A full halo CME was first detected with the SOLWIND coronagraph on 1979 November 27 and interpreted as a broad shell or bubble of dense plasma headed directly toward (or away from) the Earth (Howard et al., 1982). A broad interplanetary disturbance and a small geomagnetic storm produced by this halo CME was detected (Jackson, 1985; Webb and Jackson, 1990), supporting its interpretation as an Earthward-directed CME. Halo CMEs had only infrequently been reported in coronagraph observations of the Sun before the Solar and Heliospheric Observatory (SOHO) mission. Now, because of the high sensitivity and wide dynamic range of the Large Angle Spectrometric Coronagraph (LASCO), they can routinely be observed with SOHO/LASCO C2 and C3 coronagraphs (Plunkett et al., 1998).
CMEs are now known to be a key causal link among solar eruptions, major interplanetary disturbances and geomagnetic storms (Gosling et al., 1991; Kahler, 1992). The source regions of the halo CMEs can be studied in detail with cotemporal EIT (Thompson et al., 1998), soft X-ray (Hudson et al., 1998) and other observations of solar surface activity near the Sun's center. The existence of the associated surface activity may be used to determine whether the halo CME came from the frontside or the backside of the Sun (e.g., Webb et al., 2000a). Full halo CMEs are believed to originate less than 30 degrees of disk center of the Sun (Brueckner, G.E., et al., 1998; Zhao et al., 2001b). Thus, a frontside full halo CME with an average value of angular size must be Earth-directed, suggesting the launch of a potentially geoeffective disturbance toward the Earth. In fact, the 1997 January 6 full halo CME was used to successfully predict a geomagnetic storm. It should be noted that frontside halo CMEs, especially frontside partial halo CMEs, are not necessarily Earth-directed. The existence of an associated interplanetary disturbance near the Earth several days later may be used to determine if the halo CME was Earth-directed (e.g., Webb et al., 2000a).
Recently, it was found that all six Earth-directed halo CMEs observed by LASCO coronagraphs from December 1996 to June 1997 were associated with shocks, magnetic clouds, and moderate geomagnetic storms at the Earth 3-5 days after leaving the Sun. This suggests that magnetic cloud-like structures are a general characteristics of halo CMEs, at least during the early rise of the sunspot cycle, and also that all such frontside halo CMEs are geoeffective (Webb et al., 2000a).
Observations of halo CMEs later during the ascending phase, however, suggest that not all frontside halo CMEs produce ejecta near Earth or generate geomagnetic storms (St. Cyr et al., 2000; Cane et al., 2000). In this study, we examine the relationship between frontside full halo CMEs and geomagnetic storms from mid 1996 to the end of 1999 and find that the geoeffectiveness of the frontside full halo CMEs near sunspot maximum is weaker than near sunspot minimum. Questions to be addressed for understanding the possible cause(s) of CMEs in general and the solar-cycle effect in the geoeffectiveness of frontside halo CMEs in particular include: Does the magnetic field topology of all CMEs have the configuration of a magnetic flux rope? Would any limb CME despite its visual appearance be a halo CMEs if its source region were located near the Sun's disk center? Is the flux rope configuration a general characteristic of a halo CME at any time during the solar cycle? What is the relationship of the geoeffectiveness of halo CMEs to the location of their source regions with respect to the coronal base of the heliospheric current sheet.
In this study we locate the source regions of frontside full halo CMEs near the solar surface and study the magnetic configurations surrounding the source regions. We find that the solar cycle effect in the geoeffectiveness of frontside full halo CMEs is associated with the cyclic variation in the inclination in the base of the heliospheric current sheet. In the next section we discuss the analysis and results starting with the criteria used in associating full halo CMEs with surface activity and with geomagnetic storms, and the basic statistical results. Then we show the solar cycle variation of large-scale closed structures in the solar corona using the potential field-source surface coronal magnetic field model and synoptic magnetogram maps. Finally, we present the source locations of all the frontside full halo CMEs in terms of these maps. In the last section we summarize and discuss these results.
During the 3.5-year period from mid of 1996 to the end of 1999, the LASCO coronagraphs observed a total of 69 full halo CMEs. To identify which of the full halo CMEs were from the frontside, and to determine whether or not the frontside full halo CMEs were geoeffective, the associations of each full halo event with solar surface activity, interplanetary disturbances and geomagnetic storms were estimated. The criteria used to make these associations were described by Webb et al. (2000a, b). The solar surface activity considered to be associated with CMEs include flares and disappearing filaments in H_\alpha observations, long-duration flares, post-eruption arcade formation, depletions or `dimmings' of the coronal intensity, and bright wavefronts propagating quasi-radially from the source region in EUV observations (Thompson et al., 1998). The signatures for CME-associated interplanetary disturbances (now called ICMEs) are transient interplanetary shocks, bidirectional streaming of electrons and protons, and magnetic clouds, i.e., enhanced magnetic field strength and smooth rotation of the field orientation (e.g., Neugebauer and Goldstein, 1997). The solar wind observations were primarily from the Wind spacecraft with additional data from ACE after its operations began in August 1997.
Table 1
________________________________________________________________________ ___________________1996_____1997______1998_______1999_______1996-1999 ________________________________________________________________________
Total________________3_______17_________26________23___________69 ________________________________________________________________________
Frontside___________0________9_________15________11_____________35 ________________________________________________________________________
Associated storms__0____8+1(?)______6+6(?)_____4+1(?)_____18+8(?) ________________________________________________________________________
Fraction________________0.89-1.00(?)___0.40-0.80(?)__0.36-0.45(?)__0.51-0.74(?) ________________________________________________________________________
Among the 69 full halo CMEs, Webb et al. estimate that the source regions of 35 full halo CMEs were located on the frontside of the Sun. Table 1 lists the number of full halo CMEs observed each year from the mid 1996 to the end of 1999. The bottom two rows in Table 1 show the number of associated geomagnetic storms and its fraction with respect to the number of frontside full halo CMEs. The number followed by a question mark means that for those events the correspondence between the geomagnetic storms and the frontside halo CMEs are less certain. The fraction decreases from 0.89--1.00(?) in 1997 through 0.40--0.80(?) in 1998 to 0.36--0.44(?) in 1999 as sunspot number increases. Thus, almost all of frontside full halo CMEs during the early rise phase of the cycle are associated with geomagnetic storms, but less than half of frontside full halo CMEs near the sunspot maximum phase appear to cause geomagnetic storms. The last column of Table 1 shows the total number of the frontside full halo CMEs during the 3.5 years, the total number of the associated geomagnetic storms and the fraction between the two total numbers. The fraction of 0.51-0.74(?) is greater than 0.25 obtained for both all halo CMEs in the same period of time (Cane et al., 2000), implying that the chance for partial halo CMEs to generate storms is much less than the chance for full halo CMEs. It is understood because the chance for partial halo CMEs to encounter the Earth would be less than for full halo CMEs.
The existence of a solar cycle effect in the geoeffectiveness of frontside full halo CMEs implies that factors other than the mere occurrence of a CME directed toward Earth are important in determining the level of geomagnetic activity. It has been shown that near the maximum of cycle 21 during 1978-1979 most fast CMEs that struck the magnetosphere did not produce intense storms, because they did not have sustained periods of strong southward field (Tsurutani et al., 1988). The geoeffective solar wind structures include long intervals of large southward interplanetary magnetic field, or Bs "events" (Tsurutani and Gonzalez, 1997; Zhao et al., 1993). That there is a solar cycle effect thus suggests that almost all frontside full halo CMEs during the rising phase can produce Bs events, but near sunspot maximum less than half can do so.
CMEs are believed to originate in large-scale closed field regions with sufficient free magnetic energy to drive the CME material outward against the magnetic tension force of closed structures and solar gravity. To try to understand the cause of the solar cycle effect of full halo CMEs between the minimum and maximum phases in generating geomagnetic storms, we examine the solar cycle variation of open and closed field regions on the Sun.
It is well known that coronal open field regions or coronal holes near sunspot minimum are located at the polar regions and have opposite magnetic polarities. The boundary layer between the opposite-polarity open field lines beyond the height of the cusp points of coronal helmet streamers is the heliospheric current sheet. The base of the heliospheric current sheet near the minimum phase is approximately parallel to the solar equator. Another way of saying this is that the rotational and magnetic poles are nearly aligned. As solar activity increases, lower- latitude coronal holes appear and the polar holes shrink, and the latitude extent of the base of the heliospheric current sheet increases. Around sunspot maximum, the polar holes have disappeared and the base of the heliospheric current sheet is nearly orthogonal to the solar equator. Both the coronal holes and the base of the heliospheric current sheet can be successfully reproduced using the potential field-source surface coronal field model and the synoptic magnetogram maps over one solar rotation (Hoeksema, 1991; Wang et al., 1996; Zhao et al., 1999). Here the source surface is the spherical surface located near the height of the cusp points above which all magnetic field lines are assumed to be radial.
Figure 1 displays four panels for Carrington Rotations (CR) 1921, 1935, 1949, and 1961 corresponding to April or May of 1997, 1998, 1999, and 2000, respectively. The blue (red) dotted areas denote the foot-points of the outward (inward) open field lines near the solar surface, computed using the potential field-source surface model. The black thick lines are computed neutral lines on the source surface at 2.5 solar radii, denoting the base of the heliospheric current sheet, i.e., the boundary layer between opposite- polarity open field lines. Figure 1 indicates the cyclic variation of the open field regions and the base of the heliospheric current sheet mentioned above. Overploted on the four panels are lines consisting of blue and red segments which denote lower-lying closed field lines, i.e., those with their apexes lower than 1.25 solar radii. The blue and red segments for each line intersect at its apex. The adjacent blue-red lines are rooted in bipoles. The thin black lines are closed field lines with their apexes higher than 2.0 solar radii. Their foot points are located near the open field regions.
The large-scale closed field region with a spatial scale similar to that of coronal helmet streamers is here defined as the area between two open field regions with opposite or identical magnetic polarity. In the top-left panel, for April 1997, all large-scale closed field regions are located between two polar open field regions with opposite polarity. These regions thus represent coronal helmet streamers near solar minimum in April 1997. As shown in the panel, most of these streamers overlie three bipoles with a few having one or 5 bipoles. The apexes of their outermost closed field lines is mostly higher than 2 solar radii. In the top-right panel, for April 1998, there are large-scale closed field regions occurring between two open field regions with identical magnetic polarity as well as closed field regions consisting of odd-number bipoles. For instance, the closed field region located in the longitude range of 70 -- 90 and latitude range of 30S -- 60S occurred between two open field regions with inward (red) polarity. The new kind of closed field region consists of two bipoles. The closed field region between two outward-polarity open field regions near longitude of 360 degrees contains 4 bipoles.
The closed field regions with even-number bipoles increase as solar activity increases. In the bottom-right panel for April 2000 the even-number-bipole closed field regions occupy most of closed field areas. Examples are the closed field regions between two outward-polarity open field regions in the left half and between two inward-polarity open field regions in the right half. As opposed to the odd-number-bipole closed field regions, there is no outermost closed field line that confines all bipoles for the even-number-bipole closed field regions. The open field lines adjacent to the even-number-bipole closed field regions are in identical polarity and those adjacent to the odd-number-bipole closed field regions are in opposite polarity.
We note that the large-scale closed field regions associated with the source regions of CMEs cannot be potential because of the existence of a sufficient amount of free magnetic energy there. The field configuration above an individual bipole, for instance the bipole pointed by the arrow in the top-left panel, may be different from what shown because of the twist of the field line. The number of bipoles within each closed field region for the non-potential-like field may still be expected to be the same as that for the potential-like field due to the identical polarity distribution of the photospheric field in the area (Zhao and Hoeksema, 2000).
Observations have shown that most limb CMEs originate in coronal helmet streamers during the declining and minimum phases of the solar cycle (Hundhausen, 1993; Zhao and Hoeksema, 1996). Thus, these CMEs should overlie odd-number bipole closed field regions. On the other hand, we might expect that even-number bipole closed field regions would be the most common kind of source region for CMEs near sunspot maximum.
It is generally assumed that the source regions of CMEs are associated with solar surface activity as mentioned earlier. The surface locations of flares (and their associated active regions) and erupting prominences associated with CMEs are usually offset from the axis of the CME (Harrison, 1986; Webb, 1992). Thus, the `source location' of a CME feature can be defined to lie midway between the outer edges of the CME in a coronagraph image (e.g., Plunkett et al., 2000). For a full halo CME, its source location is expected to be located between the associated surface activity and the Sun's disk center (Zhao et al., 2001). The green `+'symbol in the first three panels of Figure 1 denotes the Sun's disk center determined using the Carrington longitude and the B angle at the onset time of the observed frontside full halo CMEs. The onset time of the full halo CME is shown at the top of each panel. The green`*' symbol adjacent to the `+' symbol denotes the location of the associated solar surface activity, determined as the angular distance between the Carrington longitude and the latitude of the observed surface activity and Sun center. All three pairs of green symbols are located within odd-number-bipole closed field regions under black closed field lines, suggesting that the source location of these full halo CMEs are all located in odd-number-bipole closed field regions, i.e., coronal helmet streamers, even near sunspot maximum.
All 35 frontside full halo CMEs (Table 1) are examined in the same way as was done for the 3 events in Figure 1. Figure 2 displays 35 panels with the 35 frontside full halo CMEs. The onset time of each event is shown at the top of each panel. As compared to Figure 1, the closed field lines are excluded for clarity. There are two `*' in one panel for some of the events, indicating the difficulty in determining a single association for the full halo CME or the possibility of a multiple association. There are a few events where the associated surface activity (i.e., for the events with onset time of 98:06:05_12:01, 98:12:18_18:34, 99:06:30_11:54) or the associated disk center (i.e., for the event with onset time of 99:12:06_09:30) are located in even-number-bipole closed field regions. However, all of the source locations of the CMEs located between the associated solar surface activity and disk center are within two opposite-polarity open field regions and close to the base of the heliospheric current sheet.
Observations of full halo CMEs and their associated solar surface activity, interplanetary disturbances and geomagnetic storms between 1997 and 1999 indicate that the fraction of frontside full halo CMEs causing magnetic storms in 1997, 1998 and 1999 range over 0.81--1.00(?), 0.40--0.80(?), and 0.36--0.45 (?), respectively. This suggests the existence of a solar-cycle effect in the geoeffectiveness of frontside full halo CMEs.
All of the frontside full halo CMEs observed in the first half of solar cycle 23, from near sunspot minimum through the ascending phase to sunspot maximum originated in coronal helmet streamers, i.e., the large-scale closed field regions lying between open field regions with opposite polarity. It has been shown that these closed field regions usually contain three bipoles, with the apexes of their outermost closed field line at a height higher than 2.0 solar radii. The source locations of the full halo CMEs is always close to the base of the heliospheric current sheet.
It has been shown that magnetic cloud-like structures are a general characteristic of halo CMEs during the early rise from sunspot minimum. Based on the similar orientation between clouds and associated filaments and the low plasma \beta of magnetic clouds, interplanetary magnetic clouds have been associated, respectively, with erupting solar filaments (Marubashi, 1986) and dark cavities (Tsurutani and Gonzalez, 1995) of three-part-structure CMEs originating from helmet streamers. On the other hand, it has been suggested that magnetic clouds are oriented approximately parallel to the base of the heliospheric current sheet (Zhao and Hoeksema, 1996; Crooker et al., 1998; Mulligan et al., 1998). The fact that all of the full halo CMEs originated in the same kind of closed field regions (coronal helmet streamers) implies that all full halo CMEs are magnetic flux ropes with their orientations parallel to the orientation of such observed structures as the coronal cavity or/and filament which are near the base of helmet streamers. Since the inclination of the base of the heliospheric current sheet is expected to be approximately parallel to the orientation of the underlying cavity or filament, the orientation of the halo CMEs may be estimated by the inclination of the associated heliospheric current sheet.
Because the source locations of most, if not all, of the observed full halo CMEs were located near the base of the heliospheric current sheet and because the inclination of the base of the heliospheric current sheet is parallel near sunspot minimum and orthogonal near sunspot maximum to the solar equator, the orientation of the associated interplanetary magnetic clouds/flux ropes will be approximately parallel near sunspot minimum and orthogonal near sunspot maximum to the ecliptic plane. This is in agreement with the solar cycle variation in the inclinations of flux ropes found in Pioneer-Venus Orbiter data by Mulligan et al. (1998), and in an empirical model outlined by Crooker (2000).
It has been shown recently that the duration and intensity of the southward IMF component within a magnetic cloud, the key coupling component to the magnetosphere, significantly depend on such characteristic parameters of the magnetic flux rope as the orientation and strength of the central axial field, the bulk velocity, and the impact distance between the rope's axis and the Sun-Earth line. The orientation of magnetic clouds is one of the most important parameters that determine the duration and intensity of the southward IMF component (Zhao and Hoeksema, 1998; Zhao et al., 2001). Specifically, there is always an interval of southward IMF component when the orientation of clouds is parallel to ecliptic plane, but there is only about a 50% chance to have a Bs event when the orientation is orthogonal to the ecliptic plan. Thus, the solar-cycle effect in geoeffectiveness of halo CMEs may be caused, at least partially, by the cyclic variation of the inclination of the base of the heliospheric current sheet.
The angular size of CMEs has been suggested to be another factor that determines whether or not a frontside halo CME can encounter the Earth (Cane et al., 2000). The fact that the geoeffectiveness of full halo CMEs in 1999 is less than 0.5 could suggest that more halo CMEs have smaller angular size in sunspot maximum than in sunspot minimum. The geometrical model of the halo CME as a cone has been used to estimate the angular size and true orientation of the central axis of the expanding halo CME (Zhao et al., 2001), providing a way to determine whether or not a frontside halo CME is a Earth-encountering CME on the basis of LASCO observations.
It is often implicitly assumed in many papers about halo CMEs that any limb CME, even those with an unusual appearance, would become a halo CME if their source region were located near disk center, so that a frontside halo CME is generally be called an Earth-directed CME, instead of an Earth-directed halo CME. Halo CMEs are interpreted as a broad shell or bubble of dense plasma ejected nearly along the line-of-sight and expanding outward from the Sun. There have been many kinds of morphological classifications for limb CMEs, such as Loop/Cavity, Cavity (but no loop), Core, Mound, Blob, Jet, and Tongue (Hundhausen, 1999 and the references therein). A 3-D shell or bubble model can be used to interpret the loop shape of limb CMEs, but cannot be used to interpret other classes such as Core, Mound, Jet. This implies that not all limb CMEs would appear as halo CME if they originated near Sun's disk center. We suggest that a flux-rope configuration of the inner magnetic field in halo CMEs is perhaps the necessary condition for a CME from near Sun's disk center to form a halo CME. The 3-D shell or bubble model is actually implicit in the characterization of mass ejections as disruptions of an arcade of closed field lines (a cavity) above a magnetic neutral line (usually a prominence) (Hundhausen, 1999). This may be why only odd-number-bipole closed field regions that have outermost closed field lines to confine inner bipoles are the source region of halo CMEs. LASCO has observed a great deal of internal structure in many limb CMEs. Concave-outward structures that can be interpreted as the back of magnetic flux ropes are observed in approximately one third of all events (Dere et al., 1999; St. Cyr et al., 2000). This ratio is the same as the ratio of the number of magnetic clouds at 1 AU to the number of bidirectional electron heat flux events (BDEs), perhaps the most common of ICME signatures (Gosling, 1990). Therefore, only one third of all limb CMEs can become halo CMEs when they are located near the Sun's disk center if the rope configuration is the necessary condition for a CME to be a 3-D shell or bubble of dense plasma.
There are no outermost closed field lines that confine the inner bipoles of even-number-bipole closed field regions. The free magnetic energy that is necessary for generating CMEs in an even-number-bipole system would be less than that in an odd-number-bipole system where the outermost closed field lines must first be opened up before lifting the inner structure. In addition, the area occupied by the even-number-bipole system near sunspot maximum is greater than that by the odd-number-bipole system. It is suggested that the even-number-bipole system is the major kind of source region for CMEs other than the magnetic cloud-like CMEs (halo CMEs near the disk center and concave-outward CMEs near the limb). The ISEE 3 electron heat flux data have shown the existence of the intrasector BDEs with polarities matching those of the sectors, and those BDEs are indeed not closely confined to the heliospheric current sheet(Kahler et al., 1999). The study of the association of limb CMEs with even-number-bipole closed field regions is in progress.
Brueckner, G. E., et al., Geomagnetic storms caused by coronal mass ejections (CMEs): March 1966 through June 1997, Geophys. Res. lett., 25, 3019, 1998
Cane, H.V., I.G. Richardson, and O.C. St. Cyr., Coronal mass ejections, interplanetary ejecta, and geomagnetic storms, Geophys. Res. Lett., 27 , 3591, 2000.
Crooker, N.U., Solar and heliospheric geoeffective disturbances, J. Atmos. Solar-Terr. Phys., 62, 1071, 2000.
Crooker, N. C., et al., Magnetic clouds at sector boundaries, J. Geophys. Res., 103 , 301, 1998.
Dere, K. P., et al., LASCO and EIT observations of helical structure in coronal mass ejections, Astrophys. J., 516, 465, 1999.
Gosling, J. T., et al., Geomagnetic activity associated with earth passage of interplanetary shock disturbances and coronal mass ejections, J. Geophys. res., 96, 7831, 1991.
Harrison, R. A., Solar coronal mass ejections and flares, Astron. Astrophys., 162, 283, 1986.
Hoeksema, J. T., Large-scale structure of the heliospheric magnetic field: 1976-1991, Adv. Space Res., 9, 15, 1991.
Howard, R. A., et al., The observation of a coronal transient directed at Earth, Astrophys. J., 263, L101, 1982.
Hudson, H. S., et al., Soft X-ray signatures of coronal mass ejections, in Coronal Mass Ejections, Goephysical Monograph Vol. 99, edited by N. Crooker et al., p. 27, Washington, DC., 1997.
Hundhausen, A. J., Size and locations of coronal mass ejections: SMM observations from 1980 and 1984-1989, J. Geophys. Res., 98, 13,177, 1993.
Hundhausen, A. J., Coronal mass ejections: A summary of SMM observations from 1980 and 1984-1989, in The many Faces of the Sun: Scientific Highlights of the Solar Maximum Mission, edited by K. T. Strong et al., p. 143, Springer-Verlag, New York, 1999.
Jackson, B. V., Helios observations of the Earth-directed mass ejection of 27 November 1979, Sol. Phys., 100 , 563, 1985.
Kahler, S. W., Solar flares and coronal mass ejections, Ann. Rev. Astron. Astrophys., 30, 113, 1992.
Kahler, S. W., N. U. Crooker, and J. T. Gosling, The polarities and locations of interplanetary coronal mass ejections in large interplanetary magnetic sectors, J. Geophys. Res., 104, 9919, 1999.
Marubashi, K., Structure of the interplanetary magnetic clouds and their origin, Adv. Space Res., 6(6), 335, 1986.
Mulligan, T., C. T. Russell, and J. G. Luhmann, Solar cycle evolution of the structure of magnetic clouds in the inner heliosphere, Geophys. Res. Lett., 25, 2959, 1998.
Neugebauer, M. and R. Goldstein, Particle and field signatures of coronal mass ejections in the solar wind, in Coronal Mass Ejections, Goephysical Monograph Vol. 99, edited by N. Crooker et al., p. 245, Washington, DC., 1997.
Plunkett, S. P., et al., Sol. Phys., 175, 699, 1997.
Plunkett, S. P., et al., LASCO observations of an Earth-directed coronal mass ejection on May 12, 1997, Geophys. Res. Lett., 25, 2477, 1998.
Plunkett, S. P., et al., Solar source regions of coronal mass ejections and their geomagnetic effects, J. Atmos. Sol.-Terres. Phys., in press, 2000.
St. Cyr, O.C., et al., Properties of coronal mass ejections: SOHO LASCO observations from January 1996 to June 1998, J. Geophys. Res., 105, 18,169, 2000.
Thompson, B. T., et al., SOHO/EIT observations of an Earth-directed coronal mass ejection on May 12, 1997, Geophys. Res. Lett., 25, 2465, 1998.
Tsurutani, B. T., et al., The future of geomagnetic storms predictions: Implications from recent solar and interplanetary observations, J. Atmos. Terr. Phys., 57, 1369, 1995.
Tsurutani, B. T., et al., Comment on `A new method of forecasting geomagnetic activity and proton showers' Planet. Space Sci., 36, 205, 1988.
Tsurutani, B. T. and W. D. Gonzalez, Interplanetary origins of storms, in Magnetic Storms, Geophysical Monograph, Vol. 98, edited by B. T. Tsurutani et al., p. 77, Washington, DC., 1997.
Wang, Y.-M., et al., The magnetic nature of coronal holes, Science, 271, 417, 1996
Webb, D. F. et al., Large-scale structures and multiple neutral lines associated with coronal mass ejections, J. Geophys. Res., 102, 24,161, 1997.
Webb, D. F. et al., Relationship of halo coronal mass ejections, magnetic clouds, and magnetic storms, J. Geophys. Res., 105, 7491, 2000a.
Webb, D.F., R.P. Lepping, N.U. Crooker, S.P. Plunkett, and O.C. St. Cyr, Studying CMEs using LASCO and in-situ observations of halo events, B.A.A.S., 32, 825, 2000b.
Zhao, X. P. and J. T. Hoeksema, Effect of coronal mass ejections on the structure of the heliospheric current sheet, J. Geophys. Res., 101, 4825, 1996.
Zhao, X. P. and J. T. Hoeksema, Central axial field direction in magnetic clouds and its relation to southward interplanetary magnetic field events and dependence on disappearing solar filaments, J. Geophys. Res., 103, 2077, 1998.
Zhao, X. P., J. T. Hoeksema and P. H. Scherrer, Changes of the boot-shaped coronal hole boundary during Whole Sun Month near sunspot minimum, J. Geophys. Res., 104, 9735, 1999.
Zhao, X. P., J. T. Hoeksema and P. H. Scherrer, Modeling the 1994 April 14 polar crown SXR arcade using three-demensional magnetohydrodynamic equilibrium solutions, Astrophys. J., 538, 932, 2000.
Zhao, X. P., J. T. Hoeksema, and K. Marubashi, Magnetic cloud Bs events and their dependence on cloud parameters, J. Geophys. Res., Accepted, 2001a.
Zhao, X. P., S. Plunkett and W. Liu, The geometrical and kinematical properties of the 1997 May 12 halo coronal mass ejection, Astrophys. J. in preparation, 2001b.
Zhao, X.P., J.T. Hoeksema, J.T. Gosling, and J.L. Phillips, Statistics of IMF Bz Events, in Solar-Terrestrial Predictions Workshop--IV, edited by J. Hruska et al., p. 712, Natl. Oceanic and Atmospheric Administration, Boulder, Vol. 2, 1993.
