 18951914.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., 1996: Atmospheric intraseasonal variability at the surface in the western Pacific Ocean. J. Atmos. Sci., 53 , 739758.

  • Zhang, C., 2005: Madden-Julian Oscillation. Rev. Geophys., 43 , RG2003. doi:10.1029/2004RG000158.

  • Zhang, C., , and M. J. McPhaden, 2000: Intraseasonal surface cooling in the equatorial western Pacific. J. Climate, 13 , 22612276.

  • Zhang, C., , and M. Dong, 2004: Seasonality of the Madden–Julian oscillation. J. Climate, 17 , 31693180.

  • Zhang, C., , M. Dong, , S. Gualdi, , H. H. Hendon, , E. D. Maloney, , A. Marshall, , K. R. Sperber, , and W. Wang, 2006: Simulations of the Madden-Julian oscillation by four pairs of coupled and uncoupled global models. Climate Dyn., 27 , 573592. doi:10.1007/s00382-006-0148-2.

    • 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
  • View in gallery

    Primary model domain for the TMM5 (D1, 21°S–21°N, 0°–360°) and the nested domain (D2, 11°S–11°N, 37°–183°E). Domains D1 and D2 have resolutions of 111 and 37 km, respectively. Control simulations include only D1.

  • View in gallery

    Time-longitude diagrams of daily (a) 850-hPa zonal wind (m s−1) from the NCEP–NCAR reanalysis and (b) precipitation (mm day−1) from the merged TRMM datasets for Case 1 (averaged over 10°S–10°N). (c), (d) Case 2. The two cases selected in this study are marked by the straight lines whose slope corresponds to an eastward-propagation speed of 5 m s−1.

  • View in gallery

    200-hPa wind anomalies from the NCEP–NCAR reanalysis during (a) 10–15 May 2002 for Case 1 and (b) 23–27 Nov 2000 for Case 2.

  • View in gallery

    Time-longitude diagrams of daily U850 anomalies (m s−1) averaged over 10°S–10°N for Case 1 from (a) NNR, (b) TMM5 single-domain simulation, and (c) TMM5 nested-domain simulation. A 3-day running mean is applied.

  • View in gallery

    (left) Mean U850 for Case 1 (10 Apr–10 Jun 2002) and (right) Case 2 (10 Nov–31 Dec 2000) from the (top) NNR, (middle) ERA-40, and (bottom) control simulations (CS1 and CS2). Contour interval is 2 m s−1.

  • View in gallery

    Vertical structure of the zonal wind anomalies (m s−1) in a 5° × 5° box centered at 0°, 90°E for Case 1 from (a) NNR and (b) control simulation (CS1). The solid (dashed) contours represent positive (negative) anomalies. A 7-day running mean is applied.

  • View in gallery

    U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1 (with constant SST), and (c) test run SST1A (with varying SST). The black lines indicate a zonal propagation speed of 5 m s−1.

  • View in gallery

    Averaged SST for (a) Case 1 (10 Apr–10 Jun 2002), (b) Case 2 (10 Nov–31 Dec 2000), and (c) their differences. Contour interval is 0.5°C.

  • View in gallery

    U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run IN_FSST1_1 starting 5 days before the control simulation, and (d) test run IN_FSST1_5 starting 5 days after the control simulation.

  • View in gallery

    U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run NO_LH1 with no latent heating, and (d) test run NO_MOIST1 with no moisture. All are averaged over 10°S to 10°N. The black lines mark the zonal propagation speed of 5 m s−1.

  • View in gallery

    Time-height diagrams of zonal wind anomalies (m s−1) in a 5° × 5° box centered at 0°, 90°E for Case 1 from (a) control simulation CS1 and (b) test run NO_LH1 with no latent heating, and (c) test run NO_MOIST1 with no moisture. The solid (dashed) contours represent positive (negative) anomalies. A 7-day running mean is applied.

  • View in gallery

    U850 anomalies (m s−1; 3-day running mean) for (top) Case 1 from (a) NNR, (b) control simulations CS1, and (c) test run FBC1A with constant lateral boundary conditions. (bottom) Case 2 from (d) NNR, (e) test run 15Oct starting from 15 Oct, and (f) test run 15Oct_fixed with constant boundary conditions. All averaged over 10°S–10°N.

  • View in gallery

    U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run MBV1A with lateral boundaries at 28°S and N, and (d) test run MBV1B with lateral boundaries at 38°S and N.

  • View in gallery

    Winds (10–15 May 2002; m s−1) at 200 hPa for Case 1 from (a) NNR, and (b) test run MBV1B with lateral boundaries at 38°S and N.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 53 53 14
PDF Downloads 16 16 6

A Numerical Case Study on the Initiation of the Madden–Julian Oscillation

View More View Less
  • 1 RSMAS, University of Miami, Miami, Florida
  • | 2 National Center for Atmospheric Research, * Boulder, Colorado
  • | 3 RSMAS, University of Miami, Miami, Florida
© Get Permissions
Full access

Abstract

A mesoscale tropical channel model is used to study the long-standing problem of the initiation of the Madden–Julian oscillation (MJO). With initial and lateral boundary conditions provided by a global reanalysis, this model is able to reproduce the initiation and gross features of two observed MJO events up to 2 months after the start of simulations. This leads to a conjecture that these two MJO events are generated by the influences from the lateral boundaries. This conjecture is supported by a series of sensitivity tests. These sensitivity tests demonstrate that the simulated MJO initiation does not critically depend on detailed characteristics of sea surface temperature (varying versus constant in time, mean distribution from boreal spring versus winter), initial conditions (within a 10-day period), the latitudinal location of the lateral boundaries (21°–38°N and S), or even latent heating and moist processes. The only factor found critical to the reproduction of the MJO initiation is time-varying lateral boundary conditions from the reanalysis. When such lateral boundary conditions are replaced by time-independent conditions, the model fails to reproduce the MJO initiation. These results support the idea that extratropical influences can be an efficient mechanism for MJO initiation. Implications of these results are discussed.

Corresponding author address: Pallav Ray, RSMAS/MPO, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149. Email: pray@rsmas.miami.edu

Abstract

A mesoscale tropical channel model is used to study the long-standing problem of the initiation of the Madden–Julian oscillation (MJO). With initial and lateral boundary conditions provided by a global reanalysis, this model is able to reproduce the initiation and gross features of two observed MJO events up to 2 months after the start of simulations. This leads to a conjecture that these two MJO events are generated by the influences from the lateral boundaries. This conjecture is supported by a series of sensitivity tests. These sensitivity tests demonstrate that the simulated MJO initiation does not critically depend on detailed characteristics of sea surface temperature (varying versus constant in time, mean distribution from boreal spring versus winter), initial conditions (within a 10-day period), the latitudinal location of the lateral boundaries (21°–38°N and S), or even latent heating and moist processes. The only factor found critical to the reproduction of the MJO initiation is time-varying lateral boundary conditions from the reanalysis. When such lateral boundary conditions are replaced by time-independent conditions, the model fails to reproduce the MJO initiation. These results support the idea that extratropical influences can be an efficient mechanism for MJO initiation. Implications of these results are discussed.

Corresponding author address: Pallav Ray, RSMAS/MPO, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149. Email: pray@rsmas.miami.edu

1. Introduction

The Madden–Julian oscillation (MJO; Madden and Julian 1971, 1972) is a dominant component of subseasonal variability in the tropical atmosphere. The importance of the MJO, as a subject challenging our intellectual understanding as well as practical prediction, has been summarized recently by Lau and Waliser (2005) and Zhang (2005). Among many unsolved problems related to the MJO, the mechanism for its initiation is one of the most puzzling and least understood. Several hypotheses have been proposed:

  • (i) Local discharge–recharge processes. Local convective instability serves as an energy source for the MJO. After being reduced during a convectively active phase of the MJO, it needs a certain time to be restored to the level to support another episode of active convection. The restoring time, as shown observationally, numerically, and theoretically, is roughly the local period of the MJO (e.g., Hendon 1988; Bladé and Hartmann 1993; Hu and Randall 1994, 1995; Kemball-Cook and Weare 2001). This mechanism attributes the most essential process of the MJO initiation to the tropical atmosphere over its “genesis region,” without a need for extra influences from the extratropics or upstream. Using a linear baroclinic model, Annamalai and Sperber (2005) found that the dry convective anomalies over the equatorial eastern Indian Ocean could force low-level circulation anomalies that help convergence over the western Indian Ocean, favoring the initiation of the next active phase of the MJO.
  • (ii) Extratropical influences. Extratropical intraseasonal perturbations have been proposed as a possible trigger for the MJO (e.g., Lau and Peng 1987). Eastward-moving extratropical disturbances can excite a variety of equatorial waves (e.g., Yanai and Lu 1983). Sardeshmukh and Hoskins (1988) found that the advection of vorticity associated with the divergent flow is important for understanding tropics–extratropics interactions. Hoskins and Yang (2000) challenged the requirement of background westerlies for Rossby wave propagation by demonstrating that the extratropics can influence the tropics directly also in the presence of easterlies. In a case study, Hsu et al. (1990) found observational evidence suggesting that Rossby wave trains propagating into the tropics from the midlatitudes may play a role in organizing deep convection associated with the MJO over the Indian Ocean. On the other hand, Kemball-Cook and Weare (2001), based on composite studies, found no evidence of systematic initiation of deep convection by the incursion of extratropical waves into the Indian Ocean. Whether the relationships between the extratropical and tropical intraseasonal signal are statistically significant is controversial (e.g., Ghil and Mo 1991). Frederiksen and Frederiksen (1997) proposed that the MJO is a tropical–extratropical coupled mode caused by moist baroclinic–barotropic instability. A recent numerical study by Lin et al. (2007) showed that tropical intraseasonal variability in wind resembling the MJO can be produced in a dry global model due solely to lateral forcing from the extratropics.
  • (iii) Upstream effects of circumnavigating waves. Circulation anomalies associated with the previous cycle of the MJO may reemerge from the west into eastern Africa and the Indian Ocean (e.g., Knutson et al. 1986; Knutson and Weickmann 1987; Lau and Peng 1987). There, these anomalies may help the initiation of the next MJO event. The time needed for the upper-level circulation anomalies to travel around the tropics sets the period of the MJO. This mechanism is evident in some modeling studies (e.g., Wang and Li 1994). It is not clear how upper-level circulation anomalies would trigger convection associated with the MJO (e.g., Bladé and Hartmann 1993), although a pressure change near the surface might be induced through mass transport by the circulation.
  • (iv) Stochastic forcing. High-frequency variability in tropical convective processes and extratropical perturbations can produce substantial variance at the planetary scales for the equatorial wave and the MJO (Wilson and Mak 1984; Neelin and Yu 1994; Yu and Neelin 1994). Such stochastic forcing is expected to have a specific spatial structure to reach optimal effect. For the MJO, such structure has yet to be observed.

The objective of this study is to further investigate the initiation mechanisms for the MJO through numerical case studies. The MJO initiation is commonly described in terms of both wind and convection. In this study, the MJO initiation is defined in terms of the appearance of MJO-associated wind only for a reason to be specified in section 2b. Also, an MJO appearing over the western Indian Ocean and then propagating eastward is considered as the MJO initiation over the Indian Ocean, irrespective of whether it is preceded by another event. Two randomly chosen MJO events (see section 2b) for this study were initiated over the western Indian Ocean.

A mesoscale model built to cover the entire tropics (i.e., a tropical channel model) is used to simulate the initiation of two observed MJO events over the Indian Ocean. The regular regional version of this model has been used by Gustafson and Weare (2004a,b) to simulate the MJO over a 2-yr period, which yielded reasonable statistics of the MJO. The tropical channel model, without the usual zonal boundaries, is a useful tool for the purpose of this study. In a regional model, any feedbacks with the rest of the globe are controlled through the boundary conditions, which allow their influences on the MJO initiation to be tested. For example, any signal related to prior MJOs can be filtered from the boundary forcing to see how external influences affect the MJO (Gustafson and Weare 2004a,b). But the removal of the MJO frequency alone cannot entirely eliminate the influence of the MJO from the boundary conditions, because the MJO may influence smaller-scale features that are coupled to convection. Another advantage of using this model is its capability of two-way nested domains with high resolution and sophisticated physics. To help identify the MJO initiation mechanisms, our approach is to simulate individual MJO events rather than MJO statistics. Our general strategy is to reproduce observed MJO initiation events and then test the sensitivity of the simulated MJO initiation to various factors, such as sea surface temperature (SST), initial and lateral boundary conditions, and convectively related processes.

Section 2 provides brief descriptions of the model, the setup of the numerical simulations, and the data used to constrain and validate the simulations. Section 3 presents diagnoses of the model simulations with an emphasis on the mean state, propagation, and vertical structure of the MJO. Section 4 describes a series of sensitivity tests, in which roles of SST, initial and lateral boundary conditions, latent heating, and moist processes are assessed. Summary and conclusions, along with the implications of this study, are given in section 5.

In short, this study has obtained supporting evidence of the lateral forcing mechanism for the MJO initiation and calls for further research attention to this mechanism that has been somewhat neglected by mainstream MJO research.

2. Model and data

a. Model

We have developed a tropical channel model based on the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (NCAR) Mesoscale Model (MM5; Dudhia 1993; Grell et al. 1995). This channel model employs a Mercator projection centered at the equator with open boundaries in the north–south direction. The model domain covers the entire tropics, with its usual east–west boundaries overlapped. Tests are conducted to ensure that simulated perturbations propagate zonally through the overlapping grids without any distortion. Especially, two test runs are made in which the overlapping zone is located over the western Pacific and Atlantic Oceans, respectively. Results from these two simulations are the same over these regions with or without the overlapping zone. The north–south boundaries can be set at selected latitudes. Hereafter, we refer to this tropical channel model as tropical MM5 (TMM5).

The dynamics and physics packages of TMM5 are the same as those of the regular regional MM5, with equations for a fully compressible, nonhydrostatic atmosphere. It retains the two-way nesting capabilities. The same options within each category of physical parameterizations (cumulus convection, moisture, radiation, planetary boundary layer, etc.) are available. The spatial differencing is centered and of second order. There are 28 unevenly spaced full-sigma levels, with the maximum resolution in the boundary layer and the model top at 50 hPa. All nested domains are activated at the initial time of the simulation. The output is taken every 3 h.

A regular regional model is subject to both boundary effects from the zonal (due to the circumnavigating waves) and the meridional (from the extratropics) directions. In comparison, the main advantage of TMM5 is that it eliminates the east–west boundaries and thereby isolates the boundary effects that arrive solely from the extratropics. For this reason, TMM5 is an ideal tool to investigate possible influences from the extratropics. A tropical channel model developed at NCAR based on the Weather Research and Forecasting (WRF) model, known as the Nested Regional Climate Model (NRCM), is conceptually the same as this TMM5. These two models form a unique pair for comparisons and cross validations.

Several simulations are performed spanning between one to several weeks. They are used to evaluate the skill of MJO simulations by TMM5 against observations and reanalyses, to assess the effects of increasing horizontal resolutions, and to select the suite of physical parameterizations that yields the optimal simulation of the MJO, but without trying to determine the accuracy and deficiency in individual schemes. Based on these tests and the work of Gustafson and Weare (2004a,b), the selected parameterizations for this study are (i) the Betts–Miller convective scheme (Betts and Miller 1986); (ii) the explicit moisture calculations using a simple ice scheme (Dudhia 1989); (iii) the planetary boundary layer (PBL) scheme of the NCEP Eta Model (Janjic 1994), and (iv) the longwave Rapid Radiative Transfer Model (RRTM; Mlawer et al. 1997) and Dudhia (1989) shortwave radiation. The Betts–Miller scheme, designed for the coarse-resolution climate models, works well perhaps because the model is mainly in the hydrostatic regime at the resolutions employed (with grid spacing of 111 and 37 km). Over the land, a five-layer soil model available within the MM5 is used.

The above chosen parameterization package and a model domain of 21°S–21°N with a horizontal resolution of 111 km (D1 in Fig. 1) comprise control simulations in this study. In the control simulations, sea surface temperature is constant in time with the same spatial pattern in the initial conditions (see section 2c). The terms constant and varying refer to time. Thus, varying SST has intraseasonal fluctuations, whereas constant SST has no intraseasonal fluctuations. An option of a two-way nested inner domain of 37 km over the Indian and western Pacific Oceans (D2 in Fig. 1) is included to assess the effect of increasing horizontal resolutions.

b. Simulation design

The general strategy of this study is to simulate the initiation of two observed MJO events, augmented with sensitivity tests to various factors that might be important to the MJO initiation. All simulations for these two events are listed, respectively, in Tables 1, 2. The first case is an MJO event in April–May 2002 (Fig. 2). This case is special in several ways. It occurs in a season in which the MJO is closest to the equator but on average weaker than in other seasons (Zhang and Dong 2004). This particular MJO event underwent a distinct intraseasonal transition between a period of convectively quiescence and low-level (e.g., 850 hPa) easterly wind dominance to one of convectively active and low-level westerly dominance. After initiation, however, it propagated eastward at a speed slightly faster than the average phase speed of the MJO (5 m s−1, marked by a straight line in Fig. 2), especially with respect to precipitation over the western Pacific (Fig. 2b). This event, therefore, is a mixture of the MJO in its initiation and early stage over the Indian Ocean and perhaps a convectively coupled Kelvin wave (Straub and Kiladis 2002) over the western Pacific. The emphasis of this study is the initiation of the MJO. Therefore, we treat this case as an MJO event initiated over the western Indian Ocean.

To assess the model’s ability to reproduce this event with different resolutions, two simulations are made from 1 March to 30 June 2002; one using a single domain (1DOM in Table 1), the other using nested domains (2DOM in Table 1). The simulation with the nested domains is better than that of the single domain, but not appreciably. In these two runs, varying SST including intraseasonal fluctuations is used. The roles of the SST variability in the simulations are later shown to be nonessential (section 4a). The initial and lateral boundary conditions are provided by a global reanalysis (see section 2c).

The starting time (1 March) of these two runs bears special implications. It is about 2 months before the initiation of the MJO phase with active deep convection and low-level westerlies in May over the Indian Ocean. The choice of such a starting time is to assess the model’s capability of reproducing the initiation of the MJO event, namely, the intraseasonal transition from low-level easterlies to westerlies (or from convectively inactive to active periods). In dynamical models, the forecast of future development of the MJO tends to have greater predictability when it is already present, compared to when it is absent at the initial time (Jones et al. 2000); even so, the model-suggested predictability limit is about 10–15 days for rainfall and about 25–30 days for upper-level winds (Waliser et al. 2003). Detailed discussions of these two runs are given in section 3a.

For computational efficiency, most sensitivity tests and their reference control simulations (CS1) start from 10 April 2002. In the control simulation, constant SST, with values at the initial time, is used. The lateral boundary condition is time varying for all the simulations, except the test run FBC1A in Table 1, where it is constant in time with spatial distribution fixed at the initial time.

The second case is an MJO event during November–December 2000 (Case 2 in Fig. 2). This is in the prime season for the MJO (Salby and Hendon 1994; Zhang and Dong 2004). This event starts around 15 November, and anomalies in precipitation and low-level zonal wind both propagate from the western Indian Ocean to the western Pacific at a speed of about 5 m s−1. Thus, this is a classic MJO event in terms of geographical location and propagation speed. The presence of La Niña conditions during November–December 2000 was responsible for below-normal SST in the west Pacific. The two MJO events are chosen from two seasons with different background SST (see section 4a). The control simulation for case 2 (CS2) starts from 10 November, when the amplitude of the MJO based on the Wheeler and Hendon (2004) index is larger than 1. Thus, one might expect the model to capture the MJO initiation efficiently. The choice of the initial date of simulation was primarily due to computational efficiency. To make sure that the model is able to capture the MJO initiation at least 3 weeks in advance, we have made another simulation that starts on 15 October (15Oct in Table 2; the corresponding constant boundary simulation is referred as 15Oct_fixed in Table 2).

An essential feature of the MJO is a pair of anomalous anticyclones (“Rossby gyres”) straddling the equator at upper levels to the west of the convection center (Rui and Wang 1990). Such Rossby gyres are evident for the two MJO cases in anomalous wind vectors at the 200-hPa level from a reanalysis product (see section 2c) over the Indian Ocean (Fig. 3). At the lower troposphere, cyclonic gyres are not as prominent as the anticyclonic gyres at the upper levels (not shown). Reproducing this pair of Rossby gyres by TMM5 will provide further confidence in the model’s application for the study of tropical waves and the MJO (see sections 3 and 4d).

In diagnoses of model simulations, we only use the wind field to measure the MJO and its initiation. This, compromising the totality of the common MJO definition in both wind and precipitation/cloud (e.g., Wheeler and Hendon 2004), makes consistent comparisons among different test runs, including ones without latent heating and moist processes (section 4c).

c. Data

The equatorial Indian Ocean is one of the poorly monitored oceans, with very few, if any, research-quality, long-term in situ observations. Model validation is made using the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research Reanalysis (hereafter NNR; 2.5° × 2.5°, daily; Kalnay et al. 1996) data and satellite observations. The European Centre for Medium-Range Weather Forecasts (ECMWF) 40-yr Reanalysis (ERA-40) winds (2.5° × 2.5°, daily; Uppala et al. 2005) is also used for comparisons. An assessment of several reanalyses (NNR, ERA40, and others) over the tropical eastern Indian and western Pacific Oceans shows considerable differences among them (Newman et al. 2000). A recent comparison of NCEP–NCAR, NCEP–Department of Energy (DOE), and ERA-40 reanalyses in the context of the MJO, however, shows that the three products provide comparable descriptions of the large-scale features of the MJO (http://climate.snu.ac.kr/mjo_metrics/menu.htm). The NCEP global tropospheric analyses (final or FNL data; 1° × 1°, 6 hourly; available from 1999) are used to provide initial and lateral boundary conditions for the model. The SST data are also from the reanalysis (1° × 1°, 6 hourly), which contain intraseasonal fluctuations (see section 4a). Precipitation data are from the Tropical Rainfall Measuring Mission (TRMM) mrged products (1° × 1°, daily) based on the TRMM Combined Instrument, including TRMM Microwave Imager (TMI) and precipitation radar data, and TRMM visible and infrared scanner (Adler et al. 2000).

3. Model validation

a. MJO initiation and propagation

Longitude-time diagrams of anomalies in zonal winds at 850 hPa (hereafter U850) for Case 1 from NNR and simulations 1DOM and 2DOM (Table 1) are shown in Fig. 4. In NNR (left), there is a strong MJO event during April–May 2002, with the eastward- propagating zonal wind anomalies switching from easterlies to westerlies on intraseasonal time scales. The eastward propagation is at a speed of about 5 m s−1 over the Indian Ocean but accelerates over the Pacific Ocean while the anomalies are weakened. Similar changes in the propagation speed can also be seen in precipitation (see Fig. 2). Simulated U850 in D1 (middle) exhibits the same intraseasonal switches between easterly and westerly anomalies and eastward propagation over the Indian Ocean. Over the western Pacific, however, it moves faster than in NNR. This problem appears to be partially remedied by including the higher-resolution nested domain D2 (right). The amplitudes of the anomalies are larger in both simulations than in NNR. Notice that the westward-propagating, synoptic-scale westerly anomalies embedded in the MJO envelope over the Indian Ocean are captured by the simulation with the nested domain (2DOM in Table 1). These detailed structures in simulated anomalies are perhaps due to the higher resolution of the model.

The most interesting result from this simulation is that the initiation of the MJO event over the Indian Ocean, namely, the intraseasonal transition between eastward-propagating easterly to westerly anomalies, is reproduced by the model at about the same time as shown by NNR 2 months after the starting time of integration. The MJO is thought to be unpredictable beyond 2 to 3 weeks (e.g., Waliser et al. 2003). If this is correct, then the reproduction of the U850 anomalies by TMM5 cannot be attributed to the initial conditions. This will be tested in section 4b. In the rest of this study, our emphasis will be on the mechanisms for the initiation of the MJO and little attention will be paid to the eastward propagation after its initiation.

The simulation with the nested domain (2DOM in Table 1) is better than that with the coarse-resolution single domain (1DOM in Table 1) in terms of the mean state as well as the initiation of the MJO event, but only slightly. For this reason and for computational efficiency, all the sensitivity tests (Tables 1, 2) in section 4 including the control simulations are performed using the single-domain D1 and covering shorter simulation time.

Next, the model’s realism is further evaluated in terms of the mean state and the vertical structure of the MJO event of Case 1 in the control simulation (CS1).

b. Mean state

The mean state has been considered vital to MJO simulations (e.g., Inness et al. 2003). For example, the MJO tends to be stronger in the models in which the seasonal cycle is stronger and mean precipitation is more realistically distributed with respect to SST (Slingo et al. 1996). Shown in Fig. 5 are U850 for Case 1 (left; 10 April – 10 June 2002) and Case 2 (right; 10 November – 31 December 2000). For Case 1, two different reanalysis products show similar patterns, except over the equatorial central Indian Ocean. The CS1 reproduces the general patterns but the biases are obvious. Strongest westerlies are found over the eastern Indian Ocean immediately north of the equator in the two reanalyses and also in the simulation. Over the western equatorial Indian Ocean, the simulated wind is stronger than the reanalyses winds. The model introduces an easterly bias of approximately 3 m s−1 in the western Pacific—but such bias in the western Pacific is not present in the CS2. The U200 over the Pacific are quite similar in all three, as are the wind directions at the 200-hPa level (not shown). Notice that there are differences of about 2 to 4 m s−1 between NNR and ERA-40 in this region (Annamalai and Slingo 1999). Easterly biases in U850 over the western Pacific are a common problem also in many global models (e.g., Zhang et al. 2006). The simulated vertically integrated water vapor content in the atmosphere and the precipitation are slightly lower than those from TMI (not shown).

c. Vertical structure

The zonal wind at 90°E (location of the strongest U850 anomalies) clearly shows a baroclinic structure in both NNR and CS1 (Fig. 6). This baroclinic structure in zonal wind distinguishes strong westerly wind events associated with the MJO from those not associated with the MJO (Fasullo and Webster 2000). The vertical overturning zonal circulation reaches its maximum during the second week of May and reverses its direction at the beginning of the third week of May. The simulation captures well the timing of these changes. The vertical tilt of the zonal wind structure documented in MJO composites (e.g., Kiladis et al. 2005) appears to be more obvious in the simulation than in NNR.

In summary, the simulations of the two control simulations are far from perfect. However, the reproduced MJO initiation appears not to be affected by the identified model biases and errors. We are confident that the model serves as a useful tool for learning about the mechanisms for the MJO initiation.

4. Sensitivity tests

The comparisons of the simulation with the observation and reanalysis in the previous section show that the gross structure for this particular MJO event, especially its initiation, is reproduced with roughly correct timing up to 2 months after the start of the simulations. This is very intriguing because the predictability limit of the MJO is believed to be about 15–20 days (Waliser et al. 2003). This MJO predictability limit is estimated using global GCMs with prescribed SST. In addition to observed SST, our TMM5 simulation is forced by prescribed lateral boundary conditions based on reanalysis. Is this MJO event reproduced by TMM5 mainly because of the particular prescribed SST, the particular initial condition, the model’s ability to generate the MJO with sheer luck, or the prescribed lateral boundary conditions? In this section, we present a series of sensitivity tests to investigate the possible role each of these factors might have played in the reproduction of the MJO event.

a. SST

It is known that SST feedback may enhance simulated MJO signals. However, this critically depends on the ability of a model to produce correct mean-state and basic MJO features (Waliser et al. 1999; Hendon 2000). While quantitative effects of the SST on initiation, maintenance, and propagation of the MJO are not well understood, it has been shown that an atmospheric model forced by intraseasonally varying SST may produce erroneous intraseasonal variations (Zheng et al. 2004). This is because atmospheric convection prefers warm sea surfaces, but sea surfaces tend to be cooled during MJO convectively active phases (e.g., Zhang and McPhaden 2000). The extent to which the reproduction of the MJO initiation as shown in Fig. 4 is sensitive to the prescribed SST is examined by the following sensitivity tests.

First, the control simulation (CS1), where SST is constant, is repeated using observed varying SST with intraseasonal variability (SST1A in Table 1). Figure 7 compares anomalous U850 in this simulation (Fig. 7c), the control simulation (Fig. 7b), and NNR (Fig. 7a). Between the two simulations, the anomalies are stronger in the control, but its intraseasonal eastward propagation is slower and the synoptic-scale westward- propagating perturbations are more obvious in SST1A. The intraseasonal transition from eastward-propagating easterly to westerly anomalies over the Indian Ocean, however, takes place nearly at the same time in both simulations as in NNR. The mean conditions and the horizontal and vertical structures are also very similar in the two simulations (not shown). To further confirm this result, simulation 1DOM in Table 1 (starting from 1 March 2002 with varying SST) is also repeated but using constant SST (SST1B in Table 1). The reproduction of the MJO initiation is not affected in this situation either (not shown). Even though the use of parameterized convection in our low-resolution simulation may not be able to respond to the SST variability correctly, we have shown in Fig. 4 that the model result is not sensitive to the horizontal resolution, at least for the MJO initiation.

The mean SST over the Indian and western Pacific Oceans is quite warm (28°–29°C, Fig. 8a) for the Case 1 MJO event in boreal spring. It is natural to ponder what would happen if the mean SST were lower. We therefore considered a different MJO event during boreal winter with lower mean SST (Fig. 8b). This is our Case 2. The difference between the mean SSTs in these two cases is larger than 1°C in most of the equatorial Indian Ocean where both MJO events started (Fig. 8c). This difference is larger than the average amplitude of observed intraseasonal SST anomalies (e.g., Zhang 1996). As in Case 1, the simulation of the MJO initiation in Case 2 is not sensitive at all to whether SST is constant or varying (CS2 and SST2 in Table 2; not shown). All these sensitivity tests convincingly demonstrate that the intraseasonally varying component of the SST is not a determining factor to the reproduction of the MJO initiation by TMM5 in these two cases.

b. Initial conditions

As pointed out in section 3, the most intriguing aspect of simulations 1DOM and SST1B is that the MJO initiation over the Indian Ocean is reproduced by TMM5 up to 2 months after the start of the model integration. Simulations by mesoscale models are thought to be critically dependent on the initial state of the atmosphere as well as the variation in the boundary conditions of the chosen domain (Warner et al. 1989). The presence of different mesoscale–synoptic scale variability in the model initial conditions (winds, moisture, SST, etc.) may lead to development of different convective cloud systems (scattered small cloud system versus large organized mesoscale system) and their upscaling capabilities (e.g., Chen et al. 1996; Mechem et al. 2006). The role of initial conditions in controlling the MJO initiation is investigated in a series of test runs, each starting at a different time within a period of up to 10 days. Both constant and varying (with intraseasoanl variability) SST surface conditions are used (IN_VSST1_n and IN_FSST1_n in Table 1). Examples are shown in Fig. 9. The sensitivity to initial conditions is obvious, in both the detailed distributions of the wind anomalies and their eastward propagation. But the central feature of the simulations, namely, the initiation of the MJO event as marked by the intraseasonal transition from easterly to westerly anomalies, remains roughly the same in these test runs. The mean states in these test runs are also very similar to the control simulation (not shown). The sensitivity tests are repeated for Case 2 (IN_FSST2_n in Table 2), which yielded the same results (not shown). The insensitivity to the initial conditions shown by the reproduction of the MJO initiation may not be surprising. In our case, the tropical channel domain is comparable to the zonal scale of the MJO. The lateral boundaries may have acted to constrain the simulations, and thereby reduce errors due to the initial conditions (Vukicevic and Errico 1990).

c. Latent heating and moist processes

Many theories and hypotheses on the MJO critically rely on atmospheric deep convection and its interaction with the large-scale circulation (see section 1). The extent to which convective activities affect the initiation of the MJO in the TMM5 simulations is examined in two test runs. In the first run, latent heating is set to zero through the entire simulation (NO_LH1 in Table 1). In the second, moisture is set to zero through the entire simulation (NO_MOIST1 in Table 1). Everything else is the same as in the control simulation (CS1). As expected, simulated wind anomalies in these two test runs are quite different from those in the control simulation (Fig. 10). The wind anomalies become much weaker and their eastward propagation much faster in these two test runs. Nevertheless, the intraseasonal transition from easterly to westerly anomalies over the Indian Ocean is still reproduced at about the same time as in the control simulation and in NNR. More interestingly, the vertical structure of the first baroclinic mode in the zonal wind anomaly over the Indian Ocean (90°E) is also reproduced after the onset of westerly anomalies in the lower troposphere (Fig. 11). This vertical structure is similar for the MJO, convectively coupled n = 1 equatorial Rossby waves, and convectively coupled Kelvin waves (e.g., Wheeler et al. 2000; Roundy 2008). These results, confirmed by the same set of test runs for Case 2, indicate that the intraseasonal variability can be initiated, at least in TMM5, in the absence of moisture and convective heating. This is consistent with a recent study by Lin et al. (2007) that reproduces, in a dry global model, tropical intraseasonal variability resembling the MJO in certain aspects. Without moist processes and latent heating, our model fails to produce the gross structure and propagating characteristics of the MJO. But the reproduction of the MJO initiation in their absence is intriguing enough to rethink the fundamental dynamics of the MJO.

d. Lateral boundary conditions

In the previous sections, we have described a series of sensitivity tests, which rule out any essential role of SST, initial conditions, and moist processes in the simulations of the initiation of two MJO events. The only feasible explanation left for the reproduction of the MJO initiation is the influence from the lateral boundary conditions. This possibility has been discussed from observational, theoretical, and numerical points of view (see section 1). To test this possibility, a simulation is conducted in which the lateral boundary condition is constant in time with spatial distribution fixed at the initial time (FBC1A in Table 1). The result is compared to the control simulation in Fig. 12 (top). In this simulation with constant lateral boundary conditions, the MJO signals almost disappear. The intraseasonal westerly anomalies in U850, seen in the control simulation and NNR, are replaced by spurious westward-propagating, synoptic-scale westerly perturbations. The timing of the transition from easterly to westerly anomalies is completely wrong compared to that in the control simulation and NNR. Three additional test runs further confirm this result. First, test run FBC1A with constant lateral boundary conditions is repeated but including the nested higher-resolution domain D2 over the Indian Ocean (FBC1B in Table 1). This does not change the result (not shown). Second, we repeat the test run with constant lateral boundary conditions for Case 2 (FBC2 in Table 2). The results were the same (not shown). Third, we repeat the constant boundary case for Case 2 starting from 15 October (15Oct_fixed in Table 2). Again, the intraseasonal transition from eastward-propagating easterly to westerly anomalies in U850 seen in the test run 15Oct disappears in the 15Oct_fixed simulation (bottom panels of Fig. 12).

All of the above sensitivity tests point to the lateral boundary conditions as the only factor for the reproduction of the MJO initiation for the two cases and suggest the importance of extratropical influences on the MJO, at least its initiation. But there is a problem: the latitudes of the lateral boundaries for domain D1 are 21°N and S. One may wonder if the key component in the lateral boundary conditions of the D1 domain responsible for the reproduction of the MJO initiation is actually part of the MJO itself, instead of an extratropical origin. To explore this possibility, we made two more sensitivity tests in which the latitudes of the lateral boundaries of the model are moved from 21°N and S to 28° and 38°N and S (MBV1A and MBV1B in Table 1). In both runs, the intraseasonal transition from eastward-propagating easterly to westerly anomalies in U850 seen in the control simulation and NNR is well captured, despite detailed discrepancies, such as reduced amplitudes (Fig. 13). The Kelvin component of the MJO, decaying exponentially with latitude, has near-zero projection at 38° latitudes. The Rossby component of the MJO, namely, the cyclonic/anticyclonic gyres, is indeed still significant at these latitudes (see Fig. 3). It is unknown and even inconceivable, however, that traces of the Rossby component alone at the lateral boundaries at different latitudes could induce similar intraseasonal transition between eastward-propagating easterly and westerly anomalies deep in the model domain. We therefore conclude that the lateral boundary conditions critical to the reproduction of the MJO initiation represent extratropical, instead of MJO, influences. With regard to the Rossby gyres, we have commented that domain D1 in the control simulation is too narrow in latitude to capture them. In the test run with the lateral boundaries at 38°N and S, the model does a reasonable job of capturing the Rossby gyres seen in the reanalysis (Fig. 14). This adds further confidence to the model’s capability of reproducing the full structure of the MJO, which most GCMs fail to do (e.g., Slingo et al. 1996).

5. Summary and discussion

A tropical channel model is constructed based on MM5 (now known as the Tropical MM5 or TMM5) to study MJO initiation for two observed cases, one in boreal spring of 2002, the other in boreal winter of 2000. These two cases are chosen randomly (except that they are from two different seasons) without any a priori knowledge of their initiation mechanisms. With the initial and lateral boundary conditions provided by a reanalysis product, this new model is integrated for several months. The model is able to reproduce the intraseasonal transition from eastward-propagating easterly to westerly anomalies in 850-hPa zonal wind (U850) over the Indian Ocean, which has been used to mark the initiation of the MJO. The simulated MJO events capture the gross features of the observed events, including their first-mode baroclinic structure and the Rossby gyres associated with the MJO.

The simulations are far from perfect. There are biases in the mean state and obvious errors in simulated MJO events. But the gross features of the two MJO events are reproduced in the control simulations. Gustafson and Weare (2004a,b) have reported that a regular regional version of MM5, with almost the same parameterization package, has the capability of producing reasonable MJO statistics in a 2-yr simulation. It is unmistakable that MM5 and its tropical channel version (TMM5) are useful tools in the study of the MJO.

One of the most intriguing results from the TMM5 simulations is that the MJO initiation is reproduced more than 30 days after the start of the simulations. This is far beyond the known MJO predictability limit based on studies using global models (Waliser et al. 2003). This has led to a conjecture that the prescribed lateral boundary conditions are responsible for the success of reproducing the MJO initiation. This conjecture is supported by a series of sensitivity simulations. They show that the reproduction of the MJO initiation

  • (i) does not depend on detailed characteristics of SST, namely, time varying or constant, and its distributions from boreal spring or winter;
  • (ii) is not sensitive to the model initial conditions (i.e., changing initial time within a range of 10 days);
  • (iii) does not even depend on latent heating or moisture processes in general, although the simulated eastward propagation of the MJO does; and
  • (iv) is not very sensitive to the latitudinal locations of the lateral boundaries within a range of 21°–38°.

The only factor found to be vital to the reproduction of the MJO initiation is the temporal variability of the lateral boundary conditions. No MJO initiation is produced when the lateral boundary conditions remain constant in time.

These results suggest that the extratropics play a role in the initiation of the MJO. Such a role has been discussed from observational, modeling, and theoretical points of view, not without controversy (see section 1). The following issues need further discussion:

  • (i) The suggested role of extratropical influences on MJO initiation by no means replaces the essence of convection-circulation interaction to MJO dynamics. It is apparent in our sensitivity runs NO_LH and NO_MOIST for both cases that without convective processes, intraseasonal perturbations in zonal wind, after being generated by lateral influences, would not propagate eastward at the observed MJO speed and therefore failed to transfigure themselves into the MJO by any definition. Our measure of MJO initiation has been limited to its wind field. In this sense, it is not MJO initiation in its totality.
  • (ii) Only two cases of MJO initiation are included in this study. Because of the small sample size, the results, although consistent with each other and with other recent modeling studies (e.g., Lin et al. 2007), are only suggestive. Other MJO events might be initiated without any lateral influences. A systematic approach is needed to assess the role of lateral influences to initiations relative to other factors for all observed MJO events.
  • (iii) The exact process by which the MJO is initialized by lateral influences is an unsolved puzzle in this study. It is unlikely that the moisture flux through the lateral boundaries plays any role, because the model is able to capture the intraseasonal transition in wind anomalies in the absence of latent heating and moist processes. The estimates of time evolution of meridional transport of zonal momentum reveals that in the absence of convective processes, the source of near-surface westerlies associated with the initiation lies along the southern boundaries. There is not much evidence for a change in the mean state over the Indian Ocean between the simulations with varying and constant lateral boundary conditions. However, how much deviation in the background state is enough to inhibit the MJO initiation is not known. Possible roles played by nonlinear scale interaction or stochastic processes cannot be ruled out. These possibilities are currently being investigated.
  • (iv) We interpret our modeling results in terms of extratropical influences on MJO initiation. In reality, it is impossible to separate completely the tropics from extratropics, because of their continuous and active interactions (e.g., Frederiksen and Frederiksen 1997). Rossby wave trains emanating from the MJO and propagating away from the equator bring with them cyclonic circulations that enhance convection, thereby affecting the global medium and extended range weather forecasts (e.g., Jones and Schemn 2000). Matthews (2008) recently categorized observed MJO events into those that are successive (preceded by another MJO event) and spontaneous (without a preceding MJO event). He identified the two MJO cases covered in this study were both successive events (A. Matthews 2008, personal communication). He found evidence that the initiation of successive MJO events is related to energy propagating from the extratropics into the tropics. But the extratropical energy source is in turn related to the previous MJO event, consistent with the idea of Frederiksen and Frederiksen (1997). The nature of the extratropical influences on the MJO initiation suggested by this study, therefore, needs to be further explored.

Given all the uncertainties and controversies regarding extratropical influences on the MJO, our interpretation of the results from this study serves three purposes. First, it explains the seeming contradiction of our results to the believed MJO predictability limit of 15–20 days (Waliser et al. 2003). That limit was derived from global model experiments as initial value problems. Our simulations are provided with prescribed evolution of lateral boundaries and thus are forced simulations, not initial value problems. Second, our interpretation points out that if a model fails to reproduce the MJO, the problem may not be single dimensional in its cumulus parameterization. Even if cumulus parameterization is indeed the main source of the problem, it may play its role in a different manner. A cumulus parameterization may be incapable of producing the heating profile needed to engage the large-scale circulation with a positive feedback to promote selective growth on intraseasonal and planetary scales so that the MJO is generated and maintained. This has been a common hypothesis for the difficulty of many GCMs to produce the MJO. For example, in our cumulus scheme tests on this domain, use of the Kain–Fritsch parameterization does not produce the MJO. A detailed diagnosis of the sensitivity due to the cumulus schemes is beyond the scope of this paper. An alternative view would be that when the MJO signal starts in the dynamic field, due, for instance, to lateral influences, a cumulus parameterization may actually work against it and weaken it. This may also explain the intriguing phenomenon that in almost all GCMs that produce certain features of the MJO, their MJO signals are always much stronger in wind than in precipitation or cloud than observed (Zhang et al. 2006). Third, our results point to a new possible practice in MJO prediction: a high-resolution domain of the tropics nested in a coarse-resolution global model. The latter is known to suffer less from deficiencies in cumulus parameterizations in the extratropics because of the strong dynamic control there. This approach of global nested domains, currently being explored by some modeling groups (e.g., NCAR), allows full tropical–extratropical interactions with a better treatment of tropical convection and much more computational efficiency than a global high-resolution model.

In conclusion, the numerical simulations by the tropical channel model suggest that lateral boundary conditions are the only critical factors for the initiation of two observed MJO events. Even though we are still trying to identify what exactly in the lateral boundary conditions are essential to the MJO initiation, the results provide supporting evidence to the hypothesis that extratropical influences are indispensable ingredients in the MJO dynamics. This hypothesis, having existed for some time but ignored by mainstream research of the MJO, deserves serious investigations along with other promising ideas.

Acknowledgments

The authors are grateful to Joe Tenerelli who helped build the model (TMM5) used in this study. Acknowledgment is made to the National Center for Atmospheric Research for the computing time used in this research. We also thank H. Annamalai and the two anonymous reviewers for providing valuable comments. The FNL data were taken from NCAR’s mass storage system. The ERA-40 and NCEP–NCAR reanalysis data were taken from the ECMWF data services and NOAA/CDC, respectively. The TRMM merged data were taken from the NASA–GSFC Web site. This study was partially supported by NSF Grants ATM9912297 and ATM0739402.

REFERENCES

  • Adler, R. F., , G. J. Huffman, , D. T. Bolvin, , S. Curtis, , and E. J. Nelkin, 2000: Tropical rainfall distributions determined using TRMM combined with other satellite and rain gauge information. J. Appl. Meteor., 39 , 20072023.

    • Search Google Scholar
    • Export Citation
  • Annamalai, H., , and J. M. Slingo, 1999: The mean evolution and variability of the Asian summer monsoon: Comparison of ECMWF and NCEP–NCAR reanalyses. Mon. Wea. Rev., 127 , 11571186.

    • Search Google Scholar
    • Export Citation
  • Annamalai, H., , and K. R. Sperber, 2005: Regional heat sources and the active and break phases of boreal summer intaseasonal (30–50 day) variability. J. Atmos. Sci., 62 , 27262748.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., , and M. J. Miller, 1986: A new convective adjustment scheme. Part II: Single-column testing using GATE wave, BOMEX, ATEX, and arctic air-mass datasets. Quart. J. Roy. Meteor. Soc., 112 , 693709.

    • Search Google Scholar
    • Export Citation
  • Bladé, I., , and D. L. Hartmann, 1993: Tropical intraseasonal oscillation in a simple nonlinear model. J. Atmos. Sci., 50 , 29222939.

  • Chen, S. S., , B. E. Mapes, , and R. A. Houze Jr., 1996: Multiscale variability of deep convection in relation to large-scale circulation in TOGA COARE. J. Atmos. Sci., 53 , 13801409.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1989: Numerical study of convection observed during the Winter Monsoon Experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46 , 30773107.

    • Search Google Scholar
    • Export Citation
  • Dudhia, J., 1993: A nonhydrostatic version of the Penn State–NCAR Mesoscale Model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev., 121 , 14931513.

    • Search Google Scholar
    • Export Citation
  • Fasullo, J., , and P. J. Webster, 2000: Atmospheric and surface variations during westerly wind bursts in the tropical western Pacific. Quart. J. Roy. Meteor. Soc., 128 , 123.

    • Search Google Scholar
    • Export Citation
  • Frederiksen, J. S., , and C. S. Frederiksen, 1997: Mechanisms of the formation of intraseasonal oscillations and Australian monsoon disturbances: The roles of latent heat, barotropic and baroclinic instability. Contrib. Atmos. Phys., 70 , 3956.

    • Search Google Scholar
    • Export Citation
  • Ghil, M., , and K. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part I: Northern hemisphere and tropics. J. Atmos. Sci., 48 , 752779.

    • Search Google Scholar
    • Export Citation
  • Grell, G. A., , J. Dudhia, , and D. R. Stauffer, 1995: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 117 pp.

    • Search Google Scholar
    • Export Citation
  • Gustafson, W. I., , and B. C. Weare, 2004a: MM5 modeling of the Madden–Julian oscillation in the Indian and west Pacific Oceans: Model description and control run results. J. Climate, 17 , 13201337.

    • Search Google Scholar
    • Export Citation
  • Gustafson, W. I., , and B. C. Weare, 2004b: MM5 modeling of the Madden–Julian oscillation in the Indian and west Pacific Oceans: Implications of 30–70-day boundary effects on MJO development. J. Climate, 17 , 13381351.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., 1988: A simple model of the 40–50-day oscillation. J. Atmos. Sci., 45 , 569584.

  • 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
  • Hoskins, B. J., , and G-Y. Yang, 2000: The equatorial response to higher-latitude forcing. J. Atmos. Sci., 57 , 11971213.

  • Hsu, H. H., , B. J. Hoskins, , and F-F. Jin, 1990: The 1985/86 intraseasonal oscillation and the role of the extratropics. J. Atmos. Sci., 47 , 823839.

    • Search Google Scholar
    • Export Citation
  • Hu, Q., , and D. A. Randall, 1994: Low-frequency oscillations in radiative-convective systems. J. Atmos. Sci., 51 , 10891099.

  • Hu, Q., , and D. A. Randall, 1995: Low-frequency oscillations in radiative-convective systems. Part II: An idealized model. J. Atmos. Sci., 52 , 478490.

    • Search Google Scholar
    • Export Citation
  • Inness, P. M., , J. M. Slingo, , E. Guilyardi, , and J. Cole, 2003: Simulation of the Madden–Julian oscillation in a coupled general circulation model. Part II: The role of the basic state. J. Climate, 16 , 365382.

    • Search Google Scholar
    • Export Citation
  • Janjic, Z. I., 1994: The step-mountain coordinate model: Further development of the convection, viscous sublayer, and turbulence closure schemes. Mon. Wea. Rev., 122 , 927945.

    • Search Google Scholar
    • Export Citation
  • Jones, C., , and J-K. Schemn, 2000: The influence of intraseasonal variations on medium-range weather forecasts over South America. Mon. Wea. Rev., 128 , 486494.

    • Search Google Scholar
    • Export Citation
  • Jones, C., , D. E. Waliser, , J-K. Schemn, , and W. K. M. Lau, 2000: Prediction skill of the Madden–Julian Oscillation in dynamical extended range forecasts. Climate Dyn., 16 , 273289.

    • 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. R., , and B. C. Weare, 2001: The onset of convection in Madden–Julian oscillation. J. Climate, 14 , 780793.

  • Kiladis, G. N., , K. H. Straub, , and P. T. Haertel, 2005: Zonal and vertical structure of the Madden–Julian oscillation. J. Atmos. Sci., 62 , 27902809.

    • Search Google Scholar
    • Export Citation
  • Knutson, R. R., , and K. M. Weickmann, 1987: 30–60 day atmospheric oscillations: Composite life cycles of convection and circulation anomalies. Mon. Wea. Rev., 115 , 14071436.

    • Search Google Scholar
    • Export Citation
  • Knutson, R. R., , K. M. Weickmann, , and J. E. Kutzbach, 1986: Global-scale intraseasonal oscillations of outgoing longwave radiation and 250-mb zonal wind during Northern Hemisphere summer. Mon. Wea. Rev., 114 , 605623.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., , and L. Peng, 1987: Origin of low-frequency (intraseasonal) oscillations in the tropical atmosphere. Part I: Basic theory. J. Atmos. Sci., 44 , 950972.

    • Search Google Scholar
    • Export Citation
  • Lau, K. M., , and D. E. Waliser, 2005: Intraseasonal Variability in the Atmosphere–Ocean Climate System. Praxis, 436 pp.

  • Lin, H., , G. Brunet, , and J. Derome, 2007: Intraseasonal variability in a dry atmospheric model. J. Atmos. Sci., 64 , 24222441.

  • 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
  • Madden, R. A., , and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29 , 11091123.

    • Search Google Scholar
    • Export Citation
  • Matthews, A., 2008: Primary and successive events in the Madden–Julian oscillation. Quart. J. Roy. Meteor. Soc., 134 , 439453.

  • Mechem, D. B., , S. S. Chen, , and R. A. Houze Jr., 2006: Momentum transport processes in the stratiform regions of mesoscale convective systems over the western Pacific warm pool. Quart. J. Roy. Meteor. Soc., 132A , 709736.

    • Search Google Scholar
    • Export Citation
  • Mlawer, E. J., , S. J. Taubman, , P. D. Brown, , M. J. Jacono, , and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102 , 1666316682.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , and J-Y. Yu, 1994: Modes of tropical variability under convective adjustment and the Madden–Julian oscillation. Part I. Analytical theory. J. Atmos. Sci., 51 , 18761894.

    • Search Google Scholar
    • Export Citation
  • Newman, M., , P. D. Sardeshmukh, , and J. W. Bergman, 2000: An assessment of the NCEP, NASA, and ECMWF reanalyses over the tropical west Pacific warm pool. Bull. Amer. Meteor. Soc., 81 , 4148.

    • Search Google Scholar
    • Export Citation
  • Roundy, P. E., 2008: Analysis of convectively coupled Kelvin waves in the Indian Ocean MJO. J. Atmos. Sci., 65 , 13421359.

  • Rui, H., , and B. Wang, 1990: Development characteristics and dynamic structure of tropical intraseasonal convection anomalies. J. Atmos. Sci., 47 , 357379.

    • Search Google Scholar
    • Export Citation
  • Salby, M. L., , and H. H. Hendon, 1994: Planetary-scale circulations in the presence of climatological and wave-induced heating. J. Atmos. Sci., 51 , 23442367.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., , and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45 , 12281251.

    • 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
  • Straub, K. H., , and G. N. Kiladis, 2002: Observations of a convectively coupled Kelvin wave in the eastern Pacific ITCZ. J. Atmos. Sci., 59 , 3053.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 reanalysis. Quart. J. Roy. Meteor. Soc., 131 , 29613012.

  • Vukicevic, T., , and R. M. Errico, 1990: The influence of artificial and physical factors upon predictability estimates using a complex limited-area model. Mon. Wea. Rev., 118 , 14601482.

    • 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
  • Waliser, D. E., , K. M. Lau, , W. Stern, , and C. Jones, 2003: Potential predictability of the Madden–Julian oscillation. Bull. Amer. Meteor. Soc., 84 , 3350.

    • Search Google Scholar
    • Export Citation
  • Wang, B., , and T. Li, 1994: Convective interaction with boundary layer dynamics in the development of the tropical intraseasonal signal. J. Atmos. Sci., 51 , 13861400.

    • Search Google Scholar
    • Export Citation
  • Warner, T. T., , L. E. Key, , and A. M. Lario, 1989: Sensitivity of mesoscale-model forecast skill to some initial data characteristics, data density, data position, analysis procedure, and measurement error. Mon. Wea. Rev., 117 , 12811310.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., , and H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132 , 19171932.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M., , G. N. Kiladis, , and P. J. Webster, 2000: Large-scale dynamical fields associated with convectively coupled equatorial waves. J. Atmos. Sci., 57 , 613640.

    • Search Google Scholar
    • Export Citation
  • Wilson, J. D., , and M. Mak, 1984: Tropical response to lateral forcing with a latitudinally and zonally nonuniform basic state. J. Atmos. Sci., 41 , 11871201.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , and M-M. Lu, 1983: Equatorially trapped waves at 200 mb and their association with meridional convergence of wave energy flux. J. Atmos. Sci., 40 , 27852803.

    • Search Google Scholar
    • Export Citation
  • Yu, J-Y., , and J. D. Neelin, 1994: Modes of tropical variability under convective adjustment and the Madden–Julian oscillation. Part II: Numerical results. J. Atmos. Sci., 51 , 18951914.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., 1996: Atmospheric intraseasonal variability at the surface in the western Pacific Ocean. J. Atmos. Sci., 53 , 739758.

  • Zhang, C., 2005: Madden-Julian Oscillation. Rev. Geophys., 43 , RG2003. doi:10.1029/2004RG000158.

  • Zhang, C., , and M. J. McPhaden, 2000: Intraseasonal surface cooling in the equatorial western Pacific. J. Climate, 13 , 22612276.

  • Zhang, C., , and M. Dong, 2004: Seasonality of the Madden–Julian oscillation. J. Climate, 17 , 31693180.

  • Zhang, C., , M. Dong, , S. Gualdi, , H. H. Hendon, , E. D. Maloney, , A. Marshall, , K. R. Sperber, , and W. Wang, 2006: Simulations of the Madden-Julian oscillation by four pairs of coupled and uncoupled global models. Climate Dyn., 27 , 573592. doi:10.1007/s00382-006-0148-2.

    • 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

Fig. 1.
Fig. 1.

Primary model domain for the TMM5 (D1, 21°S–21°N, 0°–360°) and the nested domain (D2, 11°S–11°N, 37°–183°E). Domains D1 and D2 have resolutions of 111 and 37 km, respectively. Control simulations include only D1.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 2.
Fig. 2.

Time-longitude diagrams of daily (a) 850-hPa zonal wind (m s−1) from the NCEP–NCAR reanalysis and (b) precipitation (mm day−1) from the merged TRMM datasets for Case 1 (averaged over 10°S–10°N). (c), (d) Case 2. The two cases selected in this study are marked by the straight lines whose slope corresponds to an eastward-propagation speed of 5 m s−1.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 3.
Fig. 3.

200-hPa wind anomalies from the NCEP–NCAR reanalysis during (a) 10–15 May 2002 for Case 1 and (b) 23–27 Nov 2000 for Case 2.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 4.
Fig. 4.

Time-longitude diagrams of daily U850 anomalies (m s−1) averaged over 10°S–10°N for Case 1 from (a) NNR, (b) TMM5 single-domain simulation, and (c) TMM5 nested-domain simulation. A 3-day running mean is applied.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 5.
Fig. 5.

(left) Mean U850 for Case 1 (10 Apr–10 Jun 2002) and (right) Case 2 (10 Nov–31 Dec 2000) from the (top) NNR, (middle) ERA-40, and (bottom) control simulations (CS1 and CS2). Contour interval is 2 m s−1.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 6.
Fig. 6.

Vertical structure of the zonal wind anomalies (m s−1) in a 5° × 5° box centered at 0°, 90°E for Case 1 from (a) NNR and (b) control simulation (CS1). The solid (dashed) contours represent positive (negative) anomalies. A 7-day running mean is applied.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 7.
Fig. 7.

U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1 (with constant SST), and (c) test run SST1A (with varying SST). The black lines indicate a zonal propagation speed of 5 m s−1.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 8.
Fig. 8.

Averaged SST for (a) Case 1 (10 Apr–10 Jun 2002), (b) Case 2 (10 Nov–31 Dec 2000), and (c) their differences. Contour interval is 0.5°C.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 9.
Fig. 9.

U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run IN_FSST1_1 starting 5 days before the control simulation, and (d) test run IN_FSST1_5 starting 5 days after the control simulation.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 10.
Fig. 10.

U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run NO_LH1 with no latent heating, and (d) test run NO_MOIST1 with no moisture. All are averaged over 10°S to 10°N. The black lines mark the zonal propagation speed of 5 m s−1.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 11.
Fig. 11.

Time-height diagrams of zonal wind anomalies (m s−1) in a 5° × 5° box centered at 0°, 90°E for Case 1 from (a) control simulation CS1 and (b) test run NO_LH1 with no latent heating, and (c) test run NO_MOIST1 with no moisture. The solid (dashed) contours represent positive (negative) anomalies. A 7-day running mean is applied.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 12.
Fig. 12.

U850 anomalies (m s−1; 3-day running mean) for (top) Case 1 from (a) NNR, (b) control simulations CS1, and (c) test run FBC1A with constant lateral boundary conditions. (bottom) Case 2 from (d) NNR, (e) test run 15Oct starting from 15 Oct, and (f) test run 15Oct_fixed with constant boundary conditions. All averaged over 10°S–10°N.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 13.
Fig. 13.

U850 anomalies (m s−1; 3-day running mean) for Case 1 from (a) NNR, (b) control simulation CS1, (c) test run MBV1A with lateral boundaries at 28°S and N, and (d) test run MBV1B with lateral boundaries at 38°S and N.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Fig. 14.
Fig. 14.

Winds (10–15 May 2002; m s−1) at 200 hPa for Case 1 from (a) NNR, and (b) test run MBV1B with lateral boundaries at 38°S and N.

Citation: Journal of the Atmospheric Sciences 66, 2; 10.1175/2008JAS2701.1

Table 1.

The descriptions of the simulations for Case 1 (2002 event). The horizontal resolution is 111 km for all simulations, except the nested run (2DOM) and one of the fixed boundary conditions run (FBC1B). All other sensitivity tests are conducted using single domain from 10 Apr to 10 Jun 2002. Constant and varying refer to time. Varying SST has intraseasonal fluctuations.

Table 1.
Table 2.

The descriptions of the simulations for Case 2 (2000 event). The single domain (111-km resolution; D1 of Fig. 1) is used in all simulations. Constant and varying refer to time. Varying SST has intraseasonal fluctuations.

Table 2.

1

* The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Save