Progress in astrometry and orbital modelling of planetary moons in the last decade enabled better determinations of their orbits. These studies need accurate positions spread over extended periods. We present the results of the 2014-2015 Brazilian campaign for 40 mutual events from 47 observed light curves by the Galilean satellites plus one eclipse of Amalthea by Ganymede. We also reanalysed and updated results for 25 mutual events observed in the 2009 campaign. All telescopes were equipped with narrow-band filters centred at 889 nm with a width of 15 nm to eliminate the scattered light from Jupiter. The albedos' ratio was determined using images before and after each event. We simulated images of moons, umbra, and penumbra in the sky plane, and integrated their fluxes to compute albedos, simulate light curves and fit them to the observed ones using a chi-square fitting procedure. For that, we used the complete version of the Oren-Nayer reflectance model. The relative satellite positions mean uncertainty was 11.2 mas ($\sim$35 km) and 10.1 mas ($\sim$31 km) for the 2014-2015 and 2009 campaigns respectively. The simulated and observed \textsc{ascii} light curve files are freely available in electronic form at the \textit{Natural Satellites DataBase} (NSDB). The 40/25 mutual events from our 2014-2015/2009 campaigns represent a significant contribution of 17%/15% in comparison with the PHEMU campaigns lead by the IMCCE. Besides that, our result for the eclipse of Amalthea is only the 4$^{th}$ such measurement ever published after the three ones observed by the 2014-2015 international PHEMU campaign. Our results are suitable for new orbital/ephemeris determinations for the Galilean moons and Amalthea.

Context. Bright stellar positions are now known with an uncertainty below 1 mas thanks to Gaia DR2. Between 2019-2020, the Galactic plane will be the background of Jupiter. The dense stellar background will lead to an increase in the number of occultations, while the Gaia DR2 catalogue will reduce the prediction uncertainties for the shadow path. Aims. We observed a stellar occultation by the Galilean moon Europa (J2) and propose a campaign for observing stellar occultations for all Galilean moons. Methods. During a predicted period of time, we measured the light flux of the occulted star and the object to determine the time when the flux dropped with respect to one or more reference stars, and the time that it rose again for each observational station. The chords obtained from these observations allowed us to determine apparent sizes, oblatness, and positions with kilometre accuracy. Results. We present results obtained from the first stellar occultation by the Galilean moon Europa observed on 2017 March 31. The apparent fitted ellipse presents an equivalent radius of 1561.2 $\pm$ 3.6 km and oblatenesses 0.0010 $\pm$ 0.0028. A very precise Europa position was determined with an uncertainty of 0.8 mas. We also present prospects for a campaign to observe the future events that will occur between 2019 and 2021 for all Galilean moons. Conclusions. Stellar occultation is a suitable technique for obtaining physical parameters and highly accurate positions of bright satellites close to their primary. A number of successful events can render the 3D shapes of the Galilean moons with high accuracy. We encourage the observational community (amateurs included) to observe the future predicted events.

The technique of mutual approximations accurately gives the central instant at the maximum apparent approximation of two moving natural satellites in the sky plane. This can be used in ephemeris fitting to infer the relative positions between satellites with high precision. Only the mutual phenomena -- occultations and eclipses -- may achieve better results. However, mutual phenomena only occur every six years in the case of Jupiter. Mutual approximations do not have this restriction and can be observed at any time along the year as long as the satellites are visible. In this work, we present 104 central instants determined from the observations of 66 mutual approximations between the Galilean moons carried out at different sites in Brazil and France during the period 2016--2018. For 28 events we have at least two independent observations. All telescopes were equipped with a narrow-band filter centred at 889 nm with a width of 15 nm to eliminate the scattered light from Jupiter. The telescope apertures ranged between 25--120 cm. For comparison, the precision of the positions obtained with classical CCD astrometry is about 100 mas, for mutual phenomena it can achieve 10 mas or less and the average internal precision obtained with mutual approximations was 11.3 mas. This new kind of simple, yet accurate observations can significantly improve the orbits and ephemeris of Galilean satellites and thus be very useful for the planning of future space missions aiming at the Jovian system.