The aim of the internship and the PhD thesis is to investigate the effect of entrainment, preconditioning and convective self-aggregation on cloud patterns, in connection to global warming issues, using a combination of field analysis, controlled laboratory experiments, theory and numerical simulations. Objectives — To address the organization of convection and the formation of cloud patterns, the proposal starts with two central questions. How does convection self-organize when an unstable turbulent layer is capped by a stably stratified layer? What controls the survival time against evaporation and the buoyant ascent dynamics of active clouds? The second problem requires a detailed understanding of the heterogeneities of turbulent mixing around the fractal edges of clouds in the ascending phase, i.e., the entrainment of the external fluid and thus humidity, cold, and nuclei. The environment of the clouds plays a key role in their ascending dynamics generated by the latent heat released during the nucleation of droplets. Its structure arises from two concurrent processes: long-term pre-conditioning, linked to radiative effects and overturning circulation, and the modification of thermofluidic fields by active clouds that stop their ascent and evaporate, releasing their thermal energy, nuclei, and humidity. We thus hypothesize that this local memory of previous clouds plays a key role in understanding cloud patterns as an effective spatio-temporal interaction.