Calibration and Cross-Calibration of the SNLS and SDSS Supernova Surveys

Introduction SNLS calibration Cross-calibration Conclusion

Calibration and cross-calibration of the SNLS and SDSS supernova surveys

Marc Betoule SNLS collaboration, JLA collaboration, CFHT

LPNHE

Fermilab, Apr. 2012

Betoule LPNHE Photometric calibration is the main source of uncertainty ∼ 1% reached in Regnault (2009) Need improvement for the ﬁnal release

Introduction SNLS calibration Cross-calibration Conclusion

Context and motivations Cosmological constraint from type Ia SNe −0.5 BAO+WMAP7 Most stringent constraints to date on the dark energy equation −1.0 of state parameter

68.3% +0.078 95.4% w = −1.021−0.079. w 99.7% a ∼ 2 improvement wrt

−1.5 CMB+BAO+H0

With all systematics SNLS3

−2.0 0.0 0.1 0.2 0.3 0.4 0.5 Ω m 68.3%, 95.4%, 99.7% conﬁdence contours FwCDM (Sullivan, 2011)

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Context and motivations Cosmological constraint from type Ia SNe −0.5 BAO+WMAP7 Most stringent constraints to date on the dark energy equation −1.0 of state parameter 95.4% +0.078 99.7% w = −1.021−0.079. w a ∼ 2 improvement wrt

−1.5 CMB+BAO+H0

Statistical only SNLS3 Photometric calibration

−2.0 is the main source of 0.0 0.1 0.2 0.3 0.4 0.5 uncertainty Ωm 68.3%, 95.4%, 99.7% conﬁdence contours FwCDM ∼ 1% reached in (Sullivan, 2011) Regnault (2009) Need improvement for the ﬁnal release Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

The type Ia SN Hubble diagram

24 HST

22 SDSS SNLS

20 + α (s − 1) − β C B 18 = m

corr 16 Low−z m

14 Conley et al. (2011) 0.6 0.4 0.2

corr 0.0 m −0.2 −0.4 −0.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 z The luminosity of SNIae as a function of z: a precise photometry experiment (in average) Relative calibration between photometric bands and surveys 2 large similar surveys

Betoule LPNHE Pushing this classic strategy Making the system redundant Shortening the calibration path Direct cross-calibration Cross-check of systematics

Introduction SNLS calibration Cross-calibration Conclusion

Overview

Classic (star-based) calibration HST white dwarves as primary SNLS SDSS calibration source

Megacam SDSS 2.5m Transfer to in-situ tertiary standards Landolt PT

KPNO 1.3m PT 2.5m HST

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Overview

Classic (star-based) calibration

Megacam HST white dwarves as primary SNLS SDSS calibration source

Megacam SDSS 2.5m Transfer to in-situ tertiary standards Landolt PT Pushing this classic strategy KPNO 1.3m PT 2.5m Making the system redundant HST Shortening the calibration path Direct cross-calibration Cross-check of systematics

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Outline

1 Introduction

2 SNLS calibration

3 Cross-calibration

4 Conclusion

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

SNLS

Photometric survey at CFHT 3.6 m telescope 40 nights/year over 5 years 4 deep ﬁelds of the CFHTLS

Megaprime/MegaCam 1 deg2, 36 × 2k × 4.6k CCDs → 300Mpixels

4 photometric bands gM rM iM zM

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Step I: Mapping the instrument response

Twilight observations give wrong answers Mostly reflections in the optical path (ghosts) Any optical effect affecting differently parallel and isotrope illumination (Small angle scattering, optical distortions ...) Varying filters

Number of Landolt star observations (i) 3 3 Critical in the transfer of average

2 2 calibration to SN CCD 1 Landolts and other standards observed at 1 speciﬁc positions 0

0 1 2 3 4 5 6 7 8 SN uniformly dispatched 0 CCD

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

Dithered observations of dense stellar fields 13 exposures Logarithmically increasing steps from 1.5’ to 1/2 deg 4-10 independent grid datasets /band → measure a correction δzp to the twilight flat-field 2003 35 -0.12 30 -0.12 Observation model 25

-0.03

20 -0.09 mADU(x, star) = m(x0, star) + δzp(x) 15 -0.06 m(x0) ∼ 100000 nuisance parameters δzp ∼ 100 parameters 10 -0.12 5

-0.12 0 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

-0.15 30

Observation model 25 -0.12

-0.03 mADU(x, star) = 20 m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.06

m(x0) ∼ 100000 nuisance parameters 10

-0.09 δzp ∼ 100 parameters 5 -0.12 0 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

mADU(x, star) = 20

m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.03 -0.06 m(x0) ∼ 100000 nuisance parameters 10

-0.15 δzp ∼ 100 parameters 5

-0.12 0 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

30 -0.15 -0.09

Observation model 25

mADU(x, star) = 20 -0.03

m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.06

-0.12

m(x0) ∼ 100000 nuisance parameters 10

-0.15 δzp ∼ 100 parameters 5

0 -0.12 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

30 -0.15

Observation model 25

-0.06 mADU(x, star) = 20

m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.03

m(x0) ∼ 100000 nuisance parameters 10 -0.09

δzp ∼ 100 parameters 5 -0.15

0 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

30 -0.15

Observation model 25

mADU(x, star) = 20 m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.03 -0.06 m(x0) ∼ 100000 nuisance parameters 10 -0.09 δzp ∼ 100 parameters 5 -0.15

0 -0.15 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

Observation model 25

-0.03 mADU(x, star) = 20

m(x0, star) + δzp(x) + δk(x)(g − i) 15

m(x0) ∼ 100000 nuisance parameters 10 -0.06 -0.09 δzp ∼ 100 parameters 5 -0.15

-0.15 0 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Mapping the response: the grid observations

-0.12 Observation model 25 -0.03 mADU(x, star) = 20

m(x0, star) + δzp(x) + δk(x)(g − i) 15 -0.06

m(x0) ∼ 100000 nuisance parameters 10

-0.12 δzp ∼ 100 parameters 5 -0.09 -0.12 0 -0.15 Betoule 0 5 10 15 20 25 30 35LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Intrinsic limitations

Expensive Finite number of stars External observations Limited resolution ∼ 1/2 night for all bands Limited precision Photometric time u g r

35 35 35 0.005

0.005 30 0.005 30 30

0.010 0.005 0.010

25 25 25 0.010

20 20 20 0.010 0.005 0.005 Solving for the zero point 0.010 15 15 15

10 10 10 0.015 0.005 0.005 and response map at same 0.005 0.010 5 5 5 0.005 0.005 0.010

0 0 0 time is poorly conditionned 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 i i2 z 35 35 350.015 0.010 0.010 30 30 30 Perfect photometric 0.005 0.005 0.010

25 25 25 0.005

0.010 conditions required 20 20 20

15 0.005 15 15

Radial gradient ∼ 1% 10 10 10 0.010 0.005 0.010 0.015 5 0.010 5 5

0 0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Results

Evaluating the impact of ﬂat-ﬁelding errors on the calibration Independent observations → errors average out After selection ∼ 6 useable independent photometric corrections per band All epochs agree with rms < .3% in all bands g r 0.010 0.010 0.005 0.005 0.000 0.000 0.005 0.005 − − 0.010 0.010 − −

2003-102004-032004-122005-102006-092007-032007-11 2003-10 2004-12 2005-10 2006-09 2007-03 2007-11 i z 0.010 0.010 0.005 0.005 0.000 0.000 0.005 0.005 − − 0.010 0.010 − −

2003-10 2004-12 2005-10 2006-09 2007-03 2003-10 2004-12 2005-10 2007-03 2007-12

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Step II: Anchoring to HST

SNLS HST Doubling the original path MegaCam Uncertainty on the landolt to MegaCam transformation MegaCam KPNO 1.3m Direct observations of the Landolt CALSPEC Solar Analogs with MegaCam.

A small observation program 3 standards observed along with D3 short observing group 4-5 nights / band at CFHT (5H) 5 mmag precision reached

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Agreement between SA/BD+17 calibration

(mmag) u g r i2 z dispersion between stars 27.5 5.7 5.7 0.5 4.3 dispersion between nights 12.3 4.7 6.1 5.4 11.7 oﬀset 22.9 12.5 -8.6 -1.8 -2.9 uncertainty 14.4 4.0 5.2 4.4 7.5

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Direct cross-calibration with the SDSS: a drastic shortcut

SNLS SDSS Megacam Megacam SDSS 2.5m The observation program Landolt PT Original program includes 2 ﬁelds within the S82 (SDSS36, KPNO 1.3m PT 2.5m SDSS326) observed along with HST D1 and D4.

60 D3 OB_D4 Dithered observations 50 OB_D2 40 OB_D3 Few epochs OB_D1 30 BD+284211 HZ43 20 GD153 Completed in January 2011 by GD71 dec (deg) 10 HZ4 D2 more numerous observations of SA 113 SA 107 SA95 SDSS36 0 SDSS326 SA 101 Feige110 D1

10 D1/SDSS36

D4 20 20 15 10 5 0 ra (hour)

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Measuring precise color transform

0.10 0.06 0.04

5 5 .

0 04 5 . .

0.05 . 2 2

2 0.02

u g 0.02 r

− 0.00 − 0.00 − 0.00 0 0 0 x x x

| | 0.02 |

r 0.02 u 0.05 g − − 0.04 − −0.06 0.04 0.10 − − − 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 g i g i g i 2.5 − 2.5 2.5 − 2.5 2.5 − 2.5 0.04 0.04 0.04 5 . 5 5 2 . .

0.02 0.02 2 0.02 2 2 i i z

− 0.00 − 0.00 − 0.00 0 0 0 x x x | | | i

0.02 0.02 z 0.02 2

− i − − 0.04 0.04 0.04 − − − 0.4 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 g i g i g i 2.5 − 2.5 2.5 − 2.5 2.5 − 2.5

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Checking passbands

0.10 0.06 0.04

5 5 .

0 04 5 . .

0.05 . 2 2

2 0.02

u g 0.02 r

− 0.00 − 0.00 − 0.00 0 0 0 x x x

| | 0.02 |

0.02 0.02 2 0.02 2 2 i i z

− 0.00 − 0.00 − 0.00 0 0 0 x x x | | | i

0.02 0.02 z 0.02 2

− i − − 0.04 0.04 0.04 − − − 0.4 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 0.4 0.6 0.8 1.0 1.2 1.4 g i g i g i 2.5 − 2.5 2.5 − 2.5 2.5 − 2.5

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Checking uniformity

0.04 0.04 5 5 . . 2 0.02 2 0.02 u g . . − 0 00 − 0 00 0 0

x 0.02 x 0.02

u − g − 0.04 0.04 − − 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 Ra. (deg.) Ra. (deg.)

0.04 0.04 5 5 . . 2 0.02 2 0.02 r i . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 i r − − 0.04 0.04 − − 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 Ra. (deg.) Ra. (deg.)

5 0.04 0.04 . 5 2 . 2 2 0.02 0.02 i z . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 z

2 − − i 0.04 0.04 − − 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 35.6 35.8 36.0 36.2 36.4 36.6 36.8 37.0 37.2 37.4 Ra. (deg.) Ra. (deg.)

Detect a (small) flat-fielding problem (PT related) in S82 catalog Confirm that MegaCam flat-fielding is good at the 5 permil level

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Checking calibration

Megacam SNLS SDSS Megacam SDSS 2.5m Closing the loop Landolt PT 3 independant path KPNO 1.3m PT 2.5m Do agree within announced error bars HST

g r snls5

sa band combined uncertainties sdss g 0.002 i2 z snls5 r 0.003 i2 0.003 sa z 0.006

sdss

0.02 0.00 0.02 0.02 0.00 0.02

Betoule LPNHE Introduction SNLS calibration Cross-calibration Conclusion

Conclusion

Redundant path to the HST ﬂux scale Robust assessment Precision better than 5 mmag achieved Uncertainty on the ﬂux reference itself now dominates

Pushing a classic calibration strategy, with intrinsic limitations Competition between accuracy and lost observation time Flat-ﬁelding is diﬃcult Accuracy becomes problematic around the 5 mmag level

Lessons learned Star-based photometric flat-fielding has intrinsic limitations. But it avoids the very tricky task of figuring out what the photometry do actually measure. Transfer of the average calibration between 2 fields is comparatively easier than the flat-fielding.

Betoule LPNHE Filter evolution

0.0020

0.0015 g i ) i r z No sizeable ﬁlter evolution − 0.0010 g ( / ) Color-terms between yearly catalogs

m 0.0005 − 0.0000 Evolution below .001 per unit of g-i year m ( 0.0005 − 1 order of magnitude smaller than the

0.0010 − 2004 2006 2008 CFHT/REOSC discrepency year

Betoule LPNHE SNLS color diagrams

2.5 1.5

2.0 1.0 1.5

1.0 0.5 i − z r − i 0.5 0.0 0.0

−0.5 −0.5 −0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 −0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 g−i g−i 0.03 0.03

0.02 0.02

0.01 0.01

0.00 0.00 i − z r − i −0.01 −0.01

−0.02 −0.02

−0.03 −0.03 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Betoule g−i g−i LPNHE Response uniformity: SNLS-SDSS

Flat-ﬁeld problem in S82 corrected Much better, but features remain Arguably attributatble to SDSS SDSS36

0.04 0.04 5 5 . . 2 0.02 2 0.02 u g . . − 0 00 − 0 00 0 0

x 0.02 x 0.02

u − g − 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 − − − − − − − − Dec. (deg.) Dec. (deg.)

0.04 0.04 5 5 . . 2 0.02 2 0.02 r i . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 i r − − 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 − − − − − − − − Dec. (deg.) Dec. (deg.)

5 0.04 0.04 . 5 2 . 2 2 0.02 0.02 i z . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 z

2 − − i 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 − − − − − − − − Dec. (deg.) Dec. (deg.)

Betoule LPNHE Response uniformity: SNLS-SDSS

SDSS326

0.04 0.04 5 5 . . 2 0.02 2 0.02 u g . . − 0 00 − 0 00 0 0

x 0.02 x 0.02

u − g − 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 − − − − − − − − Dec. (deg.) Dec. (deg.)

0.04 0.04 5 5 . . 2 0.02 2 0.02 r i . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 i r − − 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 − − − − − − − − Dec. (deg.) Dec. (deg.)

5 0.04 0.04 . 5 2 . 2 2 0.02 0.02 i z . . − 0 00 − 0 00 0 0

x 0.02 x 0.02 z

2 − − i 0.04 0.04 − − 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0 − − − − − − − − Dec. (deg.) Dec. (deg.)

Betoule LPNHE