Astrometry of natural satellites: the analysis of past observations for the improvement of planetary system dynamics

The current study deals with the astrometric reduction of old photographic plates : we observe in the past. The main objective is to demonstrate that it is now possible to provide high-precise astrometric positions, by reducing a new time old photographic observations. We used photographic plates of the Galilean satellites taken at the McCormick Observatory and the U.S. Naval Observatory between 1967 and 1998, and resulting in 2650 observations. Of course we had to create a complete procedure of analysis. We first digitized the entire plates archive with taking care of the process itself, not to decrease the accuracy at this point. Then we created a procedure in order to provide astrometric positions. Softwares were developed to identify, extract and correct the positions, from the digitizations. A new astrometric reduction method by linking the stars was especially developed for the photographic plates' analysis to determine high-precise (RA,Dec) astrometric positions and intersatellite positions of the Galilean satellites. We also corrected the data for various instrumental and spherical effects, such as the coma-magnitude effect or the total atmospheric refraction, due to the small number of reference stars available. For the first time with these plates, we were able to provide ICRS (RA,Dec) astrometric positions, while only relative positions could be calculated. For the Jovian system, we now provide an accuracy about of 75 mas for (RA,Dec) positions, and about of 36 mas for intersatellite positions. Note that the previous accuracy was about of 90 mas and only for intersatellite positions. More, the observations provide satellite positions and Jupiter positions too. Thus we used these results to compare star catalogs, Galilean and planetary ephemerides. They were also used to fit the last Galilean ephemeris provided by the IMCCE, with the observations.