Appendix C. Comparison with Numaguti and Hayashi (1991)

Here, we present comparison between the results of present study and those of Numaguti and Hayashi (1991), which is a previous study with almost the same setup. Both study perform numerical calculation with Kuo scheme, and the complexity of the models are at the similar level. As will be summrized below, we could not duplicate the results of the previous study in the following two points. We have to confess, at the moment, that we do not figure out the causes of the difference between the results of present study and those of the previous one.

However, actual models utilized are different; the model of Numaguti and Hayashi (1991) is, the same as HS86, the earliest version of numerical weather prediction model of Japan Meteorological Agency, while the model of the present study is a rewritten model based on that of HS86. The basic architecture of the dynamical process is the same, but there are some difference in the physical processes, and moreover there may be some difference in the values of parameters such as radiation properties.

The most important difference is the profile of SST (sea surface temperature). The profile of SST in Numaguti and Hayashi (1991) is the same as that of HS86, which is calculated as the north-south symmetric component of the zonally averaged observed real SST profile, while that of the present study is, as shown in Fig.1.2, artificial.

Longitude-time section of precipitation along the equator

Compared with the result of the corresponding experiment in Numaguti and Hayashi (1991), case kuo-c or case kuo-d in the present study is characterized with stronger wavenumber one signal and smaller number of sequences of grid-scale precipitation events. We have performed aditional simulations with various values of absorption coefficient of radiation, we have not succeeded in reproducing a similar longitude-time section of precipitation along the equator as that of the previous study.

Fig.C.1: Time-longitude cross sections of precipitation along the equator. Unit is [mm day-1]. Contour indicates 2 [mm day-1]. The region of precipitation greater than 10 [mm day-1] is hatched. (left) Numaguti and Hayashi (1991) fig.3b, (right) Case kuo-c. The same as Fig.3.4(lower left) but is adapted to the same drawing as Numaguti and Hayashi (1991) fig.3b.

Composite structure associated with grid-scale precipitation events

Numaguti and Hayashi (1991) argue that the wave-CISK dynamics contributes the maintainance of the eastward propagating grid-scale precipitation structure. The composite structure obtained in the present study referring to the eastward propagating grid-scale precipitation events is characterized by the westward phase tilt of temperature and wind fields, and temperature structure considerably dominated by the second vertical mode with two peaks in the upper and the lower troposphere, which confirm the argument of the previous study. On the other hand, there are some features different from those in Numaguti and Hayashi (1991), such as a strong downward motion in the upper level just to the east (at around 190E) of the precipitation region, and an intense warm area in the upper level to the west, etc.

Fig.C.2: Composite structure referring to the eastward propagating grid-scale precipitation events. Longitude height cross section of temperature anomaly at the equator from the longitudinal mean (unit is [K]). Arrows indicate wind velocity anomaly from the longitudinal mean (unit is [m s-1, s-1]). (left) Numaguti and Hayashi (1991)fig.8b. Contour interval is 0.2 [K], and negative region is hatched. The arrows at the lower left corner indicates ??? m/s, ?????/s. (right) case kuo-c, the same as Fig.3.10(upper right).