Spitzer Documentation & Tools
MIPS Instrument Handbook


Appendix C.      Pipeline History Log


S18.13 (No official release date):

A very small subsection of the MIPS data were reprocessed through a pipeline labeled S18.13. This pipeline is identical in every way to pipeline S18.12. It is NOT a new pipeline and contains NO changes or improvements over S18.12.


S18.12 (Release Date: Oct 09):

All the MIPS campaigns were reprocessed with S18.12 starting in early October 2009



No major changes.



For Photometric mode observations only, Prime & Parallel (all other modes remain the same) standard BCDs and enhanced BCDs will be delivered. The final PBCD is created from the enhanced data.


New products for S18.12:

          I.      Basic Calibrated Data (BCD) pipeline:

There are now two versions of the BCD products for MIPS-24 Photometry data: the original BCDs (bcd.fits) and the enhanced BCDs (ebcd.fits). The BCD and EBCD pipelines differ in the method of flat fielding, as explained in section II below. The EBCDs should, in most cases, have superior flat fielding. The mosaics are made only from the EBCDs (for data from the photometry mode). An example suite of BCD products associated with a single BCD are shown below.


BCD (original pipeline) products:

SPITZER_M1_19241216_0000_0002_6_bbmsk.fits    bcd mask SPITZER_M1_19241216_0000_0002_6_bcd.fits      bcd SPITZER_M1_19241216_0000_0002_6_bcd.log       bcd log SPITZER_M1_19241216_0000_0002_6_bunc.fits     bcd uncertainty

SPITZER_M1_19241216_0000_0002_6_dfunc.fits    difference uncertainty

SPITZER_M1_19241216_0000_0002_6_diff.fits     difference SPITZER_M1_19241216_0000_0002_6_slope.fits    slope SPITZER_M1_19241216_0000_0002_6_slunc.fits    slope uncertainty


EBCD (enhanced pipeline) products:

SPITZER_M1_19241216_0000_0002_6_ebcd.fits   enhanced bcd SPITZER_M1_19241216_0000_0002_6_ebcd.log    enhanced bcd log SPITZER_M1_19241216_0000_0002_6_ebmsk.fits  enhanced bcd mask SPITZER_M1_19241216_0000_0002_6_ebunc.fits  enhanced bcd uncertainty

SPITZER_M1_19241216_0000_0002_6_ediff.fits  enhanced difference SPITZER_M1_19241216_0000_0002_6_edunc.fits  enhanced diff. uncert.

SPITZER_M1_19241216_0000_0002_6_ermsk.fits  radhit (rejection) mask

SPITZER_M1_19241216_0000_0002_6_eslpe.fits  enhanced slope SPITZER_M1_19241216_0000_0002_6_eslun.fits  enhanced slope uncert.

Please note that for DCENUM=0, the ersmk.fits (radhit pixel rejection mask) file is not made because it is a product of MOPEX, the mosaicker pipeline, and MOPEX omits data with DCENUM=0 because they are not well-calibrated compared to all the other MIPS-24 data. There are no longer any brmsk files because the BCDs are not input into MOPEX.

The filenaming convention is



(here INSTRUMENT = M1 = MIPS-24 microns)


Calibration (cal) products:

The new calibration files for S18.12 are flats and darks. An example is given below for a specific AOR to show filenaming conventions.


Flat fields, original BCD pipeline:

Flat fields are the product of a smooth, spot-free gainflat and a ''spotmap'' matching the spot locations in each individual BCD.


Campaign gain flats. The two files below are examples of the smooth, spot-free gain flat and mask for the original BCD pipeline. These are from ''campaign flats'' made from observations of relatively blank sky at the beginning of each campaign. The AORKEY used to make these campaign flats is in the filename.

SPITZER_M1_20668928_0000_6_A31018392_fcmsk.fits     gain flat mask SPITZER_M1_20668928_0000_6_C6025416_flat.fits       gain flat

The gainflat is either made from a dedicated flat field observation at the beginning of the observing campaign or, for early mission campaigns before appropriate latent-free flats were taken, a fallback flat. The gainflats consist of three planes: the first is the gain flat and the other two planes are associated uncertainty images (for the original pipeline, they are computed in different ways and only the second plane is applied; for the enhanced pipeline, the flat field format was preserved but the uncertainty images are only computed one way and are populated in the last two planes.)


Spotmaps. The spotmap for each BCD is taken from a large library of fallback spotmaps. The spotmaps are matched independently for each individual BCD by calculating the centroid of a fiducial spot on the (pre-flattened) BCD and picking the closest match in the library to the calculated position. For the original BCD pipeline, the spotmaps have names like:



Default flatfields if spotmatching fails. If spotmatching fails for any given BCD, a fallback flat field is used that matches the CSM_PRED but is not necessarily a good match to the spot position. This flat field is a single file with the CSM_PRED value in the filename but with no x and y position of the spot. It represents the product of a spotmap and a smooth flat field (therefore it is NOT a spotmap file). An example of such a fallback and its mask are shown below (this particular example is from a medium scan observation). Spotmatching fails more frequently for scan mode observations because the spots are streaked and it is more difficult to derive good centroids, especially for regions with bright backgrounds.

mips24_flatfield_6320._1959.5.fits mips24_flatfield_6320._1959.5_cmask.fits


The filename convention for the default flatfield when spotmatching fails is:





       II.      NEW, Enhanced BCD (EBCD) pipeline.


Flatfields are the product of a smooth, spot-free gainflat and a ''spotflat'' shifted to match the spot locations in each BCD. Here, all of the BCDs in the AOR are used to compute a single spot offset (SPOT_DY) in the y-direction relative to a reference template spotflat file. This single offset shift is applied to all of the planes in the spotflat file. The shifted spotflat file is a cube, matching the spot position of the current AOR, with one plane for each mirror position used in MIPS-24 photometry mode, in ascending order of header keyword CSM_PRED (predicted cryogenic scan mirror position). After shifting the reference to match the current AOR, the appropriate planes are used for flat fielding.


The headers of these files have explanatory keywords (COMM1,COMM2, etc.) describing the nature of each file.

SPITZER_M1_19241216_0000_6_A31034185_sfunc.fits   spot-flat uncert.

SPITZER_M1_19241216_0000_6_A31034187_sfcsm.fits   spot-flat CSM_PRED

 array (1-D array)

SPITZER_M1_19241216_0000_6_A31034188_gfmsk.fits   gain flat mask SPITZER_M1_19241216_0000_6_A31034189_gflat.fits   gain flat (spot-

 free, from campaign flat

SPITZER_M1_19241216_0000_6_C6027630_sflat.fits    shifted spot flat


''Extra'' flats for campaigns using fallback rather than campaign gainflats.

For the campaigns early in the mission that require fallback flats (due to the fact that no latent-free flat fields were taken during the campaign), a set of additional flat fields appear for both original and enhanced pipelines. These are to be considered ''extra'' files that are currently archived due to how fallback files are retrieved from the database. They can essentially be ignored as they are associated with the unshifted flats and are not actually applied to the EBCDs. They will appear as follows:


For the original BCD pipeline:

mips24_flat_090909.fits          fallback gain flat for original bcd


mips24_flat_090909_cmask.fits    mask for above

For the enhanced pipeline:

mips24_flat_090909_aorgain.fits            fallback gain flat  mips24_flat_090909_aorgmsk.fits            fallback gain flat mask  mips24_flat_090909_aorspotcsm.fits         fallback CSM_PRED array

 (1-D array) 

mips24_flat_090909_aorspotmsk.fits         fallback unshifted spot

 flat mask 

mips24_flat_090909_aorspotunc.fits         fallback unshifted spot

 flat uncertainty 

mips24_flat_090909_aorspotval.fits         fallback unshifted spot

 flat cube




New dark files were made for S18.12 (Oct. 2009). They are applied to both BCDs and EBCDs. They reflect improved calibration and signal-to-noise (especially for DCENUM> 0) over older verions and were made from approximately 100 dedicated dark observations taken over many campaigns. For MIPS-24, there is no clear evidence that photons are detected in the default scan mirror position used for dark observations; thus, darks derived from the same exposure time observations (10 sec) are applied to all MIPS-24 data, regardless of the exposure time of the science data.

mips24_dark_090804_sur1.fits           dark for DCENUM=0 mips24_dark_090804_sur1_cmask.fits     dark mask for DCENUM=0 mips24_dark_090804_sur1_uncert.fits    dark uncertainty for DCENUM=0

mips24_dark_090804_sur2.fits           dark for DCENUM>0 mips24_dark_090804_sur2_cmask.fits     dark mask for DCENUM>0 mips24_dark_090804_sur2_uncert.fits    dark uncertainty for DCENUM>0