Understanding the Dynamic ISM: A "Big Data'' Comparative Analysis of Models and Observations

• Applicant: Dr. Rahul Shetty
• Subject Area: Astrophysics and Astronomy
• Term: from 2014 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 263065473

Physical processes in the interstellar medium (ISM) play a critical role in determining the structure and dynamics of nearly all environments, from the formation of planets to the evolution of galaxies. Consequently, understanding the chemical, thermal, and dynamical state of the ISM remains one of the highest priorities in astrophysical research. Observations of gas and dust can reveal details about the ISM. However, inferring the physical properties from observed fluxes is not a trivial task, as measurement uncertainties, degeneracies, and inadequate model assumptions often greatly hinder the accurate interpretation of the data. Numerical simulations provide an avenue for studying the ISM, but their comparison with observations is again impeded by statistical and systematic uncertainties. Drawing on my extensive theoretical and statistical experience, I propose to systematically investigate the ISM by employing robust statistical and numerical methodologies.Through the development of novel statistical techniques, this research will bridge the observational and theoretical domains of ISM research. Namely, I will construct hierarchical Bayesian methods to rigorously analyze multi-wavelength observations simultaneously, and compare them with state-of-the-art hydrodynamic simulations. These techniques are significantly more accurate than standard algorithms because they account for statistical and systematic uncertainties, as well as degeneracies between model parameters. The statistical models will accurately estimate various ISM characteristics, such as densities, temperatures, star formation rates, as well as their correlations. Furthermore, we will apply these techniques to synthetic observations of advanced numerical simulations, including the effects of turbulence, time-dependent chemistry, gas self-gravity, and feedback from star formation. The simulations will include diverse environments, from quiescent regions in the outskirts of galaxies, to normal environments such as the local ISM, and starbursts commonly found at high redshift. Along with radiative transfer calculations, we will model the emission from dust and gas from the simulations. The resulting synthetic atomic, molecular, and dust emission datasets will be directly compared to recent observations under the Bayesian framework. This process will allow for a comprehensive assessment of the viability the models. Similarities between the models and observations will identify those dominant physical processes operating in the ISM. Taken together, the combination of numerical simulations and a rigorous statistical comparison with the most recent observations will substantially advance our understanding of the ISM in a variety of environments across the Universe.

• DFG Programme: Priority Programmes
• Subproject of: SPP 1573: Physics of the Interstellar Medium