SASHA Simulations of alvenic solar heating

Lead Research Organisation: University of Glasgow
Department Name: School of Physics and Astronomy

Abstract

SASHA : Simulations of alvenic solar heating

Publications

10 25 50
publication icon
Molnar M (2020) Spectral Deconvolution With Deep Learning: Removing the Effects of Spectral PSF Broadening in Frontiers in Astronomy and Space Sciences

publication icon
Osborne C (2021) On the importance of Ca ii photoionization by the hydrogen lyman transitions in solar flare models in Monthly Notices of the Royal Astronomical Society

publication icon
Osborne C (2019) Thyr : a volumetric ray-marching tool for simulating microwave emission in Monthly Notices of the Royal Astronomical Society

Studentship Projects

Project Reference Relationship Related To Start End Student Name
ST/R504750/1 30/09/2017 29/09/2021
1947174 Studentship ST/R504750/1 30/09/2017 14/09/2021 Christopher Montaque Osbourne
 
Description This project has allowed for radiative modelling of the flaring solar chromosphere with more advanced techniques than has previously been possible, and has enhanced our understanding of the formation of optical spectral lines in this region of the solar atmosphere. These spectral lines are our only method by which the thermodynamic properties of the flaring atmosphere can be determined, and as such, it is important to model these as accurately as possible to both develop our theoretical models of these events and for comparison with observations. In particular, we performed the first time-dependent model of the effect of flaring radiation on an adjacent two-dimensional slab of chromospheric plasma. It was found that enhanced radiative signatures could be observed over a megametre from the flare location -- a scale that can easily be observed by modern solar observatories.

In addition to improvements in modelling, our machine learning work has enabled, for the first time, a method for directly relating observations of these flaring spectral lines to plausible model atmospheres, enhancing a previously manual forward-fitting process. With modest computing resources this method can be employed to investigate the atmospheric parameters related to the large, high-resolution spectroscopic fields of view produced by modern telescopes.

The development of the Lightweaver radiative transfer framework, which has underpinned many of these scientific developments, has removed significant barriers to the creation of novel radiative transfer simulations for non-experts and is starting to gain traction in the solar physics research community.
Exploitation Route The outcomes of this award are directly relevant to the solar and stellar physics communities, in particular the modelling of solar flares. Several recommendations for improving the treatment of physical details in solar flare models can be taken into account in the improvement of existing models, as well as the development of new modelling techniques, including the new three-dimensional radiative magnetic hydrodynamic models in development at multiple institutions.

The predictions of visibly enhanced spectral line profiles far from the flare boundary can be investigated using modern solar observations either with a dedicated set of observations taken from a modern solar observatory, or by analysis of existing data.

The machine learning technique developed over the course of this award should see wide adoption for the analysis of ground-based solar flare observations, which are now commonly observed at high cadence and resolution.
Sectors Other

 
Description STFC LTA Funding
Amount £4,307 (GBP)
Funding ID ST/R504750/1 
Organisation Science and Technologies Facilities Council (STFC) 
Sector Public
Country United Kingdom
Start 07/2019 
End 12/2019
 
Description Collaboration with National Solar Observatory including 4 month research visit. 
Organisation National Solar Observatory (NSO)
Country United States 
Sector Public 
PI Contribution This was a research trip (Aug-Dec 2019) to develop my knowledge of radiative transfer with the experts present at NSO, whilst contributing to the development of future machine learning based inversions through my machine learning and programming expertise. I contributed to future inversion projects, investigating methods of deconvolving spectral information from the instrumental point spread function using machine learning, and constructed a new python-based framework for solar radiative transfer calculations.
Collaborator Contribution I primarily worked with Ivan Milic, and we worked on developing a machine learning based model for inversions driven by a Markov Chain Monte Carlo model as well as the python radiative transfer code, and drafting a research paper describing this new framework.
Impact The primary output is the publicly available radiative transfer tool, Lightweaver, described in the software section. This has also been described in the Lightweaver research paper (listed in publications).
Start Year 2019
 
Description ISSI: Interrogation of Field Aligned Solar Flare Models 
Organisation International Space Science Institute (ISSI)
Country Switzerland 
Sector Academic/University 
PI Contribution This project aims to compare, contrast, and improve the existing field-aligned computer models of solar flares, by comparing them to each other and also to observations. I am participating in a comparison of the models using machine learning, and testing and improvements of the radiative transfer treatments used in these models. The machine learning comparison was found to not be effective, compared to the time investment needed to train the models. The comparison between the models has now occurred and shows promising agreement. I have investigated the radiative origin of the differences that have been found between the models, which has revealed a possible oversight in a popular model. We have shown that this can affect the energy balance in the simulated atmosphere and described this in a research paper. I have also investigated how to treat the modelled atomic populations with both time-dependence and partial frequency redistribution in their spectral lines simultaneously. The results of this modelling are promising. This collaboration will meet again later in 2022 and discuss the improvements made to the models, the differences that were found, along with further improvements that may be needed to prepare the models to aid in the interpretation of the large influx of data from next-generation solar observatories.
Collaborator Contribution The other contributors to this project include the original creators of the main flare simulation codes who are contributing their understanding of flare physics and modelling. Other members of the team, who are also experienced in the use of these codes are running the workshops and providing observation expertise and insight. The next step of the project is a "shoot-out" between the different codes and attempting to investigate where, and why, differences arise. The early results of this comparison show very promising agreements.
Impact Research Paper: On the Importance of Ca ii Photoionisation by the Hydrogen Lyman Lines https://doi.org/10.1093/mnras/stab2156
Start Year 2020
 
Description Research Trip to Improve Knowledge of RADYN 
Organisation University of Oslo
Country Norway 
Sector Academic/University 
PI Contribution This brief (2-week) research trip was undertaken to work with Prof. Mats Carlsson and improve my knowledge of his flare simulation code RADYN, as well as some of the techniques used in its development, and discussing avenues for further development.
Collaborator Contribution Prof. Mats Carlsson contributed many hours to explaining how the RADYN code works, in addition to productive discussions with other members of the solar group in Oslo, on the topics of radiative transfer and modern radiation hydrodynamics.
Impact None yet
Start Year 2018
 
Title Lightweaver 
Description Lightweaver is an optically thick radiative transfer framework, written in python, allowing a much simpler and faster method of constructing radiative transfer tools for a variety of solar physics use cases, i.e. synthesising the outgoing radiation from the solar atmosphere under the assumption of one-dimensional plane parallel geometry. It is developed from, and validated against the most commonly accepted reference code, RH, as well as an additional code, SNAPI, that implements the same basic method but fewer features. It has been extended to support polarised radiation, as well as optically thick radiative transfer in two dimensions and the investigation of time-dependent problems. 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact Lightweaver forms the basis of several sub-projects associated with this award, and is being used to investigate the spectral signatures of heating the solar atmosphere in a more flexible manner than was previously available. It has been employed to highlight deficits in our previous treatments of optically thick radiative transfer in solar flare models, as well as initial investigations of the effects of flaring radiation on the adjacent solar atmosphere. Due to its ease of use and flexibility, allowing projects that were not previously feasible without large modifications to existing tools, Lightweaver has rapidly gained users around the world, and is in use in at least six solar physics research groups. 
URL https://github.com/Goobley/Lightweaver
 
Title Radynversion 
Description Radynversion is the implementation of an invertible neural network in python, that can be used to infer the atmospheric parameters during a solar flare from observations of the Hydrogen-alpha and Calcium-8542 Angstrom spectral lines, under the assumption that the RADYN code is a good description of the physics occuring in a solar flare. 
Type Of Technology Software 
Year Produced 2019 
Open Source License? Yes  
Impact Radynversion has allowed credible inversions of the chromosphere in a solar flare, for the first time without the limiting assumptions of hydrostatic equilibrium and statistical equilibrium, which are unlikely to hold in flares. Initial results are described in the Radynversion paper. 
URL https://github.com/Goobley/Radynversion