Transcript C5_Hochedez - Helioseismic and Magnetic Imager (HMI)

Coronal seismology, AIA/HMI and image processing
(-: Best wishes :-)
JF Hochedez, E Robbrecht, O Podladchikova, A Zhukov, D Berghmans
SIDC @ ROB
Solar Influences Data analysis Center
Royal Observatory of Belgium
Mandate of this presentation
AIA
Coronal
Seismology
Image
Processing
EUV imaging observations and seismology
(1) in [simple] flux tube magnetic structures
Fast magneto-sonic
modes
Kink
Sausage
Slow
Magneto-sonic
(sausage) mode
Standing
TRACE 1MK 1999
(Aschwanden et al., AIA 2009?
Optical
Nakariakov
et al.) Flow
SUMER 6MK 2002
(Kliem et al.,
Wang et al.)
Propagating
Motion & brightness change
TRACE 20MK
2005
tracking
AIA 2009?
(Verwichte et al)
Loop recognition
and
EIT 1MK 1998
Cactus-like approach
Deforest & Gurman
•x-t diagrams,
Berghmans
& Clette 99,
•Hough
transform,
TRACE...
•clustering
EUV imaging observations and seismology
(2) in [other] coronal structures
Global EIT waves
EIT
EIT wave detector
Thompson et al 1998
not discussed in this talk
Prominence oscillations
but not forgotten
Oscillations and waves
during eruptions
(CME or flares)
The future? But
challenging!
Sympathetic flares
Flare detector and
Podladchikova et al (submitted)
Presentation sections
1. When Optical Flow will detect fast modes in
flux tubes
2. Loop recognition and Hough transform applied
to slow waves
3. What EIT waves can tell us about the corona
4. [Prospective] sympathetic flares.
How do they communicate?
5. Conclusions
Optical Flow
& its application to fast modes
Remaining problems with
kink oscillations
• Damping
– Test competing explanations
• phase mixing
• resonant absorption (Goosens et al 2002)
• leakage at footpoints, others…
– Too many parameters
• stratification (estimated by Andries et al 2005)
• Curvature
• variable cross-section
 More statistics needed
• Exciter(s)
– Their nature? From below? From side?
• Why so few ?
– Damping or lack of exciters?
Hopes from AIA-HMI (1/2)
• 8 bandpasses
– Longitudinal density profile (DEM tools)
– Heating profile
• Spatial resolution
– Radial density profiles: concentric shells, threads?
• 0.6”probably still too low
– Overtones (Verwichte et al 2004)
– 3D geometry with Secchi
• Loop length
• vertical vs swaying (Wang et Solanki 2004), etc.
• Full Sun FOV
– 2 pressure scale heights
• long loops with good SNR
– With temporal coverage: statistics
Hopes from AIA-HMI (2/2)
• 2s Cadence
– time aliasing repressed
– SNR  Time rebinning
– exposure time ~0.1s
• Less kinetic blurring
AIA trade-off TBD
• Stroboscopy
– Observe fast sausage waves, fast sausage oscillations, fast
propagating kink waves!
• Effective area (44x TRACE@171, 61x @194)
– See smaller disturbances.
• Presence of HMI
– Independent estimate of B (cf. too many parameters)
• Compatible with seismology? (NLFF and dynamics)
Quantify motion together with
intrinsic brightness variation
in EIT image sequences
Gissot & Hochedez, 2006
VELOCIRAPTOR
VELOCIty & bRightness vAriations maPs construcTOR
Inputs & outputs
Velocity
field
Image In(x,y)
e.g. EIT “CME Watch”
Image In+1(x,y)
Brightness
Variation
field
Hochedez & Gissot
1.
Similarity field
2.
3.
between
In(x,y) (warped)
and In+1(x,y)
Local “texture”
Residuals
Differential rotation recovered
from a couple of EIT images
(No BV estimation)
BV map of the May 12, 1997 event
Velocity map of the May 12, 1997 event
(No BV estimation)
14 July 1998 12:50:16
Presence of texture in 2 orthogonal directions
Presence of texture at least in one direction
Zoom of the previous representation
Velocity field produced by Velociraptor
Average displacement ~0.3 pixel
→ LCT not appropriate (a posteriori justification)
Velocity field corrected for global shift
Loop displacement ~0.15 pixel
Question: What are the anticipated artifacts for AIA?
OF & fast magneto-sonic waves:
Conclusions and outlook
• Velociraptor can measure sausage and kink waves
–
–
–
–
Precisely, all along the loops, systematically, Outliers?
Challenging development
Being fully calibrated
2 main problems understood and being corrected:
• Strong BV  fictive motion
• Some spurious sliding remains along loops
• Post-processing of the fields needed in order to identify
waves autonomously (1D wavelets?)
• AIA + OF  great prospect
– Sausage modes by EUV imaging?
– Flows from steady reconnections?
– Mode coupling?
Slow waves
Good overall understanding
but …
• Wave or plasma motion? (no Doppler measurements)
• Sound speed if pattern seen in several BPs
• cf. Robbrecht et al. 2001 EIT vs TRACE
• Klimchuk et al 2004:
– Their study validates classical thermal conduction damping
– But “TRACE loops are inconsistent with static equilibrium and
steady flow”
– “Observed damping times of slow mode oscillations might be a
lower limit to effective damping times, which can only be
corrected if the cooling time is known from multi-filter data.”
• Seismology is complementary to DEM
Useful image processing
for slow waves (1)
• Loop extraction (ridge detection)
Useful image processing
for slow waves (2)
• Analysis of X-T diagrams
– Hough Transform
– Clustering
– Cf “CACTUS” applied to [faint] CME detection
• in LASCO C2 & C3
Computer Aided CME Tracking -CACTus
11 November 2003
15h18
15h54
17h06
r
t
Δt
t0
EIT waves
EIT waves for coronal seismology
• EIT waves: bright fronts propagating from eruption sites
observed in EUV (SOHO/EIT, TRACE, CORONAS-F/SPIRIT,
195 Å, 171 Å, 284 Å bandpasses).
• Sometimes EIT waves propagate nearly isotropically and often
– globally.
• EIT wave speeds are usually about 150–400 km/s, typically
around 250 km/s.
• Association with chromospheric Moreton waves, waves in He I
and waves in SXR?
If EIT waves are fast
magnetosonic waves…
*
*
Wang (2000)
Wu et al. (2001)
Courtesy A Zhukov 2006
Fast magnetosonic wave speed around 250 km/s means b ~ 1 or b > 1
in the “quiet Sun” corona
Force-free approximation is not valid!
a quantitative investigation
Podladchikova & Berghmans, 2005
DIMMING & EIT wave extraction from EUV image
Brightness distribution (histogram) analysis
study of higher moments
EIT wave radial and polar analysis
Ring Analysis
radial velocities in the EIT wave
Angular-Ring Analysis
potential angular features
Skewness & Kurtosis of PDF of difference image versus time
Simultaneous peaks
+ dimming area criteria
→ EIT Waves!
Courtesy of Podladchikova & Berghmans
12 May 1997
Width
m3-m2
mmax
Both quadratic
Distances vs Time
Integrated signals vs Time
Courtesy of Podladchikova & Berghmans
Results
1.
2.
3.
4.
5.
Anisotropy even without obstacles.
Correlation with associated dimming;
Dimming contiguous to wave front in all directions
Width of the front grows ~quadratically in time;
Integrated intensity of wave front grows during > 1/2 hr
The front intensity of linear magnetosonic waves would decrease
Integrated intensity of front balances integrated intensity of the
dimmings (in early life of wave)
EIT wave = MHD wave?
Sympathetic flaring
Consecutive occurrence of flares in different
AR
Perturbation velocity from flare to flare
“to set the fire”
Vchar ~ 110 km/s
Velocity [km/s]
t < 5h.
3225 flares registered with coordinates since 01/01/2004.
Statistically complete series.
Result does not depend on time interval
Conclusion
significant number of events where one flare
“sets fire” triggering another distant flare in a
separate active region.
Propagation velocities for such perturbations
around 110 km/s.
B2X flare detector
Method: Wavelet spectrum (scale measure) analysis
Hochedez et al ’02 Solspa2 Proc., Delouille et al SoPh ’05
½ log(μ(scale))
Result: Small flares automatic detection
Relevance: Sympathetic flaring studies
At flare peak
Just before
the flare begins
log(scale)
Beauty spotter
Method: Extraction in scale space by Lipschitz coefficient
Hochedez et al 2002, Soho11 WS Proc.,
Hochedez et al 2003 Soho13 WS
Result:
BPs, brightenings and
Cosmic Ray Hits extracted
Relevance:
Oscillations in point-like structures
Conclusions
• The easy things about waves have been found.
Intelligent techniques can invigorate future research
– Prospect for eruption precursors?
• Image processing = binding agent between theory and observation
– Like an additional "telescope" for small scale physics
•
•
•
•
improve resolution
separate different processes (mutually and from noise)
extract waves or reconnection events
part intensity from velocity variations
– Like a new "microscope" for large scale physics
• Describe of important events
• "in situ sensor“, identifying the nature of events
• Uncover unexpected regularities
• For all these reasons, all detected waves should go in the SDO catalogs