modelling a river

Hi,

What is the best way to define a b.c. for a river (standing water body)? I have leak point at some depth below the bottom of a river within a groundwater.  The overlaying river is 2 m deep and then there are 3 m of sediment and finally, 10 m below is the groundwater table. So, first, a constant head boundary conditions for the model was defined on both ends to establish a hydraulic head and then ran to reach SteadyState. Next, I enabled the middle blocks and set a thin atmosphere at the top of the model and also initialize the model with the previous left and right equilibrium state and fixed them.  The model was run to reach steady state. So far, there were just the atmosphere layer, the sediment and the groundwater defined.

Next, to include the river, I converted some of the blocks into atmospheric material except with 100% SW and defined all blocks representing the river as the fixed state. I still have the atmospheric layer on the top of the river and also fixed left B.C. from the first run. After a few time steps, the simulation fails. Why is it not converging? since all the initial runs were smoothly converged, it made me think it has something to do with the river b.c. So, what is the most appropriate way to incorporate the river into the model? Specifically, should I define all the blocks representing the river as fixed state or only the bottom and upper rows, etc? (although I tried both and still failed). The data file is attached for your reference.

 Thanks,

 Alireza