## Transcript of plenary session discussion

White: All right, may we start? I think there was a principal component analysis that Andrea Dobson did, stating that you should be able to define or describe a star using four parameters, using an ideal four-dimensional phase space. Now we've gone through this exercise today, and I'll tell you: the dimensions have increased a lot [laughter] in trying to track the Sun through all these parameters that we think are various strengths, and interests, and necessities. So what I'd like to do is just move down through the questions. We've tallied the responses as best we can in the past few minutes, and the individual group leaders will have to respond to and defend choices presented on these viewgraphs. Fortunately, we haven't identified any names on these [laughter], so you won't be fingered immediately. Let's just see how far we get, and see if we can come up with some description or rationale for how we improve our selection of solar-analogous stars. So here's Question 1A...

• Question 1: The Solar Twin Search So Far: Solar twins have been selected on the basis of specific observational parameters, and there are a large number of these to choose from.
• [A] How have [list of four specific parameters] been employed in the solar analog hunt to date, and what has been the success or failure in identifying solar analogs in terms of those parameters?

White: Now [putting up Table 1A] to help you through this blizzard of numbers [laughter], we marked parameters with an asterisk where the rankings of the two groups that evaluated them, on the one-to-five scale, totaled six or more. [General laughter.] Yes, I know it's a little bit arbitrary. And I'm not going to worry about the standard deviations, I'm sorry; they're a nuisance [laughter]!

Wade: Just like an astronomer: throw out the error bars [laughter].

White: I say we throw out the whole thing and go have a drink [laughter]. Okay, effective temperature, B-V, other Johnson colors, Stromgren photometry, Cousins photometry, R'HK, composition, MK type, and luminosity are the ones that come to the top as being the ones most used. You can see the relative rankings of the others, along with some rather derisive comments at right [laughter]. For example, for energy distribution it says, Useless, but fundamental'' [laughter]!

Ayres: Um, excuse me...

White: Now, I think we...

Ayres: Excuse me-e-e-e; that was from my group in the morning, and the whole point is that from a remote sensing point of view, the energy distribution is actually the most fundamental quantity there is. But from a practical point of view, forget it! You have to rely upon color for most purposes.

White: I agree. There was continuous discussion in the groups I was in about what was a fundamentally measurable quantity, and what you had to infer by going through models. Now, there are nine of these 20 parameters that people ranked pretty highly as being widely used. Are there any comments or disagreements?

Garrison: Wouldn't you need the same comment for Mbol as you have for energy distribution?

Soderblom: And how do you get Mbol if you don't have the V magnitude?

Garrison: Right, exactly.

White: Fundamental, but useless [laughter]?

Soderblom: No, not useless. How can V be listed up there as not relevant'' and Mbol as fundamental?

Ayres: I guess we were thinking you'd have V and a bolometric correction...

Franz: And you'd better have a parallax, too.

Soderblom: I would take issue with listing temperature as a derived quantity''. Of all these different things you can determine for a star, there's not very many you can determine in a direct sense for the Sun. But temperature is one them. You can argue about whether or not you get an absolute temperature or not, but I don't care: you can get differential effects very nicely. So I happen to like Teff.

Buzasi: I'd like to point out that at least the way we were interpreting things in my group, the question was What has been useful?'', not What could be useful?''

Soderblom: Well, Teff has been useful.

Buzasi: Yes, definitely.

White: It's one of the principal things that has been used, to specify the zone on the HR diagram where we want to look. Any more comments?

Wade: Is anyone surprised?

White: I'm not particularly surprised. Well, one thing that did surprise me, and this is just out of ignorance, is the demand here for Strömgren photometry.

Skiff: It should be used more than it is.

White: This will come up over and over again: we all, in our individual fields, have these historical indices or measures that we're comfortable with and have used for a long time. As time goes on, we figure out better ways to do it, but it's hard to make the shift and still keep everything tied together, because you wind up doing it all continuously from that time on. My favorite case is the sunspot number, which drives me up the wall because it's not physically interpretable. But, it's the longest continuously-measured solar activity index we have.

Rottman: It looks like one of the groups picked B-V and the other one didn't, and the other one picked Strömgren and gave less weight to the Johnson side. Maybe it goes back to this question of what has been historically used.

Ayres: Strömgren is superior.

White: Superior, yes, but there's a lot of experience based on B-V.

Ayres: Our group said basically, there's this huge historical database based on Johnson B-V, and so naturally it has been used; however, the Strömgren colors are much better because they're not as sensitive to metallicity.

White: Everybody agree with that? [Chorus of Yes''s.]

Soderblom: We're evolving into the Strömgren system.

Wade: What about the Tycho system?

Soderblom: Well, we haven't checked it out, we haven't had a chance to absorb it yet.

White: What, you mean there's another system?

Hall: There are infinite systems [laughter].

White: So the consensus is, if we could, we'd shift to the Strömgren system from this time forward?

Ayres: Yes, exactly, definitely.

Wade: What prevents us from doing that?

Soderblom: Nothing! It's being done.

Skiff: Erik Olsen has observed all the HD stars brighter than 8.5 between A5 and G5. It's done. It's 40,000 stars. It's beautiful. It's a series of A\&A supplement papers; the first one is in volume 54, page 55.

Guinan: I know it [laughter].

Skiff: That one's got 30,000 stars. The advantage of the Strömgren system is that you can separate the metallicity...

Schleicher: Do we know the value of the Sun in that system?

Skiff: Yes, look at my poster.

Garrison: Dick, you were concerned before about there being more than four parameters. Some of these are redundant, so it may not be more than four parameters.

White: We do want to get around to the lowest potential phase space, eventually.

Radick: I think there's exactly four parameters up there, out of the nine you flagged with the asterisks. In a broad sense, there are five of them that are either temperature or some equivalent of temperature, there are two that are luminosity or equivalent to it, there's one composition and one activity. So there really are four.

White: Let's work through that...

Radick: OK, there's composition, and activity, each mentioned once. There's Mbol and MK spectral type, which are both basically luminosity, and the other five are all temperature in some way.

White: So we're down to the four fundamental ones that you always come to. We're just seeing the different names, the different techniques that have been invented for approaching the same intuitive physical quantity.

Garrison: That's wonderful. It's a good thing we have those checks on the same quantity.

Hall: So Andrea was right!

White: Well, now that we're back to four, can we move on?

Wade: Who was the second group considering rotation period? They said 3'' and the first one said N/A''.

Soderblom: The question was asking about use of this parameter to select solar analogs up to this point. We know nothing about rotation, nothing, so that's where the not applicable came from.

Wade: Fine, I agree with the N/A, so where did the second group's 3\pm0'' come from?

Soderblom: Darned if I know.

Ayres: It's only the extremes of rotation that would be important anyway.

White: All right. Question 1B says...

• Question 1: The Solar Twin Search So Far: Solar twins have been selected on the basis of specific observational parameters, and there are a large number of these to choose from.
• [B] Are there any relevant parameters that have so far been ignored (both in the literature and in the list of 20 parameters supplied to the working groups) or inadequately considered, and what would be the specific insights gained from considering them? Determine up to two that are most important.

White: It's hard to get all this on one viewgraph in a sensible way, but we'll do it anyway. [Puts up Table 2. Pause. General laughter; the first item says intelligent life''.]

White: I thought the first one was a hoot. As we filled out these viewgraphs, the thought came out that it maybe shouldn't be intelligent life, but silly life. So, what we think is missing is X-ray observations, 10-cm radio, UV spectrophotometry, rotation, Geneva photometry...now that's something I don't know, yet another system.

Soderblom: It's in some sense relatively esoteric, and we're not going to go out and do 40,000 stars all over again on a different system.

White: So we stay with Strömgren as the most reliable, best-defined system.

Soderblom: Tom, I thought we added companionship to the list.

Ayres: Yes in fact, in both of my groups, companionship came out as an important factor.

Garrison: We added companionship as well. [General babble of agreement.]

White: Yes, I just missed that in making the viewgraph, but there's an important point: were we considering companionship in the sense of binary systems, or in the sense of planets? My groups were discussing the planetary side of it as interesting for solar-type stars, but you needed to know about binary systems to discriminate against them.

Garrison: I like using the word environment.''}

Soderblom: Yes, you could be talking about extensive circumstellar material, for example.

Rottman: I'm showing my ignorance here, but what about the 10.7-cm radio entry?

Garrison: Either X-ray or radio could be used as an interesting discriminant for solar-type stars.

Ayres: Yes, but how many solar-type stars have 10.7-cm radio measurements?

Garrison: That's the point, radio is a nice one we could add as an approach for finding analogs.

Guinan: But the radio is so feeble, and X-ray would do the same job.

Garrison: But not many people have sat on a star in X-rays, to see if it's in a low activity cycle.

Ayres: You can't even detect them in the radio, though.

Skiff: You apparently can't detect them in X-rays, either, not ordinary activity?

Guinan: Yes, you can.

Soderblom: Not in survey mode.

Garrsion: Maybe radio and X-ray are redundant.

Guinan: Look, radio you just can't do. The Sun is a feeble radio source. Put the Sun out just five parsecs away, and it's gone. There's no existing radio telescope that could possibly pick it up.

White: What about UV spectrophotometry? That one got a lot of votes.

Lockwood: I was avoiding that one, the UV absolute energy distribution, because of the extreme sensitivity of the measurements, like the ones Dave Schleicher does, to discrepancies from the perfect solar analogy.''

Wade: We're not talking about transition region material here, right?

White: No, it's ground-based, so it's from 4000 Angstroms down to the ozone cutoff.

Ayres: Dick, our group was thinking more of the mid-UV, say 2000 to 3200 Angstroms.

Wade: Convection is up there as a desirable. Is that something that can be used, really, as a tool for finding solar analogs?

Ayres: You'd use line bisectors, the kind of thing that David Gray is doing -- high-resolution spectroscopy and analysis of line asymmetries.

Wade: So this would be for detailed study of the physics of the surface of the star, not a screening tool.

White: Good. Now going down to Question 2A, which is...

• Question 2: The Solar Twin Search Following This Workshop: What is the optimum method for searching for solar twins?
• [A] Which, if any, parameters must be used in any search for solar twins (determine up to four)?

White: Okay, I've listed all the groups' opinions, and what the 8 x 5'' means, for example, is that for Mbol there were eight ratings of 5'', and then there was some wag who put 4.5 [laughter]. The rest of this is pretty straightforward. A lot of votes for rotation, colors, composition got a lot of votes. The ones that fall off the wagon are mass and age. So...Tom?

Ayres: From a fundamental point of view, you really are interested in stellar evolution. Mass, age, and composition should uniquely determine the state of a star. So our group picked those, instead of, say, rotation, which is correlated with age to some extent.

Rottman: I'm a little confused by the trick question here: I read the question to mean, which four things would be the best discriminants, if you only used four.

Garrison: No, we took it as meaning which parameters would you insist upon before you declared a star a solar analog. In other words, if a reliable measure of just one of the four parameters were missing, you wouldn't be able to say the star was a solar analog.

Ayres: And the difficulty is, of course, that mass is almost impossible to determine for groups of stars, even though it's absolutely fundamental, and the same thing goes for age. We know the age of one star, the Sun. We have to move over to proxy information when we go to the stars.

Wade: Which is why you use temperature and luminosity to stick the star on an HR diagram and read off mass and age. So, in the sense that colors, and Teff, and MK type are all different ways of plotting the x-axis of the HR diagram, those have about the same total weight as Mbol.

Soderblom: I see log g up there. I would argue that log g is pretty useless.

Garrison: Well, log g is what you really get.

Soderblom: You can't get really log g very well out of spectra.

Garrison: You can't get Mbol either.

Ayres: You can't get enough accuracy on mass out of a log g determination.

Buzasi: Activity and rotation are both up there, and they're clearly related. You can argue about how tight that relation really is. Activity is nominally easier to measure. Rotation is, I would say, more fundamental.

Ayres: That's only true among stars that are fairly close to the Sun.

Buzasi: Well, that's what I'm talking about; we've already narrowed the field down. You're not measuring rotation in early A stars here. I'm talking about twins, not analogs here.

Hall: Going back to the conclusion from the first question, I notice if you allow Strömgren'' and colors'' to mean the same thing here, you get the same result.

Buzasi: A difficulty here is that some of these things are numbers that you measure, and others are techniques.

White: Okay, good. Think you can pull a summary out of this, Jeff?

[Silence.]

[General laughter.]

White: Shall we press on?

• Question 2: The Solar Twin Search Following This Workshop: What is the optimum method for searching for solar twins?
• [B] Do the parameters employed depend on the research needs of the investigator, e.g., would an investigator working with far-UV data have fundamentally different search criteria than one working in the IR, and how would these needs differ?

White: This was not controversial at all; the answer was quite simple. YES! We ran into this in both discussions, that there were clear differences in need depending on what you were looking at. It just comes down to admitting that those differences are there, and out of our list, making sure we have the fundamental parameters that really describe the star.

Soderblom: It also depends on what you're looking for, that is, your science needs.

White: Yes, the one that came up in particular was whether the star had a companion or not. Are you a SETI person, in which case discarding companionship is not the ideal criterion.

• Question 2: The Solar Twin Search Following This Workshop: What is the optimum method for searching for solar twins?
• [C] Can (or should) the essential parameters identified in [A] above be ranked in order of preference? That is, what is the optimum, properly ordered set of "filters" for identifying solar twins?

White: Now, this is the set of answers [Table 5]. You can see there are different preferences in the order here, but there's some obvious commonality, which we expect. You see Mbol and effective temperature showing up across the board, as they should. Composition appears frequently. There's another demand for Strömgren colors, and perhaps a somewhat surprising demand for rotation and activity.

Wade: It seems that sometimes the Strömgren colors are used as a temperature discriminant, and sometimes as a metallicity discriminant. For instance, in the rightmost column, you've got MK type, which is your basic temperature criterion, and then Strömgren modifying it.

Garrison: Yes, you've got MK, Strömgren, and colors, which are all ways of getting at Teff. So you can boil this down somewhat.

Wade: I'm also interested in that group's rankings: what is it about Strömgren that puts it higher than Mbol?

White: I would put Mbol higher myself.

Garrison: Well, Strömgren is well-determined and Mbol is not.

Wade: Mbol is sort of a measurable quantity.

Garrison: Aww, come on, have you determined Mbol lately?

Rottman: In five of these columns I see a lot of similarity, but in the leftmost one, we have high-resolution spectroscopy sitting in there, and I was wondering why that group put it there. What exactly did they feel you'd get?

Buzasi: I think it was our way of rolling a lot of different things into one. We were trying to squeeze as much information into our four parameters as possible. High-resolution spectroscopy is something from which you get, for example, compositions.

Wade: So how do you determine, from your spectrum, whether or not the star is a solar analog? What measurement do you make?

Buzasi: Temperature, [Fe/H].

Soderblom: How about, for things like 18 Sco, looking to see if the spectrum looks the same?

Wade: Hmm, so what do we do, a little chi by eye,'' or...

Garrison: You don't need high resolution to see if the spectrum looks the same.

Schleicher: When it comes right down to it, of the first three parameters in the rankings there, two of them have to define the star's position on the axes of an HR diagram. In every case. And the third one is composition. And the remaining one is activity, or age.

White: Well, there's the same process again, we're going down this one little track through the phase space. All right, let's go to the lottery here...

• Question 3: Optimum Candidates Rank the following stars (listed by HD number) in order of preference as solar twins, to the extent that the working group considers feasible and/or useful. Each group may add up to 2 stars to the list. The stars are: 1835, 11131, 20630, 28099, 44594, 76151, 86728, 89010, 114710, 144585, 146233, 186408, 186427, 187923A, 217014.

White: [putting up Tables 6 and 7] Well, here's the ranking. You'll see there's a big cutoff after about four stars. Within those, there's 18 Sco, 51 Peg, 16 Cyg A and B. They come in different orders. I'm a little interested in why the one group ranked B over A. But in the groups I was in, I know there was a general feeling that the top four were good solar twins, and below that we started to consider the stars more to be solar analogs.

Buzasi: Yes, we actually drew a line under those four.

White: I also see VB 64 in here twice, and that's a surprise.

Guinan: Yes, why is that there? It's an analog, not a twin.

Ayres: Well, it's photometrically very similar.

Guinan: But the age is clearly different.

Garrison: The metallicity is different.

[Numerous comments, amounting to general denouncement of VB 64.]

Guinan: When you have a twin, it isn't just height [laughter]. It's weight, age...

White: DNA, whatever.

Schleicher: It had better be the same age!

Hall: I was wondering why so many groups ranked 16 Cygni A above 16 Cygni B. After all, B was the one that Hardorp explicitly called the perfect solar spectral analog,'' and it subsequently was analyzed in great detail, so B is the one you traditionally think of as the slightly better analog.

Garrison: Who does? I never have.

Guinan: B is better. I don't think it matters. I just think people didn't care what order they put them in. I've always thought B was closer to the Sun than A.

Hall: Okay, that's what I was asking: were groups really ranking them that way, or just writing them down as a pair.

Guinan: But they're older than the Sun. They're 7 billion years old, and a bit more luminous. They're not good twins, but there hasn't been much else to use.

Garrison: Perhaps if you're looking a for a good example of what the Sun will become.

White: Okay, just to clean up the diagram, is everyone happy with VB 64 where it is?

[Further vociferous denouncement.]

Wade: You know, I don't see the Sun on that list anywhere [laughter].

Guinan: The one that really concerns me is HD 44594, which lost its twinship on the basis of a Hipparcos parallax.

Garrison: I think it should lose its twinship. It doesn't look very good.

Skiff: HD 44594 is slightly evolved.

Soderblom: Well, the two groups who put on VB 64 are clearly out of order; they were incompetent [laughter].

White: We have four excellent candidates here, and I think that represents a pretty good consensus.

Rottman: One group ranked 16 Cyg B very low, though; it's way down at number 8. Is there a reason for that?

Friel: That was our group, and the reason was that we were looking almost exclusively based on the data supplied in the table [in the meeting materials], without updating the data using more recent results. So if you just ranked them on the table data, that's where B ended up; but in our discussion we agreed it should be higher.

White: Any more comments? Well then, that about does it.

[Contents]