Learning Center
Plans & pricing Sign in
Sign Out

PowerPoint Presentation - SHARC II Data Reduction Workshop


									                                    u c
                                   Q i kTi m e™ an d a
                            YU V42 0 cod ec de com p r sso r
                           ar e ne ede d t s ee t hi s pi ct ur e.

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

SHARC-II DRW 11/08/2004
      GOALS
      Caltech success stories
      Software Requirements and Installation
           Overview of available software
           Installation overview
           Scripting with CRUSH
      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
      Lessons learned
      Tips for taking better data
      Miscellaneous notes

SHARC-II DRW 11/08/2004

         1. Transfer expertise from Caltech to other users
                 Why? To promote publication of SHARC-II data

         2. 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
              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-
      II’s abilities.

SHARC-II DRW 11/08/2004
                    SLUGS (z<0.05)
                                   •     Dunne et al. have
                                         characterized the SED of 106
                                         IRAS selected galaxies at
                                   •     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 @ 225~0.06
                               •       1.0h @ 225~0.08
                               •       0.5 Jy< S350 < 3 Jy
                               •       crush -faint -compact

SHARC-II DRW 11/08/2004
               Spitzer HLIRG (z~1.5)
                          •       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 @ 225~0.06
                              •     S350 = 226 ± 45 mJy
                              •     CHOPPED observing
                              •     sharcsolve reduction

SHARC-II DRW 11/08/2004
           Stanford sample (0.1<z<1.0)

                                   2h @ 225~0.05
                                   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 ULIRG’s 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
                Fomalhaut Debris Disk
                                  •   3.0h @ 225~0.039
                                  •   SWEEP (Lissajous)
                                  •   crush -deep reduction
                                  •   Peak fluxes: 150
                                  •   Integrated flux: 1.2 Jy
                                  •   Consistent with thin,
                                      uniform dust ring

       K. Marsh et al. (2004-5)

SHARC-II DRW 11/08/2004
             Houde et al. (2004)                     Johnstone & Bally (1999)

          rms: 0.3
          Jy/beam                                                      rms: 0.3

                                 rms: 1

                                rms: 0.3
          4 hr.                 Jy/beam
                                                   18 hr.
         1.2 mm
          PWV                                      1.0 mm
                  mosaic of BOX scans

SHARC-II DRW 11/08/2004
      Low-z Interacting Galaxy Survey
 •   Selected by IRAS 100 mm flux
     and proximity using criterion of
     Surace (2004).
 •   Perhaps analogues to z>1
 •   Lissajous scans in typically
     t225×airmass = 0.05-0.10.
 •   Observations of 14/42 sources
     complete, to be reported by J.
     Bird and D. Dowell.

                  5×1010 Lsolar

SHARC-II DRW 11/08/2004
              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 m sources marked with

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

                          •   3-4h @ 225<0.06
                          •   350 = 5 mJy

                   LE31                          LE2


SHARC-II DRW 11/08/2004
                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)

   Minimum requirements to run all the software are:
   Java 1.4.1 (for CRUSH)
   gcc version 3.2.3
   cfitsio libraries (

   Other nice software include:
   IDL or graphic (for plotting)
   DS9, GAIA (starlink) for displaying fits files.

SHARC-II DRW 11/08/2004
     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 T|F                  offset applied?
   OFFSET value                  offset value (only if OFFAP=T)
   BUNIT unit                Units of output image

   sharcgap : tests a raw sharc2 data file for timing gaps
   sharcgap startidx stopidx
   Needs to be run from the directory in which data is stored.
   It checks consecutive data points to see if they are spaced by more than 1% of the expected time.
   Expected time is 36ms, but is calculated explicitly from the first HDU.

   sharcsmooth : Performs a PSF fit to a reduced SHARC2 map.
   This will be described later in the calibration section.

SHARC-II DRW 11/08/2004
     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 Darren’s)

   sharcstat : computes basic statistics on a reduced SHARC2 map
   sharcstat file.fits
   > 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
   > sharctau -v /home/bigdisk1/sharc2-012900.fits /scr/borys/sharc/
   # 30/09/2003 12:06 0.50 0.051 0.068
   > sharctau /home/bigdisk1/sharc2-012900.fits /scr/borys/sharc/
   Note: tau225 is read from file. Fittau is at the frequency appropriate to the input taufile.

SHARC-II DRW 11/08/2004
       • 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 (7:45am) [/scr/borys/sharc/attila] >ls -la
       lrwxrwxrwx 1 borys cittgp            10 Nov 5 08:21 crush -> crush-1.34
       drwxr-xr-x 3 borys cittgp         4096 Oct 3 07:16 crush-1.33
       drwxr-xr-x 3 borys cittgp         4096 Sep 22 09:39 crush-1.33b2
       drwxr-xr-x 3 borys cittgp         4096 Oct 6 15:48 crush-1.34
       drwxr-xr-x 2 borys cittgp         4096 Aug 31 23:13 data
       drwxr-xr-x 3 borys cittgp         4096 Oct 28 00:35 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 (7:48am) [/scr/borys/sharc/code/bin] >ls -la
     -rwxr-xr-x 1 borys cittgp        6793 Aug 27 14:31 boxscan
     -rwxr-xr-x 1 borys cittgp        16548 Aug 27 14:10 sharccal
     -rwxr-xr-x 1 borys cittgp 784529 Mar 26 2003 sharcextract
     -rwxr-xr-x 1 borys cittgp 766548 Aug 27 14:31 sharcgap
     -rwxr-xr-x 1 borys cittgp 770786 Aug 27 14:31 sharclog
     -rwxr-xr-x 1 borys cittgp 770704 Aug 27 14:32 sharcsmooth
     -rwxr-xr-x 1 borys cittgp        16164 Aug 27 14:13 sharcstat
     -rwxr-xr-x 1 borys cittgp 766589 Aug 27 14:32 sharctau

SHARC-II DRW 11/08/2004
  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.SCAN2…SCANN.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

      cd /scr/borys/sharc/attila/crush
      echo “PROCESSING ic5634”
      ./crush -faint -compact -name=ic5634_1.fits 14176 14177 >! ic5634_1.log
      ./crush -faint -compact -name=ic5634_2.fits 14182-14185 >! ic5634_2.log
      ./coadd -out=ic5634.fits ic5634_1.fits ic5634_2.fits >! 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

     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
      Important Differences between
 • 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
                 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
                       SHARC-II Calibrators
   • Availability of calibration sources has always been a problem in sub-mm observations,
   particularly at shorter wavelengths (can’t 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:
   • We have used these relations to extrapolate the fluxes for all days between 2002 and
   • These calibrations are available for download for the SHARC-II web page

SHARC-II DRW 11/08/2004
SHARC-II DRW 11/08/2004
                          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
   •   Images of these fits are located on the SHARCII
       website (
   •   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
                              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

SHARC-II DRW 11/08/2004
       Keep an eye on the 225GHz and 350 micron fits…they CAN differ

                     Fits from both tippers on the same night.

SHARC-II DRW 11/08/2004
                        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.
   • CRUSH’s 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
  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.

                                QuickTime™ and a
                            TIFF (LZW) decompressor
                         are needed to see this picture.

  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.

                                QuickTime™ and a
                            TIFF (LZW) decompressor
                         are needed to see this picture.

                                                           MaiTau helps!
            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”
   • “Known” fluxes are obtained from HORIZONS.
   • CRUSH’s 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
                          Plot shows conversion
                          factor of calibrators
                          taken during the
                          2004 run.

                                  Calibration Stability

                          Conversion factor is
                          consistent to within:
                          21.3% for all
                          18.8% for Neptune
                          19.2% for Uranus

SHARC-II DRW 11/08/2004
                          Calibration is
                          consistent over a
                          wide range of

                                Calibration Stability

                          You do not need to
                          take calibration scans
                          at the same elevation
                          as your science.

SHARC-II DRW 11/08/2004
            Calibration/Tau summary

   • Tau fits: Important for understanding what is
     happening to the atmosphere during
   • 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
                                 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:

                                          A  PSF(dx, dy)  S(i  dx, j  dy)
                             W     W                                              2

                     2                         N(i  dx, j  dy)2
                            dxW dyW

  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 PSF’s 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
                     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
                       Calibration : Aperture
                                                                                          N          2
      N    = Number of pixels in map under consideration. (not                          1
      necessarily the full mapÉone may exclude edges for example)        RMS 
                                                                                        N i1
                                                                                           mi  m   
      NA   = Number of pixels within the aperture
      NS   = Number of pixels within the sky annulus                          1 S
      mi   = flux in a given map pixel, i                                S        mi
                                                                              N S i1
      F    = estimated aperture flux                                          NA
      F   = associated noise estimate                                   F   mi  S 
      S    = estimated sky background per pixel                               i1
      I    = pixel index which runs over the relevant pixels in                                          1/2
      the map                                                                              N A 
                                                                         F  RMS  N A  
                                                                                           1    
                                                                                           N S 

  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
  (you’ve 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

SHARC-II DRW 11/08/2004
                            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

                                     EXTENDED SOURCES
  PSF, since it essentially gives you Flux/beam.

  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
                     Calibrating your data
  The principles involved with calibrating SHARC-II data are applicable to all data from other
  sub-mm telescopes.

  • Good estimates of the atmospheric opacity for all science and calibration observations
  • A decent collection of calibrators (different objects, airmasses, etc.)
  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 04:54
  • PSF photometry to be used

SHARC-II DRW 11/08/2004
                     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 04:30 (~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
                             Calibration Example 2
  > ./crush -faint -compact -convolve=-1 -name=mrk331.fits 9125-9127 >! mrk331.log
  > sharcsmooth mrk331.fits mrk331_smooth.fits
  > cat mrk331.log
  crush -- Comprehensive Reduction Utility for SHARC2
       Author: Attila Kovacs <>
       Version: 1.34-1

  Scan 1: Reading /home/bigdisk1/sharc2/sharc2-009125.fits...
   Got Mai-Tau! tau(350um) = 1.3088
   83 HDUs, 16439 x 36ms frames -> 9.9 minutes total.
   Filtering 13.4Hz on noisy pixels.
   DownSampling -> 5479 frames

   [MRK331] observed at 2003-01-15T04:54:01.949
    RA = 23:48:54.0         DEC = 20:18:29.0 (1950.0)
      = 23:51:26.7           = 20:35:10.1 (2000.0)
    AZ = 277:44:54.5         EL = 57:01:57.9
    RAO = 0.0 DECO= 0.0         AZO = 0.0     ZAO =-0.0
    FAZO=-104.0 FZAO=-30.0        Rotator = 60.0 RotZero = 60.0
    Pointing Center = 16.5,6.5 Rotation Center = 18.5,8.6
    Parallactic Angle = 85.0 tau225 = 0.053 tau = 1.835
    Plate Scale = 4.93"x4.77"

SHARC-II DRW 11/08/2004
                        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.

  > ./crush -compact -convolve=-1 -name=oh231.fits 9140 >! oh231.log
  > 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 19.4±1.9 Jy (10% calibration uncertainty).
   •Hence the scale factor is: 19.4/3.4 = 5.7
   •Now we need to scale our image:

   > 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
                          Map Tweaks

          Making the maps is the first step, and
         you may need/want to perform some of
         the tweaks presented on the following

SHARC-II DRW 11/08/2004
           Tweaks-Pointing Correction

   • Two options for pointing correction:
       – Apply knowledge of improved telescope pointing model at time
         of running CRUSH:
            • crush -FAZO=-120.0 -FZAO=40.0 …
            • 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
            •   Doesn’t change WCS of image; however, pointing corrections will
                be applied in coadd

    *Attila says: Use jiggle
SHARC-II DRW 11/08/2004

   • CRUSH defaults to producing FITS images in nV,
     corrected for detector nonlinearity and atmospheric
   • Based on results for calibrator (reduced/analyzed the
     same way), one can re-scale image before or after
       – 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/arcsec**2, and Jy/sr
   • CRUSH imagetool will do these operations in future.

SHARC-II DRW 11/08/2004

   • Use CRUSH’s coadd routine:
       pushd .; cd crushdir
       coadd \
        ../data/ \
        ../data/ \
       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

   • Use imagetool to cut noisy edges off map:
       – Example:
           pushd .; cd crushdir
           imagetool -minexp=0.25 ../data/SGRASTAR.coadded.fits
       – Modifies image rather than making a new copy, unless -out option is
       – imagetool is part of CRUSH.
       – Note other options to imagetool (e.g., -clip).

SHARC-II DRW 11/08/2004
   • 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
                                 Publishable Images
    Everyone has their own way of turning maps into publication style images.
    I use IDL and the astrolib library (
    IDL is not free, but it is very versatile.

     ; read in the highest resolution image first
     CX=281 ; center of source in X dimension
     CY=282 ; center of source in Y dimension
     HW=100 ; half width of box to extract


     img_350=readfits('../../059/set_16301_smooth.fits',hdr_350) ; read in sharc image
     hastrom,img_350,hdr_350,img_350p,hdr_350p,hdr_vla,MISSING=0 ; make image match size/shape of VLA

     loadct,0                                 ; black and white color table
     gamma_ct,1.0,/CURRENT                                 ; normal gamma stretch
     img_350p=bytscl(img_350p,min=-0.05,max=0.07)                    ; choose plotting range
     set_plot, 'ps'
     device , filename='sharc.eps', /encap, xsize=10., ysize = 10., $
            yoffset = 1., BITS_PER_PIXEL=8, COLOR=1
     imcontour, img_350p, hdr_350p, levels=0, xtitle=' ', ytitle=' ',$ ; plot the axis
             charsize=1.2,charthick=3,/nodata,subtitle=' ',COLOR=0,$
     tvimage,img_350p,/overplot,/keep_aspect; display the image
       ; overlay radio contours
     imcontour, img_vla, hdr_vla, levels=[2e-4,3e-4,4e-4],xtitle=' ', ytitle=' ',$
       charsize=1.2,charthick=3,/noerase,subtitle=' ',C_THICK=3,C_COLOR=0,COLOR=0, $

     XYOUTS,10,10,'350 + VLA contours',charsize=1.5,charthick=3
     DEVICE, /close

SHARC-II DRW 11/08/2004
                  Publishable Images
                                 IDL> @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
                                 % 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
                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
         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.
                  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; don’t bother with it.

SHARC-II DRW 11/08/2004
                    Avoid NGC 2071

    Don’t Focus/Point/Calibrate With It!
SHARC-II DRW 11/08/2004
                 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
    Ripples and Mosaic Stitches in Box Scan of
       Source with Low Surface Brightness

   • crush -deep gets rid of ripples, but also some
     extended structure.
SHARC-II DRW 11/08/2004

Large scale low-level baseline
                       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.

                                                             QuickTime™ and a
                                                         TIFF (LZW) decompressor
                                                      are needed to see this picture.

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

   • Negative intensity is ~3% of peak intensity.
SHARC-II DRW 11/08/2004
        Tips for taking better data (1)
 • DSOS:
    – Use it!
        • Beam shape at 350 m becomes
          nearly independent of ZA.
    – Check that you are using correct
        • 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
        • When significantly changing ZA,
          wait for new dish setting to
          activate and settle (again with
          dsosm). Otherwise, your
          calibrator may have the wrong
        • Turn off DSOS at the end of the

SHARC-II DRW 11/08/2004
        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
   • Get started quickly with the “Cheatsheet” on the web
   • 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

SHARC-II DRW 11/08/2004
        Tips for taking better data (3)
 • “SHARC Tau”: Use Bolometer Power Mean on Bolometer Means
   window to convert to 225×airmass. This is useful in cases that skies may
   be clearing rapidly early in night.
 • Please report problems, even minor ones, to

SHARC-II DRW 11/08/2004
                 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

SHARC-II DRW 11/08/2004
Dependence of Gain on
 Atmospheric Loading
                   Instrument status
• Nasmyth move was
  successful, and we’ll
  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.
   – We’re now working on the
     micro-wires, with
     expectation of returning to
     normal by the January
   – September run was
     carried out in its entirety,
     with ~2/3 mapping speed.
          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
                          • Top: as observed, (FAZO) = 3.7″
                          • Bottom: after model applied, (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
      Temperature Effects on Pointing

SHARC-II DRW 11/08/2004
   CSO/SHARC II Sidelobes (350 m)


   • Sidelobe spots as bright as 5% of peak intensity.

SHARC-II DRW 11/08/2004

   There are a number of constraints on the scan pattern:
   1. 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.
   2. 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.
   3. 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
   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:
   > 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)
   > 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:
   > 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
        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

SHARC-II DRW 11/08/2004
                         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? (z=6.5 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.

To top