Introduction of SNID

This slides is used for a paper talk on the star cluster group's weekly meeting in KIAA-PKU. I talked about an algorithm illustrated in Blondin et al. (2007) for supernova classification -- SNID !

Transcript

1. Yuhan Yao 2017-04-21@KIAA SNID: Supernova Identification!
2. https://people.lam.fr/blondin.stephane/graphics/index.html
3. B band Maxima.
4. Type II: Hydrogen Type Ia: Si Type Ic Type Ib: He Fillippenko, 1997 And many subtypes…
5. ‘pec’=peculiar Ia-pec: 2002cx… II-pec: 1987A… And many subtypes… SNID Database
6. SNID Database 879 spectra of 65 SNe Ia 332 spectra of 19 SNe Ib/c 353 spectra of 10 SNe II Also galaxies, active galactic nuclei, stars, and novae. ⬆ In 2007 Larger now….
7. SNID Database
8. SNID Database Intrinsic Variance of https://people.lam.fr/blondin.stephane/graphics/index.html light blue: standard deviation dark blue: maximum deviation SNID Ia normal templates.
9. https://people.lam.fr/blondin.stephane/graphics/index.html Difficulty of observing SN Ia at high redshifts
10. Definition Supernova spectrum with redshift Template spectrum , Zero redshift Cross-correlation Maximize! n = 0, 1, 2, … , N-1
11. Preprocessing the SNe spectrum
12. Preprocessing the SNe spectrum
13. Definition Supernova spectrum with redshift Template spectrum , Zero redshift Cross-correlation — maximize
14. Redshift Estimate The r-Value · Spectrum Overlap Reliability
15. r-Value: correlation height-noise ratio
16. r-Value: correlation height-noise ratio
17. lap contours Redshift Estimate ❌ The r-Value · Spectrum Overlap
18. r · lap contours Redshift Estimate ❌ The r-Value · Spectrum Overlap
19. Initial and revised redshift estimates 1. All templates correlate with input, selects 10 highest peaks in . 2. Compute initial redshift , based on rlap-weighted median of all 3. Shift s(n) to , trim, search for highest peak in , find the revised redshift This time non-rlap-weighted median.
20. Save the width of the correlation peak. Redshift error 3. Shift s(n) to , trim, search for highest peak in , find the revised redshift This time non-rlap-weighted median.
21. Accuracy of Redshift & Age Simulation: choose normal Ia template spectra. Shifted: Contaminated: Added:
22. Accuracy of Redshift
23. Accuracy of Redshift
24. Accuracy of age
25. Accuracy of age
26. Type Ia normal
27. Type determination Normal vs. 1991T-like SNe Ia SNe Ia vs. SNe Ib/c Identifying SN ‘‘Oddballs’’ Using SNID for SN Identification
28. Type determination Normal vs. 1991T-like SNe Ia SNe Ia vs. SNe Ib/c Identifying SN ‘‘Oddballs’’ Using SNID for SN Identification
29. SN 1991T-like events: Local: up to 20% of SNe Ia. High redshift: goes down. Avishay, arXiv:1611.09353v1 SN 2011fe: regular Ia SNe.
30. no constraint ±0.01 ±3 both
31. Type determination Normal vs. 1991T-like SNe Ia SNe Ia vs. SNe Ib/c Identifying SN ‘‘Oddballs’’ Using SNID for SN Identification
33. Type determination Normal vs. 1991T-like SNe Ia SNe Ia vs. SNe Ib/c Identifying SN ‘‘Oddballs’’ Using SNID for SN Identification
34. Avishay, arXiv:1611.09353v1 SN 2002cx-like events: Early: resemble 1991T, Late: resemble normal. SN 1991T-like events: Local: up to 20% of SNe Ia. High redshift: goes down. SN 2011fe: regular Ia SNe.
36. – Dong, 2016, Science “SNID, a commonly used SN classification software that has a spectral library of most types of supernovae except SLSN, failed to find a good SN match. ”
37. Type determination Normal vs. 1991T-like SNe Ia SNe Ia vs. SNe Ib/c Identifying SN Ia ‘‘Oddballs’’ Using SNID for SN Identification
38. Interested? Download: https://people.lam.fr/blondin.stephane/software/snid/index.html Other methods (other than cross-correlation): Chi-square: superfit (Howell, 2005); SN-fit (Sainton, 2004); SFA (Riess, 1997) Bayesian: Kuznetsova & Connolly, 2007; Kunz, 2006 PCA: James, 2006 And improvements: Silverman, 2012, MNRAS, 425, 1789 Thank You!