24
6dFGS data quality: 6dFGS data quality: comparison of pipeline comparison of pipeline and IRAF redshifts and IRAF redshifts Lesa Moore Lesa Moore Macquarie University Macquarie University AAO 6dF Workshop 2005 AAO 6dF Workshop 2005

6dFGS data quality: comparison of pipeline and IRAF redshifts

  • Upload
    sinjin

  • View
    17

  • Download
    0

Embed Size (px)

DESCRIPTION

6dFGS data quality: comparison of pipeline and IRAF redshifts. Lesa Moore Macquarie University AAO 6dF Workshop 2005. Outline. Spectral reduction and S/N Wavelength calibration Cross-correlation redshift agreement Quality measures – S/N, Q, r Repeatability Final uncertainties - PowerPoint PPT Presentation

Citation preview

Page 1: 6dFGS data quality: comparison of pipeline and IRAF redshifts

6dFGS data quality:6dFGS data quality:comparison of pipeline and comparison of pipeline and

IRAF redshiftsIRAF redshifts

Lesa MooreLesa Moore

Macquarie UniversityMacquarie University

AAO 6dF Workshop 2005AAO 6dF Workshop 2005

Page 2: 6dFGS data quality: comparison of pipeline and IRAF redshifts

OutlineOutline Spectral reduction and S/NSpectral reduction and S/N Wavelength calibrationWavelength calibration Cross-correlation redshift agreementCross-correlation redshift agreement Quality measures – S/N, Q, rQuality measures – S/N, Q, r RepeatabilityRepeatability Final uncertaintiesFinal uncertainties Based on comparisons between Based on comparisons between

IRAF … and … 6dFDR/RUNZ processingIRAF … and … 6dFDR/RUNZ processing

Page 3: 6dFGS data quality: comparison of pipeline and IRAF redshifts

DataData

Three fields studied (A, B and C) with Three fields studied (A, B and C) with repeat observations of the B and C repeat observations of the B and C fields:fields: A: Mar 16 2002, reflection gratingsA: Mar 16 2002, reflection gratings B1: Sep 9 2002, reflection gratingsB1: Sep 9 2002, reflection gratings B2: Sep 29 2003, VPH gratingsB2: Sep 29 2003, VPH gratings C1: Sep 29 2002, VPH gratingsC1: Sep 29 2002, VPH gratings C2: Sep 18 2003, VPH gratingsC2: Sep 18 2003, VPH gratings

Page 4: 6dFGS data quality: comparison of pipeline and IRAF redshifts

MethodMethod

6dFDR on separate V and R spectra6dFDR on separate V and R spectra 6dFDR line lists6dFDR line lists

RUNZ on spliced spectraRUNZ on spliced spectra

IRAF dofibers on separate V and R spectraIRAF dofibers on separate V and R spectra My own line listsMy own line lists

IRAF xcsao on spliced spectraIRAF xcsao on spliced spectra

Batch mode processing, no heliocentric Batch mode processing, no heliocentric correction in either casecorrection in either case

Page 5: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Spectral ReductionSpectral Reduction

6dFDR reduction6dFDR reduction Sometimes requires FIT rather than Sometimes requires FIT rather than

TRAM extraction (slower)TRAM extraction (slower) Sometimes requires FLUX WEIGHTING Sometimes requires FLUX WEIGHTING

option turned offoption turned off

Page 6: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Signal to NoiseSignal to Noise

FieldField GratingGrating S/NS/N RemarksRemarks

B1 RefB1 Ref 316R ref316R ref 5.1245.124 Average Average for 68 for 68 galaxiesgalaxiesB2 VPHB2 VPH 425R VPH425R VPH 12.11912.119

B1 RefB1 Ref 600V ref600V ref 7.3137.313 Average Average for 69 for 69 galaxiesgalaxiesB2 VPHB2 VPH 580V VPH580V VPH 8.7958.795

SoftwareSoftware Field A, R dataField A, R data Field A, V dataField A, V data

6dFDR/splot6dFDR/splot 3.063.06 2.342.34

Dofibers/splotDofibers/splot 3.173.17 2.222.22

6dFDR/RUNZ6dFDR/RUNZ 10.3910.39 9.509.50

Dofibers/RUNZDofibers/RUNZ 10.2410.24 8.898.89

Cross-processed

Pipeline and IRAF reductions about equal

VPH data superior to reflection grating data (figures from 6dFDR/RUNZ)

Page 7: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Line lists andLine lists and calibration calibration Blue (V) data original line list had 14 Blue (V) data original line list had 14

lineslines 6dFDR typically locates 126dFDR typically locates 12 Software throws away 2 worst-fitting Software throws away 2 worst-fitting

(leaves 10)(leaves 10) For red (R) data For red (R) data

Line list omitted strong Ne line at 7032.41 Line list omitted strong Ne line at 7032.41 ÅÅ Arc spectra deficient of lines redward of Arc spectra deficient of lines redward of

7500 7500 Å (does not affect cross-correlation)Å (does not affect cross-correlation) Found 2Found 2ndnd order line from Hg at ~8092 Å order line from Hg at ~8092 Å

Page 8: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Calibration Calibration TestTest

Field C1 VPH reduced Field C1 VPH reduced without sky without sky subtractionsubtraction

Sky lines measuredSky lines measured 6dFDR results (6dFDR results (±0.4 ±0.4

Å) Å) equal to or equal to or superior to IRAFsuperior to IRAF

V 5577.34 Å Oxygen Line - 6dFDR

5576.55576.75576.95577.15577.35577.55577.75577.95578.15578.35578.5

0 50 100 150

Fibre

Wav

elen

gth

(Å)

R 5577.34 Å Oxygen Line - 6dFDR

5576.45576.65576.85577.05577.25577.45577.65577.85578.05578.25578.4

0 50 100 150

Fibre

Wav

elen

gth

(Å)

R 6300.30 Å Oxygen Line - 6dFDR

6299.36299.56299.76299.96300.16300.36300.56300.76300.96301.16301.3

0 50 100 150

Fibre

Wav

elen

gth

(Å)

R 6363.78 Å Oxygen Line - 6dFDR

6362.86363.06363.26363.46363.66363.86364.06364.26364.46364.66364.8

0 50 100 150

Fibre

Wav

elen

gth

(Å)

R 7243 Å OH Line - 6dFDR

7242.37242.57242.77242.97243.17243.37243.57243.77243.97244.17244.3

0 50 100 150

Fibre

Wav

elen

gth

(Å)

Page 9: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Cross-correlationCross-correlation

An earlier version of RUNZ at Epping An earlier version of RUNZ at Epping was applying the heliocentric was applying the heliocentric correction incorrectlycorrection incorrectly

RUNZ confused by noise in low S/N RUNZ confused by noise in low S/N spectraspectra

Spliced spectra much more reliable Spliced spectra much more reliable than separate R and V than separate R and V

Page 10: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Redshift AgreementRedshift Agreement

Field A spliced data

108 galaxies

y = 1.0016x + 1E-0592 galaxies

-0.1

0

0.1

0.2

0.3

0.4

-0.1 0.0 0.1 0.2 0.3

Field B1 REF spliced data

80 galaxies

y = 0.9993x + 0.0002 63 galaxies

-0.1

0

0.1

0.2

0.3

0.4

-0.1 0.0 0.1 0.2 0.3

Field C1 VPH spliced data

106 galaxies

y = 0.9989x + 0.000183 galaxies

-0.1

0

0.1

0.2

0.3

0.4

-0.1 0.0 0.1 0.2 0.3

IRAF redshifts

Pip

eli

ne

re

ds

hif

ts

Agreement ifAgreement if z| ≤ 0.0005z| ≤ 0.0005 cz| ≤ 150 km/scz| ≤ 150 km/s

Overall agreement 81% for Overall agreement 81% for 294 galaxies in 3 fields294 galaxies in 3 fields

Could still both be wrongCould still both be wrong need to check by eyeneed to check by eye

Page 11: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Quality MeasuresQuality Measures

Compared this redshift agreement Compared this redshift agreement with:with: S/NS/N RUNZ Q-ranking RUNZ Q-ranking Cross-correlation r-valuesCross-correlation r-values

as obtained from pipeline processingas obtained from pipeline processing

Page 12: 6dFGS data quality: comparison of pipeline and IRAF redshifts

S/NS/N

No strong No strong correspondenccorrespondence between e between redshift redshift agreement and agreement and S/N in separate S/N in separate R and V R and V spectraspectra

Field C1 VPH R data

106 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Field C1 VPH V data

103 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0IRA

F z

- P

ipe

lin

e z

RUNZ S/N

Field B1 REF R data

78 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Field B1 REF V data

81 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Field A R data

106 galaxies

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Field A V data

107 galaxies

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Page 13: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Q-rankingsQ-rankings

Large scatter Large scatter even with high even with high QQ

Q-rankingQ-ranking 00 11 22 33 44 55 TotalTotalss

DisagreesDisagrees 66 1212 55 2121 88 44 5656

AgreesAgrees 88 33 11 3232 7676 118118 238238} 294

Delta z vs RUNZ Q-ranking

0

30000

60000

90000

0 1 2 3 4 5 6

RUNZ Q-ranking

|IR

AF

cz

- R

UN

Z c

z| (

km/s

)

Page 14: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Q-rankingsQ-rankings Around half the “disagrees” have Around half the “disagrees” have

Q of 3, 4, or 5 Q of 3, 4, or 5

Q-rankingQ-ranking 00 11 22 33 44 55 TotalTotalss

DisagreesDisagrees 66 1212 55 2121 88 44 5656

AgreesAgrees 88 33 11 3232 7676 118118 238238} 294

Page 15: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Q-rankingsQ-rankings Around half the “disagrees” have Around half the “disagrees” have

Q of 3, 4, or 5 Q of 3, 4, or 5 Q=3 meant to imply 75% confidence Q=3 meant to imply 75% confidence

(only 60% agree)(only 60% agree)Q-rankingQ-ranking 00 11 22 33 44 55 TotalTotal

ss

DisagreesDisagrees 66 1212 55 2121 88 44 5656

AgreesAgrees 88 33 11 3232 7676 118118 238238} 294

Page 16: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Cross-correlation Cross-correlation r-value r-value

R-value is a much moreR-value is a much more useful indicator of redshift useful indicator of redshift reliability than Q-rankingreliability than Q-ranking

84% of disagrees have r<684% of disagrees have r<6 27 of 33 disagrees with 27 of 33 disagrees with

Q=3,4,5 have r<6Q=3,4,5 have r<6 72% of agrees have r>672% of agrees have r>6

Field C1 VPH spliced data

106 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0IRA

F z

- P

ipe

lin

e z

RUNZ r-value

Field B1 REF spliced data

80 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Field A spliced data

108 galaxies

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0

Page 17: 6dFGS data quality: comparison of pipeline and IRAF redshifts

SIMBAD-RUNZ difference SIMBAD-RUNZ difference vs. Q-rankingvs. Q-ranking

Q – rankings of 3, 4, 5 show large spread of Q – rankings of 3, 4, 5 show large spread of error values (4, 14, 29 data points respectively)error values (4, 14, 29 data points respectively) Have applied heliocentric correction this timeHave applied heliocentric correction this time Note bias towards high-q resultsNote bias towards high-q results

SIMBAD - RUNZ Redshifts50 Galaxies

-100000

-50000

0

50000

100000

150000

200000

250000

300000

0 1 2 3 4 5 6

RUNZ Q

SIM

BA

D -

RU

NZ

cz

(km

/s)

SIMBAD - RUNZ Redshifts

-600-500-400-300-200-100

0100200300400500600

0 1 2 3 4 5 6

RUNZ Q

SIM

BA

D -

RU

NZ

cz

(km

/s)

Page 18: 6dFGS data quality: comparison of pipeline and IRAF redshifts

SIMBAD-RUNZ difference SIMBAD-RUNZ difference vs. R-valuevs. R-value

Differences scale inversely with r-valuesDifferences scale inversely with r-values St dev: St dev: z z = 0.00012, = 0.00012, km/s km/s = 52 km/s = 52 km/s

(based on 34 galaxies whose redshift agreement (based on 34 galaxies whose redshift agreement meets criterion of meets criterion of cz| ≤ 150 km/s)cz| ≤ 150 km/s)

SIMBAD - RUNZ Redshifts50 Galaxies

-100000

-50000

0

50000

100000

150000

200000

250000

300000

0 6 12 18 24 30 36

R-value

SIM

BA

D -

RU

NZ

cz

(km

/s)

SIMBAD - RUNZ Redshifts

-600-500-400-300-200-100

0100200300400500600

0 6 12 18 24 30 36

R-value

SIM

BA

D -

RU

NZ

cz

(km

/s)

Page 19: 6dFGS data quality: comparison of pipeline and IRAF redshifts

RepeatabilitRepeatabilityy

End columns End columns are large are large discrepanciesdiscrepancies

All r>6 results All r>6 results lie within lie within z| z| ≤ 0.001≤ 0.001

Overall st.dev. Overall st.dev. zz = 0.00033 = 0.00033

km/skm/s = 98 = 98 km/skm/s

Redshift Agreement - Fields B1 REF & B2 VPH

02468

101214

-0.0

5

-0.0

05

-0.0

01

-0.0

005

-0.0

004

-0.0

003

-0.0

002

-0.0

001 0

0.00

01

0.00

02

0.00

03

0.00

04

0.00

05

0.00

1

0.00

5

0.05

>0.0

5

Redshift difference (z)

Nu

mb

er o

f g

alax

ies

68 common galaxies 32 galaxies with r > 6

Redshift Agreement - Fields C1 VPH and C2 VPH

02468

101214161820

-0.0

5

-0.0

05

-0.0

01

-0.0

00

5

-0.0

00

4

-0.0

00

3

-0.0

00

2

-0.0

00

1 0

0.0

00

1

0.0

00

2

0.0

00

3

0.0

00

4

0.0

00

5

0.0

01

0.0

05

0.0

5

>0

.05

Redshift difference (z)

Nu

mb

er

of

ga

lax

ies

80 common galaxies 49 galaxies with r > 6

Page 20: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Final uncertaintiesFinal uncertainties

NotesNotes 1. Based on 0.4 1. Based on 0.4 ÅÅ at 4000 at 4000 ÅÅ 2. Mean of “verr” from RUNZ (278 galaxies in total, 190 with r>6, possibly over-stated)2. Mean of “verr” from RUNZ (278 galaxies in total, 190 with r>6, possibly over-stated) 3. 1/√2 * st. dev.(3. 1/√2 * st. dev.(z) of repeat 6dF observations (125 galaxies in total, 81 with r>6)z) of repeat 6dF observations (125 galaxies in total, 81 with r>6) 4. Added in quadrature4. Added in quadrature kms/s = z * 300,000 assumed in all caseskms/s = z * 300,000 assumed in all cases

SourceSource General General UncertaintyUncertainty

Uncert. for r>6 Uncert. for r>6 redshiftsredshifts

±±km/skm/s ±±zz ±±km/skm/s ±±zz

calibrationcalibration11 3030 0.00010.0001 3030 0.00010.0001

Cross-correlationCross-correlation22 6363 0.000210.00021 4848 0.000160.00016

RepeatabilityRepeatability33 6969 0.000230.00023 3939 0.000130.00013

Total uncertaintyTotal uncertainty44 9898 0.000330.00033 6969 0.000230.00023

The “6dF Galaxy The “6dF Galaxy Survey” Paper Survey” Paper (2004)(2004)

82 km/s for Q = 382 km/s for Q = 3 46 km/s for Q = 446 km/s for Q = 4

Page 21: 6dFGS data quality: comparison of pipeline and IRAF redshifts

Final uncertaintiesFinal uncertainties

NotesNotes 1. Based on 0.4 1. Based on 0.4 ÅÅ at 4000 at 4000 ÅÅ 2. Mean of “verr” from RUNZ (possibly over-stated)2. Mean of “verr” from RUNZ (possibly over-stated) 3. 1/√2 * st. dev.(3. 1/√2 * st. dev.(z)z) 4. Added in quadrature, final results rounded to one significant figure4. Added in quadrature, final results rounded to one significant figure kms/s = z * 300,000 assumed in all caseskms/s = z * 300,000 assumed in all cases

SourceSource General General UncertaintyUncertainty

Uncert. for r>6 Uncert. for r>6 redshiftsredshifts

±±km/skm/s ±±zz ±±km/skm/s ±±zz

calibrationcalibration11 3030 0.00010.0001 3030 0.00010.0001

Cross-correlationCross-correlation22 6363 0.000210.00021 4848 0.000160.00016

RepeatabilityRepeatability33 6969 0.000230.00023 3939 0.000130.00013

Total uncertaintyTotal uncertainty44 100100 0.00030.0003 7070 0.00020.0002

The “6dF Galaxy The “6dF Galaxy Survey” Paper Survey” Paper (2004)(2004)

82 km/s for Q = 382 km/s for Q = 3 46 km/s for Q = 446 km/s for Q = 4

Page 22: 6dFGS data quality: comparison of pipeline and IRAF redshifts

SummarySummary

S/N superior with VPH gratingsS/N superior with VPH gratings RUNZ q-ranking RUNZ q-ranking notnot reliable indicator of z quality reliable indicator of z quality Redshift agreement (IRAF-RUNZ) scales strongly Redshift agreement (IRAF-RUNZ) scales strongly

with cross-correlation r-valuewith cross-correlation r-value Small uncertainty in Small uncertainty in calibration test calibration test Larger uncertainties in cross-correlation and Larger uncertainties in cross-correlation and

repeatability testsrepeatability tests Total uncertainties:Total uncertainties:

100 km/s general uncertainty100 km/s general uncertainty 70 km/s uncertainty for r>6 redshifts70 km/s uncertainty for r>6 redshifts

Page 23: 6dFGS data quality: comparison of pipeline and IRAF redshifts

ThanksThanks

Macquarie UniversityMacquarie University Anglo-Australian ObservatoryAnglo-Australian Observatory Wide-Field Astronomy Unit, EdinburghWide-Field Astronomy Unit, Edinburgh

Supervisors: Quentin Parker (MU/AAO), Will Saunders Supervisors: Quentin Parker (MU/AAO), Will Saunders (AAO)(AAO)

6df Galaxy Survey Team (37 members)6df Galaxy Survey Team (37 members)

References References The 6dF Galaxy Survey: samples, observational techniques and the The 6dF Galaxy Survey: samples, observational techniques and the

first data releasefirst data release, MNRAS, , MNRAS, 355355, 747-763 (2004), 747-763 (2004) This research has made use of the SIMBAD database, operated at This research has made use of the SIMBAD database, operated at

CDS, Strasbourg, FranceCDS, Strasbourg, France

Page 24: 6dFGS data quality: comparison of pipeline and IRAF redshifts