Diagnosing Coronal Heating using Computed Tomography Imaging Spectroscopy

ISSI “Why Do We Still Have a Coronal Heating Problem?” working group · 2023 · the slides as they were presented, with the movies playing

Diagnosing Coronal Heating using Computed Tomography
Imaging Spectroscopy

International Space Science Institute

Why Do We Still Have a “Coronal Heating Problem” working group

Roy T. Smart, Charles C. Kankelborg, and Jacob D. Parker

1 / 19

False-color IRIS Spectroheliogram

Si IV 1394 A

320-step dense raster

50 minutes per frame

Good FOV, but not temporally resolved



2 / 19
3 / 19

Another false-color IRIS spectroheliogram

Si IV 1394 A

AR 13449

16-step sparse raster

84 seconds per frame

Poor FOV, still not temporally-resolved


4 / 19

What could we be missing?

IRIS slit-jaw movie

C II 1330 A

Sit-and-stare program

AR 13450

0.15 seconds per frame



5 / 19
A diagram of a graph

The EUV Snapshot Imaging Spectrograph (ESIS)

A blue lines on a white background
6 / 19
A blue and yellow image Description automatically generated with medium confidenceA graph of different colored linesA screenshot of a computer gameA blue and green hexagons with white numbers

Computed Tomography Imaging Spectroscopy (CTIS) with ESIS

7 / 19
8 / 19

ESIS difference image

O V 630 A

Difference of Channels 2 and 3

Continuous explosive event activity over 5 minutes

9 / 19

Event E

Cotemporal AIA movies

Potential minifilament eruption

Complicated spectral morphology


10 / 19

ESIS Inversions

Analogue of a tomographic inversion problem

Spatial/spectral ambiguity

Limited number of angles

A grey and white striped backgroundA diagram of a mathematical equation
11 / 19

Multiplicative Algebraic Reconstruction Technique (MART) Inversion

Event C

Red/blue shifts are not cospatial or cotemporal

12 / 19

Another MART Inversion

Event D

13 / 19

Convolutional Neural Network inversion technique

Training/validation dataset

IRIS Si IV 1394 A 320-step dense rasters

Rebinned to ESIS resolution

64 training / 17 validation examples

Network Architecture

Fully-convolutional network

2 convolutional + 2 deconvolutional layers

11x11x11 kernel size

32 filters per layer

Left: Original IRIS spectrum (scaled to avg. O V width)

Right: IRIS spectrum rebinned to ESIS resolution

A comparison of a ray of light Description automatically generated with medium confidence
14 / 19
A close-up of a black and white image

Left: Spectrally-integrated IRIS raster, rebinned to ESIS resolution (ground truth)

Right: Synthetic ESIS observation calculated from the original IRIS raster

15 / 19
A collage of images of a field

Ground truth (left), reconstructed (middle), difference (right)

16 / 19
A group of images of a line Description automatically generated with medium confidence

Ground truth (left), reconstructed (middle), difference (right)

17 / 19
A graph of a graphA group of images of a graph Description automatically generated with medium confidence

Left: first four true moments vs. reconstructed moments

Right: Average validation spectrum (blue), average reconstructed spectrum (orange)

18 / 19

How can we leverage this observational capability to address the coronal heating problem?

Are there any observational signatures of nanoflares or waves that we be easier to detect over a wide FOV?

What line(s) should we observe?

What cadence, FOV, and resolution would be necessary?


19 / 19