ALPACA by chenmeixiu

VIEWS: 31 PAGES: 62

									   ALPACA
      A Superlative
Deep, Wide, Ground-based
 Optical Imaging Survey


       JEDI
  An Orbital Near-IR
Multiobject Spectrograph/
   Multiband Imager
   Advanced
   Liquid-mirror
   Probe for
   Asteroids,
   Cosmology and
   Astrophysics
                                       (alpaca)
is currently a collaboration between
Columbia University
Stony Brook University
University of British Columbia
American Museum of Natural History
Pontificia Universidad Catolica de Chile
and Cerro Tololo Inter-american Observatory
     ALPACA is a telescope and
     survey project consisting of
•   8-meter diameter, mercury primary mirror
                                        o
•   Baker-Paul 3-mirror design with ~3 field
•   Zenith-pointing, drift-scan telescope+CCD
•   ~ 1 Gigapixel mosaic, 5 simultaneous bands
                                            o
•   ~ 1000 square degree field, at Dec = -30.16
•   Nightly sampling all bands, reaching r ~ 25
•   “Real-time” and archival image processing
•   Reaching r ~ 28 in 3y survey (~ HDF depth)
Remarkable aspects of ALPACA include:
•   Diverse, superlative science return
•   Contains Galactic Center, South Galactic Pole area
•   Optimizing SNe Ia probe of dark energy dynamics
•   Extremely large AW (“etendue” or “grasp”)
•   Superlatively deep imaging, multiple bands
•   Good time sampling (time resolution Dt = 30s, 1d)
•   Simple, efficient operation; repetitive reductions
•   Simple design (few moving parts)
•   Largely established, off-the-shelf technology
•   Liquid-mirror technology now in advanced stages
•   Very cost-effective
Liquid-Mirror Telescopes




      Large Zenith Telescope (6-meter)
Time-delay integration (drift
          scan)
  • Image moves continuously across CCD due to Earth’s
    rotation
  • Charge being generated by photons is shifted
    electronically along the CCD columns at the same rate
  • Data are read continuously all night long
  • Integration time is the time taken for an image to cross
    the CCD




                                                    NASA 3m
                                                    LMT
      Liquid mirror surface waves
• Spiral waves due to turbulence in
  rotating boundary layer
• Effect of waves is to diffract light
  out of image core into a diffuse
  (I ~ r-3) halo
• Waves height is proportional to
  mercury thickness
• Concentric waves due to vibrations
• RMS wave amplitude ~ 46 nm for
  3m telescope
• Solutions for large mirrors:
   – Mylar isolation film
   – Baffles to induce air co-rotation?
           LZT performance
• 102 s exposure
  time
• ~ 1 .4 seeing
     //



• RAB ~ 23.1 (3)
  achieved so far
• RAB ~ 23.5
  expected
        Comments on Major Components
• ALPACA has no “mount” – the
  telescope is zenith-looking and is
  supported without a moving mount.
• Rotating 8-m mercury mirror can be
  constructed for under $1M, usually a
  ~$20M item.
• The telescope building is largely a
  simple tube or silo and need not
  move with the telescope. It requires
  a roll-off roof and louvers to control
  airflow for seeing control.
• Each detector surveys same strip of
  sky & photometric band every night,
  implying simple reduction stream
Proto-ALPACA imaging focal plane
•   0.86 deg diameter field (34 -> 240 deg2 strip)
•   6 CCDs, 6.7 arcmin square, 2048x2048 E2V
•   1 CCD for u,b,i,z; 2 CCDs for r
•   NASA NEOs: add two rows (18 CCDs total) for near-
    Earth asteroids (plus weak lensing, bulge
    microlensing, LSS, variable stars, etc.)



Full ALPACA imaging focal plane
    3 deg diameter field (830 deg2 strip)
    240 CCDs, 8 arcmin square, 2048x2048
    Fairchild
    deep strip, 8 columns with 6 rows of u, 4 b,
    and 2 each r, i, z
    wide strip, 8 more columns with 4 u, and 2
    each b, r, i, z
    NEO “ears”: at least 4 more columns of 2
    each of r, i
         ALPACA Survey Products (P. 1)
• Well-sampled, 5-band SN light curves (to r ~ 25 each night, r ~ 28
  each year) to discover and identify ~50000 SNe Ia and ~12000 SN
  Iab/II per year. SNe Ia mostly over 0.2 < z < 0.8 range, which is ideal
  for detailing the evolution and dynamics of dark energy
• Weak Lensing: 700 square degrees with multiband data good for
  photometric z’s
• Galaxy photometric redshift sample to r ~ 28; roughly 1 billion
  galaxies
• For galaxy clusters, should achieve same richness as SDSS cluster
  catalog (to z = 0.3) but to z = 1. Sample of ~30000 clusters
• Includes strong QSO lensing e.g., J12514-2914. Monitor 10-20
  examples.
• Map of Sculptor supercluster (z = 0.11). Novae, bright variables.
• Should find several orphan GRB afterglows per year.
         ALPACA Survey Products (cont.)
• Monitor 100,000s of AGNe to r ~ 26 for multiband variability.
• Large scale structure over 4 Gpc3 (comoving) to z = 1 and 9 Gpc3 to z
  = 1.5.
• Includes M83 (7 Mpc away, starburst); two Seyferts: NGC 2997 (17
  Mpc), NGC 1097 (17 Mpc). Follow cepheids, miras, novae, eclipsing
  variables.
• Passes through Galactic Nucleus; will find >5000 Bulge microlensing
  events per year; superlative extrasolar planet search resource.
• Many 1000s of variable stars: Galactic structure.
• Huge variety of stellar surveys.
• Discover ~50 Kuiper Belt objects per night.
• Trace near-Earth asteroids of 1 km diameter to Jupiter’s orbit,
  reconstruct orbits well within 1 AU and detect 50 m objects at 1 AU.
    (Proto-)ALPACA for Bulge Microlensing

                                             •   OGLE III has 1.3m telescope with
                                                 a 0.34 deg2 FOV, covers 90 deg2
                                                 total, spending ~60s/night per
                                                 pointing, finds about 500
                                                 microlensing events per year.
                                             •   Proto-ALPACA will spend ~160s
                                                 per star, but has collecting area 35
                                                 times greater -> will go 10x deeper
                                                 -> 30x(density of stars). Will cover
                                                 5 deg2 -> 1000 events per year
                                                 (2500/year w/ NEO).
                                             •   ALPACA will spend >500s/night
                                                 per pointing -> 3 mag deeper ->
                                                 50x(stellar density); ~30 deg2 field
                                                 -> >5000 events per year.
Microlensing follow-up groups (PLANET, FUN, MOA) want to pick the ~100 best
of these lightcurves in terms of early planet-like deviations in microlensing fit.
Luminosity Distance versus Dark Energy Density

  The distance modulus (m – M) is
cosmology dependent; distance at
given z depends on expansion
(de)acceleration and spatial
curvature. SN Ia standard candle
relation puts constraint on ~(Wde-
1.4Wm), whereas CMB anisotropy
first acoustic peak constrains Wtot,
which together currently constrain
Wde, Wm the level of a few percent.
Similar constraints are found by
comparing cosmic microwave
background constraints with Wm
from cluster masses. Gives
reasonable cosmic ages.
    SN Ia Peak Luminosity/Duration/Color Relations

•   L-relations : Dm15 (Phillips
    1993)*, MLCS (Reiss et al. 1996,
    98), stretch (Perlmutter et al.
    1997, 99), CMAGIC (Wang et al.
    2003), DC12 (Wang et al. 2005)
    relate duration (and color, shape)
    over light curve to brightness at
    maximum e.g., Dm15(B) = drop in
    brightness 15d post-maximum
    (0.5-1.5 mag).

•   Duration appears to be predicted
    primarily by mass of 56Ni. Are
    there other parameters?              of SNe Ia is from Riess et al. 1995, ApJ, 438, L17.



* MB =-19.52(+0.27) + 0.88(+0.17) [D m15(B) - 1.1] (Altavista 2003, PhD thesis)
        Parameters affecting SN Ia Luminosities
Event width (Dm15) can predict SN Ia luminosity to ~15% r.m.s., including color measures
(CMAGIC, DC12) reduce this to 7-10% r.m.s. Are the residual errors due to measurement
error? Intrinsic processes? Extrinsic? Fundamentally stochastic? Possible factors include
(some treated in publications, with disagreement on nature, size – even the sign – of effects):
    •   56Nimass
    •   Single or double-degenerate progenitor
    •   Metallicity
    •   Progenitor compositional structure e.g., C/O varying with radius
    •   Rotational velocity (rotational support influencing density structure)
    •   Magnetic fields
    •   Density structure depending on mass of progenitor before accretion
    •   Convection structure in deflagration front
    •   Viewing angle
    •   Ejection asymmetries
    •   Circumstellar interactions
    •   Varying extinction laws
    •   Weak lensing magnification variation
            SN Observations with Proto-ALPACA
•   Nightly photometry in 5 bands:
    u(310-410nm), b(415-550nm),
    r(565-745nm), i(750-1050nm),
    z(950-1050nm). ubri are spaced
    in log(), minimizing K-correction   Simulated Proto-ALPACA SN Ia lightcurves including realistic
    errors; i band for high redshift.    effects of weather & instrument (checked against Gemini ETC).
    Better exposure/sampling than
    LSST, DES, SNAP, Essence, …
•   Combined nightly sensitivity
    AB(r) ~ 25, provides S/N > 10 SN
    Ia detections in 3-5 bands for > 5
    epochs each, to z = 0.8.
•   Expect ~4000 SNe Ia & ~2500
    SNe Iab, II at this S/N level.
•   Our PUC collaborators (Minniti,
    Clocchiatti) have generous access
    to 8m-class scopes (devote ~10
    nights/year, or ~200 SN, host
    galaxy redshifts)
           Yield of Supernovae from ALPACA
                                                • In 1 night: reach u = 24.3, b =
                                                  25.1, r = 24.1 (sum to r = 25).
                                                  This implies 100 y-1 deg-2 of
                                                  SNe Ia, 20 y-1 deg-2 of SNe II,
                                                  hence (30000+6000) y-1 total.
                                                  (Note: t = 2d SN Ia is 3.8 mag
                                                  fainter than at maximum)

                                                • In one 14-night dark run: reach
                                                  u = 25.7, b = 26.5, r = 25.5
                                                  (sum to r = 26.5) => (50000+
                                                  60000) y-1 total.
                                                  (Assuming 1 arcsec seeing and
                                                  100% good weather. Limits
Number per year per deg 2 of SNe Ia (dashed),     are for 10  detections.)
II (dotted) & orphan GRB afterglows (solid –
2 models, Woods & Loeb 1998, ApJ, 508, 760)
    Depth of Multiband SN Ia Detection

                                                     ALPACA can detect
                                                   SNe Ia in a single night to
                                                   z ~ 1, and find colors in
                                                   three bands to z ~ 0.8. As
                                                   we shall see, the u band is
                                                   also crucial for SN type
                                                   identification for z < 0.4,
                                                   which ALPACA also
                                                   handles superbly.
10  thresholds for typical SNe Ia in all 4 band
after 1 night (dashed) & 14 nights (dash-dotted)
           ALPACA SN Ia Redshift Distribution
• Consider only light curves detected at
  >10 at more than 5 epochs over the
  event (sufficient for calculating Dm15
  and several higher moments of the
  flux distribution).
• We recover a huge quantity of SNe Ia
  with this and better sampling, except
  at lower redshift. This is aided,
  however, by the solid angle covered
  for z < 0.4 is doubled to the full width
  of the array (discussed later), hence
  the numbers per 0.01 in Dz is actually
  twice that shown to the right. We
  have ~10000 SNe Ia (after 3 y) in 8-
  10 bins. We will discuss in a few
   minutes what this will accomplish.
                                             “decent” means >5 points of >10 in at least 3 bands
           Photometric ID of SN Type

                                           • On the basis of color alone (7d
                                             post max) one can separate SNe
                                             Ia from all other types
                       Reddening, AV = 4
                                           • Slight degeneracy of z = 0.3 Ia
                                             with z = 0.1 Ibc is removed by
                                             different evolution of SNe types
                                             through color-color plot over
                                             event.
  Star = SNe Ia
  Circle = I bc
  Triangle = II P
  Square = II b
  Diamond = II N

Number labels = int (10 z)
                SN Color Evolution


                                     •   Red = SNe Ia
                                     •   Green = I bc
                                     •   Blue = II P
                                         Number labels = days after max.




Color evolution over SN peak easily breaks degeneracy between
z=1 SNe Ia and z=0.3 SNe I bc (and further isolates SNe Ia).
                          SN Ia Host Galaxies

  SNe Ia tend to be closely associated
with prominent host galaxy. (SNe
Iab+II sometimes associated with
disconnected star formation knots.)




             Sullivan et al. 2003
                                           Tonry et al. 2003
Accuracy and Reliability of Photometric Redshifts
    • Accuracy of photometric redshifts

        – Systematic uncertainty ∆z / (1 + z) < 6.5%

        – Photometric uncertainty

    • Reliability of photometric redshifts

        – No “outliers” out of ≈ 150 redshifts

        – “Contamination” < 1% (Lanzetta et al. 1998)
       Plan for Proto-ALPACA SN Ia Studies
•   SNe Ia can be distinguished photometrically to z > 0.8. Redshift of SN can
    be determined photometrically with Dz ~ 0.2. Longterm image accumulation
    from Proto-ALPACA can determine host galaxy photometric redshift with
    accuracy Dz ~ 0.1. Positional coincidence of host galaxy and SN as well as
    redshift consistency will establish SN redshift to accuracy Dz ~ 0.1.
•   We will have large sample (~2000/y) of SNe Ia of well-sampled, multiband
    light curves from Proto-ALPACA. We plan to subject these to principal
    component analysis to find subcategories of SNe Ia. The goal then is to
    understand the physical nature of these differences and their effects on
    standard candle relation, both of which demand spectra.
•   Minniti and Clocchiatti have access to Gemini-S, VLT and Magellan time,
    each with multiobject spectrographs (GMOS: 5.5 arcmin, ~300 slits max.;
    FORS: 6.8 arcmin, ~20 max.; IMACS: 15 arcmin, ~250 max, respectively).
    They can devote ~10 nights/y -> ~75 fields total, corresponding to ~2-6% of
    Proto-ALPACA field (hence about 80-250 SN Ia host galaxies).
•   Also, we will need time on smaller scope to establish set of secondary
    photometric standards in ALPACA strip. (Minniti and Clocchiatti will help
    acquire this SMARTS time at CTIO.)
           Are Spectroscopic Redshifts Practical?
•   Spectrograph module operating with imager
    might obtain >20 spectra at a time (every
    ~50s) per module, x10 modules => number
    of spectra ~8,000,000/y vs. 30,000/y of
    “decent” SNe Ia
•   ~50% of host galaxies require < 10
    integrations (for 5 detection, 1nm
    wavelength resolution)
•   90% of host galaxies from High-z SN Team
    (HZT) have I < 25. Typically I ~ 23.
    (Williams et al. 2003, AJ, 126, 2608)
•   Spectrograph module must have drift
    scanning slit (perforated tape – “player                        1 integration 10 integrations
    piano”) and low distortion (need ~0.1% vs.                              |          |

    ~0.2% in reality for CTIO RCSP, f/7.8 ->
    f/1.0 ). This instrument can be prototyped
                                                 Williams et al. 2003: magnitudes of HZT SN Ia hosts (0.43 < z < 1.06)
    using Proto-ALPACA.
Full ALPACA: Improving SN Ia Standard Candles
 The primary challenge is the control of
systematics. A method to improve this
might be splitting the sample of ~105 well-
sampled light curves into z bins (~0.1) small
enough that z error due to cosmological
uncertainty is insignificant, Dm< 0.1%, then
perform a principle component analysis on
~10 subsamples of ~104, and compare
different subsamples’ results to measure
evolution. Quality of the data might allow
us to explore 10-20 parameters. Finding the
covariance of luminosity in a single bin with
this parameter set should reduce the scatter
significantly, producing a detailed
luminosity model, or at least discarding
outriggers.
[Behavior of dark energy can be parameterized by its
Equation of State: pressure versus mass density w = p/ r.
w=0: “normal matter,” w= -1/3: cosmic strings,
-1/3 < w < -1: quintessence, w= -1: cosmological            see Lewis & Bridle 2002, Phys Rev D, 66, 103511
constant]
         Worked Example for 2000 SNe Ia

2000 SNe with r.m.s Dm = 0.2
mag, sampled in z via deep,
ground-based imaging. Get
maximal discrimination in dark
energy density f at redshifts 0.3 <
z < 0.8. This is true of most dark
energy models, in this case
quintessence (scalar field potential
with slow-roll) versus k-essence
(similar but with coupling to
kinetic energy term) as might help
explain why Wde ~ Wm

                                       Wang & Garnavich 2001, ApJ, 552 445
Luminosity Distance vs. Dark Energy Dynamics
 Statistical errors only: even
 assuming that we are unable to
 reduce scatter in inferred SN Ia
 luminosity below ~10% r.m.s., the
 number of SNe will allow us to
 achieve statistical errors in ~10
 redshift bins of Dm ~ 0.002. This is
 the same size proportional error one
 sees across a Dz = 0.1 bin if one
 assumes value of Wm incorrect by
 3%, roughly the uncertainty now.
 An error Dm = 0.002 is small
 compared to deviations between         Standard candle apparent brightness at moderate
                                        redshifts for different models of dark energy:
 predictions of different physical
                                        (baseline) Wde=0.7 cosmological constant – value of
 models for dark energy; 10r value      0.6 or 0.8 varies by about 0.13 in Dmag at z=1, (thick)
 of a few percent is sufficient to      pseudo Nambu-Goldstone boson, (thin) supergravity,
 differentiate most extant models.      (long dashed) pure exponential, (thick dotted) inverse
                                        tracker, (short dashed) periodic potential (Weller &
                                        Albrecht 2001, PhysRevLet, 86, 1939)
       Cosmological Measures of Dark Energy
Universal equation of state w=p/r describes expansion’s dynamics and therefore H(z).
What we actually observe are measures of H(z) and redshift integrals* over H(z), the
angle distance and luminosity distance:

Supernovae as standard candles:
                      luminosity distances dL(zi)
Baryon acoustic oscillations as standard ruler:
                      cosmic expansion rate H(zi)
                      angular diameter distance dA(zi)
Weak lensing cosmography:
                      ratios of dA(zi) / dA(zj)

*Comoving   distance              is related to expansion rate H(z):

and the observed distances (in flat Universe) dL = R0 r (1+z), dA = R0 r / (1+z)

Several independent methods will provide a powerful cross-check, and allow ALPACA to
place precise constraints on dark energy (+growth of structure via cluster counts+strong
lens delay timings+large-scale structure Alcock-Paczynski+cluster integrated Sachs-Wolfe…)
      Combining ALPACA Dark Energy Constraints
    The simplest dark energy investigation
    method sensitivities to estimate are SN
    Ia standard candles, weak lensing shear
    and baryon acoustic oscillations. To
    express dark energy dynamics, we use w
    = w0 + wa a = w0 + wa /(1+z), where wa
    describes the redshift change in w. A
    few points:
•   If SN Ia method systematics ~ 10%,
    baryon oscillations are more useful. If ~
    2%, SN are more useful, comparable to
    weak lensing constraints.
•   Current limits combining CMB
    anisotropies, LSS and SN Ia constrain w
    at the 10% level. ALPACA could              Dataset         error on: Wm     w0     wa
    improve this 5x. Limit on wa would be       SNe (2% syst.+WMAP)      0.03    0.15   1.0
    vital in distinguishing dark energy         SNe+BAO                  0.02    0.11   0.65
                                                WL                       0.02    0.20   0.57
    models.                                     SNe+BAO+WL               0.01    0.04   0.16
                                                SNe+BAO+WL+Planck        0.003   0.02   0.04
                                                Planck only              0.013   0.19   0.94
              Growth of Structure: Clusters
Clusters are the largest collapsed structure, so respond to cosmological influence
   on scales near the linear regime. This is significantly affected by dark energy
   domination. At Columbia, Wang, Haiman, Khoury & May (2004) investigate
   cluster counts dN/dz and power spectrum P(k) to constrain cosmological
   parameters, including w’ - here parameterized as wa, which is roughly
   equivalent: w = w0 + wa (1 - a). We have run these models for ALPACA.
Expected ALPACA accuracy used in conjunction with CMB anisotropies
   improves significantly on the best planned CMB-only constraints (Planck):
      1- error Planck only dN/dz dN/dz+P(k) dN/dz+P(k)+CMB
       D(Wde)          0.07        0.03      0.017           0.004
       D(Wmh2)        0.002       0.075      0.012          0.0004
       D(8 )          0.1        0.013       0.01           0.003
       D(w0)           0.6          0.4        0.2            0.05
       D(wa)           0.8         1.8        0.9              0.2
       D(Wbh2)       0.0006         ---      0.004           0.0002
       D(ns)          0.005         ---       0.05            0.002
                        CTIO Site
Cerro Tololo is in the
southern Atacama desert of
Chile, Latitude -30o.16. The
nights are 80-85% clear, and
seeing is typically at or below
1 arcsec. Site is at the “prow”
of the mountain summit,
looking into the prevailing
wind, hence best for seeing.
There are no obstructions for
many km, hence one can
                                  View from summit site looking NE into prevailing wind
expect little ground-layer
turbulence.
   Potential ALPACA Schedule
• early 2007: first light on telescope (with
  smaller CCD imager)
• early 2008: Gigapixel mosaic online
• early 2011: finish nominal imaging survey
• early 2012: start spectrographic survey?
Joint Efficient Dark-energy
   Investigation (JEDI):


    a candidate implementation
        of the NASA-DOE
Joint Dark Energy Mission (JDEM)
  JEDI Science Team:
Yun Wang (PI), University of Oklahoma
Arlin Crotts (co-PI), Columbia University
Peter Garnavich (co-PI), University of Notre Dame
William Priedhorsky (co-PI), LANL
co-Is:
Eddie Baron, University of Oklahoma
David Branch, University of Oklahoma
Edward Cheng, Conceptual Analytics
Ian Dell’Antonio, Brown University
Salman Habib, LANL
Katrin Heitmann, LANL
Alexander Kutyrev, NASA GSFC
John Mackenty, Space Telescope Institute
Harvey Moseley, NASA GSFC
Gordon Squires, Caltech
Max Tegmark, MIT
Craig Wheeler, UT Austin
Ned Wright, UCLA
    JEDI: exploiting 0.8-4 micron “sweet spot”
- lowest sky background region in ~0.3-100micron wavelengths
     - rest wavelengths in red/near-IR for redshifts 0 < z < 4




                            Background sky spectrum: Leinert 1998, A&AS, 127, 1
             What JEDI Hopes to Measure:

(1) The cosmic expansion rate H(z) as a free function
    in 0.2 redshift bins to better than 2% accuracy (using
    three independent methods: supernovae, baryon
    oscillations, weak lensing).

(2) For dark energy equation of state w(z)=w0+w’z (with
    a cutoff at z=2 such that w(z>2)= w0+2w’), measure
    w0 to better than 2%, and w’ to better than 5%
    accuracies.
                            JEDI Data:

(1) Supernovae: ≥ 7,000 type Ia supernovae with well-sampled
    light curves (every 5d) and good quality spectra.

(2) Primary baryon oscillation data: 10-100 million galaxy spectra
    (H emission line galaxies) over 10,000 square degrees, with
    0.5 ≤ z ≤ 2. The imaging over the same area serves as a
    weak lensing survey to H ~ 23.

(3) Primary weak lensing data: accurate measurements of galaxy
    shapes over 1000 square degrees to H ~ 25. The galaxy spectra
    over the same area form a baryon oscillation survey to z ~ 4.
                     Priors Assumed for
                   Dark Energy Constraints:

(1) For all three methods: assumed a flat universe, Wm is known to
    1% (or 3%) from CMB data from Planck (or WMAP).

(2) For the baryon oscillation method, assumed Wmh2 is known to
    3% fractional error, Wbh2 is known to 10%, the sound horizon at
    recombination is known to 1-2% from CMB data from Planck and
    WMAP.
   Measuring cosmic
expansion history H(z) to
  ≤ 5% accuracy using
 supernovae as standard
        candles
Uncorrelated measurement of H(z)
from simulated supernova data
(systematics included), with Wm
assumed to be known. Assuming a
Gaussian prior on Wm would
introduce some correlations in the
estimated H(z) in different z bins,
but would change the size of the
error bars little. JEDI SNe only.
h(z)=H(z)/100 km/s Mpc-1
            Wang & Tegmark (2005)
              Measuring cosmic expansion history
                 H(z) to ≤ 2% accuracy using
           baryon oscillation scale as a standard ruler
Assuming that the sound
horizon at recombination is
measured to 1.36% accuracy
(WMAP, Spergel et al.
2003). Not sensitive to
values of Wm and Wb
assumed. Data: JEDI
10,000 square degree
galaxy redshift survey.

       Blake (2005)

                      Note that the errors go opposite way of the SN case.
   Measuring baryon
oscillations as a standard
           ruler
     The Power of Three Independent
               Methods
Supernovae as standard candles:
           luminosity distances dL(zi)
Baryon acoustic oscillation as standard ruler:
           cosmic expansion rate H(zi)
           angular diameter distance dA(zi)
Weak lensing cosmography:
           ratios of dA(zi)/dA(zj)

The three independent methods will provide a powerful
cross-check, and allow JEDI to place precise constraints
on dark energy (+growth of structure via cluster counts+
strong lens delay timings+LSS Alcock-Paczynski+…)
   Measuring w0 to better than 2%, and w’ to better than 5%
                          accuracies
w(z)=w0+w’z (with a cutoff at z=2
such that w(z>2) = w0+2w’), and a
Gaussian prior on Wm with
(Wm)=0.01 (from Planck). SN
constraints (systematics included)
are derived using Fisher matrix &
including 300 SNe Ia with 0.1 < z
< 0.08 from Nearby SN Factory.
Weak lensing constraints are scaled
from Zhang, Hui, & Stebbins
(2003, Fisher matrix) for a survey
with median redshift 1, and
photometric redshift accuracy of
z=0.05. Baryon oscillation
constraints are derived using
Glazebrook & Blake’s (2005)
conservative Monte Carlo method.
            Necessity of Space Observations
Supernovae as standard candles:
1) Higher sensitivity: z > 1 SNe Ia tightens constraints on w’, tracker solutions.
2) H(z) measured as a continuous free function allows robust testing of DE models.
3) Rest J lightcurves for all SNe Ia (better standard candles – Krisciunas et al. 2004).
4) Smaller extinction corrections.

Baryon acoustic oscillation as standard ruler:
1) H(z) across redshift desert z = 1.5-2.5 significantly tightens constraints on w, w’.
2) H(z) measured as a continuous free function allows robust testing of DE models.

Weak lensing cosmography:
1) Stable and smaller point spread function a big plus over ground observations.
2) More galaxies and smaller shape noise by including red, quiescent populations

Result: Continuous H(z) to better than 2% in Dz~0.2 bins.
Comparison of JEDI and ground-based BO survey
     Required Supernova Data
• Supernova observations: 146 detection
  periods (5 days each) over 2 years
• ~ 5,000 type Ia supernovae over 12 deg2
• Photometric precision: ~ 2%
• Quality of spectra: S/N  10 with a
  dispersion of 20-40Å per pixel
           Required Galaxy Data

• Baryon acoustic oscillation measurements:
  – ~30-100 million galaxy redshifts (z2) over 10000 deg2 ;
    ~10 million galaxy redshifts (z4) over 1000 deg2
  – Redshift precision: σz /(1+z) =0.2% or better
• Weak lensing measurements (1000 deg2 ):
  – Photometric precision: ~ 1-2%
  – PSF stability: pointing stability < 0’’.02
          Observational Capability
• Observing modes:
  simultaneous imaging and multi-slit spectroscopy
• Imaging:
  FOV: ~ 0.15 deg2; wavelength: 0.8-4.2μm
  resolution: ~ 0".14 at 1μm
• Spectroscopy:
  FOV: ~ 0.35 deg2; wavelength: 0.8-3.2μm
  resolution: R ≥ 200-300
  multiplexing: ~ 50 supernova and 3000 galaxy spectra
             simultaneously
• Data rate:
  ≤ 30 Gbyte/day with compression and onboard processing
JEDI focal plane
       Strawman Mission Architecture

• Aperture of the telescope: ~1.5m
• Mission duration: 3 years
• Detectors: 32 2048x2048 HgCdTe detectors for imaging
  (0.8-4.2μm) and spectra (0.8-3.2 μm) plus eight 1024x1024
  HdCdTe detectors for longer-wavelength imaging
• Spectrograph slit mask: each spectrograph field consists
     of a 175x384 microshutter array, with a slit size 3"x6"
• Orbit: L2
• Telescope temperature: ~100 K (passive cooling achieved
     using sunshields)
JEDI Satellite
                        Microshutter Arrays:
AAS 205, [5.07] Microshutter Arrays for JWST NIRSpec., S. H. Moseley et al.

Each shutter consists
of a shutter blade
suspended on a
torsion beam (from
a support grid) that
allows for a rotation
of 90°. A motor opens
the shutters with a
specially formed
magnet as a remote
controlling tool.



2D programmable
slit mask
Strawman JEDI optical design
                           (Hawaii-1 detectors at reimaged
                           focus are suppressed for clarity)




                Dominic Benford (NASA GSFC)
                              (Hawaii-1 detectors at reimaged
                              focus are suppressed for clarity)
Dominic Benford (NASA GSFC)
Simulated JEDI spectrum of a SN Ia at z=1.7




                                     S/N ~ 10
             JEDI Mission Summary
• 1.5m-class, simultaneous imaging & multiple object spectroscopy,
  FOV: 0.5 square degree.
• Simultaneous slit spectra for ~50 supernovae and ~3000 galaxies
• Imaging in five bands (0.8-4.2 microns, spaced evenly in log )
• 4 spectrograph fields (0.8-3.2 microns), 175x384 microshutters each
• Total of 32 HAWAII-2 2048x2048 Rockwell HgCdTe detectors +
  eight HAWAII-1 1024x1024 detectors
• L2 orbit, sunshields for passive cooling
• JEDI would benefit from coordinated ground-based, optical surveys

• First two years: 7,000 type Ia supernovae (a deep galaxy survey over
  12 square degrees as a by-product)
• 3rd year: a 1000 square degree survey of galaxy redshifts (to z~4) and
  weak lensing measurements, imbedded in 10,000 square degree
  survey (redshifts measured to z~2)
• Launch scheduled for 2017
                                    Relaxing the prior on Wm


   As before, except
   gaussian prior of
   (Wm)=0.03 (from
   WMAP).

Fine point: also (Wm      h2)/Wm h2 =0.03
for baryon oscillations.

								
To top