Subject: Astrometry+Halo section of DRM

From: Dave Monet

Submitted: Sat, 19 Jul 2003 07:33:53 -0700

Message number: 154 (previous: 153, next: 155 up: Index)

Sorry for yesterday's confusion.  I circulated a crude draft of the
astrometry section of the DRM for comments, and Dennis responded to
everybody.  I have incorporated his section, deleted mine, and generally
tried to tidy things up.  As per Michael's desire to get the DRM ball
rolling, here is the current version for comments, etc.

My own opinion is that it is too long.  The DRM is not supposed to
demonstrate the scrivener's brilliance (although this is inevitable),
but rather to hit the high points with enough detail to convince the
vaguely qualified referee that LSST makes all other washday products
seem pale by comparison.  I think that Michael's suggestion of 2-3
pages per topic is pretty good, and the appended is way over this
even without the photometry and/or whatever other sections Michael
suggested for inclusion (astro.princeton.edu is asleep this morning
so I cannot review msg.131).

Fire away.  I have been volunteered to be the target, so please make
sure that I get copies of the corrections, etc.  Michael will have
a non-trivial task of merging all of these suggestions together,
and my hope is that editorial and stylistic changes can be delayed
until that time.

-Dave Monet is dgm@nofs.navy.mil

=========================================================================

Selected Topics in Stellar Astronomy

   Rather than present a shopping list of many interesting projects that
   LSST can do, we choose to present a discussion of three topics in which
   LSST will make fundamental advances.  They are

     a) structure and composition of the Galactic halo,
     b) astrometry, and
     c) <whatever else Michael thinks belongs here>.

   All of these are enabled by LSST's unique combination of sky coverage,
   repeated observations, and astrometric and photometric accuracy.

A.  Structure and Composition of the Galactic Halo

  A.1 Tracing the Smooth Luminous Halo

  From measurements of relatively local stars the current understanding 
  appears to be that the galaxy has a stellar halo of
  steeply falling density (rho proportional to r^{-3} Ivezic et al.) and
  a possible cutoff at r ~ 50 kpc. Using SDSS data, the halo
  and its inhomogeneities, are traced with stars that are brighter than
  M ~ 0 at several tens of kpc (m_r ~ 20 at the suggested
  cutoff radius). LSST will reach m_r ~ 26,
  which is fainter than the main sequence turnoff at the suggested
  cutoff radius (M ~ 6). The turnoff stars
  then become an excellent tracer of the halo (other types of stars
  that are that blue (B- V ~ 0.4) and that faint will be vastly
  outnumbered (factor?) by the main sequence stars and should be
  available perhaps out to perhaps 100 kpc. Proper motions can be used
  to further remove contaminants (foreground white dwarfs will be 
  ~100 times closer). The tracing of the halo
  can be done in a highly statistical manner by matching models with
  specific properties to the number counts and proper motion distributions.
  One should keep in mind that unlike the current status, where because 
  we are relatively data starved we need to be fairly confident of each star
  (particularly in the use of giants in the
  halo), in the LSST era much more can be done confidently in a 
  statistical manner. Even with SDSS, the structures that are being
  found are the ones we can point to the overheads (this will change with
  time once the cream has been skimmed and the data deluge recedes).
  
  Goals:
   
  1) Confirm/refute presence of cutoff at 50 kpc (trace the 3-D shape
     of the cutoff). This is simple for LSST, but see issue about colors
     and the use of colors to discriminate giants/dwarfs (below).
   
  2) Identify any further cutoffs (shell structure? multiple shells as
     seen in secondary infall calculations?). Tests of smooth infall
     vs. hierarchical.
   
  3) Measure radial behavior of stellar halo from 50 kpc to 200+ kpc.
   
  4) Measure velocity ellipsoid vs. r to constrain infall models and halo
     orbit families (important for formation/evolution modeling (deposition
     of angular momentum) and for mass profile determinations).This would
     be done with local halo populations for which the proper motions
     would be measurable (the 10-20 km/sec proper motions envisioned for the
     final survey would be ideal).

  A.2 Tracing the Unseen Halo

  LSST can play a role in the study of Galactic dark matter in at least two 
  ways: 1) it will provide dynamical information to map the inner halo via
  proper motions, 2) it will identify RR Lyrae stars at large distances to
  greatly increase the number of dynamical tracers in the outer halo, 3) 
  it will provide the most comprehensive survey of the luminous Galactic
  components. 

    A.2.1

    One of the most compelling arguments for a large dark matter halo comes
    from the fastest moving star (one assumes that is is bound and therefore
    derives a mass). It is a simple argument. By actually measuring the full
    3-D velocities of stars in the local neighborhood, we would have not just
    a single fast moving star but a sample of fast moving stars from which 
    orbits could be constructed. A self-consistent mass model would not only
    need to explain the velocity distribution function but would need to 
    tie that into the density of objects at larger radii (i.e. if we find that
    many local stars have apogalacticons of 100 kpc, there better be a
     corresponding population of stars at 100 kpc). Tracing a population in
    manner is much more constraining on a model than simply using a satellite
    galaxy (or two) at large radii. 

    A.2.2 

    Dynamical tracers at large radii are important because they provide
    independent measures of the total enclosed mass. Again, the consistency
    required by the study of these objects with the local velocity
    distribution function is well beyond anything available to date. Examining
    such studies as that by Kochanek (1996) or Wilkinson and Evans (1999),
    where the dynamics of various populations at different radii are
    combined, the tightening of the constraints by including a range of
    objects at different radial scales is evident. The current state of
    the art includes ~ 30 objects, ongoing work reaches ~100 objects over
    Schmidt plates (Clewley et al. 2002). LSST will dwarf these studies.

    A.2.3

    The interpretation of the microlensing results depends on several factors.
    The LSST can provide critical information on the radial distribution of 
    stellar lensing populations and perhaps some information on the tangential
    velocities of lensing sources. The ambiguities in the interpretation that  
    that can be introduced deviating from a standard halo can be seen in 
    the study by Geza, Evans, & Gates (1998). 

  Goals:

  1) Precise radial profile for stellar halo components.

  2) Precise measurement of the structure within the halo (what
     are the likely over/under densities along the LMC and SMC lines-of-sight?).

  3) Precise measurement of the tangential velocities (perhaps available
     only for the nearer lenses). At 1/2 the distance to the LMC (where
     lensing is favored), the proper motions will be good to 25-50 km/sec and
     so a lensing populations should be separable from the LMC stars, and a
     dispersion may be measurable if it is comparable to they of the
     Galactic halo (~ 150 km/sec).


  C. Tracing the Lumpy Halo

  To best trace the streamers we need both distances and velocities to halo 
  stars.  LSST provides the best avenues for obtaining each of these over
  large areas of the sky.  First, RR Lyrae stars will be identified throughout
  the halo and provide distances for any overdensity of stars identified as
  a potential streamer.  MSTO stars can also be used (see above).
  Second, proper motion measurements will provide some information on the
  kinematic coherence of any overdensity. The First Light observations are
  not useful for this program, but the first repeat and then the completed
  mission probe the range of expected halo velocities first to about 50 kpc
  and eventually to 200 kpc and beyond. Ironically, although radial
  velocities are generally much easier to obtain than tangential velocities,
  the all-sky nature of LSST will make proper motions easier to obtain
  than radial velocities over the same area of sky for the large number
  of stars involved. 

  With the tremendous number of stars involved, we do not need to identify
  individual streamers but rather quantify the ``lumpiness" of the 
  velocity distribution to constrain the number of streamers contributing
  to the halo population. 

  Goals:
  
  1) Completely trace streamers our to several hundred kpc. (do 
     streamers come primarily from one type of orbit family, for example,
     satellites on radial orbits). 

  2) Measure coherence of streamers over full length of arcs (a measurement
     of the roughness of the Galactic potential). 
  
  3) Measure width of streamers (a combination of Galactic potential
     roughness and the initial internal velocity dispersion of satellite 
     - important for understanding initial population). The internal
     velocity dispersion of the original object that was disrupted is expected
      to be several tens of km/sec for the most massive objects.  It is
     therefore important to obtain proper motions that are as precise as
     possible and certainly better than  several tens of km/sec to be able
     to say anything regarding the coldness of the stream and potentially
     about the original system.

  4) Measure proper motions along stream to solve for ``guiding center"
     orbit (determines potential depth and shape).

B. Astrometry

  Although not explicitly discussed by the Decadal Review, the
  ability of LSST to provide accurate positions for objects is
  a necessity.  Astrometric accuracy at the level of a few tenths
  of the size of the PSF is part of the basic processing pipeline,
  but if accuracies approaching those predicted by photon statistics
  (as limited by seeing, detector quantization, etc.), then exciting
  and fundamentally important scientific results can be obtained.

  B.1  The Parallax Survey

  LSST offers the opportunity to measure the parallax of every object
  in its field of view.  Although this sounds both obvious and trivial,
  this project has been beyond the reach of the astronomers until
  the Hipparcos satellite (Perryman et al. 1977) provided these data
  for the bright stars.  Many, most notably Murray (1986) attempted the
  measurement of the parallaxes for a large number of faint stars
  using Schmidt telescope survey plates and automated plate measuring
  machines, but the relatively large errors were encountered.  Indeed,
  the Yale Parallax catalog (van Altena 1995) lists values for only
  8,112 stars.

  LSST will change everything.  It will measure the positions for an
  estimated 1e10 stars several times per year, and the pipeline will
  produce measurements and uncertainty estimators for the position,
  parallax, and proper motion of every star it observes.  This will be
  the dawn of a new age in our understanding of the nearby stars.
  LSST will provide the first distance-limited catalog of the distances
  to all stars in the sky subject only to apparent magnitude and
  celestial coverage limitations.  In the past, studies of solar
  neighborhood have had to use magnitude, color, or proper motion as
  a surrogate for distance, at least in the selection process.
  No more.  The LSST catalog will provide the distances, and this
  will remove almost all of the selection effects that are present
  in the current studies.  This is one of the few times when the
  observational data (position as a function of time) measure the
  astrophysical parameter (distance) directly.

  The LSST survey of the nearest stars (within 10 pc) will be
  of great importance to a large segment of the community, and
  should be available after the first year of operation.  If LSST
  can observe the available sky twice per lunation, then the parallax
  and proper motion can be separated after a few months, and the
  large signal from a nearby star can be detected.  Of course,
  more observations will improve the determination and weed out
  the occasional binary system that fits the detection scheme,
  but parallaxes with enough accuracy (10% or better) for astrophysical
  follow-ups will be available in almost real time.  It is proposed
  to make this list available as quickly as possible so that further
  studies can be started before the stars leave the evening sky.
  No more will the astrometric studies of newly identified nearby
  stars (e.g., Dahn et al 2002 and Vrba et al 2003) await the curious
  process of photometric discovery.  Astrometry will be used to select
  the nearby stars without any constraints on photometry or spectroscopy.
  The discovery of L-dwarfs, T-dwarfs, and other objects with low
  luminosity are just previews of the new types of objects waiting
  to be discovered.

  B.2 The Wiggle Survey

  Binary stars offer one of the few opportunities to measure the
  mass of a star, but searches for binaries are rarely done.
  Usually, binaries are discovered by accident (corrupting an
  astrometric or spectroscopic analysis) or through directed
  surveys looking for companions around specific types of stars
  (e.g., planets around solar-type stars).  Once the binary nature
  of a system is demonstrated, many further observations are needed
  if an orbital solution is to be derived.  The Washington Double
  Star Catalog (WDS; Mason et al 2003) shows 5 systems with semi-major
  axes larger than 0.7 arcsec and periods less than 20 years.
  Indeed, the WDS lists orbits for fewer than 2000 systems.

  LSST will change all of this.  The residuals from the fits for
  position, proper motion, and parallax will be searched for
  the signature of Keplerian motion.  For systems whose orbital
  period is shorter than about a quarter of the available epoch
  difference, the periodic nature of orbital motion can be
  sensed and parameterized.  Once adequate coverage is available,
  an orbit can be computed.  Modern Fourier techniques can do this
  in an impersonal manner on all stars in the LSST archive.  A very
  exciting aspect of LSST is that the astrometry will be done in
  each of the survey passbands, and the amplitude of the astrometric
  perturbation can be measured as a function of color.  This will
  allow a statistical identification of the components involved,
  and allow the list of all binaries to be culled so that the
  important ones can be handed to the user community for additional
  observations.

  Perhaps the greatest contribution of the Wiggle survey will be
  in combination with the Parallax survey for the understanding
  of the star formation process.  The LSST data will provide
  an almost complete inventory of systems within a nearby volume,
  and the wiggles and common proper motion data can be used to
  identify the major components of each system.

=================

Contributions to Requirements Document

   1) Cadence needs to be 2 visits per field per lunation in the same
      color during the first year.  This flows from the Nearby Star
      Parallax survey's need to separate parallax from proper motion
      on the basis of the nominal 6 months of data per field during
      the first year.  After the first year, 1 visit per field per
      color is sufficient to measure parallax, proper motion, and
      wiggles from unseen companions.

   2) Sub-pixel sensor characterization.  The astrometric accuracies
      need to be at the hundredth of a pixel level per observation,
      and at the thousandth of a pixel level for the mean of several
      observations.  Once the LSST sensors have been selected, a detailed
      laboratory and observing campaign will be needed to understand
      the technology and to verify centroiding algorithms at the
      milli-pixel level.

   3) Telescope and PSF stability.  Upon review, astrometric science
      is fully enabled if the requirements imposed by the Dark Matter
      programs are implemented.

   4) Astrometric Tie to Bright Stars.  The system of J2000 is based
      on Hipparcos stars (V <= 10), and some method must exist to follow
      the mapping of the coordinate system from these bright stars
      to the deep exposures.  Neutral density filters might be needed,
      or perhaps the sensor technology allows for magnitude compensation,
      but astrometry must levy a requirement that a path for this
      mapping must exist.

   5) The Streams in the Galactic Halo project needs photometric
      data in non-Survey passbands to assist in measuring the
      luminosity and abundance of the stars.  Since variability
      is not an issue for most stars, these data can be taken
      once, presumably during the pre-Survey commissioning period.

=================

Questions that need SWG attention:

   What the relative merits of more colors to do color-color selection vs.
   more in one passband to do more precise astrometry?  The clear purpose of
   multicolor data is to separate stellar population (i.e. foreground dwarfs
   from halo giants, etc.). Some recent results using the SDSS filters suggest
   that broad band colors might do a reasonable job. The preferred narrow-band
   filters are probably out of the question for LSST.  Could they be done (at
   least in part) by a dedicated smaller telescope? This topic needs more work
   on understanding how much can be gleaned from a relatively standard
   and how much cannot.

=================

Clewley, L., et al. 200, MNRAS, 337, 87.

Dahn, C.C. et al (2002), AJ 124, 1170.

Geza, G., Evans, N.W., and Gates, E.I. 1998, ApJL, 502, 29.

Ivezic, Z. et al. 2000, AJ, 120, 963.

Kochanek, C.S. 1996, ApJ, 457, 228.

Mason, B.D., Wycoff, G.L., Hartkopf, W.I. (2003) 
 http://ad.usno.navy.mil/wds/

Morrison, H. et a. (2001) astro-hp/0111097.

Murray, C.A., Corben, P.M., Argyle, R.W. (1986) in Astrometric Techniques
 ed. H. Eichhorn and R. Leacock (Dordrecht, Reidel), p213.

Perryman, M.A.C. et al. (1997) The Hipparcos and Tycho Catalogs
(Noordwijk, European Space Agency SP-1200).

van Altena, W.F., Lee, J.T., Hoffleit, E.D. (1995) The General Catalog
of Trigonometric Catalogs (New Haven, Yale).

Vrba, F.J. et al, (2003) in preparation.

Wilkinson, M.I., and Evans, N.W. 1999, MNRAS, 310, 645.

LSST LSST LSST LSST LSST    Mailing List Server   LSST LSST LSST LSST LSST LSST
LSST
LSST  This is message 154 in the lsst-general archive, URL
LSST         http://www.astro.princeton.edu/~dss/LSST/lsst-general/msg.154.html
LSST         http://www.astro.princeton.edu/cgi-bin/LSSTmailinglists.pl/show_subscription?list=lsst-general
LSST  The index is at http://www.astro.princeton.edu/~dss/LSST/lsst-general/INDEX.html
LSST  To join/leave the list, send mail to lsst-request@astro.princeton.edu
LSST  To post a message, mail it to lsst-general@astro.princeton.edu
LSST
LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST LSST