ATLAS

1. ATLAS Asteroid Terrestrial-impact Last Alert System

2. Chelyabinsk: 2013-Feb-15 • 19m diameter, • 0.5Mton TNT • 30km altitude

3. 2012 DA 14: 2013-Feb-15 • 30m diameter • (2Mton TNT) • (10km altitude) • Closest approach 28,000km = 4.4 RE • Orbital period changed from 368 to 317 days

4. How do you design (optimize) a survey? • Science goals • Find examples of rare, interesting, or useful objects • Characterize source populations, look for trends, evolution, sub-populations, etc. • Find all objects of a certain type • Search boundary conditions • Is population of interest homogeneous and isotropic? • Timescale for variable objects and cadence • What signal-to-noise (SNR) is required? • Confusion limit or other systematics limits (e.g. exposure overhead, read noise, saturation, software, etc.) • When objects are randomly distributed and detection noise is photon-limited, survey merit is just survey area per unit time at a given magnitude and SNR. • Coverage rate, SNR, and limiting magnitude can be traded off for one another by changing exposure time or deploying more units.

5. Survey Speed • M = “survey speed” • A = aperture area • 0 = solid angle per exposure •  = PSF area = “effective noise” footprint •  = throughput efficiency •  = duty cycle •  = sky brightness • S1 = SNR per exposure •  = total solid angle covered in tcad • m = detection magnitude • tcad= cadence for covering  Tonry 2011 PASP 123, 58

6. Survey Speed • M = “survey speed” • A = aperture area • 0 = solid angle per exposure •  = PSF area = “effective noise” footprint •  = throughput efficiency •  = duty cycle •  = sky brightness • S1 = SNR per exposure •  = total solid angle covered in tcad • m = detection magnitude • tcad= cadence for covering  Tonry 2011 PASP 123, 58

7. Design Survey Speed • M = “survey speed” • A = aperture area • 0 = solid angle per exposure •  = PSF area = “effective noise” footprint •  = throughput efficiency •  = duty cycle •  = sky brightness • S1 = SNR per exposure •  = total solid angle covered in tcad • m = detection magnitude • tcad= cadence for covering  Tonry 2011 PASP 123, 58

8. Design Operation Survey Speed • M = “survey speed” • A = aperture area • 0 = solid angle per exposure •  = PSF area = “effective noise” footprint •  = throughput efficiency •  = duty cycle •  = sky brightness • S1 = SNR per exposure •  = total solid angle covered in tcad • m = detection magnitude • tcad= cadence for covering  Tonry 2011 PASP 123, 58

9. Survey Speed • M = “survey speed” • A = aperture area • 0 = solid angle per exposure •  = PSF area = “effective noise” footprint •  = throughput efficiency •  = duty cycle •  = sky brightness • S1 = SNR per exposure •  = total solid angle covered in tcad • m = detection magnitude • tcad= cadence for covering  Tonry 2011 PASP 123, 58

10. Survey Speed Better ASSASN Point size proportional to solid angle surveyed within tcad Tonry 2011, PASP 123, 58

11. Optimizing Survey Speed

12. ATLAS • Project funded by NASA to find dangerous asteroids • Two 0.5m telescopes covering 30 deg2 (5.4°x5.4°) producing ~2.5” images. • One telescope on Maui (Haleakala), one telescope on Hawaii (Mauna Loa) • Optimized for maximum survey speed (per unit cost) and fast cadence. • Broad filters (cyan~g+r and orange~r+i) provide improved sensitivity for normal search but retain color information. • 16 filters overall, including • c, o, u, v, g, r, i, z, B, V, R, I, H, [OIII] • Expect to observe ~60,000 deg2 at m~20, i.e. entire visible sky 3 times per night.

13. Oahu Maui Hawaii

14. ATLAS Components • Telescope: 0.5m f/2.0 Wright-Schmidt (DFM) • Camera: 110 Mpix Acam, 5.4x5.4 deg (IFA) • Mount: German equatorial (APM) • Enclosure: 16.5’ dome (Ash) • Software: IFA, IFA, IFA.

15. Image Detect Diff Detect Catalog Software Systems • Maintenance • Hardware status • Software version control and updates • Operations • Automatic scheduling • Robotic operations on N sites (execute schedule, monitor weather and hardware status, etc.) • Automatic reduction to classification and alerts • Image pipeline • Acquire image and metadata • Reduce image to fully clean and calibrated, detect sources • Difference image against static wallpaper, detect sources • Classify sources and process appropriately: variable, transient, moving • Auxiliary and database • Reference catalog of stars • All-sky wallpaper for each camera and filter • Database of all detections • Database of all aggregate objects

16. ATLAS Enclosures Haleakala Mauna Loa Model of ATLAS in dome

17. Camera Status • Acam1 is complete • Cryostat, controller, cabling, etc. complete • Science grade STA-1600 tested, very nice CCD. • 110 Mpix, 10e- read noise, <7 sec readout, good QE, good cosmetics • Acam2 is waiting for Telescope #2 • Cryostat, controller, cabling, etc. complete • Waiting for the perfect science grade CCD… • Expect delivery in Oct

18. DFM Telescope Status • Telescope #1 in operation • Installation on Haleakala in Jun 2015 was successful, all functions (shutter, 8 filter changer, focus) work properly • Schmidt corrector needs improvement, PSF~6” now. • Schmidt corrector replacement Mar 2016? • Telescope #2 nearly done • Only Schmidt corrector remains, installation on Mauna Loa end of 2015 5.4°

19. APM Mount status • APM mounts installed on both sites. • Haleakala • DFM telescope attached • Polar aligned, mount model • Slew, settle nearly always faster than CCD readout • Mauna Loa • “Pathfinder” telescope (Takahashi Eps-180 and IFA 4k camera) • Testbed for improved APM software • Waiting for Telescope #2

20. m-mlim Software Status • Maintenance • All components have enough maturity to safely observe, lots to do… • Operations • All components in place, lots to do… • We’ve observed many nights trusting the robot without any accidents, but we don’t yet give it permission to open • Pipeline • Satisfactory end-to-end maturity, lots to do… • Bottom line is that we have a 50-50 chance of detecting something at 5 sigma, according to photon statistics, with a false alarm rate of ~2:1. • Auxiliary • We have all-sky reference and all-sky wallpaper. We can generate new wallpaper in a single night. • Databases are maturing, lots to do…

21. Auxiliary Camera #1 • Each telescope also has a Canon 5DII with 135mm f/2 lens taking simultaneous exposures. • 300,000 deg2/nt at mlim~14, msat~6 • Fully reduced and quantified (WCS good to ~1” and photometry good to ~0.03 mag) • RGB color retained but not currently reduced

22. Auxiliary Camera #2 • Each site has a Canon 5DIII with 10mm f/4 lens taking continous exposures • All sky every 40 sec at mlim~7, msat~-1 • Fully reduced and quantified (WCS good to ~30” and photometry good to <0.1 mag) • RGB color retained but not currently reduced

23. ATLAS Met Page – 24/7 Sky Images

24. Images… Observation

25. …Wallpaper… Observation All-sky Wallpaper

26. …Differences GEO satellites Asteroid Unmasked glints Observation All-sky Wallpaper Difference

27. 1/50,000 of 1 night

28. 1/50,000 of 1 night

29. M107 globular cluster RA 248.132 Dec -13.054

30. M107 globular cluster RA 248.132 Dec -13.054

31. Wallpaper • New observations incorporated with • “wpadd 02a57243o0123c” • Complaints! Excuses! • 2014 observing ended when the (40 year old) dome on MLO broke, since fully refurbished. • El Nino has not been our friend this year… Pathfinder 2014 Acam 2015 (orange) Violet  White: noise RMS  = 20—25 mag/sec2

32. ATLAS Science • Solar system • NEO discovery, impactor warning • Unbiased phase space density of PHAs near Earth • ~103 NEO/yr, ~104 asteroid rotation, etc. • Stars • ~hr resolution light curve of all stars to m~19 • Planets: habitable planets, low mass star planets • search WD for planetary transits, Earth < 0.01 • Gravitational lensing: dark matter, H0 • near field microlensing; AGN time delays • Supernovae: SN properties and large scale flows • ~104 SNIa/yr, many CCSN, ~10 bright ultraSN/yr • AGN and black holes: AGN properties and evolution • ~hr resolution of ~105 AGN to m<19 • Gravity waves: because it’s there • E&M counterparts to events creating gravity waves

33. Tunguska Chelyabinsk ATLAS Impactor Discovery Probability • ATLAS discovery probability depends on size and survey duration • Small (<10m) only seen on last day or two. • Medium (10-140m) seen for days to weeks before impact. • Large (>140m) are often seen on orbits prior to impact. Multi-orbit Multi-night Single night

34. ATLAS Warning Time • Warning time depends on size • 140m: ~3 week • 50m: ~1 week • Missed impactors come from the Sun and the south pole.

35. 20 Lunar distances 10 D=30m Moon’s orbit 0 SUN ATLAS and NEOs • ATLAS will have about 9 NEOs of D~30m in view at any given time. • Most 30m asteroids that come within 1 lunar distance could be detected by ATLAS 1 day D>30m

36. Image Subtraction

37. Streak Detection Radon transform

38. Streak Detection

39. Asteroid Science • Light curves • At m=18 (25), ATLAS will provide ~500 photometric observations per year on ~104 asteroids. These can be processed to infer the shape and rotation. • Asteroid-asteroid collisions • There is 1 collision per day at size 10m, 1 per year at 100m. The dust from the collision will expand an unknown amount before reaching unity optical depth • At opposition in the asteroid belt, ATLAS can see a dust cloud of size 1000m, and possibly may capture such an event.

40. Variable Stars • Luminous blue variables MV = –9 to –13: • volume reaches Virgo. • Novae at MV = –7 to –9: • visible from MW, M31, M81, M101. • Miras, RCorBor, Cepheids, RR Lyrae, FU Orionis: • all variables brighter than MV < +3 are all visible throughout MW to 25kpc. ATLAS has excellent time sampling for their light curves. • Cataclysmic variables: • closer than 1kpc are visible at MV < +9, with enough time sampling for light curves and continuous monitoring for outbursts.

41. Stars • ATLAS will create light curves for ~1 billion stars with ~500 epochs each year. • “DC” light curve from images • Baseline flux is present • Confusion with nearby sources • “AC” light curve from differences • Baseline flux from wallpaper subtracted • No confusion • How do we combine these? • Science projects include • Eclipses from companions or (large) planets • Rotation signatures? • General level of (non-periodic? quasi-periodic?) variability

42. Gravitational Lensing • Microlensing • Lots of generic microlensing events • Near-field events: lensing star close enough to see lens and source separate (after a while) • m=18 expect ~40 mas/yr proper motion • Total expected is ~23 events per year at m<18, 58 events per year at m<19 • ATLAS will see ~30/yr total, and ~10/yr at high SNR and time coverage • Strong lensing • Expect ~40 AGN lensed at x3 or more, and ~7 AGN lensed at x10 or more. • These are likely to have multiple images and accessible time delays

43. Supernovae • Type Ia supernovae • ATLAS monitors half the sky at all times • ~5000/yr at z<0.1 (V<19) • ~300/yr at V<17 and ~30/yr at V<15 • Science includes • Peculiar velocities to probe dark matter • Nearby anchor for cosmology: dlum(z) • Rates • Systematics, environments, clusters, etc. • Core collapse supernovae • ATLAS will detect comparable numbers to SNIa • Low metallicity hosts ([O/H]+12 < 8.2) where CCSN may become GRBs, ATLAS will find 10/yr at 25 and z<0.04

44. 1 year Large Scale Flows • Use SNIa as standardizable candles: • ~10% distance accuracy per SNIa • ~1 SNIa each year per (30Mpc)3. • Therefore measure the distance of a shell of thickness 1,000 km/s with an accuracy of 100 km/s per year independent of distance, limited when systematics dominate (~30,000 km/s?) • Independent measure of • Monopole (Hubble bubble) • Multipole (large scale flows) • Need ~20 spectra/night…

45. AGN • There are 250,000 AGN brighter than V=19.6. ATLAS will monitor 200,000 of them • 10 day cadence 25 • 1 day cadence 10 • ATLAS will continuously monitor 40 million galaxies at V=20 • Outbursts comparable to galaxy luminosity will be seen in substantial numbers • ~20 BH stellar accretion events (MV ~ -18) per year at 0.1mag photometric accuracy

46. Static Sky • ATLAS observes most of the sky ~500 times per year • m~23.4 from one year stacked sensitivity at 5 • ~3 mag fainter than POSS • ~1 mag fainter than SDSS • similar to PS1 3pi 3 year (but only 2 bandpasses) • But note highly confused for static sources, although excellent for differencing • Unconfused for variable sources: ATLAS has a sliding sensitivity into variability structure function: • m~20.6 at 1 day, • m~21.7 at 10 day, • m~22.9 at 100 day.

47. m=18.5 m=19.4 m=17.0 4” sep ATLAS-POSS-SDSS Comparison ATLAS ATLAS

48. ATLAS: one year observation

49. SDSS

50. Papers to appear in 2016 • “Tons and Tons of Moving Object and Transient Detections” • “The Density of Asteroids Near Earth” • “Return from the Local Group to the Cosmic Microwave Background” • “10,000km Close Approach of 2016 LD44” • “Planetary Occultation of a White Dwarf” • “Five FU Orionis Outbursts” • “Five Microlensing Events” • “Five Ultraluminous Supernovae” • “Limits on Optical Counterparts to Gravity Wave Events”