Hey folks. I recently had a discussion with an acquaintance where we disagreed about the functions of media luna, specifically about the collective and dispersive impacts on surface water flow. Looking for papers or other writings about them seems to all point back to this document (PDF warning) with data that they work but not how they work. What follows is my understanding of the process, and any information that improves my understanding is greatly appreciated.

With "horns up", sheet flow slows as it encounters the obstruction - during the early parts of a precipitation event or flow event, this will cause the water to enter the root and unsaturated zones of the soil profile. The water that does not percolate into the soil continues to move at 90° to contour around the rest of the obstruction's parts while also moving downward due to gravity. As the soil moves towards its capacity for stored water, the percentage flowing through the media luna increases. The pressure from the water on the upslope side of the media luna causes the flow to disperse through the obstruction, the convex shape of which causes additional lateral movement of the water as it makes its way downslope. My intuition is that the additional resistance towards the center of the structure is part of the cause, along with an evenly distributed lower pressure gradient along the convex face.

With "horns down", the upslope water slowing capacity is similar (I would expect the water retention properties to be strongest when installed with stream banks to either side) but due to the concave nature of the shape there is a greater amount of surface area directing the flow of water towards the geometric center of the shape through the same physical processes noted above - gravity and pressure differentials. I am not strong in maths and am unsure of how greatly the pressure differential affects the water's course, but I expect that value is nonzero.

I have a sneaking suspicion that my friend was considering the upslope behavior of a less-porous obstruction while I was considering the downslope behavior of a more-porous obstruction, causing them to think I had the collection/dispersion effects backwards. I'm well practiced at being wrong though, so it's a possibility that I'm open to even if I'm not enthusiastic about it.

Thoughts?