From: clidman (clidman@eso.org)
Date: Wed Jul 23 2003 - 14:23:32 PDT
Dear All,
Please find enclosed a preliminary summary of the teleconference we had
on Monday. I hope that I have been able to summarize accurately all that was said. If you
find an error or an omission, please send me an e-mail and I'll correct the summary.
I have taken the liberty of expanding upon the discussion we had on how the equivalent
widths are defined. We had discussed this issue briefly, but I fear that we did not
dwell on the core issue long enough.
Isobel, I am sorry that I got your e-mail address wrong. It is now corrected.
Cheers, Chris.
Plots against absolute magnitude
================================
The parameter measuring a particular spectral feature should be
plotted against absolute magnitude and not Delta m15 or stretch.
We recognize that we may introduce additional scatter by using
absolute magnitudes, because of
- uncertainties in the distance estimates for SNe that
are not in the Hubble flow. (z < 0.014)
- uncertainties in host galaxy extinction corrections
However, the purpose of the paper is to discover new correlations,
and, in particular, those correlations around maximum light that can
be applied to observations of high-z SNe.
Additional Correlations
=======================
In addition to the correlations that are listed in the Gaston's
supplement, others should be tried. For example:
Absolute magnitude versus 7 + 4 + 6 + 2
Absolute magnitude versus 7 + 4 + 6
Absolute magnitude versus 4 + 2
The latter would be the most useful for high z SNe.
These EWs were chosen because low-luminosity SNe have large EWs at
maximum light and high-luminosity SNe have low EWs at maximum light.
EW definition
==============
There are two issues here:
1) The current definition of the EWs is not explained in enough
detail. The results have to be reproducible and for that to occur
other researchers will have to know how the equivalent widths are
defined.
2) The current definitions are controversial because:
i) The wavelength ranges over which the features are measured change with
phase. This is highly non-standard.
ii) The elements responsible for the features change with phase.
We briefly discussed this point at the teleconference, and I am
expanding upon that discussion. It is a point which is fundamental to
the paper and it will be an issue when the paper is given to the
collaboration for review and it will be an issue with the referee.
There are two possible ways forward.
a) We continue with the definitions as they are and argue that this a
purely empirical result.
b) We use a more standard definition of the EW and allow the wavelengths
that define the features to move over a small range of wavelength and not
drastically as they do for features 3, 4 and 7. This means that we abandon
the attempt at measuring these features after a certain date.
Additional Supernovae
=====================
It may be worth reviewing the literature to see if additional SNe can
be added. The list in Table 2. is the same set of SNe that were used
in Peter's '95 paper.
Comparison with Past Work
=========================
There is insufficient discussion or comparison with previous work on
correlating spectral features with SNe luminosity. If we do not show
that we can see correlations that have been noted in the past, then
people may ignore any correlation we do find. This is a critical point.
The following correlations have been noted in the past:
- Peter's R(Si II) and R(Ca II) ratios.
- The blueshift of SiII and CaII
Peter also noted other correlations in his '95 paper.
Statistical description of the correlations
==========================================
A statistical description of the correlations is required. This
includes correlation co-efficients and the scatter about any relation
that is fitted to the data.
Errors
======
There was a comment (attributed to Don Groom) that the errors in the
EW measurement were perhaps too small. Gaston should discuss this
directly with Don.
Correcting the indices to a common epoch
========================================
The spectral indices are generally measured over a range of
epochs. For example, the indices in figure 3 of the supplement are
measured from +/-3 days. Since we have an average measure of how these
indices evolve with time, we should be able to correct the indices to
a common epoch (in this example, the epoch of maximum light). This
may reduce the scatter in plots such as those shown in the supplement.
Multi-Parameter Correlations
============================
We should be aware that multi-parameter correlations might exist.
This archive was generated by hypermail 2.1.4 : Wed Jul 23 2003 - 13:21:56 PDT