Hi,
This abnormal water depth detection is based on a ratio between the computed
water depth and the cross sections' depth. If you have reduced the cross
section's height, then this ratio becomes larger and then it's even more likely
to get this error.
In some special cases with shallow cross sections and under drowned conditions,
it can be expected that the computed water depth will be significantly higher
than the cross section's height. In that case the solution is to allow a larger
ratio than 1. This is controlled in the 'MIKE 1D engine configuration' window.
Note that your error appears in the first time steps of the simulation, so it
can also be due to an instability, being the consequence of a significant
discrepancy between initial conditions and boundary conditions (e.g. when
applying a large discharge as boundary condition at the first time step, in an
initially dry river).