Another problem observed on some networks is the dry reach problem. In steady flow calculation there is no particular problem. On the other hand, it is impossible to do a simulation in unsteady flow on such a network. Two solutions are conceivable. The first one consists of adding a fictional discharge to a dry zone head and to subtract this discharge at the end of that zone, in order not to disturb the flow somewhere else. A second solution is to extract a subsystem from the complete system that only contains flowing reaches. This second solution is not satisfactory anymore if the dry level(s) begin(s) to flow during the simulation. This "dry bed" problem is hard to solve correctly. Cemagref is working on this problem and a next SIC version should integrate this possibility.