I made my first observation of the season of the odd INT/FU Ori star V1057 Cyg last evening and after submitting it to the database, called up the star's recent data via the Light Curve Generator. To my surprise there was a most obvious dichotomy between the reported visual and V data. The V data was found to be consistently running about 0.7 magnitudes brighter than the visual estimates. In a re-plotting, this time spanning of some thousands of days, I noted that this has been an ongoing situation of considerable duration.
Now I've been following V1057 Cyg visually for decades, in fact since its discovery. If we are to consider the visual sequence as reasonably accurate, then V1057 hasn't been seen in the mid 12's suggested by the V data in nearly two decades! One, or the other, of the sequences being employed by the two groups of observers has to be decidedly in error. I appreciate that there are a troubling number of such instances to be found in the AAVSO database and that efforts to meld observations obtaining by the two differing methods is being attempted by HQ. I also understand that these days many in the professional community would likely prefer the V data. However, considering the very long run of consistent visual observations available for this star documenting its fluctuations, currently spanning a number of decades, it is obvious that not only the V data and sequence are the ones in serious need of re-evaluation and a downward adjustment, but that the problem with this star's light curve is long overdue to have the situation addressed. If not, than those eventually utilizing or even just looking over a plotting of the data will be significantly misled regarding this rare object's long-term odd behavior.
J.Bortle (BRJ)
.
Hi John,
At least part of the difference may be explained by the fact that the V passband does not correspond exactly to what a visual observer sees. For red stars, such as V1057 Cyg, the difference may be quite large. See e.g. http://adsabs.harvard.edu/abs/1999JAVSO..27...97S.
Patrick
As Patrick mentions, V1057 Cyg is a very red star (APASS says B-V = 1.927), probably because it is embedded in a dark nebula near the North American nebula. For such stars, there will almost always be a difference between what visual observers measure and what CCD observers measure, due to the different bandpass of the eye and of the V filter. I looked at a couple of the V filter observations, and they seem to be using the most recent sequence update (about two years ago), and multiple observers agree with one another. The same thing is true for the visual observers. Both are correct; they just don't match each other.
John makes the good point that researchers need to be aware of such systematic differences when looking at long-term light curves of such stars. There is sufficient overlap in this case that you could adjust one of the light curves to fall on top of the other. This is one reason why I have been requesting continued visual observations, even of stars well-covered by CCD observers, so that we can identify systematic trends and provide guidance for researchers when looking at long-term trends.
ARne
This is similar to the situation with V838 Mon that was discussed several months ago. I'm one of the CCD "culprits" for both of these stars. It's very striking to see the offset between the V and visual curves, but there's pretty good agreement in the rises and dips so they parallel each other pretty well. It's clear that "V" != "visual" for such red objects. Even applying the modest transform correction for my system only changes the measurements by a few hundredths of a magnitude. Obviously the spectral energy distribution for this object is problematic, and the reasonably good correlation between V filter and visual magnitudes for whiter objects fails for these oddballs.
As Arne mentioned, the two data sets aren't at odds - they're just measuring somewhat different portions of the spectrum. Assuming the color index stays fairly constant over time it's a pretty simple matter to adjust one set of observations to remove the offset - the same thing is necessary even when all of the data is in same "filter." CCD observations from different observers frequently have offsets of a few hundredths of a magnitude, and visual observations from different observers may also need to be adjusted to correct for differences in instruments, techniques, and eye sensitivity. It's just one of the tasks researchers do as part of their analysis.
Shawn (DKS)
"Assuming the color index stays fairly constant over time it's a pretty simple matter to adjust one set of observations to remove the offset - the same thing is necessary even when all of the data is in same "filter."
Ah, but therein is the problem, Shawn. V1057 Cyg had been under intense visual observations for decades prior to the CCDers picking it up. I'm sure that while some few profession color index figures spread over the years can be found here and there if one searches long and hard enough, but do/will all researchers even appreciate what is going on here? What happens when, in an effort to create a homogeneous light curve from the two sets of visual and "V" data for illustrating the star's behavior in a published paper, the researcher decides to simply adjust all the older visual data upwards by THE CURRENT CCD-v difference to create a match? I'd anticipate that the resulting light curve will no longer accurately depict what V1057 really did do over the course of its last 40-odd years. It is a big potential problem I see looming on our horizon in the utilization of historic AAVSO visual data. Arne fully understands the problems, but how about the next Director, or the next?
J.Bortle (BRJ)
[quote=BRJ]
"Ah, but therein is the problem, ....
[/quote]
You seem to have identified a problem, but have you any solution for going forward? I have been a major contributor to the CCD data for this star and I've checked the resources available to me for making and reporting reasonably reliable data and can find nothing I would do differently. Would it be better (and have been better) if I stopped (or had never begun) observing this star?
J. Roe
[quote=roe]
[quote=BRJ]
"Ah, but therein is the problem, ....
[/quote]
You seem to have identified a problem, but have you any solution for going forward? I have been a major contributor to the CCD data for this star and I've checked the resources available to me for making and reporting reasonably reliable data and can find nothing I would do differently. Would it be better (and have been better) if I stopped (or had never begun) observing this star?
J. Roe
[/quote]
No, by no means stop observing this star. I certainly would never suggest that approach as a viable solution.
In the case of stars with long visual data histories, but only with recent CCD V data, what might be a possible approach would be having some provision to flag the star's database with the notation that any attempts to combine the visual and 'V' data should employ the visual as the baseline, not the CCD values. In reality we would likely have no clear idea of the star's overall past light curve behavior pattern in 'V', making it dangerous to simply adjust the visual data to link up with the far shorter data string of CCD values.
Honestly, I see more than a few variables in the AAVSO's program that suffer apparently similar distortions/wide separations between their visual and CCD 'V' light curves. In some cases the two curves simply seems to mimic, or parallel, one another a magnitude or more appart. But there are still others where spectral changes must be progressively taking place and the two light curves differ and could never be made to truly overlap and form a single definitive light curve.
In short, for oddly colored stars, or unique emission line stars, where the two observing methods give distinctly differing results, the one with the decidedly longer database should probably be employed as the standard, or baseline, in depicting the long-term behavior, not the other way around. Some stars will likely even be found that forever require depicting their activity with totally separate visual and 'V' light curves.
J.Bortle (BRJ)
Hello
I looked on the light curve and expected to see a classic bias difference between the V and the Vis light curves. While there is a bias, there is more going on here.
Whats even more troubling is the characteristic differences in the Visual and the V curves. The Visual curve seems to pretty much go along at a constant value plus or minus the scatter. When I look at the V curve, it has several minima, supported by several observers, and it does not show up in the Visual data. Could these be 2 different targets? The minima is clearly withing the capability of the visual observers, but they are not seeing it.
Gary
Wgr
John has hit upon an important point here. Visual observations have some crucial advantages over the recent detector technologies (CCD, PEP , etc.) - namely constancy and longevity. Human vision has evolved into a ubiquitous and efficient optical detection system with negligble genetic drift. It is nearly ideal for studying the long-term changes in brightness of variable stars.
Detector technologies will come and go. It wasn't too long that CCD took over from PEP. I can forsee a new set of single-photon and energy counting technologies (such as TES) soon supplanting CCD as the method of choice in Astronomy.
All these detector technologies will have different response characteristics compared to visual. Over the long term, we will have to rely on visual photometry as the underlying, constant standard which will act as a baseline to compare all the new and old technologies measurements, which come and go.
So, we must strive our utmost to never let visual photometry "die". Though it may never be the most "sexy" newest technique astronomers will have available, it is the gold standard for long term studies of optical brightness.
Mike LMK
I don't think we should get sidetracked here. The Johnson V system has been around for 60 years, and all telescopes/cameras/filters that approximate it can be transformed to within a few percent to lie on top of one another. It is not anything new; the sensor might be different, but the results are the same.
I'm also pretty sure that the (B-V) color of V1057 Cyg has stayed the same over its studied lifetime. What is more likely to be the case is the changing photometric sequence, which affects both CCD and visual observers. The two-year-old revision removed a number of red stars and improved the consistency of the magnitude steps. When these changes happen, subtle estimate differences will occur, and such changes tend to happen on the decade timescale. Therefore, any researcher looking at long-term trends on such stars has to delve deeply into the methodology, both for the visual and the CCD observations, and make adjustments. I don't think we should say that one technique gives more consistent results than another. CCD results give higher precision, and have the potential of higher accuracy, and can be compared with visual observations as long as sufficient overlap exists. What we DON'T want is for visual observers to stop observing their favorite star just because some new survey has started - make sure that there is sufficient overlap so that we can keep continuity into the future. The promise of all-sky, every-day, decade-long digital coverage remains just that - a promise, but not a reality.
Gary, I see similar periodic structure in the visual observations, just hidden beneath the higher random scatter of the observations. This again is one of those cases where averaging lots of visual observers would show low-level features that might be more easily seen in the more precise CCD measures.
Arne
If we have also CCD observations in diffrent color, such as B, we might have more information about difference btwee V and visual. I encourage multivolor, or at least two color, photometry for CCD observers.
Seiichiro Kiyota
Hi Seiichiro, hi all,
I'm going to be adding a B and an I filter (Astrodon) to my set up this month.
V1057 Cyg is one of my targets so I'll be certainly adding B filtered data to the light curve.
Douglas.