Title:	Physics of the solar atmosphere
DNr:	SNIC 2014/1-273
Project Type:	SNIC Medium Compute
Principal Investigator:	Jorrit Leenaarts <jorrit.leenaarts@astro.su.se>
Affiliation:	Stockholms universitet
Duration:	2014-09-26 – 2015-04-01
Classification:	10305
Keywords:

Abstract

The Department of Astronomy at Stockholm University has recently started a theoretical/computational solar physics group consisting of an associated professor, 2 postdocs and 2 PhD students. The aim of the group is to understand the physics of the atmosphere of the Sun. Magnetism underlies much of the activity and dynamics of the solar atmosphere. Observations with observatories such as SDO, IRIS and the Swedish Solar Telescope show an intricate structure and dynamics form the smalls observable scale to the size of the entire Sun. These phenomena are ultimately driven by the nuclear fusion in the solar centre and the enigmatic solar dynamo. The physics of the atmosphere is complex, encompassing the interaction of (polarised) radiation and matter, magneto-hydrodynamics and plasma physics. Understanding the observations in terms of the physics thus requires intricate theoretic modelling. The project "Physics of the solar atmosphere" aims to contribute to our understanding of the solar atmosphere through (1) radiation-MHD modelling of the atmosphere, (2) precise modelling of the spectral lines from these MHD models that can be directly compared to observations and (3) inversion of observations. Inversion is an iterative technique where an atmospheric structure is derived that is consistent the observed spectral line profiles. The group uses a variety of codes for which the PI and Co-I are co-developers. All codes have been run before on supercomputers, and we thus expect no difficulties porting the codes to SNIC infrastructure. We plan to perform production runs, but we also have started two projects that aim to expand the modelling of spectral lines using the Multi3d code: implementing a multi-grid scheme to speed up the computations and implementing new physics, called partial frequency redistribution, to increase the realism of the modelling. This application can be considered as a test bed. In the beginning we will focus on porting the codes to new machines and do performance tests. We will then move to production runs and large scale testing of the new Multi3d capabilities. We thus expect not to fully use our allotted CPU time during the first months, but then ramp up as various parts of the project gain traction. As the research of the group gets underway we will apply for a Large grant if we need more computing time.