SHARC-II Data reduction Workshop

1
SHARC-II Data reduction Workshop
Darren Dowell Attila Kovacs Colin Borys Darek
Lis Min Yang Jon Bird

• SHARC-II DRW 11/08/2004

2
Outline

• GOALS
• Caltech success stories
• Software Requirements and Installation
• Overview of available software
• Installation overview
• Scripting with CRUSH
• CRUSH vs SHARCSOLVE
• General SHARC-II Calibration
• Opacity (tau) estimation
• Calibrators at 350 micron
• Stability of calibration
• Calibrating your data
• Aperture Photometry
• PSF Photometry
• Example
• Map tweaks and presentation
• Pointing, Calibration, Coadding, Cropping, and
  Mosaicing
• Making publishable images with IDL
• Chopped data

SHARC-II DRW 11/08/2004
3
Goals

• 1. Transfer expertise from Caltech to other users
• Why? To promote publication of SHARC-II data
• Transfer expertise from other users to Caltech
• Why? Users tend to be familiar with the issues
  involved.
• 3. Learn about difficulties users have with
  instrument/data
• Why? To improve the system.
• 4. Improve data acquisition techniques
• Why? Improve efficiency of SHARC-II
  observations

SHARC-II DRW 11/08/2004
4
Caltech Success Stories
The next few slides present results from
observations conducted by members of the Caltech
SHARC-II group. They span a variety of flux
levels and redshifts, and are meant to illustrate
the full range of SHARC-IIs abilities.
SHARC-II DRW 11/08/2004
5
SLUGS (zlt0.05)

• Dunne et al. have characterized the SED of 106
  IRAS selected galaxies at 850?m
• Of those, only 17 were detected by SCUBA at 450
  ?m, and it was noted that the data supported a
  2-component SED fit.
• SHARC-II has detected roughly 60/65 targeted so
  far.
• They are so easy to detect that they now are done
  as poor-weather backup.

• 0.5h _at_ ?2250.06
• 1.0h _at_ ?2250.08
• 0.5 Jylt S350 lt 3 Jy
• crush -faint -compact

SHARC-II DRW 11/08/2004
6
Spitzer HLIRG (z1.5)

• In follow-up observations of Spitzer selected
  objects, we discovered an object with an apparent
  luminosity above 1013.5 L?.
• It has an SED similar to Arp220, but at at a
  redshift of 1.5.
• This object has sparked interest in Silicate
  Dropouts as a way to select high-z starbursts.

• 0.5h _at_ ?2250.06
• S350 226 45 mJy
• CHOPPED observing
• sharcsolve reduction

SHARC-II DRW 11/08/2004
7
Stanford sample (0.1ltzlt1.0)
2h _at_ ?2250.05 S350 44.1 mJy Td 40.9 K , ?
1.5

• The Stanford sample was compiled from
  cross-correlation of the faint-IRAS catalog and
  the FIRST 21cm radio catalog
• The sources are ULIRGs lying within redshift
  range of 0.1 and 1 NIR morphologies of these
  objects reveal they tend to be interacting
  systems
• FIR/submillimeter fluxes were obtained for the
  first time on these targets, so were SED fits in
  the FIR

SHARC-II DRW 11/08/2004
8
Fomalhaut Debris Disk

• 3.0h _at_ ?2250.039
• SWEEP (Lissajous)
• crush -deep reduction
• Peak fluxes 150 mJy/beam
• Integrated flux 1.2 Jy
• Consistent with thin, uniform dust ring

K. Marsh et al. (2004-5)
SHARC-II DRW 11/08/2004
9
Orion
Johnstone Bally (1999)
Houde et al. (2004)
rms 0.3 Jy/beam
rms 0.3 Jy/beam
rms 1 Jy/beam
rms 0.3 Jy/beam
4 hr. 1.2 mm PWV
18 hr. 1.0 mm PWV
mosaic of BOX scans
SHARC-II DRW 11/08/2004
10
Low-z Interacting Galaxy Survey

• Selected by IRAS 100 mm flux and proximity using
  criterion of Surace (2004).
• Perhaps analogues to zgt1 ULIRGs.
• Lissajous scans in typically t225airmass
  0.05-0.10.
• Observations of 14/42 sources complete, to be
  reported by J. Bird and D. Dowell.

51010 Lsolar
SHARC-II DRW 11/08/2004
11
Chopped High-z survey

• CHOP in azimuth 39?/1.39 Hz plus slow SWEEP
• parallactic angle rotation of 77 washes out
  negative beams in sharcsolve reduction
• 8.6 hrs, median tau225 0.044
• rms 10 mJy/beam in middle
• 2 sources are 80 mJy each
• Definite and probable SCUBA 850 mm sources marked
  with stars

SHARC-II DRW 11/08/2004
12
High-z SCUBA sources

• 3-4h _at_ ?225lt0.06
• ?350 5 mJy

LE2
LE31
LE12
LE21
SHARC-II DRW 11/08/2004
13
Software and Requirements

• CRUSH is Java based and written and maintained by
  A. Kovacs
• Available on web page
• Has been successfully used on
• Windows
• Mac OSX
• Linux
• Solaris
• SHARCSOLVE is C based and written by D. Dowell
  (phasing out?)
• Available only by special request
• Has been successfully used on
• Mac OSX
• Linux
• Solaris
• Ancillary software written by C. Borys.
• Available on web page
• Has been successfully used on
• Mac OSX
• Linux
• GCC compatible, so in principle could be compiled
  on other platforms. (C. Borys)

SHARC-II DRW 11/08/2004
14
Overview of Ancillary software 1

• header_update alters header keywords in maps
  generated by CRUSH.
• boxscan helps calculate BOX_SCAN parameters for
  SHARC-II observations.
• See SHARC-II web page for more details
• sharccal applies a calibration to a reduced
  sharc2 signal and noise map
• sharccal -c scalefactor -o offset -v -u
  units raw.fits calibrated.fits
• If no options, it will use the builtin
  calibration factor (V2JY in fits file)
• Output signal scalefactor(raw map offset)
• FITS keywords added or modified
• NAME value comment
• ---- ----------- ------------------
• V2JY scalefactor calibration factor
• CALAP T calibration
  applied?
• OFFAP TF offset applied?
• OFFSET value offset value (only
  if OFFAPT)
• BUNIT unit Units of output
  image
• sharcgap tests a raw sharc2 data file for
  timing gaps

SHARC-II DRW 11/08/2004
15
Overview of Ancillary software 2

• sharclog Gets basic information from a SHARC-II
  raw data file
• sharclog startidx stopidx
• Needs to be run from the directory in which data
  is stored.
• It scans the header of each file to provide a
  summary of the data (similar to Darrens)
• sharcstat computes basic statistics on a
  reduced SHARC2 map
• sharcstat file.fits
• gt sharcstat zw247_2.fits
• NX NY N s_mean s_stddev
  rms_mean rms_stddev
• 118 76 5049 0.00049829 0.14564
  0.023233 0.0072745
• N number of pixels with data in them (checks
  for NaN).
• S_ corresponds to mean and standard deviation of
  the SIGNAL map
• Rms_ corresponds to mean and standard deviation
  of the NOISE map.
• sharctau uses Jon Bird's tau fits to estimate
  tau for a given SHARC2 file. You also need the
  taufit files
• sharctau -v datafile taufile
• Can be run from anywhere
• gt sharctau -v /home/bigdisk1/sharc2-012900.fits
  /scr/borys/sharc/tau225_sep2003.fit

SHARC-II DRW 11/08/2004
16
Installation

• Create a convenient place for the CRUSH
  installation.
• Use logical links to point to the most recent
  version.
• CRUSH can ONLY be run from the directory in
  which it is installed.

istari (745am) /scr/borys/sharc/attila gtls
-la lrwxrwxrwx 1 borys cittgp 10
Nov 5 0821 crush -gt crush-1.34 drwxr-xr-x 3
borys cittgp 4096 Oct 3 0716
crush-1.33 drwxr-xr-x 3 borys cittgp
4096 Sep 22 0939 crush-1.33b2 drwxr-xr-x 3
borys cittgp 4096 Oct 6 1548
crush-1.34 drwxr-xr-x 2 borys cittgp
4096 Aug 31 2313 data drwxr-xr-x 3 borys
cittgp 4096 Oct 28 0035 devel drwxr-xr-x
2 borys cittgp 4096 Feb 13 2004 MaiTau

• Create a convenient place for the ancillary
  software.
• Add the directory to your PATH variable
• These programs can be run from anywhere.

istari (748am) /scr/borys/sharc/code/bin gtls
-la -rwxr-xr-x 1 borys cittgp 6793
Aug 27 1431 boxscan -rwxr-xr-x 1 borys
cittgp 16548 Aug 27 1410 sharccal -rwxr-xr-x
1 borys cittgp 784529 Mar 26 2003
sharcextract -rwxr-xr-x 1 borys cittgp
766548 Aug 27 1431 sharcgap -rwxr-xr-x 1
borys cittgp 770786 Aug 27 1431
sharclog -rwxr-xr-x 1 borys cittgp
770704 Aug 27 1432 sharcsmooth -rwxr-xr-x 1
borys cittgp 16164 Aug 27 1413
sharcstat -rwxr-xr-x 1 borys cittgp
766589 Aug 27 1432 sharctau
SHARC-II DRW 11/08/2004
17
Scripting
The necessity of running CRUSH from its install
directory makes file management slightly tricky.
Ways of manipulating output name include

• 1) -outpath OR REDUCED_MAP_PATH in crush.cfg
• This will change the path in which the file is
  saved, but not alter the name itself. i.e. it
  will keep the form OBJNAME.SCAN1.SCAN2SCANN.fits
• This filenaming structure is sometimes
  inconvenient (e.g. GAIA)
• 2) -name/path/to/mapdir/map.fits
• This will alter the name to one of your choosing.
  You can include a path here as well. If used,
  outpath is ignored.

RECOMMENDATION use scripts and the -name option
!/bin/csh cd /scr/borys/sharc/attila/crush echo
PROCESSING ic5634 ./crush -faint -compact
-nameic5634_1.fits 14176 14177 gt!
ic5634_1.log ./crush -faint -compact
-nameic5634_2.fits 14182-14185 gt!
ic5634_2.log ./coadd -outic5634.fits
ic5634_1.fits ic5634_2.fits gt! ic5634.log
sharcsmooth ic5634.fits ic5634_smooth.fits mv -f
ic5634 /scr/borys/sharc/projects/slugs/CRUSH/maps
/.
Note the 2 different ways of specifying scans to
analyze
SHARC-II DRW 11/08/2004
18
CRUSH
At this point in the workshop, Attila gave a
presentation on CRUSH. Download that separately
and review it before proceeding.
SHARC-II DRW 11/08/2004
19
Important Differences between CRUSH and SHARCSOLVE

• Pointing
• Different treatment of the case that the IRC
  Reference Pixel is not the middle of the array
  (16.5, 6.5)
• Calibration
• CRUSH and sharcsolve use completely different
  units, so cannot mix.
• CRUSH corrects for dependence of detector gain on
  detector loading, so resulting tau relations
  should look normal to SCUBA and SHARC users
  ?(SHARC II, 350 ?m) 25(?225 0.01)
• sharcsolve does NOT correct for gain change, so
  tau scaling looks too big ?(SHARC II, 350 ?m)
  32(?225 0.01)
• Chopped reduction
• sharcsolve differences with respect to chopper as
  first step.
• CRUSH treats secondary chopping as merely another
  pointing offset.
• Relative advantages of two approaches under study.

SHARC-II DRW 11/08/2004
20
Tau and Calibration

• Calibration at short sub-mm wavelengths is
  challenging, but necessary.
• In the next few slides, we present our procedure
  for estimating the atmospheric opacity, and then
  discuss the overall calibration uncertainty for
  SHARC-II
• Then we provide a more detailed example of how to
  obtain the calibrated flux for a specific
  observation.

SHARC-II DRW 11/08/2004
21
SHARC-II Calibrators

• Availability of calibration sources has always
  been a problem in sub-mm observations,
  particularly at shorter wavelengths (cant use
  BLAZARS, etc)
• We use primary calibrators (Mars, Uranus, and
  Neptune) to bootstrap the calibration of the
  secondary systems.
• For stationary objects, we can use repeated
  observations to derive averages.
• For solar system objects, we need to consider
  the changes in distance and solid angle over
  time.
• TB T1AU?r(-1/2)
• S(?) ??B?(TB)
• T1AU is derived by evaluating TB given all the
  other parameters (r is the heliocentric distance
  in AU, ? is the solid angle as seen from Earth,
  and B is the Planck function evaluated at the
  appropriate frequency (typically 350 micron).
  These values are provided by the JPL Horizons
  System http//ssd.jpl.nasa.gov/cgi-bin/eph
• We have used these relations to extrapolate the
  fluxes for all days between 2002 and 2010.
• These calibrations are available for download
  for the SHARC-II web page

SHARC-II DRW 11/08/2004
22
SHARC-II DRW 11/08/2004
23
Tau Fits

• Tau Dippers are noisy by nature (single
  measurement every 10 min).
• Fits for both the 225Ghz and 350 micron data
  exist for every SHARCII night to date.
• Least square polynomial fits over a large range
  of each night (almost always covering the entire
  observing time).
• Images of these fits are located on the SHARCII
  website (www.submm.caltech.edu/sharc).
• When reducing your data, observers should look at
  these tau fits as a FIRST step, so that they may
  determine the best fit to use and what was
  happening in the atmosphere at the time of
  observation.

SHARC-II DRW 11/08/2004
24
Tau Fits
Typical 350 micron fit. Residuals are located on
bottom of plot. Typical fit ranges from 2 to 20
hours UT. Notice that the X axis is in fraction
of a day.
SHARC-II DRW 11/08/2004
25
Keep an eye on the 225GHz and 350 micron
fitsthey CAN differ
Fits from both tippers on the same night.
SHARC-II DRW 11/08/2004
26
CRUSH and Tau

• CRUSH uses the MaiTau server to obtain the fitted
  tau.
• Parses through fit table (see below). Available
  online in conjunction with the tau fits.
• CRUSHs output will inform you if a fit for your
  file was found and what value was retrieved. Got
  Mai-Tau! tau(350um) X

By default, MaiTau looks at the 350 micron fits.
Use -taufit option to choose which fit, or not
to use a fit at all.
SHARC-II DRW 11/08/2004
27
Mai Tau success story on Mon R2
Below are maps made from individual scans of
MonR2 (provided by D. Benford). The raw tau
recorded in the file was used.
Problem Much variability.
28
Mai Tau success story on Mon R2
Below are maps made from individual scans of
MonR2 (provided by D. Benford). Opacity this
time was provided by MaiTau.
MaiTau helps!
29
SHARC-II and Calibration
Want to determine how stable that conversion
factor, and thus calibration, is over time.

• Perform aperture photometry on calibrators with
  known fluxes.
• Known fluxes are obtained from HORIZONS.
• CRUSHs default output is in Volts- constant
  Volts to Janskys applied (crush.cfg).
• By comparing known flux with CRUSH reported flux,
  we obtain a conversion factor.

SHARC-II DRW 11/08/2004
30
Calibration Stability
Plot shows conversion factor of calibrators taken
during the August-September 2004 run.
Conversion factor is consistent to within 21.3
for all 18.8 for Neptune 19.2 for Uranus
SHARC-II DRW 11/08/2004
31
Calibration Stability
Calibration is consistent over a wide range of
elevations.
You do not need to take calibration scans at the
same elevation as your science.
SHARC-II DRW 11/08/2004
32
Calibration/Tau summary

• Tau fits Important for understanding what is
  happening to the atmosphere during observation.
• Always look at the tau fits as a first step
  towards calibration and reduction.
• CRUSH calibration is now consistent to within 20
  and improving.
• Calibration is consistent over the range of
  telescope positions.
• In the next few slides, we concentrate on object
  specific calibrations.

SHARC-II DRW 11/08/2004
33
Calibration PSF
PSF Photometry is the most often used technique
for point source extraction in SCUBA maps
(particularly high-z projects). It is
mathematically equivalent to convolving with
the beam, except it also takes into account the
pixel-to-pixel noise differences. The procedure
is very straightforward. Start at a given pixel
(i,j), and calculate the following
statistic S and N are the signal and noise
maps respectively. This is simply a LLS fit, and
it is easy to derive the best fit value and error
for the PSFs amplitude, A. This can be extended
to include an offset parameter as well. The
ancillary program sharcsmooth performs this
function. It assumes a purely Gaussian, with a
default (but user settable) FWHM of 9. (see
imagetool as well) The map answers the following
statistical question what is the best fit
amplitude to a Gaussian centered at a given
pixel? There are consequences to this
assumption. i.e., for pixels near the peak of a
source, the assumption that that given pixel is
the center of a source is wrong.
Source extraction with a smoothed map is done
by setting a SNR threshold to search for sources
in the field.
SHARC-II DRW 11/08/2004
34
Calibration Aperture
Aperture Photometry is another popular choice,
most often used on sources that are readily
visible in the map, or if some other astrometric
marker is available on which the aperture can be
centered. There are 3 circular radii to choose.
In order of increasing value they are source,
inner sky, and outer sky. The annulus defined by
the last two radii are used to estimate the mean
sky level AND the scatter of pixels. The central
aperture is used to sum up the flux contained
within it (after correction for the mean sky
offset). The error on the flux estimate is
related to the RMS of the pixels in the sky
annulus, and the number of pixels in both the
aperture and annulus. Aperture radii usually
chosen via curve of groth, annuli choosen to
minimize noise while still providing a good
estimate of sky background and RMS. Given that
the pixel to pixel errors are dominated by
residuals in the sky estimation and not shot and
photon noise as they are in optical CCD work, the
equations are simpler. Note that CRUSH and
SHARCSOLVE do provide a noise map, but I have
always found that the RMS scatter in the signal
pixels is higher than what the weight map
implies. Thus I assume a uniform weight per map
pixel, and calculate this weight via the RMS of
the signal map. (implications for PSF
fitting) It is not yet clear to me that sky
estimation is necessary. CRUSH and SHARCSOLVE do
a pretty good job of returning zero for a map
mean when we look at blank sky. However,
experience has also shown that well detected
sources sometimes have a negative bowl around
them.
SHARC-II DRW 11/08/2004
35
Calibration Aperture

Important caveat This procedure assumes that the
pixels are uncorrelated. This is NOT the default
procedure for CRUSH, and one has to use
-convolve-1 to force this. Otherwise, the RMS
calculation will be lower then it is supposed to
be (youve essentially smoothed the map). If you
do not use the convolve flag, the RMS should be
increased by a factor of sqrt(N), where N the
number of pixels that fall within the area of the
convolving function. By default, we use an 8
beam, with noise calculated for 4.8 pixels,
which therefore requires a sqrt(1.33)(8/1.4)
6.6 increase in RMS (and consequently the total
error budget).
SHARC-II DRW 11/08/2004
36
Aperture vs. PSF
So which should you use, and what are some
caveats? POINT SOURCES If the PSF is varying
(ie, DSOS not functioning or not turned on),
APERTURE is probably the safer choice. If
CHOPPING, care must be taken to keep the annulus
away from the offbeams (a concern for SHARCSOLVE,
not CRUSH), hence PSF might be a good
choice. What do I use? Aperture, almost
exclusively, but use the PSF smoothed map for
presentation. EXTENDED SOURCES PSF, since it
essentially gives you Flux/beam. CAVEATS In deep
integrations, there are some issues related to
correlated sky signal still in the map. (more
from Attila)
SHARC-II DRW 11/08/2004
37
Calibrating your data

• The principles involved with calibrating SHARC-II
  data are applicable to all data from other sub-mm
  telescopes.
• Ingredients
• Good estimates of the atmospheric opacity for
  all science and calibration observations
• A decent collection of calibrators (different
  objects, airmasses, etc.)
• A CHOICE IN HOW YOU WILL EXTRACT FLUXES FROM
  YOUR DATA. What is done to the science map must
  also be done to the calibration.

• EXAMPLE Reduction of a local IRAS galaxy MRK
  331
• Scans 9125-9127, taken on Jan 15, 2003, at UT
  0454
• PSF photometry to be used

SHARC-II DRW 11/08/2004
38
Calibration Example 1
Using sharclog, (or by some other log or by
looking at the header), I find the UT time and
date the data were taken, and then go to the
SHARC-II web page to get the tau-fit plot for
that night. The data were taken at UT 0430
(0.20 fractional day). Fits look OK, so I will
not override MaiTau. Next I run CRUSH to make
the map
SHARC-II DRW 11/08/2004
39
Calibration Example 2

• gt ./crush -faint -compact -convolve-1
  -namemrk331.fits 9125-9127 gt! mrk331.log
• gt sharcsmooth mrk331.fits mrk331_smooth.fits
• gt cat mrk331.log
• crush -- Comprehensive Reduction Utility for
  SHARC2
• Author Attila Kovacs ltattila_at_submm.calte
  ch.edugt
• Version 1.34-1
• Scan 1 Reading /home/bigdisk1/sharc2/sharc2-009
  125.fits...
• Got Mai-Tau! tau(350um) 1.3088
• 83 HDUs, 16439 x 36ms frames -gt 9.9 minutes
  total.
• Filtering 13.4Hz on noisy pixels.
• DownSampling -gt 5479 frames
• MRK331 observed at 2003-01-15T045401.949
• RA 234854.0 DEC 201829.0
  (1950.0)
• 235126.7 203510.1
  (2000.0)
• AZ 2774454.5 EL 570157.9
• RAO 0.0 DECO 0.0 AZO 0.0
  ZAO -0.0
• FAZO-104.0 FZAO-30.0 Rotator 60.0
  RotZero 60.0

SHARC-II DRW 11/08/2004
40
Calibration Example 3

• Now I check the logs for that date to see which
  calibrators were done.
• In this particular case, I will only pick the
  one closest to the science observation, but you
  should reduce ALL of them and ensure that they
  seem reasonable.

gt ./crush -compact -convolve-1 -nameoh231.fits
9140 gt! oh231.log gt sharcsmooth oh231.fits
oh231_smooth.fits

• The sharcsmooth program does a PSF fit to each
  pixel, so to calibrate, I load the image in GAIA
  (or DS9) and determine the brightness of the
  peak. In this case it is 3.378 units.
• The true flux of OH231.8 is 19.41.9 Jy (10
  calibration uncertainty).
• Hence the scale factor is 19.4/3.4 5.7
• Now we need to scale our image
• gt sharccal -c 5.7 -u Jy mrk331_smooth.fits
  mrk_cal.fits
• Finally I open up GAIA, and look at
  mrk_cal.fits. By looking at the Signal and RMS
  maps, I see that the brightest pixel is
• 1.80 0.02 Jy
• In this case the calibration uncertainty
  dominates the error budget, so I simply quote 1.8
  0.2 Jy.

SHARC-II DRW 11/08/2004
41
Map Tweaks
Making the maps is the first step, and you may
need/want to perform some of the tweaks presented
on the following slides.
SHARC-II DRW 11/08/2004
42
Tweaks-Pointing Correction

• Two options for pointing correction
• Apply knowledge of improved telescope pointing
  model at time of running CRUSH
• crush -FAZO-120.0 -FZAO40.0
• See MEMO SHARC II Pointing at Nasmyth Focus
  Using CRUSH (Dowell, Nov. 2004) on web page/Data
  Analysis.
• Align images after CRUSH
• Use header_update utility (on web page under
  Software utilities)
• header_update image.fits RAP 1.0
• header_update image.fits DECP -1.0
• Doesnt change WCS of image however, pointing
  corrections will be applied in coadd

Attila says Use jiggle
SHARC-II DRW 11/08/2004
43
Tweaks-Calibration

• CRUSH defaults to producing FITS images in nV,
  corrected for detector nonlinearity and
  atmospheric absorption.
• Based on results for calibrator (reduced/analyzed
  the same way), one can re-scale image before or
  after coadd
• sharccal -c 5.0 uncal.fits cal.fits
• sharccal is on web page / Software utilities
• One can also change units of image
• sharccal -u Jy/beam input.fits output.fits
• Just updates BUNIT keyword.
• CRUSH recognizes V, nV, Jy/beam, Jy/arcsec2,
  and Jy/sr
• CRUSH imagetool will do these operations in
  future.

SHARC-II DRW 11/08/2004
44
Tweaks-Coadding

• Use CRUSHs coadd routine
• pushd . cd crushdir
• coadd \
• ../data/SGRASTAR.16938to16947.cal.fits \
• ../data/SGRASTAR.16948to16957.cal.fits \
• ../data/SGRASTAR.16960to16969.cal.fits
• popd mv crushdir/../data/SGRASTAR.coadded.fits .
• By default, images are weighted, but zero
  levels are not adjusted. (This is likely to
  change in future.)

SHARC-II DRW 11/08/2004
45
Tweaks-Cropping

• Use imagetool to cut noisy edges off map
• Example
• pushd . cd crushdir
• imagetool -minexp0.25 ../data/SGRASTAR.coadded.fi
  ts
• popd
• Modifies image rather than making a new copy,
  unless -out option is used.
• imagetool is part of CRUSH.
• Note other options to imagetool (e.g., -clip).

SHARC-II DRW 11/08/2004
46
Tweaks-Mosaicing

• To mosaic many maps with signal in them (e.g.,
  bright Galactic clouds), I find that adjusting
  the zero levels before coadd improves the
  appearance of the image.
• I find the mode of the image intensity
  distribution and subtract it off
• The mode can be found crudely with ds9.
• sharccal -c 5.0 -o 0.1 uncal.fits cal.fits

SHARC-II DRW 11/08/2004
47
Publishable Images
Everyone has their own way of turning maps into
publication style images. I use IDL and the
astrolib library (http//idlastro.gsfc.nasa.gov/ho
mepage.html) IDL is not free, but it is very
versatile.
read in the highest resolution image
first img_vlareadfits('../J1428p3526.fits',hdr_vl
a) CX281 center of source in X
dimension CY282 center of source in Y
dimension HW100 half width of box to
extract HEXTRACT,img_vla,hdr_vla,CX-HW,CXHW,CY-H
W,CYHW img_350readfits('../../059/set_16301_smo
oth.fits',hdr_350) read in sharc
image hastrom,img_350,hdr_350,img_350p,hdr_350p,hd
r_vla,MISSING0 make image match size/shape of
VLA loadct,0
black and white color
table gamma_ct,1.0,/CURRENT
normal gamma stretch img_350pby
tscl(img_350p,min-0.05,max0.07)
choose plotting range set_plot, 'ps' device ,
filename'sharc.eps', /encap, xsize10., ysize
10., yoffset 1., BITS_PER_PIXEL8,
COLOR1 imcontour, img_350p, hdr_350p, levels0,
xtitle' ', ytitle' ', plot the axis
charsize1.2,charthick3,/nodata,subtitle'
',COLOR0, XSTYLE4,YSTYLE4 tvimage,i
mg_350p,/overplot,/keep_aspect display the
image overlay radio contours imcontour,
img_vla, hdr_vla, levels2e-4,3e-4,4e-4,xtitle'
', ytitle' ', charsize1.2,charthick3,/noera
se,subtitle' ',C_THICK3,C_COLOR0,COLOR0,
XTHICK2,YTHICK2 XYOUTS,10,10,'350 VLA
contours',charsize1.5,charthick3 DEVICE, /close
SHARC-II DRW 11/08/2004
48
Publishable Images
IDLgt _at_sharc.idl READFITS Now reading 561 by
561 by 1 by 1 array HEXTRACT Now extracting a
201 by 201 subarray READFITS Now reading 201
by 201 array HPRECESS Header astrometry has
been precessed to 2000.0000 LOADCT Loading
table B-W LINEAR

• Other packages in use at Caltech
• Graphic
• GAIA (A free starlink package)

• More complicated example
• Multiple contour sources
• Astronomical coordinates
• Object labelling

SHARC-II DRW 11/08/2004
49
Chopped vs. Unchopped?

• In principle a 2-beam chopping observation
  increases the noise by sqrt(2) because of
  spending half the time on source. But this can
  be recovered by folding back in the flux from the
  off beam, as long as it lands on the array.
• In general, we have had good success chopping
  and reducing the data with sharcsolve. Attila
  only recently added chopped data support in
  CRUSH, though it seems to work. Once tested more
  rigorously, we will likely phase out sharcsolve
  completely.
• We strongly encourage people who want to chop to
  discuss it with one of us.

Chopping is best suited for observations of point
sources when the atmosphere shows signs of strong
variability. We have not yet shown that chopping
offers a substantial improvement.
SHARC-II DRW 11/08/2004
50
Comparison of Chopped Data
Stars denote sources detected by SCUBA, and 2 are
well detected by SHARC-II. In this case, CRUSH
and SHARCSOLVE both do a good job recovering the
same map.
51
Lessons learned (1)

• Poor choices for focusing telescope
• Many protostellar/UCHII things, especially
  NGC2071. Even IRAS16293-2422 is slightly
  elliptical.
• Saturn, Jupiter, Venus
• Better choices Mars (usually), Uranus, Neptune,
  Callisto (usually), Ganymede (often), Ceres,
  Vesta (sometimes), Pallas (sometimes), CRL618,
  CRL2688, IRC10216, OH231.8, ARP220
• Poor choices for pointing NGC 2071 (Use CRL618,
  HLTAU, or OH231.8 instead)
• Poor choices for flux calibration NGC 2071,
  blazars
• Moons Callisto or Ganymede is usually
  observable. Titan is hard to observe cleanly
  dont bother with it.

SHARC-II DRW 11/08/2004
52
Dont Focus/Point/Calibrate With It!
Avoid NGC 2071
SHARC-II DRW 11/08/2004
53
Lessons learned (2)

• For best looking large maps, map full area in as
  short a time as possible. Mosaics of fields
  under different conditions and scan patterns tend
  to have obvious stitches.
• The following projects have proven difficult
  embark on them at your own risk
• High dynamic range
• CSO beam at 350 microns.
• Negative artifacts surrounding bright sources.
• Faint, widely extended emission, due to ripples
  in reduced image.
• Be careful integrating total emission use
  exactly the same procedure for source and
  calibrator, and use a sky aperture.

SHARC-II DRW 11/08/2004
54
Ripples and Mosaic Stitches in Box Scan of Source
with Low Surface Brightness
12'

• crush -deep gets rid of ripples, but also some
  extended structure.

SHARC-II DRW 11/08/2004
55
M51
Problem Large scale low-level baseline
56
Bad Pixel flagging

• Bad pixels that are not properly flagged by
  CRUSH end up tracing the scan pattern on the
  final map. Sometimes this is hard to see since
  by default CRUSH smooths the output.
• I recommend reducing each scan separately with
  -convolve-1 in order to see if bad pixels are
  corrupting any of the data. If so, remove them
  from the list of scans you use to make the
  combined map. CRUSH 1.35 is meant to have a
  better algorithm for flagging bad pixels.

Right Map of MonR2 with a bad pixel clearly
influencing the output. (map courtesy of D.
Benford)
57
Negative Halo Surrounding Bright Sources

• Negative intensity is 3 of peak intensity.

SHARC-II DRW 11/08/2004
58
Tips for taking better data (1)

• DSOS
• Use it!
• Beam shape at 350 mm becomes nearly independent
  of ZA.
• Check that you are using correct procedure
• Wait until telescope near outside ambient
  temperature, then init. (Assumes that telescope
  at that time has the same shape as during the
  holography.)
• Make sure you get into agent mode before
  observing. Easy to check with dsosm monitor
  screen.
• When significantly changing ZA, wait for new dish
  setting to activate and settle (again with
  dsosm). Otherwise, your calibrator may have the
  wrong PSF.
• Turn off DSOS at the end of the night.

SHARC-II DRW 11/08/2004
59
Tips for taking better data (2)

• UIP source catalogs
• Enter before you get to 14,000 ft.
• Check the coordinates and equinox with VERIFY.
• Please use the existing CAL_ sources for
  extrasolar calibrators. This makes it easier for
  the staff to identify pointing/calibration scans
  to study, e.g., pointing model. The CAL_
  sources are now loaded automatically from
  USERSHARCSHARCCAL.CAT upon INST SHARC2.
• Get started quickly with the Cheatsheet on the
  web page.
• Focus and pointing tend to drift early in the
  night. Check every 45 minutes until 8 PM.
  After that, focus is usually stable (check every
  2 hours). Still should check pointing every 1
  hour, especially if observing faint sources.

SHARC-II DRW 11/08/2004
60
Tips for taking better data (3)

• SHARC Tau Use Bolometer Power Mean on
  Bolometer Means window to convert to
  t225airmass. This is useful in cases that skies
  may be clearing rapidly early in night.
• Please report problems, even minor ones, to
  sharc_at_submm.caltech.edu

SHARC-II DRW 11/08/2004
61
Miscellaneous Notes
In the next few slides, we present some
information that is related to the workshop and
SHARC-II in general, but was tangential to the
material already presented.
SHARC-II DRW 11/08/2004
62
Dependence of Gain on Atmospheric Loading
63
Instrument status

• Nasmyth move was successful, and well stay there
  for foreseeable future.
• In early September, a glue joint in the amplifier
  for rows 9-12 separated, breaking 300
  micro-wires in the process.
• The glue joint has been repaired and reinforced.
• Were now working on the micro-wires, with
  expectation of returning to normal by the January
  run.
• September run was carried out in its entirety,
  with 2/3 mapping speed.

64
Sept 2004 Pointing T-Terms

• The move to Nasmyth forced a revisit of how we
  calculate pointing corrections. See the
  documentation on the SHARC-II web page for a full
  report.
• Top as observed, s(FAZO) 3.7?
• Bottom after model applied, s(FAZO) 2.1?
• Model fit and method of correcting data web
  page memo
• Use CRUSH exclusively to apply the model if not
  pointed on center of array (16.5, 6.5).

SHARC-II DRW 11/08/2004
65
Temperature Effects on Pointing (Hiroko)
SHARC-II DRW 11/08/2004
66
CSO/SHARC II Sidelobes (350 mm)
160?

• Sidelobe spots as bright as 5 of peak intensity.

SHARC-II DRW 11/08/2004
67
BOXSCAN

• There are a number of constraints on the scan
  pattern
• It turns out that by starting the scan from the
  center of the map, the map size cannot be
  perfectly square, otherwise the scan will hit a
  corner, and the algorithm will fail.
• There is also a choice in how far apart the
  intersection points are.  You want them to be
  less than the size of the SHARC array (to ensure
  the whole map gets sampled), but having them too
  close together means it will take longer to
  complete a scan.  
• The scan rate should be fast enough to close out
  the scan in a reasonable time, but not so fast
  as to smear the beam.  Also, faster scans mean
  it is harder for the telescope to handle rapid
  change in direction at the boundaries.  
• Want a rectangular pattern instead of a box?
• This is straightforward, but ask me for details.

SHARC-II DRW 11/08/2004
68
BOXSCAN
   

• To help plan your scan, download, and compile the
  program boxscan.c.
• Here is an example of it in action I want to
  scan a 10x8 arcminute region, relatively finely.
• I start with a 10 arcsec spacing, and a 40
  arcsec/sec scan rate
• gt boxscan 600 480 10 40
• BOX_SCAN  600.000  471.429 40.0 45       23.33
  minutes
• WARNING scantime exceeds 15 minutes
• Hmmm...that won't work.  Scan length is too long.
   I'll try a larger spacing (20.0 arcsec)
• gt boxscan 600 480 20 40
• BOX_SCAN  600.000  457.143 40.0 45       11.31
  minutes
• That's better, but now I wonder if 40arcs/sec is
  too fast.  I'll try an even larger spacing and a
  slower scan speed
• gt boxscan 600 480 25 30
• BOX_SCAN  600.000  494.118 30.0 45       13.20
  minutes
• Enter the scan command dirently into the UIP, and
  set the integration time as suggested (perhaps
  adding on 0.2 minutes for overhead). Wondering
  what the "45" is in the "BOX_SCAN" output?  It is
  the angle at which the scan starts. This should
  always be 45.  Consult with us if you want to try
  other things.

SHARC-II DRW 11/08/2004
69
SHARC-II/CRUSH Simulations
Tom Tyranowski (visiting student) wrote a data
simulator, which we use to test the performance
of CRUSH. If you are interested in using it,
contact us.
100 mJy Ring surrounding compact star in 1
hour 10' x 10' Billiard Ball Scan. Imperfect
cleaning of faint large scale structures
Source Fluxes Recovered within 1
100 mJy Compact Lissajous Sweep in 1
hour Imperfect cleaning of faint large scale
structures
SHARC-II DRW 11/08/2004
70
To-Do list

• Get rid of negatives around bright sources.
• Find mode of an image
• Calibration cheatsheet
• example graphic link/script on web page
• Improved CRUSH default -chopped reduction
• Memory problem in 1.34? (z6.5 source)
• Bolometer scrambling test
• CRUSH email exploder
• Resume simulations release simulation code.
• sharcsolve (for Johannes) independent tau for
  each scan
• Hoping for better pixel flagging in v. 1.35
• Update CAL catalog in UIP.