I wanted to mention one thing about a recent paper which used the ICON general circulation model to simulate states of radiative convective equilibrium over ocean.  The authors ran the model on 5 different domains of varying total area.  The largest simulation was 256 times larger than the smallest.  That’s an impressive range!  The authors convincingly show that the domain size affects many aspects of the simulated atmosphere, but the one I want to mention is the physical morphology of the aggregated, high moisture regions (their figure 6).  The larger the domain size, the more structure there is in the moisture field.  I’ve done a lot of work recently on the links between total moisture and cloud behavior, and I can say that all that variation in the moisture field is critical for clouds.  We convective theorists have based a lot of recent thinking about the interaction of convection with it’s environment on the results of RCE simulations, but very few people have thought about the sensitivity of these ideas and theories to basic simulation properties (like domain size or shape) on the kinds of scales these authors do.  I agree with the authors that their results suggest a few new avenues of inquiry that could be rather revealing.

Source: Radiative convective equilibrium as a framework for studying the interaction between convection and its large-scale environment – Silvers – 2016 – Journal of Advances in Modeling Earth Systems – Wiley Online Library