[vsnet-alert 10924] Re: re FWD ATEL 1892

[qso at Safe-mail.net]

-------- Original Message --------
From: Michael Linnolt <linnolt at hawaii.edu>
To: qso at Safe-mail.net
Cc: vsnet-alert at ooruri.kusastro.kyoto-u.ac.jp
Subject: Re: [vsnet-alert 10912] Re: re FWD ATEL 1892
Date: Tue, 06 Jan 2009 11:23:12 -0800

> Ok, Well thanks a lot for clarifying all the conditions and assumptions under which this transformation should be good. I should add that when doing calibrations, it is important to include a sufficient range of different stars and observers/equipment to make sure you are acounting for many of the important variables, such as: Filter charcteristics, detector responses, optical system characteristics(especially coatings for such a wide range in wavelengths) Source emission characteristics, interstellar and atmospheric absorption differences between 0.5, 1.2 and 2.5u... etc.<

Indeed, but why mention these things?  These are universally applicable matters irrespective of whether someone uses a "V" magnitude from YB6 or some ~ V from CMC14/2MASS, so it adds no extra to which way to do it or what system to use.

Where do you think Bessell and others get their transformations from?  Instead of using a handful to at most a few dozen Landolt and/or Stetson stars I tend to use thousands from Brian Skiff's critically compiled loneos.phot, and instead of doing ever increasing polynomials I tend to stick to the first linear fit.

I mean, you hear folk bang on about "must use Bessell" instead of the original formulation for converting Tycho VT and BT to Johnson V, but when one goes to read the paper itself Bessell himself says something like it only being about 0.01 difference for most normal colour range stars.  Same with some of the SDSS to V conversions, improvements are in hundredths of a magnitude only.

But I don't do anything different than these people do, it's all basic standard simple stuff, you know, nothing clever.

Folk have been resistant for years in using CMC14 magnitudes, no matter who tells them, but at the drop of a hat they use USNO B1.0 and USNO A2.0 red magnitudes in plethora of ATELs, GCNs and MPECs, and call it photometry (which technically is correct, photometry is photometry, it says nowt about quality and accuracy inherent to the results).  Not all of them are recalibrated later when some decent field photometry comes along.  For MPECs, probably none.

And it was clear that as soon as folk saw these V magnitudes in a nice column, marked V, in vizier versions of NOMAD, which has a rabid flocking to it for astrometric solutions for again no reason they can ever justify (usually it's claimed it has stuff specially selected for astrometric qualit, the best of the constituent parts, whereas the NOMAD readme itself directly declares that it has not been critically generated), that they'd soon start using it for the magnitudes, and if there's a V, they'd take it.  Why?  I've no idea.  Yale plate mags usually match up to about +/- 0.3 of a mag from true, nearly as bad as usno mags, however that's the Yale themselves solutions calibrated against Tycho2.  The YB6 solution is an independent USNO NOFS calibration (probably against Tycho2 again), but it's unpublished and undocumented, so how knows?.

Yes, the stars in question have to be pretty much continuum at the passbands being viewed upon (well, except for photometric systems intended to measure line strengths, of course), and a big Halpha spike in the r' band will be well dodgey.

But this has no special relevance to r' and J and Ks, it's true for everything involving red mags for Halpha, which ain't that common incidentally, and all emission lines and troughs are true for all stars and all passbands.

Your caveats are granted true, but bear no specific relevance to the current context.

Don't use nomad.  Or do use nomad if you want, but don't make unjustifiable claims upon the poor thing.

Cheers

John

NB Brian noted problems with using CMC14 quite near to the Galactic Plane, one thing I've forgotten to mention is that the CMC14 to 2MASS crossmatching is a bit generous, and at times for |b| very near 0 the 2MASS magnitudes in the CMC14 catalogue will be spurious as the crossmatch is wrong.

Of course, rather than allowing for that and taking that into consideration, most folk will just see that as a good reason not to bother using it, and use some V from somewhere obscure instead, like nomad yb6.  Fascinates me does this science of astronomy, in practice, it really does.  If folk can't automate it and do it without thinking, they don't want to play most of the time it seems, whilst always crying out for assistance to be more scientifically rigorous.  Lots of paper reading and very little empirical testing goes on too, you'd be surprised what you find with a bit of the old empiricism re the datasets and papers out there.

Look how many CCD photometrists use the simple principle of the check star.  Very few.  It's seen as some sort of concept only needed because pep was old fashioned and weak or something.  Nope, it was to make sure the comparison star was constant.  Just as applicable to CCD.

So if you do half a dozen 'check' stars, determine the variables mag using 6 comparison stars, which cmc14 may well furnish you with in a field, you can take a mean and a standard deviation, and/or see a blatantly obvious outlier and not use it.

Transforms from one system to another in the literature, if you read the papers they come from, involve matching the mags of the new dataset against landolt primary standards or stetson secondaries.  I tend to do it against loneos.phot, which of course carries these as well as Henden stuff, and I have checked the CMC14 route against Henden CV sequences, as also with SDSS derived V.  Often within 0.02.  Which is a damn lot better than 0.7, for example.

But of course we'll have to wait and see if anyone measures the true V of this comparison star afore we know.  And if successful, well, then of course, that'll be only one instance, so no doubt negated as a fluke, and the original YB6 V of 13.3 forgotten about.