Welcome to the second edition of the CGPS newsletter. The first contained many articles outlining the plans of the various groups and teams. Now read on to find out what we've all been doing during our summer vacation! Remember also that whenever you have something exciting to share with the rest of us you need not wait until the next newsletter, simply email us all at firstname.lastname@example.org.
Just a few technical comments about the set-up of this newsletter. As with the June issue, the author names of each article have been linked to their email addresses, making it conveniently easy to send them relevant queries or comments. As well, in Lloyd Higgs' article about the final mosaics, various internal links have been inserted to make navigating to-and-from or around the figures easier. I am not a web wizard so if you have any suggestions about set-up, please pass them on.
As of yet, no progress has been made on the issue of bright sources away from the phase centre (A1/1A and O4). We are going to get this ball rolling soon, since we need to have a strategy in place before we embark into the E, F or G fields where Cass A will be bright!
The synthesis beast chugs along at an impressive rate. The survey area is now 44% observed, in 48% of the allotted time.
There are the occasional bolts of cosmic radiation that seem to embed gotcha's in the processing requiring me to assume my wizard's garb and apply a modicum of magic. These acts only occur when smugness creeps into the system. I have a suite of unfriendly programs which I work with in these rare circumstances.
I enjoy being the first person to see some of the amazing output from this ambitious project. I often drag unsuspecting astronomers in to view these maps, they are only allowed to escape when the appropriate 'Gee Whiz' is forthcoming.
A preliminary (non-short-spacings added, rough flux registration, no position registration) mosaic has been completed for 14 fields centered around l=142°. This mosaic is the first to be completed using the correct "doppler skew" correction and took three weeks worth of computer time to generate. The resulting fits file is 1.3 Gb (yes, that's GIGA) in size. Processing of 9 additional fields on either side of this mosaic are either completed or well in progress. A 23 field mosaic, covering fields "U7" -> "W9" is envisioned.
Additionally, reduction of the "O4" field, which contains Cygnus-A within the field of the telescopes, is ongoing. The extreme brightness of Cygnus-A makes reduction of this field difficult, and will require extra work to make the quality of the resulting data comparable to that of the rest of the survey. It is concievable that it will require application of self-calibration models derived from the continuum data, which has not been attempted for any other field. These problems have delayed re-reduction of the fields at the Cygnus end of the mosaic.
The IGA was developed under an Archival Data Processing (ADP) grant from NASA with support from Infrared Processing and Analysis Center (IPAC). The basic HIRES algorithm used to develop the Atlas is described in Aumann, H.H., Fowler, J.W., and Melnyk, M., 1990 ( A.J.,99, 1674). The modifications needed to run HIRES on a parallel super-computer are described in Y. Cao, T.A. Prince, S. Terebey, and C.A. Beichman, 1996 (PASP, 108, 535). The characteristics of the Atlas are given in Y. Cao, S. Terebey, T.A. Prince, C.A, Beichman, 1997 (ApJ.Supp. 111, 387).
The IGA incorporates several important improvements from standard HIRES processing at IPAC. Foremost is improved destriping and zodiacal emission subtraction, which lead to reduced artifacts, better ability to discern low-surface-brightness features and the ability to mosaic images without edge discontinuities. The IGA is well suited to high-resolution studies of extended structure, and will be valuable for a wide range of scientific studies, including: the structure and dynamics of the interstellar medium (ISM); cloud core surveys within giant molecular clouds; detailed studies of HII regions and star-forming regions; determination of initial mass functions (IMFs) of massive stars; and study of supernova remnants (SNRs). The IGA will be especially useful for multi-wavelength studies using the many Galactic plane surveys that have similar (1') resolution.
The IGA images consist of maps made with 1 or 20 iterations of the HIRES algorithm, corresponding to either no non-linear processing (1 iteration), or to the maximum amount (20 iterations) of non-linear processing deemed to be reasonably free of artifacts. We emphasize that the spatial resolution within the maps varies with the details of the scan coverage for a particular area of the sky. The beam maps provided with each field are essential for assessing the angular resolution at various positions in the maps. For more discussion of the quality of the maps, the nature and number of artifacts, please see the two articles by Cao et al.
The directory structure on tapes 1 through 9 is as follows:
The file naming convention is as follows:
In the TYPE/pPLATE_bBAND directory:
Software and documentation are located on Tape 0 in appropriately named directories.
The software catalog contains the besthires program that will return the best field name given input coordinates. The besthires program located in the software/bin directory was compiled on a Sun IPX running Solaris. If this binary file does not work on your machine, you must recompile the C programs in the software/besthires directory via the following procedure:
[Editor's note: The above is the README file which accompanies the IGA distribution. The only modifications brought to it for the purpose of this newsletter were to add in links to the IPAC homepage as well as to the two postscript files. Both Peter Martin at CITA and Russ Taylor at UofC have copies of the tapes described. The IGA, once regridded and reprojected to the CGPS grid, will be made available to consortium members through the UofC lab. Individual regions can also be obtained from the IGA homepage and the whole set can be obtained from the NSSDC by contacting David Lesiawitz.]
The final CGPS data products will be issued to the world in some mosaic form such that the data can be fairly readily visualized and analysed on a work station. Clearly, the data must be broken down into reasonable ``chunks'' that still give good representation of large structures. Moreover, the adopted pixel sizes should well represent detailed structure, such as radio point sources. In order that multi-wavelength comparative astronomy can easily be done, all data products will be issued on the same grids, although there could well be two representations for the lower-resolution data (greater than 3 arc-minutes), one on the basic CGPS grids (where the data would be greatly over-sampled) and one on a coarser grid.
In June 1997 the CGPS Management Committee asked me to look into a scheme of mosaics which would meet these criteria. We will soon be ready to produce semi-final data products and it would be useful to know in detail the mosaic specifications. I produced such a scheme, based on two sets of mosaics -- one of 5° x 5° areas and one of 15° x 15° areas. If one wishes to see this proposal in detail, it can be obtained by anonymous ftp from DRAO (drao.nrc.ca) in the directory pub/lah/gps. The file is ``mosaic_proposal.ps'' and it is available in both ``gzipped'' (0.28 Mb) and straight PostScript form (1.71 Mb).
The Management Committee approved this proposal in July, with some slight modifications. I have prepared several memoranda giving detailed specifications of the mosaics and how they relate to the fields observed with the DRAO Synthesis Telescope, so I will only give a brief overview in this news note.
The basic set of mosaics consists of 36 partially overlapping areas,
each about 5° x 5° in extent and defined in Galactic coordinates, as
shown in Figure 1. They
are arranged in pairs whose centres are separated by 4° in Galactic
latitude, and the 18 sets of pairs are separated by 4° in Galactic
longitude. The areas overlap by about 1° in each coordinate. These
mosaics are given codes such as ``MA1'' where the ``M'' denotes ``mosaic'',
the ``A'' denotes that the DRAO A fields are the major contributor to the
mosaic, and the ``1'' designates the lower-latitude mosaic of the pair
of mosaics at that longitude. (The terminology breaks down slightly
in the region of the I and J fields -- they both contribute about equally
to the pair of mosaics termed ``MIJ1'' and ``MIJ2''.) The grids for these
mosaics are 1024 x 1024 pixels with a pixel size of 0.005°. The
relationship between one of these mosaics and the corresponding DRAO fields
is shown in Figure 2.
This shows the contribution of observed
21-cm line fields to the mosaic. The fields are defined by a cut-off radius
of 93 arc-minutes, or about the 10 percent level of the primary beam.
(skip to next paragraph)
Figure 1: The basic set of mosaics
(back to text)
Figure 2: An HI mosaic
(back to text)
For 1420-MHz continuum data, a somewhat larger cut-off radius will be
used: 100 arc-minutes, or about the 6.6 percent level. This gives a
somewhat smoother variation of the instrumental noise over the mosaic.
In Figure 3,
the corresponding field coverage is shown.
Typically, 16 -- 19 fields must be included in the mosaicing operation,
which means that ``supertile'' will have to be run in a multiple-pass
mode if it is used for the processing.
(skip to next paragraph)
Figure 3: A 1420 MHz continuum mosaic
(back to text)
For 408 MHz, where the DRAO field-of-view is so much greater, many more
fields will contribute to a given mosaic. A field cut-off at the 10 percent
level is proposed for the 408-MHz data, corresponding to a radius
of 291 arc-minutes. The 408-MHz field coverage for the
mosaic MA1 is shown in Figure 4.
In fact, 39 different
fields contribute to this mosaic.
(skip next paragraph)
Figure 4: A 408 MHz mosaic
(back to text)
Two of the CGPS datasets (the H I data and the CO data) will be ``data cubes'', and a decision was needed regarding the final number of spectral channels to be produced. The DRAO fields are being observed with a range of central velocities, but a uniform set of channels for each final mosaic is desired. One could either regrid the spectral data to give 256 channels covering the full observed range of LSR velocities with a channel separation of 0.875 km/s for example, or combine the DRAO fields using 272 spectral channels having the observed channel separation of 0.8245 km/s. The Management Committee has adopted the latter approach in order to maintain the lowest ratio of sample interval to spectral resolution for the two datasets.
In order to present the lower-resolution data more effectively, a second
set of six mosaics will be produced. These are 1024 x 1024 with a 0.015°
pixel size (about 15° x 15° in extent) and are spaced in Galactic
longitude at 12.5° intervals. These mosaics are coded ``C1''
to ``C6'', and are shown in Figure 5.
will be no confusion between the designation of these mosaics and that
for the DRAO C fields! The 408-MHz fields
that contribute to mosaic C2 are shown in
The number of fields to be processed to create
this mosaic is 68!
(skip to next paragraph)
Figure 5: Mosaics for low resolution data
(back to text)
Figure 6: A 408 MHz mosaic
(back to text)
More details of the mosaics and the contributions of observed DRAO fields to them are given in the above-mentioned memoranda, which can be obtained from the DRAO anonymous ftp area in directory pub/lah/gps. The detailed specification of the mosaics is in file ``overview_mosaics.ps'' (0.22 Mb) and its ``gzipped'' counterpart (0.04 Mb). Listings of the fields contributing to each of the mosaics are given in file ``detail_mosaics.ps'' (0.10 Mb) and its ``gzipped'' counterpart (0.03 Mb). Graphical presentations of field coverage for the H I mosaics are in ``atlas_hi_mosaics.ps'' (0.99 Mb; 0.14 Mb gzipped), for the 1420-MHz continuum mosaics in ``atlas_1420_mosaics.ps'' (1.01 Mb; 0.15 Mb gzipped), and for the 408-MHz mosaics in ``atlas_408_mosaics.ps'' (1.38 Mb; 0.24 Mb gzipped).
Now that the final mosaics have been defined, we just have to tackle the huge processing task of averaging the DRAO data onto these grids, and of regridding the other datasets onto them. This should keep us busy for the next few years!
The DRAO export package has been successfully ported to the Linux operating system running on a PC. Binaries compiled in static mode are available. Interested people should contact Tony Willis.
As part of an ongoing project to investigate the interstellar environments of filled-center (FC; a.k.a. "Crab-like" or "plerionic") SNRs the HI environment of 3C58 is being explored. The purpose of this work is to test the hypothesis that FC SNRs lack associated shells because they lie in low density regions of the ISM.
3C58 is, after the Crab Nebula itself, the most-studied FC SNR. It is thought to be the remnant of the historical SN of AD1181, thus making it roughly the same age as the Crab. It is also thought to lie at a similar distance as the Crab, but is slightly larger in both angular and physical dimensions. Unlike the Crab, 3C58 does not have a detected pulsar (although an X-ray point source within the SNR may betray the existence of a neutron star) and is only 1/30th as bright in the radio. Both the Crab and 3C58 are visible in the optical and X-ray, but 3C58 is notably absent in the IR, possibly due to a spectral turnover seen above 10 GHz. The similarities in the distance and age of 3C58 and the Crab, as well as their differences in other regards, make comparing and contrasting these two objects a natural extension of the ongoing work.
Despite being less studied than the Crab, 3C58 is in some ways a superior object for our study. Being fainter than the Crab means that fewer artefacts are created in the HI maps, making analysis easier. In addition, lying at l=131.5°, the velocity field of the HI in the direction towards 3C58 is much more spread out than for the Crab which, at l=184°, shows very little velocity spread with distance (due to the Galactic motion being primarily tangential to the line of sight).
As a result of an earlier, low resolution, study of the HI environment of 3C58, an 8° square region, centered on the SNR, was imaged in HI using the DRAO 26m telescope. This low resolution data was obtained before the 26m's spectrometer upgrade to 256 channels, and has less velocity resolution than the CGPS short-spacings will. These data have been added to the existing CGPS/DRAO data, but required that the CGPS velocity resolution be degraded to match the 26m data.
3C58 lies near the edge of the CGPS/DRAO "A1/1A" field; because of this, the signal-to-noise in the region immediately to the north-east of the SNR is much too low for analysis. As it is likely that any interaction between the SNR and its surroundings will occur in this region, this severely limits the usefulness of the present data. Further analysis will require the mosaicing of the present data with data from adjacent fields. Data from the CGPS Pilot Project lie immediately to the east of the A1/1A field, but do not overlap the region of interest.
Further analysis will thus require observations of overlapping fields. The Y9 and Y6 fields provide the necessary overlap, to the north and east respectively. These fields are not scheduled to be observed for several months, and a change in the observing strategy may push their observation until the year 2000.
There has been no substantive change since the last newsletter. See the article in the June newsletter for details on the state of the project.
There has been no substantive change since the last newsletter. See the article in the June newsletter for details on the present state of the project.
The promised JCMT observations are still "in progress". Other progress on this project awaits the appearance of this data. See the June newsletter for the present state of this project.
The region above the W4 chimney, at higher latitudes than the pilot project data, has been observed with the DRAO telescopes. The new data include three CGPS fields (X4, X9, Y4) and two non-survey fields (KM and KN, for those who like to know these things). There is the possibility of an additional field being observed, depending on the results obtained with the five listed above. The objective is to investigate whether the structure remains open at higher latitudes or if it closes off as suggested by Dennison, Topasna & Simonetti (ApJ 474, L31).
Synthesis telescope observations were carried out last winter. The 26m data for KM were collected earlier this summer, before the telescope decided to be contrary, and the KN data were recently gathered. The HI shortspacings for X4, X9 and Y4 will be observed within the ``normal'' CGPS stream within the next few months.
Processing is underway. KN C21 processing has been declared done, including the addition of shortspacing information from the Stockert survey following the time-honoured DRAO cookbook recipe (the method for adding shorspacings to the CGPS has yet to be finalized as far as I know). The KM C21 image is being more than a little finicky and will require some fancy processing footwork apparently. The rounds of selfcalibration and modcal have just begun for the 408 MHz data and will undoubtedly be complicated by the grating rings from Cas A (which is itself off the map for a 2048x2048 20" increment image) and some man-made interference, in addition to W3 and 3C58. The HI has been entirely processed for KM and KN, again using the old standard technique. Processing of X4, X9 and Y4 fall within Sean's and Brad's domains and won't be done anytime soon (recall that outside proposers are to receive the appropriate subsection of the fully processed CGPS mosaics rather than the original data); I have volunteered to process them myself but due to uniformity considerations, they would have to be reprocessed by the DRAO Ops team anyway. Stay tuned to this newsletter to find out if I end up doing my own processing anyway because I'm too eager to see the results!
In July Jo-Anne Brown (PhD student), Marta Peracaula (post-doc), and Russ Taylor (all from University of Calgary) visited DRAO to discuss the various issues related to the analysis of CGPS polarization data, as set out in the previous newsletter (briefly, ellipticity of the feeds, off-axis instrumental effects, and ionospheric Faraday rotation). A plan for dealing with these effects was formulated, with Jo-Anne and Marta to work with Andrew to address them. Once these algorithms are in place and tested then bulk processing of the CGPS polarization data to yield final images can begin.
There has also been some discussion of late among the CGPS Polarization Group of the possibility of implementing a polarimeter system to operate at 408MHz. The antennas are already able to be switched remotely between receiving right and left circular polarization, so an investigation of a time-sharing system to gather the required cross-polar correlations on all baselines with minimal changes to the existing system is being undertaken by Leonid Belostotski (Masters student, U.Alberta). Results from this work will allow us to assess the severity of spurious instrumental terms, after which further work may then be required on calibration algorithms, as the first order approximations used at 1420MHz may no longer be valid. Finally, a means of calibrating polarization angles - particularly in light of the stronger Faraday rotation at this longer wavelength - needs to be investigated, since the standard polarization calibrator at 1420MHz (3C286) cannot be used at 408MHz.
Finally, the Letter to Nature discussing the lenticular polarization feature towards the W5 HII region (Gray, Landecker, Dewdney & Taylor) received favourable comments from the referees, and the revised manuscript is currently back with Nature. The second paper discussing the polarization from the rest of the pilot survey region (Gray, Landecker, Dewdney, Taylor, Willis, Normandeau) is nearly ready for submission to Astrophysical Journal.
The data paper for the FCRAO CO Survey of the Outer Galaxy was submitted to ApJ Supplement in late June. Pending the acceptance of the paper, the data will be sent to the NASA Astrophysical Data Center and the NCSA Astronomy Digital Library. The CO data were sent to Calgary in July to facilitate early access for the GPS consortium. Presently, as released, the archive is comprised of 2 dimensional vlsr-longitude FITS images at the 606 observed latitudes.
In August, Heyer and Terebey submitted the paper, ``The Anatomy of the Perseus Spiral Arm: 12CO and IRAS Imaging Observations of the W3/4/5 Cloud Complex'' to ApJ. This paper investigates:
This summer, Heyer and Snell have decomposed the entire CO data cube into an ensemble of discrete objects. The Survey yielded over 8400 objects. Many of the identified objects are simply clumps within larger cloud complexes while others are isolated molecular features. From the measured line widths, sizes, and CO luminosities, we find the following relationships:
Mike Fich and Doug Johnstone are computing a model for the spiral structure of the Galaxy using HII regions as tracers for the velocity field. The hope is to have a global velocity-distance model at each longitude to be used by the gps teams for determining distances from the HI cubes.
At present we are computing the best fit global spiral structure and hope to have significant results soon. Keep your fingers crossed.
Shantanu Basu, Doug Johnstone, and Peter Martin are writing up a paper on the dynamics and ionization structure of the HI bubble observed by Normandeau, Taylor, and Dewdney in the CGPS Pilot project. The semianalytic model originally used by Kompaneets to study blast waves in an exponential atmosphere is adapted to model a wind-blown bubble in the stratified interstellar medium. The paper contains a description of some general principles necessary for understanding the dynamics of an expanding bubble and the associated ionization structure in a stratified atmosphere. The Kompaneets model can be used to determine the mean scale height of the ambient medium as well as the age of the bubble. The ionization structure also places constraints on the ambient density near the cluster. The surface brightness of the shell is estimated, as well as the fraction of ionizing photons which escape to the halo. The prescription used here can be applied to any observed bubble that is blown by the effectively continuous energy output of stellar winds or multiple supernovae. Application to the W4 bubble shows that the mean scale height of the ambient gas around the cluster is 25 pc for a cluster distance of 2.35 kpc. The age of the bubble is estimated to be about 2 Myr, consistent with the notion that the bubble is blown by stellar winds from a very young OB association in which no supernovae have yet occured.
Phil Komljenovic, an undergraduate at York University is just completing a summer project in which he has developed bubble models which are more sophisticated than the Kompaneets model. He will continue to work with the Theory Group this fall and expects to be able to perform a more detailed parametric study of the W4 region.
The CGPS Visualization Group met in Calgary on Monday, July 14, 1997, to discuss software and data issues pertaining to visualization of the CGPS data. In attendance were Sean Dougherty, Jayanne English, Steven Gibson, Richard Gooch, Denis Leahy, Russ Taylor, and Brad Wallace. Richard Gooch is the author of the Karma software package, which has been chosen by Consortium members to be the primary tool for the scientific visualization of three-dimensional datasets from the Canadian Galactic Plane Survey.
The purpose of the meeting was twofold: to allow Richard to explain the internal workings of his software in person to those most interested in using and extending it for specific data applications, and to give the same parties an opportunity to communicate their needs and concerns with regard to the ongoing evolution of the Karma package. Both goals were met. A presentation was given by Russ Taylor on the general nature of the Survey, the types of data it will generate, and what tools would be useful in the analysis and display of this data. Richard followed this with his own presentation on Karma's design philosophy and internal mechanics, which led into an extended discussion among those present about current needs and possible future developments of the software.
In particular, a facility requested by Sean Dougherty and Brad Wallace to read the DRAO data format directly was added by Richard shortly thereafter, greatly reducing the overhead involved in dealing routinely with very large datasets while they are being processed. Several other features of greater complexity were planned for future development as collaborations between Richard Gooch and Steven Gibson which may be of general interest to the larger Karma user community, but will definitely be of great utility to CGPS scientists. These include: the ability to integrate position-velocity spectra in the perpendicular spatial direction in order to enhance signal-to-noise; a package of spectral analysis tools for examining 3-D spectral line data in arbitrary data subsets; a suite of user-controlled image annotation features with scripting capabilities, such as marking and labeling major and minor axes of objects; and enhanced, generalized color table representation and control, e.g., for custom colormap generation. Jayanne English generously offered to advise and assist on the last two projects after being introduced to C programming in the Karma environment during her Calgary visit. Richard also remained after the meeting for a few days to aid the start of the annotation project and work on the DRAO data format reader.
Commentaires ou suggestions? Comments or suggestions?
Ecrivez à / Write to email@example.com