I have the same question on my project. Using HEC-RAS v5.07 my losses across the lateral weir are far in excess in what the upstream cross section says is in the channel. Or, if it’s not, the amount deducted from the channel flow in the next downstream cross section is not the same as what is shown as loss to the weir.
Interestingly, I downloaded HEC-RAS v4.1 and opened the same file and ran the steady flow analysis with all the same calculation parameters (flow optimization turned on, etc.) and the calculations work out and are reasonable. I’m wondering if the lateral weir calculation engine is working correctly in the HEC-RAS versions since the addition of 2D capabilities. Perhaps you HAVE to model a 2D storage area and can’t just assume loss from the system or return at a downstream cross section?