Frontside halo coronal mass ejections (CMEs) are generally considered as potential candidates for producing geomagnetic storms, but there was no definite way to predict whether they will hit the Earth or not. Recently Moon et al. suggested that the degree of CME asymmetries, as defined by the ratio of the shortest to the longest distances of the CME front measured from the solar center, be used as a parameter for predicting their geoeffectiveness. They called this quantity a direction parameter, D, as it suggests how much CME propagation is directed to Earth, and examined its forecasting capability using 12 fast halo CMEs. In this paper, we extend this test by using a much larger database (486 frontside halo CMEs from 1997 to 2003) and more robust statistical tools (contingency table and statistical parameters). We compared the forecast capability of this direction parameter to those of other CME parameters, such as location and speed. We found the following results: (1) The CMEs with large direction parameters (D ! 0:4) are highly associated with geomagnetic storms. (2) If the direction parameter increases from 0.4 to 1.0, the geoeffective probability rises from 52% to 84%. (3) All CMEs associated with strong geomagnetic storms ( Dst À200 nT) are found to have large direction parameters (D ! 0:6). (4) CMEs causing strong geomagnetic storms (Dst À100 nT), in spite of their northward magnetic field, have large direction parameters (D ! 0:6). (5) Forecasting capability improves when statistical parameters (e.g., ''probability of detection -yes'' and ''critical success index'') are employed, in comparison with the forecast solely based on the location and speed of CMEs. These results indicate that the CME direction parameter can be an important indicator for forecasting CME geoeffectiveness. Subject headingg s: solar-terrestrial relations -Sun: coronal mass ejections (CMEs)

Abstract:We investigate two abnormal CME-Storm pairs that occurred on 2014 September 10 -12 and 2015 March 15 -17, respectively. The first one was a moderate geomagnetic storm (Dst min ∼ -75 nT) driven by the X1.6 high speed flare-associated CME (1267 km s −1 ) in AR 12158 (N14E02) near solar disk center. The other was a very intense geomagnetic storm (Dst min ∼ -223 nT) caused by a CME with moderate speed (719 km s −1 ) and associated with a filament eruption accompanied by a weak flare (C9.1) in AR 12297 (S17W38). Both CMEs have large direction parameters facing the Earth and southward magnetic field orientation in their solar source region. In this study, we inspect the structure of Interplanetary Flux Ropes (IFRs) at the Earth estimated by using the torus fitting technique assuming self-similar expansion. As results, we find that the moderate storm on 2014 September 12 was caused by small-scale southward magnetic fields in the sheath region ahead of the IFR. The Earth traversed the portion of the IFR where only the northward fields are observed. Meanwhile, in case of the 2015 March 17 storm, our IFR analysis revealed that the Earth passed the very portion where only the southward magnetic fields are observed throughout the passage. The resultant southward magnetic field with longduration is the main cause of the intense storm. We suggest that 3D magnetic field geometry of an IFR at the IFR-Earth encounter is important and the strength of a geomagnetic storm is strongly affected by the relative location of the Earth with respect to the IFR structure.