Colloquium Abstract - Xue - 2024Apr19
April 19, 2024
11:00am Mountain
Ceci Xue (MIT)
Molecular Spectra Diagnostics with a Bayesian MCMC Approach
Abstract
The recent improvement in receiver technology within modern facilities has enabled us to efficiently perform wide-band and high-sensitive molecular line surveys. To better extract the information from these wide-band spectral data, we introduce a molecular signal diagnostic tool coupling non-LTE radiative transfer models and a Markov Chain Monte Carlo (MCMC) scheme. In contrast to a canonical least-squares fit approach, MCMC analyses allow a more efficient exploration of the physical parameter space and provide access to the parameter's probability distribution, which can be used to characterize the confidence intervals and covariances between parameters.
In this talk, following a brief introduction to Bayesian statistics, we will present a case study demonstrating the analysis of molecular line observations from ALMA using this tool. Built upon RADEX (van der Tak et al. 2007), our tool features novel implementations to support multiple components along the line of sight and allow Bayesian inference about physical characteristics. We will share the first detection and mapping of the Class I methanol maser at 84 GHz toward the north region of Sagittarius B2 molecular cloud. Our analyses include locating the regions where the maser emission originates and assessing their observed spectral profiles respectively. The results suggest a chained two-component model for explaining the intense methanol Class I maser emission toward a region with weak continuum background radiation. In addition, our diagnostic tool also supports thermal line modeling. This module leveraging the MCMC approach will be applied to the spectral line survey, GOTHAM, to conduct a statistical study of the molecular census of the molecular dark cloud TMC-1. Recent surveys since 2020 have reported the detection of 60 new molecular species in this region, making it the most prolific source for interstellar molecular discoveries.
Local Host: Emmanuel Momjian
