Thank you, Luis. The connection is an upstream boundary condition for a 2D flow area. It joins a storage area to the 2D flow area. I am not seeing the option to use 2D equations that would usually be an option for a structure internal to a 2D flow area. The weir does not overtop, either. For the concrete culvert, bell end upstream and protruding, I’ve got a entrance loss coefficient of 0.2 and an exit loss of 1.0. I am using 6.3 for this model. I’m not sure this got to the information you asked about.
Taking another look at the model, I realize I did not oversize the receiving cell at the pipe outlet. That could be causing some instability if the cell fills too quickly. I tried increasing the cell size, but not by a large amount. The fluctuation is still present.
Checking the Froude number now – it varies from 0.65 to about 2.0 at the outlet, so maybe the fluctuation is simply a matter of fluctuating tailwater conditions. Not sure.