You can do ab initio calculations to see how a chunk of moist atmosphere behaves including all the relevant effects. It will within a fairly short period evolve away from its initial conditions towards a dynamic equilibrium compatible with the solar forcing and energy balance.
It is just impossible to do that for the entire planet at once. A chunk of atmosphere is now realistically within the remit of simulation.
This sort of modelling is common and uncontroversial in astrophysics where it is used to model things like relativistic jets and accretion disks. We don't know the initial conditions there either but we can still model them and then compare the simulations with reality.