MEGAMORPH. Measuring messy galaxies with a non-parametric component. Steven Bamford. Marina Vika, Boris Häußler, Alex Rojas QNRF

Transcription

1 MEGAMORPH Measuring messy galaxies with a non-parametric component Steven Bamford Marina Vika, Boris Häußler, Alex Rojas QNRF

2 Two distinct components disks MEGAMORPH spheroids

3 MEGAMORPH More meaningful profile fits to galaxies in very large imaging surveys

4 MEGAMORPH Making being reckless more reliable

5 Galaxies are simple

6 Galaxies are simple Sérsic profiles still work well

7 Galaxies are messy

8 Messiness is in M/L Zibetti, Charlot & Rix, 2009, MNRAS, 400, 1181

9 Messiness is in M/L Zibetti, Charlot & Rix, 2009, MNRAS, 400, 1181

10 Wuyts et al. 2012, ApJ, 753, 114 Messiness is in M/L

11 What can we do? data smooth model

12 What can we do? data smooth model + something else

13 Dust and wider wavelengths... not yet. 3D radiative transfer models but still smooth

14 original image, I Non-parametric components create working image, W = I start fitting parametric model, P, to working image, W continue fitting parametric model, P, to working image, W create new working image: W = I N blank initial non-parametric model, N create residual image: R = I P update nonparametric model: N ~ filter(r) gradual build-up wavelet filtering SED homogenisation

15 Non-parametric components Tunable parameters: schedule for build-up of nonparametric image wavelet filtering threshold SED homogenisation on/off Reasonable defaults.

16 Multi-band Optional SED homogenisation

17 GALFITM input file

18 GALFITM input file

19 Some specific examples Using a non-parametric component to remove spiral arms SED homogenisation on

20 data standard model data - nonparam non-param model mask standard residuals nonparam image nonparam residuals NGC 4321 SDSS z=0.05

21 data standard model data - nonparam non-param model mask standard residuals nonparam image nonparam residuals NGC 4321 SDSS z=0.05

22 data standard model data - nonparam non-param model component SEDs mask standard residuals nonparam image nonparam residuals NGC 4321 SDSS z=0.05

23 Specific examples Using a non-parametric component to remove bar features and other oddities SED homogenisation on

24 NGC 5850 SDSS r-band original data - intermediate scale

25 NGC 5850 SDSS z=0.05 artificially-redshifted data

26 nonparametric image

27 NGC 5850 SDSS r-band original data - faint scale

28 NGC 5850 SDSS r-band original data - intermediate scale

29 NGC 5850 SDSS r-band original data - bright scale

30 NGC 5850 SDSS r-band original data - bright scale - zoom

31 NGC 5850 SDSS r-band nonparametric image - zoom

32 General parameter improvement Standard versus non-parametric SED homogenisation on

33 sersic index General parameter improvement 1e standard Standard 4.0 sindex_serr_ e Nearby galaxies, artificiallyredshifted to z= mag_serr_1 magnitude Both datasets masked

34 sersic index General parameter improvement 1e nonparam Nonparam 4.0 sindex_serr_ e Nearby galaxies, artificiallyredshifted to z= mag_serr_1 magnitude Both datasets masked

35 sersic index General parameter improvement 1e nonparam Nonparam standard Standard 4.0 sindex_serr_ e Nearby galaxies, artificiallyredshifted to z= mag_serr_1 magnitude Both datasets masked

36 Specific examples Using a non-parametric component instead of masking neighbours SED homogenisation off

37

38

39

40

41 General parameter improvement Non-parametric as alternative to masking SED homogenisation off

42 magnitude unmasked fit General parameter improvement 20 nonparam standard mask vs nomask nonpar mask vs nomask magr_ Nearby galaxies, artificiallyredshifted to z=0.05 Both datasets unmasked magr_1 magnitude standard masked fit

43 General parameter improvement 1e nonparam standard mask vs nomask nonpar mask vs nomask Re unmasked fit reffr_2 1e e e Nearby galaxies, artificiallyredshifted to z=0.05 Both datasets unmasked 5.0 1e reffr_1 Re standard masked fit

44 n unmasked fit General parameter improvement 1e nonparam standard mask vs nomask nonpar mask vs nomask sindexr_ e Nearby galaxies, artificiallyredshifted to z=0.05 Both datasets unmasked e sindexr_1 n standard masked fit

45 A stringent test Exponential + Sérsic Standard

46

47

48

49

50 A stringent test Exponential + Sérsic Non-parametric

51

52

53

54

55

56

57 MEGAMORPH Non-parametric component Simple to turn on in GALFITM Effective at removing problematic galaxy features Automatic inconsistency masking Works well in specific cases (Less well in some cases) Overall produces more meaningful measurements Demonstration paper in prep. Software available:

58 Sérsic profiles will be around for a long time yet