Three-Dimensional Power Spectra of GONG High-Cadence Magnetograms

1
Three-Dimensional Power Spectra of GONG
High-Cadence Magnetograms

• F. Hill, J. Bolding, R. Clark, K.
  Donaldson-Hanna, J. Harvey, G. Petrie, C. Toner
  T. Wentzel

2
GONG Magnetograms

• Obtain one per minute, 24-7, average duty cycle
  of 0.87
• Noise level 3G per pixel per minute
• Recent installation of new modulators reduced
  zero-point background to 0.1 G
• Hourly synoptic products
• Here look at 1-min merged magnetograms

3
Merging

• Obtain as many as four simultaneous images for a
  given minute need to merge
• Tested histogram equating method (H. Jones)
• Take two magnetograms
• Form cumulative histogram of field on disk
• Find linear transformation to produce nearly
  identical histograms
• Apply to images
• Found that this does not work when considering a
  time series of images with variable seeing
• Currently using a very simple scheme
• Register and circularize simultaneous images
• Simply average

4
The movie

• Sequence of 3200 images/minutes starting on July
  1, 2006 at 0 UT
• Fill factor of 0.95 (148 missing images)
• Missing images are not interpolated across,
  simply skipped
• 32 square patch tracked and remapped using GONG
  ring pipeline
• Patch tracked symmetrically across central
  meridian
• Tracking at photospheric rate, not magnetic rate,
  so some drift

5
One frame
Contains quiet area, sunspot, and
network/decaying active region
6
The movie

• Highly time compressed 53 hours in about 10 sec
  (20,0001 compression)
• Things to look for
• Evershed flow from spot
• Flickering field in spot (probably seeing and
  line blending in strong fields)

7
3-D Power Spectra

• Analyze field in three 16 areas
• Spot region (lower right)
• Network/decaying region (lower left)
• Quiet region (upper left)
• Compute 3-D FFT, then power

8
Slices at constant frequency of 31.25 µHz (Period
of 8.9 hr)
Power offset due to drift from tracking rate.
Once corrected, should be able to measure outflow
velocities.
9
Slices at constant frequency of 3333.33 µHz
(5-min band)
10
Closer look at anisotropic oscillations
Use entire 32 area for analysis mixes
different field characteristics, increases k
resolution.
2917 µHz
3125 µHz
3333 µHz
3542 µHz
3750 µHz
11
Causes?

• Instrumental
• Velocity undoubtedly leaks into B
• Polarization measurements taken at different
  times?
• Spatial offset of polarization measurements?
• Processing
• No problems in spectrum computation
• Merging?
• Interpolation?
• Solar
• Magnetic field selects a propagation direction?
• ??

12
Conclusions

• GONG continual high-cadence merged magnetograms
  available shortly.
• Good for rapid changes (e.g. flares, flows,
  oscillations).
• 3-D power spectra useful for magnetic field
  studies.
• Strange anisotropic behavior of oscillations?

13
Postcsript

• While at Melbourne, R. Bogart managed to process
  some MDI 1-min magnetograms
• He found similar results incomplete rings in a
  quiet area
• Need to get larger sample in many different
  situations of activity to understand