Target Selection Work Week Report on Standard Stars
This is an updated posting of the spectroscopic standard star
module in "target". The first draft of this dates to that
put together by Burles, Kron and Schlegel at the target selection
work week in March 1999. However, this document was never posted,
only checked into target/doc/www.
This describes the targetting info for the following objects:
* guide stars
* SPECTROPHOTO_STD
* REDDEN_STD
* HOT_STD
* SKY
* light traps
The files in the "target" code to which this refers are:
target/etc/standardsParamsRead.tcl
target/examples/tsStandardsParams.par
target/include/taStandards.h
target/src/standards.c
We describe how priories are set for the spectro-photo standards,
though I'm not sure yet if/when/how they are applied.
-David
----------------------------------------------------------------------
STANDARD STAR MODULE
----------------------------------------------------------------------
This module chooses objects for the following categories:
(A) guide stars (GUIDE_STAR, sectarget=64)
(B) spectrophotometric standards (SPECTROPHOTO_STD, sectarget=32)
(C) reddening standards (REDDEN_STD, sectarget=2)
(D) hot stars (HOT_STD, sectarget=512)
(E) blank sky fiber positions (SKY and COHERENT_SKY)
Aside from the sky fibers, all catagories in the standard star module
make the following cuts using the OBJC_TYPE and OBJC_FLAGS as measured
by PHOTO:
OBJC_TYPE == 'star' - select stars
NOT AR_DFLAG_BRIGHT - exclude duplicates
NOT AR_DFLAG_CHILD - exclude children of blends
NOT AR_DFLAG_BLENDED - exclude any object with children as well
as any objects with multiple peaks
NOT AR_INTERP - exclude objects with saturated pixels,
cosmic rays, or bad columns
NOT AR_EDGE - exclude objects near the edge of frames
We also select objects with a minimum fiber magnitude in r'-band:
FIBERCOUNTS[r'] < spectralLimit = 20.5
----------------------------------------------------------------------
A. GUIDE STARS (GUIDE_STAR, sectarget=64)
Guide stars need to be selected according to the following
general criteria:
1) their positions must be accurately on the system of the
main targets (galaxies and quasars) - hence, they must not
be too bright.
2) there must be a moderately large number of them -
say > 50 per plate - to ensure several options for covering
the the plate area (this detailed assignment is done
downstream in PLATE)
3) to avoid stars likely to have high proper motion, we should
use color cuts to select against red dwarfs and white dwarfs
4) the guider is filtered in a band that approximates g', so
stars should be selected according to g' magnitude.
Stars bluer than g-r = 0.3 are unusual, so this is a sensible
blue limit. Similarly, few stars are redder than g-r = 1.4,
and those should be eliminated.
However, many low-luminosity stars have g-r < 1.4. To
eliminate these, we have a cut also in r-i. According to Lenz
et al., working with the Gunn-Stryker atlas, M0 V stars have
r-i = 0.7, and we choose to select only stars bluer than this.
If r = 14.25 (the nominal bright limit for good astrometry),
then V = 15.0. Adopting Mv = 8.8 for MO V, then the distance
is 170 pc. For a transverse velocity of 25 km/sec, the proper
motion in 5 years is 0.16 arcsec. While this is not very small,
this is a "worst case" example (and it is one star out of 10).
We want to include stars as red as this to enlarge the pool of
potential guide stars and to make the overall color distribution
of the guide stars more like the colors of galaxies.
To eliminate peculiar stars, we also limit stars to being
redder than r-i = 0.
Summary: the code selects candidate guide stars according to
the following criteria. All magnitudes and colors are apparent,
i.e., NOT corrected for reddening.
13.0 < g' < 15.5
0.3 < g'-r' < 1.4
0.0 < r'-i' < 0.7
-0.4 < i'-z' < 1.0
The magnitude limit was originally 14.0 < g' < 15.5, but those
stars were too faint for the guider! Also, the i-z color-cut
was added later, which will only exclude objects with bad photometry.
Priorities are assigned to first select stars closest to the red
limit in g'-r' (ie, 1.4) in order better to match the color of
typical galaxies. The priorities are rank-ordered in this color.
>From run=259 col=3 frames=50:249, we find 229 guide stars. This
corresponds to 33/deg^2 or 232 per plate. The galactic coordinates
span (l,b) = (126,-63) to (176,-50).
In view of this large number of selected stars, we should consider
making the following modifications:
cutting at r'-i' < 0.6 instead, to eliminate stars with higher proper motions
cutting in u'-g' to eliminate white dwarfs
----------------------------------------------------------------------
B. SPECTROPHOTOMETRIC STANDARDS (SPECTROPHOTO_STD, sectarget=32)
We select relatively bright F dwarfs as spectrophotometric
standards. Ideally we want the low-metallicity ones in order
to have spectra that are smooth and homogeneously so. Until
we gain some experience, we cannot be assured that color
selection will necessarily result in a homogeneous set of stars
in terms of metallicity. Therefore, we will presume that we
will have a range of metallicity, but this will be recognized
by the line strengths. Thus we should target several stars per plate.
The magnitude limit is in g-band only:
15.5 < g' < 17.0
The bright limit has been adjusted (from 14.0 originally) to
keep very bright objects off the spectrographs, since the scattered
light can compromise neighboring fibers.
The faint limit ensures that distant F dwarfs
are included, but not so faint that the S/N suffers.
The color selection is as follows. First, isolate only stars in
the box defined by:
0.6 < u'-g' < 1.2
0.0 < g'-r' < 0.6
Of these, choose only those stars above this line in color-color space:
g'-r' > 0.75 * (u'-g') - 0.45
This selects stars that are on the blue edge of the main sequence
in u'-g'. These should be the low-metallicity (halo) stars.
These magnitudes and colors are all supposed to be intrinsic, i.e.
corrected for reddening.
Priority will be given to the bluest stars selected, with u'-g' < 0.8.
There will only be a few such stars per plate.
In detail, we give highest priority to those stars with colors
closest to:
u'-g' = 0.80
g'-r' = 0.30
r'-i' = 0.10
This is meant to select those stars close in color to BD+17, which
David Hogg has computed to have these colors in the 2.5-m system:
u-g = 0.934
g-r = 0.280
r-i = 0.101
i-z = 0.013
Specifically, the priorities are set in the range [1,100] from:
priority = 100 - 100 * sqrt( (ug - 0.80)^2 + (gr - 0.30)^2 + (ri = 0.10)^2 )
Any priorities less than 1 in the above formula are set to 1.
>From run=259 col=3 frames=50:249, we find 45 spectrophotometric stars.
This corresponds to 6.5/deg^2 or 46 per plate. The galactic coordinates
span (l,b) = (126,-63) to (176,-50).
______________________________________________________________________
C. REDDENING STANDARDS (REDDEN_STD, sectarget=2)
There are two types of reddening standards, one of which is
plentiful and the other of which is rare. It makes sense to
pursue both types.
The signature of reddening is very clear when considering
ensembles of stars in color-magnitude or color-color diagrams
because of the existence of steep gradients (the halo main-
sequence turn-off, the saturation of g'-r' colors for
late-type dwarfs). The sampling of the reddening material on
the sky is good because these stars are common, and in principle
we can use a large range of magnitude. But, the success of
this technique depends on controlling for systematic population
(metallicity) differences in different directions. Spectroscopy
of some stars is intended to support a reddening determination
that is otherwise essentially photometric.
This category of star can be considered simply as a fainter
extension of the F (sub)dwarfs selected for spectrophotometry.
These stars *are* the stars that operationally define the blue tip
of the halo main sequence. All of the selection criteria are the
same as for (B) above, except the magnitude range (again, corrected
for reddening using the DIRBE dust maps) is
17.0 < g' < 18.5
The magnitude range is therefore contiguous with that for the
spectrophotometric stars. Assuming an absolute magnitude of 4.3
for these stars, the distance would be greater than 4 kpc, which
seems adequate. Ten of these stars will be observed
with "reserved fibers".
The priorities are assigned the same as for the spectrophotometric standards.
>From run=259 col=3 frames=50:249, we find 84 reddening F stars. This
corresponds to 12/deg^2 or 85 per plate. The galactic coordinates
span (l,b) = (126,-63) to (176,-50).
______________________________________________________________________
D. HOT STARS (HOT_STD, sectarget=512)
Alternatively for reddening determinations, we can take advantage of
stars that are so hot that they approximate a thermal source of very
high temperature. The temperature can be derived from helium line
strengths and He/H line ratios, and an intrinsic color determined
from models; the observed color then yields the reddening. We will
select these stars according to the following de-reddened magnitudes
and colors:
-1.5 < u'-g' < 0.0
-1.5 < g'-r' < 0.0
14.0 < g' < 19.0
The blue limits on the u-g and g-r colors are to cull bad targets.
These are expected to primarily be subdwarf O stars. Many of these
will be targeted for other reasons (esp. as QSO targets), so few fibers
will be needed. These stars will be tiled. We find approx. 1 to 3
per plate.
______________________________________________________________________
E. BLANK SKY FIBER POSITIONS (SKY and COHERENT_SKY)
PHOTO selects 5 blank sky positions in each frame. We pass along
the positions of all of these positions in the target code, for
a total of approximately 150 per plate.
----------------------------------------------------------------------
LIGHT-TRAP STARS
----------------------------------------------------------------------
Assuming absolute worst-case reflection, we want to eliminate all
stars with V < 8.5. This yields less than 1% of the sky background
due to scattered light for the stars that remain.
At high latitude we expect about 5 stars per plate that have
light-trap holes. We don't want too *few* holes, since we want
at least a couple holes per plate to keep the drilling operation
as routine as possible.
There is enough margin in the V = 8.5 limit that we don't need to
worry about stars of extreme color (which are rare anyway).
Proper motion is not an issue, even for nearby stars moving relatively
rapidly, and for a 5-year time interval.
We will use the Tycho catalogue to identify the stars meeting the
V < 8.5 criterion. The necessary reformatting has been done.
It is unlikely that light-trap holes will eliminate science targets
because even for the V = 8.5 stars, the region on the CCDs masked
out by light is large enough that PHOTO is unlikely to find objects
that close. Still, the pipeline needs to resolve any collisions
between light-trap holes and science targets.