This study provides additional explanations for some of the experimental findings of Chao (2000) and Chao and Chen (2001, 2004) concerning the mechanisms responsible for the latitudinal location of the ITCZ (intertropical convergence zone) in an aqua-planet model. These explanations are given in terms of attractors that are caused by the earth’s rotation and sea surface temperature (SST) latitudinal distribution. These explanations are then used to explain the origin of some of the systematic errors in the GCM (general circulation model) simulation of ITCZ precipitation over the ocean, including the origin of the “double-ITCZ bias.”
When the influence of the SST distribution is excluded, the earth’s rotation, by itself, can generate either a single rotational ITCZ attractor at the equator or a double rotational ITCZ attractor at approximately 15° N&S, depending on the model physics. When a double rotational ITCZ attractor is generated, if it is excessively strong due to problems with the model physics, the addition of the attraction due to the SST distribution will not be able to turn a double ITCZ into a single ITCZ in the non-equinoctial seasons. This results in a “double-ITCZ bias,” with the component of the double ITCZ on the opposite side of equator from the SST latitudinal peak being stronger than the component on the same side of the equator as the SST latitudinal peak. On the other hand, when a single rotational ITCZ attractor over the equator is generated, a single ITCZ exists between the equator and the SST latitudinal peak (assuming there is only one SST latitudinal peak), which is too close to the equator year-round. In fact, to obtain a good ITCZ simulation, the model should have a double rotational ITCZ attractor of moderate strength.
ITCZ precipitation systematic errors are highly sensitive to the model physics, and by extension, the model horizontal resolution. A few possible methods of alleviating systematic errors in the GCM simulation of the ITCZ precipitation--such as rain re-evaporation, cumulus momentum transport, and an extra condition on the cumulus parameterization scheme--are discussed in the context of experiments using a recent version of the Goddard Modeling and Assimilation Office’s Goddard Earth Observing System (GEOS-5) GCM. One of the important findings is that the ITCZ precipitation systematic-error problem should be studied in conjunction with other problems in the GCM simulation, such as the problem in simulating the tropical wavenumber-frequency power spectrum.
The contribution of this paper is mainly conceptual. This paper’s findings, along with those of Chao (2000) and Chao and Chen (2001, 2004), contribute to building a theoretical foundation for ITCZ study that may lead to the eventual diminution of ITCZ precipitation systematic errors in GCMs.