Explicit Simulations of the Intertropical Convergence Zone

458 VOLUME 61J O U R N A L O F T H E A T M O S P H E R I C S C I E N C E S

q 2004 American Meteorological Society

Explicit Simulations of the Intertropical Convergence Zone

CHANGHAI LIU AND MITCHELL W. MONCRIEFF

National Center for Atmospheric Research,* Boulder, Colorado

(Manuscript received 8 April 2002, in final form 8 September 2003)

ABSTRACT

The intertropical convergence zone (ITCZ) is one of the most important components of the global circulation.In order to understand the dynamical processes that regulate its formation, latitudinal preference, and structure,explicit two-dimensional numerical modeling of convection on an equatorial beta plane was conducted with anonhydrostatic cloud-system-resolving model. The model was forced by energy fluxes associated with constantsea surface temperature (SST) and by horizontally homogeneous radiative cooling.

Two distinct patterns were identified for the spatial distribution of convective activity in the Tropics. The firstwas characteristic of enhanced off-equator convection, namely, a double ITCZ-like morphology (one more salientthan the other) straddling the equator during the early period of the integration. The second featured enhancedequatorial convection, namely, a single ITCZ-like morphology on the equator during the later quasi-equilibriumperiod. Diagnostic analysis and two additional experiments, one excluding surface friction and the other havingtime- and space-independent surface fluxes, revealed that the wind-induced surface flux variability played anessential role in the development and maintenance of the equatorial maximum convection. Surface friction waslargely responsible for the early asymmetric convective distribution with respect to the equator in the controlsimulation and acted to flatten the convective peaks.

One important discrepancy from observations concerned the too-weak trade wind convergence around en-hanced convective regions. This unrealistic feature suggested that, as well as the meridional dynamics, latitudinalSST gradients, large-scale forcing, and other physical processes regulate the observed ITCZs.

1. Introduction

The intertropical convergence zone (ITCZ) refers tothe narrow and approximately east–west-oriented beltof concentrated vigorous cumulonimbus convection inthe Tropics and represents the ascending branch of themeridional Hadley circulation. Observations show thatthe ITCZ typically resides between the latitudes 48 to128 away from the equator over warm oceanic regions(e.g., Waliser and Somerville 1994). However, the pre-cise latitudinal location varies greatly with season andlongitude. Moreover, a pair of ITCZs straddling theequator also occurs (e.g., Lietzke et al. 2001; Zhang2001; Halpern and Hung 2001; Liu and Xie 2002), aswell as the common situation where a single ITCZ islocated at or away from the equator.

The physical mechanisms regulating the formation andlatitudinal preference of the ITCZ have been a subjectof numerous observational, theoretical and numericalmodeling investigations. The earliest attempts tried to

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

Corresponding author address: Dr. Changhai Liu, National Centerfor Atmospheric Research, P. O. Box 3000, Boulder, CO 80307-3000.E-mail: [email protected]

relate the spatial distributions of sea surface temperature(SST) to the spatial structure of tropical convection, mo-tivated by the observed high correlation between con-vective enhancement and warm SST forcing (Bjerkneset al. 1969; Graham and Barnett 1987). This hypothesisis supported by a number of numerical simulations (e.g.,Pike 1971; Manabe et al. 1974; Goswami et al. 1984).

Although it has a strong influence on observed trop-ical convection and ITCZs, SST forcing alone cannotexplain all observed features, and considerable variationexists in the relationship of convection to the underlyingSST distribution. For instance, many observational stud-ies showed that the highest SST is often not collocatedwith the ITCZ (e.g., Ramage 1974; Sadler 1975; Has-tenrath and Lamb 1977; Lietzke et al. 2001). In addition,general circulation modeling (e.g., Hayashi and Sumi1986; Hess et al. 1993; Waliser and Somerville 1994)showed that double ITCZs develop straddling the equa-tor even if the SST maximum is at the equator. Fur-thermore, a well-defined ITCZ can still occur in nu-merical simulations with globally uniform SST (e.g.,Sumi 1992; Chao 2000; Kirtman and Schneider 2000),implying that inhomogeneous SST forcing might not benecessary. From these observational facts and modelingstudies, it is concluded that dynamical processes likelyplay an important role in regulating the observed ITCZs.

Charney (1971) put forward an explanation for the

Unauthenticated | Downloaded 12/14/21 06:49 AM UTC

15 FEBRUARY 2004 459L I U A N D M O N C R I E F F

ITCZ formation in terms of two competing processes,namely Ekman pumping and moisture availability. Theformer is proportional to the Coriolis parameter and thusincreases poleward, whereas the latter generally increas-es equatorward. Therefore, their combined effect resultsin the moisture supply peaked at a finite distance fromthe equator, accounting for an off-equator ITCZ.

Several hypotheses for the off-equator ITCZ wereproposed based upon low-level convergences. Holton etal. (1971) argued that the ITCZ is favored at the latitudeswhere the frequency of zonally propagating disturbanc-es equals the Coriolis frequency, and maximum bound-ary layer convergence occurs. This theory was furtheredby Chang (1973) who showed that the mechanism isindeed responsible for the development of the ITCZ ina numerical model. In contrast, observational studiesfound that the ITCZ does not owe its existence to zon-ally propagating synoptic-scale disturbances, in thesense that it would still exist in their absence (Gu andZhang 2002). Lindzen (1974) found that there exists alongitude-independent wave–CISK (conditional insta-bility of the second kind) mode with a period of about5 days. This oscillation, when coupled with the travelingwaves of the same period generated by the inhomoge-neous surface conditions, leads to the maximum low-level convergence several degrees away from the equa-tor. Based upon linear shallow-water theory, Waliser andSomerville (1994) also demonstrated that the largestlocal low-level convergence is produced when a heatsource is located at a finite distance from the equator.

Recent studies showed that the character of tropicalconvection is affected by the cross-equatorial pressuregradients and radiative–convective instability. Tomasand Webster (1997) noted that, in regions of a substantialcross equator surface pressure gradient, a local anticy-clonic circulation exists on the low pressure side of theequator, and the flow meets the criterion for inertialinstability. It was hypothesized that the atmospheric re-sponse to the absolute vorticity distribution results in alow-level convergence–divergence doublet, which isimportant in determining the strength and location ofconvection. On the other hand, Raymond (2000) ex-amined the radiative role of convective clouds in thedynamics of the Hadley circulation. The modeling re-sults demonstrated that cloud–radiation interaction isstrong enough to drive a circulation that is comparablein intensity to the observed global mean Hadley cir-culation.

Studies of the ITCZ are further complicated by thesensitivity of the modeled ITCZ location and structurein GCMs to convective parameterizations. For example,in aquaplanet modeling with globally uniform SST,Chao (2000) and Chao and Chen (2001) found that adouble ITCZ can evolve into a single one by simplyswitching the relaxed Arakawa–Schubert parameteri-zation (Moorthi and Suarez 1992) to the moist convec-tive adjustment scheme (Manabe et al. 1965). The highsensitivity of the ITCZ behavior to convective param-

eterization was also reported in aquaplanet models withan equatorial SST maximum (Hess et al. 1993; Nu-maguti 1993).

The uncertainty resulting from convective parame-terization is a serious disadvantage in using GCMs toinvestigate the physical processes governing the ob-served ITCZs. In this study, we adopt an explicit ap-proach—cloud-system-resolving modeling (CSRM)—that avoids the use of any convective parameterization.Presently, computer capability limits our simulations totwo spatial dimensions; namely, a latitude–height crosssection across the equator on an equatorial beta plane.We have two major objectives. First, we examine pat-terns of convective activity generated solely under theinfluence of the earth’s rotation and, in particular, wheth-er concentrated tropical convection (i.e., the ITCZ-likefeatures) resides either at or off the equator. Second, weexamine the mechanisms responsible for the simulatedITCZ-like concentrated tropical convection.

The numerical model and design of the numerical ex-periment are described in the next section. In section 3,the simulation results are detailed. Section 4 is devotedto the physical interpretation of the simulated convectivefeatures through diagnostic analyses and two sensitivityexperiments. The paper concludes in section 5.

2. Numerical model and experimental design

We use the two-dimensional Eulerian version of thenonhydrostatic Eulerian/semi-Lagrangian anelasticmodel (Smolarkiewicz and Margolin 1997). The 16 000km 3 24 km (north–south oriented) computational do-main represents an equatorial beta plane bounded withrigid walls and centered on the equator. The horizontaland vertical grid spacing are 5 km and 0.3 km, respec-tively. Note that the 5-km grid length is a practical com-promise made to enable longtime, large-domain simu-lations. It allows the mesoscale organization of con-vection to be treated explicitly, but not individual con-vective cores. This grid size has been used in previousstudies of tropical convection (e.g., Held et al. 1993).Rigid and free-slip vertical boundary conditions are em-ployed at the top and bottom of the domain. At thelateral boundaries, the potential temperature and watervapor mixing ratio equal their respective domain-av-eraged values at every time step. Otherwise, undesiredlarge temperature and water vapor gradients could occurnear the lateral boundaries when the modeled atmo-sphere departs sufficiently from the initial state. A 500-km-wide and a 6-km-deep absorbing layer at the lateraland top vertical boundaries, respectively, damp propa-gating gravity waves that could otherwise be reflectedunrealistically.

The bulk cloud microphysical parameterization usesa two-category warm rain scheme (Grabowski and Smo-larkiewicz 1996) and a two-category ice scheme (Gra-bowski 1999). The domain lies over an ocean having aconstant SST (302.5 K). The surface moisture and sen-



sible heat fluxes are calculated using a simplified versionof the Tropical Ocean Global Atmosphere CoupledOcean–Atmosphere Response Experiment (TOGACOARE) surface flux algorithm (Fairall et al. 1996).Rather than using cloud-interactive radiation, a time-independent and horizontally uniform radiative cooling(a constant value of 21.5 K day21 below 12 km de-creasing linearly to zero at the model top) is specified.This simplification facilitates the interpretation of thesimulation results and exploration of the fundamentalphysics of ITCZs.

The simulation starts from a resting atmosphere. Theinitial temperature and moisture field are based uponthe averaged condition over the Intensive Flux Array(IFA) during the 19–26 December 1992 period ofTOGA COARE. Small random perturbations of the po-tential temperature and water vapor mixing ratio fieldsinitiate convection.

3. Description of simulation

Our 100-day simulation enables the simulated at-mosphere to attain a state of statistical quasi-equilibri-um. Because the trade wind is not prescribed, the mod-eled ITCZ-like circulation must result solely from theatmospheric response to tropical convection. As indi-cated in the Hovmoller diagram of surface precipitationrate (Fig. 1), the convective pattern in the Tropics duringthe early period of integration is dramatically differentfrom the late quasi-equilibrium stage, corresponding tooff-equator and equatorial ITCZ-like features, respec-tively. These two distinct phases are now described.

a. Off-equator maximum convection stage

The off-equator ITCZ-like pattern roughly corre-sponds to the first 40 days of integration. Characteris-tically, the convective activity displays a dual maximumstraddling the equator, but is not symmetric across theequator: one is more salient than the other. The space–time surface precipitation distribution in Fig. 1 illus-trates the latitude-dependent behavior of convective ac-tivity. First, convection at high latitudes is frequent, butrandomly distributed and short-lived. In contrast, trop-ical convection is comparatively strong, persistent, or-ganized and aggregated. Second, vigorous convectionin the Tropics is preferentially off-equator (at least sev-eral hundred kilometers away from the equator): it rarelyoccurs near the equator. Third, convection is asymmetricabout the equator and preferentially develops in theNorthern Hemisphere. Finally, once vigorous convec-tion occurs on one side of the equator, convective ac-tivity is commonly suppressed on the other side. An-other noticeable feature is the large clear region sur-rounding the tropical convective systems.

The precipitation rate distribution averaged over thefirst 40 days (Fig. 2a) shows that the enhanced con-vective region is concentrated at approximately 1200

km from the equator in the Northern Hemisphere, andthe secondary peak is located roughly 500 km from theequator in the Southern Hemisphere. In contrast, pre-cipitation in the neighborhood of the equator is sup-pressed significantly, consistent with the temporal andspatial rainfall distribution. Outside the tropical area,the precipitation intensity does not show marked spatialdependence, except in the outermost 500-km-widesponge layers, where convection is suppressed com-pletely. Figure 3a further demonstrates the meridionalconvective variability in terms of the spatial distributionof cloud fraction. In calculating the cloud amount, 100%cloudiness is assumed over a grid box when the totalcondensate (i.e., the sum of cloud water, rain water, andcloud ice) exceeds 0.01 g kg21. The fractional cloudi-ness clearly displays an off-equator maximum with avalue exceeding 25% about 1500 km from the equatorin the Northern Hemisphere, a secondary extremum inthe Southern Hemisphere, and an equatorial minimum.This spatial distribution is well-correlated with the pre-cipitation variability in Fig. 2a.

Figure 4 shows the perturbation fields averaged overa period of 10 days (from days 16 to 25) based on thehourly archived dataset. The fields are smoothed witha 500-km running mean filter to eliminate the smaller-scale noise and highlight the organized features. Thetropical meridional circulation in Fig. 4a displays astriking asymmetry with respect to the equator. In theNorthern Hemisphere, equatorward flow prevails at up-per levels and poleward flow at lower levels, leading toupper-level convergence and low-level divergence at theequator. The opposite circulation pattern occurs at thepoleward flank of the concentrated convection about1200 km from the equator. The flow structure supportsa deep off-equator ascent and a wide equatorial descent(see the shading) and is consistent with weak equatorialconvection and active off-equator convection. A similar,albeit much weaker, meridional circulation occurs in theSouthern Hemisphere and is evidently related to theweak convective maximum there. It should be pointedout that although a Hadley-like circulation develops, themeridional flow at both low and upper levels is ratherweak and localized, in contrast with some previous mod-eling results (e.g., Waliser and Somerville 1994; Ray-mond 2000) using cumulus parameterizations. However,a strict comparison is meaningless because our modelhas neither latitudinal SST variation nor cloud–radiationinteraction, unlike the GCMs. Another noticeable fea-ture is the equatorial southerly flow at 7–11 km andnortherly flow at 4–6 km. The existence of midleveldivergence around the equator illustrates that the sub-siding warm air partially moves away from the equatorbefore reaching the lower troposphere. The similar dou-ble-celled meridional circulation behavior has been doc-umented in the zonally symmetric tropical atmospherefrom observations (e.g., Johnson et al. 1999; Mapes2001). Although instantaneous perturbations are sub-



FIG. 1. Space–time distributions of surface precipitation rate during (a) days 1–25, (b) days 26–50, (c) days 51–75, and (d) days 76–100.The light and dark shading correspond to rainfall intensity greater than 1 and 10 mm h21, respectively. The equator is located at the centerof the domain.



FIG. 2. Spatial distributions of surface precipitation rate averagedover (a) the early 40-day integration and (b) the late 60-day integra-tion. The field is smoothed with a 500-km running mean filter.

FIG. 3. Spatial distributions of cloud amount averaged over (a) the early 40-day integration and (b) thelate 60-day integration. The light, moderate dark, and heavy dark shading correspond to cloud fractiongreater than 5%, 15%, and 25%, respectively.

stantial, the mean meridional circulation in high lati-tudes is weak, irregular and incoherent.

The most striking feature in Fig. 4b is the deep equa-torial easterly wind prevalent throughout the tropo-sphere, peaking just below the tropopause. Character-istically, outside the Tropics the wind perturbation iseasterly at low levels and mostly westerly at upper lev-els. The two subtropical westerly jets in the upper tro-posphere are intimately associated with the effects of

the Coriolis torque on the poleward flow of the merid-ional circulation. This is consistent with the conceptualexplanation that the westerly jets originate from the an-gular momentum transports of Hadley cells (e.g., Heldand Hou 1980; Lindzen and Hou 1988). In comparisonwith the counterparts in the real atmosphere; however,the westerly jets are at the wrong latitudes and muchtoo weak because the observed subtropical jets are in-fluenced strongly by both stationary and transient eddyflux convergences, which are absent in our two-dimen-sional model setup.

The temperature field in Fig. 4c is dominated by wide-spread cooling relative to the initial state. The warmestair resides over the equator, resulting from adiabaticsubsidence, and the coldest occurs about 6500 km fromthe equator in each hemisphere, maintaining a weakequator-to-pole temperature gradient. (The reversedtemperature gradient near the boundaries is artificial,the result of the lateral boundary conditions and damp-ing zones.) Figure 4d features weak moist perturbationsin the planetary boundary layer and strong dry pertur-bations in the free troposphere. The equatorial regionand the adjacent tropical atmosphere undergo prominentdrying due to compensating subsidence, which is a re-sponse to off-equator vigorous convection.

The pressure distribution in Fig. 4e hydrostaticallycorresponds to the temperature perturbation and exhibitsa relatively low pressure in the lower troposphere anda relatively high pressure aloft in the Tropics. The lowestsurface pressure occurs in the concentrated convection



FIG. 4. Physical fields averaged from days 16 to 25. (a) Meridionalwind (1 m s21 contour interval), (b) zonal wind (2 m s21 contourinterval), (c) temperature perturbation (1-K contour interval), (d) wa-ter vapor mixing ratio perturbation (0.5 g kg21 contour interval), and(e) pressure perturbation (25-mb contour interval). The white anddark shadings in (a) and (b) correspond to vertical velocity less than22.5 3 1023 m s21 and greater than 2.5 3 1023 m s21, respectively.All fields are smoothed with a 500-km running mean filter.

FIG. 5. Evolution of the space-averaged precipitation rate (solidline) and CAPE (dashed line) over a 1500-km-wide area centered atthe equator during (a) days 50–75 and (b) days 75–100.

region, whereas the surface highs are positioned about6500 km away from the equator in accordance with therespective warmest and coldest locations. The pressuregradient is mainly pole-to-equator at lower layers, re-versing at upper levels.

b. Equatorial maximum convection stage

After the simulation achieves a quasi-equilibriumstate, the aforementioned off-equator ITCZ-like patterntransforms into a single equatorial ITCZ-like morphol-ogy as evinced in the space–time precipitation distri-bution (Fig. 1). Long-lasting convection, or convectionaggregated on mesoscales, is concentrated in a narrowarea around the equator, whereas flanking active con-

vection is scarce. Convective activity is not continuousnear the equator. Instead, groups of convective clustersare intermittent with a regular period of about 2 days.Approximately, each episode persists for 1 day, fol-lowed by a 1-day suppression. The periodic enhance-ment and suppression of convective activity are a resultof the destruction and recovery of convective availablepotential energy (CAPE). This is clearly demonstratedby the out-of-phase correlation of precipitation andCAPE spatially averaged over a 1500-km-wide regionaround the equator in Fig. 5. Once deep convectionbreaks out, the atmosphere is stabilized via latent heat-ing in the middle and upper troposphere and downdraft-induced evaporative cooling in the boundary layer: con-vection cannot last forever over a limited area. Duringa suppressed episode, the persistent moisture and sen-sible heat transports from the underlying warm ocean,as well as the imposed radiative cooling and horizontalmoisture convergences in the lower troposphere, restoreCAPE gradually. On average, the CAPE peaks approx-imately 15 h earlier than the precipitation.

Figure 2b presents the spatial distribution of the sur-face precipitation rate averaged over the last 60 days(from days 40 to 100). Enhanced convection extendsapproximately 1000 km across the equator, and the sur-rounding convective activity is substantially reduced inthe tropical region: a precise reversal of the earlier con-vective activity. The strong equatorial convection is alsoreflected in cloudiness, which has a salient maximumof over 35% at the equator, and a minimum of less than10% in the nearby latitudes (Fig. 3b).

Figure 6 illustrates the atmospheric structure duringthe quasi-equilibrium stage. The tropical meridionalwind in Fig. 6a displays a wavelike vertical structureof wavelength half of the tropospheric depth and is op-posite in sign in the two hemispheres. The attendantconvergence–divergence and ascent–descent distribu-tion are exactly opposite to those in the off-equator ac-tive convection phase. They are consistent with en-hanced equatorial convective activity that draws inwarm and moist boundary layer air from both sides ofthe equator, transports it upward, and detrains it near



FIG. 6. Same as in Fig. 4, but for the averages from days 81–90.

the tropopause. We speculate that the weak divergencearound the 5-km level is due to the shallower convection(cumulus congestus) with tops near the melting level(Johnson et al. 1999; Liu et al. 2001a,b), while the over-lying weak convergence is attributable to the adjacentdescending warm and dry air that could not reach thelower troposphere and diverges in the middle tropo-sphere. Alternatively, this two-cell vertical structure wasinterpreted as a consequence of the vertically variedstatic stability in the tropical troposphere by Mapes(2001). In high latitudes, the mean meridional wind isweak and does not possess coherent large-scale circu-lations as in the off-equator convective maximum.Again, the meridional wind is weak and localized, sug-gesting that a moderate SST gradient and/or other phys-ical processes, such as interactive radiation and large-scale forcing, are necessary to generate realistic tradewinds and Hadley circulations.

The zonal wind in Fig. 6b is much stronger than themeridional component. Interestingly, the equatorialeasterly wind and flanking westerlies prevail regardless

of whether enhanced convection occurs at or off theequator. Unlike the earlier off-equator maximum con-vection episode, however, the equatorial easterly is uni-formly distributed in the vertical due to convective mix-ing and does not posses a jetlike structure. The highlatitudes are dominated by easterlies in the lower tro-posphere except in the vicinity of the lateral boundaries,similar to the enhanced off-equator convection stage.

The temperature perturbation in Fig. 6c is character-ized by a warming relative to the initial atmosphere inthe lower tropical troposphere and cooling elsewhere.The prominent two warm centers are located at about1000 km from the equator, obviously a result of thesubsidence warming surrounding the equatorial con-vection. The cold center is situated at about 1500 kmfrom the lateral boundary, an artifact of the lateralboundary conditions as indicated earlier.

The water vapor perturbation in Fig. 6d is character-istic of shallow moistening in the lowest 1 km, under-lying a deep drying layer in high latitudes, mostly com-parable to the off-equator convective maximum periodin Fig. 4d. Within the tropical area, upward convectivetransport of moist boundary layer air leads to a relativelymoist region around the equator, whereas nearby com-pensating downward motion causes significant off-equa-tor drying. This pattern is distinct from the early partof the simulation.

The pressure perturbation in Fig. 6e exhibits oneequatorial trough and two midlatitude ridges at approx-imately 6500 km from the equator near the surface. Thepressure gradient is mostly directed toward the equatorin the lower troposphere. The upper-level pressure gra-dient is relatively weak and variable.

c. Formation of the equatorial easterlies

An equatorial easterly flow occurs irrespective of thelatitudinal position of the convective peak (see Figs. 4band 6b). Figure 7a shows the evolution of this tropo-spheric easterly wind in terms of the zonal-average overa 1000-km domain centered at the equator. Several char-acteristics are salient. First, the development of theequatorial easterly wind starts in the upper troposphere,gradually expands downward and reaches the surface atabout 25 days into the simulation. Second, an easterlyjet is maintained in the upper troposphere during thefirst 40 days. Third, vertically uniform easterlies prevailand there is little shear throughout the troposphere dur-ing the enhanced equatorial convection.

In order to quantify the physical processes in the de-velopment of the equatorial easterlies, the zonal mo-mentum budget is investigated. In a two-dimensional(latitude–height) anelastic framework, the zonal mo-mentum equation is given as

]u 1 ]ruy 1 ]ruw5 2 2 1 f y 1 D , (1)u]t r ]y r ]z



FIG. 7. Evolution of the equatorial easterly wind averaged over a 1000-km-wide area centered at equator for (a) control simulation and(b) simulation without surface friction. The shading scale is shown to the right (in m s21).

where the notations are standard, and the overbar rep-resents the spatial average.

Figure 8 displays the zonal wind tendency due tohorizontal momentum flux convergence (21/r)(] /]y),ruyvertical momentum flux convergence (21/r)(] /]z),ruwand Coriolis torque , from days 10 to 60. In the off-fyequator concentrated convection stage, the horizontalconvergence term (Fig. 8a) is the major source for theupper-level easterlies. However, the upper-level easterlymomentum originates from the equatorward meridionalflow outside the budget region through the Coriolis forceas inferred from Fig. 4a. The vertical convergence term(Fig. 8b) accounts for the easterly wind generation inthe middle and lower troposphere. Evidently, the grad-ual downward extension of the easterly momentumsource during the early 25 days is consistent with thepersistent downward expansion of easterly flow asevinced in Fig. 7a. After the maximum convection isdisplaced to the equator (roughly 40 days into the in-tegration), the two branches of equatorward meridionalflow associated with the double-celled vertical circu-lation in Fig. 6a produce easterly momentum by theCoriolis torque. The easterly momentum subsequentlyconverges near the equator and gives rise to two easterly

momentum sources located near the surface and at 7–10 km (Fig. 8a), respectively. The equatorial convectiontransports the easterly momentum upward, maintainingthe easterly flow in the 2–6-km layer and above 11 km(Fig. 8b).

Although the Coriolis force is essential to the easterlymomentum generation outside the budget domain (notshown), its contribution within the budget domain is ofsecondary importance because of the small Coriolis pa-rameter (Fig. 8c). As expected, the Coriolis torque pro-duces easterly momentum in the equatorward meridi-onal flow and westerly momentum in the poleward flow.The subgrid diffusion u is weak except in the planetaryDboundary layer where it acts to reduce the easterly windspeed (not shown).

The foregoing analysis suggests the following phys-ical basis for the formation and sustenance of the equa-torial easterlies. During the early stage, active off-equa-tor convection transports low-level mass upward anddeposits it in the upper troposphere. The detrained massmoves toward the equator from both hemispheres, asevinced by the meridional wind in Fig. 4a, and generatesan equatorial easterly wind at upper levels due to theturning action of the Coriolis force. The meridional flow



FIG. 8. Evolution of the equatorial zonal wind tendency by (a)horizontal momentum flux convergences, (b) vertical momentum fluxconvergences, and (c) Coriolis torques averaged over a 1000-km-wide area centered at the equator during days 10–60. Contour intervalis 1 m s21 day21.

converges over the equator, causing equatorial subsi-dence. This slow descent transports the easterly mo-mentum and gradually extends the easterly wind intolower layers and eventually down to the surface. Duringthe quasi-equilibrium stage, convective activity is con-centrated at the equator, and the meridional circulationis reversed. In contrast to the early period, easterly mo-mentum is generated in the lower troposphere and alsowithin an elevated convergence layer, converges towardthe equator and is subsequently transported verticallyby attendant convection. Because of the strong, fast, andefficient convective transport, the well-mixed easterlywind is striking during the equatorial ITCZ episode. Itshould be pointed out that the above explanation maybe only applied to our two-dimensional framework. Inthe real world, the transient meridional circulation as-sociated with the seasonal cycle is crucial for main-taining the climatological easterly wind in the equatorialupper troposphere (Lee 1999).

4. Physical interpretation

As reviewed in the introduction, quite a few hypoth-eses have been proposed for the observed off-equatorITCZs. However, most arguments do not apply to oursimulated off-equator maximum convection in a two-dimensional equatorial beta plane with uniform SST.Moreover, none of the previous theoretical argumentsare relevant to the single equatorial convective peak

occurring in the quasi-equilibrium stage. In the follow-ing, thermodynamical and dynamical arguments are pre-sented for the equatorial and the off-equator ITCZ mor-phologies.

a. Surface friction

Surface friction influences ITCZs and attendant con-vection in various ways. First, the friction-induced Ek-man pumping, which is a function of the Coriolis pa-rameter, preferentially supports high-latitude convection(Charney 1971). Second, according to Chao and Chen(2001), surface friction increases the surface energy in-take of the boundary layer air converging toward con-vective areas because of a longer inward-spiraling paththan in the frictionless situation, which also favors high-er-latitude convection. Finally, surface friction decreas-es the wind speed near the ground, affecting the wind-induced surface flux variability, which is a mechanismresponsible for the simulated strong equatorial convec-tion (see later discussion).

To quantify the potential role of surface friction inthe simulated convective behavior, we performed a 100-day sensitivity experiment, which is identical to the con-trol run except that the surface friction is excluded. Asexpected, absence of the friction enhances the windspeed near the surface. Increased moisture and sensibleheat transports from the underlying warm ocean lead toan overall stronger convective activity compared to thecontrol experiment. Nevertheless, the integration againhas two distinct phases, namely off-equator and equa-torial concentrated convection in terms of the Hovmollerdiagram of surface precipitation rate (not shown). Theinstantaneous convective behavior and the spatial de-pendence are barely affected although the off-equatorconcentrated convection only lasts for about 1 month,roughly 10 days shorter than in the control simulation.

The close similarity between the two experiments isfurther demonstrated by comparing the time-averagedspatial distributions of precipitation rate in Fig. 9 againstthose in Fig. 2. The convective peak is more pronouncedin the frictionless simulation. Another apparent featureis that the early double convective peaks straddling theequator, although less persistent, are comparable in in-tensity unlike their counterparts in the control simula-tion with one much stronger than the other.

When surface friction is absent, the low-level easterlymomentum in the converging inflow toward the equa-torial region is amplified. This intensified momentumgeneration results in a stronger tropical easterly flow asindicated in Fig. 7b. The intensified surface wind leadsto large surface energy fluxes and thus favors convectivedevelopment in the equatorial region. This is consistentwith the aforementioned swifter transition from off-equator to equatorial regime compared to the controlsimulation. Note that the equatorial east wind tends togradually intensify with time and does not realize asteady state during the 100 days of integration.



FIG. 9. Spatial distributions of surface precipitation rate averagedover (a) the early 30 days and (b) the late 70 days in the experimentexcluding surface friction. The field is smoothed with a 500-km run-ning mean filter.

b. Wind-induced surface flux variability

In addition to the specified radiative cooling, the sur-face heat and moisture fluxes are the only physical pro-cesses that can generate CAPE. To quantify the corre-lation between the surface flux and convective activity,Fig. 10 presents the space–time distributions of the sur-face sensible and latent heat fluxes. There is a significantcorrespondence between convective activity and surfacefluxes. During the off-equator convection, the moistureand sensible heat fluxes frequently peak at the same off-equator locations of enhanced precipitation, whereas theminimum fluxes near the equator are consistent with theequatorial convective minimum. During the active equa-torial convection, marked maximum fluxes are mostlypositioned at the equator, an exact reversal of the earlyepisode.

Because of the assumed uniform SST, the spatial var-iations in surface fluxes can only result from inhomo-geneous wind speed, air temperature, and water vapormixing ratio at the surface. The surface air temperatureand water vapor (not shown) both display weak equator-to-pole gradients outside the tropical location through-out the integration. For the most part, the equator isslightly warmer than the adjacent environment duringthe early period, while the opposite occurs during theequatorial-maximum convection. The relatively warmenvironment surrounding the peak convection is a directconsequence of the adiabatic heating associated with thecompensative subsidence. In comparison with the sub-tropical region, the equatorial atmosphere is relativelydry in the early stage and moist in the quasi-equilibriumstage. Accordingly, the time–space distribution of sur-face water vapor cannot explain both the minimumequatorial moisture flux during the off-equator concen-trated convection and the maximum equatorial moisturefluxes during the equatorial concentrated convection.

The preceding discussion reveals that the water vaporand temperature distributions contradict the surface fluxdistributions. In other words, an inhomogeneous surfacewind may be important in regulating the energy trans-ports from the underlying warm ocean, as demonstratedin space–time distributions of surface wind speed (Fig.11) and evolution of the spatially averaged surface fluxand wind speed over a 1500-km equatorial area (Fig.12). Evidently, a positive correlation exists between sur-face wind and heat flux intensity; in general, large windspeeds are accompanied with large surface fluxes. Aswift transition is noticeable around 25 days into thesimulation, as well as the variations on short time scalesassociated with convective modulations. Before theeasterly wind gets firmly established near the surface,the wind speed maximum is positioned at more than1000 km from the equator in the Northern Hemisphere,correspondent with the surface flux maximum in theconcentrated convective area, whereas the weak equa-torial wind is well correlated with the surface flux min-imum there. Quantitatively, the two individual windcomponents (not shown) are comparable except in theTropics, and both exhibit spatial patterns similar to thetotal wind speed. The meridional wind accounts for mostof the total wind maximum, but the weak zonal windnear the equator causes the significant minimum in thetotal wind speed. For the most part, we speculate thatthe weak equatorial zonal wind is due to the small Cor-iolis parameter and accounts for the weakened equa-torial surface fluxes, which is undesirable for equatorialconvection and indirectly beneficial to the enhancementof off-equator convection.

During the period of equatorial concentrated convec-tion, the wind speed peaks at the equator and decreasesrapidly toward high latitudes. Apparently, the enlargedsurface fluxes around the equator occur after the east-erlies reach the surface. On average, the peak windspeed is roughly 150% as large as the value outside theTropics. It is the strong equatorial wind speed that isresponsible for the surface flux maximum therein. Thisimplies that the development of the enhanced equatorialconvection results from wind-enhanced equatorial sur-face fluxes.

c. Uniform surface fluxes

The foregoing analyses suggest that wind-inducedsurface flux variability is a viable mechanism for thedevelopment of the simulated ITCZ-like features. Dur-ing the early period of simulation, the weak Coriolisacceleration generates the weak zonal flow around theequator. This leads to an equatorial surface flux mini-mum which, in turn, subdues the equatorial convection.As demonstrated in Fig. 7a, an equatorial easterly windgets gradually established in the upper troposphere andspreads downward during the off-equator ITCZ period.When the equatorial easterly reaches the surface, theequatorial surface flux is increased and eventually ex-



FIG. 10. Space–time distributions of surface fluxes during (a) days 1–25, (b) days 26–50, (c) days 51–75, and (d) days 76–100. The light,moderately dark, and heavy dark shadings correspond to values greater than 100, 140, and 180 W m 22, respectively.



FIG. 11. Space–time distributions of surface wind speed during (a) days 1–25, (b) days 26–50, (c) days 51–75, and (d) days 76–100. Theshading scale is shown at the upper-right corner (in m s21).



FIG. 12. Evolution of the space-averaged surface fluxes (solid line)and surface wind speed (dashed line) over a 1500-km-wide area cen-tered at the equator.

ceeds values at other locations. The attendant increasein CAPE leads to the concentrated convective activityat the equator.

In order to explore whether the zonal-wind-inducedreduction or enhancement in the equatorial surface flux-es is essential, we conducted one more 100-day sensi-tivity experiment, which is the same as the control sim-ulation except that spatially uniform and time-invariantsurface fluxes are specified. This setup excludes the ef-fects of surface wind speed and thermodynamic vari-ability on the fluxes and thus completely eliminates thephysical process related to the wind-induced surface fluxvariations. Specifically, it eliminates the wind-inducedsea–air heat exchange (WISHE) mechanism (Emanuel1987; Neelin et al. 1987). The constant sensible andlatent heat fluxes are set equal to 12 and 120 W m22,respectively, which approximates the space- and time-averaged quantities in the previous simulation.

The space–time distribution of surface precipitationrate in Fig. 13 and the temporal mean distribution inFig. 14 display persistent, symmetric off-equator en-hanced convective activity, a pattern opposite to theconcentrated equatorial convection in the preceding twosimulations. This double-ITCZ-like pattern is main-tained throughout the integration. Tropical convectionaggregates and alternates between the two hemisphereswith a period of about 2 days. Convection usually beginsnear the poleward side of the equatorial waveguide andnew convective systems periodically develop a finitedistance ahead of old ones on the equatorward side. Theenvelope of the loosely coupled convective systemsmoves toward the equator and persists for 1 to 2 days.

The persistence of suppressed equatorial convectionillustrates that the wind-induced surface flux enhance-ment plays a critical role in the formation of the singleequatorial convective maximum during the quasi-equi-librium stage of the control and friction-free simula-tions. In addition, the wind-induced surface flux vari-ation is responsible for the weaker off-equator meanrainfall during the early period (not shown) than in thecontrol experiment (Fig. 2a) and the sensitivity exper-iment without surface friction (Fig. 9a).

5. Concluding discussion

The ITCZ is a key feature of the tropical large-scalecirculation. The physical processes governing its main-

tenance, latitudinal position and structure are still poorlyunderstood. Understanding the responsible mechanismsis further complicated by the strong dependence ofITCZs in GCMs on cumulus parameterizations. Also,GCMs contain complex interactions that are difficult tounambiguously quantify. To simplify the physics andovercome this vexing uncertainty, we adopted an ex-plicit approach using a numerical model with 5-km hor-izontal resolution. Although not fine enough to resolveindividual convective cells, this resolution reasonablycaptures the mesoscale organization of convection. Tofacilitate interpretation, the computational domain isover an ocean of constant and uniform SST, and hori-zontally uniform radiative cooling is imposed.

Two distinct convective patterns in the Tropics areobtained during the 100-day integration. The first pat-tern is characteristic of a pair of enhanced convectivebands displaced from the equator during the early stage(one more pronounced than the other), and is compa-rable to the observed off-equator ITCZs. The secondpattern, which corresponds to enhanced equatorial con-vection, prevails during the later stage of the integration,resembling a single ITCZ at the equator. However, theattendant meridional circulation is weak and spatiallynot as extensive as seen in observations. This impliesthat large-scale forcing, latitudinal SST gradient andadditional physics are essential for more realistic ITCZs,at least in a two-dimensional framework.

Several physical processes are explored. First, the im-pact of surface friction is quantified via a friction-freesensitivity experiment. The comparison with the controlsimulation illustrates that the two contrasting phases areindependent of surface friction. The main frictional ef-fect is to weaken the convective activity and destroy thesymmetric structure in the off-equator active convectionstage. Second, the spatial variability in surface fluxesdue to local wind speed enhancement/reduction is ex-amined. During the early part of the 100-day integration,the small Coriolis parameter near the equator results inthe weak zonal wind and thus contributes to the smallequatorial surface fluxes. This wind-induced surfaceflux minimum suppresses equatorial convection and,therefore, benefits off-equator convection indirectly.When the equatorial easterly wind reaches the surface,the surface fluxes around the equator are substantiallyenhanced due to the zonal wind intensification. Thiswind-induced surface flux enhancement supports theequatorial convective activity, eventually leading to thedevelopment of a single ITCZ-like feature at the equator.The role played by the wind-induced surface flux var-iations in controlling the tropical convection is con-firmed by an additional experiment in which the surfacefluxes are uniformly prescribed. In particular, the off-equator concentration of convective activity is a robustfeature in this sensitivity experiment, suggesting thatthe wind-induced surface flux enhancement is crucialto the formation of an equatorial ITCZ-like feature.

The off-equator ITCZ-like feature occurs in all of the



FIG. 13. Same as in Fig. 1, but for the experiment with uniform surface fluxes.



FIG. 14. Spatial distributions of surface precipitation rate averagedover the 100-day simulation with uniform surface fluxes. The fieldis smoothed with a 500-km running mean filter.

three simulations, but none of the existing hypothesesprovide an adequate interpretation. Recent work by Liuand Moncrieff (2002) suggests that the simulated off-equator peak convection results from two competingdynamical mechanisms associated with the latitudinalvariation of the Coriolis parameter. The first is rotation-induced trapping and accumulation of compensatingsubsidence warming and drying in response to convec-tive heating: this mechanism favors an equatorial ITCZ.The second concerns the rotation-induced strengtheningof the low-level convergence in the heating area, whichprefers an ITCZ displaced from the equator. These twoconflicting physical processes compromise to displacethe maximum convection a finite distance from the equa-tor. Although this theoretical argument is attractive, fur-ther studies are needed to substantiate its relevance tothe simulated off-equator ITCZ-like pattern herein.

Because of our highly idealized experimental setup,it is not surprising that the model results differ from oreven contradict some observed facts. For example, aconvective minimum near the equator is often observedin the western Pacific warm pool where quite uniformSST exists, distinct to the equatorial convective maxi-mum in the quasi-equilibrium stage. Another noticeablediscrepancy from observations concerns the magnitudeof shallow warm clouds; the model warm clouds accountfor roughly 14% of total rainfall, less than the 20%contribution documented from the Tropical RainfallMeasuring Mission (TRMM) satellite (Nesbitt et al.2000).

The two-dimensional (i.e., zonal symmetric) assump-tion excludes zonally propagating waves and three-di-mensional circulations that are potentially important inregulating the observed ITCZs. It also excludes three-dimensional mesoscale convective systems and quasi-two-dimensional, shear-parallel convective systems. Inaddition, two-dimensional simulations could differ sig-nificantly from zonally averaged three-dimensional sim-ulation results, and how this would affect the results ofthis study is unknown. These limitations are likely re-sponsible for the aforementioned unrealistic features inthe simulations. Nevertheless, a two-dimensionalCSRM identifies fundamental physics of the observedITCZ behavior and provides a basis for three-dimen-

sional explicit studies, which will be feasible in the nearfuture.

Acknowledgments. We would like to thank the anon-ymous reviewers and editor G. Kiladis for their con-structive comments. The National Center for Atmo-spheric Research is sponsored by the National ScienceFoundation. This work is partially supported by NASATRMM Grant NAG5-7742.

REFERENCES

Bjerknes, J., L. J. Allison, E. R. Kreins, F. A. Godshall, and G.Warnecke, 1969: Satellite mapping of the Pacific tropical cloud-iness. Bull. Amer. Meteor. Soc., 50, 313–322.

Chang, C. P., 1973: A dynamical model of the intertropical conver-gence zone. J. Atmos. Sci., 30, 190–212.

Chao, W. C., 2000: Multiple quasi-equilibria of the ITCZ and theorigin of monsoon onset. J. Atmos. Sci., 57, 641–651.

——, and B. Chen, 2001: Multiple quasi-equilibria of the ITCZ andthe origin of monsoon onset. Part II: Rotational ITCZ attractors.J. Atmos. Sci., 58, 2820–2831.

Charney, J. G., 1971: Tropical cyclogenesis and the formation of theintertropical convergence zone. Mathematical Problems of Geo-physical Fluid Dynamics, W. H. Reid, Ed., Lectures in AppliedMathematics, Vol. 13, American Mathematical Society, 355–368.

Emanuel, K. A., 1987: An air–sea interaction model of intraseasonaloscillations in the Tropics. J. Atmos. Sci., 44, 2324–2340.

Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S.Young, 1996: Bulk parameterization of air–sea fluxes for Trop-ical Ocean–Global Atmosphere Coupled-Ocean Atmosphere Re-sponse Experiment. J. Geophys. Res., 101, 3747–3764.

Goswami, B. N., J. Shukla, E. K. Schneider, and Y. C. Sud, 1984:Study of the dynamics of the intertropical convergence zone witha symmetric version of the GLAS climate model. J. Atmos. Sci.,41, 5–19.

Grabowski, W. W., 1999: A parameterization of cloud microphysicsfor long-term cloud-resolving modeling of tropical convection.Atmos. Res., 52, 17–41.

——, and P. K. Smolarkiewicz, 1996: Two-time-level semi-Lagrang-ian modeling of precipitating clouds. Mon. Wea. Rev., 124, 487–497.

Graham, N. E., and T. P. Barnett, 1987: Sea surface temperature,surface wind divergence, and convection over tropical oceans.Science, 238, 657–659.

Gu, G.-J., and C.-D. Zhang, 2002: Westward-propagating synoptic-scale disturbances and the ITCZ. J. Atmos. Sci., 59, 1062–1075.

Halpern, D., and C.-W. Hung, 2001: Satellite observations of thesoutheast Pacific intertropical convergence zone during 1993–1998. J. Geophys. Res., 106, 28 107–28 112.

Hastenrath, S., and P. J. Lamb, 1977: Some aspects of the circulationand climate over the eastern Atlantic. Mon. Wea. Rev., 105,1019–1023.

Hayashi, Y.-Y., and A. Sumi, 1986: The 30–40 day oscillation sim-ulated in an ‘‘aqua-planet’’ model. J. Meteor. Soc. Japan, 64,451–466.

Held, I. M., and A. Y. Hou, 1980: Nonlinear axially symmetric cir-culations in a nearly inviscid atmosphere. J. Atmos. Sci., 37,515–533.

——, R. S. Hemler, and V. Ramaswamy, 1993: Radiative–convectiveequilibrium with explicit two-dimensional moist convection. J.Atmos. Sci., 50, 3909–3927.

Hess, P. G., D. S. Battisti, and P. J. Rasch, 1993: The maintenanceof the intertropical convergence zones and the large-scale cir-culation on a water-covered earth. J. Atmos. Sci., 50, 691–713.

Holton, J. R., J. M. Wallace, and J. A. Young, 1971: On boundarylayer dynamics and the ITCZ. J. Atmos. Sci., 28, 275–280.



Johnson, R. H., T. M. Rickenbach, S. A. Rutledge, P. E. Ciesielski,and W. H. Schubert, 1999: Trimodal characteristics of tropicalconvection. J. Climate, 12, 2397–2418.

Kirtman, B. P., and E. K. Schneider, 2000: A spontaneously generatedtropical atmospheric general circulation. J. Atmos. Sci., 57,2080–2093.

Lee, S., 1999: Why are the climatological zonal winds easterly inthe equatorial upper troposphere? J. Atmos. Sci., 56, 1353–1363.

Lietzke, C. E., C. Deser, and T. H. Vonder Haar, 2001: Evolutionarystructure of the eastern Pacific double ITCZ based on satellitemoisture profile retrievals. J. Climate, 14, 743–751.

Lindzen, R. S., 1974: Wave-CISK in the Tropics. J. Atmos. Sci., 31,156–179.

——, and A. Y. Hou, 1988: Hadley circulations for zonally averagedheating centered off the equator. J. Atmos. Sci., 45, 2416–2427.

Liu, C., and M. W. Moncrieff, 2002: An explicit numerical study ofthe intertropical convergence zone in an aqua-planet with uni-form sea surface temperature. Preprints, 25th Conf. on Hurri-canes and Tropical Meteorology, San Diego, CA, Amer. Meteor.Soc., 313–314.

——, ——, and W. W. Grabowski, 2001a: Explicit and parameterizedrealizations of convective cloud systems in TOGA COARE.Mon. Wea. Rev., 129, 1689–1703.

——, ——, and ——, 2001b: Hierarchical modeling of tropical con-vective systems with explicit and parameterized approaches.Quart. J. Roy. Meteor. Soc., 127, 493–515.

Liu, W. T., and X. Xie, 2002: Double intertropical convergencezones—A new look using scatterometer. Geophys. Res. Lett., 29,2072, doi:10.1029/2002GL015431.

Manabe, S., J. Smagorinsky, and R. F. Strickler, 1965: Simulatedclimatology of a general circulation model with a hydrologiccycle. Mon. Wea. Rev., 93, 769–798.

——, D. G. Hahn, and J. L. Holloway Jr., 1974: The seasonal variationof the tropical circulation as simulated by a global model of theatmosphere. J. Atmos. Sci., 31, 43–83.

Mapes, B. E., 2001: What’s two height scales: The moist adiabat and

the radiative troposphere. Quart. J. Roy. Meteor. Soc., 127,2353–2366.

Moorthi, S., and M. J. Suarez, 1992: Relaxed Arakawa–Schubert: Aparameterization of moist convection for general circulationmodels. Mon. Wea. Rev., 120, 978–1002.

Neelin, J. D., I. M. Held, and K. H. Cook, 1987: Evaporation-windfeedback and low-frequency variability in the tropical atmo-sphere. J. Atmos. Sci., 44, 2341–2348.

Nesbitt, S. W., E. J. Zipser, and D. J. Cecil, 2000: A census of pre-cipitation features in the Tropics using TRMM: Radar, ice scat-tering, and lightning observations. J. Climate, 13, 4087–4106.

Numaguti, A., 1993: Dynamics and energy balance of the Hadleycirculation and the tropical precipitation zones: Significance ofthe distribution of evaporation. J. Atmos. Sci., 50, 1874–1887.

Pike, A. C., 1971: Intertropical convergence zone studied with aninteracting atmosphere and ocean model. Mon. Wea. Rev., 99,469–477.

Ramage, C. S., 1974: Structure of an oceanic near-equatorial troughdeduced from research aircraft traverses. Mon. Wea. Rev., 102,754–759.

Raymond, D. J., 2000: The Hadley circulation as a radiative–con-vective instability. J. Atmos. Sci., 57, 1286–1297.

Sadler, J., 1975: The monsoon circulation and cloudiness over theGATE area. Mon. Wea. Rev., 103, 369–387.

Smolarkiewicz, P. K., and L. G. Margolin, 1997: On forward-in-timedifferencing for fluids: An Eulerian/semi-Lagrangian nonhydro-static model for stratified flows. Atmos.–Ocean, 35, 127–152.

Sumi, A., 1992: Pattern formation of convective activity over theaqua-planet with globally uniform sea surface temperature(SST). J. Meteor. Soc. Japan, 70, 855–876.

Tomas, R. A., and P. J. Webster, 1997: The role of inertial instabilityin determining the location and strength of near-equatorial con-vection. Quart. J. Roy. Meteor. Soc., 123, 1445–1482.

Waliser, D. E., and R. C. J. Somerville, 1994: Preferred latitudes ofthe intertropical convergence zone. J. Atmos. Sci., 51, 1619–1639.

Zhang, C., 2001: Double ITCZs. J. Geophys. Res., 106, 11 785–11 792.


Documents

Explicit Simulations of the Intertropical Convergence Zone