r.sim.water: Depths

[xycarto]

I am curious about the behaviour of depth outputs for r.sim.water in GRASS GIS. When I run a simulation for an extended period of time, my model outputs what seem to be unrealistic depths for certain locations. .

Inputs:

• 5m DEM with hydro-correction to fill sinks
• 5m Mannings raster with rivers having an "N" value of 0.05. Mannings average is 0.15

Perimeters run:

r.sim.water elevation=$BASE dx=dx-$BASE dy=dy-$BASE depth=depth-rast disch=disch-rast niterations=720 rain_value=100 nproc=16 nwalkers=10000 man=man_coef

Running the above produces occurrences like seen below:

In this case, the output depth raster is suggesting the dark blue location has a water depth of almost 25m. The channel at that location is only 4m deep.

Does anyone know what causes behavior like this in the model?

[xycarto]

Closing in favour of a different discussion