From: Robert A. Knop Jr. (robert.a.knop@vanderbilt.edu)
Date: Fri Mar 14 2003 - 05:49:17 PST
On Thu, Mar 13, 2003 at 06:25:53PM -0800, Andy Howell wrote:
> Forgive me for only noticing this now, but your template on p. 9 has the
> V reaching maximum at
> day 1 relative to Bmax. While there is a large distribution in
> differences between day of Vmax and Bmax,
> it is more typical for there to be a two day delay. For R there may be
> an even bigger difference, but you also have only a 1 day delay. Since
> you are doing simultaneous fits, if you are systematically off by a day,
> I would expect this to lead to a systematic error in stretch (at least)
> that could bias the results. Gerson's V template peaks at day = 2.
Well, it is what it is. If you plot the templates (see attached file),
you will note that B and V are very close in rise, but that V drops
slower; in other words, V is "flatter to the right" than B, which
effectively gives you a little slope (to that direction) in the dates
that will be near maximum. It also means you'll get a different day of
max-- probably later-- if you throw out the top couple of days from the
peak, and just extrapolate the few days before and the few days after to
see where it looks like they'll join.
As for a systematic error in stretch -- I'm not worried about that.
Even if my peak day (i.e. day where V-band magnitude reaches peak) is a
day too low, it's incorrect to think of it as the whole template being
shifted a day to the left. Thus, even if the peak of the template
doesn't show up in the expected place, the whole rest of the template
won't be systematically off by a day. I generated the template by
fitting it to the low redshift data. As such, the declining part ought
to still be in the right place (as that's where most of the data is); if
the peak day is off, it means that the shape of the lightcurve near the
peak is off, not that the whole thing is shifted.
More importantly, though, was the way I generated the template: I fit
the B and V low redshfit data simultaneously to a model which included
the Gerson-B template and a bunch of spline knots that allowed the V
template to vary, but which used the same stretch. Each supernova was
only given one stretch; as the B template was fixed, that would control
what stretch was given the supernova. Therefore, the resultant V
template should correspond to a s_B=1 supernova. Gerson did V all by
itself, so there's no guarantee that his s_V=1 is the same as his
s_B=1. You need to keep this in mind when comparing my V template and
his.
-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 - 05:49:19 PST