C3. Thermal Studies: Techniques

1
C3. Thermal Studies Techniques
Happy Valentines Day!
2
Science Context DEM

• Observed pixel value p in the ith channel at
  pixel coordinate ?
• where DEM(T, ?) describes conditions in the
  corona
• With AIA, well have the best chance yet to
  obtain meaningful solutions for the DEM
• 6 narrow EUV bandpasses ? broader T coverage
• high spatial resolution ? study the thermal
  properties of elemental structures
• high cadence ? study evolution of coronal
  structures at the shortest relevant timescale
• simultaneous observations from Solar-B EIS XRT,
  GOES SXI, STEREO SECCHI,
• Improving spectral codes and inversion techniques

3
Agenda

• Modeling Assumptions and Uncertainties
• Inversion Techniques
• Data Products
• DEM Science
• AIA DEM recovery challenge status update
• Wanted practical ideas and questions, not
  necessarily solutions
• Well try to limit the first three topics to 20
  minutes each
• Please try to hold your comments for the
  appropriate topic!

4
C3.1 Modeling Assumptions and Uncertainties

• What are the pros and cons of current spectral
  codes (CHIANTI, APEC, MEKAL, etc.)?
• Whats the best way to handle abundances and
  ionization equilibrium calculations?
• How significant are opacity effects due to
  optical depth of emitting plasma along the lines
  of sight?
• How do we factor in uncertainties in the
  instrument calibration?
• How do all these uncertainties propogate into DEM
  analysis?
• How can we design our DEM tools to take into
  account all of these questions?

5
C3.3 Data Products

• What sort of data products should we provide, and
  at what level?
• Nominal DEM solutions
• How are they calculated (algorithm, assumed
  abundance, etc.)?
• What resolution/cadence?
• How are uncertainties calculated and provided?
• Temperature maps
• Derived from DEM solutions, but how, exactly?
• How are they displayed?
• To answer this question, we need to determine
• computational power required for different data
  products and solution methods
• useful/reasonable levels of modularity and
  flexibility for DEM tools provided by the
  instrument teams
• graphical methods for communicating the
  information of DEM solution/uncertainty sets
  across images with millions of pixels
• value of instrument-team provided DEM products
  (near real-time cadence) versus end-user tools
  (greater flexibility, processing)

6
AIA DEM Recovery Challenge

• Basic Procedure
• We provided instrument response functions and 24
  simulated sets of multispectral AIA observations
  participants found the DEM functions that best
  reproduced those observations, and compared them
  to the targets
• www.lmsal.com/boerner/demtest/
• So far, we have responses from
• J. Kaastra, SRON
• S. Gburek
• M. Siarkowski
• V. Kashyap (SAO)
• more may be coming more are welcome!
• full results will be posted on Friday

7
Simulated Observations

• 6 DEM variants
• plotted at left
• one-parameter tuning
• most are fairly smooth
• 4 sets of parameterizations
• Case 1-6 used provided CHIANTI T responses
  (Feldman abund, Mazzotta ion eq.) no systematic
  errors
• Case 7-12 used CHIANTI T responses with
  different abundance (Meyer) ionization (Arnaud
  Rothenflug) no errors
• Case 13-18 used CHIANTI T responses added
  pseudorandom calibration and atomic physics
  errors
• Case 19-24 no errors, but used APEC T responses

8
Non-uniqueness

• 4 perfect solutions from Marek Siarkowski
• black target
• red Withbroe-Sylwester solution
• blue/green/orange genetic algorithms

9
Preliminary Conclusions

• DEM reconstruction is hard
• Solutions are extremely sensitive to error
• Solutions are extremely non-unique
• This sort of test is potentially valuable
• A lot of people have simulated observations, then
  recovered their own observations
• This is not as powerful a test as it could be
• We will continue this experiment, and are
  interested in suggestions for how to do it better
• Focus on algorithms vs. spectral codes vs.
  systematic errors
• Develop an understanding of the differences in
  the spectral codes
• You can help!