• Betts, A. K., , and M. J. Miller, 1986: A new convective adjustment scheme. Part II: Single column tests using GATE wave, BOMEX, ATEX, and arctic air-mass data sets. Quart. J. Roy. Meteor. Soc., 112 , 693709.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., , and M. J. Miller, 1993: The Betts–Miller scheme. The Representation of Cumulus Convection in Numerical Models of the Atmosphere, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 107–121.

    • Search Google Scholar
    • Export Citation
  • Chao, W. C., 2000: Multiple quasi equilibria of the ITCZ and the origin of monsoon onset. J. Atmos. Sci., 57 , 641652.

  • Chao, W. C., , and B. Chen, 2004: Single and double ITCZ in an aqua-planet model with constant sea surface temperature and solar angle. Climate Dyn., 22 , 447459.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , and J. D. Neelin, 1996: Linearization of a longwave radiation scheme for intermediate tropical atmospheric models. J. Geophys. Res., 101 , 1512915145.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , and J. D. Neelin, 1999: Cirrus detrainment–temperature feedback. Geophys. Res. Lett., 26 , 12951298.

  • Chou, C., , and J. D. Neelin, 2004: Mechanisms of global warming impacts on regional tropical precipitation. J. Climate, 17 , 26882701.

  • Chou, C., , J. D. Neelin, , and H. Su, 2001: Ocean–atmosphere–land feedbacks in an idealized monsoon. Quart. J. Roy. Meteor. Soc., 127 , 18691892.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106 , 447462.

  • Hack, J. J., , W. H. Schubert, , D. E. Stevens, , and H-C. Kuo, 1989: Response of the Hadley circulation to convective forcing in the ITCZ. J. Atmos. Sci., 46 , 29572973.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., , M. E. McIntyre, , and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Joly, A., , and A. J. Thorpe, 1990: Frontal instability generated by tropospheric potential vorticity anomalies. Quart. J. Roy. Meteor. Soc., 116 , 525560.

    • Search Google Scholar
    • Export Citation
  • Kirtman, B. P., , and E. K. Schneider, 2000: A spontaneously generated tropical atmospheric general circulation. J. Atmos. Sci., 57 , 20802093.

    • Search Google Scholar
    • Export Citation
  • Kuo, H. L., 1973: Dynamics of quasi-geostrophic flows and instability theory. Adv. Appl. Mech., 13 , 247330.

  • Lin, X., , and R. H. Johnson, 1996: Kinematic and thermodynamic characteristics of the flow over the western Pacific warm pool during TOGA COARE. J. Atmos. Sci., 53 , 695715.

    • Search Google Scholar
    • Export Citation
  • Magnusdottir, G., , and C-C. Wang, 2008: Intertropical convergence zones during the active season in daily data. J. Atmos. Sci., 65 , 24252436.

    • Search Google Scholar
    • Export Citation
  • Mechoso, C., and Coauthors, 1995: The seasonal cycle over the tropical Pacific in coupled ocean–atmosphere general circulation models. Mon. Wea. Rev., 123 , 28252838.

    • 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 results. J. Atmos. Sci., 51 , 18761894.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , and N. Zeng, 2000: A quasi-equilibrium tropical circulation model—Formulation. J. Atmos. Sci., 57 , 17411766.

  • Nieto Ferreira, R., , and W. H. Schubert, 1997: Barotropic aspects of ITCZ breakdown. J. Atmos. Sci., 54 , 261285.

  • Raymond, D. J., , G. B. Raga, , C. S. Bretherton, , J. Molinari, , C. López-Carrillo, , and Z. Fuchs, 2003: Convective forcing in the intertropical convergence zone of the eastern Pacific. J. Atmos. Sci., 60 , 20642082.

    • Search Google Scholar
    • Export Citation
  • Schubert, W. H., , P. E. Ciesielski, , D. E. Stevens, , and H-C. Kuo, 1991: Potential vorticity modeling of the ITCZ and the Hadley circulation. J. Atmos. Sci., 48 , 14931509.

    • Search Google Scholar
    • Export Citation
  • Sumi, A., 1990: Pattern formation of convective activity over the aqua-planet with globally uniform sea surface temperature. Meteorological Research Rep. 90-1, Division of Meteorology, Geophysical Institute, University of Tokyo, 124 pp.

    • Search Google Scholar
    • Export Citation
  • Wang, C-C., , and G. Magnusdottir, 2005: ITCZ breakdown in three-dimensional flows. J. Atmos. Sci., 62 , 14971512.

  • Wang, C-C., , and G. Magnusdottir, 2006: The ITCZ in the central and eastern Pacific on synoptic time scales. Mon. Wea. Rev., 134 , 14051421.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., , V. O. Magaña, , T. N. Palmer, , J. Shukla, , R. A. Tomas, , M. Yanai, , and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103 , 1445114510.

    • 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
  • Yu, J-Y., , and J. D. Neelin, 1997: Analytic approximations for moist convectively adjusted regions. J. Atmos. Sci., 54 , 10541063.

  • Yu, J-Y., , C. Chou, , and J. D. Neelin, 1998: Estimating the gross moist stability of the tropical atmosphere. J. Atmos. Sci., 55 , 13541372.

    • Search Google Scholar
    • Export Citation
  • Zehnder, J. A., , and D. M. Powell, 1999: The interaction of easterly waves, orography, and the intertropical convergence zone in the genesis of eastern Pacific tropical cyclones. Mon. Wea. Rev., 127 , 15661585.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., , J. D. Neelin, , and C. Chou, 2000: A quasi-equilibrium tropical circulation model–implementation and simulation. J. Atmos. Sci., 57 , 17671796.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    An ITCZ breakdown case triggered by easterly waves: the anomalous wind (vectors) and surface relative vorticity (shading) on (a) 6, (b) 8, (c) 10, and (d) 12 August 2000 derived from QuikSCAT wind. Units for vorticity are 1 × 10−4 s−1. The average wind is obtained from all available QuikSCAT data in June–August.

  • View in gallery

    The dry experiment of the ITCZ breakdown on day (a) 5, (b) 7, (c) 9, and (d) 11. Contours (negative dashed) indicate relative vorticity at 850 hPa in units of 1 × 10−5 s−1. Vectors indicate the wind field at 850 hPa. The x and y axes are relative longitude and latitude, respectively. The area of prescribed heating is indicated by the thick dashed line.

  • View in gallery

    The moisture-on experiment of ITCZ breakdown on (a),(c) day 3 and (b),(d) day 7, showing relative vorticity (contours, negative dashed) and wind (vectors) at 850 hPa (left) and precipitation (right). The prescribed heating is indicated by the thick dashed line in (a) and (c).

  • View in gallery

    As in Fig. 3 but for the all-physics-on experiment of ITCZ breakdown.

  • View in gallery

    The evaporation (grayscale), surface wind speed (contours), and the wind vectors of the all-physics-on experiment of ITCZ breakdown on (a) day 7 and (b) day 16. The zero wind speed line is thickened. The contour interval is 5 m s−1 and the unit of evaporation is W m−2.

  • View in gallery

    The development of the tail (expt 1). Experimental design is the same as the all-physics-on experiment (Fig. 4) but for a shorter heating area, indicated by a dashed line in (d). The contour field is relative vorticity; vectors indicate the wind field at 850 hPa. The box in (e) indicates the area for budget analysis.

  • View in gallery

    The time evolution of each term in the temperature and moisture equations of (expt 1 − ctrl). The x axis is model day. The y axis in (a)–(c) is the heating rate (K day−1). The surface wind speed (solid, left y axis, m s−1) and the first baroclinic divergence (dashed, right y axis, 1 × 10−5 s−1) are shown in (d).

  • View in gallery

    Time evolution of the energy budget for (a) expt 1, (b) expt 2 (suppress horizontal moisture convergence), (c) expt 3 (suppress surface evaporation), (d) expt 4 (warmer SST), (e) expt 5 (warmer SST, suppress horizontal moisture convergence), and (f) expt 6 (mixed layer ocean). The x and y axes are the model day and the heating rate (K day−1), respectively.

  • View in gallery

    The surface wind speed (m s−1) (solid lines, left y axis) and mode-1 divergence (1 × 10−5 s−1) (dashed lines, right y axis). Thick lines indicate the experiment in which the Kelvin wave is filtered; thin lines are from expt 1. Values shown here are averaged over the box of budget analysis.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 53 53 6
PDF Downloads 29 29 1

Breakdown and Reformation of the Intertropical Convergence Zone in a Moist Atmosphere

View More View Less
  • 1 Research Center for Environmental Changes, Academia Sinica, and Department of Atmospheric Sciences, Chinese Culture University, Taipei, Taiwan
  • | 2 Research Center for Environmental Changes, Academia Sinica, and Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan
  • | 3 Research Center for Environmental Changes, Academia Sinica, Taipei, Taiwan
© Get Permissions
Full access

Abstract

The effects of moisture on the intertropical convergence zone (ITCZ) over the eastern Pacific on the synoptic time scale are investigated using an intermediate complexity atmospheric circulation model, the quasi-equilibrium tropical circulation model (QTCM1), on an aquaplanet.

The dry simulation shows results consistent with those of simple dynamic models, except that a slightly stronger heating rate is needed owing to different model designs. In the moist simulations, the most important result is the formation of a tail southwest of a vortex during and after the ITCZ breakdown. This tail may extend zonally more than 60° longitude and last for more than two weeks in an idealized simulation. In the eastern North Pacific, this phenomenon is often observed in cases that involve easterly waves. In a sense, the formation of the tail suggests a possible mechanism that forms an ITCZ efficiently.

This study shows that the surface convergent flow induced by a disturbance initializes a positive wind–evaporation feedback that forms the tail. In the tail, the most important energy source is surface evaporation, and the latent heat is nicely balanced by an adiabatic cooling of the ascending motion. In other words, the energy is redistributed vertically by vertical energy convergence.

The lifespan of the tail is controlled by the propagation of tropical waves that modify the surface wind pattern, leading to a decrease in surface wind speed and corresponding surface fluxes. It may explain the absence of the tail in some of the events in the real atmosphere.

Corresponding author address: Chia-chi Wang, Department of Atmospheric Sciences, Chinese Culture University, 55 Hwa-Kang Road, Taipei, 111, Taiwan. Email: wang1794@rcec.sinica.edu.tw

Abstract

The effects of moisture on the intertropical convergence zone (ITCZ) over the eastern Pacific on the synoptic time scale are investigated using an intermediate complexity atmospheric circulation model, the quasi-equilibrium tropical circulation model (QTCM1), on an aquaplanet.

The dry simulation shows results consistent with those of simple dynamic models, except that a slightly stronger heating rate is needed owing to different model designs. In the moist simulations, the most important result is the formation of a tail southwest of a vortex during and after the ITCZ breakdown. This tail may extend zonally more than 60° longitude and last for more than two weeks in an idealized simulation. In the eastern North Pacific, this phenomenon is often observed in cases that involve easterly waves. In a sense, the formation of the tail suggests a possible mechanism that forms an ITCZ efficiently.

This study shows that the surface convergent flow induced by a disturbance initializes a positive wind–evaporation feedback that forms the tail. In the tail, the most important energy source is surface evaporation, and the latent heat is nicely balanced by an adiabatic cooling of the ascending motion. In other words, the energy is redistributed vertically by vertical energy convergence.

The lifespan of the tail is controlled by the propagation of tropical waves that modify the surface wind pattern, leading to a decrease in surface wind speed and corresponding surface fluxes. It may explain the absence of the tail in some of the events in the real atmosphere.

Corresponding author address: Chia-chi Wang, Department of Atmospheric Sciences, Chinese Culture University, 55 Hwa-Kang Road, Taipei, 111, Taiwan. Email: wang1794@rcec.sinica.edu.tw

1. Introduction

The intertropical convergence zone in the eastern Pacific has been observed to be highly active on the synoptic time scale from summer to fall (Wang and Magnusdottir 2006). Compared to the conventional view of the ITCZ in climatology, in which the ITCZ is the average of many individual convective cells, we are focusing on a different time scale in which these two points of view do not contradict each other. In other words, the climatological ITCZ represents statistical information of the synoptic-time-scale ITCZ, including both individual convective cells and organized zonal structures.

ITCZ dynamics have been studied for more than a decade in several studies. Researchers have been interested in its evolution, preferred location, vertical structure, and interactions with other tropical phenomena. The eastern Pacific is a relatively simple region in which to study the dynamics of the ITCZ compared to other ocean basins, such as the western Pacific, where the structure and behavior of the ITCZ are strongly influenced by the warm pool and monsoon circulation (e.g., Lin and Johnson 1996; Webster et al. 1998). The eastern Pacific ITCZ exhibits strong modes of variation throughout its life cycle, including undulation, breakdown, and reformation or dissipation, on the synoptic time scale. The ITCZ can be disturbed by easterly waves that propagate from the Atlantic basin (Zehnder and Powell 1999) or self-destroyed by the effect of barotropic instability (Nieto Ferreira and Schubert 1997). Both processes generate tropical disturbances that have the potential to develop into tropical cyclones.

A series of modeling studies has focused on understanding the breakdown process. It is found that heating in the midatmosphere, usually from convection, can induce unstable flows in the lower atmosphere (Hack et al. 1989; Schubert et al. 1991) because the heating may generate a positive (negative) potential vorticity (PV) anomaly in the lower (upper) level and cause the sign of the meridional PV gradient to be reversed on the equatorward (poleward) side of the heating (Hoskins et al. 1985, ch. 7). This PV gradient sign reversal induces flow instability (the so-called Rayleigh instability theorem).

Nieto Ferreira and Schubert (1997) and Wang and Magnusdottir (2005) both used dry models to simulate an ITCZ breakdown. The former successfully simulated the basic features of an ITCZ breakdown in the lower troposphere by prescribing a mass sink in a shallow water model, whereas the latter extended the simulation to three dimensions and investigated the effects of background flows on the breakdown process by adding a heat source in a primitive equation (PE) model. Once the vorticity strip is induced by the forcing, the major axis of the vorticity strip tilts counterclockwise due to its own cyclonic flow. As the perturbation grows with time, a series of vortices roll up, intensify, and eventually break apart. The number of vortices depends on the width of the strip and the wavelength of the most unstable mode in the flow (e.g., Kuo 1973; Joly and Thorpe 1990). This evolution captures the basic characteristics of the ITCZ breakdown on the synoptic time scale. Both studies conclude that barotropic instability is the main contribution to an ITCZ breakdown. The background flow may accelerate or decelerate the evolution of the ITCZ breakdown, depending on the direction of the flow. If the background flow enhances the horizontal wind shear, the ITCZ structure becomes more unstable and vice versa. The background flow can advect disturbances produced from the breakdown (Wang and Magnusdottir 2005). However, a recent work (Magnusdottir and Wang 2008) that analyzed reanalysis data showed that these disturbances originating from the ITCZ propagate as a wave packet with a much faster phase speed (−8 m s−1) than the background flow in the eastern Pacific (−2.6 m s−1). Their results show that the ITCZ has a pronounced wavelike characteristic.

Another characteristic of the ITCZ is its strong tendency to reform. Wang and Magnusdottir (2006) surveyed the frequency of ITCZ breakdown events using operational analysis data and satellite data. They found that the life span of an ITCZ breakdown event varies from several days to three weeks and the formation of a new ITCZ is rather fast. With a new ITCZ forming within a couple of days, another ITCZ breakdown may occur and produce numerous tropical disturbances. However, the hypothesis of an ITCZ formation (Chao 2000; Chao and Chen 2004) does not explain the time scale of its reformation. Other processes may be involved in speeding up the formation of the ITCZ. It is possible that the fast reformation of the ITCZ is related to its own breakdown, which provides initial disturbances to induce the wind–evaporation feedback. This concept will be elaborated in section 3.

Generally, tropical regions are moist and warm, and trade winds converge at the location of the ITCZ. However, it is not clear how a large-scale convergence zone develops and organizes efficiently. The ITCZ simulation in the eastern Pacific in general circulation models (GCMs) is not satisfactory (e.g., Mechoso et al. 1995). In some models, the ITCZ has seasonal migrations in the meridional direction, which is not consistent with observations, and some models produce a double ITCZ structure that is too strong to be consistent with observations. A fundamental understanding of the ITCZ dynamics and its interactions with other components is needed in order to solve the problems of ITCZ simulations in GCMs correctly.

Some studies have tried to explain the formation of the ITCZ through dynamical processes. Using an atmospheric general circulation model (AGCM) forced by a prescribed sea surface temperature (SST), Sumi (1990) found that the organization of the ITCZ is sensitive to the rotation rate of the earth. Kirtman and Schneider (2000) further coupled an AGCM to a mixed layer ocean model and concluded that the sufficient condition to form an ITCZ and the corresponding large-scale circulation is the earth’s rotation (i.e., the latitudinal variation of the Coriolis force), while the SST pattern and the solar flux may modify the actual location of the ITCZ. Chao (2000) and Chao and Chen (2004) suggest that there are two effects from the Coriolis force: one pulls the convection toward the equator and the other pulls the convection to the pole. Under uniform SST and heat flux conditions, the competition between these two effects determines the location of the ITCZ. The time scale of organizing an ITCZ is at least 10 days based on Chao’s hypothesis because the second effect of the Coriolis force involves the moist convective scheme in the model.

In this study, the effects of moist processes associated with ITCZ breakdown and reformation will be investigated. The quasi-equilibrium tropical circulation model 1 version 2.3 (QTCM1V2.3) developed at the University of California, Los Angeles (Neelin and Zeng 2000; Zeng et al. 2000), is employed to investigate the synoptic-time-scale ITCZ in a moist atmosphere. The outline of this paper is as follows: section 2 briefly describes the model and experimental design, section 3 summarizes the main features of the ITCZ, section 4 describes the results, and sections 5 and 6 are the discussion and conclusions.

2. Model and experimental design

a. The quasi-equilibrium tropical circulation model

In this study, an intermediate atmospheric model (Neelin and Zeng 2000), which is usually coupled with a simple land surface model (Zeng et al. 2000) and a mixed layer ocean model (Chou et al. 2001), is used. Under a quasi-equilibrium moist convective constraint, typical vertical structures of the atmospheric profile (Neelin and Yu 1994; Yu and Neelin 1994) can be derived from the Betts–Miller deep convection scheme (Betts and Miller 1986, 1993). This intermediate atmospheric model uses these analytical solutions for deep convection as leading basis functions for a Galerkin expansion in the vertical dimension. Thus, total fields at a given vertical level p can be written as
i1520-0469-67-4-1247-eq1
where Tr(p) and qr(p) are the spatially uniform reference profiles of temperature and moisture for a typical deep convection in the tropics; a1, b1, and V1 are the typical vertical profile of deep convection; v0 and v1 are the barotropic and baroclinic components of horizontal winds v; and T1 and q1 are the temperature and moisture anomalies projected onto the a1 and b1 profiles, respectively. Temperature T and moisture q are in energy units, with heat capacity at constant pressure Cp and the latent heat per unit mass L absorbed. With a single vertical structure of deep convection for temperature and moisture, this quasi-equilibrium tropical circulation model is referred to as QTCM1. In the tropics, QTCM1 is similar to a general circulation model running with the Betts–Miller deep convection scheme. Away from the tropics, on the other hand, QTCM1 is roughly equivalent to a two-layer model.

A cloud–radiation package that includes a linearized longwave radiation scheme (Chou and Neelin 1996), a simplified shortwave radiation scheme (Zeng et al. 2000), and an empirical cloud–radiation scheme is used to calculate effects of deep convection, cirrocumulus/cirrostratus (CsCc), and cirrus clouds (Chou and Neelin 1999). Simple bulk formulas are used to calculate surface fluxes, such as the momentum flux, evaporation, and sensible heat flux (Zeng et al. 2000). A mixed layer ocean with prescribed Q flux (divergence of ocean heat transport) is used to couple with the atmospheric model.

b. Moisture and moist static energy budgets

We are interested in the moisture budget and moist static energy balance in the simulations. Thus, the moisture and thermodynamic equations are briefly introduced and the details, including momentum equations [Eqs. (5.1) and (5.2) in Neelin and Zeng 2000], can be referred to Neelin and Zeng (2000) and Chou and Neelin (2004).

The vertically integrated moisture equation in the model is
i1520-0469-67-4-1247-e1
where angle brackets denote a mass integration through the troposphere; therefore, 〈v · q〉 is the vertically integrated moisture convergence associated with horizontal flows v. The moisture stratification Mq is defined as 〈Ω∂pq〉, where Ω is a typical vertical profile of the vertical motion of deep convection (Neelin and Yu 1994; Yu et al. 1998). Thus, the Mq · v1 is the moisture convergence associated with vertical motions ω, that is, −〈ωpq〉. The summation of these two convergence terms is the three-dimensional moisture convergence. Overall, the moisture tendency is determined by moisture convergence (in the flux form), moisture sink 〈Qq〉, and surface evaporation, evap.
The vertically integrated temperature equation is in a form similar to the moisture equation along with convective heating 〈Qc〉, radiative fluxes Frad, and surface sensible heat flux FTs on the right-hand side, as shown in the following:
i1520-0469-67-4-1247-e2
The temperature convergence is also decomposed to the horizontal component 〈v · T〉 and the vertical component Ms · v1 in which Ms, the gross dry stability, is defined as −〈Ω∂ps〉, where s = T + ϕ and ϕ is geopotential.
Consequently, the moist static energy equation, which combines the moisture and temperature equations, can be written as
i1520-0469-67-4-1247-e3
with the net energy flux into the column Fnet = FTs + evap + Frad. Moist static energy is h = s + q, so M = −〈Ω∂ph〉 = MsMq. Note that the convective heating 〈Qc〉 is compensated by the moisture sink 〈Qq〉; that is, 〈Qc〉 = −〈Qq〉.

c. Experimental design

There are two sets of experiments. The first set (Table 1) aims to understand the effects of moisture in the breakdown process. The experiments are forced by a prescribed heating that initiates the circulation of deep convection. The spatial dimension of the prescribed heating is 10° latitude and 120° longitude (centered at 10°N) and it is added to the temperature equation for the first five days to mimic the diabatic heating in the ITCZ. The heating rate smoothly increases to the target value and decreases to zero within half a day at the beginning and the end of the heating period to minimize the shock of turning the heating on and off. Damping of the model is scale-selected to be much stronger on small-scale perturbations; therefore, the shocks are not observed in our simulations. The column-integrated heating rate 〈Qc〉 is given as 8 K day−1, corresponding to 6.5 K day−1 at 600 hPa.

Each experiment is a pair of runs; one is the control run without the prescribed heating and provides the basic state for the experiment run that has the prescribed heating. All variables shown in this study are deviations from the control run.

The first experiment (dry) is a dry atmosphere on an aquaplanet, and the result is compared to that of previous work (Wang and Magnusdottir 2005). The moisture is turned off and all surface momentum fluxes and energy fluxes; that is, evap, Frad, and FTs are blocked out during the simulation so that the boundary conditions from the surface are not affecting the atmosphere.

The second experiment (moisture-on) has moisture and surface evaporation, evap, turned on while all other physical processes (surface momentum fluxes Frad and FTs) remain off. This experiment is a transition step to the third experiment (all-physics-on) that has all physical processes turned on to compare with real events. Several intermediate experiments are performed as well to investigate the effect of each physical process, but the results show that they are of minor importance and will not be discussed here.

The SST pattern is a uniform 300 K between 30°N and 30°S and is linearly decreased to 273 K at the poles. The SST value is chosen to represent the relatively cooler SST in the eastern Pacific compared to the western Pacific. The effects of the meridional SST gradient along the edges of 30°N/S should be minor and are neglected in this study. The minimum wind speed that accounts for subgrid wind contributions to surface fluxes is chosen to be 2 m s−1, which is smaller than that in the standard released version 2.3. The horizontal resolution is 1° × 1° in order to depict the narrow structure of the convergence zone and the tropical cyclones. The model domain is global in longitude and extends to 78.5° north and south. The north–south boundaries are solid wall, which means the meridional velocity is set to zero there. The model is run in the perpetual mode using the solar condition for July. Viscosity coefficients are smaller than the model default settings by two orders of magnitude to retain weak initial perturbations in the mean flow.

The second set of experiments, listed in Table 2, is designed for the energy budget analysis and to investigate the importance of each process. The dimension of the prescribed heating is 10° latitude and 20° longitude. Such a heating domain generates a single cyclone (e.g., Kuo 1973; Joly and Thorpe 1990) and avoids possible multicyclone interaction. The basic state of each experiment is provided by its own control run, which is a moist atmosphere driven by the same solar condition and SST pattern or the same ocean model. Different terms in the moisture and temperature equations are suppressed in a series of experiments by removing the specific term from the equation.

Experiment 1 has the prescribed heating to initiate the disturbance with none of the processes suppressed. The horizontal moisture convergence 〈v · q〉 is suppressed in experiment 2. The purpose of this experiment is to examine the importance of horizontal moisture convergence. Surface evaporation, evap, which is the major source of energy, is suppressed in experiment 3. The horizontal temperature convergence 〈v · T〉 is examined as well, but the contribution of this term is very small and is not discussed here.

The SST pattern is the same as that of the first set of experiments for experiment 1 to experiment 3. Two cases with warmer SST (experiments 4 and 5, 305 K) and a case coupled with a mixed layer ocean (experiment 6, SST is 300 K initially) are also performed to gain an insight into the air–sea interaction associated with ITCZ breakdown.

Experiment 7 is specially designed to filter the equatorial Kelvin wave in order to provide a comparison with experiment 1 to show the effects of large-scale wave propagation. The wind field at +120° relative longitude is set to use the values from a control run without the prescribed heating. The Kelvin wave signal is filtered as it propagates through the “wall” at +120°.

3. The observed ITCZ over the eastern Pacific Ocean

Figure 1 shows a case of ITCZ breakdown triggered by easterly waves from 6 to 12 August 2000. The shading is relative vorticity (1 × 10−4 s−1) derived from the QuikSCAT surface wind and the vector indicates wind direction of the anomalous wind. The mean of surface wind is derived from all available (1999–present) QuikSCAT wind from June to August. A noticeable feature is the intensification of the long tail southwest of the vortex around 20°N, 120°W on 10 August during the cyclone formation. The number of pixels that have relative vorticity ≥2 × 10−4 s−1 within the area 6°–14°N, 150°–120°W is counted. The averaged value of these high vorticity pixels is computed as well. The number of pixels increased from 440 (6 August) to 512 (8 August), and the average value increased from 4.6 × 10−4 to 6.4 × 10−4 s−1. From 8 to 10 August, the number of pixels slightly decreased to 506; however, the average value increased to 8.6 × 10−4 s−1. These numbers indicate that the tail significantly intensified from 6 to 10 August. On 12 August, the number of pixels decreased to 471, and the average value decreased to 3.8 × 10−4 s−1 since the convergence zone was weakening. A new ITCZ appeared at the coast of Central America on 8 August (Fig. 1b) and extended westward gradually within a few days (Fig. 1c) as the vortices and the tail moved westward and poleward. After the vortices broke off from the tail, the tail connected with the new ITCZ and formed a longer structure (Fig. 1d) that later dissipated.

There are some differences between model simulations and observations. In dry model simulations, the ITCZ is forced by a prescribed heating. Since the prescribed heating is rather smooth in spatial dimensions and the large-scale environment is simple, the disturbances are mostly equal-sized and have similar intensity. However, each breakdown event appears differently in the real atmosphere. Some cases resemble the dry simulation; some cases are very different. The most obvious difference is the long tail that is not observed or addressed previously in dry model simulations. However, this type of ITCZ breakdown occurs quite frequently in the eastern Pacific during summer (Wang and Magnusdottir 2006). Wang and Magnusdottir cataloged that this type of breakdown was due to westward propagating disturbances (WPDs) because easterly waves can be traced prior to the breakdown events. With the disturbances coming from the east, it seems reasonable to have part of the ITCZ remain undisturbed at the west of the newly formed vortex. However, the coming direction of the WPDs cannot fully explain the existence of the tail or its intensification. In this particular case, the tail may be seen as the remaining part of the ITCZ. However, the continuous intensification and extension of the tail implies that some moist processes may be responsible for the formation of the tail.

The tail is hardly observed in another type of ITCZ breakdown, which is self-induced by the barotropic instability of the flow, the vortex roll-up cases (Wang and Magnusdottir 2006) in the real atmosphere. This type of ITCZ breakdown often occurs in locations close to the central Pacific, while the WPD cases occur in locations close to Central America. From observations, such as in weekly or monthly precipitation, the value is lower in the central Pacific than in the eastern Pacific during summertime in the Northern Hemisphere. This suggests that the large-scale environment close to the central Pacific is less favorable for the development of deep convection compared to the eastern Pacific. One of the possible mechanisms that may suppress the formation of the tail is eastward propagating waves, such as Kelvin waves or the Madden–Julian oscillation (MJO).

4. The role of moisture in the ITCZ breakdown and reformation

a. Dry atmosphere

The dry experiment simulates the warming effect from latent heat release in deep convection without considering influences from moisture. Figure 2 shows that the result of this experiment is qualitatively the same as those using the shallow-water model (Nieto Ferreira and Schubert 1997) and the dry PE model (Wang and Magnusdottir 2005). These three models have fundamental differences in the vertical dimension. In the shallow-water model, it is the lower troposphere that is simulated, whereas in the dry PE model the vertical profile of the prescribed heating is designed to roughly follow that of the apparent heat Qc. In QTCM1, the vertical dimension is prescribed and can be reconstructed according to the quasi-equilibrium balance when analyzing the result. Because the phenomenon of interest is mostly deep convection with a clear dipole structure in the vertical direction (e.g., Fig. 8 in Wang and Magnusdottir 2005), its vertical dimension can be reasonably approximated by QTCM1. By comparing the results from these models, it is evident that a highly truncated model in the vertical direction is still valid for the phenomenon of interest.

Another characteristic of QTCM1 is that small perturbations tend to dissipate quickly because of a stronger diffusion by of the equations. This may result in a slightly stronger heating rate (6.5 K day−1 at 600 hPa) than in previous studies (5 K day−1) that is needed to initiate and support the circulation throughout the simulation. Nevertheless, the heating rate only affects the evolution speed of the breakdown process, not the dynamics.

The negative relative vorticity on the equatorward side may be an artificial response due to the number of vertical modes in the model. A similar structure is observed in the composite ITCZ that is reconstructed by limited wavenumbers and frequencies (Magnusdottir and Wang 2008). Apparently, this artificial response influences neither the dynamical processes nor the overall results. In the real atmosphere, negative absolute vorticity in the Northern Hemisphere will dissipate immediately because of inertial instability.

b. Effects of moisture

The moisture-on experiment is a transition step to the experiment with full physical processes. Figure 3 shows the moisture-on ITCZ evolution in which the radiation, sensible heat flux, and surface momentum fluxes are off. The variables shown are relative vorticity (1 × 10−5 s−1), wind at 850 hPa on day 3 (Fig. 3a) and day 7 (Fig. 3b), and precipitation (mm day−1) on the same days (Figs. 3c and 3d). The major difference after adding moisture to the atmosphere is the appearance of the tail (an elongated area of positive relative vorticity) at the southwest of the leftmost vortex. The same feature can be seen in the precipitation field (Fig. 3d), surface fluxes, and cloudiness (not shown).

The intensity of the ITCZ (in terms of vorticity) is increased due to latent heat release from the moisture feedback after the circulation is initiated. This feedback speeds up the evolution of the ITCZ breakdown; however, that same acceleration can be achieved by providing stronger heating in a dry simulation. Therefore, the breakdown processes are not directly related to moist processes. Another effect of moisture is the continuous latent heat supply that maintains the deep convection. Thus, either a smaller heating rate (about 3.6 K day−1) or a shorter heating period is sufficient to produce an ITCZ breakdown and the corresponding tail.

The location of the precipitation band in Fig. 3d is not exactly the same as that of the vorticity band because in the model the precipitation pattern is in phase with the large-scale convergence (Yu and Neelin 1997) and it is not necessary to match the vorticity field. However, a reasonable relationship between the vorticity field and thermodynamical field can be expected. The ITCZ location defined by dynamical fields may be slightly more poleward than that defined by thermodynamical fields.

A comparison of the dry experiment (Fig. 2) and the moisture-on experiment (Fig. 3) reveals that moisture is the dominant factor in formation of the tail. Other factors such as radiation, sensible heat flux, and surface momentum fluxes (the all-physics-on experiment, Fig. 4) merely modify the shape and intensity of the tail.

In the all-physics-on experiment (Fig. 4), since the total effect of radiative fluxes is cooling and the surface momentum fluxes decrease wind speed in the free atmosphere through surface friction, the intensity of the ITCZ is in between that of the moisture-on and the dry experiment. However, the tail appears to be stronger and extends farther than in the moist-on experiment. This is because, in QTCM1, the surface wind speed is enhanced by the effect of momentum fluxes that transfer energy from flows in the free atmosphere to the surface wind in the boundary layer. In this experiment, the strength of the vortices is weaker than in the moisture-on experiment because of radiative cooling and momentum fluxes. Thus, these weaker vortices are less axisymmetric and are connected by fine filaments in the relative vorticity field. This representation is actually often observed in real events, which makes the point of breakdown unclear and difficult to define.

The tail may extend more than 60° longitude and last for more than two weeks regardless of the strength and dimension of the initial heating. Once the vortex dissipates or breaks off from the tail, the tail can be seen as a new and weaker convergence zone. In our simulations, this new convergence zone is too weak to go through another ITCZ breakdown cycle, probably because the QTCM1 does not include shallow convection that is nonprecipitating and can accumulate energy in the atmosphere. Therefore, the convergence zone is dissipated by the large-scale stability gradually if there is no further forcing. In addition, the SST pattern is fairly simple in this study, leading to a very weak trade wind (about 0.65 m s−1) in the background.

Without further breakdown, the lifetime of the tail is mainly controlled by large-scale wave propagation instead of energy consumption in the atmospheric column. Convection in the tail is completely suppressed by the passage of the equatorial Kelvin wave (zonal wavenumber 1), induced by the initial heating, and propagates eastward around the globe.

Figure 5 shows the evaporation (grayscale), surface wind speed (contours), and wind vectors at the surface on (a) day 7 and (b) day 16 of the all-physics-on experiment. The surface easterly phase of the Kelvin wave can be seen around +90° to −150° (across the 180° meridian) in all three fields on day 7 when the tail is most intensified and extensive. On day 16, the surface wind speed around the tail has been compensated significantly by the surface easterly phase of the Kelvin wave. The corresponding surface flux and the rainband (not shown) are shortened as well. The influence of the Kelvin wave will be further discussed in the discussion section.

c. Budget analysis

The second set of experiments (Table 2) with a short heating area (20° longitude in the zonal direction, indicated by a black dashed line in Fig. 6d) is performed for budget analysis. Figure 6 shows the spatial pattern of relative vorticity, wind fields at 850 hPa (Figs. 6a–c), and precipitation (Figs. 6d–f) of experiment1. A box from the equator to 10°N and from −60° to −10° relative longitude (indicated by a black dashed line in Fig. 6e) that covers most of the tail is chosen for budget analysis. The exact location of the box does not alter the general result. This single-vortex experiment also shows that the generation of the tail is only related to the appearance of a local disturbance. An ITCZ breakdown happens to be one of the processes that provide tropical disturbances.

Figure 7 shows the temporal evolution of anomalous (i.e., experiment 1 − ctrl) energy fluxes into the atmospheric column. All of the fluxes and processes are shown in terms of heating rate (K day−1) in Figs. 7a–c. Generally speaking, moisture in the column atmosphere is provided by evaporation, evap, and the moisture stratification term Mq · v1 and is removed mostly by precipitation 〈Qq〉. A very small amount of moisture is balanced by the horizontal component of the moisture convergence −〈v · q〉. Another interpretation of the moisture stratification term Mq · v1 is that it represents the vertical component of the moisture convergence (vertical velocity × vertical moisture gradient), meaning that the moisture is redistributed by vertical motion. In other words, in deep convective areas the low-level inflow brings in moist air (compared to the air on upper levels), the vertical motion transports moisture upward, and the upper-level outflow transports out drier air. In this process, the mass flux is conserved; however, the moisture convergence is positive because the upper-level dry air is replaced by the lower-level moist air, leading to an increase in precipitation.

The temperature tendency is determined by convective heating 〈Qc〉 and the dry stability term −Ms · v1 (Fig. 7b) as well as the horizontal component of the temperature convergence −〈v · T〉, the sensible heat flux FTs, and radiative fluxes Frad (Fig. 7c). Note that the dry stability term is plotted in the opposite sign to better compare with convective heating 〈Qc〉. The latter three components are an order of magnitude smaller than the first two components, and these three processes are less important, as expected. By day 5 the temperature is also influenced by the prescribed heating that is advected by the horizontal motion as shown by the −〈v · T〉 term; however, the amount is small. In the end, the most dominant terms are the convective heating 〈Qc〉 and the dry stability terms −Ms · v1.

Similar to the moisture equation, the dry stability term represents the vertical component of the temperature convergence. In deep convective areas the vertical motion replaces warmer air on the upper levels with cooler air from the lower levels, which results in a cooling effect (i.e., a decrease in energy) in the column that is also the adiabatic cooling from ascending motion.

The net radiative forcing is cooling, as expected. Note that the radiation flux Frad shows a positive anomaly in the first two weeks. This is because the values shown here are anomalies from its control run. The variation of the radiative flux is dominated by the longwave radiation associated with deep convective clouds, whereas the shortwave radiation remains almost unchanged (not shown).

Both the spatial and temporal patterns of surface evaporation are closely related to surface wind speed that is initially induced by the vortex. Figure 7d shows the temporal evolution of surface wind speed (solid line, left y axis) and the mode-1 divergence · v1—in other words, the vertical motion (dashed line, right y axis). The evolution of surface evaporation, evap, and sensible heat flux, FTs, is highly correlated with surface wind speed. Figure 5 also shows that the spatial pattern of surface wind speed matches that of surface evaporation. Variables associated with vertical motions, such as convective heating, cloudiness (which influences Frad), dry stability, and gross moisture stratification, are well correlated with the mode-1 divergence. In addition, Fig. 7d shows that the surface wind speed is highly correlated with vertical motion in deep convective areas. After the convection dies out, these two variables are not necessarily correlated.

Figures 8a–d show the time evolution of each flux in the moist static energy equation (3) for experiment 1 to experiment 4, respectively. For a better visualization of the results, the fluxes with minor importance (Frad, FTs, and −〈v · T〉) are summed together. When the temperature and moisture equations are added together, precipitation −〈Qq〉 and convective heating −〈Qc〉 cancel each other out. The moist static energy term M · v1 (dashed line) represents the combined effects of temperature and moisture on energy changes in the atmospheric column due to vertical motion. As mentioned before, the dry static energy term has a cooling effect, while the moisture stratification term has a warming effect. It is found that the cooling effect is always slightly larger than the warming effect in the tropics (Yu et al. 1998), which means that the atmosphere is generally stable from a large-scale point of view. The upward motion that decreases and redistributes energy in the atmospheric column can roughly balance the energy source (Fnet in this study).

For experiment 1 (Fig. 8a), the secondary cooling process is the horizontal moisture convergence −〈v · q〉 (dotted line). Although this process accounts for about one-third of the total cooling, it is of less importance than the vertical convergence. Moreover, a stronger vertical motion can also result in a similar magnitude of cooling; thus, its role can be replaced by stronger vertical motions.

In experiment 2 (Fig. 8b), the horizontal moisture convergence −〈v · q〉 is suppressed. It is evident that the vertical motion is enhanced slightly (not shown) because the column atmosphere gets warmer. The warming further enhances the upward motion that increases the vertical moisture and temperature convergence. In other words, more moisture is transported upward, and the adiabatic cooling is enhanced to balance the warming (Fig. 8b, dashed line). Precipitation and convective heating become stronger as well (not shown). The residual energy of the energy tendency term in experiment 2 is therefore larger than that of eperiment 1, leading to a stronger tail in terms of precipitation. However, the lifetime of the tail is still controlled by the Kelvin wave propagation.

Experiment 3 further investigates the role of surface evaporation. When the major energy source, evap, is prescribed by a small constant, the strength of the vortex is much weaker, and the surface wind speed within the box decreases remarkably, leading to a weaker surface sensible heat flux (not shown). The most important influence is that the convection within the tail fails to develop. The magnitude of both surface wind speed and vertical motion is much smaller and all of the corresponding fluxes decrease significantly (Fig. 8c). The surface wind speed and vertical motions are also poorly correlated, as expected (not shown). Overall, a very weak cooling from the vertical motion is balanced by the surface sensible heat flux.

The importance of other processes such as cloud radiation, horizontal temperature convergence, and surface sensible heat flux are also examined, and the results show that these bear no significant influence on the development of the tail (not shown).

In the warmer SST experiment (experiment 4, Fig. 8d), the atmospheric column is warmer due to more surface evaporation. Therefore, stronger cooling is expected to compensate the warming. Interestingly, the horizontal moisture convergence −〈v · q〉 accounts for stronger cooling than the moist stability term −M · v1. In the warmer situation, both the horizontal moisture gradient and surface wind speed increase (not shown). Therefore, the effect of cooling (drying) due to lower-level inflow is intensified. The horizontal moisture convergence −〈v · q〉 is then suppressed in a warmer SST situation (experiment 5, Fig. 8e) to investigate the importance of this process. In experiment 5, a similar enhancement of vertical motions that leads to more vertical moisture transport and stronger adiabatic cooling, as in experiment 2, balances the warming. This experiment further shows that the convection is still dominated by surface evaporation and vertical energy convergence.

When the QTCM1 is coupled with a mixed layer ocean (experiment 6, Fig. 8f), the result is qualitatively the same as that of experiment 1, except that the lifetime of the tail is shortened by a few days. This difference is caused by a smaller energy supply from the ocean when the SST is cooled by surface fluxes. The decrease in its lifetime is not strong enough and the tail still exists much longer than in observations. Nevertheless, the oceans have a negative feedback to development of the ITCZ.

The suppression of the equatorial Kelvin wave is shown in experiment 7 (Fig. 9). Thin lines represent the surface wind speed (solid, left y axis) and vertical motion (dashed, right y axis) from experiment 1, and thick lines are from experiment 7 for the same variables. The tail in experiment 7 clearly lasts longer than that in experiment 1.

5. Discussion

a. The interaction between moisture and dynamics

The major difference between a dry simulation and a moisture-on simulation of the first set of experiments (Table 1) is the appearance of the tail on the southwest side of the vortex. The low-level convergent flow induced by the vortex initiates weak vertical motions. The consequent response of this circulation includes surface fluxes (more precisely, the evaporation) that provide energy to the atmosphere and the upward motion that redistributes and releases the energy. This process further enhances circulation and forms a positive wind–evaporation feedback. Therefore, a long tail develops on the southwest side of the vortex where the surface convergent flow is strongest. The surface convergent flow is a typical response to forcing in a resting state (e.g., Gill 1980). As shown in the budget analysis (Table 2), the energy that supports the development of the tail comes from latent heat release, while the moisture is provided by surface evaporation. Therefore, the tail does not exist in a dry simulation. Some additional experiments with stronger heating (10 K day−1 or higher) in a dry condition were performed and the tail development was not seen. In fact, both the low-level convergent flow and moisture are necessary conditions for the development of the tail. An increase in the SST will change the energy balance between different processes (i.e., the absolute value of horizontal moisture convergence becomes larger than the vertical energy convergence in experiment 4, Fig. 8d). However, the importance of evaporation and vertical energy convergence remains the same.

Another difference between the dry and moist atmosphere is the moisture feedback from latent heat release. Given the same initial forcing in the atmosphere, the response of a moist atmosphere is stronger and the convection lasts longer than that of a dry atmosphere. Because of this moisture feedback, a much weaker heating (2 K day−1 maximum at 600 hPa, not shown) is sufficient to generate a tail with the same spatial extent and a slightly weaker intensity than the result of experiment 1. Therefore, the role of tropical disturbances is twofold. A disturbance, such as an easterly wave, may perturb the ITCZ and trigger a breakdown if the ITCZ already exists; however, it may also induce a long tail soon after initiation of the low-level convergent flow if no ITCZ presents. In a sense, this can be one of the processes that organize an ITCZ. In the situation that an easterly wave approaches an existing ITCZ (as Fig. 1 shows), the tail not only enhances the strength of the ITCZ but also extends the length of it. The formation and development of the tail suggest a possible mechanism for a much more efficient ITCZ reformation, while it generally takes more than a week to form an ITCZ with an exaggerated SST forcing (in QTCM1). The existence of the tail also modifies the conditions of the atmosphere and provides a more unstable environment for convection to occur.

b. Comparison with observations

Although this is an idealized study, the result is well supported by a field measurement conducted in the eastern Pacific ITCZ. In the 2001 Eastern Pacific Investigations of Climate (EPIC 2001) (Raymond et al. 2003), it is found that the two largest terms contributing to the entropy budget in a selected control volume near the surface are surface fluxes and the volume-top downdraft flux, meaning that the energy supplied by surface fluxes is balanced by the cooling effect from vertical motion. We also show that the amount of horizontal moisture convergence is minor, consistent with observations. The radiative cooling on a synoptic time scale is minor as well.

In the real atmosphere, it is known that the propagation of tropical waves, such as Kelvin waves and the MJO, may suppress or enhance convection by modifying the surface wind pattern. In our study, the suppression from a Kelvin wave is clearly shown in the spatial pattern of surface wind speed (Fig. 5) and the decrease in energy fluxes, shown in Fig. 7, around days 15–20. Experiment 7 (Fig. 9) further confirms that, when the Kelvin wave is filtered, the tail lasts longer until it gradually dissipates owing to large-scale stability and model damping. Although there is no MJO in our simulations, it is known that the MJO signal propagates to the central Pacific, and then only the upper-level signals remain to go into the eastern Pacific. This may explain why in observations the tail is often observed in the eastern Pacific and does not appear in all of the ITCZ breakdown events.

Additionally, the experiment coupled with a mixed layer ocean shows that, when SST drops because of surface flux exchange, the life span of the tail is shortened by a few days. However, the role of the ocean, such as ocean dynamics, needs further in-depth study.

6. Conclusions

The effects of moisture on synoptic-time-scale ITCZ dynamics over the eastern Pacific are investigated using an intermediate atmospheric circulation model. In this model, the vertical structures in deep convective regions are constrained by the quasi-equilibrium closure. The dry simulation shows results consistent with those using simple dynamical models except that a slightly stronger heating rate is needed. The results show that the ITCZ breakdown is dominated by dry dynamical processes.

In the moist simulations, the moisture feedback provides latent heat release to warm the atmosphere, and the most intriguing result is the formation of the tail at the southwest of a vortex during and after ITCZ breakdown. In the eastern North Pacific, the tail is observed in summer and fall. Our study shows that the low-level convergent flow induced by a disturbance, either generated within the ITCZ through barotropic instability or propagated from the east, may result in the development of a long tail in a moist atmosphere.

A positive wind–evaporation feedback is necessary to enhance the initial circulation within the tail. The surface wind speed induces evaporation that provides moisture to the atmosphere, and the upward motion redistributes the moisture vertically. The low-level convergent flow and vertical motions induced by the disturbance then lifts the moist air to the condensation level and releases latent heat to warm the atmosphere. This warming is nicely balanced by adiabatic cooling of the ascending motion.

The horizontal moisture flux plays a secondary role in balancing the evaporation. In a warmer situation, its contribution increases significantly; however, the process of horizontal moisture convergence is still less important than the vertical convergence. By suppressing the horizontal flux, the atmospheric column gets warmer and enhances the vertical motion that further intensifies the vertical moisture transport. Thus, the effect of horizontal processes can be easily compensated by altering vertical motions, but the opposite situation is not possible.

Although the intensity of the tail is weak, it persists for more than a week with or without the vortex breaking off. The tail may be seen as a new ITCZ and the formation process may explain the quick ITCZ reformation in the eastern Pacific. In our study, the new ITCZ does not go through another ITCZ breakdown cycle, possibly owing to model design; and this requires further investigation. Other factors that influence the lifetime of the tail are the passage of tropical waves and changes of SST through a surface energy exchange. It has been suggested that eastward propagating large-scale waves in the tropics can enhance or suppress local convection and modulate the location of tropical cyclones by modifying the surface wind pattern. In our study, the suppression from an equatorial Kelvin wave is clearly seen in the temporal and spatial patterns. The influence of a change in SST appears to be small in our study. However, the ocean model here is rather simple and does not consider many important features that we may find in the eastern Pacific. A more detailed study focusing on the role of the ocean is ongoing.

Acknowledgments

We thank two anonymous reviewers and the editor for comments that improved the manuscript. This work was supported by National Science Council, Taiwan.

REFERENCES

  • Betts, A. K., , and M. J. Miller, 1986: A new convective adjustment scheme. Part II: Single column tests using GATE wave, BOMEX, ATEX, and arctic air-mass data sets. Quart. J. Roy. Meteor. Soc., 112 , 693709.

    • Search Google Scholar
    • Export Citation
  • Betts, A. K., , and M. J. Miller, 1993: The Betts–Miller scheme. The Representation of Cumulus Convection in Numerical Models of the Atmosphere, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 107–121.

    • Search Google Scholar
    • Export Citation
  • Chao, W. C., 2000: Multiple quasi equilibria of the ITCZ and the origin of monsoon onset. J. Atmos. Sci., 57 , 641652.

  • Chao, W. C., , and B. Chen, 2004: Single and double ITCZ in an aqua-planet model with constant sea surface temperature and solar angle. Climate Dyn., 22 , 447459.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , and J. D. Neelin, 1996: Linearization of a longwave radiation scheme for intermediate tropical atmospheric models. J. Geophys. Res., 101 , 1512915145.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , and J. D. Neelin, 1999: Cirrus detrainment–temperature feedback. Geophys. Res. Lett., 26 , 12951298.

  • Chou, C., , and J. D. Neelin, 2004: Mechanisms of global warming impacts on regional tropical precipitation. J. Climate, 17 , 26882701.

  • Chou, C., , J. D. Neelin, , and H. Su, 2001: Ocean–atmosphere–land feedbacks in an idealized monsoon. Quart. J. Roy. Meteor. Soc., 127 , 18691892.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106 , 447462.

  • Hack, J. J., , W. H. Schubert, , D. E. Stevens, , and H-C. Kuo, 1989: Response of the Hadley circulation to convective forcing in the ITCZ. J. Atmos. Sci., 46 , 29572973.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., , M. E. McIntyre, , and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Joly, A., , and A. J. Thorpe, 1990: Frontal instability generated by tropospheric potential vorticity anomalies. Quart. J. Roy. Meteor. Soc., 116 , 525560.

    • Search Google Scholar
    • Export Citation
  • Kirtman, B. P., , and E. K. Schneider, 2000: A spontaneously generated tropical atmospheric general circulation. J. Atmos. Sci., 57 , 20802093.

    • Search Google Scholar
    • Export Citation
  • Kuo, H. L., 1973: Dynamics of quasi-geostrophic flows and instability theory. Adv. Appl. Mech., 13 , 247330.

  • Lin, X., , and R. H. Johnson, 1996: Kinematic and thermodynamic characteristics of the flow over the western Pacific warm pool during TOGA COARE. J. Atmos. Sci., 53 , 695715.

    • Search Google Scholar
    • Export Citation
  • Magnusdottir, G., , and C-C. Wang, 2008: Intertropical convergence zones during the active season in daily data. J. Atmos. Sci., 65 , 24252436.

    • Search Google Scholar
    • Export Citation
  • Mechoso, C., and Coauthors, 1995: The seasonal cycle over the tropical Pacific in coupled ocean–atmosphere general circulation models. Mon. Wea. Rev., 123 , 28252838.

    • 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 results. J. Atmos. Sci., 51 , 18761894.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , and N. Zeng, 2000: A quasi-equilibrium tropical circulation model—Formulation. J. Atmos. Sci., 57 , 17411766.

  • Nieto Ferreira, R., , and W. H. Schubert, 1997: Barotropic aspects of ITCZ breakdown. J. Atmos. Sci., 54 , 261285.

  • Raymond, D. J., , G. B. Raga, , C. S. Bretherton, , J. Molinari, , C. López-Carrillo, , and Z. Fuchs, 2003: Convective forcing in the intertropical convergence zone of the eastern Pacific. J. Atmos. Sci., 60 , 20642082.

    • Search Google Scholar
    • Export Citation
  • Schubert, W. H., , P. E. Ciesielski, , D. E. Stevens, , and H-C. Kuo, 1991: Potential vorticity modeling of the ITCZ and the Hadley circulation. J. Atmos. Sci., 48 , 14931509.

    • Search Google Scholar
    • Export Citation
  • Sumi, A., 1990: Pattern formation of convective activity over the aqua-planet with globally uniform sea surface temperature. Meteorological Research Rep. 90-1, Division of Meteorology, Geophysical Institute, University of Tokyo, 124 pp.

    • Search Google Scholar
    • Export Citation
  • Wang, C-C., , and G. Magnusdottir, 2005: ITCZ breakdown in three-dimensional flows. J. Atmos. Sci., 62 , 14971512.

  • Wang, C-C., , and G. Magnusdottir, 2006: The ITCZ in the central and eastern Pacific on synoptic time scales. Mon. Wea. Rev., 134 , 14051421.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., , V. O. Magaña, , T. N. Palmer, , J. Shukla, , R. A. Tomas, , M. Yanai, , and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103 , 1445114510.

    • 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
  • Yu, J-Y., , and J. D. Neelin, 1997: Analytic approximations for moist convectively adjusted regions. J. Atmos. Sci., 54 , 10541063.

  • Yu, J-Y., , C. Chou, , and J. D. Neelin, 1998: Estimating the gross moist stability of the tropical atmosphere. J. Atmos. Sci., 55 , 13541372.

    • Search Google Scholar
    • Export Citation
  • Zehnder, J. A., , and D. M. Powell, 1999: The interaction of easterly waves, orography, and the intertropical convergence zone in the genesis of eastern Pacific tropical cyclones. Mon. Wea. Rev., 127 , 15661585.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., , J. D. Neelin, , and C. Chou, 2000: A quasi-equilibrium tropical circulation model–implementation and simulation. J. Atmos. Sci., 57 , 17671796.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

An ITCZ breakdown case triggered by easterly waves: the anomalous wind (vectors) and surface relative vorticity (shading) on (a) 6, (b) 8, (c) 10, and (d) 12 August 2000 derived from QuikSCAT wind. Units for vorticity are 1 × 10−4 s−1. The average wind is obtained from all available QuikSCAT data in June–August.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 2.
Fig. 2.

The dry experiment of the ITCZ breakdown on day (a) 5, (b) 7, (c) 9, and (d) 11. Contours (negative dashed) indicate relative vorticity at 850 hPa in units of 1 × 10−5 s−1. Vectors indicate the wind field at 850 hPa. The x and y axes are relative longitude and latitude, respectively. The area of prescribed heating is indicated by the thick dashed line.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 3.
Fig. 3.

The moisture-on experiment of ITCZ breakdown on (a),(c) day 3 and (b),(d) day 7, showing relative vorticity (contours, negative dashed) and wind (vectors) at 850 hPa (left) and precipitation (right). The prescribed heating is indicated by the thick dashed line in (a) and (c).

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 4.
Fig. 4.

As in Fig. 3 but for the all-physics-on experiment of ITCZ breakdown.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 5.
Fig. 5.

The evaporation (grayscale), surface wind speed (contours), and the wind vectors of the all-physics-on experiment of ITCZ breakdown on (a) day 7 and (b) day 16. The zero wind speed line is thickened. The contour interval is 5 m s−1 and the unit of evaporation is W m−2.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 6.
Fig. 6.

The development of the tail (expt 1). Experimental design is the same as the all-physics-on experiment (Fig. 4) but for a shorter heating area, indicated by a dashed line in (d). The contour field is relative vorticity; vectors indicate the wind field at 850 hPa. The box in (e) indicates the area for budget analysis.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 7.
Fig. 7.

The time evolution of each term in the temperature and moisture equations of (expt 1 − ctrl). The x axis is model day. The y axis in (a)–(c) is the heating rate (K day−1). The surface wind speed (solid, left y axis, m s−1) and the first baroclinic divergence (dashed, right y axis, 1 × 10−5 s−1) are shown in (d).

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 8.
Fig. 8.

Time evolution of the energy budget for (a) expt 1, (b) expt 2 (suppress horizontal moisture convergence), (c) expt 3 (suppress surface evaporation), (d) expt 4 (warmer SST), (e) expt 5 (warmer SST, suppress horizontal moisture convergence), and (f) expt 6 (mixed layer ocean). The x and y axes are the model day and the heating rate (K day−1), respectively.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Fig. 9.
Fig. 9.

The surface wind speed (m s−1) (solid lines, left y axis) and mode-1 divergence (1 × 10−5 s−1) (dashed lines, right y axis). Thick lines indicate the experiment in which the Kelvin wave is filtered; thin lines are from expt 1. Values shown here are averaged over the box of budget analysis.

Citation: Journal of the Atmospheric Sciences 67, 4; 10.1175/2009JAS3164.1

Table 1.

The first set of experiments with the setting of the physical processes in QTCM1. The length of the prescribed heating area is 120° longitude, and the width is 10° latitude centered at 10°N.

Table 1.
Table 2.

Experimental designs for budget analysis: the process that is suppressed and the SST pattern in the experiment. There is no prescribed heating in all control runs (also see a detailed explanation in the text). The length of the prescribed heating area is 20° in longitude.

Table 2.
Save