Transcript Document

Looking for the CME Onset:
A 10 Year CDS Campaign
Richard Harrison, Rutherford Appleton Laboratory
• An example of a
long-term CDS
campaign
• An excuse to
indulge in a bit of
CDS history
SMM CME Onset Campaign
OSO/Skylab – first CME observations did hint that there was NOT a
one to one relationship.
SMM investigations from 1982 showed more inconsistencies - first
real opportunity to explore flare-CME relationship with reasonable
daily planning, resolutions, co-pointed instruments etc… Seemed
straightforward enough, but opened a can or worms!!
Using today’s images…
We are asking a very basic question. How do we
relate one of these…
… to one of these? What is the relationship
between the two major solar active phenomena?
The flare-CME relationship
Basic stuff! If CME is corona ‘blast’ from flare then:
 Flare would sit under core of CME;
 Onset of flare & CME launch would coincide;
 Scale of flare & CME would be consistent;
 Would expect near one to one flare-CME occurrence.
 Didn’t turn out to be that simple…
The flare-CME relationship Pre-SOHO conclusions
 There is a strong association between flares and CMEs – but not a
one to one association;
 The onset of a CME associated with a flare appears to occur at any
time within several tens of minutes of the flare onset;
 The scale sizes of CMEs and flares are very different (~45o and
<10o, respectively). Since the footpoints of CMEs do not expand
outward, this suggests that the flare site is too small – unless there is
very early expansion;
 The flare tends to lie within the span of the CME and may often lie to
one side;
 The CME source region is commonly much larger than an active
region or flare, though it frequently encompasses an active region.
The flare-CME relationship Pre-SOHO conclusions
“The flare and CME are both consequences of the same magnetic
‘disease’. They do not cause one another but are closely related.
Their characteristics are the results of local conditions, and thus
we may witness a spectrum of flare and CME properties which
are apparently unrelated, even resulting in events without the
flare or CME component.”
The flare-CME relationship
Understanding the CME onset and the flare-CME relationship
are critical for understanding the impact of solar activity on the
Earth and of processes such as mass acceleration and
reconnection, which are fundamental stellar processes. An
ideal candidate for the new SOHO mission.
Enter SOHO….
December 1984 – ‘European Space Science Horizon
2000’
“… cornerstone consisting of two projects, an
observatory at the L1 point and a multipoint space
plasma physics mission.”
Enter SOHO….
Solar and Heliospheric Observatory
The most sophisticated solar observatory ever built!
- ESA Cornerstone mission (with NASA)
- launched (Atlas rocket) - 2 Dec 1995
- 12 instruments to study:
• solar interior
• solar atmosphere
• solar wind
- UK involvement: CDS, part
of LASCO and scientific
involvement in most of
payload.
- 1.85 tonnes, 3-axis
stabilized, Sun-pointed
Enter SOHO….
Orbit of Sun-Earth L1 Lagrangian
point - 1.5 million km Sunward of the
Earth - 1/100th of the way to Sun
• constant view of Sun - no eclipses
• no ‘contamination’ due to Earth
environment
• sunward sentinel
Enter CDS….
The CDS heritage at
RAL:
- Alan Gabriel
- Bruce Patchett
- CHASE and MSSL
CDS Operations
EJECT studies started in mid-1996, to be repeated many, many
times over coming years – mainly as JOP67:
- Mosaic of three 4 arcmin fields
- 10 s exposures
- 4x240 arcsec slit (60 locations)
- Cadence 50 min(!)
- Six emission lines:
He I 584 Å (20,000 K)
O V 629 Å (250,000 K)
Mg IX 368 Å (1 million K)
Fe XVI 360 Å (2 million K)
Si X 347/356 Å (1.3 million K)
- Play off between cadence and plasma diagnostic tools! BUT, first
real opportunity to obtain plasma diagnostics of CME source?
CDS Operations
The flare-CME relationship – Sept 23 2001
CME Onsets
The Events of September 23, 2001
12:13 UT
13:03 UT
13:53 UT
14:43 UT
15:33 UT
2 million K Fe XVI 360 Å line
The flare-CME relationship – Sept 23 2001
12:13 UT
13:03 UT
13:53 UT
14:43 UT
15:33 UT
1 million K Mg IX 368 Å line
The flare-CME relationship – July 25 1999
13:54 UT
14:06 UT
15:54 UT
The flare-CME relationship – July 25 1999
13:54 UT
The flare-CME relationship – July 25 1999
-
Pre-flare ascending loops
Coronal dimming
Pre-flare CME onset
CME source larger than flare
The flare-CME relationship – Dimming Mass
CME Onsets
Date
Dimming mass
(DEM/Si X) [kg]
CME Mass
[kg]
Jul 16 1997
4.3x1010/1.3x1011
5x1010
May 8 1999
1.1x1012/4.2x1012
3x1011
Jul 25 1999
7.4x1011/3.4x1012
3.5x1012
Feb 19 2000 1.1x1014/2.7x1014
1.1x1012
Aug 19 2000 6.4x1011/1.8x1012
4.7x1011
 Onset of dimming and CME ‘coincident’.
 Location of dimming under ascending CME.
- Unique method for identification of CME source material.
- Plasma diagnostic analyses of the source providing
information on onset process.
The flare-CME relationship – Dimming
CME Onsets
• CME Onset Studies, Harrison, R.A., 1997, ESA SP-404 (proc. 5th
SOHO Workshop), 85.
• A spectroscopic study of coronal dimming associated with a coronal
mass ejection, Harrison, R.A. and Lyons, M., 2000, Astron. Astrophys.
358, 1097.
• Coronal Dimming and the Coronal Mass Ejection Onset, Harrison,
R.A., Bryans, P., Simnett, G.M. and Lyons, M., 2003, Astron.
Astrophys. 400, 1071.
• SOHO Observations Relating to the Association Between Flares and
Coronal Mass Ejections, Harrison, R.A., 2003, Adv. Space Res. 32,
No. 12, 2425.
• On the Coronal Mass Ejection Onset and Coronal Dimming, Howard,
T.A. and Harrison, R.A., 2004, Solar Phys. 219, 315-342.
The flare-CME relationship – Dimming Mass
• CME-related dimming first seen using Skylab, also seen with Yohkoh as
well as EIT and CDS – but only CDS has addressed this with detailed
spectral analysis.
• Many studies identify patches of dimming within active regions - how
does that relate to a 45 degree CME? The scale ‘problem’ must be
addressed. Is there also a line-of-sight problem?
• If the dimming region identifies the critical low coronal source region
then we can analyse the source plasma in the lead up to onset. We can
also explore the possibilities of early CME prediction or on-disk CME
prediction.
CME Prediction?
Can we predict a CME onset utilising an algorithm based on the
number of pixels which show declining intensity in selected
emission lines?
If we can do this successfully on
the limb, we can do it on the disk.
Basic scheme scan Mg IX and
Fe XVI EJECT mosaics from 1996
to date (several hundred) using
automated procedure.
For each, if contiguous set of
pixels (predefined minimum
number) shows decrease in intensity
beyond specified limit, define a CME alarm.
Compare CME alarms with LASCO event lists.
CME Prediction?
Preliminary results for 100 EJECT runs (1999-2003) –
Ω = CME alarm parameter (defined by number of contiguous pixels
decreasing in intensity beyond specified limit)
• 47 CMEs in periods covered - BUT
some will be from behind limb
• Even some behind limb will be seen
as coronal depletion, so expect >50%.
• Thus, if we are catching over 50% of
the CMEs, we are doing well.
• Need balance between false alarms
(as low as possible) & fraction
predicted.
CME Prediction?
Preliminary results for 100 EJECT runs (1999-2003) –
What if it was random??
• 47 CMEs in period.
• 100 runs - equal chance YES/NO, would get 50 alarms.
• Of 47 events, equal chance of alarm, expect to ‘catch’ 24.
HIT RATE – successful prediction/total alarms
FALSE HIT RATE – false alarms/total alarms
No.
alarms
CME
predicted
False
alarms
CMEs
not
predicted
HIT
RATE
FALSE
HIT
RATE
Fraction
CMEs
predicted
Fraction
CMEs
not
predicted
Random
Case
50
24
26
23
48%
52%
50%
50%
Ω > 0.25
23
18
5
29
78%
22%
38%
62%
Ω > 0.2
37
27
10
20
73%
27%
57%
43%
Ω > 0.15
54
37
17
10
69%
31%
79%
21%
CME Prediction?
Preliminary results for 100 EJECT runs (1999-2003) –
No.
alarms
CME
predicted
False
alarms
CMEs
not
predicted
HIT
RATE
FALSE
HIT
RATE
Fraction
CMEs
predicted
Fraction
CMEs
not
predicted
Random
Case
50
24
26
23
48%
52%
50%
50%
Ω > 0.25
23
18
5
29
78%
22%
38%
62%
Ω > 0.2
37
27
10
20
73%
27%
57%
43%
Ω > 0.15
54
37
17
10
69%
31%
79%
21%
• Want high hit rate and low false hit rate – but note that some over the limb events
you cannot see – but also few missed CMEs. What is OK? (>75%, <20%, <30%
from limb obs.??)
• Need to do considerably better than random case!
• Is it better to predict few, but with confidence – i.e. issue less alarms but be sure
that they achieve, say, better than 75% success, knowing you will miss some?
Future Missions for CME Research?
STEREO – Twin spacecraft out of the Sun-Earth line
Two NASA spacecraft, orbiting the Sun
– one leading the Earth, one
following, providing two views of the
Sun and the space between the Sun
and Earth. Due for launch in 2006.
 First 3D views of a star
 First views of solar ejected clouds
impacting the Earth
 UK involvement: (i) Novel CCD camera
systems on all remote sensing systems
aboard the spacecraft; (ii) Leadership of
the unique Heliospheric Imager
instrument, a wide-angle telescope
system, to image solar clouds in
interplanetary space
Future Missions for CME Research?
SDO – The Solar Dynamics Observatory
NASA’s flagship of the Living with
a Star programme – a large
Earth-orbiting high-resolution
solar observatory. Due for launch
in 2008.
 High-resolution imaging,
helioseismology and magnetic
mapping of the Sun, will allow
detailed understanding of the
complex solar atmosphere and
its magnetic fields.
 UK involvement: CCD
camera systems, developed
from the STEREO programme,
are incorporated in the US-led
instrument package.
Conclusions?
CDS has allowed a spectroscopic dimension which has enabled
significant advances in understanding the dimming phenomenon – and is
well suited to determining prediction algorithms… Next steps?
- Analysis of plasma prior to dimming (having identified the source?)
- Completion of dimming ‘alarm’
- SOHO, STEREO, Solar-B, SDO
Watch this space!