Minutes of the 1st Sunday Afternoon Discussion

Liese vZ	Could someone please go over the basic parameters of ALFA?
Karen O.	100 MHz x 4096 x 7 beams x 2 pols x 3-level If you can expand the WAPPs to 200 MHz, you get same thing If you want 9 level, you lose channels.
Liese	What is velocity resolution per channel?
Riccardo	5 km/s
Wim vD	We reserved a day and a half for discussions because NAIC would like to have a document or at least minutes of this meeting that discuss our requirements. Lots of things need to be discussed. If we do this in plenary, then it will take a long time. So one idea is to take today to discuss and hopefullly reach consensus, and then tomorrow have subcommittees discuss specific topics.
Riccardo	I have a problem with the meaning of "consortium". A consortium is a group of people who get together to carry out a coordinate effort, to commit resources, to make sure a project is successful. The nature of this workshop is too distributed. 50 people probably cannot actually do this. What is our responsibilility? I think it is to lay out the science objectives and discuss requirements for hardware. We also need to specify the software needs, especially for things that NAIC does not have available. And, we need to define the data products that will make ALFA surveys a major legacy for the Arecibo Observatory. In this way, what we are talking about is not part of a user-driven proposal paradigm. Rather, these surveys constitute a major telecope time investment. Surveys have to be undertaken as large, coherent efforts that are not the same as piecemeal PI projects. Furthermore, the survey teams have to serve much broader community and the efforts must lead to the production of robust software and best quality data products also on the shortest possible timescales. Note though that while new to AO, this paradigm is not unusual to astronomy. Think for example about the SIRTF legacy program, the VLA FIRST survey, etc. We need to consider how those functioned and what their experience can teach us. To me, an important function of this workshop is to explore ideas on how we can achieve these goals. For example, together we can probably come up with a better survey definition than a single person might come up with, and we can hope to foster alliances that may lead to formation of consortia which in turn will define h/w & s/w requirements based on the science goals.
Ed S.	I have a couple of questions. I've spent a lot of time on committes for HST before it existed. Should the "consortium" be arranged on any one scientific topic? That is, should there be a single consortium or should there be a competition? The same circumstance arises in other fields, like medicine and genomics. There are papers with 30 names, and it sort of works. I think it is probably better to have only one consortium if that can work. But on the other hand, what about small very specific projects, e.g., mapping HI in the outer layers of spiral galaxies? How do you do those when they don't fit into big surveys? Do they or not? Should folks who want smaller things done apply to NAIC or apply to consortium?
Jon D.	I find this very strange way to proceed. Don't you need to first define the unique science? We need a set of scientific objectives; once you define them, then you define the strategies that will accomplish those.
Wim	We can discuss the scientific objectives first.
Eli B.	But if NAIC is planning to produce a survey to reach a broad community, then science doesn't matter.
Jon	Science has to dictate the surveys. We should start with that, and then the consortia will sort of fall into place.
Wim	Who wants to work on what will fall out? Riccardo proposed three kinds of surveys: ALFALFA Virgo ZOA Let's let these make a first group. So what do you want to get out of it? Don Campbell said at lunch that 25% of the observing time on the telescope/year might go to ALFA surveys. That stretches out the timescale considerably. We could also argue that it is only useful if you give 50% of the time to ALFA. What about synergy with other consortia?
Jon	I'd like to focus on what you can do with Arecibo that you cannot do with something else? Can Riccardo outline his scientific goals again?
Riccardo	The science goals behind the simulations that I discussed in my talk are: Determination of the HIMF to faintest levels. Determination of its dependence of environment. Mapping the distribution of luminous and dark matter. Determining the gas-rich membership of nearby groups. Identifying the population of gas rich systems in Local Group and periphery of Milky Way. Looking for (rare) OH megamasers near z=0.25. Being surprised! Note that the ZOA survey in the simulation was specifically designed piggy back with the P-ALFA pulsar galactic plane survey. I think we can't afford at this point to discuss the surveys in great detail. But we do need to discuss them in general, so we can figure out what they are actually going to require.
Wim	Right, we need to supply information to NAIC.
Eli	Do you need 3 different surveys to make statements about your scientific goals?
Riccardo	Yes. For the lowest masses, you need to cover large sky area. Arecibo will have much better resolution than the other surveys; that will immediately enhance the results vs. HIPASS/HIJASS. But remember that we also need to be (a bit) realistic. It would be great to detect solar masses at 15,000 km/s but we'll all be dead by the time the survey is finished. We want to get results on a short time scale. Once you start integrating for a long time per point, the total survey time per given sampled volume becomes large.
Karen O.	To determine the HIMF, we want to get to to solar masses at a distance of 10 or 20 Mpc.
Jon	It would be good to integrate longer to investigate the low end of the HIMF might be different?
Riccardo	But as you increase the integration time, you necessarily have to cut down the sky area.
DJ P.	What volume do you need for HIMF studies to look for variations with environment?
Steve S.	Can you look at different strips within HIPASS?
Lister	It's very hard to do with current data. Not just for the mass range between and , but also to . You really need something like ALFALFA.
Peppo G.	A wide Virgo survey would let you go to explore low mass volume. If you go to 810 degrees, you are 20 degrees away from Virgo.
Karen O.	But then you are looking in a biased region.
Riccardo	But you also get the foreground and background. A combination of small and wide, narrow and deep is the best. That's why I discussed three surveys. Perhaps we would want to do a shallower Virgo survey over a wide area, and more deeply, a strip off to one side.
Jon	Another approach would be to do a deep strip that went around a wide part of the sky across Virgo.
Noah B.	I am very interested in the Virgo survey such as that Riccardo proposed. Lots of other (multiwavelength) work is also being done on the Virgo region, so you ought to do the whole cluster.
Steve	But, velocities of Virgo are very had to interpret.
Brent	You do want to get solar masses in certain regions at 1000 km/s. The best place is in the CanVen region.
Karen O.	We don't need a single survey; several will do.
Lister	Riccardo, can you review how much telescope time is required for ALFALFA and how many galaxies you expect to detect?
Riccardo	Over the Arecibo sky, ALFALFA would detect about 10,000 galaxies out to 6000 km/s in 1200 hours, with variance depending on the HIMF.
Karen O.	We might want to ask what we gain relative to HIPASS in doing something like ALFALFA.
Ed	Is the optical community trying to do the same kind of surveys that ALFA will do?
Judith I.	We might also want to do deep surveys around nearby galaxies looking for outflows, HVC's, etc. Targets might be the largest galaxies within 10 Mpc or even closer.
Riccardo	Remember that we actually do know something about the density field in the nearby universe from the available density maps, so we actually can design surveys to look in different density regimes to look for environmental differences.
Noah	There is a void behind Virgo.
Riccardo	There is also the void in front of Pisces-Perseus, in the other part of the sky.
Brent	Let me toss out another idea. I have speculated that there are groups of dwarf galaxies within 5 Mpc. In a few regions, we see 4 or 5 galaxies over scales of 300 - 400 kpc, with very small velocity dispersion so they do appear to be bound groups. If we included those regions in a survey to look for invisible targets, you'd have to survey around 100 sq. deg. for each group.
Riccardo	Remember that for a fair number (like 400 nowadays) of galaxies, there are accurate distances from Cepheids, tip of the red giant branch, etc.
Brent	I can think of 2 groups in the AO dec range. One is around 0 hours, and one is around 7 hours.
Wim	Ok, so there seems to be some consensus for a large shallow survey and also for some deep targeted regions. And maybe a strip over the sky that might or might not go over Virgo.
Karen O.	Perhaps what might work well for Arecibo would be to do ALFALFA but also agree on some declination strip for a deeper survey and also focus more on that in selected regions.
Riccardo	To get low mass objects, you need to maximize solid angle coverage.
Steve	But you have to ask over what distance do you want to detect lowest masses?
Riccardo	But for the same total survey time, solid angle is better.
Wim	Perhaps each strip could be a subset of the ALFALFA survey, but cover some region deeper.
Riccardo	Every time you restrict solid angle, you restrict your volume.
Wim	OK, there seems to be consensus on the need for both a shallow survey and a deep survey.
Karen O.	Don Campbell said that of the 4380 hours of astronomy time, maybe 30% might be for ALFA. ==> say 1200 ==> say 300 hours/year per consortium. So we might really may benefit from piggybacking.
Wim	Riccardo's table shows how much time we need.
Jon	Do you need to detect 10000 galaxies?
Riccardo	There is no simple way to trade high mass and low mass galaxies. If you cut down area, you don't get the low masses.
Jessica R.	It doesn't help you to go deeper in same amount of time because you cut solid angle.
Mary P.	Can you do shallow piggy back with GALFA?
Riccardo	That is why I asked if a shallow survey with 5 km/s resolution would be interesting to the GALFA group.
Mary	I think 5 km/s would be enough for an all-sky survey. Of course, the galactic plane would need better velocity resolution.
Riccardo	Another way to "cheat" is to try to figure out how to drive down detections to lower sigma.
Wim	Are you saying we really can't do it?
Riccardo	I don't know. That's something we need to explore, detection algorithms.
Wolfram	The question is: what is the most efficient way to discover? Other motivations say it would be worth it to do a very deep survey in a small part of the sky, say an "ultra deep" survey.
Phil P.	So it comes down to how deep is deep? Is a pulsar-piggyback ZOA enough?
Karen O.	In 4 minutes, you get 1 mJy rms out to 40,000 km/s.
Riccardo	But that's because you also do an OFF scan. Another option is to do simple drifts where you get 12 sec.
Steve	There are different numbers quoted for ALFA vs HIPASS and HIJASS.
Riccardo	Of course, the numbers depend on the assumptions on Gain, Tsys, the number of beams, the velocity resolution, etc.
Liese	Important difference will be velocity resolution. While 5 km/s is marginally ok, for dwarf galaxies, 2.5 km/s would be better. Perhaps that would also make for greater synergy with GALFA at high latitude.
Noah	If you drift, you can probably pinpoint location in RA better; that would be good for making optical identifications.
Steve	What is the minimum velocity width for a dwarf irregular? I know of nothing less than 25 km/s.
Eli	Can someone remind us of the specs of the GALFA spectrometer?
Karen O.	At this point, it is purely a proposal to GALFA for a machine that gives 5 MHz bandwidth for 2 polarizations with 1 KHz resolution.
Wim	Lister, in your talk today you had viewgraph for what ALFA should do; will that shed light on the choice of survey?
Lister	I think ALFALFA as Riccardo discussed it fills in nicely. We need to do such a shallow survey. As to drift scanning and the accuracy of positions, as long as you Nyquist sampling, it doesn't matter. But the important thing to keep in mind is that you need to Nyquist sample.
Riccardo	If the S/N is high enough, you can centroid.
Lister	But, will you get high enough S/N in a shallow survey like ALFALFA?
Riccardo	You can centroid, but not resolve. Another point is that, if you drift, the pointing accuracy would be optimized. When you drive the telescope, (e.g. at slew rate), what is the pointing accuracy? If you have ever been on the platform when the telescope is moving, you know how much it shakes.
Phil	It looks good enough, for 3.5 arcmin beam, as long as you do not accelerate at high Z.A.
Daniel A.	Drifting also means a lot less less wear and tear on the telescope.
Phil	Standing waves do not move if you drift.
Wim	So far, the discussion has been mostly on ALFALFA. It seems to me counter-intuitive to use most sensitive telescope on earth for a shallow survey. Maybe the percentage of telescope time for ALFA should be more than 30%.
Christian B.	Remember that GALFA will also do high latitude surveys.
Riccardo	It is going to be a matter of priorities. Of course every one would like to do the whole sky.
Jim	Remember that only 4.5 hours per day can be used for a galactic survey, but it's the galactic survey that is highest priority for the pulsar community.
Steve	So if we piggyback, who gets charged the time?
Riccardo	The allocation of telescope time will be driven by science. We should think like scientists not like bookkeepers.

A coffee break was taken at this point.