After a few years non active in variable star observing a decided to go back. I was browsing some lightcurves and found that a note I made years ago (and not only me) is still to pollute the lightcurves. It disturb me that there are observers that are estimating mira stars on practical daily basis. I you take the lightcurve of R And for the last 2 years, you can see that some observers have a lot of observations. There is an observer how has 90 observations of a mirastar in 2 years and they are concentrated around maximum and magnitude 12. This is not correct. Some of these observations are done day after day and I also found two on one day by the same observer.
I know this is a hobby but one they should take serious.
I don't understand why the AAVSO doesn't warn these people that they have to respect the cadence of 5-7 days. Maybe the AAVSO wants to have the greatest and most filled database but I should want a database with qualitatively observations, not quantitative.
I don't know what kind of warning are you suggesting, but any observer who has read the Manual for Visual Observing (chapter 6, "Planning an Observing Session") should be aware of this.
Then this observer clearly haven't read the manual
I understand your point. However, if the observer was trying to determine and compare any flaring of this mira at maximum and minimum, then this would be a perfectly proper observation cadence. So I always let the observation cadence support the science--and sometimes we don't know what to expect. So some probing is always good science/research. We should not be using some rule of thumb, unless we have no science in mind, and are just observing to support the light curve.
Gary
Gary, I don't think this is a case of flaring.
Have you talked to the observer, or are you assuming?
Amen to that, Gary.
If the phenomenon being observed is strictly periodic, then then Nyquist criterion applies, and we need observe it only twice per period to determine that period. But when applied to the real world of variable star observing, that would be an extremely shortsighted program, for several reasons.
1. All measurements are prone to error. Within limits, increasing the frequency of measurements reduces the error variance (though alas not any inherent bias).
2. If we knew that the stars were absolutely periodic then a single, highly accurate series of measurements, over some small number of periods, would suffice to characterize the period forever. But the stars are evolving, and either we do not know the time scales or we need data to support the theories of period evolution. More rapid cadence supports detection of period evolution.
3. If there are short term variations, which I think Gary is alluding to, then the cadence needs to be sufficiently fast to catch those variations.
4. On the other hand, visual observations should not be made so frequently that the observer has recall of the last measurement, because if she or he does, it introduces a form of bias, which in the machine learning community anyway is referred to as "selection bias." (A bit off topic, but related: One should not look at recent data before conducting a session for that same reason. If you know that Joe Blow measured it at 13.252V an hour ago, you will have a strong urge, whether conscious or not, to validate that measurement.) The upshot of this is that my personal practice, which mostly involves CVs and SRs in my visual program, is to limit myself to one observation per star per night. But all that goes out the window when electronic measurements are being made.
CS,
Stephen
Have you simply contacted the observer to ask them why they are observing more often than you like? Maybe they have a reason you haven't thought of.
Mira observers are not required to hold to a cadence of one observation per 5-7 days. For example, some Miras have a period of less than 180 days, and a slow cadence would be bad strategy.
Why would AAVSO "warn" observers to take less data, when in fact most Miras are woefully under-observed? Encouraging observers to spread out their observing efforts across more targets might be helpful, yes. But I would not take kindly to a "warning" that I'm giving too much.
HHU, I fully agree with your take on the situation, something I have posted about here a number of times in recent years...apparently to no avail. Back in the day rules were much more strictly followed. In MWM's time it was clearly stated that if more than one observation per Mira star were submitted within a 7 day period by the observer that the additional observation would simply be discarded and not included into the AAVSO's database. The intent was to better balance the weight of observer input. To my knowledge, this was indeed enforced. However, the situation has become progressively more lax as the years have gone by, until now it has become something of a fairly common practice. Just as you, I see the same folks submitting strings of nightly observations of not only Miras, but bright slow varying, or even nearly constant, red variables as well. What is particularly disturbed is that those violating the rules are often the newer observers, whose level of accuracy in their observations often leaves much to be desired, weigh increasingly heavily in the mean light curve of a given star. Their impact on light curves also grows more significant as the number of contributing AAVSO observers steadily shrinks year by year. Being mostly relative newbies, there is little excuse for them not having read the Manual if their intent was truly to contribute to our database. Clearly, the only purpose of those failing to follow the cadence guidelines is to build up totals, 30 years of editing the AAVSO Circular taught me that! In no way is it any attempt on the observer's part to supposedly detect nigth-to-night small amplitude variations when one is still a relative novice.
BrooksObs
"Building up totals" is very far from the only reason to deviate from the old and too-inflexible cadence guidelines.
For example: the Miras I monitor vary in period from 118.5 days (FN Ori) to 611 days (RU Tau). Obviously there is no single cadence that can cover them all, so we can dismiss that idea from the outset. Indeed, for each of the > 200 Miras I cover, I in fact adjust each Mira's individual cadence for its very own VSX period. This does not seem "lax" to me, even if it causes some very fast Miras to be (properly) observed every other night.
Yes, we may have a problem with daily observations where they are not warranted. Yes, it may sometimes arise from observers' inexperience--asking them why they're doing it is probably in order. They may well have a reason; and if they don't, let's make it a teaching opportunity by the LPV Section. But to callously trash others' hard-won Mira observations simply because there's been a gap of less than 5 days, without even considering the Mira's period or other perfectly good reasons one might do frequent observations (e.g., phase recovery, fine structure in part of the light curve, or even questionable classification as a Mira), would damage AAVSO's database--and discourage observers both experienced and not--far more than would a redundant submission here and there.
Perhaps a more constructive approach: display a warning on the submission review page whenever a LPV/Mira observation follows an gap (since the user's own last observation) of less than, say, 1% of its (putative) period, complete with a reminder link to the appropriate Manual page and maybe the e-mail address of the responsible Section contact for any clarification.
Eric, I can hardly imagine that you are the only AAVSO member observing the majority of your 200, or so, Miras. That being true, it is not glaringly necessary for you to arbitrarily adjust cadence such that your observations become the dominant input on the star. Such over-observation is exactly what I pointed out in my previous post and something I see as doing more harm than good to our light curves.
As I already pointed out, HQ never found it necessary to encourage the disregarding of the prescribed cadence of once per 7-10 day interval for decades and, in fact, the figure of 7 days for Miras was boldly represented on-line within the past 5 years by our recently retired director. I have a copy of it covering all the common types of variables on the AAVSO program pinned to my computer. So it makes no sense for members to be making up their own cadenence, especially with intervals of only 1 or 2 days separating estimates, which I see with surprising frequency when examining lightcurves.
And under no circumstances should an observation cadencce rate of a mere 1% of a given Mira star's period ever be regarded as acceptable. Unless you are the sole observer following that particular star, I see anything less than a 7 day cadence a waste observing time, if for no other reasons than introducing bias into the derived estimate figures and giving excessive weight to a particular observer's input when compiling a mean magnitude for a given interval.
J.Bortle (BRJ)
Agree that Mira cadences < 1% of period are almost always unnecessary. And agree that Mira observers who habitually breach this guideline should be called on it. Again, maybe a good role for the LPV section.
But that aside, adjusting observing cadence to Mira period is the very most efficient approach, yielding the most real info for the fewest number of submitted data points. While the traditional "one-size-fits-all" cadences are better than no guidelines especially for beginners, they are not optimal when Mira periods range so widely. I'm not "disregarding" the published cadences, I'm adjusting them to fit reality. Now, other observers may consider applying a reasonably faster cadence to a fast Miras "over-observation" and a "waste of time". Great, they need not observe that way.
Since you brought it up: it makes no difference to me whether I am the sole observer for a given Mira. My software can schedule an observation only if the time since the most recent reported CCD observation in V exceeds a period-fractional threshold (yes, I check > 200 Miras before every observing night). And that gap threshold is always >1% of the period, almost always 2-3%.
So depending on what others are doing: for some well-observed Miras I end up observing mostly to fill gaps [AH Ser, AL Boo (fast Mira), AQ Aur], and for others I end up submitting most or all CCD observations [AD Dra, AI Cam, AI Her, AK And, AL Cep, AL Dra, AL Per (fast Mira), AM Gem]. I invite you to check some of these targets in LCG to judge for yourself whether my cadences are unreasonable or my observations merely cluttering up AID.
Whatever works, for all of us to construct a good lightcurve together. Should another CCD observer suddenly move into a Mira with regular cadence, I automatically give way. If they abandon the Mira, even temporarily, I automatically fill in without delay. It's unclear how one's scheduling could be more efficient than that.
Yes, I fully agree that my cadence and scheduling adjustments are not "glaringly necessary". But I'm not aiming for minimum acceptable practice. I'm aiming at the best possible aggregate lightcurves.
Thanks John (BRJ), you are only one who is understanding me. In the days of Janet, this won't be possible. Anyway, I don't want to discus this anymore. I made my point.
Cheers to all, Hubert
But where is the problem? It is always possible to "correct" this problem (if it is really one) by downsampling the datapoints for some observers by randomly dropping data points, when you are analysing observations (not dropping them from the DB). In this sense, no real damage can be done to the data of any star by "too frequent observations".
I guess some obsevers try to pinpoint the exact time of minimum/maximum to compare with long time trends.
Also if you happen to live in a region of this planet where clear skies are not the rule but the exception, you think twice about skipping a good night when the next good night might be a week or two in the future. You just observe every clear night that you get. around a maximum/minimum.
All that isn't irrational to me.
Also I must say that I think it's a bit rude to start a discussion, wait for someone to support your stance and then declare the discussion closed for you and leave the room, so to speak. Discussion requires being open to other opinions or it's just standing on a soapbox.
Just my 2 cents
EDIT: PS.. It should also be noted that back in "the old" days, the situation was a bit different: when observation records were kept in filing cabinets->punchcards->tapes->diskettes, entering and storing observations came at a non-trivial cost, and there was a rationale for being a bit economical about observation cadence.
HB
I agree entirely with Hubert and I also perceived yet this kind of cadence in some observers. I ever tried to keep the aavso cadence along the years I've been observing and in some cases nobody knows for shure if , in fact, these observers are more focused in keep a 'highter number of observations', sacrifying the aavso recomendations. Miras with low periods deserves more cadence but I perceived that this happening in stars with longer periods. I know that this is a polemic discussion but I never conformed myself with this kind of behaviour, that come not only from the novices but also from the experienced ones. In the past I testemonied a case of an observer that made diary Mira observations cadence, including under severe moonlight. He acquired 'thousand of estimatives' in few time. This observer, knowing in antecipation that the Mira was clearly much under his instrumental limit, insisted in report a negative estimative everyday along months. Its easy, no?
Eric, in perusing the light curves of the stars you cited in your post, I note that your own observing cadence in nearly every case does not violate the 7 day cadence rule. Likewise, these seem to be largely stars that others are not following at all, making you their sole observer. This thread was started to address the excessive - nightly or almost nightly - observation of Miras and related variables by individual observers of stars which were already receiving modest coverage by others. The two situations are worlds apart. I see no reason that you would then be attempting to justify any unnecessarily rapid cadences (you had suggested even 1% of a star's period in you posts) when you seem to be abiding by the AAVSO's recommended cadence interval of 7, or sometimes more, days in nearly every instance yourself! And while I agree with a cadence of 3% of the star's period is fine for most Miras with periods over 250 days, it quickly becomes decidedly too short for those with significantly shorter periods.
J.Bortle (BRJ)