The onset time and source region of the 12 May 1997 halo-type CME have been inferred (Thompson et al., 1998). We examine the change of the photospheric magnetic field in the vicinity of the source region before and after the onset time using magnetic observations from SOHO/MDI. Consistent with our earlier study of the CMEs, no obvious change pattern that is associated with the launch of the CME could be identified.
The dimming regions occurred during the CME can not be reproduced as open field regions by the coronal field model that has been used to successfully reproduce coronal holes. The magnetic field underlying the dimming areas is not predominatly of single polarity. Both suggest that the magnetic field configuration in the dimming area is different from the configuration in coronal holes, and that the dimming is caused by the magnetohydrodynamical process taking place purly in the corona.
We have examined the change of the photospheric magnetic field during two halo-type CMEs (06May1997 and 07Apr1997) and found that there was no obvious pattern of the change that may associated with the launch of the CMEs. The 12May1997 halo-type CME has large dimming areas. Is the dimming associated with the change of large-scale photospheric magnetic field?
The onset time and source region of the 12 May 1997 halo-type CME have been inferred to be 04:35 UT and centered at N23 W07 (Thompson et al., GRL, 25, 2465, 1998). This work tries to figure out whether or not the onset of the CME is associated with the change in the photospheric magnetic field, and whether or not the dimming area has similar magnetic configuration to that of the coronal hole.
Flares and CMEs are believed to be fuled by the free magnetic energy. Most prediction scheme of solar flare are builded on the complexity of photospheric magnetic field configuration in active regions. The high-cadence MDI observations of the photospheric magnetic field provide an unique oppotunity to examine the possible relationship between the CME-flare and photospheric magnetic field.
Figure 1a display the location of the CME-associated flare in EIT 195 A and the full-disk MDI magnetic image just before the onset time of the CME. The images are 1024x1024 pixels with 2 arcsecond per pixel. The 06:34 EIT image indicates the maximum dimming area after the onset of the CME. The dimming areas are similar to polar coronal holes in both their size and darkness. It has been suggested that the dimming is the trasient coronal hole (Watanabe et al., 1992). To search for the time variation of the magnetic field in the vicinity of the source region we first examine the magnetic field in the region borded by the white lines that contains the maximum dimming area (See Figure 1b).
Figure 2a displays the line-of-sight field observed just before (01:40 and 03:10), during (04:57) and after (06:28) the onset of the CME. The EIT images show the brightenning and the dimming observed firstly during the CME. Figure 2b displays the "source region" of the CME in full solar disk. It appears that no significant change in the photospheric field can be identified in Figure 2.
To find out any change in the photospheric field during the CME, the difference image of the photospheric field is produced using a series of MDI images. Figure 3a shows three difference images for the full disk (left column) and for the "source region" (right column). On the top of each panel, the observational time of two successive MDI images and pixel number with which the second image is left-shifted with respect to the first image. The pixel number is zero for the three difference images here, meaning that the effect of solar rotation does not excluded. There are several pairs of strong field (the white-black pair and the green-yellow pair). They occurred nearly same location, suggests that the pairs of strong field in the difference images are basically, if not all, caused by the solar rotation. Figure 3b is the same as Figure 3a but with a shift of 6 pixels that makes the cross-correlation of two successive images reaching maximum. There are still several pairs of strong field relative to the background. These pairs of strong field might still be the effect of the rotation, rather than the newly emerging ephemeral region, since they occurred in all three difference images at nearly identical places. If they are caused by the decaying or emerging magnetic flux, this change is certaily not a sufficient condition that causes or triggers the CME because the change pattern in the three difference images, i.e., before, during and after the onset, are basically the same.
Figure 4 displays the dimming in EIT 195 A images and the time variation of the dimming intensity (Courtesy of Dr. Barbra Thompson). The dimming region was formed around 05:07 UT and reached maximum near 06:34 12 May 1997. It lasted for more than 6 hours. If the field lines are monotonously open to the interplanetary space, like the field configuration in polar coronal holes, the magnetic field should be current-free.
The daily change of the coronal hole boundary has been partly reproduced (Zhao, Hoeksema and Scherrer, 1998) using the potential field-source surface model and the proxy of the instantaneous, global photospheric field distribution constructed with 96-minite full-disk MDI magnetograms. Figure 5 shows how to construct the global photospheric field distribution at 12May1997_06:28UT (the bottom) by using the monthly synoptic magnetogram map (the top) and the remapped full-disk magnetogram at the time of interest (the middle).
It is expected that the dimming area might be reproduced using the aforementioned method and there could be a set of field lines changing from close to open around the dimming area if the dimming is indeed a "trasient coronal hole". Figure 6 displays the calculated change of field lines taking place between the times of 03:16 and 06:28 UT may 30, 1998, i.e., between the times of just before the onset and of the time of maximum dimming. Also shown in each panel is the foot points of open field lines (the dotted areas) The left and right columns of Figure 6 show the change of field lines from close (c2o_c) to open (c2o_o) and from open (o2c_o) to close (o2c_c), respectively. There are more than 10 field lines that change from close to open. But there are fiels lines that change from open to close during the onset as well. All the changing field lines are located at the equatorward edges of the polar open field regions (holes). These change of field lines probably are associated with the evolution of the coronal hole boundary (Zhao et al, 1999), though the possibility of their association with the CME dimming cannot completely excluded. Figure 7a and Figure 7b display the changing field lines taking place before the occurrence of dimming and after the maximum dimming, respectively. The same as in the case of onset time, there are both close-to-open and open-to-close field lines, and all changing field lines are located at the edge of polar holes. It favours the suggestion that the change of field lines is associated with the evolution of coronal holes.
Figures 6 and 7 are obtained on the basis of the spherical harmonic coefficients inferred from the global photospheric field distribution. Is it possible that the dimming is caused by the change of small-scale field in the source region of the CME? Figure 8a, and Figure 8b show the field lines calculated using the Green-function method (Sakurai's code) above the dimming areas. Figure 8 shows again that no any open field line is obtained both before and after the onset of the CME.
Figures 6, 7, and 8 are obtained by upward integration starting at the solar surface. According to the potential field-source surface model, all open field lines must reach the source surace. Figure 9 shows the footpoints of open field lines (dark area) calculated from the source surface downward along each field lines using the global photospheric field distribution at the times cover the whole day of 30 May 1998, as shown on the top of each panel. There is no calculated open field regions corresponding to the dimming area in all of panels (see the bottom panel of Figure 1)
Observations have shown that coronal holes occur in the area where the magnetic elements are predominatly of one polarity. On the large-scale photospheric field observed by WSO, coronal holes are located in the unipolar magnetic region. The top panel in Figure 10 is the global photospheric field distribution at the time of peak dimming with grid spacing of 1 degree, showing that the dimming is located in the region not predominated by one polarity. The middle and bottom panels are obtained by lowering the spatial resolution to 5 degrees and 10 degrees, respectively. The dimming areas are not located in the unipolar magnetic region.
It is suggested from Figures 6--10 that the dimming during CMEs are associated with magnetohydrodynamic processes purely in the corona.
The same as the case of the Jan. 6 and Apr. 7, 1997 CMEs, there is no obvious change pattern of the photospheric magnetic field observed that is associated with the onset of CMEs.
The dimming during the CME suggests the decrease of the plasma density and/or temperature, the same as the case of coronal holes. However, the magnetic field configuration within the dimming area may be different from the configuration in coronal holes. If it is not the case, the dimming would be reproduced as the footpoint area of the open field lines and would be located in the unipolar magnetic region. It appears that the cause of the dimming, and probably the onset of CMEs as well, is associated with the dynamic change of the magnetic field and plasma purely in the corona, instead of the photosphere.
