Some Strategies for Discovering Minor Planets

Some Strategies for Discovering Minor Planets

By Andrew Lowe


We live in an exciting time for minor planet research. Advances in hardware and software have allowed us to observe and discover asteroids that were considered too faint for detection only a few years ago. Of course, this ability is not restricted to amateurs. The big professional surveys such as LINEAR, LONEOS, NEAT, Spacewatch, CSS, and others sweep large areas of the sky every night and are very successful at discovering asteroids.

Although Spacewatch set up an asteroid detection system in the mid-1980s with new CCD technology, the pace of asteroid discoveries accelerated in 1997 when LINEAR commenced operations. The success of the LINEAR concept spurred similar systems: Lowell Observatory's LONEOS, NASA/JPL's NEAT, CSS, and others. These surveys submit tens of thousands of observations of known and unknown minor planets to the Minor Planet Center (MPC) every day. Every night, the MPC's DOU (Daily Orbit Update) contains hundreds of orbits for new discoveries and updated orbits for newly-discovered objects.

It is perhaps only natural to take the attitude that the big surveys are discovering everything in the sky. After all, LINEAR has been in operation for eight years, and every main-belt asteroid (MBA) has had one favorable opposition during that time. And it's true that all of the reasonably bright objects have been found over and over again. Nevertheless, I would suggest that it is possible to discover asteroids in spite of LINEAR and the other big surveys. In fact, in this article I'll talk about strategies for not only maximizing your chances of finding an asteroid, but also using the big surveys to help follow-up your discoveries.

How do I know what I talking about? Well, in my first month using the New Mexico Skies Observatory's Takahashi Epsilon E250 Astrograph, a 0.25-m f/3.4 telescope, I discovered six asteroids. In my second month, I added twelve more. I'm convinced that with the proper equipment and a focused approach, anyone can do the same.

Here are some points to get us started:

1. The more fields you shoot, the better the chance of discovering something. While you don't have to be out every clear night, you do need to put in a concerted effort. And knowing where to shoot will increase your odds of making your own discovery instead of proving up someone else's.

2. You need to be able to image fainter asteroids. If your system doesn't allow you to image mag. 19 objects, you aren't going to find much. A 8" (0.20-m) scope is probably too small. To find anything faint, you'll spend too much time shooting long exposures. I get good results with the 0.25-m Epsilon. If you have a bigger scope, all the better.

3. Shoot multiple exposures and stack them at a typical asteroid velocity. The stacking will build up signal and reduce random noise.

4. Reshoot the same field at least three times over the evening to detect any moving objects among the stars. But don't reshoot the exact same coordinates. Dither (or slightly shift) the field north and south 10" so that any defects related to the CCD chip are easily recognized.

5. Review your fields as soon as possible, certainly within 24 hours, and get the astrometry submitted to the MPC. Other surveys may be shooting the same area of the sky, and if you delay in sending in your results, you may find that someone else has beaten you to the discovery.

6. Keep on top of your discoveries. If LINEAR photographed your latest discovery last night, there's no need to shoot it again tonight. Try to reobserve if a discovery hasn't been observed in a week. Get a third night for any objects still in Vaisala class (an assumed orbit based on only two nights of astrometry). Remember that the MPC will attempt to link a new asteroid to its database of unknown objects, but if the orbit arc is less than 30 days or so, the orbit won't be accurate enough to do a proper linkage.

Let's get down to details.

Everyone knows that the moon plays a big part in sky brightness. While it is possible to find asteroids when the moon is up, it's a real challenge, and by the time the moon is within a few days of full, basically the sky is so illuminated that it's impossible to find anything new. After a few days past full, the moon is starting to rise after the end of twilight, and you can shoot a few fields in darkness. These fields are, however, well past opposition, so they are unlikely to contain any new objects. If you have discoveries from the previous month, now is the time to shoot them. They may not have been observed for ten days or more, and they may have strayed some distance from their predicted locations.

As the days pass, the moon is rising later and later. This is prime discovery time. Areas just ahead of the opposition point contain asteroids that have been brightening for at least ten days in the glare of the moon. The big surveys are waiting to pounce. If you can get there one or two nights earlier and submit your discoveries to the MPC, the big surveys will follow and confirm your discoveries soon after.

Here's one of the most common questions I'm asked; "Is it better to submit one-night or two-night discoveries?" To which I respond; "It depends." If you submit a discovery with two nights of astrometry, it has higher priority compared to a discovery with only one night of astrometry. Let's say you are shooting in an area where you know LINEAR was active the night before. If you something find new on one night and send it in, all you've done is proven up a discovery for LINEAR. If you want to keep this one-night discovery, don't submit it right away. Wait and get a second night. Submit both at the same time and with a bit of luck, you'll get the discovery credit and LINEAR provides a bonus third night.

That's how it should work in theory. In practice, however, things are always more complicated. Perhaps LONEOS sweeps through the area the same night as your first observation and submits the second night for the LINEAR object. Then you'll lose it fair and square.

In the case when you are shooting as soon as the moon is out of the way, it's a good idea to send in one-nighters. The big surveys will be passing through in days, and if you've swept as much of this area as possible, they will provide the second-night confirmation.

Now let's say it's a week or two later. The moon's been out of the way for a while. The big surveys have made a few passes across the opposition area, and you find that one of them was in your area a few days ago. What to do now? If you find something new, wait and get a second night. Submit the two nights together. You'll have priority over the isolated single-nighters, and you'll be able to use them to improve the orbit.

In a way, it's easy to know when to shoot -- no twilight, and not too close to the time of full moon. But where to shoot? If you are going to spend a good part of the night looking for new asteroids, it makes sense to take a few minutes earlier in the day to get prepared. While no one can tell exactly where a discovery is waiting to be found, you can improve your chances if you know where not to look. Let's start with that approach first.

Unless you are trying to find an asteroid with an unusual orbit, your best bet is to search along a band within 10 of the ecliptic. Sure, there are some objects further from the ecliptic (my 1991 TJ15 was recovered in Jan. 2005 at a declination of +72!), but if you are trying to find something new, you'll have better success closer to the ecliptic. Another suggestion: ignore the area past opposition. Asteroids in this area were being picked off in the previous lunation by the big surveys. Even if one did manage to escape detection, it's already starting to fade out, and you'll never observe it long enough to determine an accurate orbit.

So that leaves the area ahead of opposition as your prime discovery hunting ground. Where exactly you decide to shoot is up to you. But let me offer a few pointers. If I can get an asteroid's span-of-observations up to 30 days, then I feel that its orbit is in good shape to be recovered at a subsequent opposition, or to be used by the MPC for precovery efforts (when observations at a previous opposition are shown to link to the asteroid). Now it's a lot easier to find a faint asteroid 20 days after opposition when it's already discovered and you know where to look, compared to 20 days before opposition when you are looking for unknown objects. One strategy I follow with a new discovery field is to shoot about 10 days ahead of opposition, and then hope that any discoveries are bright enough to follow for the next 30 days, to about 20 days past opposition. Of course, the specifics will depend on weather and the phase of the moon, but it's a good starting point.

Let's suppose this plan is reasonable for you. How do you find out where this area of the sky is? One approach I use is to type in initial coordinates in the MPC's MPChecker at http://scully.harvard.edu/~cgi/CheckMP . This service will give you a list of known asteroids in the field. Choose one of them and enter it into the Minor Planet Ephemeris Service (MPES) at http://cfa-www.harvard.edu/iau/MPEph/MPEph.html . The ephemeris for this object will show an upcoming maximum in the elongation from the sun, which roughly corresponds to opposition. Compare that date of maximum elongation to today. Move higher in right ascension if you need a later date, lower in right ascension if you need an earlier date. Eventually, you'll find a point about 10 days ahead of opposition where you stand a good chance of finding something. If you've got one of the many planetarium programs for PCs, you can do the same thing if you've got some asteroid elements loaded.

Unless you have a 0.5-m to 1.0-m scope, there's another fact of asteroid hunting that you need to consider. In all likelihood, the big surveys have been sampling that same area ahead of opposition for their NEO work, and they've been reporting any detections of MBAs. If you want to prove up discoveries that they made days or weeks earlier, that's fine, but otherwise you're going to have to avoid areas where they've been shooting. There are two resources that will help you here:

1. http://scully.harvard.edu/~cgi/SkyCoverage.html is the link to the MPC's Sky Coverage Plots. These plots will show you where the big surveys have been recently shooting. The color coding by date is particularly useful, because it will show you the areas that have been shot on consecutive or nearly-consecutive days. These should be avoided, because discoveries from the first date will be linked to those on the second date, leaving nothing for you to find. If a particular area has been shot on only one night, then it might provide you with an extra night of astrometry if you submit two nights of astrometry for a discovery.

I made my first discovery of 2005 -- 2005 AC11 -- by choosing a field in a gap in the sky coverage plot. Unfortunately, this strategy was a bit too successful. The only contribution from the professionals was one night of astrometry from CSS, and it duplicated one of the eleven nights of astrometry that I acquired. 2005 BA, a discovery I made while tracking 2005 AC11, was even worse. The big surveys completely missed it, and I ended up getting all of the eleven nights of astrometry.

2. Remember that list of asteroids that popped out of MPChecker when you were trying to find the area of sky 10 days ahead of opposition? Let's go back to that list. The right-hand column, which is labelled "Further observations?", will contain a range of dates when additional astrometry would be useful. If the range of dates starts with today's date, this means that the object hasn't been observed at this opposition, or hasn't been observed for over a month. So perhaps a survey hasn't been through the area in some time. But don't just focus on one asteroid. Look at the whole list. If it's too short, choose a bigger "Radius of search" in the selection menu. If all the objects show today's date in the "Further observations?" column, it means that the area hasn't been scanned yet by the big surveys. Otherwise, take the list of asteroids and put it into MPES. Turn on the "Show residuals blocks", and check when the asteroids were last observed. If it was less than ten days ago, chances are that the MPC will link your one-night discovery with any one-nighters from the earlier date. You will need at least a two-night discovery to take priority over any earlier one-nighters. Also note that if the asteroids were observed on more than one night over an interval of less than 10 days, then any potential discoveries have probably already been linked. You can confirm this by checking the "Orbit" column in the MPChecker output. Any object showing a V is classed as Vaisala, which means it has been observed already on two nights during the present opposition.

This is a real mountain of information, so let's condense it down to the essentials. Use MPChecker for the status of a field that you might be interested in shooting. If the asteroids in the listing have today's date at the start of the "Further Observations?" column, then they haven't been observed in the last month, and there may be discovery candidates in the field. As a final test, see if there are any objects with V in the "Orbit" column. If not, then there are no asteroids observed on two nights in the field. You can see that MPChecker is a good way of determining if a field may have good potential for discoveries.

Once you have determined where to shoot, you need to get good-quality images. The bigger your scope the better, but there is hope even with a smaller instrument. The Takahashi Epsilon E250 0.25-m (10") f/3.4 Astrograph at the New Mexico Skies Observatory can easily reach to R=19.5. You need to shoot multiple shots of the same area of the sky, then blink them in order to show any potential movers among the stars. Single exposures don't give me adequate signal, so I shoot five exposures of 90 seconds each, one after the other. Then I stack them together at the velocity of a typical MBA in the field. I get the velocities from the MPChecker listing. Specify speed and direction, not separate motions in Right Ascension and Declination. Then get a second and third stack at least 15 and 30 minutes later.

There are many great software packages for processing the data and measuring the positions of asteroids. I've used Astrometrica myself for many years, and I'll refer to it in the processing steps. It's a great program for the beginner.

I like to book two hours of observing time with the Epsilon, and shoot four discovery fields. Each stack below is a sum of five exposures of 90 seconds each:

  • 1. Shoot field#1 stack#1
  • 2. Shoot field#2 stack#1
  • 3. Shoot field#3 stack#1
  • 4. Shoot field#4 stack#1

  • 5. Shoot field#1 stack#2
  • 6. Shoot field#2 stack#2
  • 7. Shoot field#3 stack#2
  • 8. Shoot field#4 stack#2

  • 9. Shoot field#1 stack#3
  • 10. Shoot field#2 stack#3
  • 11. Shoot field#3 stack#3
  • 12. Shoot field#4 stack#3

    The multiple aiming-and-shooting processes can get complicated, so it's great that the Epsilon has a scripting option that allows you to prepare your observing procedure ahead of time in a simple text editor. I copy and paste the commands in the script area, start the script, and let the scope follow the commands for the next two hours. Very convenient!

    A critical step when acquiring the data is to dither, or slightly offset in position, the center of the field each time you reshoot. For example, I will shoot stack#1 at the nominal position. When I return for stack#2, I offset the field 10" to the south. When I finish with stack#3, I offset the field 10" north from the first stack. Any defects on the CCD chip will jump up and down during blinking, so you can immediately discount them as possible asteroid images.

    Let's skip ahead a few hours. You've acquired your data and it's been transferred to your hard drive, ready for processing. Before you start Astrometrica, download the latest version of MPCORB.dat. I've never been able to download from the MPC because of their security settings, but http://mpcorb.klet.org/ always works for me. MPCORB.dat is a listing of all the asteroids that you are likely to encounter on your discovery field. More in a moment. Fire up Astrometrica, and take a look at your "File/Settings...". Did you save ASTORB.dat to the same path that's listed in for MPCORB under "Environment"? If not, reload the file, change the path, or transfer the file. Nothing like working with a version that you find out later is a year old. We'll assume that you've reduced FITS files before in Astrometrica, and your numbers for the scope and the CCD are good. What about "Time in File Header"? If you have it set to "Middle of Exposure", you've made your first mistake, and joined the ranks of at least one unnamed ICRAR member who submitted astrometry with incorrect times to the MPC (hint: his initials are A.L.). The keyword DATE-OBS (which actually contains the time and date) in the header refers to the start-of-exposure. So you have to select "Start of Exposure".

    Have you ever shot on one side of the meridian, then on the other side, and then can't figure out why the Astrometrica parameters for one side don't work for the other? It happened to me on my first night on the Epsilon. I was shooting some asteroid targets in the east, then Comet C/2004 Q2 (Machholz) in the western half of the sky. After scratching my head for a while, I figured out that for the western views you had to invert them by switching on the "Flip Horizontal" and "Flip Vertical" options. The Epsilon is mounted on a German Equatorial mount, and has to be flipped around when it crosses the meridian. So you are going to need two different configuration files. In the eastern half of the sky I use a file named "Normal start.cfg", using the default orientation and the start-of-exposure. In the western half of the sky I use "Reverse start.cfg".

    You might be thinking at this point that there could be problems if some of the files for the discovery field were shot in the east and some were shot in the west. Well, you are quite correct. If you want to blink the files, they have to be processed using the same configuration, so you can't blink east files and west files. If you have MaxIm, then you can rotate the western files by 180 and then you are all set. If you don't have MaxIm, then you will have to plan your evening to have all the files shot on one side of the meridian. Not so easy if you have clouds and you have to abandon your carefully-prepared script and shoot when clear areas open up.

    If you've followed my shooting procedure, you'll have fifteen individual FITS files; five for each of the three individual stacks for the discovery field. Load your dark and flat frames. Click on "Track & Stack..." to select the first five exposures for stack#1. When you get to the "Coordinates, Tracking and Stacking" box, you need to add a generic asteroid speed and direction (P.A.) in the "Object Motion" section. These are the numbers you should have written down when you were using MPChecker earlier in the evening to figure out where to shoot. Since I never remember to write them down, I'll assume you forgot too. So go back and redo that step. Then let the program do the stacking. When the stack appears, click within the field to get the mid-time of the five stacks. Write it down. If you have something to submit later, you'll need this time as a check.

    At this point, it's useful to note that there is a bit of divergence if you are shooting MBAs vs. NEOs. In the case of the fast-movers, you'll be measuring from the display. Our NEO hunters Robert, Jeffrey, and others have shown that Astrometrica can handle the huge trailing of the reference stars and produce some excellent astrometric results. For MBAs, though, I prefer to save the stack display as a new FITS file. I find that I return often to the field to check for new objects. It's a bother to have to reload the raw data, darks, flats, recheck the proper stacking velocity again, and then restack. I'll repeat the whole stacking process for stack#2 with exposures 6 to 10, then stack#3 with exposures 11 to 15. After saving the three stacks, you'll have three files with names that should tell you something about their processing. Here are the three stacks that contained the discovery images of 2005 CD62 and 2005 CE62:

  • SUM - DISCOVERY4 1 2005-02-14 06-39-12 0.65 295.0 NORMAL.fits
  • SUM - DISCOVERY4 2 2005-02-14 07-13-53 0.65 295.0 NORMAL.fits
  • SUM - DISCOVERY4 3 2005-02-14 07-48-27 0.65 295.0 NORMAL.fits

    SUM tells me that they are stacked files. DISCOVERY4 was the last of the four discovery fields. 2005-02-14 is the UTC date. 6:39:12, 7:13:53, and 7:48:27 are the three UTC mid-times. 0.65 295.0 is the speed and direction I used for stacking, and NORMAL reminds me that it was shot in the eastern sky.

    Since these new FITS files are referenced to mid-time, you need to use new configurations in Astrometrica where the "Time in File Header" is set to "Middle of Exposure". And yes, you need a different set for the eastern sky and the western sky. So you will be working with four configurations, as follows:

  • Normal start.cfg -- raw data for stacking, acquired in the eastern sky
  • Reverse start.cfg -- raw data for stacking, acquired in the western sky
  • Normal middle.cfg -- stack, eastern sky
  • Reverse middle.cfg -- stack, western sky

    It's been a lot of work to get to this point, but your perseverence is about to pay off. Load the three stack FITS files, and confirm that the mid-time for each corresponds to the time you wrote down at the end of the stacking processing.

    More to come ...