14.1 $uv$ tables

After calibration with CLIC, the calibrated data may be stored in a particular file called a `$uv$ table'. This is useful because much of the data in the CLIC data file are not needed any more: atmospheric parameters, total powers, image sideband visibilities, data from other receivers may be discarded at this stage. All that counts is: the data that are needed to describe the source itself, the sky frequency that was observed, ... One may for instance create a $uv$ table for the continuum and one for each line that was observed.

These $uv$ tables are just special GILDAS tables suited for $uv$ data handling that are created by CLIC. Mapping consists of transforming these tables into something more meaningful for the astronomer, either images or numbers like positions, flux densities, sizes, etc. However a good part of the data evaluation and analysis can be directly performed on the $uv$ data itself, before performing any of the complex operations involved in creating an image (Fourier transform and deconvolution). Direct analysis of the $uv$ data is the subject of this Lecture.

14.1.1 $uv$ table contents

A $uv$ table is a file in the Gildas Data Format, of dimensions [3 $N_\mathrm{\scriptscriptstyle C}$+7, $N_\mathrm{\scriptscriptstyle V}$], for $N_\mathrm{\scriptscriptstyle C}$ spectral channels and $N_\mathrm{\scriptscriptstyle V}$ visibilities. The $3\ensuremath{N_\mathrm{\scriptscriptstyle C}}+7$ lines contain:

1. $u$ in meters
2. $v$ in meters
3. Scan number
4. Observation date (integer CLASS day number)
5. Time in seconds since above date
6. Number of start antenna of baseline
7. Number of end antenna of baseline
8. First frequency point (real part)
9. First frequency point (imaginary part)
10. First frequency point (weight)
11. Same for second frequency point, and so on

Thus for a given scan with $N_\mathrm{\scriptscriptstyle A}$ antennas, $\ensuremath{N_\mathrm{\scriptscriptstyle A}}(\ensuremath{N_\mathrm{\scriptscriptstyle A}}-1)/2$ visibilities are recorded.

The table header has the standard form of a GILDAS Image. The header is available (for instance) by declaring:

 GRAPHIC> SIC\DEFINE HEADER T co10.uvt READ
 GRAPHIC> EXAMINE T%

For a table named co10.uvt. Some keywords convey a more precise meaning for $uv$ tables:

T%NDIM should be 2

T%DIM contains 3 $N_\mathrm{\scriptscriptstyle C}$+7 and $N_\mathrm{\scriptscriptstyle V}$

T%RA,T%DEC coordinates (radians) of the pointing center (the center of the primary beam).

T%A0,T%D0 coordinates (radians) of the phase tracking center (a point source at this point should have zero phase); they are identical to RA and DEC when a table is first produced.

T%EPOCH The epoch of those coordinates. Should be 2000.0

T%VELOFF, T%VELRES The velocity of the reference channel, and the channel separation in velocity units (km/s)

T%RESTFRE, T%FREQRES The rest frequency, and the channel separation in frequency units (MHz)

T%CONVERT[1,1] the reference channel

T%CONVERT[1,2] the actual observing frequency at the reference channel (MHz); the one used to scale angular displacements from $u,v$ coordinates in meters.

One may also examine directly the header by typing simply :

 GRAPHIC> HEADER co10.uvt

14.1.2 How to create a $uv$ Table

$uv$ Tables are created by CLIC using the command TABLE.

A set of commands to create a $uv$ table may look like:

! Reset the default options:
  SET DEFAULT
! find the useful scans:
  FILE IN 21-JAN-1998-H126
  SET SOURCE IRC+10216
  SET RECEIVER 1
  SET PROCEDURE CORRELATION
  SET QUALITY AVERAGE
  FIND
! calibration options:
  SET AMPLITUDE ANTENNA RELATIVE
  SET PHASE ANTENNA RELATIVE INTERNAL ATMOSPHERE
  SET RF ANTENNA ON
! table creation:
  SET SELECTION LINE LSB L01
  TABLE HCN NEW /FREQUENCY HCN  88631.85 /RESAMPLE 19 10 -27 2.12 V

All but the last two commands should be familiar at this point.

• The first new command, SET SELECTION, prepares the last one TABLE. It selects that the next table to be created will be a line table (i.e. with more than one spectral channel). The lower sideband data will be used, and only the first subband of the correlator: L01.
• The last command TABLE, actually creates the table named hcn.uvt. The rest frequency, 88631.85 MHz, is set to be the reference used for the velocity scale. The data will be resampled to a velocity grid of 19 channels; the reference channel 10 will correspond to the LSR velocity -27 km/s; the channel spacing will be 2.12 km/s. Without /RESAMPLE, one would have got all the channels in the subband L01 with their original velocity separation. Without /FREQUENCY, the rest frequency present in the data (in the observing procedure) would have been used.

Using /RESAMPLE, one may avoid creating tables with too many channels (by discarding unused parts of correlator subbands) and choose the resolution that is actually needed.

If the data is spread on several files, one may go on by opening the other files, finding the data scans, and appending to the table:

  FILE IN 12-FEB-1998-H126
  FIND
  TABLE
  FILE IN 21-FEB-1998-H126
  FIND
  TABLE
...

(the arguments to TABLE need not to be repeated).

For continuum tables one may use:

SET SELECTION CONTINUUM DSB L01 TO L05 -
  /WINDOW 214405 214726 217476 217796 217837 217875
TABLE CONT-1MM NEW

Here we are using data from all the line subbands, but only in the three frequency windows: 214405 to 214726 MHz, 217476 to 217796MHz, and 217837 to 217875. This is of course to avoid the line emission of some molecules.

A standard menu is available under the CLIC main menu (``Create a UV Table''). After execution, a specific procedure is created to keep track of the options and parameters used. This procedure can subsequently be edited to add new data files (data files can also be added from the menu).

Figure 14.1: ``Create a UV Table'' menu in CLIC