mwvgroup / csp-cmfgen

A comparison spectral SNe models against CSP data
https://mwvgroup.github.io/nir-comparison/
0 stars 0 forks source link

Supernova Model Comparison in the NIR

A comparison of CMFGEN supernova simulations against data from the Carnegie Supernova Project (CSP), particularly in the NIR.

Project Goals

Due to the increased observation time required for spectroscopic observations, the number spectroscopically observed supernovae that are publicly available is significantly smaller than those with photometric observations. This imbalance is expected to worsen over the next decade with large scale photometric surveys like the Large Synoptic Survey Telescope (LSST) coming online in 2021. This is a particular challenge when modeling the astrophysical channels that result in a Type Ia supernova (SNe Ia). Spectroscopic observations are crucial in constraining SNe Ia model due to their increased wavelength resolution.

We attempt to predict how well we will be able to constrain spectroscopic SNe Ia models using the predominantly photometric data that will be available in coming years. As a baseline, we use observations taken by the Carnegie Supernova Project primarily due to its employment of simultaneous spectroscopic and photometric observing campaigns. This includes spectroscopic data published in the first CSP data release (DR1) and photometric observations published in the third (DR3).

Our analysis is broken down into four key steps: two spectroscopic and one photometric. We start by comparing modeled spectra against flux values from DR1 observations. This serves as a baseline for understanding how well the models we consider represent our observed targets on a physical level. Second, we supplement this baseline by comparing the predicted pseudo equivalent widths of different features considered in the DR1 release paper. Thirdly, we perform light curve fits of DR3 photometry and compare our ability to fit in the optical vs the NIR. Finally, we consider how well the color evolution of each model matches the photometric data.

Running the Analysis Pipeline

The analysis pipeline for this project can be accessed either from within a python environment or from the command line. The python interface recommended when trying to extend or reproduce individual steps of the analysis. For usage examples see the notebooks/ directory. The command line interface is intended for running complete sections of the analysis. For more instructions see python run_analysis.py -h.

Notes on the Models

A copy of the CMFGEN models in ascii format is provided in the ascii_models/ directory. A version of these models ported for use with SNCosmo is available via the analysis.models module (see the docs for more info). It is important to note that these two models are not identical. The wavelength grid in the asccii models changes with phase, where as the SNCosmo models have been interpolated onto a common wavelength grid. Both models represent the flux of a supernova at 1 kpc for a given time and wavelength.