Stripped envelope supernovae classification at low spectral resolution

Abstract

Future astrophysical photometric surveys, like the Rubin Observatory Legacy Survey of Space and Time, will discover tens of thousands of astrophysical transients and hundreds of supernovae every night. This puts a significant strain on existing spectroscopic resources and motivates studies that may lead to increasing the efficiency of spectroscopic follow up. While other supernovae classes can be distinguished photometrically, Stripped Envelope Core Collapse Supernovae require spectra for classification. We attempted to classify Stripped Envelope Core Collapse Supernovae subtypes at increasingly low resolution to find a critical resolution value at which classification is not possible. Surprisingly, the accuracy of machine learning classifier, which is around 75% at the original R ~ 800 resolution, only decreases slightly and accuracy scores close to 50% are observed even at resolution R ~ 20. Upon investigating the low resolution spectra and testing the He signature in SNe subtypes Ib and IIb at phase 15 ± 5 days, we found encouraging evidence of information retained in the signatures associated to the same spectral features used for high-resolution classification. Further investigation in low-resolution SNe classification is required which may lead to engineering recommendations and improvements in the throughput of large photometric surveys by maximizing the efficiency of follow-up studies.