Single and double ITCZ in an aqua-planet model with constant sea surface temperature and solar angle

Abstract

It has been known for more than a decade that an aqua-planet model with a globally- and temporally-uniform sea surface temperature and solar isolation angle can generate intertropical convergence zones (ITCZ). Employing such a model, previous studies have shown that one of several means can be used to change between a single ITCZ over the equator and a double ITCZ straddling the equator. These means include switching to a different cumulus parametrization scheme, making changes within the cumulus parametrization scheme, and changing other aspects of the model such as horizontal resolution. Here, an interpretation of these findings is offered. In an aqua-planet model with globally and temporally uniform sea surface temperature and solar isolation angle, the latitudinal location of an ITCZ is the latitude where a balance exists between two types of attraction, both resulting from the Earth’s rotation. The first attraction pulls the ITCZ towards the equator and is not sensitive to changes in model design. It is directly related to the Coriolis parameter, which provides stability to the atmosphere. The second ssattraction pulls the ITCZ poleward and is sensitive to changes in model design. It is related to the convective circulation, modified by the Coriolis force. A balance between the two types of attraction is reached either at the equator or more than 10° north and south of the equator, depending on the shape and magnitude of the attractions. A balance at the equator yields a single ITCZ over the equator, whereas a balance north and south of the equator yields a double ITCZ straddling the equator.

Appendix 1

1.1 Difficulties with previous theories of the latitudinal location of the ITCZ

The experimental findings enumerated in the Introduction present a good means to test the various theories of the latitudinal location of the ITCZ. Waliser and Somerville (1994) and Tomas and Webster (1997) have reviewed these theories. For example, Charney’s (1971) explanation for the latitudinal location of the ITCZ involves a balance of two components. One component is the efficiency of Ekman pumping (or frictionally induced low-level convergence), which favors a larger Coriolis parameter or location at the poles. The other component is the higher moisture content (or temperature) of the tropics, which favors a location at the equator. The second component does not exist under U-SST-SA conditions, therefore Charney’s (1971) explanation results in ITCZs at the poles under these conditions, in contradiction to our model results. Hence, one must conclude that Charney’s (1971) first component is either incorrect or incomplete. His first component is essentially derived from a zonally symmetric version of the conditional instability of the second kind (CISK) theory for tropical cyclogenesis. The exact theoretical reason(s) for failings of the CISK theory is still not a settled question, although some very persuasive discussions have been presented (Emanuel et al. 1994).

Holton et al. (1971) also based their theory to explain the latitudinal location of ITCZs on the efficiency of Ekman pumping. Instead of using zonal symmetry, they considered the fact that the ITCZs consist of many propagating synoptic waves. Their conclusion was that the Doppler-shifted frequency of the synoptic waves is equal to the Coriolis parameter. However, as pointed out by Lindzen (1974) regarding diurnal oscillation, there is no observational evidence of maximum precipitation intensity at 30° latitude, as predicted by Holton et al. (1971). Moreover, Holton et al. (1971) imply that at any latitude, there can only be one wave frequency. Observations, however, show that an ITCZ contains more than one wave frequency (Takayabu et al. 1996, their Fig. 3). Also, just as with Charney (1971) theory, the fact that under U-SST-SA conditions, switching cumulus convection schemes leads to switching locations by the ITCZ, cannot be explained by Holton et al. (1971), whose theory is not sensitive to different cumulus convection schemes.

Waliser and Somerville (1994) theory is also built on frictionally induced low-level convergence, and it has no provision to account for the dependence of the latitudinal location of the ITCZ on the cumulus convection scheme, either.

Lindzen (1974) wave-CISK theory, first discussed by Hayashi (1970, 1971), does not rely on Ekman pumping (or frictionally induced low-level convergence), but there is also no provision to account for the dependence of the ITCZ latitudinal location on the cumulus convection scheme.

Tomas et al. (1999) work, is an extension of that of Tomas and Webster (1997), and emphasizes the role of the cross-equatorial surface pressure gradient in determining the latitudinal location of the ITCZ. Although they realized from observations that there are two modes of ITCZ latitudinal locations (equatorial and off-equatorial), which correspond to different magnitudes of the cross equatorial surface pressure gradient, how these two modes arise is not explained. Since the cross-equatorial surface pressure gradient is a model-produced quantity, and cannot be specified externally, their model, which is relaxed to a specified cross-equatorial surface pressure gradient, does not lead to a self-contained theory to account for the latitudinal location of the ITCZ. Also, their article does not discuss the sensitivity of ITCZ location to the cumulus convection scheme.

In summary, the theories described here, when applied to U-SST-SA conditions, do not explain or predict the dependence of the latitudinal location of the ITCZ on the cumulus convection scheme. Other theories of the ITCZ, such as those of Xie and Saito (2001) and Philander et al. (1996), invoke continental distribution and ocean–atmosphere interaction, respectively, and are thus not applicable to the problem at hand.