These plots are intended for reference purposes only and are often not the best images which could be made from the input data. They do not give definitive results for detection experiments! Fuller results may have been published already.
You can obtain the data to reprocess yourself by web links (where available) or by emailing merlin_archive@jb.man.ac.uk The Archive Policy explains when data becomes available, and if you use MERLIN data please add this acknowledgment.
The flux scale for all data has been set in Jy and a preliminary bandpass correction applied. These can be modified by the user. All further calibration and editing (performed in AIPS) is entirely reversible.
The maps presented here result from automatic processing of blocks of up to a few weeks of data. However your source may have been observed in one or more separate blocks (listed separately on the first search results page). For non-variable sources you may be able to combine data from several epochs. Alternatively some sources e.g. XRB may show artefacts characteristic of intra-day variability.
If your source can be self-calibrated and has good uv coverage then you may achieve a 1 sigma noise level of the order of 60 micro-Janskys/beam at C-band and L-band (apart from very bright sources which may produce dynamic-range limited maps). At the other extreme if your source has «50000 visibilities, much bad data, or is variable during observations, the noise can be an order of magnitude or more worse.
TOTAL INTENSITY
The minimum error in absolute position for a target observed in phase-referencing mode is constrained by the accuracy of the phase-reference source position, which is almost always <40 mas, often only a few mas. In addition, the relative error in the position of a bright peak is roughly proportional to the beam size divided by the signal to noise ratio (see maps) - better for targets with good uv coverage, possibly worse under adverse conditions where even compact sources show very rapid phase rates (see plots of phase solutions).
In general the flux scales here are accurate to ~5%. When considering source structure, note the beam shape (bottom left of map plots). Symmetric artefacts may occur around sources at low Dec. or with poor uv coverage (few visibilities) or with bad data (see plots of calibration solutions and of uv amplitudes). In single-channel data it is possible that extended structure may be poorly reproduced and components towards the edges of the larger maps may be distorted.
If you reduce the data more carefully, including
you may be able to
POLARISATION
Up to 2-3 weeks data is processed at any one time. Changes in instrumental polarisation properties should be small during the period chosen but ionospheric effects may change dramatically. Thus, in these plots the residual polarisation errors may be a few % of the peak flux but the polarisation angle errors may be »10°.
In order to obtain reliable polarisation measurements, in addition to the total intensity considerations, you should reprocess the data using
which should give
This is an example of the AIPS procedures used to produce these maps. There may be small deviations from this for any given data set.
See the MERLIN User Guide for information on data reduction and jargon.
1. The maps and uv plots shown on the web pages are made from data squashed to a single channel with calibration applied. These are suitable for refining the mapping or continuing with the calibration etc. of sources bright enough to self-calibrate and not needing wide-field mapping.
The simple single channel data are usually contained in the following files:
a) PHASE-REF.SPLIT and TARGET.SPLIT with the flagging and amplitude, phase and
polarisation corrections applied, all good channels averaged.
b) If the target was self-calibrated, TARGET.MULTI containing CL2 with
phase corrections.
c) The clean maps of the PHASE-REF and TARGET made from b) and c)
2. If your target is too weak to self-calibrate, if you are interested
in polarisation or wide-field mapping, if the data needs more editing
or for other reasons you want to start at an earlier stage, you want
multi-channel data. Such data have had the flux scale and a basic
bandpass correction applied.
a) TARGET.SPLAT will contain (as available) TARGET, PHASE-REF, POINTs,
3C286 in multi-channel format with extension tables:
AN table containing antenna positions and polarisation
leakage corrections (if all 4 Stokes present)
FG table containing flagging (if any)
SU and NX listing sources (in J2000 coordinates) and times of
observation relative to the first data in the file (day 0)
The flux of the POINT-CAL used is set.
CL1 - initial empty CL table with entries every 0.5 min
If you want to repeat the phase and amplitude calibration from scratch use TARGET.SPLAT (you could use single-channel maps as starting models and you could squash the data to a single channel before calibration).
You can use 3C286 to correct the polarisation angle (do not further modify the AN table supplied with TARGET.SPLAT in this case). Or, if you want to repeat the corrections for polarisation leakage too you can replace the TARGET.SPLAT AN table with that from PHASE-REF.SPLIT (single-channel data).
See the MERLIN User Guide for more details of data reduction.