Survey strategy


    1.  Survey goal

    The goal of the survey is to discover asynchronous binary asteroids among small NEAs, MCs, and inner MBAs, and to do it in a controlled way that will allow to simulate selection effects and biases in obtained sample. For that we use the photometric technique (Pravec et al., 2006, Icarus 181, 63-93; see also Pravec et al, 2000, Icarus 146, 190-203; Merline et al, 2003, Asteroids III, 289-312).


   2.  Target selection criteria, required photometric accuracy

    Asteroids with D<10 km (H≥12) in near-Earth, Mars-crossing, and inner-main belt (a<2.50 AU) orbits will be targeted. Smaller objects are to be preferred but the minimal requirement of photometric errors of 0.03 mag or better is to be held.

    An observational window of the asteroid for the given station should last for a few weeks; observational window lasting for a week or less is too narrow, as it might be difficult to get sufficient follow-up for the asteroid, if discovered to be binary, in the narrow window.

    During the observational window, the asteroid should be observable at airmass lower than 2 for at least a few hours on each night and the photometric accuracy of 0.03 mag or better should be achievable during it.


    3.  Observing strategy

    There are a few levels (ways, modes) how individual stations can contribute. (Note: It should be quite natural that a particular stations may switch between different modes, even on a scale of days, according to conditions and circumstances --- e.g., weather forecast, telescope time availability.)

    3.1 Two or more good nights in a row available

    - Select a new object. An optimal target is such whose conditions do improve or stay constant for several days at least; don't select a target whose conditions are worsening for you in any sense soon after you start to work it.

    - Cover it on two nights in detail and then see what your reduction shows. Plan further nights if needed, until its period and lightcurve is well established (U=3 though U=2+ might suffice in some cases). 

    - If unable to take follow-up nights while the object's lightcurve is still not reliably established, let other observers know so that someone can take over it (see also below).

    - If seeing something suspicious, communicate the data quickly so that a detailed look can be given and further observations can be planned at other stations and/or with more powerful instruments.

    3.2 A single (isolated) night available

    - Ask other observers involved in the Survey what objects they suggest you to follow up. Such observations should help to establish a lightcurve for the particular object faster and/or with a higher reliability, therefore freeing hands of observers working in the Mode 1 above to turn to a new target sooner.

    3.3  Have a telescope available for another program but able to do critical observations to help with interesting or difficult targets

    - Let other observers know on your availability to perform such observations on particular nights. They may or may not ask you to observe an interesting target depending on a status of their current work. Your critical observations, especially if in different or better conditions (larger instrument, a different longitude) might help to resolve a difficult case therefore saving a larger effort of the others in less favorable conditions.


    4.  Reduction and analysis strategy

    It is essential to have the data reduced fast. As a general rule, the data should be reduced and analyzed BEFORE a following night observations are started, i.e., usually in less than 12 hours since the last observation was taken. In that way, a planning of the next night observations is possible.


    5.  Communicating data while a campaign on one object is in progress

    - If observations of the object develop normal (i.e., data consistent with a single period, no anomaly seen, expecting to be able to take a further night if needed), you don't need to communicate the actual data to others in any way during your work of the object.

    - If you have a suspicion, do communicate it to an "analyst-on-line" (see below) so that a quick action can be taken if needed. (Of course, if you see something really interesting, not just suspicious, you should call the group for collaboration right away, not waiting for an opinion of the "AOL".)

- For objects called for collaborative work, a "principal observer/station" is assigned (see below) and he/she/it collects the data and assures they are handled as needed. Data for the collaboratively worked object should be sent to the "PO/S" as soon as they are reduced.


    6.  Who is "analyst-on-line"; "principal observer/station"; "coverage data collector"

    I plan to assure that there is always one person available on any day serving as an "analyst-on-line". He should be an experienced data analyst familiar with various cases and situations that may happen. He will receive requests for checking suspicious/unusual data, communicate back to the observer with analysis and suggestions on how to go on, and may also call other observers for help if agreed with the observer. For objects worked collaboratively, he will be ready to help the PO/S (who collects data for the particular object, see below). As a general rule, he should reply within 24 hours to the request sent, though we will try to reply faster usually. For some time now, the "AOL" will be always a member of my Ondrejov group; I plan that Adrian Galad will be in duty as the "AOL" on at least 25% and perhaps even more than 50% of days (depending on how we agree on his involvement while he stays at Modra Observatory for two weeks per month) and I should serve for the rest of days. In near future, however, I would like to involve one or two more people; three or four should be an optimum number to cover the duty. The "AOL" can be reached at this email address:

    The "principal observer/station" is by default the one who started the observations (took the first two or more nights data) for the given object. In specific cases, the original PO/S may give over the role to someone else, e.g., due to a limited available time or limited ability to follow up the particular object. Other observers contributing or planning to work the collaboratively worked object should always communicate with the "PO/S"; it will be at his discretion what information he/she/it sends to other observers or to the whole group.

    A "coverage data collector" is a person who receives final data from observers (or from PO/S) for finished objects in the "asynchronous binary asteroids range" (set to P<5 h at this time). A function of the "CDC" is to prepare the data for their subsequent use in simulations of the survey. He may also communicate with the observers or the PO/S to clear uncertain points. At the beginning, the "CDC" will be my assistant Adrian Galad. Since the work of the "CDC" is not time-critical, he may not reply immediately; when being off or busy with other things, it may take even a week or two until he gets to prepare your received data for their subsequent use in the survey simulations.


    7.  Internal communication in the group

    A basic mean of communication between members of the group is email. I consider that it should be useful to create an "egroup" (e.g., on Yahoogroups, or elsewhere if there is a better service available --- feel free to suggest) so that communications to the whole group can be made to a single address. I also intend to set up a single address for the "AOL" so that your request for checking interesting data goes always to the right person. We may also put some information that don't have to be kept strictly within the group to web pages. (E.g., I think that the information from this email might be adapted to a web page.) Other services might go trough suitable web pages we (or you) might set up in future.


    8.  External communications on worked objects

    Communications to persons outside the group on results and data for a particular object is always at a discretion of the PO/S for the object.


    9.  Publications of data and results

    Once observations of a particular object are finished, the derived results should be "pre-published" by the PO/S on a suitable web page. It might be either your own web page (even one devoted to the individual object if it is particularly interesting, or one with summary data for a number of objects), or it might be or (for objects where Ondrejov is PO/S, or for results that I or another member of my group re-analysed). Just be sure that if you set up a new page (not just adding a new item to an already-existing page) you communicate the URL to the group. We should also have one page listing URLs to all the webpages with the "pre-published" results --- we will put it together soon.

    A regular (paper) publication of the results should be prepared by the PO/S later on. It should be understood that the paper publication (in MPB, P&SS, Icarus, A&A, and that like) may be done as soon as in several months or as late as several years after the observations are done --- it is sometimes not bad to wait with a paper publication for some time, until a proper way of all the things involved with it is established. The "prepublished" summary data, however, should be published on the web pages rather fast --- it is the way all potential users (outside this group) can hear on your basic results fast, may ask you for further analyses or for a joint publication of your data etc. In any case, it should be always at a discretion of the PO/S to decide where and when he/she wants to publish the results in a paper form, though some friendly agreement with collaborating observers will be always good.

    Discovery announcements on new binaries should be published in a special fast way. In each case where we are 100% (or nearly so) sure that it is a binary, the PO/S should publish the result on an IAUC. The line-charges can be paid from my grant (or from other grants we may obtain in future). Probable binaries (e.g., like Hovland) should be published fast as well, being it an IAUC or MPML (depending on a particular case). A summary of estimated parameters will be given on the web page or a similar one.

    Once in a couple years (depending on how fast the sample increases), I plan to do a statistical analysis of properties of the binary asteroids population and their trends with size, orbital, and other parameters. The first work of that kind (on NEA binaries) is to be finished soon; a short overview of some of the results has been presented at the 36th DPS meeting, see the presentation available on

    I encourage you to consider further ways how you/we would like to publish results obtained in the course of the Survey. (Note: As a general rule, the first paper publication on a particular object should have at least one observer from each participating station on the particular object included as co-author. Further publications using the data for the particular object should always refer to the first paper publication.).


    10.  Some concluding remarks

    10.1 Who are included in this group?

    The group consists of selected people who are currently very active in the field of asteroid photometry or directly involved with the data analyses, and who have a serious interest and ability to detect binary asteroids in a regular way (i.e., in an efficient way that makes the data suitable to model observational biases). A good group spirit is also needed, since we need to work collaboratively and in a friendly environment, so a lot of points have to be "assumed" rather than discussed out extensively again and again. (We need to keep the email communication under some control, so that to be able to concentrate on new and important points most of the time :-) I expect that the group will grow in following weeks, months, and years, it should be a natural way. Feel free to suggest at any time who else you think should be included.

    10.2 Are relative (differential) data usable?

    Yes, they are. Well, the lack of absolute calibration (or at least linkage with other data) means that it is even more important to obtain long series (full night if possible) so that a magnitude scale zero point can be estimated with low uncertainty.

    10.3 What is a minimal recommended length of one series?

    There is no general rule but uncalibrated series shorter than 2 hours are often not easily interpretable, so they should be avoided.

    10.4) How fast objects can be observed?

    If you can link individual series (separated by significant field shifts) to a common magnitude scale for the whole night, then a rate of apparent motion is not a critical factor. Otherwise, where the individual runs have to be used as independent series, I recommend to observe only objects which you can run with the same set of comparison stars for at least two hours (see the point 3 above).

    10.5 What is a minimum and what is an optimum number of nights for one object?

    A minimum number of nights are two.

    Actually, the rule is that each rotational phase has to be covered at least twice. It may occasionally happen that you cover an object for a full (long) night and find that you have covered a short period twice already. Then, if you were entirely sure that the period estimate is correct and if you saw a pretty single periodicity, it might be formally OK to call the object "finished" and to move to another target on a following night. On the other hand, a second night should be still very useful, basically because it gives a (statistically) better coverage of a (longer) orbital period of a potential satellite, therefore increasing a probability to detect it. So, two nights coverage is recommended as a minimum.

    An optimum number of nights is not so straightforward to be given. The rule to cover each rotational phase (at least) twice should be considered there as well. From a point of view of a potential satellite detection, a probability of its detection increases with increasing number of nights, but the rate of the increase lowers with the higher number of nights. In other words, the relative gain of taking a 6th night is significantly lower than that of a 3rd night. Taking four (long) nights appears to be about an average for optimum number of nights for one object. In any case, it has to be decided case by case. I would recommend to contact the "AOL" for opinion how to go on if you already covered five nights and not sure if taking more nights is justified.

    10.6 Is it worthwhile to revisit a finished object later on (in a following lunation, or an apparition)?

    From a point of view of the Survey for asynchronous binary asteroids, it is actually more efficient to observe new objects rather than to re-observe older ones (where just a normal, single-period behavior has been found). Well, in specific cases, it might still be useful to re-observe the old object, especially if there is a special interest in the given object for another project. In such case, the rule that each rotational phase has to be covered twice should also be held, though a (long) single night coverage of two rotation cycles might be (unlike in the case of a new object mentioned in point 5 above) OK for such follow-up observations.

    10.7 What sampling rate should be used?

    A typical mutual event is 2-h long, but the deepest part (where there is the best opportunity to detect the event well) is shorter, sometimes even shorter than 1 hr. Thus, in order to be sure that a detected attenuation is real and not just an observational error, a reasonable sampling rate is ~6 points/hr, i.e., one point per 10 minutes. So, if your instrument can switch easily between two (or multiple) targets, it might be actually possible to observe two (or more) objects at once. On the other hand, a denser coverage rarely harms, as then you can beat a noise with the larger number of data. So, it is basically a matter of your preference (and experience with your system's performance) if you target one object with a frequent sampling, or two objects with a lower sampling.


Last update: 2006 June 28

Back to the main page