Abstract
Trans-Neptunian Objects are distant, faint and poorly known solar system objects. Stellar occultations are currently, the only way to precisely determine some physical characteristics of these objects, such as the shape/size, the multiplicity or an eventual atmosphere. The prediction of stellar occultations requires both accurate astrometry of stars and accurate ephemeris. The current methods of prediction use a constant offset compared to JPL ephemeris. The offset is calculated from the most recent observations as the mean difference between observations and the ephemeris. This method assumes that the offset remains constant over a certain period. In this study, we perform a new ephemeris with available observations and observations done for offset determination. In this context, we have developed a dynamical model of the motion of asteroids (NIMA), fitted to observations and determined a new ephemeris. The difference between NIMA and JPL ephemerides is close to the offset at the date of observations but this difference then varies over time, according to a periodic one-year term and a secular term. For some objects, the offset method may remain accurate when the time between observations and occultation is short or when the offset's variations remain small. For other TNOs, the difference sharply increases making inaccurate predictions in the future. Consequently, new ephemerides should be used to make predictions. Finally, occultations also provide accurate astrometric positions and therefore new constraints on the TNO's motion. We analyze the contribution of previous occultations in the improvement of TNO'S ephemerides.
