oppley

I have a simple but large 2d only model with direct rainfall applied to create runoff. I have introduced embankments to protect certain areas from flooding using some zlns. But with the direct rainfall approach (which I cannot move away from for reasons beyond the scope of this query) I obviously get ponding behind the embankments in the areas I am trying to protect and sometimes this is deeper than the flooding from the river without the embankments. Because of the level of study I want to introduce something to allow flow to pass through the embankments as would be allowed by the downstream flood level (to all intents and purposes a flapped culvert). But given the nature of the model that would a very onerous process, so to get to the question... can unidirectional flow be set up within the 2d domain without having to introduce any 1d components?

If in some locations the width of a 1D channel is less than 2x the grid cell size, there are no cells to deactivate. The coupled cells on the left and right bank are adjacent cells. In this instance is it appropriate to include a THIN Z LINE to block flow from one flood bank to the other? Or are more agricultural methods required to artificially "widen" the channel to ensure there are cells to deactivate?

We have an ISIS TUFLOW model that is proving a little problematic and any pearls of wisdom would be gratefully received. The 1D network represents a small, narrow, steep channel with many closely spaced structures. We have rationalised the set up and simplfied the system so that the ISIS model works ok, but when coupled with the TUFLOW model it is taking longer than "normal" to stabilise the coupled model. Before we seriously consider alternative modelling options, does anyone have any tips for such models? Will modification of any of the ISIS or TUFLOW parameters aid stability for small steep channels embedded into a fine TUFLOW grid??

I am after some general advice about modelling breaches. We have successfully been using variable geometry to define breaches in fluvial and coastal defences. I have heard several times recently that TUFLOW shouldn't be used for breach modelling and I'd like to ascertain if this is a generalisation or whether there is some validity behind it. I can appreciate that if I want to model the minutiae of the breach propogation and the flow processess immediately downstream of the breach then a dedicated breach modelling package or 3D modelling is possibly more appropriate. But is saying that TUFLOW shouldnt be used for breach modelling a generalisation like saying 1D models shouldnt be used for representation of floodplains? or is it more likely to be from the mouths of those with a vested interest in selling different software to us?? Or should we be more careful when assuming confidence in the breach model results?