Abstract

The variability in a two-level nonlinear atmospheric model is examined. The model domain is spherical. The sole forcing is a zonally symmetric parameterization of the December-February insulation. An extended 1500 day run is carefully analyzed.

Despite the absence of any asymmetric forcing the model produces a large amount of low-frequency variance and very red temporal spectra in subtropical and polar latitudes. The spatial signature of these low-frequency disturbances is that of quasi-stationary Rossby wave trains. It is suggested that the ubiquity of quasi-stationary Rossby wave trains in this model with no stationary asymmetric forcing is a consequence of energy cascade from the scale of baroclinic instability. The inertial cascade of energy toward larger spatial scales that is characteristic of geotrophic turbulence is terminated at a wavenumber where wave dispersion becomes as important as advection. This is precisely the scale at which Rossby waves are stationary. Hence, the cascade of energy from the scale of baroclinic instability to larger scales deposits energy preferentially into quasi-stationary Rossby waves.

The tropical variance in this model is dominated by an abundance of mixed Rossby-gravity waves that are driven from the extratropics. Most extratropical waves are seen to be dissipated at their low-latitude critical line if they propagate into the tropics.

Abstract

A spectral analysis of winds analyzed and initialized at the European Centre for Medium-Range Weather Forecasts reveals an abundance of power in the 850 mb meridional wind along the equator with periods near four days. The power is mostly in the westward propagating component.

Using high-pass filtered data it is shown that the waves have westward phase and eastward group propagation relative to the mean wind. The longest wavelengths are found over the Pacific Ocean, while the shortest are found over the convectively variable regions of Indonesia, South America, and Africa. Mean phase speeds at 850 mb are positively correlated with the mean wind on the equator at 500 mb and below, and negatively correlated with the mean wind above that level. The effective advecting zonal wind of the disturbances seems to be the density weighted average of the lower troposphere.

The structure of the disturbances bears resemblance to the expected structure of an equatorially trapped mixed Rossby-gravity wave over the central Pacific and Atlantic oceans, although the anomalies, while statistically significant, are extremely small. The outgoing longwave radiation (OLR) pattern is consistent with the flow field, suggesting that the waves are not merely a model artifact. Over the Atlantic there is a mode well defined by the zonal wind at the equator, but the OLR pattern is not consistent. Over the far western Pacific, there is evidence of meridional propagation from Northern Hemisphere midlatitudes. North of the equator there is meridional propagation at every longitude.

The strongest disturbances are primarily confined to the lower half of the troposphere, but at many longitudes there is evidence of a weak first baroclinic-mode structure within the troposphere. North of the equator the structures are barotropic.

Effective equivalent depths are estimated by comparing dispersion characteristics with mixed Rossby-gravity dispersion curves. Where the assumption of a mixed Rossby-gravity mode is believed to be valid, the equivalent depths are found empirically to lie between 1–60 m.

Abstract

Onset of the Australian summer monsoon is identified each year (1957–87) using the wind and rainfall record at Darwin. Onset is defined as the first occurrence of wet, 850 mb westerly winds. Composites of atmospheric fields at stations in and about the Australian tropics are constructed relative to the onset date at Darwin.

The composite onset is accompanied by the development of a convectively driven, baroclinic circulation over northern Australia. Upper tropospheric easterlies expand about the equator and the subtropical jet shifts poleward at onset. This behavior is interpreted as a transient southerly shift of the local Hadley circulation concurrent with the development of an upper level anticyclone over northern Australia.

The composite onset coincides with the arrival of an eastward propagating convective anomaly. The anomaly originates in the southern Indian Ocean, propagates eastward at 5 m s⁻¹ and is detectable as far east as the date line. An eastward propagating zonal wind anomaly also is detectable at tropical stations east and west of Darwin. These features are indicative of the 40–50 day oscillation and thus the composite onset is concluded to coincide with the traversal of the oscillation across northern Australia. The composite onset is further shown to coincide with the first occurrence of the convectively active 40–50 day oscillation during each southern summer.

Abstract

The tropical intraseasonal (30–50 day) oscillation manifests itself in the Australian summer monsoon by a pronounced modulation of the monsoonal westerlies. These 30-50 day fluctuations of the monsoonal westerlies are coherent with rainfall and OLR across northern Australia. The OLR fluctuation originates in the Indian Ocean and systematically propagates eastward at 5 m s⁻¹, consistent with previous studies of the intraseasonal oscillation.

The detailed evolution of the intraseasonal oscillation of the monsoon is studied via composites of upper air data in and about the Australian tropics. During the summer periods 1957-87, 91 events were identified at Darwin, Australia. The composite oscillation at Darwin has a very deep baroclinic structure with westerlies extending up to 300 mb. The westerly phase lasts about ten days and lags a similar duration rainfall event by about four days. During the westerly phase, the upper troposphere is warm and the extreme lower troposphere is cool. This structure is consistent with midtropospheric latent heating and lower tropospheric cooling due to evaporation of falling rain. The magnitude of the composite oscillation at Darwin is about 5 m s⁻¹ in zonal wind, 1 m s⁻¹ in meridional wind, 0.5°K in temperature, 5 mm rainfall per day, and 10% in relative humidity. The oscillation at Darwin is readily traced as far west as Cocos Island and as far east as Pago Pago.

Above northern Australia, enhanced synoptic scale variability develops during the wet-westerly phase of the oscillation. Analysis of a single station record precludes documentation of the structure of these synoptic fluctuations. In the Northern Hemisphere midlatitudes, a wave train in 500 mb heights appears to emanate from the longitude of the Australian tropics during the wet-westerly phase. The magnitude of this wave train is only about 50 m while the wave train undergoes a systematic evolution as the tropical convective anomaly moves west to east, no sense of dispersion from a localized low-latitude heat source is evident.

Abstract

The signature of 4–5-day period Rossby–gravity waves is searched for in the tropical convection field across the Indian-Pacific oceans. The convergence/divergence field of these waves in the lower troposphere is anticipated to produce an antisymmetric fluctuation in tropical convection. Antisymmetric fluctuations of tropical convection are shown to exhibit a pronounced spectral peak at a 4–5-day period only during boreal fall and only within about 30° longitude of the date line. The peak amplitude occurs around 7.5° latitude. These fluctuations propagate westward at 15–20 m s⁻¹ with zonal wavelength of about 7000&–9000 km. The fluctuations of convection are coherent and out of phase with the equatorial meridional wind, which also exhibits a pronounced spectral peak at a 4–5-day period in the lower troposphere near the date line. The antisymmetric zonal wind also is strongly coherent with the antisymmetric convective fluctuations in this region. The horizontal distributions of the 4–5-day power and coherence of the winds and convection are consistent with that produced by a convectively coupled Rossby–gravity wave that is confined near the date line.

The localization of the convectively coupled Rossby–gravity wave activity near the date line during boreal fall is postulated to be due to the unique meridional distribution of sea surface temperature at this location. The equatorial minimum flanked by maxima at about 5°–10° latitude is thought to encourage antisymmetric convection, which interacts efficiently with Rossby–gravity waves. The fall maximum in convectively coupled Rossby–gravity wave activity is consistent with these unique sea surface temperatures occurring only during fall.

Abstract

Thermal forcing feedback is proposed to be an important mechanism in middle and high latitudes in determining the low-frequency variability of the stationary wave structure. The total diabatic heating in the atmosphere is not due solely to the fixed longitudinally varying heat sources but also depends on the flow field itself. As a first approximation to this complex process, a heat flux which is proportional to the low-level temperature of the atmosphere is incorporated into a multi-level. steady-state, linear primitive equation model on a sphere. It is shown that, for deep vertical distributions of middle- and high-latitude diabatic heating, the inclusion of this feedback significantly amplifies the local and remote response. For shallow vertical distributions of middle- and high-latitude heat sources, the significant increase of amplitude is confined to the local response while the remote response is damped. The remote response due to tropical forcing is dramatically damped regardless of the vertical distribution of the heating.

Abstract

Two questions related to the intraseasonal variability of tropical convection and circulation remain controversial. 1) To what degree is the convective component of the Madden–Julian oscillation (MJO) a standing oscillation? 2) Is the eastward propagating circulation anomaly of the MJO coherent with a standing oscillation in convection?

In an attempt to settle these issues, the authors undertake a series of statistical analyses of gridded outgoing longwave radiation and winds to quantify the magnitudes of the propagating and standing components of convection and their coherence with the propagating component of the circulation. They demonstrate that no dominant standing oscillation in convection can be identified. Instead, intraseasonal variability of convection is dominated by an eastward propagating mode, which the authors interpret as the convective signal of the MJO. This propagating component accounts for almost all of the convective variance that is coherent with the eastward propagating disturbance in the zonal wind, which is a traditional measure of the MJO. Analysis of synthetic time series illustrates that an impression of a standing oscillation in convection may come forth because of the modulation of the eastward propagating convective disturbance by an amplitude envelope with maxima in the eastern Indian and western Pacific Oceans and a minimum over the maritime continents.

Abstract

Interhemispheric differences of the Madden–Julian oscillation (MJO) are investigated in a linearized primitive equation model. Heating is prescribed from the observed life cycle of the MJO, in which anomalous convection is concentrated in the Eastern Hemisphere. The dynamical response in the Eastern Hemisphere has the form of a forced disturbance that involves Kelvin and Rossby components. These dynamical components propagate eastward along with the prescribed heating and have zonal wavenumber-2 structure, in accord with observed behavior. The behavior in the Eastern Hemisphere also resembles that emerging from frictional wave-CISK, in which heating follows autonomously from the circulation. When the prescribed heating collapses near the date line, the Rossby component dissipates. The Kelvin component, however, continues to advance across the Western Hemisphere, where it propagates at more than twice the speed of the disturbance in the Eastern Hemisphere. The disturbance in the Western Hemisphere, which likewise is in accord with the observed behavior, can be understood as the radiating response to transient heating confined to the Eastern Hemisphere.

Abstract

The spectral character of tropical convection is investigated in an 11-yr record of outgoing longwave radiation from the Advanced Very High Resolution Radiometer to identify interaction with the tropical circulation. Along the equator in the eastern hemisphere, the space–time spectrum of convection possesses a broad peak at wave-numbers 1–3 and eastward periods of 35–95 days. Significantly broader than the dynamical signal of the Madden–Julian oscillation (MJO), this quasi-discrete convective signal is associated with a large-scale anomaly that propagates across and modulates time mean or “climatological convection” over the equatorial Indian Ocean and western Pacific. Outside that region the convective signal is small, even though, under amplified conditions, coherence can be found east of the date line and in the subtropics. Having a zonal scale of approximately wavenumber 2, anomalous convection propagates eastward at some 5 m s⁻¹ and suppresses as well as reinforces climatological convection in the eastern hemisphere. The convective signal amplifies to a seasonal maximum near vernal equinox and, to a weaker degree, again near autumnal equinox, when climatological convection and warm SST cross the equator.

Contemporaneous records of motion from ECMWF analyses and tropospheric-mean temperature from Microwave Sounding Unit reveal an anomalous component of the tropical circulation that coexists with the convective signal and embodies many of the established properties of the MJO. Unlike anomalous convection, that dynamical signal extends globally around the Tropics. The anomalous circulation differs fundamentally between the eastern and western hemispheres. In the eastern hemisphere, subtropical Rossby gyres and zonal Kelvin structure along the equator flank the convective anomaly as it tracks eastward, giving the anomalous circulation the form of a “forced response.” In the western hemisphere, the dynamical signal is composed chiefly of wavenumber−1 Kelvin structure, which has the form of a “propagating response” that is excited in and radiates away from anomalous convection at some 10 m s⁻¹. Kelvin structure comprising the propagating response appears in 850-mb and 200-mb zonal winds even when the convective signal is absent, albeit with much smaller amplitude. In contrast, the signal in 1000-mb convergence appears only when accompanied by anomalous convection, which suggests that convergence in the boundary layer is instrumental in achieving strong interaction with the convective pattern.

Abstract

A composite life cycle of the Madden–Julian oscillation (MJO) is constructed from the cross covariance between outgoing longwave radiation (OLR), wind, and temperature. To focus on the role of convection, the composite is based on episodes when a discrete signal in OLR is present. The composite convective anomaly possesses a predominantly zonal wavenumber 2 structure that is confined to the eastern hemisphere. There, it propagates eastward at about 5 m s⁻¹ and evolves through a systematic cycle of amplification and decay. Unlike the convective anomaly, the circulation anomaly is not confined to the eastern hemisphere.

The circulation anomaly displays characteristics of both a forced response, coupled to the convective anomaly as it propagates across the eastern hemisphere, and a radiating response, which propagates away from the convective anomaly into the western hemisphere at about 10 m s⁻¹. The forced response appears as a coupled Rossby–Kelvin wave while the radiating response displays predominantly Kelvin wave features.

When it is amplifying, the convective anomaly is positively correlated to the temperature perturbation, which implies production of eddy available potential energy (EAPE). A similar correlation between upper-tropospheric divergence and temperature implies conversion of EAPE to eddy kinetic energy during this time. When it is decaying, temperature has shifted nearly into quadrature with convection, so their correlation and production of EAPE are then small. The same correspondence to the amplification and decay of the disturbance is mirrored in the phase relationship between surface convergence and anomalous convection. The correspondence of surface convergence to the amplification and decay of the convective anomaly suggests that frictional wave–CISK plays a key role in generating the MJO.