y than fast ones. This is consistent with the observational results obtained by Zhang et al. (2003), that all four Earth-encountered limb CMEs originated from the east. On the other hand, since the most of the EFHCMEs are fast events, the range of the longitude distribution given by the theoretical model is E40◦, W70◦, which is well consistent with the observational results (E40◦, W75◦)

The shock compression of the preexisting southward directed magnetic field can enhance a geomagnetic disturbance. A simple theoretical model is proposed to study the geoeffectiveness of a shock overtaking a preceding magnetic cloud. Our aim is to answer theoretically the question how deep the shock enters into the cloud when the event just reaches the maximum geoeffectiveness. The results suggest that the minimum value of Dst* decreases initially, then increases again while the shock propagates from the border to the center of the cloud. There is a position where the shock compression of the preceding cloud obtains the maximum geoeffectiveness. In different situations, the position is different. The higher the overtaking shock speed is, the deeper is this position, and the smaller is the corresponding Dst∗min. Some shortcomings of this theoretical model are also discussed.

d in daily averages of Ap index for geomagnetic disturbances from the World Data Center (WDC) at the International Association for Geomagnetism and Aeronomy (IAGA) are also examined for the same four-year time span. By Fourier power spectral analyses, the CME data appears to contain significant power peaks at periods of ～358±38, ～272±26, ～196±13 days and so forth, while except for the ～259±24-day period, X-ray solar flares of class>～M5.0 show the familiar Rieger-type quasi-periods at ～157±11, ～122±5, ～98±3 days and shorter ones until ～34±0:5 days. In the data of daily averages of Ap index, the two significant peaks at periods ～273±26 and ～187±12 days (the latter is most prominent) could imply that CMEs (periods at ～272±26 and ～196±13 days) may be proportionally correlated with quasi-periodic geomagnetic storm disturbances; at the speculative level, the ～138 ±6-day period might imply that X-ray flares of class>M5.0 (period at ～157±11 days) may drive certain types of geomagnetic disturbances; and the ～28±0:2-day periodicity is most likely caused by recurrent high-speed solar winds at the Earth's magnetosphere. For the same three data sets, we further perform Morlet wavelet analysis to derive period-time contours and identify wavelet power peaks and timescales at the 99 percent confidence level for comparisons. Several conceptual aspects of possible equatorially trapped Rossby-type waves at and beneath the solar photosphere are discussed.

A multiple-magnetic-cloud (Multi-MC) structure formed by the overtaking of two successive coronal mass ejections(CMEs) in the heliosphere is studied by using a 2.5-D MHD simulation. This simulation illustrates the process of the formationand propagation of two identical CMEs, which are ejected with speeds of 400km s−1 and 600 km s−1 respectively and initiallyseparated by 12h. The results show that it takes ～18h for the fast cloud to catch up with the preceding slow one, thenthe two clouds form a Multi-MC structure that arrives at 1 AU three days later. The fast cloud is slowed down significantlybecause of the blocking by the preceding slow one. This implies that the travel time of a Multi-MC structure is dominatedby the preceding slow cloud. Moreover, most primary observational characteristics of Multi-MC at 1 AU are well representedby the simulation. In addition, by combining observations, theoretical model and the simulation results, differences betweenMulti-MC and other types of in-situ observed double-flux-rope structure are addressed. A comparison of Multi-MC to coronalmass-ejection cannibalization near Sun is also given.