Note: the Scattered discussion runs the gamut. There is discussion of K-corrections, the Nature paper, the histogram of Omega and how it's always changing, more K-corrections, the first 35 SNe, the first seven SNe, K-corrections, correlated errors, templates, the use of metaphor in British renaissance drama and poetry, the stretch, extinction, and, finally, K-corrections. I kid you not; all of these were going on at the same time. It was almost impossible to follow. I haven't even tried to reorganize the notes.
Saul was talking on the phone with Brad Schaefer, who "can't tell us anything" but hears that we had a proposal in for NICMOS time... and perhaps we got some time... well, OK, he didn't quite say it straight like that, Saul is interpolating. We can only hope that the interpolation is robust.
Also related to Brad Schaefer, a Venezuela 1m telescope is supposed to go online soon, which will observe 300 square degrees to magnitude 21, which will discovere hudreds of telescopes down out to z=0.3. (What is the time scale for this?) Susanna suspects that the site is not very good; it's on the landward side of the mountains. Susanna also says that the original intent of the telescope was to do macrolensing. Charles Baltay is the PI they say. Yale is putting in a proposal for more buckage; theoretically we are going to see this proposal to give them suggestions.
Next news: the Nature paper was "accepted," although there are comments from both referees. Saul has talked to Science and told them that our paper is accepted. They are apparently going to use the Hubble picture from our other z=0.83 SN (SN9621) to show what a SN at z=0.83 looks like. They will use the Hubble diagram from the other group. Greg is worried to make sure we credit the supernova and the Hubble image correctly for the Science image.
Anyway, the Nature paper is going to take a little bit more work in order to get it out, because of the referee comments. (Peter and others have guessed that the second referee is Nick Sunszetff [sic].) One of the things which could be the most work is separating out the systematic error from the statistical error, including the reddening. Re: things getting too long, Leslie Sage says that we aren't in trouble yet, but Peter is dubious about his ability to count.
Timescale... we'd like to get this out the door (again) early next week. Not like the last ApJ paper where after it was accepted, we sat on it for several months; this time we have the other group breathing down our neck. Sage isn't positive that we'll fit it in before the end of the year, although it would be nice if it appeared the week before the AAS meeting, or even sooner. We shall see; we'll need to get it back to him.
Next question: can we put the paper on the net now that it is accepted? Saul thinks that it would make sense to put a revised version which says Nature accepted on the net. (We should make sure to have unambiguous permission from Nature before that.)
Next 35 SNe, what will it take to convince ourselves that we have something we can trust. It would be really nice to get something out the door... and do we really believe that the answer for Omega is zero.
First, we have to figure out why some numbers have changed. Second thing is the errorbars, are we comfortable with what they are and where the come fromm. Third thing is the template and trusting that we have something we can use. Finally, we will have to reconcile what we have with what we used to find. In particular, we will have to worry about the K-corrections, and how much they change the results of the first seven SNe and our current SNe.
Peter's K-correction table, if he gets it into any kind of shape, we can already test out to see how the first seven SNe change.
Peter says taht in order to do the K-corrections correctly, you need to know the color of the supernova at each epoch. (Peter bluens or reddens (tilts) the spectrum in order so that the integrated broadband colors match what we "know" is right for a SN of the given stretch at the given epoch.) In order to know the color of the supernova, you have to know the relationship between the stretch and the color at different epochs. Peter is waiting for this information from Alex, who still seems to be converging. Peter will set up the machinery to deal with it, and we will set up SNMINUIT's Libengut fitter to see how much the old SNe change.
(The originally problem was that all of the first seven supernovae were on one side (the high side) of a distribution of values of Omega caluclated by each of our supernovae individually.)
Peter and Saul figured out that the change in the K-corrections could give changes of 0.06 to 0.09 magnitudes in the SNe. Also, one SN which was rejected as having an unreasonable stretch (too narrow) comes back into being usable with the new K-corrections. We also want to go back and worry about the Malmquist bias.
Reynald mentions that we also need to worry about the Malmquist bias on the nearby data. The other group claims that there may be a huge Malmquist bias on the nearby data. Both groups are using the same nearby data, so that is a point which is the same between the two projects, so both might get Omega wrong the same way for that reason.
Saul wonders if perhaps the Science reporter should be told that the low redshift stuff is a place where the biggest question mark remains.
Reynald suggest doing a two parameter fit using just our data (not the nearbys). This should give a bigger errorbar, but should be less sensitive to Malmquist bias, and independent of bias in the nearby SNe. It would be worth doing, in any event.
Reynald and Susanna mention that other studies of Omega (not including Lambda), e.g. of Baryonic matter, get between 0.2 and 0.4. Greg points out that these people have a galaxy bias fudge factor which can futz thigns up to a factor of 2.
Malmquist correction: extreme thing to do is to pretend that all of the nearbys are maximally biased. Change all of their magnitudes by 0.1 (make them dimmer) or even 0.17 (a whole sigma). That would change our value of Omega (in which universe) by something like 0.4! The effects should be less than this, but this is a limiting case.
Another idea is to take the Hamuy set; throw out low-z ones affected by peculiar velocity; throw ought higher-z ones which will probably be more affected by Malmquist bias. The problem with this was that the search was done in a bunch of different runs with different charateristics.
However, first, of course, we need to worry about differences when we reanalyze the old data.
Note, first seven SNe and the Nature paper. What if we only use the ones which have spectroscopic confirmation? How much does this change things? (Chew on that for a while.)
Better K-corretions, should come in in the next paper after the nature paper is the one that should introduce these new, better systematics.
Change in the histogram between August and October. A couple changed by 0.1, one changed by 0.2 (which one?). These went in the direction of getting fainter. Several got brighter because of including extinction in our galaxy. Gerson and Rob are going to try to track some of this down.
Peter has a routine written to interpolate K-corrections; given a time and a color, he spits out a K-correction. Alex's template program will use this.
Alex says one problem is that the matrices are getting so big (e.g. 600x600) that when he does the inversion (with SVD via IDL or one that comes from Greg) there is precision loss. For peculiar velocities, Alex has been putting in 600km/s as 1sigma due to this on the redshift; all the points of one supernova have giant correlated errors, which can give negative determinants, and which is causing trouble. Popular opinion seems to be that 300km/s would be better. We shall see if this helps things.
(This meeting is degenerating in to a very fragmented collection of little discussions and factoids.)
Where are we? Templates. Right-o. Greg is working on speeding things up, Alex is working on getting rid of little troubles that he's having. What Alex says is that he's working on including the nearby ones so that we can really get the colors of the supernovae.
I assume that everything is going to work perfectly by Friday.
-S. Perlmutter
Saul is worried about Greg and Alex keeping in touch after Friday and that things will continue to work smoothly.
He's a smart guy.
-A. Kim about G. Aldering
We seem to be back to talking about K-corrections. (This meeting has been all over the place.) The issue is how easy it would be to patch Peter's K-corrections into the current Libengut fitting program. The danger is that the new K-corrections are stretch sensitive, whereas the program has the implicit assumption that the K's are not stretch sensitive, so it's not clear how easy it would be to patch in the new K-corrections. Various methods of doing this are under discussion. It's a priority.
Peter says that below z=0.5, you stretch the K-correction. (Well, we've oscillated between stretching the template and not stretch the K, and stretchign both.) Peter says that above z=0.5, you want to divide by the stretch instead of multiplying.
Huh?
-M. Kim and R. Knop
This, I guess, is an approximation to what will eventually be Peter's real K-corrections. (Peter later gave a brief explanation of why this would happen.)
Alex points out another thing: time-dependent R values. Alex says it's already there? In any event, we need to worry about it.
Meta. This is segueing into a discussion by Matthew about overall organization. Mostly, the issue is documentation: making clear what is being done at each step, so that people don't make assumptions about what happened that is wrong. It does need to be written down.
Sigh.
Right. Peter says that all you need to know right now is that for any given spectrum, the K-correction at a given redshift depends on the observed color of that spectrum. (Right features, wrong tilt.)
Next issue: our set of 35 SNe, we still do want to track down everything we can about the host galaxies. Do we have a subpopulation of ellipticals and/or spirals? Robert Q. has started working on this. There are still clearly problems with the photometry, but he's plotted color of host (not K-corrected) vs. stretch of supernova. There does seem to be some sort of trend. There are reasons to worry about the R-I sign, and things are all preliminary; however, right now it looks like the trend is _backwards_ (higher stretch in redder (elliptical) galaxies). Work is proceeding on this.
Another question: did anything hear anybody from Mike about the clustering. Apparently he was doing stuff for Gordon, and has been stuck for a couple of weeks.
Greg and Robert will also look through the images to see if you can pick out morphology of these galaxies from our data.
We need a catalog of hosts... perhaps coadding images would help this as well. (That would be the next step; the first step is to just use the best image we have once the SN is gone.) Perhaps Patricia will work on this.
Back to the top: new student from Portugal is here, Patricia Castro. She's here for 10 months, doing her master's thesis work.
While Patricia is building this, Greg can look at the images to establish morphology.
The run starting. Who's going when where.
CFHT run, coming up in <2 weeks. Susanna is going to run over there, fly the flag, and help get the data back. We're going to see what we can get done without brinigng the rest of the group to its knees. We need to get those reference images in so we can start knowing what their data look like. (We also have to get APM/APS fields for this.) Then we have to deal with the new images as they come in. We will perhaps, 3 weeks later, do spectroscopy followup at CFHT with the same team who will be doing spectroscopy followup of their stuff (we throw in a couple of additional slits to multislit).
Question: do we want to try to do a one-day turnaround on this? Scary.
Runs and funding; do we have to buy our tickets through Sato? It's not clear. Is CfPA still funding? It's not clear. The lab rule seems to be you must go through Sato for Domestic flights, so perhaps it doesn't matter for international flights.
Minimum configuration:
Additional notes on this: Maybe Sebastien will also go to CTIO at the end of December. Maybe Greg or Susanna or somebody will go to Keck instead of Saul. We will see if Chris Lidman can go to CTIO on Jan 21 along with Peter.
Dinner to say good-bye to Alex: place on Telegraph, crossstreet M.... well, hell, Alex will send us an E-mail telling us the deal.