© The Authors 2024

1 Introduction

Commission 20 of the International Astronomical Union (IAU) recommended in 1991 that its Working Group on Natural Satellites should create an international center of astrometric data on the natural satellites. Under the auspices of the IAU, this center would work to make these data widely available to researchers for the purpose of building or testing ephemerides. Thus, the Natural Satellites Database (NSDB) was created as a result of collaboration between the Sternberg Astronomical Institute of the Moscow State University, in Moscow, Russia, and the Institut de mécanique céleste et de calcul des éphémérides – UMR 8028 du CNRS, Paris, France. The fact that the two countries were actively observing the natural satellites of the planets was the main reason. Many observational data have been accumulated in the NSDB, that were sometimes published in journals that were inaccessible for researchers.

A first description of the NSDB was published by Arlot & Emelyanov (2009). Since then, the database has been significantly expanded and updated with new data. Now it consists of three parts: (1) Astrometric observations of the natural satellites of the planets from Mars to Neptune, as well as observations of the satellites of the dwarf planet Pluto, (2) astrometric observations of the satellites of dwarf planets and asteroids and (3) photometric observations of the satellites of Jupiter, Saturn and Uranus during their mutual occultations and eclipses used for astrometric purposes. The latter type of observations has a specific format that is described in the corresponding “Content” files.

The quantitative and qualitative development of the NSDB necessitates a new description which is given in this paper.

2 Astrometric observations of natural planetary satellites

A data unit in the NSDB corresponds to homogeneous observational results published in one paper or in one data source. A data unit is represented as a pair of plain text files, “Content” and “Data.” The Content describes the observations and the Data file lists the observations themselves. This pair of files is suitable and sufficient for use in any software that determines satellite orbits. The basic principle is that the data are added to the database exactly as they were published and the description contained in the Content file is extracted from the paper.

We accumulated observations from the following sources:

1. The published papers.

2. The data files provided as supplementary information to online versions of journal papers.

3. Direct contact with observers.

4. Observations available in other databases, such as those provided by the Pulkovo Observatory¹, the Minor Planet Center², and the Flagstaff observatory³.

Some of the sources are difficult to access. As an example, we can mention the publication in the Proceedings of the observatory in Kazan, Russia, which contains tables with observations of the main Saturnian satellites that were made in Kazan observatory in 1982 (Kitkin, 1986), as well as the papers deposited in the All-Russian Institute of Scientific and Technical Information. Observations from these sources had to be added manually to the database since they were not available in digital format.

As mentioned above, the observations are grouped into the so-called “data units.” The work spent on adding observations to the database was proportional to the number of data units, not to that of observations. Some units contain only one satellite observation and they were added to the database when it was appropriate to use them.

At the present time, 225 satellites of planets from Mars to Pluto are known. All satellites may be classified into groups. Table 1 gives the number of data units and the number of observations for each group.

As new observations are published, we add them to our database⁴. The status of the number of individual observations distributed over different satellite groups is given in Fig. 1. The data in Table 1 and Fig. 1 are given for the state of the NSDB as of March 1, 2023.

The observations from our database are used to determine the satellite orbits. For this purpose, the ephemeris server MULTISAT for planetary satellites was created by Emel’yanov & Arlot (2008)⁵. Since 2008, the MULTI-SAT server has been significantly developed. In particular, the ephemerides of all planets, of the Sun and of the Moon have been added. The server is especially designed for observers, and thus it allows one to determine the residuals between any number of observations provided by the user and theoretical positions.

The NSDB database provides all the information necessary to build dynamical models of the motion of natural planetary satellites. Note that most of the observations made using space telescopes are now included in the database. For groundbased observatories, the IAU code for the observation location is given. The same is true for the space observatories which have their own IAU codes: the Wide-Field Infrared Survey Explorer (WISE, C51), Hipparcos (248), the Hubble Space Telescope (HST, 250), and the Gaia space observatory (258). If the observing site is WISE or Gaia, the Data file also gives the geocentric or barycentric coordinates of the spacecraft at the time of observation. In addition to observations made from space observatories that have their own IAU codes, observations from the Mariner, Viking, Phobos, Mars Express, and Cassini probes are also available. The coordinates of the probes are given when they are provided by the source of observations. Even when the coordinates are not given, they can be retrieved from the corresponding SPICE kernels.

The observations from Gaia are specific in that their accuracy is much better than that of ground-based observations, but the accuracy depends on the scanning directions on the sky. The accuracy in the scanning direction is better, and Gaia observations therefore need to be listed with the position angle of the scanning direction. The experience using Gaia observations of faint planetary satellites with the aim to determine their orbits is described in Emelyanov et al. (2023). The accuracy of observations made at the Gaia space observatory is discussed further, in Section 5.

Fig. 1 Number of observations (N) in NSDB in different years for some groups of satellites.

3 Astrometric observations of asteroid satellites

The database of astrometric observations of asteroid satellites is a new feature of the NSDB. Asteroid satellites are particular in that their apparent angular separations from their primaries are smaller than 1 arcsec. Thus, specific methods of observations are used such as adaptive optics and speckle interferometry. Classical astrometric observations are possible only for very few asteroid satellites.

Because it is difficult to observe these objects, not many publications of their observations have been made. The authors of these publications used to determine the orbits of satellites based on their own observations alone.

The section of the NSDB that is devoted to asteroidal satellites contains 170 data units for 104 satellites. In total, we have 2340 astrometric positions of asteroid satellites.

Currently, 517 asteroid satellites are known. However, all we know about most of them is that they exist, since there are not enough astrometric positions to determine their orbits as well as the masses of both the satellites and their primaries.

Fig. 2 provides statistics on the number of observations for the asteroid satellites. Each point in the figure corresponds to one individual satellite. The vertical axis gives the number (N) of satellite observations. The satellites are positioned so that the number of observations decreases from left to right. The satellite Linux of asteroid (22) Kalliope has the maximum number of observations, 260. We have only one observation each for 28 satellites. For 48 satellites 10 or more observations are available.

Emel’yanov & Drozdov (2020) attempted to determine the orbits of all asteroid satellites based on all available observations. This attempt is in constant progress as new observations appear. The results of this work led to creation of the AsterSat ephemeris server Emel’yanov & Drozdov (2021), which is a unique tool that provides the user with many opportunities for working with ephemerides of asteroid satellites. Orbits can currently be determined for 66 asteroid satellites.

The section of the NSDB dealing with asteroid satellites is provided with two additional web-pages. The first web-page (“Summary”) lists the data units for each asteroid system. The second web-page (“All binary asteroids”) provides some orbital and photometric parameters of each system. These pages can be accessed via the corresponding links given at the top of the main page⁶.

Fig. 2 Number of observations (N) for different satellites of asteroids.

4 Photometric observations of the satellites during their mutual occultations and eclipses

As presented in our previous paper (Arlot & Emelyanov 2009), the NSDB contains some photometric observations from which astrometric information may be extracted: the occultations of Solar System objects, and especially natural planetary satellites. They are eclipses and occultations of the Galilean satellites by Jupiter (observed since Galileo in 1610, but NSDB contains observations that only begin in 1652, the earliest observations cannot be used). The same events as they occur in other planetary systems with other satellites are also proposed although they are rarer. Other occultations are mutual occultations and eclipses between the Galilean satellites as well as between other natural planetary satellites. These observations were performed only recently because a computer is required for accurate predictions of these events. The final occultations are occultation of stars by planetary satellites.

This observation consists in observing the intensity of the light flux received from the satellites before, during and after the event. The minimum intensity is thought to correspond to the minimum distance, and the value of the minimum is converted into the minimum distance in arcsec. Since the measurement is mainly a timing, the accuracy of the observation is much better than that of a direct classical astrometric measurement of an angle as shown in previous analyses of these observations (Arlot et al. 2014; Arlot & Emelyanov 2019). When we enter these observations in the NSDB, we first enter the time of the minimum light flux, as well as the light flux for the moment of time as observed data. However, we understood that the apparent disks of the satellites are not uniform surfaces (because of the albedo, phase angle, etc.) which results in an error in the deduced positions of the gravity centers of the objects, the positions of which were used for dynamical purposes. Then, we used to add the complete observed light curve in the NSDB, which provides photometric information on the albedo of the satellite surfaces. This allows the user to improve the astrometric accuracy of the observation.

Sometimes the observers reduce the lightcurves and publish astrometric positions that are deduced from photometric observations. In these cases, observations of both types are added to the NSDB.

Observations from space probes allow us today to build very accurate ephemerides, so that the albedo of the satellite and not the astrometric positions might be considered as unknown values. Photometric light curves are carefully added to the NSDB. A better knowledge of the albedos should help us to improve the determination of the center of mass relative to the photometric center which is very valuable for dynamical studies. The albedos that can be used for this purpose are not provided by space probe observations since the probes do not observe the satellite at the same phase angle as ground-based observatories.

5 Accuracy of the observations in the NSDB

The coordinates of the outer satellites of Jupiter, Saturn, Uranus and Neptune are measured relative to the stars. Therefore, the absolute coordinates, that is right ascension and declination, are given in the NSDB for these satellites.

When old observations are used, the following features must be taken into account. The accuracy of old observations was limited by the accuracy of the star catalogs such as FK4 and FK5. Old observations had systematic errors. In particular, for observations made prior to 1940, an additional correction of 0.75 arcsec must be applied to all right ascensions. Emelyanov (2005) confirmed that this correction improves the fit to nearly all early observations. The author also showed that after taking systematic errors into account, the root mean squares (rms) for the Jovian satellites J6–J13 are within the limits of 0.55 and 0.92 arcsec. Table 2 is taken from Emelyanov (2005) and reproduces the residual statistics (rms) for the satellites J6–J13 for different observation periods. Some crude observations have been discarded here.

It follows from Emelyanov et al. (2022) that the accuracy of ground-based observations of the newly discovered outer satellites of Jupiter and Saturn is in the range between 0.3 and 0.6 arcsec.

Our comparison of ground-based observations with the most accurate models of motion of the outer satellites of Uranus and Neptune shows that the observational accuracy is in the range of 0.2–0.6 arcsec.

Old data are mainly photometric data because the deduced accuracy is better than that of classical astrometric measurements that were made at the same time. As we described in the previous section, providing the original observed light flux will help us to improve the accuracy through a new reduction using new tools or new complementary data. Similarly, it seems to us that we should provide the original support of the astrometric data in order to allow the user to improve the astrometric accuracy through a new reduction. For astrometric data, the original support is either a photographic plate or a CCD or CMos image. We focus on photographic plates because they are older and warrant a new reduction. Several projects of digitizing photographic plates Robert et al. (2021) have made digitized plates available through FITS files. Each user of the NSDB would be able to choose which observations warrant a re-reduction using new methods and the Gaia reference star catalog provides the current astrometric accuracy to these old observations.

The coordinates of the inner satellites are mainly measured relative to the major satellites. Therefore, for these satellites, the NSDB gives relative coordinates. The observational accuracy of these satellites is limited by the interference from the planet brightness. A comparison of the observations of these satellites with the most accurate motion models shows that the observational accuracy is between 0.04 and 0.26 arcsec. This applies both to ground-based observations and to those made using the Hubble Space Telescope.

Our comparisons of the ground-based observations of the satellites of Mars, the Galilean satellites of Jupiter, the main satellites of Saturn, Uranus, and Triton (a satellite of Neptune) with the most accurate models of motion show that the observational accuracy is in the range 0.05–0.6 arcsec.

The launch of the Gaia space observatory started a new era in astrometry in which the accuracy of star coordinates increased by thousands of times. A significant improvement in accuracy was also expected for the coordinates of the Solar System bodies. Emelyanov et al. (2023) determined the orbits of six distant satellites based on observations from Gaia. Some specific moments in using observations of the satellites made by Gaia were demonstrated. These peculiarities stem from the scanning motion of Gaia, in particular from the fact that the observational accuracy is significantly different along and across the scanning direction. As expected, Gaia observations proved to be more precise than those made from Earth, which results in more accurate satellite ephemerides. The accuracy of the satellite coordinates measured in the along scan-directions is 0.003–0.004 arcsec.

6 Future and projects for NSDB data base

Most of the tasks proposed in our previous paper were fulfilled. The main goal now is to propose access to the archive observation records in the form of numerical files of digitized astrometric photographic plates corresponding to each NSDB data unit. For this purpose, a link will be provided with PADC database⁷ archiving the images of photographic plates. Most of the photographic plates corresponding to the astrometric observations of NSDB are currently still available and the work of digitizing plates has started at the NAROO center Robert et al. (2021).

7 Conclusion

The NSDB data base continues to propose data to users. Observational material in the form of FITS images will be provided soon and is already available on request. This new challenge will allow the NSDB to continue to be a unique tool for dynamical studies.

Data availability

NSDB is accessible online at http://www.imcce.fr/nsdc (IMCCE) and http://lnfm1.sai.msu.ru/neb/nss/html/obspos/index.html (trilingual version of SAI).

Acknowledgements

This work was supported by IMCCE and Paris Observatory and by the SAI of the Lomonosov University in Moscow.

References

All Tables

All Figures

	Fig. 1 Number of observations (N) in NSDB in different years for some groups of satellites.
In the text

	Fig. 2 Number of observations (N) for different satellites of asteroids.
In the text