Large Synoptic Survey Telescope

1
Large SynopticSurvey Telescope
David Wittman, UC Davis
www.lsst.org
2
Vital Statistics

• 8.4 m telescope
• 3-mirror design provides good imaging over 10
  deg2 FOV
• ugrizy survey over 20,000 deg2
• rapid readout and slew/settle time for fast
  cadence
• multi-PB archive for massively parallel
  astrophysics
• 14M for DD next four years
• first light 2012-2013

3
(No Transcript)
4
Historical Context
Note angular resolution also tends to increase
with time.
5
Why?

• Two compelling problems require a survey both
  deep and wide
• Dark energy must measure geometry and/or growth
  of structure out to
• redshift 1-2
• Potentially hazardous asteroids (PHAs) survey
  must go faint, fast, often
• The same survey can address these plus many other
  science goals
• best-ever Galactic map, solar neighborhood
  inventory, the transient
• universe

6
Compelling?
7
Compelling?
dark energy theory is off by 10120 gt there's
something VERY IMPORTANT we don't understand
8
Optical Configuration
Monolithic M1 and M3!
9
(No Transcript)
10
Camera Design

• 4-element corrector
• 65 cm focal plane
• 10 deg2 FOV
• 3 Gpix
• 200 CCDs
• 32 channels/CCD
• 2s read time

11
Survey Speed Deep, Wide, Fast
12
Superb Images Over Entire Field
(define PSF)
13
LSST Final Three Sites
San Pedro Martir
Cerro Pachon
Las Campanas
Site Evaluation
14
Project Baseline Schedule
First Light
Submit MREFC
Telescope
Camera
Data Mngt
15
Key ProjectsFully Exercise Capabilities of
System

• dark energy and dark matter
• weak lensing
• baryon acoustic oscillations
• supernovae 250k Ia/yr!
• solar system inventory
• transient optical sky
• mapping the Milky Way

Resulting dataset suitable for tomorrow's
questions
Common database, common cadence!
16
LSST Dedicated Survey

• 51-band Survey grizyu
• Sky area covered 20,000 deg2 0.2
  arcsec / pixel
• Each 10 sq.deg FOV revisited 200 times/band
• Time resolution 20 sec
• Limiting magnitude 26.5 AB magnitude _at_10?
• 24 AB
  mag in 10 seconds
• Photometry precision 0.01 mag requirement,
  0.005 mag goal
• Galaxy density 50 galaxies/sq.arcmin
• 3 billion galaxies with color redshifts
• Time domain seconds years
• Massively parallel astrophysics

17
LSST Catalog

• 3 billion galaxies
• each with 500 attributes measured
• each with 200 time samples
• 106 moving objects
• 108 variable objects (50,000 supernovae per
  month, variable stars, gamma-ray bursts, ??)

18
Deep digital survey of sky each week 18 terabyes
imaging data per night Unique science
opportunities Celestial cinematography
19
10-6 of the LSST area!
20
Don't Forget Dark Matter!
21
(No Transcript)
22
Lensing independent of dynamics, baryon content,
star formation history
Strong lensing on axis, high resolution, densest
regions of universe
Weak lensing off axis, low resolution, all
regions of universe, statistical
23
(No Transcript)
24
First DLS Mass Map
2o
Do we trust this? Next slide...
25
(No Transcript)
26
(No Transcript)
27
Photometric Redshifts from Multiband Imaging
28
Photometric redshifts Example

• Big Issues
• training sets and bias
• priors

29
Photo-z scatter and expected N(z)
N(z) from degraded HDF
30
Lensing in 3-D with photometric redshifts
Mass map using high-z sources
Mass map using low-z sources
Result from lensing alone cluster at z0.3.
Spectroscopic value 0.27
31
20,000 deg2 Shear Power Spectra
z1100
z3.2
z3.2
z3.0
z0.6
z0.6
z0.2
z0.2
Blanco shear systematics floor
LSST shear systematics floor
32
sheared image
? 4GM/bc2
b
DS
DLS
?
shear
DLS
?????????
4GM/bc2
DS
33
PSF Systematics gtgt Lensing Signal
Before After
34
PSF Systematics Short Exposures
SuPrimeCam dense star field
Wittman, ApJL 632, 5 (2005)
35
Short Exposures PSF Residuals
Wittman, ApJL 632, 5 (2005)
Don't require long exposure limit!
36
Controlling Systematic Errors
37
Controlling Systematic Errors

• Quality, quality, quality
• in situ metrology
• wavefront sensors throughout the focal plane
• observe in r and i when seeing lt 0.7
• prompt (real-time) reductions and integrated
  quality assessment pipeline
• boresighted cirrus monitor
• in situ response calibration
• compact design minimizes flexure

38
Controlling Systematic Errors

• Chop, chop, chop
• each galaxy will have hundreds of exposures in
  each filter
• range of camera rotations
• range of pupil rotations
• range of field centers
• 10-year time frame
• compact design fast slew/settle times

• Quality, quality, quality
• in situ metrology
• wavefront sensors throughout the focal plane
• observe in r and i when seeing lt 0.7
• prompt (real-time) reductions and integrated
  quality assessment pipeline
• boresighted cirrus monitor
• in situ response calibration
• compact design minimizes flexure

39
Wavefront Sensors Maintain Image Quality
20 Curvature Sensors
3.5 FOV ? 64 cm ?
X
X
X
X
X
X
X
X
X
X
X
X
Shack Hartman Sensor
Raft
X
X
X
X
X
X
X
X
X
X
X
X
40
End-to-End Optical Simulation
Peterson (SLAC), Jernigan (SSL) ray tracing
through atmospheric turbulence, perturbed optics,
detector (incl. diffusion) Actual
WIYN perturbations
41
Detector Model (Andy Rasmussen, SLAC)
Refraction for light entering the Si
surface Photon interaction (wavelength and
temperature dependent) Lateral diffusion due to
finite electric field Validation Process
ongoing to maintain consistency with
sensor group
42
Piece of UDF through 12 layers of
atmosphere, Complete LSST optics, and detector
with complete wavelength dependent
effects, fairly high fidelity in some areas, and
does this in 1/2 hour
And that's just for 15 seconds of LSST time!
43
Operational Simulator

• Includes
• real weather/seeing data
• slew/settle time
• dome rotation
• cable wrapping
• operational DB very important
• for assessing systematics!

Will have multi-TB of simulated LSST data in a
few years, ready for statistical analysis!
(K. Cook et al)
44
Massively Parallel Astrophysics

• Dark matter/dark energy via weak lensing
• Dark matter/dark energy via supernovae
• Galactic Structure encompassing local group
• Dense astrometry over 20000 sq.deg rare moving
  objects
• Gamma Ray Bursts and transients to high redshift
• Gravitational micro-lensing
• Strong galaxy cluster lensing physics of dark
  matter
• Multi-image lensed SN time delays separate test
  of cosmology
• Variable stars/galaxies black hole accretion
• QSO time delays vs z independent test of dark
  energy
• Optical bursters to 25 mag the unknown
• 5-band 27 mag photometric survey unprecedented
  volume
• Solar System Probes Earth-crossing asteroids,
  Comets, TNOs

45
Focus task transients
46
The Transient Universe Image differencing
technique
47
Transients from the Deep Lens Survey
Supernova
Unknown R 23.1
Asteroid (slow-moving)
48
Asteroid (fast-moving)
49
Near-real-time transient data http//dls.physics
.ucdavis.edu/transients.html
50
Rare Events
51
(No Transcript)
52
Source Modeling Using Full Dataset(as opposed to
coadd)

• PSF estimation errors random, not systematic
• Weights better-seeing images properly
• Computationally feasible?

53
Source Detection in Time Domain(example courtesy
Phil Marshall, SLAC)

• time delay of strongly lensed variable source
  yields Hubble constant
• most strongly lensed sources are weakly varying,
  so measuring time
• delays is difficult
• supernovae have strong, well-defined time
  variation, but it would be
• extremely rare to see one strongly lensed
• with LSST, one can find extremely rare objects
  (1000 lensed SNe)

54
LSST Statistical Challenges

• All of classical survey problems (e.g. source
  detection) plus time domain
• Time domain effects on survey quality e.g. one
  object splits into two on
• a good night and three on a very good night.
• Image reconstruction/stacking/compression
• Classification, especially of transients. Need
  VERY small false alarm rate!
• For moving transients, matching of detections
  across nights and weeks
• Fast algorithms for n2 (and worse) problems (e.g.
  tree codes)
• Robust estimation and outlier identification
• Data mining finding rare objects/events finding
  unexpected correlations
• Data quality assessment
• Operations/scheduling
• Inference from LSST other surveys (related to
  Virtual Observatory)
• how to drink safely from a firehose

55
(No Transcript)
56
Data Management

• 1/3 of the project, equal
• to telescope and camera
• subsystems.
• summary catalog will be 10 TB
• full dataset 100 PB
• 10-20 TB precursor archive will
• grow to include simulations
• build on existing data center

57
(No Transcript)
58
Color-color diagram reveals extreme stars
?
59
Brown dwarf (L type) discovery
V
B
R
Z'
In 7 subfields (2.5 square degrees) we have
found 1 L dwarf candidate (V26.77, R23.94,
z20.30 and which no detection in B).