What Controls the Magnetic Field Configuration of Interplanetary Coronal Mass Ejections ?

In this work we address the question of the classification of interplanetary coronal mass ejections (ICMEs): magnetic cloud (MC) or ejecta (EJ). Using 186 shock-associated ICMEs from 1997 to 2006, we have examined three possible causes : (1) magnetic complexity characterized by sunspot number, (2) CME direction characterized by CME angular distance (the angle between the CME cone axis and the sky plane), and (3) CME-CME interaction characterized by the number of halo CMEs. First, the annual fraction of MC is poorly anti-correlated (R=-0.36) with annual sunspot number. Second, more than half of the CMEs that originated near the central meridian produced EJs and the distribution of CME angular distance for 38 EJs is not much different from that for 16 MCs. Third, the annual fraction of MC is well anti-correlated (R=-0.78) with the annual number of halo CMEs. In addition, we also searched for candidate of interacting CMEs according to temporal and spatial closeness by considering all halo CMEs during the same period. As a result, we find that the annual fraction of interacting CMEs is well correlated (R=0.87) with the annual number of the halo CMEs as well as anti-correlated (R=-0.85) with the annual fraction of MCs. The contingency table between CME-CME interaction and MC occurrence also shows a good statistical result (Prediction of detection 'yes' is 0.88, and Critical Success Index is 0.62), which is better than that for the halo CME-storm relationship. Our results imply that the CME interaction is mainly responsible for their observed structure (MC or EJ) in the interplanetary medium.