Re: "Error floor"

From: Lifan Wang (lifan@panisse.lbl.gov)
Date: Fri Mar 14 2003 - 11:15:15 PST

  • Next message: Pilar Ruiz-Lapuente: "Re: News from the high-z world"

    Would not this error floor introduce unreasonably large errors at the
    late time light curves ? At day 40, the supernova is about 3 magnitude
    fainter, i.e. the flux is about 0.063 times that of the peak. The error
    floor then implies late time errors are around above 0.1 mag and increases
    at later epoch. I don't think this is the correct way of treating the
    data, as the measurement can definitely be better than the error floor.
    Also, if you fit to a pure power law to many of the nebular phase data,
    you actually find that the observations agree nicely for many of the
    nearby SNe. This suggests that for most SNe, there is no justification
    for assuming different errors. It is not correct that we are allowed
    to play with the error because of our troubles with stretch fits.

    What is wrong with just make a cut in date ? We all know that stretch does
    not apply to data in the nebular phase, so why not just accepted it and
    apply stretch only to pre-nebular phase data ?

    Lifan
    > From owner-deepnews@listserv.lbl.gov Fri Mar 14 10:54:38 2003
    > X-Authentication-Warning: listserv.lbl.gov: majordom set sender to owner-deepnews@listserv.lbl.gov using -f
    > Organization: Lawrence Berkeley National Laboratory
    > X-Mailer: Mozilla 4.75 [en]C-CCK-MCD LBNL V4.75 Build 1 (Windows NT 5.0; U)
    > X-Accept-Language: en
    > MIME-Version: 1.0
    > To: "Robert A. Knop Jr." <robert.a.knop@vanderbilt.edu>
    > CC: deepnews@lbl.gov, hstpaper@panisse.lbl.gov
    > Subject: Re: "Error floor"
    > Content-Type: text/plain; charset=us-ascii
    > Content-Transfer-Encoding: 7bit
    > Sender: owner-deepnews@lbl.gov
    >
    > I'm a little puzzled about the choice of 0.007 times the maximum flux
    > point. I guess this ensures that the tails don't wag the main lightcurve
    > fit, i.e. this is a way of giving a little more weight to the data near
    > maximum in the cases were the later error bars are just too good for the
    > errors in the tempate's tails. --Is that the goal? Perhaps we should think
    > of this really as an error bar on the template (which we might know
    > something about) rather than as modified error on the data?
    >
    >
    >
    > "Robert A. Knop Jr." wrote:
    >
    > > I'm about to go through and modify all my hamuy and riess low-z data
    > > files to introduce an "error floor".
    > >
    > > As per our discussion, I will set a minimum error equal to 0.007 times
    > > the maximum flux point in a given lightcurve data file. If any
    > > lightcurve point in that data file has a flux error smaller than this, I
    > > will replace that error bar with my minimum error bar.
    > >
    > > I think this is the procedure we agreed was best. Is there a reference
    > > on why this is a good idea, so that we can skip writing the
    > > justification?
    > >
    > > -Rob
    > >
    > > --
    > > --Prof. Robert Knop
    > > Department of Physics & Astronomy, Vanderbilt University
    > > robert.a.knop@vanderbilt.edu
    >



    This archive was generated by hypermail 2.1.4 : Fri Mar 14 2003 - 11:15:35 PST