Summary of NICMOS flatfielding documents
                           Dec 12, 2007


    (One outstanding question: what about wavelength calibration?  
     See notes at end)
              

NICMOS ISR-95-005: Characterisation of NICMOS Array Flat-field Response.
  - used a SPARE array, not actual flight arrays
  - tested at Steward in 1995
  - flashlight on wall provides illumination
  - used many filters, broad and narrow
  - very large change in dynamic range of variations
       short wavelength -- factors of 2
       long wavelength -- factors of 10-20 percent?
  - important to locate spectral regions in which the
       pixel-to-pixel sensitivity changes rapidly,
       that's where color terms will be most important
  - conclude errors in flatfield response are around 4 percent


NICMOS ISR-95-006: Effects of the NICMOS Array Flat-field Response on  
  - numerical simulations of how the flat-field response measured
       for SPARE array might affect stellar photometry
  - sources were a 10,000 K blackbody and 700 K blackbody
  - typical error created by imperfect flatfielding <= 3 percent
  - biggest errors in passbands in which response changes
       most rapidly with wavelength


NICMOS ISR-98-002: First NICMOS On-Orbit Flat-Fields: Results from SMOV Data
  - real on-orbit flats with internal lamp
  - dithered positions in 16-point pattern, then subtracted
       lamp-off exposures at same positions, to subtract background light
  - differences 2-3 percent between pre-flight and in-flight flats
  - common structure in diff passbands
  - some spots caused by debris on detector surfaces
 


NICMOS ISR-98-011: NIC1 Narrow-band Earth Flats
  - exposures of Earth, streaked, in 1997
  - combine exposures taken at different streak angles
  - only possible in narrow-band, as wide-band saturates 
  - "super-shading" is special detector effect triggered only by
        (very intense) Earth exposures, must correct for it
  - some features visible in Earth-flat divided by lamp-flat
  - reduce stellar images with the two types of flats
        photometry differs at 1-2 percent level, at most
        only 2 isolated stars used in test, though


NICMOS ISR-99-002: Color Dependence of NICMOS Flatfields
  - how to create a flatfield appropriate for a source
        which is very different in color than the source(s) used
        to create ordinary flatfields?
  - method A: interpolate on graph of sensitivity vs. wavelength
        to the wavelength of the source
  - method B: given spectrum of source, combine weighted average
        of narrowband flats to match source spectrum
  - can use either method
  
  Hmmm.  Create flats for individual sources?  Not a good idea
for SNAP, methinks.



NICMOS ISR 2002-004: SM3b Science Flats
  - based on internal lamp flats in 2002
  - same set of dithered positions, with lamp-on minus lamp-off 
         to get rid of background light
  - no obvious changes relative to earlier flats
  - some new "grot" due to stuff falling on detectors




NICMOS ISR 2005-002: Grism Sensitivities and Apparent Non-Linearity
  - use the "ordinary" flatfields to correct the raw grism images
  - not sure what else to say here ...


NICMOS ISR 2007-002: NICMOS Time Dependent Flat-fields
  - examine internal lamp flatfield images taken over span 2002-2007
  - find general decrease in QE of about 1 percent over 5-year period
  - some large-scale spatial variations grow with time 
  - changes could be due to a change in temperature of detector over time



    
--------------------------------------------------------------------------

  Wavelength calibration -- how?

      - McCarthy et al., 1999ApJ...520..548M, states
           "The rms accuracy of the wavelength calibration for 
               G141 is about 0.002 micrometers " = 2 nm = 20 Angstroms.
           "Repeated calibration exposures have shown no changes 
               larger than 3% in the wavelength zero point, 
               or 1.5% in the dispersion."

              Does one use internal lamps for wavelength calibration
              of NICMOS?

      - no mention of wavelength calibration in NICMOS Handbook (???)