GBOT Asteroids
Presentation of the
GBOT asteroid finding programme

Asteroid 0012820 "Robin Williams"
Discovered by Tautenburg Observatory in 1978
and re-observed by GBOT with the VST on the 13th February 2015

The GBOT asteroid task force

  • GBOT office (Heidelberg): Martin Altmann
  • GBOT data centre (Paris) Sebastien Bouquillon, Christophe Barache, Francois Taris, Teddy Carlucci, William Thuillot, Rene A. Mendez (Universidad de Chile), Damya souami
  • Others: Paolo Tanga (GBOT asteroid expert, GaiaFUN-SSO, Nice), Jon Marchant (detection method. Liverpool telescope, LJMU), Tim Lister (LCOGT)
  • External connections: GaiaFUN-SSO (W. Thuillot, IMCCE & B. Carry OCA), asteroid finding group of the Starkenburg Sternwarte, Heppenheim (Matthias Busch), Skybot (Jerome Berthier, IMCCE).

The GBOT telescopes

GBOT currently uses 4 telescopes of the 2-3 m class, located in Chile, Australia, the Canary Islands and Hawaii (USA):


Since the task of GBOT is to astrometrically track the Gaia satellite, as it oscillates around the Sun-Earth L2 point, the project obtains significant data of fields located on or near the Ecliptic. The fields containing Gaia are roughly located opposite to the Sun, which means that all Solar System Objects (SSOs), such as asteroids, are near their opposition, and are thus near their smallest distances to Earth. Therefore such fields are suited to detect asteroids with a much fainter absolute magnitude (the term absolute magnitude is defined for SSO as the brightness an object has at a distance of 1 A.U.) than in other fields. Also asteroids near opposition are fully illuminated when seen from Earth, away from opposition the hemisphere (assuming a spherical object, for non-spherical objects the principle is the same) facing earth can have quite a fraction which is not illuminated by the Sun. This effect is not as pronounced as for the Moon or the inner planets but can amount to 16% missing in Mars, 1% in Jupiter, and somewhere inbetween for the main belt asteroids, which are located between Mars and Jupiter (see Fig. 1).

Fig. 1: Phases of outer solar system bodies as seen from Earth, demonstrating that objects with smaller radius orbits show a greater phase variation than those with larger orbits. The plot shows Sun (yellow), Earth (red) and two outer bodies (Blue and Green) on different size orbits, each of them seen at three different phase angles in respect to the Earth. The black thick straight lines show the geometric normals to the Sun, i.e. the terminator of Solar illumination, the red lines the geometric normals in respect to the line Earth-object, i.e. the part of the objects which can be seen from Earth. For an object in opposition, the two coincide (therefore the red lines are not given), the phase is always full, while the angle between the terminator and the geometrically visible horizon, α is always larger (except for opposition and conjunction when it is exactly equal) for the blue object on the smaller orbit than that of the more distant green object. This means that the illuminated fraction (which in terms of angle defines as 180 degrees - α varies more for the green object than for the blue one.

GBOT serendipitously records such objects on its data, especially the data obtained with the 1 square degrees field of ESO's VST is thought to contain several such objects on every frame. Therefore the GBOT group began to exploit this potential, also because in principle the GBOT operating system, i.e. database and pipeline already offered all the infrastructure required. The adaptations that were necessary (e.g. expand the astrometric reduction of the VST data to all 32 detectors rather than stick to the one detector recording Gaia) were completed in record time, and from End of February the whole asteroid detection system was in essence fully operational. At current all data obtained from January 2015 on (The decision to embark on an asteroid finding programme was decided on during the 9th GBOT meeting in Turin, October 2014) is analysed this way, now the older 2014 data gets the same treatment. However most important is the timely analysis and reporting on asteroids on recent data. Here it is important to report new detections as soon as possible after the data is taken, so that follow up observations, which usually can not be conducted by GBOT itself can ensue. GBOT's prime task is to track Gaia, and since Gaia moves on average 1 degrees per day (360 degrees in one year) there is only a small overlap of the fields covered by GBOT's VST observations of consequtive nights. Therefore we have to rely on third party observations for comfirmation purposes. By carefully tweaking all processes involved in the download/reduction and analysis process, we can now report new observations usually within 12 hours. Reporting is done on this page as well as to the MPC and also the GaiaFUN-SSO network.

The GBOT asteroid finding programme is not (just) a sportive exercise, but it has also the potential for significant science implications. This is fundamentally due to two reasons:
  1. The first, is the significant amount of asteroids whose orbit is poorly known, with uncertainties on the predicted positions of several arcseconds or more.
  2. The second - which might seem surprising - is that the precise brightness of asteroids is very poorly known. Usually, to derive the asteroid albedo (a fundamental surface property linked to composition) an accurate measurement of its size is needed, along with its absolute magnitude "H". H (not to be confused with the absolute magnitude of stars, which has the same purpose, i.e. a measure of a normalised brightness of an extrasolar object but is - for reasons of convenience - defined differently) is defined as the brightness, normalized at 1 AU distance from both Sun and Earth, that the object would have if it was seen exactly at opposition. While size can be estimated by thermal IR measurements (as WISE did for 100,000 objects), H requires a more complex procedure. In fact, an asteroid is rarely observed exactly when the angle between the observer and the Sun, as seen from the asteroid ("phase angle") is 0°. As a consequence, its value H must be estimated by observing it at several phase angles, and then by fitting a model to the observed magnitude-phase variation. The closer the observations are at opposition, the better the resulting value for H. While this an acceptable complication, in reality the largest majority of the asteroid photometry available in the archives is simply of quality so poor that the final H can be off by several tenths of a magnitude! This is basically due to the fact that asteroid astrometry has always been the priority, not accurate absolute photometry (requiring additional, painful calibrations!). Even neglecting other complications such as photometry variations due to the asteroid shape, such errors in the available data bases are so large that today we can say that only a very small number of asteroids (about 30!) has a precise H known.
How can Gaia and GBOT help?
Of course, Gaia performs a very accurate photometry of all the sources on the sky. However, it won't be able to catch Main Belt asteroids at phase angles smaller than 10-20°, as it never observes at opposition. On the other hand, GBOT does exactly that. By targeting Gaia, all the asteroids serendipitously observed are around opposition. Its accurate photometry is thus a must for complementing the one by Gaia, at small phase angles! Also, the information on the collected data is being stored in such a complete way that once Gaia has delivered its photometry, a full, more accurate calibration of all magnitudes will be possible. Recent studies also show that the phase-magnitude curve shape (in particular, its slope) is also related to the spectroscopic type (i.e. to composition) of the asteroids. However, to be really exploitable for physical characterization, this relation must be calibrated on a large and accurate sample of photometric measurements. Gaia and GBOT, coupled together, can thus really open a new perspective in this direction, too!
There is also another reason for pursuing the GBOT asteroid finding programme. Since we are observing asteroids of a rather wide brightness range, the asteroid programme is is and will be (especially when we start using Gaia astrometry for our reductions) really useful for testing and improving the astrometric precision of the pipeline for moving objects thus optimising the precision of the determination of Gaia's position.

The detection process

When newly obtained data has been downloaded from the observatory archive and incorporated into the GBOT database, the GBOT reduction pipeline is invoked by the GBOT operation centre member on duty, and the astrometric reduction, primarily to measure the position of Gaia (our prime objective) but now also to find and identify asteroid candidates. After the general astrometric reduction, which involves, source extraction (and x,y,mag coordinate determination), catalogue retrieval, and astrometric solution finding for each image/subimage, all orphaned sources, i.e. those sources, which do not have an entry in the reference star catalogue, are subjected to a 3 parameter (x,y,time) Hough transform (see e.g., which leads to an automatic identification of asteroid candidates, which are characterised by being in chains on subsequent images. Since one observing sequence on the VST consists of 10 exposures, we ideally have chains of 10 points, which allows for a quite robust automatic detection. Even if due to the faintness of the object, the object running over a star, the object being close to the edge of a detector or passing over a known bad region, we do not have all 10 detections, the result is still quite reliable. Nevertheless, prior to publication, the candidates get inspected manually, and those of dubious nature rejected. Additionally we check whether an object has already been identified and submitted to the MPC. This cross-identification is not a trivial task, and it is being done, using the Skybot search system and comparisons of positions, speed and brightness. Overall problematic fields are fields which contain very bright stars, and thus a host of optical artifacts, which get detected by detection algorithms, and crowded fields, such as those in the central Milky Way region, through which Gaia ventures in June every year.

After the detection and verification process has been completed, the results are submitted to the MPC, the GaiaFUN-SSO network etc., where they are ready for follow up observations. These are required to verify an object, and to improve it's orbit. While the astrometry GBOT derives is certainly of high quality, it generally only covers a 20 min period in one night, meaning that the leverage to fix the orbit is rather small, based on this alone

First Results and statistics

Current number of asteroids detected by GBOT:

Known : 38300
New : 21329
All : 59629

On the 1 square degrees VST data usually between 10 and 80 asteroids are found for each night, about half of them previously discovered objects, and half newly found ones. Both are valuable, since observations of previously discovered objects may present second or third epoch observations required to confirm recently found minor bodies. The vast majority are Main Belt Objects (MBO's), one discovery was shown to be a Mars Grazing object (from a Martian's viewpoint, this would be a NMO ("Near Martian Object"). Right now no NEO (Near Earth Object) has been found, but this is probably just a matter of time. One notable appearance was the asteroid 12820 Robinwilliams, named after the recently deceased great actor Robin Williams.

Please note that the term "new" means that the GBOT team could not associate an object with objects already in the MPC database. However this does not necessarily mean (while probably being true for the vast majority of candidates pre classified as "new", especially given the brightness range of >21 for most objects) that the object has not yet been discovered before. Bad astrometry of the original discovery epoch can easily lead to such uncertainties in the orbit that the coordinates of observations of one and the same object so not seem to be connected. Additionally an object could have been first observed years ago, and thus may not be easily linked to new observations like those of the GBOT team. Therefore a certain (unknown but probably not very large) fraction of the "new" objects are most likely re-observations of objects discovered prior to our observations and susequently lost.

Fig. 2: Number of asteroids found on every night of 2015. The blue parts of each bar show asteroids which are currently not previously in the MPC database, while the green parts those objects, which could be identified as previously observed.

Fig. 3: Histogram showing the nightly bounty for 2015. Depicted is the number of nights with a given number of objects found. Again, the blue shows "new objects" (see Fig. 2 and text), green the previously known objects and the red histogram the total.

Fig. 4: Magnitude range of the GBOT asteroid programme's discoveries in 2015. Again, the red histogram shows the total, the blue "new objects" and green the previously known objects.

Fig. 5: Phase angle of known asteroids when observed by GBOT in 2015 (colors are to distinguish the different groups of asteroids). The average phase angle is about 4 degrees, i.e. very close to opposition.