Sky Mapper Brian Schmidt Mike Bessell Stefan Keller

Sky. Mapper Brian Schmidt, Mike Bessell, Stefan Keller, Patrick Tisserand, Gary Da Costa and Paul Francis Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 2

What is Sky. Mapper? • • • 1. 35 m telescope with a 5. 7 sq. degree field of view Fully Autonomous observing To conduct the Southern Sky Survey: • • Five year Multi-colour (6 filters) Multi-epoch (6 exposures, each filter) 2π steradians Limiting mag. g~23 Aiming for regular operations this year Summary of program: Keller et al. 2007 PASA 24, 1 Total Cost: Hardware US$11 M oftware: ~$1. 5 M Slide Dedicated Science/Operations: ~0. 6 M/yr Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 3

The Sky. Mapper Telescope Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 4

Telescope Optics 0. 75 m secondary 0. 6 m fused silica asphere 1. 3 m primary Telescope – Focal length & f/ ratio. 2 x 0. 45 m fused silica spherics Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics 16224. 75 mm f/4. 78 Fermilab April 2012 Slide 5

The Sky. Mapper imager and CCDs • • 6 Filter slots (< 10 second exchange time) Bonn shutter (2 ms accuracy) Exposure homogeneity <0. 3% at 0. 1 sec 32 E 2 V CCD 44 -82 devices: 2048 x 4096 15 micron pixel CCDs • Broadband coated • 40 micron (thick) devices • • • Reduced fringing, inc. red response, without bad blue 16384 x 16384 0. 5” pixels Using new Pan Starrs controllers (Onaka at If. A) Readout in ~20 seconds through 64 64 channels (300 kpix/s) Readnoise ~7 e Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 6

The Southern Sky Survey (>=75% of all available time) • 2π coverage: shallow survey. Photometric and astrometric Calibration of the Southern Sky in 6 -filters (3 -epochs per colour) • 2π coverage: ~4000 fields observed in six filters, six times per filter to a (13 -22 mag) • Cadence: hours, days, weeks, months, years • Photometry to 0. 02(0. 01 aspiration) mag globally (g>18. 5) • astrometry to 50 (15) mas • 36 images of each object over 5 years – ⇒ proper motions to ± 5(2) mas/yr. (i. e. σvtan=25 km/s at 2. 5(6) kpc) – ⇒ parallax ± 5(2) mas (i. e. 20 pc (50) σd=10%) • survey complete in 5 years using 75% of telescope time Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 7

moon Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 8

Sky. Mapper Optimised for Stellar Astrophysics but SN program driver for precision photometry • T, log(g), and Z are encoded in the spectrum of each star • Using filters we can isolate portions of the spectrum • In designing our survey we sought to optimise our ability to determine three important stellar parameters (T, log(g), Z) • so Sky. Mapper not only complements survey efforts in the northern hemisphere but enables us to tackle important astrophysics in an exciting new way. Bessell et al. 2011 PASP 123, 789 Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 9

Expected Main Survey limits u v g r i z 1 epoch 21. 5 21. 3 21. 9 21. 6 21. 0 20. 6 6 epochs 22. 9 22. 7 22. 9 22. 6 22. 0 21. 5 Sloan Digital Sky Survey comparison 22. 0 22. 2 21. 3 20. 5 AB mag. for signal-to-noise = 5 from 110 s exposures Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 10

Cadence • 1 st epoch : all filters consecutively (colour + short term variability in uv) • 3 first epochs in (g, r) in less than 7 days : for Astrometric and photometric short term variation (TNO + RR Lyrae/Cepheids+young SN): • (i, z) spread out to measure parallax over the year. – in total, 160, 000 sq-degrees observed per year to g~22 Use a “Score” algorithm for field priority - weighted by amount of time field is observable in survey’s remaining time. Deal with “Quick Data Quality Check” + Database Quality to validate each field meets criteria to be in main survey. Take care : distance of the Moon, Planets, Sky conditions, Satellites. . . Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 11

Calibration Plans • Conduct Shallow Survey • in mainly photometric conditions cover the southern sky with 3 x (30, 5, 5, 10, 15) exposures: 12 -17 th mag • primary standards: Seven new STIS spectrophotometric stds • 0. 3 sec exposure for ~9 mag standards • During shallow survey observe the highest two reference fields every 90 minutes • Anchor the deeper Main Survey to the Shallow Survey photometry and astrometry • Enables the Main Survey to proceed under non-photometric conditions. • Self Calibration via overlaps and the Main sequence (in low dust areas) Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 12

Calibration strategy Calibration procedures following Regnault et al. (2009) A&A 506, 999 for Mega. Cam on CFHT. • Bias subtraction, linearity correction • Flat fielding, scattered light and fringe removal 1) generate median twilight flatfields in all colors 2) generate superflats from median filtering only night-sky pixels 3) subtract scaled twilight flat from superflat to give fringe flats in I and z 4) scale and subtract fringe frame from observations 5) generate true flatfield by dithering standard fields in 8 offsets NS and 8 offsets EW. This to be done every ~6 months. • Use Sky. DICE to monitor mirror reflectivity and filter transmission ~daily. Clean mirror when advised. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 13

Sky. DICE System of direct illumination • • • Calibrated Photodiodes (NIST) 25 – 30 LEDs 3000 – 11000Å Precision of 0. 1% between 4000 and 9000Å Monitoring instrument ~0. 01% Permanently mounted on dome Nicolas Regnault - LPNHE Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 14

Photometric reference fields • During commissioning establish a series of photometric reference fields around the sky. • DDT application will be made for 15 STIS spectra to be measured to <1% precision 3000Å – 10200Å • Fields will be centered on 7 southern STIS stds • Six fields at 4 hour intervals between -18 S and -49 S. • One at -87 S to ensure pole region well calibrated. • These fields will provide calibrated photometry over the entire Sky. Mapper field of view. • + Six northern SDSS stds and 3 equatorial stds. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 15

Need for precise spectrophotometry Our goal is to tie photometry in both hemisphere to an absolute spectrophotometric system - This would serve the basis of Sky. Mapper's photometric system, and is the basis of the calibrations necessary to undertake supernova cosmology experiment to a high degree of precision. Currently, uncertainties in the overall calibration of SN to the fundamental standards is the largest single source of systematic uncertainty in the SN experiments. These data would be used for current SNLS/SDSSII work, as well as Sky. Mapper, and DECCAM, and eventually LSST calibration. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 16

Problems with NGSL spectra Currently there are two spectrophotometric catalogs from HST - NGSL, and CALSPEC. We had hoped to use the NGSL objects as our fundamental standards - but it appears that due to slit effects, there remain uncertainties (~ 2%) in the relative flux as a function of lambda. Sally Heap will discuss the NGSL stars at this meeting. The CALSPEC stars, which used a wide slit and were well centered, do not suffer from this problem, but have other drawbacks that make the current ensemble difficult to use to achieve our goals Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 17

CALSPEC spectra drawbacks • Very few moderate colored stars which can be tied directly to normal stars in the survey. The very blue stars typically require extrapolation from the bulk of the objects we wish to calibrate. • There are no appropriate southern stars which we can use to tie down our 1/2 of the sky. • Very few stars that can currently be tied to an external photometric system (Hipparchos now but GAIA is coming). This provides an overall normalisation and homogenisation to the spectra - instead of simply relying on the fidelity of HST. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 18

Choosing new secondary standards • It would be good to extend the HST calibrated stars to cover a wider color range of objects to ensure that bandpasses of our systems are correctly realised. • It is also important that the spectra can be well modeled so that accurate fluxes can be synthesized across a wide wavelength range. • It is generally accepted (eg. Bohlin) that hot DA white dwarfs are the best and most reliable to model, but the smooth spectrum and neutral colors of metalpoor FG-stars (subdwarfs), provide additional benefits and they also seem to be well modeled. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 19

Southern STIS Spectrophotometry Name Hip RA 2000 Dec 2000 V/B-V Hp/err Te/logg [Fe/H] HD 9051 6894 01 28 46. 4 -24 20 25 8. 93 0. 82 9. 079 0. 002 4842 1. 97 -1. 81 NGSL HD 31128 22632 04 52 09. 0 -27 03 50 9. 13 0. 497 9. 243 0. 002 5825 4. 30 -1. 50 MILES HD 74000 42592 08 40 50. 5 -16 20 38 9. 67 0. 431 9. 762 0. 002 6166 4. 19 -2. 02 MILES HD 111980 62882 12 53 15. 1 -18 31 20 8. 38 0. 55 8. 463 0. 002 5600 3. 7 -1. 20 MILES HD 160617 86694 17 42 49. 3 -40 19 12 8. 72 0. 456 8. 824 0. 002 5920 3. 60 -1. 96 HD 200654 104191 21 06 34. 5 -49 57 48 9. 09 0. 582 9. 214 0. 002 5160 2. 55 -2. 82 HD 185975 100968 20 28 18. 7 -87 28 20 8. 11 0. 69 8. 239 0. 002 5870 4. 4 0. 00 circumpolar Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 20

Northern STIS Spectrophotometry Name Hip RA 2000 Dec 2000 V/B-V Hp/err Te/logg [Fe/H] BD+21 0607 19797 04 14 35. 2 +22 21 06 BD+54 1216 40778 08 19 22. 6 +54 05 15 BD+29 2091 52771 10 47 23. 0 +28 24 03 BD+26 2606 72461 14 49 02. 3 +25 42 12 BD+02 3375 86443 17 39 45. 8 +02 24 59 BD+17 4708 9558 22 11 31. 0 +18 05 33 G 24 -03 98989 20 05 44. 4 +04 02 54 Feige 110 115195 23 19 58. 4 -05 09 56 9. 23 0. 362 9. 73 0. 484 10. 26 0. 506 9. 73 0. 437 9. 94 0. 458 9. 46 0. 459 10. 46 0. 48 11. 83 -0. 30 9. 327 0. 002 9. 812 0. 002 10. 373 0. 002 9. 819 0. 002 10. 051 0. 003 9. 569 0. 002 10. 579 0. 002 11. 753 0. 007 Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics 6300 4. 0 5900 4. 2 5700 4. 50 5875 4. 10 6140 4. 31 6000 4. 00 5870 4. 0 DAwk -1. 67 SDSS -1. 70 SDSS -1. 98 SDSS -2. 48 SDSS -2. 15 SDSS -1. 65 SDSS CALSPEC -1. 8 equatorial CALSPEC equatorial Fermilab April 2012 Slide 21

Caveats g~12 is the bright limit for 5 sec focussed exposures. This necessitates ~0. 2 sec exposures for 8. 5 mag or out-of-focus imaging. Scintillation at 0. 2 sec is < 0. 01 mag Shutter imprecision (0. 2 ms) and homogeneity (0. 3% at 0. 1 sec) should not be an issue. Issues with defocussing – curve of growth problems? Needs exploration. Useful to establish fainter spectrophotometric standards with ground-based telescopes (Tonry). Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 22

The Shallow Survey • Determine photometric status of each night by performing aperture photometry of all point sources and compare with 2 MASS color-magnitude relations. • Observe two different standard reference fields every hour to provide zero-points for each chip and illumination correction. • Calculate extinction and derive zeropoints. • Monitor seeing, photometric zp, sky-background. • Observe different passbands consecutively. • Use SExtractor with aperture twice fwhm, then correct to infinite aperture, to generate standard magnitudes. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 23

Shallow Survey Filter Exp time SN=7 SN=50 u 30 (dark time) 17 15 v 30 (dark time) 18. 5 16. 5 g 5 18 16 r 5 18 16 i 10 17. 2 15. 2 z 15 17. 2 15. 2 Data to be acquired in photometric and non -photometric time. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 24

Main survey photometry • Use stars from Shallow Survey to put instrumental photometry on the standard instrumental system. • Combine images and create object catalog. At high galactic latitude do for all bands, at low use I band. • Extract transients by comparing individual frame with combined frame, after masking out cataloged objects, and locating objects with appropriate PSF. • Determine PSF shape for each object in each image in a filter. Use as basis for galaxy/star separation. • Determine galaxy shape parameter (PA and ellipse isophot) and measure magnitudes (Source Extractor III). Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 25

Acknowledgements Before Magnier & Cuillandre (2004, PASP 116, 449) and Regnault et al. (2009 A&A, 506, 999) widefield photometry was not considered by many people to be capable of delivering better than 3% photometry. But we are all here this week discussing how to achieve 1% photometry. A paradigm shift. We thank all those who have worked to make this happen. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 26

Flats before and after The hexagonal structures that were visible in the middle and the corners have gone. The new flat is actually flat. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 27

Pinhole test - edge of field These features have completely disappeared thanks to the velvet tape stuck on the carbon fiber collar in front of the camera and the 10 mm wide ring painted on the outside of the 3 rd CLA lens. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 28

Pinhole test - on axis Using the velvet tape on the inside of the CLA and secondary baffles, we have removed the bright lines due to the baffle junctions. The black disk placed in the middle of the secondary has been useful to eliminate the reflection of the secondary itself on the CLA lens. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 29

Sample v filter sky flat Waffle pattern from laser annealing. Removed by flat fielding. Seen in u and v frames. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 30

Telescope oscillation Satellite trails with cryocooler on and off. Seeing 2” when on. 1. 3” when off. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 31

Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 32

Sky. Mapper Filter Set Ex-atmosphere Bessell et al 2011 PASP 123, 789 Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 33

Observational Scheduler The scheduler must respond to photometric/seeing conditions during a night. Maximize the science output 1 st epoch : all filters consecutively u, v, g, r, u, v, i, z, u, v (color + short term variability in uv. 3 first epochs in (g, r) in less than 7 days : for Astrometric and photometric short term variation (TNO + RRlyrae/Cepheids): (i, z) spread out parallax over the year. Use a “Score” algorithm for field priority - weighted by amount of time field is observable in survey’s remaining time. Deal with “Quick Data Quality Check” + ANUSF Data. Base Quality to validate each field meets criteria to be in main survey. Take care : distance of the Moon + Planets, reduce Airmass. . Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 34

Some Science Themes • What is the distribution of large Solar-System objects beyond Neptune? • What is the history of the youngest stars in the Solar neighbourhood? • How far does the dark matter halo of our galaxy extend and what is its shape? • Gravity and metallicity for on order of 100 million stars the assembly and chemical enrichment history of the bulge, thin/thick disk and halo? • Extremely metal poor stars. • Undiscovered members of the local group – Sculptor, Cen too… • accurate photometric calibration of galaxy redshift surveys: 2 d. F/6 d. F. • bright z>6 QSOs probes of the ionization history of the Universe. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 35

Structure of outer Halo of galaxy with giants, BHBs and RRs Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 36

Blue Horizontal Branch Stars A-type MS and BS BHBs Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 37

Structure of outer Halo of galaxy with giants, BHBs and RRs The SDSS technique Sagittarius stream Use a set of colour and spectroscopic indices to isolate BHBs Extend to 60 kpc The Sky. Mapper technique Photometric BHB selection to 130 kpc with 5% contamination + RRs obtained from time series …and lines of sight to HVCs From Sirko et al. 2004 Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 38

Extremely Metal-poor Stars in the Halo • Goal: find the first stars to have formed in the Universe: tell us about the assembly and chemical enrichment of the Galaxy • v-g is dependent on the level of metal line blanketing in the blue continuum ✓ not perturbed dramatically by C-enhancement, chromospheric emission as affects objective-prism surveys F Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics G Fermilab April 2012 Slide 39

Sky. Mapper High-z QSOs • Detection is simple: i-band dropouts. • Contaminants: L+T dwarfs • SDSS tells us we can expect of order 50 objects z<20. 5 in the SSSS – Detection limited by our i sensitivity (not z!) – Proper motions can help remove most L+T dwarfs – as will J, H, K photometry – initial follow-up with Wi. FES+2. 3 m and IRIS 2+AAT Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 40

Search for extreme LSB MW satellites Jerjen, Willman, Mateo. . . • Chemical abundance measurements of individual stars to learn about SN yields, Pop-III stars? , and how smallest galaxies form. Sky. Mapper will cover an uncharted 20000 sq-degrees 0. 7 mag deeper than SDSS. UMa, 100 kpc, -6. 8? Discovered by Willman et al. 2005 from SDSS data Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 42

Sky. Mapper And a whole lot more • All Sky Tully-Fisher peculiar velocity survey when combined with ASKAP (250000 galaxies) • input catalog for HERMES (+Teff and g) • QSOs (Bright QSOs for dz/dt, dα/dz, TPE) and input to Lyα-BAO experiment • large sample of hyper-velocity stars in halo based on gravity+metallicity • Transiting planet discrimination (HAT-South based at RSAA) • Discovery of rare stars (e. g. R Cor Bor, AM CVn) Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 43

up to 25% of time for other things • Hα of Southern Sky • Mg 5100 filter for K Giants (maybe not needed) • GRB, Gravity Wave, Radio Transient TOOs • SN Search override • High Cadence Variability surveys Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 44

Southern Sky Poor Seeing Time Survey Time Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 45

Data Release Deliverables to the Outside User: -Data (epoch, RA, DEC, mags, galaxy shape info, …) to be available through a web-served interface which provides catalogs over a user defined area -Images to be available through a web-served interface which provides images over a user defined area (1 degree max, cut of a 2 -degree TAN projection across the sky), or individual frames. Data release will occur after extensive data validation: -Shallow survey data after closure in RA and trial application to concurrently obtained main-survey data -FDR Main Survey 3 epochs all filters -SDR Main Survey 6 epochs all filters Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 46

Non-ANU parties can gain access to data during data characterisation period: • Request is forwarded to Project Scientist for Sky. Mapper. Project Scientist will ensure Project does not overlap/interfere with existing projects. • Establish a memorandum of understanding (MOU) between the external investigators and the School regarding the usage of data for specific purposes and authorship policy (inclusion of Sky. Mapper Development Group). External groups will be expected to visit Stromlo (we will try to support) as part of the MOU. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 47

Problems encountered Problems solved • Scattered light in telescope - paint, masks, baffles • Air-conditioners, dry air - more • Fail-safe remote observing - new single board computer • Gradient in bias - software Remaining problems • M 2 resonant vibration Currently 48 sec between exposure, aim is 30 sec. Read-noise is 7 e on most CCDs. 100 hz cryo-cooler induced resonance in M 2 causes image wobble. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 48

Current status In operation for last 2 weeks. Still some issues with TAROS. Still working on a few CCDs. Dampers ordered. Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 49

Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 50

Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 51

• Nicolas Regnault & Julien Guy Research School of Astronomy and Astrophysics Research School of Astronomy & Astrophysics Fermilab April 2012 Slide 52