Return to Home Page
(fwd) Mira flares (Greaves) The following message is from John Greaves: === I've been following this topic, on and off, for some time now, since I read the Hipparcos Venice Symposium paper on it about 5 years ago. The "examples" out there are still mostly at the anecdotal level, and the following is somewhat anecdotal too. As there is some current level of interest on the matter, and possibly the bestest chance so far of possible follow ups through either survey data or personal interests, I'll put some things forward, just for information/viewpoint's sake. Firstly, however, I'll ask if anyone has ever come across a paper reviewing "Vanadium Stars" or similar. I came across, and read, this sometime during the past couple of years, and then totally managed to lose all trace of it. I can't even remember if any of the stars were variables, as I was reading it on an unrelated matter. It was a review of stars with spectral class M8 to M10 (Miras can have these spectral classes, but using only during minima. The M8 to M10 range is more or less defined by the "condensing" out of VO in significant amounts due to the cool photospheric temperatures invilved). Okay... When I read the Venice paper (amongst many others in the Symposium proceedings: it is a truly weighty tome ;) ) it piqued my interest, because around that time had been busy ploughing through LPV lightcurves en masse. I had, over time, downloaded data for a few hundred objects from the publicly available AFOEV data held at the CDS ftp, and analysed about 200 Miras and a 100 semiregular lightcurves. For certain interesting cases I usually augmented these with VSOLJ (via vsnet ftp) and BAAVSS (by request) data, obtaining up to a century's worth of data for some objects. There are selection effects to consider in terms of flaring. The occasional 'outlier', or lone bright observation, is not unknown in such lightcurves, but they are not common. When considering their validity it is necessary at times to see if they are from a normally reliable and consistent observer. However. There is also a 'negative' selection effect, in the sense that it is a result of things _not_ being selected. Although observers are warned about bias, have observers a tendency not to forward discrepant observations? On the other hand, and again in terms of selection pressures, these flarings have a tendency to occur around maximum. This is a special part of the light curve, in that it engenders more interest, and comes within the observing capabilities of more people using varying instrumentation (eg binocs instead of their normal light bucket). Minima of some Miras, especially before bigger telescopes became more widely available, just aren't/weren't covered at all. Equally, the Hipparcos Epoch Photometry that has somewhat scarily been used to note these flarings is also selecting only at the bright end, by default. Incidentally, HEP should be checked against Tycho Epoch photometry at all times, and BT and VT should always be checked against each other to see if both flare at the same time (even if not to the same extent), over and above the usual consideration re quality and background brightness flags etc, and then serious doubts still retained about the whole thing. A way around some of these problems would be to check individual datasets from long time observers to see if any such brighter than usual observations occur. Gary, I know you've been following some handful of Miras for a coupla decades or so now. When you've time, could you have a quick visual look at a couple to see if you can see evidence of occasional brighter than usual events? Also, it'd useful to know if they do tend to be nearer maximum or whether they occur elsewhere in the lightcurve, or at minimum. (That is, if you've bothered to plough this far through this ;) ). Summary on this bit: There are selection effects that cause flaring to be likely to be more evident around maxima in Miras. Is flaring actually concentrated around this time, which would be important observational evidence, or worse, are maxima occasions when flaring- emulating errors are going to occur. In the context of the latter part of the summary: a collection of flaring events, if found, should be corellated against lunar cycle to ensure there is no effect due to that (I believe full moon has some effect on the measuring of red stars when red stars are bright, due to cones being preferentially used to rods at these background light levels, and their widely differing response to red light, such that cones would see the star brighter than rods would. That needs checking though, I always get confused on which is which with this). Highly seasonal and circumpolar stars will be viewed at low elevations, the former always, the later part of the time. Again, simply testing an apparently flaring observation against local elevation (difficult if you don't have a rough idea of the observer's site) at that time with, say, a planetarium/sky charting package should soon show if any correlation is occuring here. Note that seasonal altitude variations, with an annual value, and/or lunar variations, with a monthly value, combined with the long periods between maximal brightness that LPVs have, can lead to _apparent_ brightenings occuring only at the infrequent intervals when all circumstances are timed just right. That is, a similar pattern to that suggested for the flarings. Concentrating on evidence from long runs from individual experienced observers should get around this, but nothing is certain. Basically, it would be useful if a handful of stars could be selected for inclusion in a monitoring program, as there are quite a few Mira out there, and the normal suggested rate is to view them once a fortnight, if not as infrequently as once a month. More frequent observation probably needs a small sample to target. Visual observers like to view as many objects as possible given kind weather, if only to make up for the times they couldn't observe. It's a personal reward thing. Of course, the comments so far have been relative to visual observations, and especially archival ones. At present I _personally_ haven't sufficient confidence in the available sky surveys (ASAS3 and TASS) to be happy with using them for something like this. This is not meant disparagingly. To show something new and unknown (well, mostly) you have to have a lot of confidence in precluding other, more mundane, alternatives. I've had a bit of bother with ASAS3 data from time to time, so I don't always trust it implicitly, and remember we're trying to decide whether something is a glitch or a phenomenon here. With regards to TASS Mk IV stuff, there is a nice little trick where you can check adjacent similarly coloured stars to see if any interesting event also occurs in them. If yes, possible image/processing/atmosphere problems exist, if no, well then the phenomenon is likely real. However, most of the current observing regimes of the TASS (well, currentlly mostly Tom's cameras') survey are insufficiently dense enough (time resolution wise) to be of effective use in this matter. Though they can be suggestive (as can the ASAS3 stuff). Increasing passage of time will make both these systems more valuable, as time baselines extend. It is as usual a matter of time. However, if these flaring events are on very short time scales within each observing session, the current survey practices (often dictated by necessity) of daily or less frequent short sessions may still mean detection is hit and miss, and evidence not necessarily unambiguous. [After all, it is not within these surveys' remits to look specifically for flaring Miras.] These events need not only to be shown to exist (or not for that matter), but if they do exist, their distribution within the lightcurve needs to be delineated. I make the last point in the context of eventual models. If the events occur at a specific point in the phase cycle of these pulsators, a mechanism may be easier to track down. This is why I asked about the 'Vanadium star' paper. I've been all over the ADS and elsewhere trying to refind it, but with no luck. A short list of vanadium stars, or more properly stars of spectral type mostly between M8 to M10, especially variable ones, could be useful as a small target sample. If vanadium oxide is going to be considered the source of these flarings, then these stars may be more prone to showing flaring. However, I suppose that depends on the temperatures involved. M8-M10 stars may not get hot enough to cause the effect. Indeed, if the bias towards flarings occuring at maxima is real and not a observational selection effect, this may be the case, as not an unfew number of Mira descend that 'late' in their spectral class near minimum. This has turned out a lot more long winded than I had hoped, but I hope it gives some clues and thoughts to those interested. If some are going to pursue this they need to define the problem a bit first, as some serious, continual, high time resolution monitoring is going to be needed, and for this practical consideration dictates that a small sample of likely candidates needs to be shortlisted. Hope I've not confused the issue (or myself) too much. Personal feeling, based on some anecdotal evidence and circumstantial evidence, and 'feel' for Mira lightcurves (analysis gives one a totally different perspective on archival datasets that makes some of the issues normally voiced on said totally irrelevant), I'd say the phenomenon was real enough and simply mostly missed due to the short duration of each event when compared to traditional observing regimens. The flares aren't especially brighter than usual in the visual after all. I've been wrong afore, mindst ;) Currently, Vello Tabur's comments at the bottom of http://vsnet.tip.net.au/~vello/varstar2/notes.htm are the best simple non-circumstantial recent evidence out there, and as this occured during a "secondary" maximum in this non-standard Mira lightcurve, it is not necessarily extendable to all Mira stars.. That ain't a lot to be going on with. Cheers John John Greaves
Return to the Powerful Daisaku Nogami
vsnet-adm@kusastro.kyoto-u.ac.jp
