From: Robert A. Knop Jr. (robert.a.knop@vanderbilt.edu)
Date: Thu Feb 27 2003 - 11:38:07 PST
Here's some quick-n-dirty data on the dispersion of B-V at Bmax.
Note that what I've done is very different from what Serena did in her
paper. There, she was looking at colors on specific days. I'm looking
at the results of colors from lightcurve fits at t=0. Quite different
things.
First: take *all* low-z supernovae used in the HST paper (Hubble flow
Riess + Hamuy). Apply K-corrections, and extinction corrections basedon
the E(B-V) from my fits. The color dispersion I get is 0.09 (a bit more
than Serena's day 0 dispersion, but remember that I'm doing a different
thing).
Hamuy + Riess: dispersion is 0.09
Hamuy alone: dispersion is 0.04
Riess alone: dispersion is 0.14
==> Serena already sort of discusses this in her paper, saying that
there is evidence for a systematic difference in the two sets. Table
7 in my paper already suggests this. The systematic difference may
be worse for lightcurve fits than for Serena's "just look at the
colors" method. Anecdotally, I saw that the Riess set didn't as
often give good fits to the lightcurve template. Part of this could
be extinction: a number of the Riess supernovae were *very* reddened;
with the reduced number used in the paper from the Hubble flow, they
have a bigger effect.
Finally, I looked at the dispersion throwing out the most egregiously
reddened supernovae: I threw out everything with either the Milky Way or
the Host E(B-V)>0.1
Hamuy + Riess : 0.03
Hamuy alone : 0.03
Riess alone : 0.03 [ only six events left ]
The great improvement when throwing out the most egregiously reddened
supernovae lends some support to Greg's argument that there may be only
a very small intrinsic color dispersion (though I'd be surprised if it
were truly strictly zero), but that there is a dispersion in the
reddening law (and/or we have a systematic error in the reddening law).
Even though this isn't where I got the number, this also justifies my
exrectum 0.03 that I used for intrinsic dispersion. However, it may
also be worth trying fits that use 0 intrinsic dispersion (at least in
B-V, perhaps sticking with something in U-B) and an uncertainty on Rb.
-Rob
-- --Prof. Robert Knop Department of Physics & Astronomy, Vanderbilt University robert.a.knop@vanderbilt.edu
This archive was generated by hypermail 2.1.4 : Thu Feb 27 2003 - 11:38:42 PST