Long-Period Variables and the AAVSO

As described in my last post, I recently became interested in variable stars.   The American Association of Variable Star Observers (AAVSO) provides a terrific amount of information and guidance about observing variable stars, and the underlying science.   My own excursion into this area began with a somewhat randomly chosen variable known as XZ Cyg, an RR Lyra – type variable with a magnitude amplitude of 1.3 in the Visual band and a period of 0.46 days.   As a beginning observer, I have become increasingly interested in selecting variable stars as part of a personal “observing program.”    There are tens of thousands of catalogued variables.  Where to begin?  A number of personal considerations include:

- the nature of the light curve – for me, and I think for most observers, a certain degree of unpredictability is more interesting.

- The star needs to be obviously visible from my northern latitutude (42 degrees N), and I have trees and houses to contend with as well!

- The magnitude range needs to fall within the observing range of my available equipment which includes:

a) my naked eye, which achieves a limit of 5.95m here in the light-polluted skies of metro-west Boston area,
b) Pentax 10×50 binoculars (mag limit approx 9.50 – 10.0)
c) 10″ Meade SCT (mag limit approx. 13.5-13.7).

To better appreciate the distribution of variable star properties, we can turn to The General Catalog of Variable Stars (GCVS) which currently lists over 40,000 variables.   By coincidence, the AAVSO announced one week ago that they were going to start a section dedicated to long-period variables.   These LPV’s include the famous Omicron Ceti, or “MIRA”, a pulsating red giant star with an amplitude of over 8m, and a period of about 332 days.   One of the first goals of the AAVSO’s LPV section is to identify a selection of stars to be the focus of their observing program and in so doing, provide guidence and direction to the AAVSO member community.    In short, they want to make sure that observers are well-aligned with the scientific objectives of the astronomical community.   Towards that end, the AAVSO has also put together data on all of the LPV’s in their database including the number of observations as well as the number of known publications about each star.   In combination with the GCVS catalog data, one can gain some insight into how observers choose their stars. Note: Click on each image to see full size!

1.  The figure below plots the 70 LPVs having more than 20,000 observations a piece.   The larger the circle, the more observations.   The redder the circle, the more scientific references.   One notices immediately that almost all of the most popular stars occur in the northern celestial hemisphere (>0 degrees declination) because most of the observers live in the Northern hemisphere.   And while in theory my horizon might extend down to -48 degrees if I lived on a volcano in the middle of the ocean, the practical limit here in New England is probably about -20 degrees.   It is worth noting that the six southern exceptions, which includes Omicron Ceti itself, all achieve naked eye visibility at maximum, perhaps enabling more frequent observations.

LPVs with greater than 20,000 Observations

Also interesting is the fact that these 70 LPV’s (out of 16,860 in the AAVSO database) garner over 35% of the LPV observations (2,414,998 /6,856,617).

2.   The next figure plots 59 of the (scientifically) hottest stars,  each having more than 200 references.   I’ve also highlighted (dark green small circles) those stars having fewer than 1000 observations as potential neglected candidates.   One is QX Pup, the focus of a recent AAVSO observing campaign, others however have very small amplitudes or are primarily infrared-band sources.   A strange example of a “neglected star” is U Her, but I believe this is simply a data error in the AAVSO spreadsheets, as there are in fact over 30,000 observations of this star. [11/15/08 Addendum - 18 stars reported to have fewer than about 1000 observations and more than 100 scientific references all have greek letter names - so this is probably a confusion between u (greek mu) and U, for example.   Thanks due to John Greaves for pointing out the likely cause of the discrepancy.]

The scientifically hottest LPVs

3. The above figure does suggest that observation counts aren’t always aligned with the scientific interest in the star based on the number of known references -  see figure 3 below:  R-squared = 0.3029.

Observations -vs- References

Sometimes the discrepancy is simply the result of the fact that the scientific interest in the star is largely outside the scope of its variability.   For example, CW Leonis with over 1500 references is believed to be a rare example of a proto-planetary nebula and surrounded by a cloud of water-bearing comets.

4.  Another major factor influencing observation counts is the maximum magnitude.   This in itself is probably not too surprising, but the steepness of the curve is remarkable.

Number of observations -vs- Peak magnitude

Note in particular that below 10m, my own binocular limit, few stars have more than 10,000 observations  a piece.

5.  We look now at the diversity of M-type stars.    In trying to find interesting LPV candidates for my own observing program, I thought a good starting point would be to better understand how such stars are distributed in terms of period and amplitude.   Being an impatient novice variable observer, I thought it would be cool to note large changes in brightness over the shortest time possible.   Here I have included only M-type stars with magnitudes in the V band (ignoring infrared and photographic), and excluding stars flagged in the GCVS as having uncertain (or missing) min/max magnitude or type.   Again, size is by number of observations in the AAVSO database, and redder means more references.

M-type, V-Band, Amplitude -vs- Period

There does seem to be a shared preference for larger amplitude stars, though not a strong preference for shorter periods.  If you plot number of observations -vs- the Amplitude/Period ratio (plot not shown), there does not appear to be a strong correlation.   On a side note, one of the goals of astronomical data mining is to identify interesting outliers.    R Vulpeculae (marked) along with the southern star W Puppis have  the highest amplitude/period ratios for an M-type star.   R Vul varies over 7 magnitudes in just 137 days!   Lowly V0384 Persei has an Amplitude/Period ratio a factor of 10 smaller (2.9m range in 535 days) which makes it interesting in its own right, though understandably not to your average observer.   Repeating the excercise for other types of variables having well-defined periods is interesting too.

Final thoughts:

Many other kinds of analyses are possible.   I would like to see the AAVSO identify not only the total number of observations, but also the number of unique observers as way of better understanding general community interest.   I’d also recommend additional documentation of the spreadsheets including details on how the observation and reference counts are performed so that the data can be further checked.   I’d also love to see this data provided for every star in the AAVSO database.    I hope the above analysis gives some improved insight into how individuals choose an observing program and what stars might be of interest both to observers and the scientific community.

AAVSO/RJB