[News] Potential Data Problems

[Dr. Mark Neeser]

Hi Roeland,

The answer, I think, is quite simple.  Those who claim
fringing is easily corrected have only ever corrected
data obtained under good conditions.  Those who claim
that it is very difficult have tried to fringe-correct
data obtained during variable sky/non-photometric
conditions.

In how far the AW pipeline is willing to address the
latter conditions is open for discussion.  Fringe correcting
under poor conditions can be a lot of work for relatively
little gain.  Since we have a service observing instrument,
it is easy to stipulate that I and, particularly, z' bands
only be observed during stable conditions.

Some points regarding Roberto's comments:

    Yes, the fringing pattern IS stable in SHAPE and POSITION.
    It has to be, since the pattern is intrinsic to the detector.
    It is the varying thicknesses of the CCD strata created during
    the thinning process that creates constructive and destructive
    interference akin to a fabry-perot.  What is NOT constant is
    the intensity of the fringe pattern.  Also, this intensity
    variation is, unfortunately not constant across the pattern.
    Why?  Two major component of sky emission, for example OH and O
    emit at differing frequencies.  Some part of the chip will have
    a thickness that is an integral number of the wavelength of OH,
    say, while another part will a half-wavelength off.  Therefore,
    the former will be brighter, the latter darker.  But, O, and
    other night sky lines have different wavelengths, so, they will
    produces bright and dark features at differing positions along
    the fringe pattern, particularly since the lines are do not
    vary in sync with one another.  Therefore, a simple normalization
    of the fringe pattern does NOT work under variable sky conditions
    (as I stated in my last e-mail).  The OH and the O must be scaled
    separately as they lighten/darken the fringe pattern at differing
    positions!

That said.  I would suggest that we use Roberto's simple scaling
procedure, and limit I/z observations to stable conditions.
Perhaps, at a later time, if the need arises, we can refine the
de-fringing procedure with more advanced scaling routines.

Cheers,

Mark






On Wed, 11 Jun 2003, Roeland Rengelink wrote:

> Hi,
>
> Re: the de-fringing algorithms
>
> I'm slightly worried about the conflicting reports on:
>
> 1) The stability of the WFI fringing pattern
> 2) The complexity of the required algorithm for fringe-subtraction
>
> Does anybody understand why different groups get different results?
>
> Cheers,
>
> Roeland
>
> _______________________________________________
> News mailing list
> News@astro-wise.org
> http://listman.astro-wise.org/mailman/listinfo/news
>