[Aavso-photometry] What error to report in WebObs
Brad Walter
bwalter at activepower.com
Fri Dec 24 00:32:04 EST 2004
In WebObs, when reporting error of measurements taken using a std
photometric filter (e.g. V band), should the error of the color sequence
values on the charts be included in the reported error, or should the
error only include the sigma from our own observations?
It also seems to me that using the error resulting from the SNR of the
Object star alone is inadequate. The stochastic error of the reference
star (comp) must also be included. where sigma(total) =
SQRT(sigma(obj)^2 + sigma(ref)^2). It further seems to me that you can
get the sigma of the reference star in a couple of ways. If you are
taking measurements from only one image (single integration or multiple
integrations combined), You would use the SNR of the comp star. If you
are taking measurements from a number of images however, would it not
also be valid, and more accurate, to arrive at the sigma for the comp
star by using the std deviation of the check star errors from the
series? Specifically:
sigma(err)^2 =alpha*(sigmabar(ref)^2 + sigmabar(chk)^2), where
sigmabar(ref) and sigmabar(chk) are the means of the sigmas of the ref
and check stars derived from the SNRs of the individual measured
images, and
alpha is a proportionality constant arbitrarily assigned to fit the
data, i.e.
alpha = sigma(err)^2/(sigmabar(ref)^2 + sigmabar(chk)^2).
Then
sigma(ref) =SQRT((sigma(err)^2 - alpha*sigmabar(chk)^2)/alpha)
I would think the second method would be more precise since it is
derived from a "mean of means" calculation. Can anyone comment on the
validity and accuracy of these approaches.
One final issue: should scintillation error be included in the error
reported in WebObs? I would think that it must be since it is probably
the most important stochastic error in evaluating time series
observations, even though it's probably not significant for absolute
magnitude measurements made from stacked images taken over a period of
hours.
Brad Walter
Director of Product Marketing
Active Power
2128 West Braker Lane
Austin, TX 78758
Phone: (512) 744 9414
Fax: (512) 836 4511
bwalter at activepower.com
More information about the Aavso-photometry
mailing list