I research type Ia supernova (SN Ia) empirical modeling under Professor Michael Wood-Vasey and in collaboration with Dr. Alex Kim at the Lawrence Berkeley National Laboratory. Dr. Kim and I have developed a new two color-component model that disentangles intrinsic and extrinsic phase-independent SN Ia spectra variation through a novel application of geometric algebra. This model is implemented with the programming language Stan and trained using a Hamiltonian Monte Carlo sampler. 

Our incomplete theoretical understanding of SN Ia progenitor systems and explosion dynamics necessitates the use of empirical SN Ia models trained from observation.  Separating phase-independent SN Ia variational modes has historically proven particularly challenging, and our original application of geometric algebra provides a powerful framework to decoupling these modes. Indeed, all past empirical SN Ia models have either made strong assumptions about extrinsic variation (such that all phase-independent variation is due to dust extinction) or have fit for a single effective ‘color’ template that captures all modes of phase-independent variation. Our model is the first to instead fit simultaneously for two color templates, recovering expected dust-sourced variation and dust-minimal variation lining up with known SN Ia spectral features. 

These findings complement recent strides made in nonlinear and phase-dependent modeling while constraining progenitor scenarios. We have also found that including intrinsic color-dominated SNe Ia when estimating effective selective extinction can drive down one’s estimate, partly explaining the historically low fit Rv values from past SN Ia analyses. 

Previously in the Wood-Vasey working group, I wrote a paper summarizing our results regarding SN Ia host galaxy bias dependence on observation method and host property fitting technique. 

We found that the host galaxy bias and resulting mass/sSFR step bias in standardized absolute magnitudes is not influenced by observation type or the technique used to estimate host properties.

(9/2022)