Information for Astronomers
Information for Astronomers
The raw data of the 100m Effelsberg telescope is stored in MBFITS-Format. In Effelsberg the files are located in the directory /daten/Raw which should be available on every Observer-PC. Older data can be found in /daten/Raw/Raw-YYYY-MM.
Every 30 minutes the raw data is synced to Bonn. It is accessible in /eff/data/. More details on the location of data can be found here under Data storage and archive
The subscans are combined to a MBFITS map that should be processed with the “toolbox”. This can be done on the “observer7” PC in Effelsberg. For MPIfR observers the login can be done with their own LDAP account. Please add the directories /opt/bin and /soft/astro/nod/Ubuntu14_32bit/bin.gfortran to your path. Note: If you want to use a different PC than observer7, which is running on Ubuntu Version 14, you have to adjust the nod2 path to use the appropriate version compiled for your operating system.
# For bash export PATH=$PATH:/opt/bin:/soft/astro/nod/Ubuntu14_32bit/bin.gfortran # For csh set path = ($path /opt/bin) set path = ($path /soft/astro/nod/Ubuntu14_32bit/bin.gfortran)
For all other observers a common account called “obs2” can be used (ask the operator for the actual password). The toolbox performs some basic baselevel subtraction and filtering of strong radio-frequency interference (RFI) signals. Some parameters can be modified by hand if needed (see Reduction of pointing measurements for toolbox options). Transformation into NOD2 format, restoration of multi-horn data, transformation into right ascension/declination coordinates and Gaussian fitting for calibrators is done by dedicated automatic pipelines developed for each receiver.
The pipeline scripts can be found in /opt/actual_scripts/ There are scripts for cal maps that are called ketXXcal.script and perform an additional Gaussian fitting to check the (polarization) calibration. Scripts for target maps are called ketXXmap.script. The XX is some acronym for each receiver mostly the wavelength in cm or mm. For the 6cm receiver there are scripts with "db" for double beam and "sb" single beam reduction. Just copy the script in your directory and call it with the scan number to be reduced.
e.g. for a 11cm calibration map with scan number 0340:
|21cm PFK calibration map
|21cm PFK target map
|11cm SFK calibration map
|11cm SFK target map
|6cm SFK double beam calibration map
|6cm SFK double beam target map
|6cm SFK single beam target map
|3.6cm SFK calibration map
|3.6cm SFK target map
|2.8cm SFK calibration map
The resulting raw maps in NOD2 format are provided as files named “mpXXXX.nod” (where XXXX is the 4-digit “scan number”). These include a map of sidereal times, a map of parallactic angles, followed by quartets of maps of the 4 Stokes channels (R, L, U and Q) for each horn and for each frequency channel. The data format is NOD2 and can only be reduced using the NOD2 software.
As circular polarization is generally very weak, the maps in R and L are similar and can be averaged in the later steps of data reduction.
Automatic 2-D Gaussfits are provided for the maps of calibrators, stored as “GSXXXX” (single horn) and “gsiuqXXXX” (multi-horn, after restoration). These can be used for calibration of the flux scale and polarization angle, and to check the telescope pointing.
NOTE: NOD2 assumes that the maps are in B1950 coordinates, but single-horn maps can be observed in J2000 without problems.
The raw maps include artifacts which need to be removed:
This can be done with the NOD2-based software package called “Ozmapax”. In Effelsberg you can find it on the "observer3", user "nod2". The program is started with the command Oz.
Ozmapax is steered by keywords or key-numbers. The most important are:
NOTE: for removing one point in a map, specify the row and column, or mark a very small area around the pixel center with the cursor, or use “mpfiltb”.
“presse” is a powerful but tricky option which needs care and some experience. “presse” smoothes the map with an elliptical beam, fits the differences between the original and smoothed map and corrects all base-levels. The smoothing beam should be slightly larger than the telescope's half-power beam width in the scanning direction and about 3x the HPBW perpendicular (non-scanning) direction. “presse” needs specification of a threshold for the rejection of strong sources in the map. Start with first-order fits and not too many iterations. Be careful: Second-order fits may remove real emission!
NOTE: If “presse” does not work properly, the map has to be re-loaded with “mapin”.
More details about “presse”: see Sofue & Reich 1979, AAS 38, 251
“gauss” or “gaus2” are needed to measure the total flux of a source, to determine its polarization angle (from Gauss-fits in U and Q) and to check the telescope pointing with help of strong sources in the individual maps.
NOTE: Only positive sources can be fitted. Negative sources in U or Q need scaling by -1 before fitting, see below.
NOTE: “smooth” needs the value of the actual beam which is not always correctly stored in the header of the map. This is why “smooth” asks for the correct values (in RA and DEC).
NOTE: “smooth” does not change the scale. In case of scaling in Janskys per beam, the scale needs to be corrected by applying scale with the value: scale=(new beamsize/old beamsize)2.
NOTE: “smooth” can be applied only to maps in Stokes I, U and Q – but never to maps in polarized intensity, polarization angle and Faraday rotation measure!
NOTE: “fchop” may change the resolution, so that “scale” needs to be applied (see above).
NOTE: only needed once if a set of maps e.g. from 4 Stokes channels is reduced.
NOTE: “openout” does NOT actually write out the map!
The reduced maps are then available under the name specified in “outname”.
These have to be run separately for Stokes I (i.e. R and L together) and U and Q. Preparation of the several text files (“input file”) with separate input parameters for I and U and Q is recommended.
NOTE: “poldeniuq” stores the polarization angle (PA/1000) in the maps 4 (PI) and 5 (p*1000) as the fractional part. Hence only the integer part contains information on PI and PC. In deep maps of weak sources, the PI intensities can be very low so that clipping the fractional part in “poldeniuq” reduces the dynamic range or may even remove all information. Scaling the maps by a large a factor (e.g. scaling to microJy/beam) is strongly recommended!
NOTE: Check whether signal-free areas in the PI map have an average value of zero. If not, repeat “poldeniuq” with a corrected value for the rms noise.
Maps of polarized intensity can also be computed with the “algebra” option 6 of “Ozmapax”, but without correction of the noise bias, and hence should be used only for a rough estimate and not for the final map.
NOTE: Integration or averaging is only allowed for maps of the Stokes parameters I, Q and U, but NOT for maps in PI, PA, p, spectral index and RM!