Geologically, the only reasonable conclusion is that the lavas were erupted subaerially, or perhaps emplaced as sills close to the surface in some cases.
To first order, the magnitude of the exchange flow over the sill is determined by the horizontal momentum balance within the basin.
The limited horizontal extent of the forcing results in a buoyancy gradient across the sill and an associated circulation.
For practical reasons the experiments were performed 'upside-down' with the sill attached to the upper surface and buoyancy forcing at the lower surface.
We therefore speculate that a highly insulating foam of desorbed water vapour developed at the top of the sill.
Such layering is consistent with a 5-7 km thick underplated, basaltic zone of mafic sills that have intruded into the deep crust.
Near the sill we shall assume that mixing is negligible and that the sill exchange is composed of two distinct layers.
Include a continuous sill flashing with panned-up ends and inboard leg; the waterproofing layer should counterflash the inboard leg of the flashing system.