Asteroid Photometry

Over the years, I have done many facets of amateur astronomy including visual observing, film photography, CCD imaging, variable star observing, and occultation and graze timings.  In recent years, I have developed an interest in asteroid research including astrometry - the recording of positions to determine orbits; and photometry - the recording of brightness to determine rotational periods.

What got me started in asteroid work is an article written by Dennis di Cicco of Sky and Telescope magazine on just how easy this is to do from the city. Dennis related that fact that he could make new discoveries every night.  Things have changed since that article was written several years ago. With the alarming realization that Near Earth Asteroids pass close to the Earth on a regular basis, several professional search programs have come on line in order to identify all of these close asteroids over the next ten years.  These automated professional programs regularly sweep the sky on every clear night and have found most of the brighter asteroids that amateur astronomers used to find. Now, unless you are away from city lights and/or operate a larger telescope, you are not likely to find any 'New' asteroids (I say that knowing that amateurs still find new asteroids every day). 

There is more work than ever to be done. There are now more than 60,000 numbered asteroids, those with a well-known orbit. Rotational periods have only been determined for about a thousand of these asteroids. A brief description of the necessity for measuring asteroid light curves by Dr. Alan W. Harris can be found here.

I subscribe to the Minor Planet Mailing List which gives a lot of good ongoing information on the subject of measuring asteroids. The loose group of professionals and amateurs that subscribe to this list, also host an annual workshop. A workshop on photometry (which I am on the organizing committee) is hosted by the Society for Astronomical Sciences (formerly the IAPPP Western Wing).

Current Program 

Normally installed in the observatory is a C11 on a Losmandy mount controlled by an Astrometrics GOTO System.  Currently, a C14 on a Paramount temporarily resides in the observatory until it is moved out to the desert sometime next year.  An SBIG ST9 CCD camera is used with a focal reducer providing a F/5.8 focal length.

The main reason for configuring the instrument, as it exists, is to allow for automated data gathering.  The telescope takes an image of an object, slews to a different section of the sky, takes and image of another object, and so on.  The data can then be reduced while the telescope is operating, or the following day.  The automatic nature of the equipment allows for much greater data being gathered than ever before.  On a typical night, 80 to 150 CCD images are gathered of various asteroids and comparison stars.  The software also controls the Optec filter slider. 

Obtaining asteroid light curves is done by obtaining observations of an asteroid over several nights or weeks and charting the change in brightness on an X-Y chart.  By charting where the peaks and the valleys on the light curve lay, you can find a repeating pattern that allows you to determine how fast the asteroid rotates.

The average main belt asteroid looks something like a potato. Therefore, during each rotational period, it shows two elongated sides to us, and two shortened ends towards us.   During those times, the light curve will show two bright peaks, and two dim troughs. This average asteroid will rotate once in an eight-hour period, allowing you to see most of the curve with one night's data. Others, can be much more difficult. Some of these pesky asteroids will take several days to rotate just once.  That means you need night after night of data to build the lightcurve.   Even worse, is when the asteroid’s rotational period is close to 24 hours.  Here the asteroid will show nearly the same face to you every evening making it difficult to complete the entire lightcurve.  In this case, a collaborative effort with observatories spaced around the Earth is advantageous. I often collaborate with Glenn Malcolm, Bob Koff, Stephen Brincat or Brian Warner

Typical Observing Run

The target Asteroid is first chosen determining which targets are within range of the telescope and are favorably placed for a long, overnight run. Asteroids with well-known lightcurves are eliminated by referring to Minor Planet Lightcurves Parameters list posted at the Brian Warner's CALL Site. I tend to pick dimmer asteroids, in the 14th magnitude range, to work on, even those the curves may have a bit more scatter than the brighter objects.

 

The observatory is in heavily light polluted, suburban skies.  Exposures of 90 to 180 seconds will result in an adequate signal-to-noise ratio on asteroids down to 15th magnitude.  A longer exposure requires guiding, and adequate guide stars are not always available.   

 

Accurately measuring the asteroid brightness is a tricky task. You have to compare its brightness to that of accurately measured stars. The distribution of those accurately measured stars is much sparser than you might think. In a ten-degree square section of the sky, there might be a couple of stars that have had accurate photometry done on them. Furthermore, they certainly don’t lie in the field of view of your target asteroid, which then has the nasty habit of moving from field to field, night after night.  Also, those stars might not be in the range of brightness of your asteroid.  All of the rest of those stars plotted in your star charts are only good to within ½ magnitude – if you're lucky!

The reduction program I use, Canopus, by BDW Publishing, requires an initial measurement of high accuracy stars to set the Magnitude/Intensity Relationship.  This measures the performance of the system and allows for proper interpolation of the measured results.  I slew the scope to a field close to the asteroid with stars of medium to high accuracy magnitudes. These fields are taken from the LONEOS catalog produced by Brian Skiff at Lowell Observatory, Flagstaff, AZ. The system loads an image of the LONEOS field, aligns the image to a catalog of stars, and measures every known star in the field.  Then it discards measurements that are substantially off the predicted magnitude.  The result is a sequence of measurements of between 4 and 30 stars in the red magnitude band.   From this, a linear relation of measured intensity vs. magnitude is computed and applied to subsequent measurement.

The level of accuracy you must obtain is quite stringent. The typical asteroid has an amplitude (change in brightness) of just a few tenths of a magnitude. The professionals like to therefore measure to 1/100th of a magnitude in order to generate a smooth, repeatable lightcurve. The amateurs who endeavor to do this work, although seeking to obtain the Holy Grail of ‘1% Photometry’, are usually satisfied in getting a 2% - 3% of a magnitude accuracy.

The C11 is controlled using MPO Connections, a telescope/camera control and scripting program written by Brian Warner Publishing.  The C14 and Paramount are controlled by TheSky and CCDSoft from Software Bisque and scripted by a custom program from Brian Warner.

 

Analyzing the Light Curve

Finding a match on the rotational period is also very tricky.   I use Brian Warner's Canopus software as it is the only Windows based software I am aware of that will do this function. Since the asteroid moves across different fields over the course of the study, different comparison stars are used from night to night.  As I discussed, these comparison stars do not have well-established magnitudes.  Canopus allows for a zero point adjustment to the average of the comparison stars used in order to prove a good match between segments of the lightcurve.  This technique is most effective on short period asteroids where recognizable features in the lightcurve can be found on most observing runs.

The rotational period determination is accomplished using a routine developed by Dr. Alan Harris at JPL. This performs a Fourier Analysis on the data, allowing different parameters such as number of harmonics, period, size of period steps, etc. to be held constant while others are varied. Finally, it does a least squares fit to determine the most likely period.

Now, you must apply judgment in picking the best fit. As I previously mentioned, most asteroids show a classic two-peaked light curve, which is symmetrical in shape. Often, the program will try to fit a three or four peaked lightcurve, lopsided to one side. While this is not impossible, it is highly unlikely. In an instance such as this, try an incremental harmonic order scan to reduce the possibilities. Perhaps you just need more data.

email: rstephens@foxandstephens.com