By warming the surface and stimulating the release of water by plants through photosynthesis, solar radiation drives the turbulent transport of heat and moisture from the land surface and consequently the formation of clouds. Clouds then partly absorb or reflect the incoming solar radiation, casting shadows on the surface and thus locally reducing the amount of incoming radiation. In this project, our goal is to better understand these interactions between the land surface, clouds and radiation. To reach this goal, we start by developing a fast radiation model using an innovative combination of machine learning and GPU computing.