From: Greg Aldering (aldering@panisse.lbl.gov)
Date: Sun Mar 16 2003 - 13:26:24 PST
Last night I was comparing Rob's fits to the Hamuy SNe against those
done for P99. The main thing I noticed was that for P99 we let the
baseline float in both B and V for *all* the Hamuy SNe. That is, the
full prescription for the Hamuy fits was:
cut at 60 days
fit B with floating baseline
fit V with floating baseline and vertical scale, but all else fixed by B fit
At the time of P99 we were not as cognizant of the issues in using stretch
at the latter times and/or the idea that SN may have true variations in the
peak/tail ratio. Thus, I believe the concept of the floating baseline was
many to allow for host galaxy missubtraction. I think we were driven to this
because otherwise the fits could be quite poor. Note tha in general the
offsets are less than 0.01*fpeak , but there are some SNe were the offsets
reach 0.03*fpeak. The cases where Rob let the baseline float were usually
cases where the P99 fits showed the largest baseline offsets.
Let me comment further by saying that the template Rob is using does seem
to do a slightly better job for several SNe. It is often very slightly
worse in terms of chi**2, but I think this can generally be traced to the
floating baseline - when Rob didn't include it his fits could be worse than
P99, where it was included. Of course the most common case is that Rob's
fit is worse in B and better in V, because he fit B and V simultaneously.
Here are comments on a few SNe:
93O - Rob's B-fit is much worse than P99; no floating baseline
92P - Rob's B is worse, but V is better; no floating baseline plus
extra data point > 60 days AND and extra point at +48 days
that P99 didn't have!
92br - Rob's B is worse, but V is better; no floating baseline
92bo - Robs' V is much better; Rob let baselines float
92bl - Rob's V a little better; one point after 60 days; no baseline
92bh - Rob's V is much better; no baseline
92bg - Rob's V slightly better; 3 points after 60 days; no baseline
92bc - Rob's B and V much better; ~ 9 points after 60 days; no baseline
(I think this is an example where the new template is probably helping)
92al - Rob's V is better; 7 points after 60 days; no baseline
92ag - Large group of B-band points at 40 days that are not in P99 fits!!!!
Note that the differences in the main fit parameters can be huge
compared to the quoted uncertainties. The color changes in particular
are very scary, being 0.1 mag quite often! That's 0.4 mag after
extinction correction (at least to the extent that the color changes
are not systematically in one direction and then absorbed into the
intrinsic color).
- Greg
This archive was generated by hypermail 2.1.4 : Sun Mar 16 2003 - 13:26:26 PST