From: Chris Lidman (clidman@eso.org)
Date: Sat Mar 06 2004 - 12:12:36 PST
Hi Greg,
Thank you for your clear constructive comments.
The classification of the candidates in the paper is based on
the VLT spectra alone without any input from other observables
whatsoever.
I agree that many of the unknown candidates could very well be
Type Ia SNe, but from the VLT spectra alone, Andy and I cannot
tell. I have no objection in including lightcurve and/or other
information that could help us with the classification. If we
choose to do this, then I propose that I re-label column #3 in Table 4
with the phrase "Spectral Classification" and add another column, or
use the "Comments" column, to add additional information.
It is difficult to estimate how many would have light curve
information, but my guess would be that it would be somewhere between
half and three-quarters of all the supernovae in the paper.
One could add information from the search itself, e.g. what
was the percentage increase in the candidate, the significance,
etc. It is a point that we can discuss during next weeks
teleconference.
The success rates of the different runs is interesting and is
a reflection of the different search strategies. Searches that
targeted z~1 SNe are much less successful, if one uses the
percentage of candidates that were spectrally classifiable as a
yardstick for success. For these runs, the S/N ratio of most
of the spectra are very low (some are below 1), so all the spectra
may very well be consistent with a Type Ias. Indeed, if I were
to turn the question around and ask myself how many candidates
are not clearly not Ias, then the answer is probably less than 10 out
of 40 candidates.
I have not discussed the search strategies of each of the runs
that are listed in Table 1, but this is something that we
could add to the paper, although I suspect that different
people may have different ideas as to what each of search
strategies were.
In regards to individual candidates that you note in
the paper, I will revise the classification. I expect that
different people will have different degrees of confidence
to the classifications that Andy and I have given.
At this stage, I'd like to keep the "spectroscopic"
unknowns in the paper. If this can be done by adding
a little bit of additional information, then I think that
the end result will be a better paper.
Cheers, Chris.
On Sat, 2004-03-06 at 00:06, Greg Aldering wrote:
> Hi Chris,
>
> On my flight over to Hawaii I read you paper on the VLT spectroscopy.
> Overall I found it well-written and easy to understand.
>
> My one main concern centers around the objects which are "unknown."
> First, there are a very large number of them, suggesting that we are
> taking spectra of spurious candidates. While this may be true in for a
> minority of our searches, reporting it is clearly bad for business and
> should be considered carefully! If the "unknowns" are objects which
> were finally classifiable at our other telescopes, this should be noted
> (maybe with a comment on why the VLT spectrum was unable to produce a
> classification). Given the way we work at the telescope it is quite
> possible we stopped the observations early because of external
> developments.
>
> In other cases of "unknowns" it would be very important to report on
> whether lightcurves confirm the reality of the candidate and whether it
> is a likely SN. Then the paper itself should address whether or not the
> VLT spectrum is consist, e.g., is it simply too low S/N or is the
> spectrum inconsistent with the lighcurve (suggesting a pointing error)?
> In particular, the cases where an IAUC name is assigned, but where we
> report the type as "unknown" will be very confusing to readers (and could
> set-off Dan Green).
>
> So, although the paper does exactly what it says - reports onf VLT
> spectroscopy of SN candidates - the broader context can not be ignored.
> A simple alternative is to not report on the "unknowns" at this stage,
> or if they were confirmed elsewhere.
>
> Overall, I expect this will be a nice paper!
>
> - Greg
>
This archive was generated by hypermail 2.1.4 : Sat Mar 06 2004 - 12:13:12 PST