LowCoronal and Coronagraphic Images: Their Complementary Roles in Understanding GeoEffective Eruptio

1
Low-Coronal and Coronagraphic Images Their
Complementary Roles in Understanding
Geo-Effective Eruptions
N. V. Nitta, Lockheed Martin Solar and
Astrophysics Lab.
It has been established that coronal mass
ejections (CMEs) are the primary driver of severe
space weather. By definition, CMEs are observed
by white-light coronagraphs. However, a question
arises as to whether modern coronagraphs (e.g.,
SOHO/LASCO) are sensitive enough to capture all
the Earth-directed and potentially geo-effective
CMEs. In the recent list of interplanetary
coronal mass ejections (ICMEs) during 1996-2002
by Cane and Richardson (2003), 30-50 of ICMEs
have no associated halo CMEs. Even though no
features in X-ray and EUV images are known to be
100 correlated with CMEs (Hudson and Cliver
2001), these images are useful for understanding
the origin of geo-effective CMEs.
2
CME Source Regions

• The source region of a fast CME, as observed by
  LASCO, can usually be located unambiguously in
  EUV/X-ray images taken within 1 hour before the
  first detection of the CME.
• If we do not see in EUV/X-ray images any
  signatures (e.g., waves, dimming, post-eruption
  arcades, etc.) that often accompany CMEs, we may
  conclude that the fast CME comes from backside.
• But how well is the source region identified, if
  the CME is slow and diffuse, and its first
  detection by LASCO C2 comes hours later than a
  CME-associated phenomenon?

3
Fast CMEs
Figure 1 Bastille Day 2000 event observed by
LASCO, EIT and TRACE
Perhaps nobody dares to question the link between
the CME and the flare/filament eruption on 14
July 2000. Note that low coronal images combined
with magnetic field data/extrapolation, not
coronagraph images, help us understand its origin
and its possible geo-effectiveness. We already
know from the major SEP event (starting as early
as 1050 UT) that the eruption drove an IP shock.
LASCO data are useful for predicting the shock
arrival at Earth (combined with DH type II burst
data) but not in this event because of too much
contamination with SEPs.
4
Slow CMEs
Figure 2 LASCO height-time plot superposed with
the GOES X-ray light curve. Extrapolation of
the height-time relation (noisy!) puts the CME
departure time at 0305 UT from the active region
at S17 W19. But this fit ignores the data point
at 0440 no CME was detected above 2Rsun
(red). In fact, EUV/X-ray images suggest no
CME-like activity until the flare onset at 0431
UT, so the true height-time relation may be
something like the blue line.
If we blindly use the LASCO height-time relation
without reference to the data coverage or to low
coronal images, we often conclude the CME onset
time to be in the middle of nowhere. This has a
danger of unduly separating CMEs from what we
know in the corona.
5
Use of Height-Time Relation
Figure 3 Tracing the ejection from the low
corona. (a) Height-time plot superposed with the
GOES X-ray light curve. Linear fit to the LASCO
C2/C3 height-time relation puts the CME lift-off
time to be 200 UT, well before the flare onset.
Inclusion of the C1 data makes a better link
between the CME and the flare X-ray ejection.
However, plotting the ejections in X and Y
directions separately (in b and c) indicates that
things are not so simple.
To relate the phenomena in EUV/X-ray images with
CMEs, it is important to observe 1.1-2 Rsun.
According to the above figure, however, (1) CMEs
may be highly non-radial in the beginning (then
the height may not be a good representation),
or (2) CMEs observed by C2/C3 may be physically
separate from the low coronal counterparts.
6
Earth-directed CMEs Halo CMEs
How can we be sure that a halo CME is
Earth-directed, when we do not see characteristic
patterns for CMEs in the low corona?
Figure 5 As predicted, a clear magnetic cloud
started on 1997 January 10 and triggered a
reasonably severe storm. Hence the importance of
halo CMEs established to predict storms.
Figure 4 Partial halo CME on 1997 January 6.
This occurred during a CDAW, where it was
predicted that the ejected material would be
observed near Earth 3-4 days later and that it
might cause a geomagnetic storm.
Figure 6 (Movie) But the solar signatures in
Yohkoh/SXT images around the estimated CME
liftoff time are minimal (no EIT data). There was
a bigger eruption on the previous day, but no
halo CME was detected. Movie not included in
CD. Get it from the workshop ftp site or from
here.
7
Alternative Scenario
If we do not start from the halo CME ICME
association as given, the following scenario may
be a possibility.

• The magnetic cloud on January 10 originated from
  the eruption (with LDE and filament
  disappearance) on January 5.
• The January 6 halo CME was a backside event, not
  related with the disappearance of a filament, as
  noted by Webb et al., in the NW part of the
  region (note that the halo CME was seen mostly
  over the S limb).
• Neither filament eruption on January 5 nor
  January 6 caused a CME intense enough to be
  observed by LASCO.

Figure 7Correlation between the maximum ICME
speed and the shock transit speed, which was used
by Webb et al. (1998) to argue that the eruption
on January 6 is a better candidate for the
magnetic cloud on January 10 than that on January
5. But the scatter is already large, and even
the case for January 6 makes an outlier.
But SXT may have failed to detect key signatures
from the eruption on January 6, which may have
indeed been responsible for the halo CME. These
signatures may be found only with simultaneous
images in a broad temperature range. Or can
coronagraph images tell us convincingly whether
the CME was front-sided?
8
Sensitivity of Coronagraphs
Figure 8 On 1997 May 10, MURI region AR 8038
produced an eruption less global than the one on
May 12. LASCO did not observe a CME on May 10.
When geo-effectiveness is studied in terms of
halo CMEs, the sensitivity of the instrument to
motions off the plane of the sky is seldom
addressed, as if the halo CME were a necessary
condition. But in the ICME list of Cane and
Richardson (2003), many ICMEs are not associated
with LASCO CMEs.
9
Weird Geo-effective Events
Figure 9 Some geomagnetic storms are not
associated with fast CMEs like the Bastille Day
2000 event. Zhang et al. (2003) argued that they
are attributable to faint slow partial halo CMEs
from the over the east limb. But during the
periods of interest, multiple eruptive events
occurred closer to the disk center but without
CME association. For example, the cusp-like
appearance in SXT images often corresponds to a
post-eruption arcade. Why are they not seriously
considered?
This may again be related to the issue of the
sensitivity of coronagraphs. We should identify
the coronal manifestation of ICMEs, at least
magnetic clouds which may have a more organized
origin than other kinds of ICMEs.
10
CMEs and Non-eruptive Flares
X-ray/EUV images occasionally reveal ejections.
But some of them do not move long distances and
are not associated with CMEs.
Figure 11 Wind/WAVES dynamic spectrum for 1997
November 27, overplotted with the GOES X-ray
light curve. Note that the X-class flare at
1210 UT was very quiet in these wavelengths.
However, there was a metric type II burst.
Figure 10 SXT images of X-class flares without
CMEs. The first two events show slow (ejections as indicated by thin arrows, but they
do not move beyond the overlying loop structures
as indicated by thick arrows. The third event
does not even show ejections, but slowly (km/s) rising loops.
11
EUV Ejections and CMEs
Figure (Movie) 13 Similar example. This one
seems to be a hesitant (?) ejection. A slow
CME seems to be associated. The white feature in
the encircled area, rather than the filament
itself, may be the CME signature. Movie not
included in CD. Get it from the workshop ftp
site or from here.
Figure 12 (Movie) TRACE example of
unsuccessful ejection, not leading to a CME.
The motion decelerated and halted (Ji et al.
2003). Movie not included in CD. Get it from the
workshop ftp site or from here.
Filament eruptions not associated CMEs are
characteristically decelerated. There is a
variety to this class of ejections. The
filament eruption in the right example is not
spectacular, but it may be responsible for the
slow CME observed an hour later.
12
Dimming in Multiwavelength Images
Figure 14 Coronal dimming in a sigmoidal region
as observed by SXT, SXI and EIT. Images
dominated by hot plasma (SXT and SXI B_THN_A and
P_THICK) show dimming only in Area 3. Area 2
shows decrease in flux but it is due to a
brightening around 1300 UT in most images but
EIT and SXI OPEN. Area 1 shows dimming only in
EIT and SXI OPEN. The quicker recovery in SXI
OPEN images seems to be due to flare brightening.
Note the different behaviors of SXI OPEN and
P_THN_A even though their temperature responses
are close. We need a broad temperature coverage
to tell mass depletion (due to the CME) from
temperature effects.
It appears that dimming in EIT images is a
reliable indicator for mass depletion due to
CMEs. But it is important to supplement it with
data that sample different temperatures so that
we can distinguish real mass loss from
heating/cooling.
13
Conclusion
We need part II of
which should focus on disk events, calibrating
both coronagraphic and low coronal data.