All that is required now is that the data be rebinned onto a rectangular grid with the rebin task. If necessary it is possible to enter Az/El pointing corrections by using change_pointing:
The time for the pointing corrections must be in LST (they also must be entered in chronological order). The pointing offset is assumed to vary linearly with time. Here I have assumed good pointing at an LST of 12h (the pointing observation before the map) and a small shift 50 minutes later when another pointing observation was taken (the shift can be found by using pointsum). It is probably best that the pointing offset is measured directly from the image by first regridding the data in Az/El coordinates and then using, for example, the KAPPA centroid task to find any offset.% change_pointing n59_sky_lon SURF: run 59 was a MAP observation of 3c279 SURF: observation started at LST 12 19 59 and ended at 12 28 40 SURF: no pointing corrections found CHANGE_POINT - Do you want to change the pointing correction data > y POINT_LST - The sidereal time of the pointing offset (hh mm ss.ss) /!/ > 12 00 POINT_DAZ - The azimuth pointing correction to be added (arcsec) > 0 POINT_DEL - The elevation pointing correction to be added (arcsec) > 0 POINT_LST - The sidereal time of the pointing offset (hh mm ss.ss) /!/ > 12 50 POINT_DAZ - The azimuth pointing correction to be added (arcsec) > 1.1 POINT_DEL - The elevation pointing correction to be added (arcsec) > -0.9 POINT_LST - The sidereal time of the pointing offset (hh mm ss.ss) /!/ >
A number of questions need to be asked before regridding the data :-
Currently, three methods are available. The data can be regridded with a weighting function, interpolated using spline fitting or by calculating the median data value in each cell of the output grid.
Bessel, Gaussian and linear weighting functions are available; in theory the Bessel function interpolation should give the best results but in practice the Gaussian or linear functions should be used (they are much faster and less affected by edge effects). The Gaussian function should be used if you are interested in beam shape (since it is easier to work out what is going on when you convolve a JCMT beam with a Gaussian than when you convolve it with a cone).
The MEDIAN regridding technique can be used if many data points are available (since for small output grids at least one input data point [preferably many more] must be available in each cell in the output to avoid bad pixels). If fewer points are available (only a few integrations) consider using larger cells or use the KAPPA routines fillbad or glitch to interpolate over the holes.
The spline fitting algorithms are experimental and have not been thoroughly tested - please use with care.
The data can be rebinned in the following coordinate systems:
| NA | Nasmyth (SCUBA) coordinate frame |
| AZ | Azimuth-Elevation offsets |
| PL | Moving source (e.g. planet) |
| RB | RA/Dec B1950 |
| RJ | RA/Dec J2000 |
| RD | RA/Dec epoch of observation |
| GA | Galactic coordinates (J2000) |
The first two coordinate systems are fixed on the telescope so that the source rotates during long observations. They are most useful for taking beam maps (AZ) or examining the properties of the SCUBA bolometers (NA). Obviously AZ and NA contain no astrometry information. The PL coordinate system should be used for moving sources (e.g. planets or comets) where the RA and Dec of the source is changing with time; offsets from this moving centre are calculated and no astrometry information is stored. The remaining coordinate systems correct for source rotation and do have associated FITS World-Coordinate-Systems (WCS) astrometry information [19,20].
The default map centre will be the map centre of the first map entered into rebin modified to the epoch of the output map if necessary. coordinates used as the map centre of the This question is not asked if a NA, AZ or PL coordinate system is being used.
The regridded image can be in any pixel size. The main point is that account is taken of the beam sizes: approximately 7 arcsec at 450 microns and 14 arcsec at 850 microns. The on-line system regrids with 3 arcsec pixels. Obviously the regridding takes longer the smaller the pixel size that is requested but only becomes a real problem if BESSEL regridding is used. Linear interpolation should be fast (less than 10 seconds) for most reasonable pixel sizes.
The map can now be made with rebin (in this case using linear interpolation, J2000 coordinates, 1 arcsec pixels and default map centre, looping is turned off since I am only regridding one map):
% rebin noloop REBIN_METHOD - Rebinning method to be used /'LINEAR'/ > SURF: Initialising LINEAR weighting functions OUT_COORDS - Coordinate sys of output map; PL,AZ,NA,RB,RJ,RD or GA /'RJ'/ > SURF: output coordinates are FK5 J2000.0 REF - Name of first data file to be rebinned /'n59_sky_lon'/ > SURF: run 59 was a MAP observation of 3c279 with JIGGLE sampling SURF: file contains data for 4 exposure(s) in 3 integrations(s) in 1 measurement(s) WEIGHT - Weight to be assigned to input dataset /1/ > SHIFT_DX - X shift to be applied to input dataset on output map (arcsec) /0/ > SHIFT_DY - Y shift to be applied to input dataset on output map (arcsec) /0/ > SURF Input data: (name, weight, dx, dy) -- 1: n59_sky_lon (1, 0, 0) LONG_OUT - Longitude of output map centre in hh (or dd) mm ss.ss format /'+12 56 11.17'/ > LAT_OUT - Latitude of output map centre in dd mm ss.ss format /'- 05 47 22.1'/ > OUT_OBJECT - Object name for output map /'3c279'/ > PIXSIZE_OUT - Size of pixels in output map (arcsec) /3/ > 1 OUT - Name of file to contain rebinned map > n59_reb_lon WTFN_REGRID: Entering second rebin phase (T = 0.9061 seconds) WTFN_REGRID: Entering third rebin phase (T = 3.682912 seconds) WTFN_REGRID: Regrid complete. Elapsed time = 4.055644 seconds.
If more than one map is available the extinction corrected data (with or
without sky removal) can all be added into a single map at this stage. The
parameter IN can be supplied with one new map at a time or via a text
file (§![[*]](crossref.png)
). Each input data set can be shifted by setting
SHIFT_DX and SHIFT_DY (this shift is in arcseconds on the
output grid cf. pointing corrections which are in Az/El offsets) and assigned
a relative weight with the WEIGHT parameter. rebin does understand
SCUBA sections (§) so it is possible to select part of an
observation for regridding at this time.
In addition to rebin there are three closely related tasks (in fact they all use the same code): bolrebin will regrid each bolometer individually, intrebin will regrid each integration into a separate file and extract_data will write the data to a text file before regridding. Note that the output file for bolrebin and intrebin is an HDS container [21] rather than a simple NDF. For example, if the OUT file is test.sdf the images will be accessible as NDFs via test.h7, test.h8 etc (or test.i1, test.i2 for intrebin).
![\includegraphics[width=\textwidth]{sun216_3c279.eps}](img10.png)
|
At this point the map can be displayed with, say, KAPPA display.
Fig. shows the 850 micron image of 3C279 rebinned in RJ
coordinates with the long wave bolometer array overlaid (note that
scuover displays the array at zero jiggle offset).
Fig. was made as follows (note that this requires
psmerge [22] in addition to KAPPA's display):
% display n59_reb_lon axes lut=$KAPPA_DIR/bgyrw_lut device=epsfcol_p MODE - Method to define the scaling limits /'SCALE'/ > LOW - Low value for display /-0.01095889788121/ > HIGH - High value for display /0.022901531308889/ > % scuover prompt MSG_FILTER - Messaging level /'NORM'/ > DEVICE - Name of graphics device /@xwindows/ > epsfcol_p Current picture has name: DATA, comment: KAPPA_DISPLAY. Using /scuba/maps/sun217/n59_reb_lon as the input NDF. EXT - Name of (extinction corrected) demodulated data file /'n59_sky_lon'/ > SURF: file contains data for 4 exposure(s) in 3 integration(s) in 1 measurement(s) INTEGRATION - Integration number /1/ > EXPOSURE - Exposure number /1/ > COL - Colour of annotation /'red'/ > white NAME - Display bolometer name (else number)? /TRUE/ > % psmerge -e gks74.ps gks74.ps.1 > 3c279.eps
In general, for faint sources it would now be necessary to go back to the extinction corrected (or sky-removed) data so that any bad bolometers and integrations can be turned off (using change_quality and SCUBA sections - rebin can be used to test a section before committing the change), different sky bolometers chosen or new pointing corrections added. Once complete the data can be calibrated - planet fluxes can be obtained using the FLUXES package and work is progressing on a list of secondary calibrators (see e.g. [24]).
SURF -- SCUBA User Reduction Facility