Skip to main content

Comments on one aspect of J.Percy's paper

We're currently performing upgrades to the AID. During this time WebObs, LCG and any other services that access the AID will be unavailable. We apologize for the inconvenience.
BRJ's picture
BRJ
Offline
Joined: 2010-07-24

I applaud Percy's recent paper that encourages continuing visual observations. As many older members appreciate, there has been no bigger supporter of the AAVSO's visual program over the past several decades than myself. However, one has to use caution in what they are encouraging among observers, lest the true purpose of the observer's motivations go off in the wrong direction.

My case in point concerns a passage in Percy's paper addressing the observation of nebula variables, where he states:

"Back in the 1970's, some AAVSO visual observers began observing these stars. They tend to occur in specific star-forming regions, so they can be observed very efficiently. The observers were able to make many thousand observations of them each year, and thus rank high on the list of top observers. Finally, Director Janet Mattei declared that visual observations of T Tau stars [meaning specifically the Orion Nebula variables] would be devalued by a factor of ten in the annual observer totals. The observations languished, unvalidated."

While Percy and his students would seem to have derived some apparent value from these data after encouraging the Director to validate a portion of them, I would say that one must be very cautious in drawing any conclusions from these data.

Back in the 1970's I participated in the nightly observations of these stars. It was soon recognized that a number of the comparisons stars themselves were variable and the remainder on the charts really were not very good at all. Compounding these problems was the fact that in many instances while the comparison stars were seen on a relatively dark sky background, the variables were viewed against a much brighter nebula veil. That situation, alone, renders truly accurate magnitude estimates highly questionable, if not outright impossible, especially to small fractions of a magnitude. As a result, I soon dropped my observation of these stars to concentrate of more worthwhile stars.

I would also have to perhaps question the motivation of several of the observers of the time who were making brightness estimates every ten minutes or so over the course of hours. This resulted in thousands of observations likely made more for the purpose of generating numbers than a real desire to gather scientifically valid data. Director Mattei also recognized this and after a time curtailed the tallying method with regard to observer totals. I would further point out that when this situation came about most of the prolific nebula variable observers dropped their monitoring programs, to my mind clearly indicating what their actual motivation was.

What is always needed in AAVSO programs is a certain degree of oversight and feedback that keeps observers on the straight and narrow. Over the course of the three decades I edited the AAVSO Circular I saw a number of otherwise promising observers redirect their observing programs specifically to the purpose of generating ever greater numbers, rather than useful data. The situation is still with us today, and is seen in some of today's prolific observers making nightly observations of bright Mira stars, something that serves only to skew the final combined data.

I have long felt that far greater AAVSO emphasis needs to be placed on at what intervals various types of variables should be observed. Excessive observation of slowly varying stars, especially when specifically done for the purpose of generating numbers, needs to be curtailed as it is harmful in several ways.  I know that in the era of Director Margaret Mayall such guidelines were set on submitting observations. Those data that violated the rules was dropped. Perhaps such stringent policing might not be a bad idea still today. 

BRJ

 

 

Big universe, finite collection capability: must prioritize
KTC's picture
KTC
Offline
Joined: 2010-12-08

What is always needed in AAVSO programs is a certain degree of oversight and feedback

...

I have long felt that far greater AAVSO emphasis needs to be placed on at what intervals various types of variables should be observed.

When faced with such a big universe, and trying to study it with finite resources...the only logical choice is to make prioritized recommendations for a collection strategy.  (That includes recommended cadence, filters, signal-to-noise ratio, etc.)

Will some of the recommendations turn out to be non-productive?  Of course...that's because we're still studying the universe.  Incomplete knowledge means non-optimum recommendations will happen.

But that's far better than no recommendations at all.

(I don't have a favorite type of star.  I have a favorite process:  science.  That implies plenty of feedback and reassessment...repeat as need.)

re comment on Percy +over observing
paw's picture
paw
Offline
Joined: 2010-08-07

Hi BRJ

I unfortunately see another similar issue today; logging hundreds of <11.0 for a star that never goes brighter than 12.5 mag or so. Like logging <-4.0 for Betelguese! It could happen. But it's not my business, it's HQs.

It is so easy today for HQ to send an email and ask why a person is observing in a certain way, and there may be good reasons I don't know about, that should be the way to do it.

Alan PAW

P.S. I observe some old novas at  <13 or 14, and the reason is that more old novas are showing DN outbursts over time. Its my time to waste! Its only a minor part of my program.

Over-observing vs. accuracy
lmk's picture
lmk
Offline
Joined: 2010-07-23

This cadence topic has been discussed before on the old discussions. My feelings about it haven't changed, really. I dont believe its any significant concern, PROVIDED, every observation is made to the maximum accuracy and in unbiased fashion.

From a purely statistical point of view, the more observations the better, PROVIDED, they are all "independent" and free of systematic error. Observing a Mira daily, which changes by less than 0.05 magnitude or so, isn't the best use of one's time, but as long as the observer doesn't bias his observation by the expected trend, and makes each such observation as accurately as every other observation, it does no harm, and maybe a little good, since you are increasing the "N" value. More "N", the count of data points, tends to improve statistics by SQRT(N). Thats always good.

Its only bad if the observer adds his bias of the known trend of the star, to each observation. I personally dont know how often that happens, and to what degree it is a problem. John Bortle is of the opinion that observers cannot be trusted, and thus should not do a daily cadence on a Mira, but is this fear borne out by the facts?

Mike LMK

Data quality in general
PKV's picture
PKV
Offline
Joined: 2010-07-26

This is indeed an interesting thread.  Percy's paper was very good and seems to bode well for the future visual observations.  I knew about the past T Tauri observations and the 90% penalty in the past and now know the reason.  Thanks BRJ.  In reading Doc Kinne's very recent paper, I was amazed to see that Headquarters at one time gave feedback about bad data points and seemed to offer remedies and guidance, real time QC'ing of the submitted data so to speak.  And we had a data validation project that QC'ed our older AID data from the beginning up through 2001.  Now it seems that data quality may again be suffering again.

While observer quidance was a minor recent survey issue, the data quality in general and the quality of the AID were even lesser surveyed concerns.  But if one takes a look at selected light curves on the LCG, one sees that potential problems may exist:  the usual scatter on some visual light curves which may be of no major concern, a scattter of +-0.3 or more on some Johnson V magnitudes which may be problematic and over-observation by some select observers is apparent.  There has been an explosion of CCD time series data over time, most of it good, but some it of questional scientific value.  As a result of CCD time series data, several CCD observing awards have gone by the wayside.

One can step back and ask many important questions.  "What is the state of data quality and data QC in the AID?  Is guidance and feedback desireable for providers of poor data?  Is data management of submitted observations required?  Do we continue on "autopilot," accepting all data that is submitted and filter and QC the data in a subsequent data analysis stage only?  What is the solution to over observing?  What are the true quidelines of observation intervals for Mira's, SR's and RV Tau's, as well as other stars like P CYG, GAM CAS and others?  How many samples are required for one cycle of a light curve that looks like a sine wave?  Are observing totals affecting data quality?  Have we now grown complacent and is our current data quality suffering?  All interesting questions.

The AID is the most prized asset of the our organization, followed by its people (members and observers).  Observer training, the CHOICE program, CCD schools and continued education and communication should increase our observational data quality over time.  Currently, are we satisfied with the status quo?  I feel that there is always room for improvement in our data, services and activities (TQM) at the AAVSO.  Even more so with the quality of the AID. 

Kevin B. Paxson - PKV 

 

 

 

 

Percy Paper
SET's picture
SET
Offline
Joined: 2010-07-23

Friends,

I find John Percy's paper VERY encouraging to visual observers. I really get the impression that John feels visual observations of eclipsing binary stars (my specialty) are still important. 

I think that John Bortle has some good points in his post. I have never observed nebular variables since I began observing in 1973. I do remember talking with a few observers who did and they expressing their frustration with the comp stars and they sensed their own observations were somewhat questionable as far as good quality for science.

John, I'm glad you are chipping in on the Visual forum. I actually was going to email you to be a regular contributor to the forum. I really respect you and your years of experience. I for one know I can learn a lot from you. Thanks.

 

Chris Stephan

Robert Clyde Observatory

Sebring, Florida  USA

Comments on Percy's Paper
BRJ's picture
BRJ
Offline
Joined: 2010-07-24

SET wrote:

John, I'm glad you are chipping in on the Visual forum. I actually was going to email you to be a regular contributor to the forum. I really respect you and your years of experience. I for one know I can learn a lot from you. Thanks.

Chris Stephan

Robert Clyde Observatory

Sebring, Florida  USA

Thanks, Chris. You'll probably be seeing me post on the new forums, at least from time to time, to offer what advice and guidance I can to visual observers.

BRJ

Pointless "Fainter Thans"
BRJ's picture
BRJ
Offline
Joined: 2010-07-24

paw wrote:

Hi BRJ

I unfortunately see another similar issue today; logging hundreds of <11.0 for a star that never goes brighter than 12.5 mag or so. Like logging <-4.0 for Betelguese! It could happen. But it's not my business, it's HQs.

It is so easy today for HQ to send an email and ask why a person is observing in a certain way, and there may be good reasons I don't know about, that should be the way to do it.

Alan PAW

P.S. I observe some old novas at  <13 or 14, and the reason is that more old novas are showing DN outbursts over time. Its my time to waste! Its only a minor part of my program.

 

Indeed, Alan, this is something I've also begun to notice fairly recently as a growing practice among a few bigtime data contributors when I've been perusing MyNewsFash and VSNET. Quite honestly it is just foolishness whose only purpose must be to generate empty numbers. In the case of CVs, reporting a negative observation fully a magnitude, or more, above a given star's brightest recorded maximum is of absolutely zero value and indeed should be discouraged in some manner by the folks at HQ.

Currently I don't see this practice as a major problem, but I do see increasing emphasis among some of today's observers to report enormously high annual totals generated in odd ways. Haste in observing (as in making negative estimates that you know must be true, simply becasuse they are brighter than the star ever gets) is just wasted time, and worse yet, potentially a lowers confidence in what the observers' lightcurves show by implying a lack of serious effort on the part of the observer.

J.Bortle   (BRJ)

AAVSO 49 Bay State Rd. Cambridge, MA 02138 aavso@aavso.org 617-354-0484