The knowledge of the solar spectral irradiance in the UV and its variation in time is a key problem in aeronomy but also in climatology and in solar physics. While the Extreme UV (10-121 nm) range is important for thermosphere/ionosphere specification, the Far UV and Middle UV ranges are essential for climate modelling. However, the continuous monitoring of the UV irradiance is a difficult task. Space instruments are indeed suffering from ageing but also signal contamination of many kinds. Because of the lack of long-term measurements of the whole UV range, most thermosphere/ionosphere and climate models rely today on proxies for the solar irradiance, which may however not reflect very well the variability. As an alternative, we proposed in this work to use a few radiometers with properly chosen passbands in order to reconstruct the solar UV irradiance. Using a multivariate statistical approach, we first characterize the high redundancy as well as the different features of the solar UV irradiance. With four passbands from already existing instrument, we test our concept: the solar UV flux is reconstructed with a relative error of about 20%. This work proposes then to define a new kind of instrument, which may use wide bandgap materials as detectors selecting moreover the spectral range. Filters are indeed very sensitive to the degradation. This thesis finally proposes to model the impact of the solar UV flux on the atmosphere of Ganymede. We predict some atmospheric emissions in the framework of future space mission to Jupiter.
