Examples of Point Source Spectral Extraction with SPICE

SPICE runs the same software that is used in the SSC pipeline. When run with all the default options selected, it produces the same extracted spectra available from the archive. There are many science programs that will need to optimize one or more aspects of the extraction, or to perform additional reduction of BCDs before extraction. On this page, we give a few basic examples processing to be done before and with SPICE.

NOTE: When BCD data are modified prior to spectral extraction, it is important to preserve the FITS header information because some keywords are used by the extraction pipeline.


Example: Extracting an object not in the default FOV

In low resolution, the pipeline searches for the brightest source near the requested position. A common use of SPICE is to extract another source in the slit. The location of the extraction window is controlled by the Ridge module. It is specified as a percentage along the slit, in the +x direction. The same location is used for all spectral orders. So, to extract another source in the slit:

  1. Identify in which spectral order(s) the new target is located.

  2. If the target is not in the default spectral order, then change the order in Profile. For example if it is in SL1 but the original target is in SL2, then change "Low Res Order" from "Default" to "1st Order". The default order can be determined, for example, from the Field-of-View keyword in the FITS header, which is also displayed at the top of SPICE in the FOV window.

  3. Run Profile

  4. View the Profile output and measure the approximate location of the object along the slit (in the x direction), shown in percent on the Profile graph. Use this location as input into the Ridge module by setting "Percent setting" to "Manual" and entering the position in the "PERCENT" box.

  5. Run Ridge. Modify the ridge center and re-run as necessary.

  6. Run Extract and Point Source Tune to extract and calibrate your spectrum.


Example: Sky Subtraction

The SSC pipeline performs background subtraction by differencing the nods of the pipeline coadded data products (the *.bksub.fits products in the pbcd subdirectory). However, users may need to perform their own sky subtraction for a variety of reasons before extracting spectra. Very faint sources (pdf), in particular, may require processing before coaddition and so new sky subtraction will need to be performed. Here we provide three examples of sky subtraction in the low resolution modules:

  1. Extracting the sky spectrum. One option for sky subtraction is to follow the standard longslit technique of extracting a ``sky spectrum'' at one or more blank positions along the slit. The 1D sky spectrum can then be subtracted from the target spectrum. SPICE allows the user to view extraction windows to choose blank regions and then extract them. The extraction is performed following the procedure outlined for other targets.

  2. Differencing the Nods. Staring mode observations are taken in nodded pairs, with telescope offsets performed along the slit between observations. The size of the offset (one third of the slit length) is large enough that compact objects have no overlapping pixels in the two dispersed images. Thus, a standard ``chopping'' technique can be used taking the difference between the two nod positions before extracting. This is the method used by the pipeline for post-bcd products.

    The difference can be taken in any standard way, for example in IRAF or IDL. Care must be taken to ensure that the FITS header of the source, as opposed to the sky, is retained. After sky subtraction, the target can be extracted in the standard way.

  3. Subtracting the other subslit: Many programs obtain spectra in both low resolution spectral orders (i.e. SL1 and SL2). The orders are observed separately in the two subslits. Thus, for compact objects in sparse fields the unused subslit often measures the sky. In most cases, both nods of the unused slit can be coadded to create a higher signal to noise sky image. This 2D subslit sky can be subtracted from the target dispersed image before extraction.


Example: Changing the Fatal Bitmask

The *bmask.fits file provides status information for each pixel in the input image. During extraction, pixels masked as ``fatal'' do not contribute to the spectrum. By default, the online pipeline defines bits 7,12,13,and 14 as fatal. In the Extract module, the user can change the definition of fatal bitmask conditions using the pull down menu. This change may improve the fidelity of features in the final spectrum or, more often, the cosmetic appearance of the continuum. Some users will modify the bitmask to reflect their own evaluation of pixel status and may wish to correspondingly change the definition of which bits are fatal.

The definition of bit-setting conditions in the bmask.fits file are given in online help within SPICE. For a full discussion of the bitmask, see the IRS Data Handbook and the IRS Pipeline Handbook.


SPITZER HOME || SPITZER SCIENCE || INSTRUMENTS || SCIENCE USER SUPPORT || SEARCH

help@spitzer.caltech.edu

This file was last modified on Fri Aug 29 14:39:44 PDT 2008.

California Institute of Technology Jet Propulsion Laboratory NASA