Use of times series or fixed water level

I am the supporter and user of SWMM5. Recently I built a small model using SWMM5 and found some weird outputs. The model build is attached herewith for your reference.

The model is built for run with time series of inflows and downstream water levels. To clearly demonstrate the findings, 2-hour duration is selected for running only. The simulation option is Dynamic Wave with inertial terms kept.

After the model run, I have compared the total inflow at upstream end node with that at outfall node for each reach, see summary attached.

Since the inflow at upstream end during the 2-hour simulation period is constant therefore the inflow at downstream outfall is expected to be constant as well. But this is only true for reach DP3~DP3-out. For reaches DP5~DP5-out and DH6~DH6-out the inflow at downstream outfall is totally different from that at the upstream end, see attached captured Plots of Inflow Time Series.

Does anyone can expalin this problem?