Saul isn't here yet. Everybody keeps trying to make somebody else go get him.
We're now deciding that we're all going to take the name Alex. There are a huge number of people here, some new, so we're going to do the whole introduction thing. We've got Michael, a grad student who's considering joining the group. Brenda Frye, is a new post-doc as of yesterday. She's now going to do all the work while the rest of us retire to Tahiti. For the first month, though, she'll still be finishing her thesis back on campus. Tom York is also new as of two days ago. He's a Physicist who's going to be with us at least through December. He's from Ithaca NY, although he's not really connected with Cornell at this point. He's formerly of high energy experiment, netrinos and so forth.
The rest of us then introduce ourselves.
Brenda is going to start by looking into the HST data, and the problems having to do with charge transfer, and so forth. She's also probably going to be dealing with analysis of nearby supernovae as they come in.
Tom is still just trying to look around and get some idea of what goes on in the group. He's going to try to look into the question of dust and how it will affect our supernovae at vast distances.
Susana tells us that she sent the final plans for Chew's Ridge down.... she hopes to hear back soon, so that we might actually be able to move the robotic telescope to the new site. (Well, building a building first.)
The NY Times reported on some meeting in Chicago. Saul is telling us about it. It was a meeting mostly of theorists. They had Bob Kirshner, apparently, talking about the results of the supernova stuff. Saul said he talked to Alan Guth. Of the inflationary theory people, some people say that it doesn't matter what the data says, they were right (of course), and this just confirms, but anything would confirm. Others are holdouts saying you need a flat universe. There was also, apparently, much highly abstract discussion about things like what happened before the Big Bang (and here I always thought there was no before). Most of this seems to be irrelevant to anybody who has anything to do with actual data.
Saul is still trying to talk to people to find out if there was any actual content that was relevant to us.
Data Schmata
--R. Knop
It's February. It might even be submitted this February. Gerson gave a talk on it last Friday, and he'll give us some highlights.... He has the Hamuy data compared to our composite lightcurve. One thing is that the errors are so small that he can't get a good chisquare. Saul wants to know whether the near neighbors are scattered by the errorbar, or by more, so we can evaulate how good the errorbars really are.
Gerson also says that the outliers in stretch, when plotted on a residual plot (i.e. difference between data and curve), show some real systematic deviations. This only really happens for stretch less than 0.8 (one at 0.5, two at 0.7); basically all of our data has s>0.8. Peter says that for narrow stretch supernovae, there are huge titanium lines that are probably what is mucking up the stretch relationship. Those lines do not exist for the broader stretch supernovae at all. (The ones at stretch 0.5 are in a very different ionization state than the ones at stretch 0.8-1.2; Peter guesses that the Titanium lines will start to become important below stretch of 0.75, but I think that number is off of the top of his head.) Conclusion: it seems to be reasonable to limit ourselves to 0.8-1.2.
Greg is raising the issue that he's worried about the fact that the chisquare goes to hell if you fit to weighted averages of points instead of individual points. Don doesn't think this is a problem, and Rob and Peter are arguing that the correlations between points mean that combining things in a weighted average leads you to underestimate your errorbars.
Don is going to talk about some template stuff as well. He's been messing with the template as well. He shows us some warts added to the template, which it looked like the data wanted. His template is zero, then a parabola (in flux) going up to -9 or -10 days. Then there's a spline fit out to 50 days, using 9 points to draw the splie through. Then at the end there's an exponential tail. He's plotted his thing on top of the Leibengut template, and we see the warts there.
He's done this just fitting our data; he hasn't thrown the Hamuy in. Saul is concerned, saying that we ought to make sure it works with Hamuy as well. Don's worried about systematic differences in the Hamuy, and systematics in the data. He wants to redo all this when we have finished our nearby deal that's currently trying to kill all of us.
Saul wonders why there's two spline knots right at -10 days. There are also two knots at 50 days. Don says he wanted to constrain both the height and the slope at the two ends. The IDL routine he was using didn't let him do this in an easy way. So, he put in these four in to constrain those two things. Then, when he was doing his fit, he only let the other five vary. At each point, he matches function, derivative, and second derivative. (There may only be a discontinuity in third derivative and higher.) (See two paragraphs down.)
The fits were all done to the unaveraged complete dataset; Don is showing us plots with 0.5 day averaging. He plots new template, data, and Leibengut relative to the new template. You can see the two "Gerson" warts. He played with putting the knots various places to minimize the data residuals.
Don also wanted a smooth curve, not something wiggily like from Reiss Press, and Krishner. So, he looks at the thing and tries to avoid huge excursions in the curvature. The curvature is allowed to have a discontinuous slope at a knot.
Matthew wants to know the requirement at each knot. Don says that he definse the positions of the knots. (He did play with numbers, and needed at least five.) With the six knots, he adjusted the amplitudes to minimize the sum of the squares of the residuals. (I.e. chisquare.) All of these techniques are similar to what Don used in order to create the K-correction spline curves.
What Don created is in good agreement with what Gerson created. Don says he didn't like Gerson's because there was a curvature glich at the end.
Don points out that there is a difference in height by 0.3% between his and Leibengut's curves. There is also a shift in time 0.2 days earlier. Don says that this is because experimentally, the time of maximum is not well constrained. Don says that he has a hard time attaching an error to this. Also, there are problems saying the number of days before maximum for the explosion, since wrangling the curve around makes the maximum ill-defined. Don suggests using the half-height position as a time reference, since it is better defined.
Saul wants to see a two-point correlation on these templates. Knowing the brightness on one day, how correlated is it with the brightness one day off; i.e., how smooth should the thing really be? Saul wants to get this numerically.
There are open ended things associated with this that I can work on forever; I have other papers to get to.
--D. Groom
(Words to live by.)
Gerson says that he does see these sort of correlations for the Hamuy data, but nothing obvious for ours. However, to really know it, we'd probably have to put all of our data in. Peter points out that you have to put in the observational correlations if you want to get the right answer on this.
EROS is taking the first search images tonight. They'll be going for the next two weeks.
Greg has added a lot of things to the nearby page. There is a guide to spectroscopy and photometry, and there is information about how to report supernovae. Alex Kim had the very same day sent questions....
Greg says that we're 95% ready to deal with a supernovae as it comes in. He and Robert are working on an automated scheduler.
Greg says that since last week, in conversation with the NEAT people at JPL and Shane Burns, it looks like the NEAT search won't happen, certainly in February, and probably not in March either.
We're starting to get data from Spacewatch. They'll start going back in a month.
Rob's been looking at the KPNO 0.9m Mosaic data. Greg says look at the nearsearch page for background. Rob already found that there was a coordinate error in some of the data on the first night. Greg says that with this finding, we've already proven our worth to them.... Greg thinks that all of this is going smoothly. Greg will talk again to Chris Smith on February 18. We've set up arrangements for this disk shipping thing. (They're going to send back an 18GB disk each night.) We may get data in either raw or reduced format.
Saul wonders if we should send anybody to the mountaintop to make sure that disks are getting written and so forth. Saul's worrying about that having enough priority. Greg doesn't think they would "not get around" to writing our disks because they are too busy, but he can't rule it out. We'll see.
We do finally have YALO time, a whole lot of it. 14 nights from Portugal, and some from Brad Schaefer. There's a lot of confusion on the Portugese end about how things are going to work, but Greg thinks he may have finally cut through all of that, and has found somebody he can talk to who sounds very flexible (Charlie Baylen, although I have no clue if I've spelled that right). There are problems with YALO; they have a Stromgren U filter, not a Johnson U filter, but the S/N isn't as great. Also, the R filter is a wideass bizarre thing that's nothing like Harris R. It's really R+I. It's going to be interested dealing with all of this stuff in different kinds of filters. It will be a way to test the K-corrections.... We will have to be very careful about how we put the YALO data in with everything else. Greg thinks this means we're going to have to get redundant B-band coverate, i.e. don't lean entirely on YALO on any given night.
Greg says that we didn't get all the time we had hoped for at Mt. Laguna.
Greg believes that most of the scheduling information has been settled. He's figured out who's going when, and there are opportunities for people to go along on various runs.
Alex Conley is building more evidence that the published geometric distortion for WFPC that he's been using is wrong.
Alex Lewin is starting to get snminuit \emph{et al.} to work with the other group's data. She hasn't done any fits yet. Gerson questions if we have their correlations, which we don't. Saul says, however, that he thinks the other group didn't use correlations.
The CCD project got an NFS grant for $225,000/year over 3 years, plus $75,000/year of matching funds from Shank. At last this project, which has been anemic, will start to be able to do some of the things we've been wanting to do.
H's plotting his efficiency curves, the modelled quantum response. Don has now added in modelling the absorption of Si as a function of temperature. The band gap increases as you cool things, meaning the high wavelength cuts off. The red edge of the efficiency curve has moved down about 55 nanometers. This coefficient should be negligible for even the I-band filter, but will be very important for the Z-band.
Don notes that normally when people do z-band, they're using the CCD cutoff, and it's a very narrow sort of deal.
Don overlays a standard Site CCD, from some paper or another. These look to be pretty red... 20% out at 1 micron. However, this is at room temperature. There is still a 55nm shift when you cool the thing, so it's not quite as good as it looks. (Also, these thin CCDs have fringing out in the I and Z bands, whereas Don's CCDs will not have this.)
Don notes that when we're using Z-band, we're going to have to do a CCD temperature correction, and we're going to have to keep track of that in the data headers. He also says don't believe any fool's corrections, but rather we should get data to measure the effect.
Peter is doing most of the HST morphology stuff, working a little with Richard Ellis. He's making a list of 45 supernova host galaxies that we definitely want to do, and half a dozen others that we'll see. There is no way to prioritize the list; they'll do whichever ones the coordinates work for. He's going to do 3 dithered 500s exposures. (This is all in "snapshot" mode.)
Richard has lost his office, computer, and E-mail, apparently, so he wants to talk to Peter on the phone. The thing is due Feb. 18, and Peter thinks they are on track to get it in.
The only other thing to consider is whether we want to double up, by putting extra exposurs on 9784 (say in lieu of the first 7), to see if we can get a host with STIS (which should be better than WFPC).
Saul wants somebody to touch base with Doug on our other proposal to make sure we don't have to put in a placeholder for that. (This is the 20 orbits or so to do z>1 supernova.) The other question is if there is any budget thing we have to do this week. Saul tells Peter to ask our coordinater to find out.