Fitting accretion disk models to optical spectra of supermassive black holes: faster exploration of parameter degeneracies with nested sampling
PI: Charlotte Ward (Astronomy and Astrophysics)
Tuition and remaining salary can be covered by the PI’s start-up funds
The structure of gas disks that are accreted by supermassive black holes (SMBHs) can be revealed by optical spectroscopy (Chen &Halpern 1989). Efforts to fit models of accretion disk geometry and emissivity to large spectroscopic datasets are currently being limited by two computational challenges: the parameter space of disk properties and how they map to observable data is large and multi-degenerate, and it is computationally intensive to calculate the predicted spectrum from a given set of parameters with existing codes (Ward et al. 2024). This makes it difficult to fully explore the parameter space via comprehensive sampling of the posteriors. In this project, we aim to a) implement a faster version of pre-existing code that models the observed spectrum for a given set of disk parameters in python, jax or equivalent and b) implement a nested sampling framework to enable fast modeling of spectra and full characterization of posteriors. We aim to release this software publicly on Github, potentially with an accompanying paper submitted to the Journal of Open Source Software.
Expertise/skills of interest:
● Programming experience with python, jax, or alternative languages, as well as experience with Github. Ability to port a simple Fortran code to python. Ability to test and improve runtime performance and generate unit tests.
● Familiarity with nested sampling packages such as Ultranest.
● Useful but not required: familiarity with astronomical spectroscopic datasets.
Expectations:
● Post-comps graduate student or postdoc with at least some experience and/or training in: (1) Astronomy & Astrophysics, Physics or a related field; and (2) Applied Math, Computer Sciences, Data Sciences, IST, Statistics or a related field.
● Write code organized with appropriate documentation that will be released on Github.
● Weekly meeting and project updates with faculty advisor. Participate in group meetings every month.
Goal: Develop a faster version of a pre-existing Python and Fortran package that can predict a spectrum based on a given set of disk parameters in python/jax, and demonstrate efficient nested sampling for model fitting. We aim to release this package on Github and submit a JOSS paper on the package.
Specific Objectives:
● Port an old version of disk modeling code (currently Fortran code with an f2py wrapper so that users can interact with it via python) to be fully implemented in python and/or jax. Design the code so that it can explore multiple parameter sets in parallel to increase computational efficiency. Demonstrate sufficient speed for nested sampling with Ultranest or equivalent.
● Produce a public github repo with example notebooks where the modeling package is implemented on example spectra.
● If time permits, write a JOSS paper to accompany the public release of the code.
Engagement:
As a new faculty member looking to build connections with ICDS, Ward will participate in ICDS seminars and workshops, and will explore potential collaborations with ICDS co-hires, particularly those who are part of the Center for Astrostatistics & Astroinformatics.