MPO Canopus - Ensemble Photometry
One of the big "buzz words" these days in photometry software is
"ensemble photometry." There can be many definitions but, in general, this can
be taken to mean that two or more comparison stars are averaged to find a single value.
This value is then used to find the difference between the target and the value - thus
"differential photometry." How many stars are used and how they are averaged is
where the big differences derive.
Honneycut produced a sophisticated method that does not require all stars in the
original comparison star set to be in every image and gives weights to each star. A
simpler approach, one used in MPO Canopus, is to use up to five stars and find the average
of the instrumental or derived magnitudes. In either case, the error in the value
decreases with the sqrt(N), so that doubling the number of stars reduces the error to
about 0.7. Going from one star to four stars cuts the error due to random noise to about
half. By having a more stable reference set, the differential values have less random
noise, thus improving the overall photometry.
Another significant benefit to using more than one star is that you have one or more
"backups" should a selected star be variable. One of the important features that
makes MPO Canopus stand out for photometry is that not only does it use multiple
comparisons but it plots the comparison star data so that you can check on the quality of
the comparisons.
The screen shot above shows the Sessions form's Comparison Plots page. Here, the raw
data of the #1 comparison has been plotted. The curve shows low noise and the expected
brightening of the star as the field rose higher in the sky. However, just the raw plots
alone may not be enough. Canopus goes a step further by doing differential photometry on
the comparisons, a result of which is show in the screen shot below.
In this case, the average of the other comparisons used in the given session was found
and then subtracted from the value for Comparison #1. Again, the scatter is low, about
0.02m and, more important, the general trend of the plot is a flat line, meaning that this
comparison was not variable (and - most likely - neither were any of the others).
Should you ever see an average plot like this:
You should be very suspicious! At the least, remove this comparison from the set and,
better yet, measure it as a target after you're done with the current target.
Using this feature of MPO Canopus has resulted in a number of variable stars being
discovered and then studied, again using MPO Canopus, to determine the type and period of
the variable.
