SDSS Photometric Calibration Revisited A look in the rear view - - PowerPoint PPT Presentation

▶

Apr 11, 2023 252 likes •461 views

SDSS Photometric Calibration Revisited A look in the rear view mirror John Marriner 18 April 2012 Cosmology with SN The desire to determine SN magnitudes from measurements of different parts of the spectrum is a driver for precise

SLIDE 1

SDSS Photometric Calibration Revisited A look in the rear view mirror

John Marriner 18 April 2012

SLIDE 2

Cosmology with SN

 The desire to determine SN magnitudes from measurements

f different parts of the spectrum is a driver for precise

photometric calibration.  Comparison of different experiments requires an “absolute” calibration.  Photometry is typically difference photometry relative to field stars.  SN magnitudes vary by ~7%

 Errors are reduced by averaging many SN.  Random calibration errors are a minor concern.  Even small systematic biases can cause problems.

SLIDE 3

SDSS SN Calibration Strategy

 Report “native magnitudes”  Measure filter response (relative throughput)  Calibrate absolute response with standard stars whose

 Spectrum is well measured.  Photometric response of the SDSS telescope is measured.

 Measurements of the standard stars are reported as an “AB offset”  Consider only stripe 82

SLIDE 4

Details

 The Photometric Telescope (PT)

 Measures atmospheric extinction based on primary standards (the USNO standards).  Measures unknown field stars for use as secondary calibration standards.

 The 2.5 m Survey Telescope is calibrated using the secondary calibration standards measured by the PT.  The 2.5 m Telescope is normally operated in drift scan mode.

SLIDE 5

Filter Measurements

From Doi (2010)

SLIDE 6

More Filters

SLIDE 7

Flat-field Stability

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

# #$#" #$#6 #$#7 #$#8 #$#% #$#9

:;4/<(,--'3/*+/='3.>.*+-

? 2 3 * @

SLIDE 8

Rms Measurement Variation

SLIDE 9

PT Flat Field

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-12)+./,3)1/01-04,-.+150,-/,-6178)62))9:

!"#% !"#$ " "#$ "#%

(,;;)2)-*)1,-1<.6-,/=8)

7!:

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-12)+./,3)1/01-04,-.+150,-/,-6178)62))9:

!"#% !"#$ " "#$ "#%

(,;;)2)-*)1,-1<.6-,/=8)

7!:

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-12)+./,3)1/01-04,-.+150,-/,-6178)62))9:

!"#% !"#$ " "#$ "#%

(,;;)2)-*)1,-1<.6-,/=8)

7!:

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-12)+./,3)1/01-04,-.+150,-/,-6178)62))9:

!"#% !"#$ " "#$ "#%

(,;;)2)-*)1,-1<.6-,/=8)

7!:

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-12)+./,3)1/01-04,-.+150,-/,-6178)62))9:

!"#% !"#$ " "#$ "#%

(,;;)2)-*)1,-1<.6-,/=8)

7!:

!"#$ " "#$ "#% "#& "#'

()*+,-./,0-123)45))67

!"#"$ !"#"8 " "#"8 "#"$ "#"9

:055)*/,0-12;.4-,/<3)7

SLIDE 10

What is the Ideal Star Flat Dither Pattern?

 Rely on dome flats for high spatial scales.  Dither by 1/2, 1/4, 1/8, …?

f x

( ) =

aneinkx

∑

f x + h

( )− f (x) =

an 1− einkh

( )einkx

∑

⇒ an = 1 1− einkh e−inkx f x + h

( )− f (x)

[ ]

∫

dx

SLIDE 11

PT Response Map

SLIDE 12

A Tale of 2 Catalogs

 Stripe 82 Coadd (Ivezic)

 Uses PT response for magnitude (r-band)  Uses stellar locus for colors  Multiple measurements, outliers & variables rejected

 DR8 (Ubercal)

 “Insensitive” to PT response  Color uncalibrated  Single epoch catalog

SLIDE 13

Coadd/DR8 Comparison

SLIDE 14

PT Observation Times

!"# !$# !%# # %# $# "#

&'()*+,-./0-'10+23/(4//-5

# % $ " 6 7# 7% 7$

8'9/+2:8+)1;4-5

SLIDE 15

Comparison Coadd/DR8

fits as a function of declination

SLIDE 16

Variability of the Calibration

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$#6 !#$#7 # #$#7 #$#6

8,2+*-91'/1*::'3'+('/0;.,11!&<=5

!%%><?>!@# !7%><?># %><?>@# @%><?>A# !A#><?>A#

035

SLIDE 17

All 5 Bands

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$#6 !#$#7 # #$#7 #$#6

8,2+*-91'/1*::'3'+('/0;.<11!&=>5

!%%?=<?!@# !7%?=<?# %?=<?@# @%?=<?A# !A#?=<?A#

025

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$#6 !#$#7 # #$#7 #$#6

8,2+*-91'/1*::'3'+('/0;.,11!&<=5

!%%><?>!@# !7%><?># %><?>@# @%><?>A# !A#><?>A#

035

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$#6 !#$#7 # #$#7 #$#6

8,2+*-91'/1*::'3'+('/0;.,11!&<=5

!%%><?>!@# !7%><?># %><?>@# @%><?>A# !A#><?>A#

0*5

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$" !#$#% # #$#% #$"

6,2+*-71'/1*88'3'+('/09.,11!&:;5

!%%<:=<!># !?%<:=<# %<:=<># >%<:=<@# !@#<:=<@#

075

!" !#$% # #$% "

&'()*+,-*.+/01'23''45

!#$#6 !#$#7 # #$#7 #$#6

8,2+*-91'/1*::'3'+('/0;.,11!&<=5

!%%><?>!@# !7%><?># %><?>@# @%>3,>A# !A#><?>A#

0B5

SLIDE 18

AB Offsets

Standard u g r i z Solar

0.0660

0.0202 0.0052 0.0207 0.0125 Solar Error 0.0068 0.0035 0.0027 0.0057 0.0070 BD+17°4708

0.0541

0.0158

0.0021

0.0099

0.0054

WD Average

0.0455

0.0173 0.0037 0.0168 0.0059 WD Error 0.0149 0.0014 0.0030 0.0035 0.0057 Error u g r i z PT Measurement 0.0026 0.0015 0.0013 0.0025 0.0018 Color Transform 0.0029 0.0009 0.0008 0.0009 0.0017 HST Measurement 0.0056 0.0031 0.0022 0.0050 0.0066 SDSS Filters ? ? ? ? ? HST Calibration ? ? ? ? ?

SLIDE 19

Conclusions

 Calibration is only as good as the weakest link (PT filters & flat field).  Atmospheric extinction measurements are needed more frequently than once per night.  Star flats are critical for precise flat-fields.  “Frequent” calibrations are necessary (filters & flat field? & atmospheric extinction?)  Redundancy is a key to understanding

 Many standard stars  Multiple overlapping “star flats”  Cross-calibration with other surveys

 A transparent calibration technique is important.

SLIDE 20

SDSS Photometric Calibration Revisited A look in the rear view mirror

John Marriner 18 April 2012

Cosmology with SN

 The desire to determine SN magnitudes from measurements

photometric calibration.  Comparison of different experiments requires an “absolute” calibration.  Photometry is typically difference photometry relative to field stars.  SN magnitudes vary by ~7%

 Errors are reduced by averaging many SN.  Random calibration errors are a minor concern.  Even small systematic biases can cause problems.

SDSS SN Calibration Strategy

 Report “native magnitudes”  Measure filter response (relative throughput)  Calibrate absolute response with standard stars whose

 Spectrum is well measured.  Photometric response of the SDSS telescope is measured.

 Measurements of the standard stars are reported as an “AB offset”  Consider only stripe 82

Details

 The Photometric Telescope (PT)

 Measures atmospheric extinction based on primary standards (the USNO standards).  Measures unknown field stars for use as secondary calibration standards.

 The 2.5 m Survey Telescope is calibrated using the secondary calibration standards measured by the PT.  The 2.5 m Telescope is normally operated in drift scan mode.

Filter Measurements

More Filters

Flat-field Stability

Rms Measurement Variation

PT Flat Field

What is the Ideal Star Flat Dither Pattern?

 Rely on dome flats for high spatial scales.  Dither by 1/2, 1/4, 1/8, …?

f x

( ) =

aneinkx

∑

f x + h

( )− f (x) =

an 1− einkh

( )einkx

∑

⇒ an = 1 1− einkh e−inkx f x + h

( )− f (x)

[ ]

∫

dx

PT Response Map

A Tale of 2 Catalogs

 Stripe 82 Coadd (Ivezic)

 Uses PT response for magnitude (r-band)  Uses stellar locus for colors  Multiple measurements, outliers & variables rejected

 DR8 (Ubercal)

 “Insensitive” to PT response  Color uncalibrated  Single epoch catalog

Coadd/DR8 Comparison

PT Observation Times

Comparison Coadd/DR8

fits as a function of declination

Variability of the Calibration

035

All 5 Bands

AB Offsets

Conclusions

 Many standard stars  Multiple overlapping “star flats”  Cross-calibration with other surveys

 A transparent calibration technique is important.

Stellar Color/Color Plots