On the Fly Mapping at the FCRAO 14m Telescope

The OTF program shares several keywords with the Observatory MAP program but have modified definitions.

With the START OTF command, the antenna moves to the REF position and makes a calibration measurement (unless NCAL=0) which is immediately followed by a reference measurement. The antenna then begins to raster across the source starting at the upper (lower) left corner of the desired map depending if the sign of DY is negative (postive). At each TSAM interval, the MODCOMP sends an interrupt to the spectrometers to dump the accumulated data. It also sends a packet of positional information which gets tagged along with the data. Data are also accumulated as the antenna is turning around to scan across the next row. Once the antenna has traversed NBREF rows, a second reference measurement is taken for this set of data and one otf data file is written for each IF or LO configuration. If the antenna has traversed NCAL rows since the last CAL measurement, a CAL observation is activated followed by another reference observation. If a CAL is not necessary, then the ending reference observation of the previous set is used as the first reference of the subsequent set. The map is completed when the reference pixel in the array has offset LY+2*BY*RAMPY arcminutes from the starting position.

The Data File Selection environment facilitates the consolidation of raw otf files to be processed and creates the Input List file required by the Data Assessment and Regrid Data environments. The user sets some basic criteria through which the data files are sorted and filtered. These criteria are Source Name,Starting Scan Number, Ending Scan Number, IF number, and UT Date Interval. One can view the filtered list of file names by left clicking on the Filtered Raw Data List button. To view the names and parameters of the entire set of raw data available on the disk, one left clicks on the All Raw Data button.

Once the list of files are displayed, the user can select all of the files by left clicking on the All button. To select individual files, left click on the file name in the displayed list. One or more files can be selected with consecutive clicks. To enter the highlighted files into the OTF file list, left click on the right arrow button. To remove file names from the list, highlight the file name in the OTF file list window and left click in the left arrow button. To remove all files from the list, left click on the None button.

Finally, one needs to either create an Input File List file or append / overwrite an existing file. Enter the name of the Input File List and left click on the Save List file button. When you subsequently enter the Data Assessment or Regrid Data environments, the Input File List value will already be filled with this file name.

Figure 4.1. Display of the Data File Selection tab in otftool

The names of the raw otf files are read into the program and displayed on the right side of the Data Assessment Window. To view the basic header of a given file, click the left mouse button on the name of the file (this highlights the name) and then double click the right mouse button. An auxiliary window pops up to display the header information. The Data Assessment tab displays the parameters which can be set by the user to inspect the quality of the data. These are the Reference Weighting Scheme, Assessment Mode, and baseline parameters.

The Reference Weighting Mode (Equal | Biased) sets how the two bookend reference observations within the file are weighted to each on source spectrum. The equal weighting mode implies that the two reference observations for a given pixel are simply averaged to form a single reference spectrum for all data obtained with that pixel. A biased weighting applies a unique weight to each reference spectrum which reflects its time proximity to the on-source measurement. For a scan with NDUMPS, the reference weighting for the ith dump of the spectrometer is,

Weight for Reference 1 = 1.0-(i-1)/ndumps

Weight for Reference 2 = (i-1)/ndumps

An optional parameter is Set Horn Mask. This allows the user to either include or exclude selected pixels in the file. The default is to include all horns in the file for processing. The sequence of numbers to the right of the Set Horn Mask button is the bit array which reflects the mask (recall that bit arrays are evaluated from right to left such that pixel 1 corresponds to the last digit on the right and pixel 32 corresponds to the first digit on the left). To set the horn mask, left click on the Set Horn Mask button. This pops up an auxiliary window with a check box for each horn. One can quickly reset the mask by clicking on either the Select All Horns or Deselect All Horns buttons. Once the horns are selected, click on the Dismiss button to remove this popup window.

Currently, there is only one mode to gauge the data quality. This is the RMS mode in which the rms of each spectrum contained in each file is calculated. The user needs to set the baseline parameters ( baseline order, number of windows, window intervals, and window units (channels | velocity)) to determine the standard deviation of antenna temperature values in backend channels which lie outside the window intervals. These baseline parameters are applied to each file in the input file list excluding those files for which the user has specfically defined the parameters in the Edit Selected function (see Section 4.1.2) or the command line otfedit program (see Section 4.2.2).

To view the mean rms for each pixel in the array for a selected file in the list, left click on Baseline Selected button. A progress bar appears in the status window to show the progression through the file. To view the statistics over all pixels in all files in the input file list, left click on Baseline All button. Two progress bars appear in the status window - one to show the progression through the file and a second to show the steps through the input file list. When the function has stepped through all dumps of all selected files, an auxiliary window pops up to print a summary of the average rms values in a matrix for each horn and each file. There is an option at the bottom of this auxiliary window to plot the average values for each pixel.

Figure 4.2. Display of the Data Assessment tab in otftool

To visualize the raw data and to edit the data on a file by file basis, one can select a file in the input list (click left mouse button on the file name) and left click on the Edit Selected button. Alternatively, select a file and right click on the file name which pops up a window with the choices (HEADER | EXAMINE). Select EXAMINE. The editing/examine option launches a pgplot window which displays a greyscale image of the data from the first selected pixel of the mask with velocity on the x axis and dump index on the y axis. The program then waits for specific keystrokes with the cursor position which allow the user to inspect the data (plot raw spectra) and to define regions of data to be ignored by subsequent analyses or regridding. The keystroke/cursor options are summarized in Table 4.1.

The user is reminded that the baseline subtraction in only within the memory of the program. The results are not written into the otf file. The baselines are dynamically removed during the regridding step described in Section 4.1.3.

For the "w" keystroke, the user sets one or more windows (Maximum of 5 windows) within which the spectrometer channels are not included in the calculation of baseline polynomials. Two consecutive "w" keystrokes define a window. The "x" keystroke escapes the window utility. Vertical green lines highlight each defined window.

The "m" keystroke enters the mask utility in which the user can set a range of rows/indices which contains suspect data to be excluded from the regridding process. Two consecutive "m" keystrokes define the rows of data between the cursor positions to ignore. The defined region is highlighted with red hashing. Consecutive "u" keystrokes reset the index mask between the two cursor positions. The "r" keystroke resets the entire mask for the displayed pixel and the "R" keystroke resets the mask for the entire file. When reset by the u,r, or R keystrokes, the red hashing disappears.

Since the raw otf data files are readonly and one is discouraged to simply copy these files from the /RAW directory to the user's current directory, the "w", "m", "u", "r", "R" keystrokes create or modify an extension header file (filename.X) in the user's current directory. This file contains the baseline parameters and indice masks defined by the user in this editing step. The baseline parameters override those set by the user within the otftool and are applied to this file only. This extension header is inspected in all subsequent calculations upon the data contained in the file.

On the Fly mapping generates large volumes of data sampled on an irregular grid. Regridding this data takes all of the redundant measurements of the otf map and constructs the end product - convolved spectra on a regular grid. Currently, otftool can write discrete, convolved spectra into a CLASS file or construct a fits data cube to be imported into IRAF, IDL, or other image processing packages. The X and Y limits of the output grid are determined from information contained in the headers of the otf files. The ramp areas are not included in the final grid. The input otf files are generated from one or more setups of the OTF data collection program for a given source. Note that the regridding program checks whether the otf files have the same source name, central position, and alignment of the spectra. If the spectra are misaligned or the files do not share the same (0,0) position, the regridding process aborts. There is a limit to the size of the output grid to be less than 536 Mbytes which is sufficient to hold a 2048 channel x 256 x 256 pixel grid. To construct a larger data cube, the user will need to first make smaller mosaics and recombine these in a later step.

The Regrid Data tab of the otftool displays the menu for both required and optional parameters which affect the construction of convolved, regridded spectra (see Fig 4.2). Several of these parameters are shared with the Data Assessment tab (Input list file, Reference Weighting Scheme, Set Horn Mask, and baseline parameters). These are redisplayed in the Regrid Data window. The user needs to select the Output File Type (FITS | CLASS) and the name of the Output File. These should contain the full name as no default extension (i.e .bur or .fits) is appended. For CLASS output file type, a CLASS file is created if it does not yet exist. If the file already exist, the new spectra are appended. Since the output spectra are a convolution of many spectra with different scan numbers and horns, the observation number is set to the sequence in which it is written into the CLASS file. For each spectrum, a weighted average system temperature and an effective integration time are stored in the header.

Two critical required parameters which need to be set by the user are the Output Cell Size in arcseconds and the Spatial Filtering Kernel. The width of the convolution function is set to be twice the cell size to generate a Nyquist sampled grid . Figure 5 shows the various spatial filter functions and Table 4.2 lists the formula for each function. The default Spatial Filtering Kernel is Jinc*gaussian*Jinc (kopt=1) function.

Figure 4.3. Spatial Filter Functions for regridding data.

There are several optional parameters which affect either the regridding process or simply provide additional information in the output headers.

The Set Horn Mask is described in section 4.1.1. When regridding, the mask defined by the Set Horn Mask and the index masks generated from editing the data are both examined to determine if a spectrum is to be convolved into the output grid.

Linename is an optional string parameter which is inserted into the CLASS and fits headers. It is a useful keyword to efficiently sort and gather spectra with the same line name.

Once the required and optional parameters are set, the user initiates the regridding process by left clicking on the Compute OTFMAP button. Two progress bars appear in the Status Window. The first progress bar shows the fraction of the otf file processed and the second progress bar shows the steps through the input file list. When completed, a window pops up to inform the user that the regridding is finished.

Figure 4.4. Display of the ReGrid Data Tab in otftool

The otfmap program is a command line function to regrid OTF data exactly as otftool but without the graphical user interface frontend. Therefore, it is useful to place into shell scripts to process many individual maps in batch mode or if the user does not have an X Window interface or for those who simply do not like GUIs. It operates like any UNIX utility with flags followed by arguments.

For example, to regrid the otf data files contained within L1551.lis while removing the first order baselines determined outside of the velocity window of 3 to 8 km/s, flatfielding, tsys noise weighting with output directed to

otfmap -f L1551.lis -t fits -o L1551.fits -cell 25 -wt equal -kopt 1 -flatfield yes -base yes -windtype velocity -wind 3:8 -order 1 tsysweight yes

otfmap -f L1551.lis -t fits -o L1551-HORN5.fits -cell 25 -wt equal -mask 5 -kopt 1 -flatfield yes -base yes -windtype velocity -wind 3:8 -order 1

One could write a script which successively executes otfmap with a different mask and writing the output to a different file to inspect the consistency of data between array elements.

To write the regridded spectra to a CLASS file

otfmap -f L1551.lis -t class -o L1551.bur -cell 25 -wt equal