• Chou, C., , and J. D. Neelin, 2003: Mechanisms limiting the northward extent of the northern summer monsoons over North America, Asia, and Africa. J. Climate, 16 , 406425.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , J. D. Neelin, , and H. Su, 2001: Ocean-atmosphere-land feedbacks in an idealized monsoon. Quart. J. Roy. Meteor. Soc., 127 , 18691891.

    • Search Google Scholar
    • Export Citation
  • Cook, K. H., , and A. Gnanadesikan, 1991: Effects of saturated and dry land surfaces on the tropical circulation and precipitation in a general circulation model. J. Climate, 4 , 873889.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1998: Land-sea geometry and its effect on monsoon circulations. J. Geophys. Res., 103 , 1155511572.

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

  • Godbole, R., 1977: The composite structure of the monsoon depression. Tellus, 29 , 2540.

  • Hahn, D., , and S. Manabe, 1975: The role of mountains in the South Asian monsoon circulation. J. Atmos. Sci., 32 , 15151541.

  • Hoskins, B., , and M. Rodwell, 1995: A model of the Asian summer monsoon. Part I: The global scale. J. Atmos. Sci., 52 , 13291340.

  • Krishnamurti, T., , M. Kanamitsu, , R. Godbole, , C-B. Chang, , F. Carr, , and J. Chow, 1975: Study of a monsoon depression. I: Synoptic structure. J. Meteor. Soc. Japan, 53 , 227239.

    • Search Google Scholar
    • Export Citation
  • Molnar, P., , and K. Emanuel, 1999: Temperature profiles in radiative-convective equilibrium above surfaces at different heights. J. Geophys. Res., 104 , 2426524271.

    • Search Google Scholar
    • Export Citation
  • Pauluis, O., 2001: Axially symmetric circulations in a moist atmosphere. Preprints, 13th Conf. on Fluid Dynamics, Breckenridge, CO, Amer. Meteor. Soc., CD-ROM, 4.12.

  • Plumb, R., , and A. Hou, 1992: The response of a zonally symmetric atmosphere to subtropical forcing: Threshold behavior. J. Atmos. Sci., 49 , 17901799.

    • Search Google Scholar
    • Export Citation
  • Privé, N. C., , and R. A. Plumb, 2007: Monsoon dynamics with interactive forcing. Part I: Axisymmetric studies. J. Atmos. Sci., 64 , 14171430.

    • Search Google Scholar
    • Export Citation
  • Rodwell, M., , and B. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc., 122 , 13851404.

  • Rodwell, M., , and B. Hoskins, 2001: Subtropical anticyclones and summer monsoons. J. Climate, 14 , 31923211.

  • Thorncroft, C., , and B. Hoskins, 1994: An idealized study of African easterly waves: I: A linear view. Quart. J. Roy. Meteor. Soc., 120 , 953982.

    • Search Google Scholar
    • Export Citation
  • Xie, S-P., , and N. Saiki, 1999: Abrupt onset and slow seasonal evolution of summer monsoon in an idealized gem simulation. J. Meteor. Soc. Japan, 77 , 949968.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , and C. Li, 1994: Mechanism of heating and the boundary layer over the Tibetan Plateau. Mon. Wea. Rev., 122 , 305323.

  • View in gallery

    Model framework for cases of increasing complexity. (top) Axisymmetric setup in the latitude–height plane with subtropical continent. (center) Three-dimensional framework with zonally symmetric landmass on the latitude–longitude plane. (bottom) Three-dimensional framework with zonally asymmetric landmass on the latitude–longitude plane.

  • View in gallery

    (top to bottom) Steady-state (100-day time mean) streamfunction, subcloud moist static energy, and precipitation fields. (a),(c),(e) Two-dimensional cases; (b),(d),(f) three-dimensional cases with symmetric continent. Subcloud moist static energy is shown in units of 105 J. Precipitation is shown in units of mm day−1. For streamfunction, solid contours indicate counterclockwise flow and dotted contours indicate clockwise flow; the contour interval is 5.0 × 109 kg s−1 in (a)–(d) and 1.0 × 1010 kg s−1 in (e), (f). Cases with weak forcing and uniform warm ocean, THF0 = 130 W m−2, are shown in (a) and (b). Cases with uniform warm ocean and global circulation, THF0 = 140 W m−2, are shown in (c) and (d). Cases with cross-equatorial circulation and summer-like ocean, THF0 = 140 W m−2, are shown in (e) and (f).

  • View in gallery

    Steady-state zonal mean fields (100-day mean) as a function of land surface forcing strength for cases with uniform warm ocean. Dashed line with squares represents the axisymmetric case; solid line with circles represents the three-dimensional case with zonally symmetric continent. (top) Absolute global maximum streamfunction strength of counterclockwise meridional cell. (bottom) Minimum 150-hPa absolute vorticity between 6° and 64°N.

  • View in gallery

    Wave spectra for 1000-mb meridional velocity in case with symmetric continent, THF0 = 125 W m−2. Speed is in m s−1, and negative wavenumbers indicate westward propagation. (a) Spectra at 2°N, with contour interval of 3.0 × 10−10 m s−l × Δc−1, where Δc is the unit phase speed interval of 1.0 m s−1. (b) Spectra at 42°S, with contour interval of 6.0 × 10−8 m s−1 × Δc−1.

  • View in gallery

    Composite of 1000-mb winds (arrows) and precipitation (shading, mm day−1), calculated in a frame of reference moving westward with the disturbance, for the case with zonally symmetric continent and THF0 = 130 W m−2. Mean is over days 327 to 387.5.

  • View in gallery

    Hovmoeller diagram of precipitation at 18°N (mm day−1). First 100 days are transition to summer forcing state. (a) Case with zonally symmetric continent with summer SST profile, THF0 = 130 W m−2. (b) Case with asymmetric continent (land between 0° and 180°) and summer-like SST, THF0 = 140 W m−2.

  • View in gallery

    Wave structure of persistent anomaly for case with zonally symmetric continent, THF0 = 130 W m−2. (a) Composite vertical velocity (ω), with contour interval of 0.1 Pa s−1, and dotted contours indicate ascent. (b) Composite eddy temperature, with contour interval of 1 K. (c) Eliassen–Palm flux (not composite); contours indicate divergence, with contour interval of 30 m2 s−2. Panels (a) and (b) represent a composite of a vertical slice at 18°N calculated in a frame of reference moving westward with the persistent anomaly over days 327.5 to 387.5.

  • View in gallery

    Comparison of subcloud moist static energy for cases with and without eddies. Solid line shows 1000-hPa h (J) for axisymmetric case with uniform warm ocean, THF0 = 140 W m−2. Dot–dash line shows zonal mean 1000-hPa h for three-dimensional case with zonally symmetric continent and warm ocean, THF0 = 140 W m−2.

  • View in gallery

    1000-hPa winds (m s−1) and precipitation (mm day−1), for case with asymmetric continent and THF0 = 140 W m−2. (a) Case with uniform warm ocean; (b) case with summer-like ocean SST; (c) case with summer-like ocean SST and thin walls at eastern and western coastlines.

  • View in gallery

    Global zonal mean meridional circulation for three-dimensional cases with asymmetric forcing; contour interval is 5 × 109 kg s−1. (a) Aquaplanet case with ΔT = 2.25 K, SST maximum at 24°N, 50-day time mean. (b) Continental island case with THF0 = 180 W m−2, 100-day time mean.

  • View in gallery

    1000-hPa winds (m s−1) and moist static energy (J), as in Fig. 9 for case with asymmetric continent and THF0 = 140 W m−2. (a) Case with uniform warm ocean; (b) case with summer-like ocean SST; (c) case with summer-like ocean SST and thin walls at eastern and western coastlines.

  • View in gallery

    500-hPa ω for case with asymmetric continent and warm ocean, THF0 = 140 W m−2. Dotted contours indicate ascent; contour interval is 2.5 × 10−3 Pa s−1.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 52 51 13
PDF Downloads 39 39 7

Monsoon Dynamics with Interactive Forcing. Part II: Impact of Eddies and Asymmetric Geometries

View More View Less
  • 1 Program in Atmospheres, Oceans, and Climate, Massachusetts Institute of Technology, Cambridge, Massachusetts
© Get Permissions
Full access

Abstract

The roles of eddies and forcing asymmetry in the dynamics of the large-scale monsoon circulation are investigated with a general circulation model. The net impact of eddies is found to be a slight weakening of the zonal mean monsoon circulation. The eddies strongly impact the momentum budget of the circulation, but the qualitative behavior of the monsoon flow is not substantially altered. The introduction of asymmetric forcing reveals the limitations of axisymmetric studies in representing the fully three-dimensional monsoon. Advection of low subcloud moist static energy air from the midlatitude oceans is seen to strongly impact the subcloud moist static energy budget in the continental subtropics, limiting the poleward extent of the monsoon. The advection of low moist static energy air must be blocked by orography, or the source of low moist static energy air must be removed, in order to induce strong precipitation over the subtropical landmass. An equatorial SST gradient is needed to induce a cross-equatorial meridional monsoon circulation. The location of the maximum subcloud moist static energy remains a good indicator for the limit of the monsoon.

Corresponding author address: Nikki Privé, Cooperative Institute for Research in the Atmosphere, NOAA/ESRL, R/GSD, 325 Broadway, Boulder, CO 80305-3337. Email: nikki.prive@noaa.gov

Abstract

The roles of eddies and forcing asymmetry in the dynamics of the large-scale monsoon circulation are investigated with a general circulation model. The net impact of eddies is found to be a slight weakening of the zonal mean monsoon circulation. The eddies strongly impact the momentum budget of the circulation, but the qualitative behavior of the monsoon flow is not substantially altered. The introduction of asymmetric forcing reveals the limitations of axisymmetric studies in representing the fully three-dimensional monsoon. Advection of low subcloud moist static energy air from the midlatitude oceans is seen to strongly impact the subcloud moist static energy budget in the continental subtropics, limiting the poleward extent of the monsoon. The advection of low moist static energy air must be blocked by orography, or the source of low moist static energy air must be removed, in order to induce strong precipitation over the subtropical landmass. An equatorial SST gradient is needed to induce a cross-equatorial meridional monsoon circulation. The location of the maximum subcloud moist static energy remains a good indicator for the limit of the monsoon.

Corresponding author address: Nikki Privé, Cooperative Institute for Research in the Atmosphere, NOAA/ESRL, R/GSD, 325 Broadway, Boulder, CO 80305-3337. Email: nikki.prive@noaa.gov

1. Introduction

It was proposed in the companion paper (Privé and Plumb 2007, hereafter Part I) that the location of the deep ascent branch of an angular momentum conserving circulation is strongly tied to the distribution of subcloud moist static energy. If a localized region of subtropical forcing is strong enough to induce such a meridional circulation, the dividing streamline at the boundary of the circulation, where there is no horizontal advection in the free troposphere, should be located in a region of zero vertical shear of zonal wind. In thermal wind balance with a moist adiabat in the vertical, the only regions that have no vertical shear of zonal wind are those with no meridional gradient of subcloud moist static energy, hb. Therefore, given a local maximum of subcloud moist static energy, the poleward boundary of the circulation will be collocated with the maximum hb, and the large-scale ascent and greatest precipitation will occur near and slightly equatorward of the hb maximum. It was found that the location of the monsoon was in good agreement with this theory in a series of axisymmetric modeling studies.

In Part I, the nonlinear theory of Hadley circulations was investigated as it applies to the large-scale dynamics of the monsoon in an axisymmetric framework. The focus of Part I was on the impact of interactive forcing on the dynamics of the monsoon; it was found that advection of subcloud moist static energy by the overturning circulation strongly impacted the location of the monsoon as well as the strength of the circulation. Axisymmetric studies are limited by the exclusion of zonally asymmetric forcing and horizontal eddies; they cannot address the observed asymmetry of the monsoon. In this paper, the behavior of the monsoon in three-dimensional flow is explored.

Some studies have sought to understand the monsoon circulation by examining the dynamical response to a specified local heat source. These studies have emphasized the role of tropical waves in determining the large-scale flow. The linear Gill (1980) model frames the atmospheric response to a localized off-equatorial heat source in terms of forced Kelvin and Rossby waves. Using a similar model, Hoskins and Rodwell (1995) found that the large-scale circulation associated with a specified diabatic heating is in good agreement with observed flow, suggesting that the dynamics are principally linear. Rodwell and Hoskins (1996) suggested that Rossby waves generated in the monsoon region suppress convection to the northwest of the monsoon, causing east–west asymmetry. The Rodwell–Hoskins effect and Gill-type circulations are completely neglected in an axisymmetric framework. An important limitation of these studies is the lack of interaction between the forcing and the induced flow; in fact, the diabatic heating is strongly dependent upon the circulation, and is itself a part of the dynamical response to external forcing.

The advection of moist air into the monsoon region by the large-scale flow is a controlling factor in determining the strength and location of the monsoon. The relationship between the moisture supply and the monsoon geometry has been emphasized in a number of studies that allow interaction between the circulation and the forcing. Using a simple rectangular continent, Cook and Gnanadesikan (1991) demonstrated that rainfall is difficult to achieve over the interior of a continent due to negative feedbacks with the surface hydrology. Dirmeyer (1998) found that the monsoon over a subtropical continent of limited longitudinal extent is supplied moisture primarily by easterly flow along the eastern coastline. The monsoon is weakened when a continent spanning 360° longitude is used, in which case the moisture supply by easterlies from the ocean is lost. Xie and Saiki (1999) found that baroclinic waves generated by a low-level easterly jet along the coastline play an important role in monsoon onset by initiating precipitation over the landmass. Their results suggested that without these baroclinic waves, monsoon onset would be more difficult or even impossible. Xie and Saiki also found that the low-level cyclonic circulation over the continent contributes to the east–west asymmetry of the monsoon by suppressing moist convection to the west, while encouraging rainfall in the east. Chou et al. (2001) suggested that the advection of air with relatively low moist static energy from the cool midlatitude oceans acts to suppress moist convection over the western continent and limits the poleward extent of the monsoon. In Part I, it was shown that advection of air with relatively low moist static energy from the tropical oceans has a strong impact on the location of the monsoon.

A model of intermediate complexity is used here to bridge the gap between simple models and the fully three-dimensional interactive monsoon. The wealth of feedbacks present in full GCM studies makes diagnosis of the monsoon behavior extremely difficult. The approach taken here is to start from a simple framework and gradually increase the complexity of the model, so as to reveal the impact of individual mechanisms upon the monsoon. Understandability is maintained by using simple representations of continental geometry and idealized physical parameterizations in a full GCM.

In Part I, an axisymmetric model was used to explore the role of interactive forcing in the dynamics of the monsoon. While axisymmetric theory is useful for developing an understanding of the basic physical mechanisms that drive and affect the monsoon, it is unclear how the simplifications that are involved impact the results. In the present study, the three-dimensionality of the monsoon circulation is investigated. The model used is the Massachusetts Institute of Technology (MIT) General Circulation Model (MITGCM), as in Part I, to which the reader is referred for details. First, a two-dimensional setup on the latitude–height plane with subtropical continent is tested to provide a baseline for comparison. The model is then expanded to three dimensions (Fig. 1), but the surface conditions and continental geometry are kept zonally symmetric; comparison of these results with the two-dimensional case will reveal the role of eddies in the dynamics of the monsoon. Finally, a continent of limited longitudinal extent is introduced to explore the effects of zonally asymmetric forcing on the large-scale monsoon circulation.

2. Impact of eddies

The role of eddies in the dynamics of the large-scale monsoon is studied by comparing the circulation in a purely axisymmetric model framework to the circulation in a case where eddy motions are allowed, but zonal symmetry of the forcing is maintained. South of the coastline at 16°Ν is an ocean with the SST a function of latitude only. Two different ocean temperature profiles are used: a uniformly warm ocean with SST of 302 K at all locations, and a summer-like SST with maximum temperature at 8°Ν:
i1520-0469-64-5-1431-e1
The uniform warm ocean setup is tested first to isolate the dynamical response to a local subtropical forcing, then the summer ocean temperature distribution (1) is used. Over the land surface, a proxy for the net downward radiative flux (THF) is prescribed as a function of latitude with maximum flux at the southern coast of the continent:
i1520-0469-64-5-1431-e2
i1520-0469-64-5-1431-e3
where LHF is the latent heat flux, SHF is the sensible heat flux, ΔTHF is 50 W m−2, and THF0 is varied from 120 to 150 W m−2. The strength of the surface forcing is manipulated by varying THF0. The distribution of surface forcing given by (3) is intended to represent a summer mean forcing rather than solsticial forcing.

The two-dimensional experiment results show three different dynamical behaviors, depending on the forcing magnitude. For weak land forcing (THF0 ≤ 120 W m−2, not shown), a shallow dry circulation forms with little precipitation over the subtropical continent. At intermediate forcing levels (125 W m−2 ≤ THF0 ≤ 130 W m−2; Fig. 2a), a deep but meridionally narrow circulation occurs with ascent and precipitation over the land and subsidence over the nearby ocean. A cross-equatorial circulation with ascent and precipitation over the land and subsidence in the opposite hemisphere occurs for sufficiently strong forcing (THF0 ≥ 135 W m−2; Fig. 2c); this circulation is close to angular momentum conserving in the free troposphere (Fig. 3). This transition from local to global circulation is in agreement with the nonlinear theory of Hadley circulations (Plumb and Hou 1992). When a summer-like SST distribution (1) is included, the response over the continent is not significantly affected, although a cross-equatorial circulation is always present due to the SST gradient at the equator (Fig. 2e). The streamfunction shows a tendency to jump at the equator, with a precipitation maximum in the Southern Hemisphere; this type of behavior is described in detail by Pauluis (2001), and is a result of dynamical limits on cross-equatorial flow in the mixed layer.

When the model is expanded to include a third dimension with symmetric landmass, the zonal mean circulation is quite similar to the two-dimensional results overall. The transition between three different circulation regimes is observed, although stronger land forcing is needed to incur a cross-equatorial circulation in comparison to the two-dimensional cases. A deep but local circulation occurs for forcing of THF0 = 135 W m−2, with only a shallow dry circulation for weaker forcing, as seen in Fig. 2b for THF0 = 130 W m−2. The transition to a global meridional circulation occurs at THF0 = 140 W m−2 (Fig. 2d), in comparison to a threshold at THF0 = 135 W m−2 in the axisymmetric case (Fig. 3). Above this threshold, however, the geometry and magnitude of the circulation is remarkably similar to the two-dimensional cases with the same forcing, both in the cases with uniform warm ocean and with summer-like ocean temperatures (Fig. 2f). In contrast to the meridional streamfunction, the zonal wind field (not shown) is noticeably altered by the presence of eddies, with weakening of both the upper-level easterly jet at the equator and the westerly midlatitude jets. The zonal mean cross-equatorial circulation in the strong forcing cases (THF0 ≥ 140 W m−2) is far from angular momentum conserving, as indicated by the nonzero absolute vorticity (Fig. 3b). Although a clear transition between local and global circulations is seen in both the two- and three-dimensional cases, threshold behavior of the circulation strength as predicted by the nonlinear theory is not observed (Fig. 3a). However, the subcritical forcing range is narrow, so that threshold-type behavior is difficult to discern.

Various types of eddy motions are observed in the three-dimensional cases. The midlatitude westerly jets show eastward-propagating disturbances of wavenumbers 5–7 (Fig. 4b), which produce midlatitude rain bands. In nonmonsoonal cases, slowly westward-propagating waves in the meridional flow field are more common over the equatorial ocean (Fig. 4a) with wavenumbers ranging from 2 to 7. When a strong zonal mean monsoon circulation is present (THF0 ≥ 140 W m−2), the subtropical precipitation field (not shown) features widespread moderate precipitation with little wave activity, although there are regions of localized precipitation that may remain stationary or may progress eastward or westward.

The most interesting wave behavior over the subtropical continent is seen when the land forcing is slightly weaker than the level needed to induce zonally extensive deep convection over the subtropical landmass. In these cases, the land remains arid but the atmospheric temperature is elevated, so that there is a poleward-increasing temperature gradient in the lower troposphere. A low-level easterly jet forms along the coastline, centered near 700 mb, and baroclinic disturbances are generated from the jet. A most striking example is seen for the case with THF0 = l30 W m−2 and summer-like SST distribution: a single, very long-lived warm-core anomaly forms along the coastline (Fig. 5), which propagates westward at 3.25 m s−1. A Hovmoeller diagram of the precipitation along the coast is shown in Fig. 6a, illustrating the remarkable persistence of the anomaly and its westward progression. The anomaly shows strong cyclonic flow in the lower troposphere (Fig. 5) around a surface low, with enhanced precipitation in the region of southwesterlies on the east side of the anomaly, and arid conditions on the west side. The anomaly is warm core in the lower troposphere, with ascent collocated with the positive eddy temperature (Fig. 7) and poleward flow (not shown). There is ascent on the eastern side of the disturbance (Fig. 7a), with subsidence to the west; the vertical velocity field tilts eastward with height from the surface to 750 mb, with little tilt above 750 mb. The Eliassen–Palm flux signature (Fig. 7c) is also indicative of a baroclinic wave structure, with downward flux originating near the core of the easterly jet at 700 mb, 20°–30°N.

The westward-propagating waves may be compared with observed and modeled monsoon depressions. Observed monsoon depressions feature a cold-core structure in the lower troposphere overlain with a warm core aloft (Godbole 1977; Krishnamurti et al. 1975), with an eastward tilt of the wave structure with height. The precipitation maximum is observed to be in the southwest sector of the monsoon depression, while in the MITGCM, the precipitation maximum is in the southeast sector (Fig. 5). The wave structure is perhaps more reminiscent of African easterly waves over a dry inland region (Thorncroft and Hoskins 1994), where moist convection is favored only in regions of poleward flow on the east side of the wave. Xie and Saiki (1999) found westward-propagating waves to be instrumental in onset of the monsoon in a GCM with simple continent. The waves generated in the MITGCM are in many ways similar to the waves found by Xie and Saiki: baroclinic origin from instability of a low-level easterly jet along the continental coastline, westward propagation, eastward tilt of the waves with height; however, the structure of the waves is different. Xie and Saiki’s onset waves were baroclinic in nature in the lower troposphere and a convective structure in the upper troposphere, with vertical motion collocated with warm temperature anomaly only in the upper troposphere. The waves seen in the MITGCM are warm core throughout the entire depth of the troposphere.

The circulation strength as a function of the land forcing is illustrated by the solid line in Fig. 3. The net effect of eddies appears to be a weakening of the monsoon in comparison to the two-dimensional cases: stronger land surface forcing is needed to induce a global mean monsoon in the three-dimensional cases. The meridional circulation is not as conservative of angular momentum when eddies are included, as indicated by the upper-tropospheric absolute vorticity, which does not closely approach zero even for a global circulation (Fig. 3b). In the cases with summer-like ocean temperatures, the eddies do not seem to strongly affect the strength of the cross-equatorial cells associated with the SST distribution, as noted by comparison of Figs. 2e,f.

Why is a stronger land surface forcing needed to induce a monsoon in the three-dimensional cases with symmetric continent than in the two-dimensional cases? The poleward boundary of the circulation cell is collocated with the maximum in subcloud moist static energy, as previously discussed. The subcloud moist static energy maximum is weaker and located slightly closer to the equator in the three-dimensional cases in comparison to the two-dimensional cases (Fig. 8). The eddies tend to redistribute hb, decreasing the moist static energy near the maximum (near the coastline), and increasing hb near minimums (inland). The hb maximum decreases in magnitude and shifts slightly closer to the coastline in the presence of eddies.

3. Asymmetric continent

Zonal asymmetry of the forcing is introduced by limiting the longitudinal extent of the continent to 180°, with ocean covering the opposite hemisphere (bottom panel in Fig. 1). The location of the coastline and the distribution of land surface forcing remain the same as in the axisymmetric and zonally symmetric cases.

A uniformly warm ocean with SST of 302 K at all latitudes is tested first to isolate the dynamical response to localized forcing. For weak forcing (THF0 ≤ 125 W m−2), the maximum subtropical precipitation (not shown) remains over the ocean, with only weak rainfall along the immediate coastline. The majority of the coastal rainfall is associated with westward-propagating waves generated in a low-level easterly jet in thermal balance with the poleward-increasing temperatures in the lower troposphere at the coastline. These waves are very similar to the baroclinic waves seen in the zonally symmetric cases (Figs. 5, 7), although they die out after reaching the western coastline, with new waves forming along the eastern and central continent. In these cases, the subcloud moist static energy (not shown) is maximum over the tropical ocean, near the coastline.

With stronger land surface forcing (THF0 ≥ 130 W m−2), a precipitation maximum (Fig. 9a) forms over the subtropical continent, with greatest rainfall over the western half of the landmass. The subcloud moist static energy (Fig. 10a) develops a strong maximum over the interior of the continent, with highest values over the western continent. A low-level anticyclone forms over the Northern Hemisphere ocean, with cyclonic flow and strong southerly inflow over the subtropical continent. Although there is significant large-scale ascent and rainfall, the meridional overturning circulation in the continental region is weak and does not span the equator, with subsidence over the Northern Hemisphere ocean (Fig. 11). This circulation is far from angular momentum conserving, and remains weak even for the strongest surface forcing levels—the subsidence region over the tropical ocean is not sufficient to compensate for the upward mass flux in the ascent region. Instead, a local region of strong subsidence occurs to the northwest of the monsoon region, over the eastern ocean. This subsidence region is reminiscent of that described by Rodwell and Hoskins (2001) as a feature induced by Rossby waves generated in the monsoon region.

The lack of a cross-equatorial circulation in the case with uniform warm ocean is in contradiction to the circulation associated with observed monsoons. To test the atmospheric response to a localized subtropical heat source, a series of aquaplanet cases are performed with a local maximum of SST in the subtropics, and uniform SST elsewhere. The SST perturbation takes the form
i1520-0469-64-5-1431-e4
i1520-0469-64-5-1431-e5
where ϕ0 is the latitude of the maximum perturbation SST and λ is the longitude. When the SST perturbation is located relatively close to the equator (ϕ0 ⩽ 24° for ΔT = 2.25 K), the localized forcing induces a global, cross-equatorial circulation with ascent and strong precipitation near the maximum SST and subsidence in the Tropics and opposite hemisphere (Fig. 12b). However, as the SST perturbation is shifted poleward, the cross-equatorial flow weakens and then ceases, leaving a localized circulation similar to that seen in the previously discussed continental cases. These results are in agreement with the axisymmetric results shown in Part I, wherein the cross-equatorial circulation weakened and became local as the forcing was moved poleward. While the subcloud moist static energy maximum is collocated with the SST maximum in an aquaplanet setup, the subcloud moist static energy maximum tends to shift poleward with increased land forcing in continental cases, as discussed in Part I. At the strong land forcing levels, which might induce a cross-equatorial circulation, the hb maximum has shifted far inland, and only a local overturning circulation results.

To test this hypothesis, an island continent with coastlines at both 16° and 32°N is tested with a uniformly warm ocean and very strong surface forcing of THF0 = 180 W m−2. The ocean to the north of the northern coast prevents the hb maximum from moving out of the subtropics even for strong surface forcing; for a continent extending poleward, the highest subcloud moist energy shifts poleward into the midlatitudes. Figure 10b shows that a cross-equatorial circulation occurs for the island case, similar to that seen in the aquaplanet case (Fig. 12a). These results suggest that a cross-equatorial meridional circulation may be induced by a localized subtropical forcing, but that stronger forcing is needed to do so with an asymmetric continent. This case also illustrates that factors which control the hb distribution can strongly impact the large-scale circulation.

A summer-like ocean SST distribution that is cooler in the midlatitudes is introduced to increase the realism of the model setup. The results are somewhat surprising: even for the cases with strongest land forcing, precipitation over the subtropical continent is weak (Fig. 9b), and there is no deep meridional circulation over the land. The ITCZ and Hadley circulation remain located over the tropical ocean, associated with the SST maximum at 8°N. The southeastern corner of the continent is the only subtropical region which receives significant rainfall. Over the central and western continent, rainfall is associated with westward-propagating baroclinic waves (Fig. 6b) of a low-level easterly jet, as seen in previous cases. This result is quite different than that found by Xie and Saiki (1999) for the perpetual summer case, which featured strong precipitation over the continental interior. The different response between the two models may be the result of the different surface forcing: in Xie and Saiki’s model, the perpetual summer case is forced by 1 June insolation, which has much stronger forcing over the continental interior than that prescribed by (3) in the MITCGM.

If the SST maximum is shifted poleward to coincide with the coastline, removing the small dip in ocean temperatures between 8° and 16°N, moist convection (not shown) becomes more widespread along the immediate coast, especially over the western continent, but the continental interior remains arid. This SST distribution is similar to that in the Indian Ocean, with SST increasing poleward; the results show that maximum SSTs near the coastline are not sufficient of themselves to induce an inland monsoon.

The subcloud moist static energy field (Fig. 10) holds clues to the pronounced differences in the response between the summer-like ocean cases and the uniformly warm ocean cases. With a summer-like SST distribution, there is a large pool of low moist static energy air over the midlatitude ocean. Westerlies in the midlatitudes carry this low energy air over the continent, decreasing the moist static energy levels over interior of the continent. The maximum moist static energy levels remain over the tropical ocean equatorward of the coastline, with a rapid decrease in moist static energy poleward of the coastline. However, when the ocean is warm at all latitudes, the advection of oceanic air over the continent has a much weaker effect on the moist static energy budget of the continental interior, and a moist static energy maximum can form inland. The impact of advection of low moist static energy air on the monsoon has been described by Chou et al. (2001), with similar results.

To assess the idea that the advection of oceanic air with low moist static energy is responsible for the weakening of the monsoon and the decrease in subcloud moist static energy over the continent, an additional case is tested with thin walls to block advection from the ocean. The thin walls extend from the surface to 700 mb, and simply prevent flow between adjacent grid boxes. The low height of the walls should limit the direct impact on the flow to lower-tropospheric levels only. Two walls are added: one along the west coast of the continent, and one along the east coast; flow across the southern coastline is not altered.

The resulting boundary layer moist static energy (Fig. 10c) is much greater across the entire continent in comparison to the cases without walls, especially in the subtropics near 30°N, and along the eastern edge of the continent where the flow impinges upon one of the thin walls. The zonal mean hb maximum over the continent is located slightly inland near 20°N. The time mean precipitation field (Fig. 9c) shows greater precipitation along the entire southern coastline of the continent, and also over the eastern third of the continental interior in comparison to the corresponding case without a wall. Over the eastern continent, rainfall is widespread and persistent, and appears to be associated with large- scale ascent induced by the thin wall. Along the western subtropical continent, precipitation is dominated by westward-propagating disturbances. The zonal mean meridional circulation over the continent features a single cross-equatorial cell, with deep ascent ranging from the Southern Hemisphere Tropics to the subtropical continent.

These results illustrate the very strong impact of low-level advection on the distribution of subcloud hb and correspondingly on the monsoon. This implies that factors that affect the advection of hb may also have a strong effect on the monsoon; these include both factors that influence the form of the large-scale flow and those that influence the moist static energy content of the upstream atmosphere.

4. Discussion

In Part I of this work, the location and extent of the monsoon was shown to be strongly dependent upon the distribution of subcloud moist static energy. In this study, the influence of the large-scale flow on the subcloud moist static energy is seen to greatly affect the monsoon circulation.

The existence of eddies in the midlatitudes and equatorial regions has relatively little impact on the zonal mean steady meridional circulation in the cases with symmetric continent, although the upper-tropospheric zonal wind jets are weakened significantly. The eddies act to redistribute moist static energy in the subcloud layer, bringing the moist static energy maximum closer to the equator, and weakening the magnitude of the maximum. This results in a circulation that is slightly weaker and whose ascent branch is closer to the equator than in the axisymmetric cases (with no eddies). Even though the meridional circulation does not conserve angular momentum as closely as in the axisymmetric cases, a transition from local to global circulation occurs as predicted by the nonlinear theory.

The introduction of asymmetric continental geometry strongly impacts the large-scale monsoon circulation. When the longitudinal extent of the continent is limited to 180° with uniform ocean temperatures, the meridional circulation remains confined to the summer hemisphere for even the strongest land surface forcing. Although the meridional circulation is relatively weak, there is strong deep ascent and considerable precipitation over the continent. Much of the compensating subsidence occurs not in the descent branch of the meridional circulation, but in a region to the northwest of the monsoon. This subsidence region is similar to the Rossby wave–induced descent found by Rodwell and Hoskins (1996, 2001).

Although the meridional circulation does not closely follow the axisymmetric theory of angular momentum conserving (AMC) circulations when an asymmetric landmass is used, the subcloud moist static energy distribution remains a good indicator of deep moist convection over the continent. Deep ascent and moist convection occurs over the landmass only when the subcloud moist static energy is greater over the continent than over the tropical oceans. The ascent and precipitation occur on the equatorward side of the moist static energy maximum, which is the location predicted by theory (Privé and Plumb 2007). As long as the boundary of the meridional circulation has zero vertical shear of zonal wind, the relationship between the subcloud moist static energy distribution and the location of ascent and precipitation prescribed by the theory should hold. Even if the zero wind shear argument is not valid locally, the tropospheric temperature and wind fields must remain in agreement with the local gradient of subcloud moist static energy in the monsoon region. The theory is a dynamical constraint on the location of the boundary of the meridional circulation; the precipitation is strongly favored by the vertical motion associated with the meridional circulation, but may also be triggered by convective instability and high subcloud moist static energy.

The impact of advection of moist static energy is revealed in the cases with zonally asymmetric continent. For the case of uniformly warm oceans, advection of relatively low subcloud moist static energy from the tropical ocean across the southern boundary of the continent causes the maximum hb and the monsoon precipitation to shift poleward. Although the strongest forcing is located near the coastline, the hb maximum moves further poleward with increased land forcing strength, because of the resultant increased advection of low moist static energy air from the tropical ocean. This poleward shift of the monsoon prevents the formation of a zonal mean cross-equatorial circulation, so that the monsoon circulation is limited to the Northern Hemisphere. Chou et al. (2001) used a continent that was centered on the equator, and did not observe this type of behavior. Only in the “island” case, where poleward movement of the hb maximum was limited, was a cross-equatorial meridional circulation achieved with an asymmetric continent.

In the cases with summer-like SST distribution, the cool midlatitude oceans are a source of low moist static energy air, which is advected into the continental interior by the prevailing westerly winds and by cyclonic flow over the landmass. This advection drastically reduces the subcloud moist static energy over the continental interior and along the western subtropical continent, so that the moist static energy is greater over the ocean than over the continent in these areas. A deep meridional circulation does not form over the land, and the precipitation along the central and western continent is weak even for the strongest forcing levels. Over the eastern continent, the low-level flow carries tropical ocean air inland, and a maximum in moist static energy forms over the subtropical continent for strong forcing, allowing deep ascent and steady precipitation only in the southeastern corner of the continent. Over the central and western continent, where the land is relatively dry, a low-level easterly jet forms in thermal wind balance with the poleward-increasing temperature gradient; baroclinic disturbances form in this jet and propagate westward along the coastline and provide the only source of precipitation over the central and western continent.

These results support the work of Chou et al. (2001) and Chou and Neelin (2003), who found that ventilation of the monsoon region by advection of low moist static energy air limits the poleward and westward extent of the monsoon. The analytic theory linking the monsoon location to the subcloud moist static energy distribution gives a framework to explain why the advection of hb is so influential. It is doubtful that the Rodwell–Hoskins effect of subsidence due to Rossby waves is solely responsible for the strong asymmetry of the monsoon in the cases with summer SSTs: if it were, the cases with uniform SSTs, which have a much stronger monsoon, should also show significant east–west asymmetry, but they do not.

When a thin wall is used to block advection of midlatitude air into the monsoon region, a cross-equatorial meridional circulation develops with deep ascent over the subtropical asymmetric continent. When low hb air is thus blocked, a maximum in sub-cloud moist static energy forms over the subtropical continent, with precipitation and deep ascent over the continent. This ascent region is joined to the cross-equatorial Hadley cell, although stronger ascent occurs over the equatorial ocean. This illustrates the importance of advection of subcloud moist static energy on the total moist static energy field and on the large-scale monsoon dynamics. Zonal asymmetry of the large-scale flow clearly has a severe impact on the applicability of axisymmetric theory to the monsoon circulation.

The model studies have revealed that the formation of a monsoon with significant inland precipitation and strong meridional circulation is not trivial in the asymmetric case. A maximum in subcloud moist static energy is difficult to achieve over the interior of the continent in the absence of some mechanism to shield the subtropics from the advection of low energy air from the midlatitude oceans. Additionally, tropical ocean temperature gradients at the equator are needed to encourage a robust cross-equatorial meridional circulation.

The subcloud moist static energy may help to explain why the Asian monsoon is the largest and strongest of the observed monsoon regions. In the summer months, the upper-tropospheric temperature field features a strong maximum over the Tibetan Plateau, which has long been presumed to be major factor in causing the Asian monsoon to be the most intense of the global monsoons (Hahn and Manabe 1975; Yanai and Li 1994). Molnar and Emanuel (1999) have shown that the radiative convective equilibrium temperature at a given pressure level is greater over an elevated surface than over a surface at sea level; this helps to explain the strong maximum in subcloud moist static energy over the Himalayas. The results of this study further suggest, however, that the blocking effect by the Tibetan Plateau may have a very significant impact on the moist static energy distribution, by shielding India and Southeast Asia from inflow from the Asian midlatitudes. The South American monsoon may also be affected by blocking of the low-level flow by orography. A pool of high subcloud moist static energy forms over Amazonia during the summer months in European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis data (not shown); this region is sheltered from influx of low energy air from the ocean by the Andes Mountains to the west, and the Brazilian Highlands to the southeast.

In contrast to Asia, the Australian monsoon is confined only to the coastal region, with the strongest moist convection occurring over the tropical ocean. The January long term mean ECMWF reanalysis data indicates that Australia has a maximum in subcloud moist static energy at the coastline (not shown), with moist static energy decreasing rapidly inland over the interior desert. Advection of low moist static energy air from the nearby oceans keeps hb levels low over the interior, banishing the monsoon to the immediate coastal region. The results of the model cases with asymmetric continent and summer-like ocean temperatures (Fig. 9b) are reminiscent of the Australian monsoon.

Acknowledgments

This work was supported by National Science Foundation Grant ATM-0436288; N. Privé received support from a National Science Foundation Graduate Research Fellowship. Thanks to Kerry Emanuel, Elfatih Eltahir, John Marshall, Chris Hill, Olivier Pauluis, Jean-Michel Campin, and Ed Hill for helpful discussions. Helpful comments from Shang-Ping Xie and an anonymous reviewer led to significant improvements of this manuscript.

REFERENCES

  • Chou, C., , and J. D. Neelin, 2003: Mechanisms limiting the northward extent of the northern summer monsoons over North America, Asia, and Africa. J. Climate, 16 , 406425.

    • Search Google Scholar
    • Export Citation
  • Chou, C., , J. D. Neelin, , and H. Su, 2001: Ocean-atmosphere-land feedbacks in an idealized monsoon. Quart. J. Roy. Meteor. Soc., 127 , 18691891.

    • Search Google Scholar
    • Export Citation
  • Cook, K. H., , and A. Gnanadesikan, 1991: Effects of saturated and dry land surfaces on the tropical circulation and precipitation in a general circulation model. J. Climate, 4 , 873889.

    • Search Google Scholar
    • Export Citation
  • Dirmeyer, P. A., 1998: Land-sea geometry and its effect on monsoon circulations. J. Geophys. Res., 103 , 1155511572.

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

  • Godbole, R., 1977: The composite structure of the monsoon depression. Tellus, 29 , 2540.

  • Hahn, D., , and S. Manabe, 1975: The role of mountains in the South Asian monsoon circulation. J. Atmos. Sci., 32 , 15151541.

  • Hoskins, B., , and M. Rodwell, 1995: A model of the Asian summer monsoon. Part I: The global scale. J. Atmos. Sci., 52 , 13291340.

  • Krishnamurti, T., , M. Kanamitsu, , R. Godbole, , C-B. Chang, , F. Carr, , and J. Chow, 1975: Study of a monsoon depression. I: Synoptic structure. J. Meteor. Soc. Japan, 53 , 227239.

    • Search Google Scholar
    • Export Citation
  • Molnar, P., , and K. Emanuel, 1999: Temperature profiles in radiative-convective equilibrium above surfaces at different heights. J. Geophys. Res., 104 , 2426524271.

    • Search Google Scholar
    • Export Citation
  • Pauluis, O., 2001: Axially symmetric circulations in a moist atmosphere. Preprints, 13th Conf. on Fluid Dynamics, Breckenridge, CO, Amer. Meteor. Soc., CD-ROM, 4.12.

  • Plumb, R., , and A. Hou, 1992: The response of a zonally symmetric atmosphere to subtropical forcing: Threshold behavior. J. Atmos. Sci., 49 , 17901799.

    • Search Google Scholar
    • Export Citation
  • Privé, N. C., , and R. A. Plumb, 2007: Monsoon dynamics with interactive forcing. Part I: Axisymmetric studies. J. Atmos. Sci., 64 , 14171430.

    • Search Google Scholar
    • Export Citation
  • Rodwell, M., , and B. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc., 122 , 13851404.

  • Rodwell, M., , and B. Hoskins, 2001: Subtropical anticyclones and summer monsoons. J. Climate, 14 , 31923211.

  • Thorncroft, C., , and B. Hoskins, 1994: An idealized study of African easterly waves: I: A linear view. Quart. J. Roy. Meteor. Soc., 120 , 953982.

    • Search Google Scholar
    • Export Citation
  • Xie, S-P., , and N. Saiki, 1999: Abrupt onset and slow seasonal evolution of summer monsoon in an idealized gem simulation. J. Meteor. Soc. Japan, 77 , 949968.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , and C. Li, 1994: Mechanism of heating and the boundary layer over the Tibetan Plateau. Mon. Wea. Rev., 122 , 305323.

Fig. 1.
Fig. 1.

Model framework for cases of increasing complexity. (top) Axisymmetric setup in the latitude–height plane with subtropical continent. (center) Three-dimensional framework with zonally symmetric landmass on the latitude–longitude plane. (bottom) Three-dimensional framework with zonally asymmetric landmass on the latitude–longitude plane.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 2.
Fig. 2.

(top to bottom) Steady-state (100-day time mean) streamfunction, subcloud moist static energy, and precipitation fields. (a),(c),(e) Two-dimensional cases; (b),(d),(f) three-dimensional cases with symmetric continent. Subcloud moist static energy is shown in units of 105 J. Precipitation is shown in units of mm day−1. For streamfunction, solid contours indicate counterclockwise flow and dotted contours indicate clockwise flow; the contour interval is 5.0 × 109 kg s−1 in (a)–(d) and 1.0 × 1010 kg s−1 in (e), (f). Cases with weak forcing and uniform warm ocean, THF0 = 130 W m−2, are shown in (a) and (b). Cases with uniform warm ocean and global circulation, THF0 = 140 W m−2, are shown in (c) and (d). Cases with cross-equatorial circulation and summer-like ocean, THF0 = 140 W m−2, are shown in (e) and (f).

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 3.
Fig. 3.

Steady-state zonal mean fields (100-day mean) as a function of land surface forcing strength for cases with uniform warm ocean. Dashed line with squares represents the axisymmetric case; solid line with circles represents the three-dimensional case with zonally symmetric continent. (top) Absolute global maximum streamfunction strength of counterclockwise meridional cell. (bottom) Minimum 150-hPa absolute vorticity between 6° and 64°N.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 4.
Fig. 4.

Wave spectra for 1000-mb meridional velocity in case with symmetric continent, THF0 = 125 W m−2. Speed is in m s−1, and negative wavenumbers indicate westward propagation. (a) Spectra at 2°N, with contour interval of 3.0 × 10−10 m s−l × Δc−1, where Δc is the unit phase speed interval of 1.0 m s−1. (b) Spectra at 42°S, with contour interval of 6.0 × 10−8 m s−1 × Δc−1.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 5.
Fig. 5.

Composite of 1000-mb winds (arrows) and precipitation (shading, mm day−1), calculated in a frame of reference moving westward with the disturbance, for the case with zonally symmetric continent and THF0 = 130 W m−2. Mean is over days 327 to 387.5.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 6.
Fig. 6.

Hovmoeller diagram of precipitation at 18°N (mm day−1). First 100 days are transition to summer forcing state. (a) Case with zonally symmetric continent with summer SST profile, THF0 = 130 W m−2. (b) Case with asymmetric continent (land between 0° and 180°) and summer-like SST, THF0 = 140 W m−2.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 7.
Fig. 7.

Wave structure of persistent anomaly for case with zonally symmetric continent, THF0 = 130 W m−2. (a) Composite vertical velocity (ω), with contour interval of 0.1 Pa s−1, and dotted contours indicate ascent. (b) Composite eddy temperature, with contour interval of 1 K. (c) Eliassen–Palm flux (not composite); contours indicate divergence, with contour interval of 30 m2 s−2. Panels (a) and (b) represent a composite of a vertical slice at 18°N calculated in a frame of reference moving westward with the persistent anomaly over days 327.5 to 387.5.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 8.
Fig. 8.

Comparison of subcloud moist static energy for cases with and without eddies. Solid line shows 1000-hPa h (J) for axisymmetric case with uniform warm ocean, THF0 = 140 W m−2. Dot–dash line shows zonal mean 1000-hPa h for three-dimensional case with zonally symmetric continent and warm ocean, THF0 = 140 W m−2.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 9.
Fig. 9.

1000-hPa winds (m s−1) and precipitation (mm day−1), for case with asymmetric continent and THF0 = 140 W m−2. (a) Case with uniform warm ocean; (b) case with summer-like ocean SST; (c) case with summer-like ocean SST and thin walls at eastern and western coastlines.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 12.
Fig. 12.

Global zonal mean meridional circulation for three-dimensional cases with asymmetric forcing; contour interval is 5 × 109 kg s−1. (a) Aquaplanet case with ΔT = 2.25 K, SST maximum at 24°N, 50-day time mean. (b) Continental island case with THF0 = 180 W m−2, 100-day time mean.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 10.
Fig. 10.

1000-hPa winds (m s−1) and moist static energy (J), as in Fig. 9 for case with asymmetric continent and THF0 = 140 W m−2. (a) Case with uniform warm ocean; (b) case with summer-like ocean SST; (c) case with summer-like ocean SST and thin walls at eastern and western coastlines.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Fig. 11.
Fig. 11.

500-hPa ω for case with asymmetric continent and warm ocean, THF0 = 140 W m−2. Dotted contours indicate ascent; contour interval is 2.5 × 10−3 Pa s−1.

Citation: Journal of the Atmospheric Sciences 64, 5; 10.1175/JAS3917.1

Save