Visits 3, 4, and 5 are listed by APT as 3 orbits each, whereas we were only awarded 2 orbits each. However, in each
case the 3rd orbits only contains the buffer dump of the final exposure (which ends during the 2nd orbit). If this is not
legal, please contact us.
Our science (measuring proper motions) requires that we reobserve all of our targets in Cycle 15 (this time has been
awarded). The PSF degredation caused by 2-gyro guiding might significantly degrade our results. Thus, we request
that our Cycle 13 observations be scheduled as early in the cycle as possible so that our Cycle 15 observations can
take place as early as possible. Our Cycle 15 observations will require precisely the same roll angle as the Cycle 13
observations and so will have to take place on nearly the same day of the year. So the dates of the Cycle 13 observations
will dictate the dates of the Cycle 15 observations (the narrow CVZ windows of Targets 3, 4, and 5 will also mandate
this).
----- Submission 1 (Fri May 14 20:27:21 GMT 2004) -----
Assigned ID: 956
----- Submission 1 (Sat Jan 24 00:57:52 GMT 2004) -----
Space Motions for the Draco and Sextans Dwarf Spheroidal GalaxiesWe will use the powerful astrometric capabilities of HST to measure proper
motions for the Draco and Sextans dwarf spheroidal galaxies that will yield
tangential velocities accurate to about 30 km/s. These two galaxies are the
last inside a galactocentric radius of 200~kpc without measured proper motions.
Knowing their orbits is critical for our understanding of the low-luminosity
satellites of the Milky Way. In particular they are critical for understanding why
Ursa Minor has survived tidal disruption on its plunging orbit and how Carina
formed a large intermediate-age stellar population despite its small mass.Piatek et al. (2002b; P02) and Piatek et al. (2003) describe our method for deriving proper motions
for dSphs with WFPC2 or STIS imaging. In brief, we use images at three epochs of fields which
contain a QSO that provides the ?zero-velocity? reference point. This proposal will use the same
methods as our previous program, except with ACS, to measure the proper motion of the Draco
and Sextans dSphs. Recently, SDSS (Rave et al. 2003) and a member of our original team (Harris
& Munn 2003) have independently discovered several QSOs behind Draco. The 2DF team has
found QSOs behind Sextans (Croom et al. 2001). We have used these data to select QSOs that are
within or just slightly beyond the core radius and have magnitudes fainter than 19.8. Both of these
conditions are required to ensure enough stars belonging to the dSph in the field with adequate
S/N. This yields three fields in Draco and two in Sextans. Unfortunately, no field contains more
than one QSO. Multiple fields both reduce the final uncertainty by averaging the measurements
and provide critical checks on the accuracy the proper motions. If the multiple measurements do
not agree within their uncertainties, then something is wrong. Three fields potentially allows the
rejection of one discrepant measurement.
We will observe each field in each of cycles 13, 14, and 15. Having three measurements instead
of two again provides a check on discrepant data. The advantage of the checking provided
by having multiple fields and epochs is discussed more fully in Piatek et al. (2003). We cannot
take more than one epoch per year because of the constraints imposed by the HST roll angle. Our
observations at second and later epochs are designed to reproduce the locations on the CCD that
the QSO and stars had in the first-epoch. This approach removes errors caused by geometrical
distortion in the camera, which alters the apparent separation between a QSO and a star. Such
distortion is not characterized accurately enough to simply correct for its effect. We have obtained
good results with this approach and so strongly prefer to continue with it for these new observations.
P02 shows that applying the above method to observations at three epochs spread over two
years in a single field produces a proper motion that is accurate to about 20 milliarcsec/century
in each coordinate. Data for three fields in Draco and two in Sextans will yield uncertainties in the
tangential velocity, vt, of at most 30 km/s and 35 km/s, respectively.
Observations in cycles 13 and 15 yield our desired two-year baseline. The cancelation of
SM4 means that it is somewhat unlikely that HST will still have three-gyro guiding (or even be
operating at all) for cycle 15. In these circumstances, our cycle 14 epoch provides insurance that
we will obtain proper motions, albeit with uncertainties larger by a factor of two. Our assumed
uncertainties are based on our experience of what can be achieved with STIS and WFPC2. It is
possible that the larger field and improved throughput of ACS will yield more accurate proper
motions. Our improved understanding of how to dither, discussed later in this section, should also
produce a small reduction in the final uncertainties. Thus, we feel a result based on a one-year
baseline would still be a large improvement over no knowledge of the proper motion at all. If we
are granted observing time, we will work with STScI to schedule our first epoch observations as
early in cycle 13 as possible. We would also explore the feasibility of using the imaging with an
increased jitter in one direction that will result for two-gyro guiding.
Anderson & King (1999, 2003; AK), Kuijken & Rich (2002; KR), P02, and Piatek et al.
(2003) have demonstrated that it is possible to measure positions for stars and QSOs accurate to
about 0.01 pixel or better using HST and from these to derive a proper motion. All of these studies
derive a point-spread function (PSF) which is subsequently used to measure the positions of stars
and QSOs. An accurate PSF is essential for deriving accurate positions. Its derivation in undersampled
images requires a large number of stars with high S/N (we use S/N > 15) distributed
throughout an image. The two features of our work on dSphs that distinguish it from studies of
globular clusters (AK) or field stars (KR) are: 1) the dSph fields contain several tens to a few
hundred stars compared to thousands or tens of thousands in fields towards a globular cluster and
2) the requirement for the presence of a ?zero-velocity? reference point, a QSO, in the field. Point
1) requires multiple, dithered exposures to increase the sampling of the PSF. Point 2) restricts the
number of dSph stars in the fields with adequate S/N because of the requirement that the QSO be
unsaturated.
The average magnitudes of the QSOs are in the range V = 18.9 ? 20.5. Our adopted exposure
times are conservative to avoid having the variability of the QSO saturate its image. The ideal
(conservative) exposure time produces the largest number of stars with S/N > 15 while aiming to
keep the brightest QSO pixel about 1/3 of full well. We can often come close to this ideal.
We have opted to use the ACS WFC for the new fields in Draco and Sextans. The rationale
for the choice of the ACS WFC over the ACS HRC, which has smaller pixels that better sample
the stellar images, is as follows. We use our observations of Ursa Minor with STIS to estimate
the number of stars expected and their S/N with different observing strategies? Ursa Minor has
a similar stellar density to Draco and Sextans and the QSO in this Ursa Minor field is similar in
brightness (V =20) to those in the Draco and Sextans fields. In two CVZ orbits we obtained 6
images at each of 8 dither positions with an exposure time of 176 s each. In a single image, there
are 17 stars with a S/N > 20 (these have a positional uncertainty of 0.05 pixel or less in each
coordinate) and 35 stars with a S/N > 10 (uncertainties of 0.1 pixel or less). The ETC shows
that ACS HRC with the F606W filter is about as efficient as the unfiltered STIS. However, HRC
has only a quarter of the field of view of STIS, so only 4 stars with S/N > 20 are expected to
be in the field and this is too few to construct an accurate PSF. Even the unsupported unfiltered
mode of HRC would only gain a factor of two in the number of stars (and, in this mode, the PSF
is degraded). An exposure time that is four or more times longer might bring the number of stars
up to a marginally acceptable level, but the observing strategy described below using ACS WFC is
more efficient and will still yield excellent proper motions.
We prefer to use ACS WFC and the F606W or F775 filter over STIS because: 1) It has an
8x larger field (we cannot measure the separation of the two WFC CCDs and so will use the data
from just one). 2) It has a more constant PSF across the field (confirmed by Anderson at the last
HST Calibration Conference). 3) Its geometrical distortions will be accurately characterized (Jay
again). 4) Its throughput is three times larger. 5) Finally, it will probably have fewer damaged
pixels. The pixels of the WFC are the same size as those of STIS. The STIS data produce good
results, therefore WFC should do so too. The WFC has a larger distortion across the field than
STIS does. As was discussed above, placing the stars and QSO on nearly the same pixel at every
epoch minimizes the impact of this distortion. This placement requires that all epochs of a given
field be taken with the same roll angle of the telescope. The roll angle is not pre-determined, but
will become fixed after the first-epoch observations. Although the associated overhead is high, the
requirement that stars be placed back on the same pixels is critical.
An exposure time of 340 s with WFC and the F606W filter results in about three times the
number of electrons for a comparable star than does the exposure for Ursa Minor using STIS
described above. Because of the larger field and higher throughput, such a WFC image will have
about 250 stars with S/N > 20 per field in Draco and Sextans. This number of stars is comparable
to, or greater than, the largest number of stars which we have achieved with STIS for any dSph.
The exposure time of 340 s is just long enough to avoid the serious 5.8-minute overhead per image
due to serial buffer dumps. For QSOs brighter than V = 20, we want shorter exposure times to
keep the QSO image at 1/3 of full well. In this case, we will switch to a narrower filter, keeping the
exposure time as close to 340 s as possible. The brightest QSO in Draco causes the biggest problem
? for this field we will use the F775W filter. This approach is more efficient than incurring the
buffer dump overhead. We could reduce the minimum exposure time to avoid the buffer dump by
reading out only part of the image (since we will use the data from only 1 chip), but this has its
own overhead. Thus, we prefer to change filters, using one WFC chip for astrometry and placing
data for both chips in the archive. For the faintest QSOs in our fields, an exposure time of 495 s
puts the brightest QSO pixel at 1/3 of full well. To keep the exposures for a single field to 2?3
orbits, we compromise at 400 s. The resulting S/N for the QSO and dSph stars is still higher than
those obtained in our Ursa Minor STIS field.
For the exposure times which we use, our experience is that hot pixels corrupt many more
pixels than cosmic rays do. We also find it unnecessary to reconstruct cosmic-ray-cleaned images.
We fit for the positions of objects in all of the images and simply ignore those instances that are
too strongly affected by corrupted pixels. Cosmic-ray splits are useful for identifying cosmic rays,
but our software is able to identify corrupted pixels without them. Thus, the dithering between
each image will combine integer-pixel offsets -- to reduce the number of instances when an object
is lost to hot pixels -- with fractional pixel offsets -- to increase the number of samples of the
PSF. This approach produces more independent samples of the image of each QSO or star with
fewer images (hence, fewer readout times) than does our previous scheme of 3 images at 8 dither
positions used with STIS. Draco is in the CVZ and 18 (21) exposures of length 400 s (340 s) will
fit into two orbits. For Sextans, three non-CVZ orbits are enough for 12 - 14 such images. This
scheme produces at least 1.5X the number of samples for each object than what we have been
getting with STIS. It also obtains at least 20% more electrons per comparable object than the Ursa
Minor STIS observations. These ACS observations will require a total of 12 orbits in each cycle --
2 orbits for each of the 3 Draco fields and 3 orbits for each of the 2 Sextans fields.
In summary, our program will require 12 orbits in cycle 13, 12 orbits in cycle 14, and 12 orbits
in cycle 15. The total is 36 orbits spread over the three cycles.STELLAR POPULATIONS12falsefalsefalsefalsefalseSUPPORTEDGALAXYDWARF SPHEROIDALQSOThe QSO J1012-0130 is centered in this field. The field also contains stars of the
Sextans dwarf spheroidal galaxy.J200010 12 43.7-01 30 43.0J2000GALAXYDWARF SPHEROIDALQSOThere is a QSO centered in the field. The field also contains stars of the Sextans
dwarf spheroidal galaxy.J200010 13 56.1-01 43 20.0J2000GALAXYDWARF SPHEROIDALQSOThere is a QSO centered in the field. The field also contains stars of the Draco dwarf
spheroidal galaxy.J200017 21 48.3+57 58 5.4J2000GALAXYDWARF SPHEROIDALQSOThere is a QSO centered in the field. The field also contains stars of the Draco dwarf
spheroidal galaxy.J200017 20 52.31+57 55 14.4J2000GALAXYDWARF SPHEROIDALQSOThere is a QSO centered in this field. The field also contains stars of the Draco dwarf
spheroidal galaxy.J200017 19 34.43+57 58 49.8J2000ACS/WFCImagingNoF606WACS/WFCImagingNoF606WACS/WFCImagingNoF606WACS/WFCImagingNoF606WACS/WFCImagingNoF775W
(Tic-Visit
:proposal-id 10229
:visit-id "01"
:proposal-version -1
:trans-version "20040225"
:date "14-May-2004 13:00:13.0"
:trans-errors '(:WARNINGS 0 :ERRORS 0 :HEALTH-AND-SAFETY 0
:FATAL-ERRORS 0 :CONTINUABLE-FATALS 0)
:number-of-orbits 3
:saa-model "27"
:si-bright-earth-limb-angle 20.0
:data-volume 4263360
:tdrss-contact-duration 0
)
(Tic-Obset
:proposal-id 10229
:visit-id "01"
:obset-id "01"
:pcs-mode "FGS"
)
(Tic-Orbit
:proposal-id 10229
:visit-id "01"
:orbit-number 1
:pcs-time 365
:si-vis-utilization 2835
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
(Tic-Orbit
:proposal-id 10229
:visit-id "01"
:orbit-number 2
:pcs-time 300
:si-vis-utilization 2900
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
(Tic-Orbit
:proposal-id 10229
:visit-id "01"
:orbit-number 3
:pcs-time 300
:si-vis-utilization 2900
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
trueC:\Documents and Settings\pryor\My Documents\Research\dsph-pm\cycle13\OP-cache\cycle13-v01.xml
(Tic-Visit
:proposal-id 10229
:visit-id "02"
:proposal-version -1
:trans-version "20040225"
:date "14-May-2004 15:55:40.0"
:trans-errors '(:WARNINGS 0 :ERRORS 0 :HEALTH-AND-SAFETY 0
:FATAL-ERRORS 0 :CONTINUABLE-FATALS 0)
:number-of-orbits 3
:saa-model "27"
:si-bright-earth-limb-angle 20.0
:data-volume 3410688
:tdrss-contact-duration 0
)
(Tic-Obset
:proposal-id 10229
:visit-id "02"
:obset-id "02"
:pcs-mode "FGS"
)
(Tic-Orbit
:proposal-id 10229
:visit-id "02"
:orbit-number 1
:pcs-time 365
:si-vis-utilization 2835
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
(Tic-Orbit
:proposal-id 10229
:visit-id "02"
:orbit-number 2
:pcs-time 300
:si-vis-utilization 2900
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
(Tic-Orbit
:proposal-id 10229
:visit-id "02"
:orbit-number 3
:pcs-time 300
:si-vis-utilization 2900
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 1920
:low-sky "N"
:pcs-visibility 3200
:pcs-gap 0
)
trueC:\Documents and Settings\pryor\My Documents\Research\dsph-pm\cycle13\OP-cache\cycle13-v02.xml
(Tic-Visit
:proposal-id 10229
:visit-id "03"
:proposal-version -1
:trans-version "20040225"
:date "14-May-2004 15:56:36.0"
:trans-errors '(:WARNINGS 0 :ERRORS 0 :HEALTH-AND-SAFETY 0
:FATAL-ERRORS 0 :CONTINUABLE-FATALS 0)
:number-of-orbits 3
:saa-model "27"
:si-bright-earth-limb-angle 20.0
:data-volume 5400256
:tdrss-contact-duration 0
)
(Tic-Obset
:proposal-id 10229
:visit-id "03"
:obset-id "03"
:pcs-mode "FGS"
)
(Tic-Orbit
:cvz T
:proposal-id 10229
:visit-id "03"
:orbit-number 1
:pcs-time 365
:si-vis-utilization 11196
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 0
:low-sky "N"
:pcs-visibility 0
:pcs-gap 0
)
trueC:\Documents and Settings\pryor\My Documents\Research\dsph-pm\cycle13\OP-cache\cycle13-v03.xml
(Tic-Visit
:proposal-id 10229
:visit-id "04"
:proposal-version -1
:trans-version "20040225"
:date "14-May-2004 15:57:15.0"
:trans-errors '(:WARNINGS 0 :ERRORS 0 :HEALTH-AND-SAFETY 0
:FATAL-ERRORS 0 :CONTINUABLE-FATALS 0)
:number-of-orbits 3
:saa-model "27"
:si-bright-earth-limb-angle 20.0
:data-volume 5400256
:tdrss-contact-duration 0
)
(Tic-Obset
:proposal-id 10229
:visit-id "04"
:obset-id "04"
:pcs-mode "FGS"
)
(Tic-Orbit
:cvz T
:proposal-id 10229
:visit-id "04"
:orbit-number 1
:pcs-time 365
:si-vis-utilization 11196
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 0
:low-sky "N"
:pcs-visibility 0
:pcs-gap 0
)
trueC:\Documents and Settings\pryor\My Documents\Research\dsph-pm\cycle13\OP-cache\cycle13-v04.xml
(Tic-Visit
:proposal-id 10229
:visit-id "05"
:proposal-version -1
:trans-version "20040225"
:date "14-May-2004 15:58:16.0"
:trans-errors '(:WARNINGS 0 :ERRORS 0 :HEALTH-AND-SAFETY 0
:FATAL-ERRORS 0 :CONTINUABLE-FATALS 0)
:number-of-orbits 3
:saa-model "27"
:si-bright-earth-limb-angle 20.0
:data-volume 5400256
:tdrss-contact-duration 0
)
(Tic-Obset
:proposal-id 10229
:visit-id "05"
:obset-id "05"
:pcs-mode "FGS"
)
(Tic-Orbit
:cvz T
:proposal-id 10229
:visit-id "05"
:orbit-number 1
:pcs-time 365
:si-vis-utilization 11196
:dark-offset NIL
:dark-utilization 0
:max-dark-utilization 0
:low-sky "N"
:pcs-visibility 0
:pcs-gap 0
)
trueC:\Documents and Settings\pryor\My Documents\Research\dsph-pm\cycle13\OP-cache\cycle13-v05.xml