An email to Eric:

Eric,

something we will have to discuss is the role of the bottom drag in the bottom layer. Because I did not know what else to do, I decided to see if the results were sensitive to the level of linear bottom drag in the bottom layer. For me, because there is no stress between layers, the flow in middle layer should have a free-slip condition at its two interfaces. But it is not clear to me what is the effect of the bottom drag on the wave field: Is the level of drag equivalent to a partial reflection of the wave onto the bottom? Within zero drag, we have perfect reflection onto the bottom, with infinite drag, we have perfect reflection onto the /interface/ between the middle and bottom layer (and the model is basically a 2-layer model) and with values of drag in between, a partial reflection?

In any case, I run two experiments, one with half the drag coefficient (

exp2_w6_diffbotfric), the other with 10 times the drag coefficient (exp2_w6_diffbotfric2). In the last case, the flow in the middle and upper layers are negligible compared to the two other cases. So, the bottom drag has definitively an important effect. What is that effect and what should the drag coefficient be, are questions we need to answer. In particular, can the drag coefficient be used to stabilize the recent experiments with weak horizontal mixing?Francois