Future devices like JT-60SA, ITER and DEMO require quantitative predictions of pedestal density and temperature levels, as well as inter-ELM and ELM divertor heat fluxes, in order to improve global confinement capabilities while preventing divertor erosion/melting in the planning of future experiments. Such predictions can be obtained from dedicated pedestal models like EPED, and from non-linear MHD codes like JOREK, for which systematic validation against current experiments is necessary. In this paper, we show progress in the quantitative validation of the JOREK code using JET simulations. Results analyse the impact of diamagnetic terms on the dynamics and size of the ELMs, and evidence is provided that the onset of type-I ELMs is not governed by linear MHD stability alone, but that a nonlinear threshold could be responsible for large MHD events at the plasma edge.