On the Scalar Field Tomography for the Solar Corona |
Maxim Kramar
(The Catholic University of America, NASA-GSFC)
|
Shaela Jones
(University of Maryland & NASA-GSFC, USA)
|
J. Davila
(NASA-GSFC, USA)
|
B.Inhester
(Max-Planck-Institut fuer Sonnensystemforschung, Katlenburg-Lindau, Germany)
|
| Abstract |
| Since the solar corona is optically thin, coronal observations are essentially integrated over the line-of-sight (LOS). It is therefore impossible to resolve the structure of the corona along the LOS if observations are provided from a single view direction. Observations from different view positions are necessary to reconstruct 3D coronal structure and is the essence of the tomography inversion method. When having observations only from a single view direction, a rigid rotation of the coronal density structures with the Sun about the ecliptic must be assumed in order to apply the tomography technique. As a consequence, only structures which are stationary over half a solar rotation can be reconstructed. The STEREO observations allow us to reduce this stationarity assumption.
Depending on the size of the reconstructing domain, the number of the observations, data noise, etc, the reconstruction usually is not unique. To stabilize the solution, we introduce regularization into the inversion which is essentially a smoothness factor for the density reconstruction. The introduced regularization is not isotropic in order to prevent smoothing the radial structures often observed in the corona.
Unlike usual scalar field tomography, which is essentially two-dimensional, our regularization concept and also the tilt of the Sun\'s axis with respect to the ecliptic requires to perform operations over the entire 3D domain during every iteration, which increases the computational resources needed for the inversion. To increase computational performance and resolution of the reconstructing domain, we have implemented a MPI-parallelization for the inversion code. The code is going to be used with STEREO COR1 data.
|