The article talks about 1.2km horizontal resolution. Is it a 3d grid? If so what is the vertical resolution? Or is the vertical dimension integrated within 2d boxes?
Yes, these are fully 3D simulations. The vertical resolution is slightly harder to describe because models like these typically use more complicated vertical coordinates than just "meters above ground". E.g. a technical article on the ICON-A formulation [1] says that it uses a "hybrid sigma" coordinate - so something that follows a combination of pressure and terrain. For a very high-resolution simulation that resolves clouds, you'd need to crank this up to O(100 m) in the bottom of the model, but it can space out as you get higher in the atmosphere.
