Dark Energy Survey Survey Strategy

1
Dark Energy Survey Survey Strategy

• James Annis
• Experimental Astrophysics Group
• Fermilab

2
Survey Strategy WBS 1.3
Dictionary

• Survey planning
• Science requirements and technical specifications
• Survey observing strategy
• Calibration
• Photometric calibration
• Astrometric calibration
• Photometric redshift calibration
• Survey data simulations
• Catalog level simulations
• Image level simulations
• Mock data challenge
• Mock analysis challenge
• Survey operations
• Defined as after commissioning

3
Survey Planning (WBS 1.3.1)

• Science requirements and technical specifications
• Survey observing strategy

4
Science Requirements (WBS 1.3.1)
Science goals drive the science requirements,
science requirements flow down to technical
requirements.

• 5000 deg2 in the South Galactic cap
• overlap SPT
• redshift surveys
• Photometric redshifts
• To z1.0 and ½ L with dz lt 0.05
• all z lt 1 SPT clusters
• g,r,i,z
• PSF sufficient for weak lensing, FWHM lt 0.9
• 40 deg2 supernova area
• Time domain, 1 hour every 2 days

• 10-sigma limiting magnitude
• g24.6 500 sec
• r24.1 500 sec
• i24.0 900 sec
• z23.6 1600 sec
• Calibration Accuarcy
• 10 single image
• 1 final coadd
• 2 spectrum convolution

DES Technical Flowdown Document v1 Dec
2003 v3 May 2004
5
Science Requirements Footprint

• Extinction Map, Centered SGP
• Midnight on Nov 1 at CTIO
• Black lines
• Equatorial coords
• Green lines
• Footprint
• 5000 sq-degrees total
• In order of priority
• 4000 sq-deg SPT
• 250 sq-deg SDSS stripe 82
• Photo-z areas
• 700 sq-deg overhead
• Alama, VLT, APEX, ACT

6
Survey Strategy Design Goals

• Efficiently attain required depth
• Minimize photometric calibration errors
• Scientifically interesting data at natural points
  in the survey
• After year 1
• internal goal, we reserve the right to be
  affected by weather
• After year 3
• After end of survey

7
Survey Strategy

• Decisions
• Area as function of time
• Area more important than depth
• Entire survey area the first year
• Work during all phases of moon
• g,r during dark time
• i,z during bright time
• Moonlight has little effect

• Instrumental Factors
• Telescope slew time 35 sec
• Instrument read time lt 20 sec
• Implies 2 filters per position to minimize
  overhead
• Data Quality
• Airmass limit lt 1.5
• Seeing limit lt 1.1

8
Survey Strategies

• Baseline Strategy
• 2 filters per night
• 2 tilings of survey area per year, per filter
• 100 second exposures
• Calibration to 2 the first year
• 5 tilings is break point
• Calibration is 1
• g,r filters done
• i,z filters increase exposure time (years 4-5)
• Standard stars during night
• Alternate Strategy
• 1 tiling of survey are per year, per filter
• Divide total integration by 5 to find exposure
  time
• Calibration to lt5 first year
• Standard stars during poor seeing nights

We are exploring alternatives.
9
Survey Strategy Characteristic Magnitudes

• Note nonlinear behavior of limiting magnitude as
  function of time
• Knee at transition from S/N time to S/N
  ?time
• Reach characteristic magnitude in O(100) seconds

10
The Main Survey Strategy Table
N8 gal/sq-arcmin
N12 gal/sq-arcmin
N16 gal/sq-arcmin
N20 gal/sq-arcmin
Catalog completeness
N28 gal/sq-arcmin
Weak lensing depth maximal
11
Survey Calibration (WBS 1.3.2)

• Photometric Calibration
• System response functions
• Relative calibration
• Absolute calibration
• Astrometric Calibration
• Photometric Redshift Calibration

12
Large Area Survey Photometry

• Unique properties of large surveys
• Single stable instrument
• Huge homogeneous photometric data set
• System defined by 108 magnitudes of the survey
• Placing onto a standard system is unimportant
• Though transformations to standards is very
  useful.

• The aim of calibration in large surveys
• The magnitudes may be calculated by convolving a
  spectrum with good spectrophotometry with the
  system bandpasses, and
• The magnitudes vary only by 2.5log10(f1/f2),
  independent of position
• f1/f2 are the ratio of the photon fluxes
• The magnitudes have a well-defined absolute
  zeropoint.

13
The magnitudes may be calculated by convolving a
spectrum with good spectrophotometry with the
system bandpasses

• Full system response measurement
• 1nm resolution
• Through all components of the instrument
• Focal plane array
• Filters
• Corrector
• Primary
• If without this, input beam has correct f/number
• Aimed at final system testing prior to ship

14
The magnitudes vary only by 2.5log10(), where
are the ratio of the photon fluxes, independent
of position

• Relative photometry
• Use overlapping images of stars to place all
  images on same relative system.
• 1000s of stars per overlap
• Precision very high, limited by systematics
• Overlapping tilings
• Allow reduction of systematics

1 tiling 3 tilings 3 more
tilings
15
Relative Photometry Simulation
 

• INSTRUMENT MODEL
• A multiplicative flat field gradient of amplitude
  3 from east to west
• A multiplicative flat field gradient of amplitude
  3 from east to west
• An additive scattered light pattern with a
  amplitude from the optical axis, 3 at the edge
  of the camera
• An additive 3 rms scattered light per CCD
• Solution
• Simutaneous least squares solution to the
  underlying relative photometry given the
  observations

scaling bar is 0.20 mags to 0.20 mags
16
The magnitudes have a well-defined absolute
zeropoint

• Absolute photometry
• Transfer standard star magnitudes
• Need fainter standards (Tucker and Smith)
• Questions
• How many transfers to tilings?
• When are standards taken?
• What spatial pattern?

Constant airmass tracks, std at end Standards on
poor seeing night, on sparse hex grid
17
Absolute Photometry Simulation
 

• INSTRUMENT MODEL
• As before
• NIGHT MODEL
• g band, 10 extinction
• Full photometry model of atmosphere
• Plus a linear gradient in k over night
• Induces RA gradients
• Constant airmass tracks, 4 standards night
• Assuming each standard calibrates track to 5
• Solution
• Simutaneous least squares solution to the
  underlying absolute photometry given the
  observations

scaling bar is 0.20 mags to 0.20 mags
18
Quality Assurance on Photometry

• Use principal axes of stellar locus to check
  colors
• Photometry calibrates mags, not colors
• sigma_o is 0.003 mags
• Check on CCD scales, and on smaller scales,
  looking for systematics

19
Survey Simulations (WBS 1.3.3)

• Survey observation simulations
• Catalog level simulations
• Image level simulations

Huan Lin will cover in detail in the next talk
20
Mock Data Reduction Challenge (WBS 1.3.4)

• Task reduce 1 year of imaging data
• Input 1 Survey year of Image level simulations
• From Survey Simulation (WBS 1.3.3)
• Transfer to NCSA
• Run through Survey data pipelines
• Transfer to Fermilab
• Deliver to Mock Analysis Challenge (WBS 1.3.5)
• Results
• Integration testing of pipelines
• Data throughput testing
• Plan
• Staged 1 night, 1 month, 1 year
• Aim at 1 year in starting in mid 2007

21
Mock Data Analysis Challenge (WBS 1.3.5)

• Task analyze 1 year of data
• Input 1 Survey year of pipeline processed
  simulations
• From Mock Data Reduction Challenge (WBS 1.3.4)
• Make available to Science Teams
• Coordinate science team efforts to
• Integrate science codes to catalogs
• Analyze 4 key project science goals
• Results
• The season after the first year of observing will
  be spent on arguing the science, not developing
  science codes.
• Plan
• Staged start with catalog level simulations,
  then to MDRC data
• Aim at full test starting in 2008

22
Survey Strategy WBS 1.3
Fermilab FTEs
Fermilab Computing

• Year 1 (FY2005)
• Scientist 1 FTE
• Year 2
• Scientist 1 FTE
• CP 0.5 FTE
• Year 3
• Scientist 1 FTE
• CP 1.5 FTE
• Year 4
• Scientist 1 FTE
• CP 1.0 FTE

• Year 1 (FY 2005)
• Year 2
• 10 TB
• Year 3
• 50 compute nodes
• 50 TB
• dCache front end
• 100 TB Enstore
• Year 4
• 100 compute nodes
• 100 TB
• dCache front end
• 300 TB Enstore

23
Summary

• Survey Planning
• Science requirements and technical specifications
• Survey observing strategy
• Calibration
• Photometric calibration
• Survey data simulations
• Catalog level simulations
• Image level simulations
• Mock data challenge
• Mock analysis challenge