With multiple sources of Mannings values being presented for 2D modeling, there is a problem with the location where a 2D model begins and and 1D model ends. There are a couple requirements in order to obtain stability at the interface of a 1D to 2D model:
1) Mannings values must match
2) 1D cross section must be as close as possible to the 2D mesh
the problem with this is that the interpretation of mannings values in 2D is greatly different from 1D…
2D Grass= 0.15-0.3
1D Grass=.03-0.35
That trend continues for every land cover type
So when a model has to transition from 2D to 1D or vis-versa, the large difference in mannings value causes a problem, this forces the user to pick either the 1D or 2D models for the entire model. If you chose to only match at the first few intersecting cross sections results will be odd.