The MDL-file (model components)

Contents

• Contents
• Inspecting the list of model components
• Source component parameters
• Overview of model `handling' options.
  • Calibration
  • Sorting
  • Editing
• Conversion to a uv-model for the .SCN file

Inspecting the list of model components

******* Insert new script here ******

The above model is in `local mode', which means that no reference position or reference frequency is known. This information may be taken from a SCN-file (use NMODEL option CONVERT), and stored with the model in the MDL file. The model then looks like this:

******* Insert new script here ******

The coordinates may als be viewed in RA and DEC:

******* Insert new script here ******

Note that not all source parameters (see below) are visible here. They may be inspected in the log-file whenever the model has been used.

Source component parameters

The source components in a NEWSTAR model have the following parameters:

• I: total flux in W.U. (even when extended or polarised)
• l, m: offset from the reference point (map centre) in arcsec
• ID: identification number (1,2,3,...) in the list.
• Q, U, V: Stokes Q, U, V in percent (of I)
• long, short, PA: long and short axes (full halfwidth in arcsec) of an elliptic gaussian extended source, and the position angle of the long axis (degrees North thru East).
• si: spectral index (). Only used if reference frequency is known.
• rm: Faraday rotation measure (). Only used if reference frequency is known.
• flags: (e.g. a `proper' source (0) or a CLEAN component (1))

NB: CLEAN components:
- Are not corrected for beam smearing in SELFCAL.
- Are confined to map grid postions, which may result in poor subtraction.
- Their positions cannot be "updated" automatically after SELFCAL.
Therefore, use ``proper sources'' for strong point sources, and CLEAN components for weak extended sources (but only if necessary).

Overview of model `handling' options.

In various NEWSTAR programs (NMODEL, NCALIB, NMAP etc) the user is able to handle the list of model components in various ways. The keyword often used is MODEL_OPTION:

• READ/WRITE: Read/write from/to an external MDL file
• CLEAR: Clear the source component list, while resetting reference coordinates
• ZERO: Empty the source component list, but keep the coordinates of the field centre and frequency.
• SHOW/LIST: Show source list on terminal screen, or both terminal and LOG-file. See example above.
• RSHOW/RLIST: Show source list in RA/DEC coordinates See example above.
• TOT: Show source list statistics (for a specified range of sources). The result looks like:

  
  Sources at epoch 1950 at 05:38:43.51, 49.49.42.8, 1417.248 MHz

  
  10 sources (0 deleted) with 4453.700 W.U. (Max= 1900.000, Min= 2.200)
  

• ADD: Add sources to the list by hand.
• CALIB: Convert the source list by scaling intensities and/or moving l,m positions.
• EDIT: Edit the sources in the list (an amplitude of zero will delete the source)
• FEDIT: Edit a `field' (parameter) for a range of sources
• MERGE: Combine sources components that have the same position
• SORT: Sort the source list in decreasing amplitude (sorting will always precede a write)
• FSORT: Sort on a specified `field' (parameter) in the source list
• DEL: Delete sources
• DNCLOW: Delete non-CLEAN components with low amplitudes
• DCLOW: Delete CLEAN components with low amplitudes
• DAREA: Delete sources in specified area

In many cases, the user may specify sources in the list upon which some `handling' options is to act. The specification may be in the form of a SOURCE_LIST of id-numbers, separated by comma's (e.g. 3,6,7,2,34). If the user answers with , a SOURCE_RANGE is asked, i.e. the first and last id-number of a contiguous sublist:

SOURCE_LIST (Source number list) = :
SOURCE_RANGE (Source number range) = "":

means all sources

Calibration

If the flux or position of one of the sources in the list is known, it may be used as a calibrator. In the example, all 10 fluxes of the example model used above are multiplied by a factor (2000/1900), to make the flux of the first one equal to 2000 W.U. At the same time, all positions are shifted.

******* Insert new script here ******

Sorting

The source components may be sorted in many different ways. The option SORT orders all components in order of decreasing flux. This operation is performed before each WRITE, so that the components in an MDL file are always sorted this way!.

MODEL_OPTION (READ,WRITE,CLEAR,ZERO,SHOW, ) = QUIT:

FSORT allows the sorting on a range of criteria. In this example, the soorces are sorted in the order of decreasing m-coordinate:

****** Insert new script here *****

Other possibilities are:
- I: amplitude
- L,M: l or m
- LM,ML: l and m, or m and l
- ID: identification
- Q,U,V: Q or U or V
- SI,RM: spectral index or rotation measure
- LA,SA,PA: long or short axis, position angle
- BITS: bits (i.e. if source is extended or has polarisation)
- TYP: source type (0 is the standard)
- CC: clean component
- TP2: reserved
- DIST: distance to a specified centre
- POL: polarised intensity

Editing

Model components may be edited in various ways, using `handling' options CLEAR, ZERO, ADD, MERGE, DEL, DNCLOW, DCLOW (see the overview above). The handling options EDIT and FEDIT allow specific parameters (`fields') to be edited separately. EDIT acts on a specified source, and FEDIT acts on a range of sources. The possibilities are:

- I: amplitude
- L,M: l or m
- ID: identification
- Q,U,V: Q or U or V
- SI,RM: spectral index or rotation measure
- LA,SA,PA: long or short axis, position angle
- BITS: bits (i.e. if source is extended or has polarisation)
- TYP: source type (0 is the standard)
- CC: clean component
- TP2: reserved

Conversion to a uv-model for the .SCN file

The calculation of uv-model data (for comparison with uv-data in NCALIB)i s time consuming. Therefore, a calculated uv-model is saved in the .SCN file for which it is calculated, together with the source list used. In all programs that use the model data it can be specified how the model calculation should be done, and if the calculation should be saved. The relevant question is MODEL_ACTION, which expects a list of three answers. The first one can be one of:

• merge: replace the model saved in the SCN node with the one specified by the user. However, first compare these two lists, and only add the difference to the saved model data. I.e. make the calculation as short as possible if the new list differs only slightly from the saved one.
• add: replace the saved model with the sum of the saved one and the one specified by the user. The model calculated on the basis of the user specified list is added to the saved data.
• new: replace the saved model by the model specified by the user, and calculate a completely new set of model data
• temporary: do not use any data in the SCN node, or write anything, but use the data based on the list specified by the user
• increment: use the saved model data, and add to it the model data based on the user specified list, but do not save anything

The second answer can be BAND or NOBAND, and specifies if in the model calculation source data should be corrected for bandsmearing to match the actual data better. The third answer can be TIME or NOTIME to indicate the use of integration time smearing.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.