by Ernst Pollmann » Sat Aug 26, 2017 9:42 am
Dear Marc,
from my point of view it seems to be not easy and not meaningful too, to define average values or standard deviations, since the variability of V&R seems to be (out of eclipse) stochastically.
That´s why our out of eclipse base line is not explicit a strict straight line. It is more or less a tendency line, smoothed on the variability. And our V&R values fit very well in that tendency. Because of that I think we can speak only of a qualitative assessment than of a statistical evaluation. So the question is at all, whether a statistical statement makes sense in this certain case.
Nevertheless, there are error ranges, at least in evaluation, that´s clear. The overall accuracy of the EW and PH determination of V&R depends on:
1) the quality of the spectra taken during the observed night
2) the reduction process (spectrum processing) of the individual observers
3) the determination in the normalized spectrum.
To 1):
I am not able to estimate this order of magnitude, but my own experience says, this is approx. in the order of magnitude (+/-) 3% in EW for one single observation.
To 2):
I am also not able to estimate the accuracy of the spectrum processing of the given observer. This will remain an unknown factor (at least for me).
To 3):
Here I can speak only about the reproducibility of normalization, and about the accuracy in definition of the integration wavelength area. This reproducibility lays in the order of magnitude (+/-) 5% in EW. This is the only thing I can do when drawing the monitoring diagrams: to insert that kind of error bars in EW and PH.
OK, a V/R plot will comes in the afternoon or tomorrow. Concerning to recalculate some ephemeries we should wait what our further observations will say.
Ernst