From: Greg Aldering (aldering@panisse.lbl.gov)
Date: Thu Mar 20 2003 - 14:30:44 PST
Hi Rob,
Yes, it is likely that the P99 K-corrections and intrinsic colors are
not internally consistent. I know that for Bmax-Vmax Peter and I looked
at the Bmax-Vmax of lots of local SNe and I chose a color offset
(using Peter's slope) which hugged the blue edge of the distribution and
giving most weight to SNe in ellipticals. This was not forced to agree
with the K-corrections. I don't recall how the U-B color was determined.
Regarding R_R, take a look at
/home/astro34/aldering/njnunes/minuit/total_select_extinc.pro
To the best of my recollection, this is the program that was used to
assign R_R to each SN. It is based on calculations done by Peter -
perhaps he can recall in more detail what procedure was used. I think we
can guarentee that the light was treated as red light, not as blue light!
This effect surfaced late in the analysis, but was discussed pretty
thoroughly at the time. But, I agree that the effect seems awfully large
in retrospect.
Greg
On Thu, 20 Mar 2003, Robert A. Knop Jr. wrote:
> The "extinc" column from the old gersontable. I am having trouble
> getting that number out of the current FSD E(B-V) values that I have.
>
> In particular, there is a statement in P99 that I have questioned before
> and which I shall question again, as I simply do not believe it:
>
> These values of R_R range from 2.56 at z=0 to 4.88 at z=0.83
>
> This is A_R/E(B-V) due to various different supernova spectra, used for
> calculating Galactic extinction.
>
> This CAN NOT be right, and I hope it's not what we really did -- if so,
> then we biased ourself seriously at high redshift.
>
> Think about it. We're always talking R-band light... whatever the
> sueprnova's redshift, by the time it reaches our galaxy, it's done
> redshifting and is at whatever wavelength we will observe it at. As
> such, the light being extincted by the galaxy's dust is all red light.
> As such, you wouldn't expect a value too different from 2.56. Now, it
> can vary, because the shape of the spectrum (i.e. the weighting of the
> different parts of the R filter) will vary a lot depending on how much
> it's been shfited to the red. BUT, it's completely ridiculous that you
> might get an R_R from *any* weighting across the R band which is
> *greater* than the quoted R_B=4.14. Unless extinction laws are highly
> non-monotonic, there is *no* weighting of the effective R_lambda across
> the wavelenghts of the R filter that will give you a larger R_R than
> what you get from taking any reasonable average across the B filter.
>
> The only way I can see this huge range happening is if you *deredshift*
> the filter before calculating the extinction, rather than *redshifting*
> the spectrum. Physically, though, the latter is what happened, and the
> latter is the right way to calculate your extinction. We are observing
> red light, and it's red as it goes through our galaxy; the dust in our
> galaxy does not care that that light may have been blue when it was
> first emitted five billion years ago.
>
> (Obviously, host galaxy extinction corrections should be done using the
> z=0 R_B value, since there the light *is* blue, and the dust in *that*
> galaxy doesn't care if by the time we observe it it's red.)
>
> If we did what we said we did, we've got pretty serious problems in P99.
> At least some of the differences I'm seeing in my mb numbers compared to
> the P99 mb numbers come from the galactic extinction, and the difference
> between O'Donnell and Cardelli laws aren't enough to explain it. The
> error in the procedure I describe would make a pretty big difference,
> however.
>
> -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 Mar 20 2003 - 14:30:45 PST