From: Chris Lidman (clidman@eso.org)
Date: Sun Aug 22 2004 - 12:03:21 PDT
Hi Alex,
Thanks for the comments.
On Sat, 2004-08-21 at 19:41, Alex Kim wrote:
> Chris,
>
> I went through the VLT spectroscopy paper. Just a few questions.
>
> You mention that S/N is calculated for each spectrum but this is not
> shown in the plots. Any reason for this?
>
It would not add much if they were included, so I did not put
them into the paper. If someone asks for the data, I will give
them both the SN spectrum and the error spectrum.
> The plots show supernova spectra corrected for host-galaxy
> contamination. Was the host galaxy determined from the Howell-Wang fit
> or from spectral reduction?
>
This is done from the Howell-Wang fit. In most cases, the SN is heavily
blended with the host, so it is not possible to separate them with
simple extraction techniques. There are a few techniques that
can spatial deconvolve 2-d spectra and this is something that might
be worth investigating in the future.
> In section 5, where you discuss the challenge of spectroscopic
> confirmation for z>1 or in SN2002gj part of 3.1, it may be nice to say
> (if it is true) that although the Ia* spectra alone are not {\it
> sufficient} to type the supernova, they are {\it necessary} for
> unambiguous identification. The reason I mention this is because
> someone could look at figure 2, see only one Ia at z>1, and decide that
> it's not worth alloting more VLT time for high-z sn observations. But
> if it is true that spectra can unambigously type the Ia* in conjunction
> with the epoch from the light curve, it will be easier to justify more
> observing time.
>
This is a good question. It is true that the spectra of high-z SNe
are sometimes not sufficient. However, there is one case in this paper
and there are several cases that were observed with Subaru and
Keck, where the spectra is sufficient.
Are they necessary? If one requires an accurate redshift and if one
wants to use them for cosmology then my answer is yes. For the
purpose of precision cosmology, a photometric redshift (if you can
detect the host) is, in my opinion, not good enough.
If one just wants to identify the type, then one needs
a well observed light-curve in several colours. It is not clear
to me how well one can type SNe from the photometry. This is something
that Essence or the SNLS will tell us.
I will think about this point a bit more and how it might be added to
the paper.
Cheers, Chris.
> Alex
-- European Southern Observatory Alonso de Córdova 3107, Vitacura Casilla 19001, Santiago 19 CHILEPh. +56 2 463 3106 FAX +56 2 463 3101
This archive was generated by hypermail 2.1.4 : Sun Aug 22 2004 - 12:03:44 PDT