Peter did Monte Carlo simulations for a possible nearby search. He wanted to look at what the effect of the reference/search gap would be. He assumed a limiting magnitude of 22.3 in the V band, required a 15% increase (note that this isn't quite right, because he didn't consider the host). He put in some distribution and photometry error, but ignored reddening. For the 7-day case, you never find any more than a day or two after max. However, you sacrifice all the others. Right now, the 7-day gap gets everything before max that the 21-day does, but putting in the host will probably change that.
The search was volume limited up to about 0.25. The gap cut didn't effect the redshift cut at all (because of the fact of being volume limited).
In other words, we're going deep enough for the redshifts we're interested in that 7 days is a good enough gap, modulo things like stretch and putting in host galaxy light over which we have to detect the supernova.
Thought: do two refs, wait a week, do two news, then hope that there is an instrument change so that two days later we can get spectra.
Nature reprints: $145 vs $1225 for 100 color reprints. It goes under the heading of color regardless of how many pages of color. 1000 copies would be $590. We will have to query all the co-authors asking how many they want. We'll also talk to the Nature preprint people about using our black and white image instead of a black and white converted version of the color.
Next issue, if we can get Branch's piece added in, it's another 1-2 pages. One question is if we can get it added in rather than as a separate deal... but Peter says it is cheaper to order them separately.
Issue of passing out our figure from the poster with the omega-lambda confidence from all 40 supernovae. Should we let a figure go into Science? One suggestion is to let one go which is marked heavily to be preliminary. The idea is that no journal is really going to reject it, and that the numbers will change a little bit by the time they show up in a paper. Gerson notes also that Science can report on a conference, and that this was shown at a conference.
Peter says that we should only do this if all authors agree to the wording in Science. Of course, timing would be a problem getting the information back. Peter volunteers to be devil's advocate and object to things.
Saul shows the figure (confidence region on omega_lambda-omega_m plane). which we showed already, with the dotted lines for 0.2 magnitudes of systematic error. Now Saul is moving more toward 0.25 magnitudes of green and dotted uncertainty.
Another plot: Histogram of E(B-V) Host (i.e. Milky Way has already been taken out). There is a whole peak which are reddened the same amount (E(B-V)=0.09), which is A_B (rest)=0.35. The narrowness of the peak Saul draws is very perverse; why would so many SNe be redenned by exactly the same amount? Rob wants to see this plot by run, which might show up some correlations if there are zeropoint problems. Peter says that the difference in the SCP and Hamuy distribution of stretches could explain the oddness. We have a chunk of supernovae for which there are no Hamuy supernovae with that stretch, so the extrapolation of color vs. stretch could be bogus. So, we need to plot just the SNe which have a stretch in the range of known Hamuy stretches, to see what that distribution of E(B-V) looks like.
Another issue: stretch in I is stretch in R times 1.06. Suppose you choose a too narrow template: you have to raise it to fit most of the I-band points, which could make the supernova look dimmer. Also, hopefully all the K-corrections are stretched now, so that there won't be a problem with that. However, if there is a peak offset, then that could be coming in to it. This is another point which needs to be addressed.
(K-corrections to this secretary are very, very, very scary.)
NOTE FOR SEBASTIEN: Make sure that all of this, including K-correction errors, gets into the error budget.
IF (a big if right now) the big peak is true, then really we are on the shifted contours of lots of reddening rather than the "green" contour we've been thinking we are on.
Note about MW extinctions: B&H, not Finkbeiner, was used. Gerson asserts that it changed our galaxy and Hamuy the same so that our answers didn't change. However, we should look into Finkbeiner vs. B&H with this weird E(B-V) peak in mind... could the peak be the ones where it did make a difference?
Greg suggests that we should look at the host galaxy colors of our SNe hosts vs. the Hamuy hosts.
New issue: Saul plots up the current constraints on this plot from cosmic background radiation. The extent of the confidence region is perpendicular to ours. Current constraints are big, George hopes the constraints to be a lot smaller a year from now.
Saul started pulling together the paper we're doing this for, to find out if it will fit in an ApJ letter. His impression is that it might just fit into a letter.
Discussion over whether or not we have to include a table of all the data. Obviously, that won't fit in a letter. Peter believes that we can't have a letter without at the same time submitting a paper that has all the data. Others disagree, saying that so long as we refer to a forthcoming paper with all of the data and more details, we can still get the letter published.
Greg sent Chris the schedule for tonight. Things look good at CTIO. Things do not look so good at INT.
Greg is still having trouble with IRAF, so he's looking at the HST images "the hard way." He's sorted out the supernovae from the cosmic rays. In three of the four cases, everything is very clean. In one case, if there is host light, it's a very small, compact host. In another case, he wasn't sure he was able to identify the supernova.
NICMOS data is still a problem. There is some question of whether or not we need to change our strategy for the rest of the points on the current supernovae.
Gerson took our SNe and the Hamuy SNe and moved them all as if we were looking at them at a z of 0.5. He then averages all of the data. Three cuts, Hamuy (purple), ours below z=0.48 (green), ours above z=0.48 (red). These are weighted averages, each point being weighted based on the error on that point. All of these points have been stretch corrected (Gerson asserts you are better off if you do stretch them). Around the peak there's about 0.2 magnitudes difference between the Hamuy data and our data. Of course, one must interpret this in terms of how big Omega is. He didn't choose any cosmology, which seems to be a choice of the cosmology Omega_M=1, Lambda=0... is it? Now we're trying to figure that out. Well, OK, he chose some cosmology or another, and the difference would indicate that the cosmology is incorrect. His idea is to change the cosmology until the things match up.
Matthew: for some of the cleaning, Nelson figured out that some of the residual fringing we saw and such was more of an additive effect than a multiplicative effect. The result isn't obvious, but on a couple of frames it's consistently tighter if you treat the problem as an additive problem. All of this effects the Jan 4 BTC R-band data, which has some kind of extra light in it. If it's additive, then the surfacing program can take it out.