MIKE-SHE and MIKE-HydroRiver model run error “Water level exceeded maximum…more the 4.0000 times cross section height”

Question

I am trying to run a coupled MIKE-SHE - MIKE-HydroRiver (MIKE 1D) model (2020 version) on a daily timestep. The model gets through the preprocessing stage without a problem but then shows an error message in the first timestep and stops running. The error is to do with the river water level in the MIKE-Hydro model being extremely high, however there was no rain in the first timestep and the initial water levels input in MIKE-Hydro were not deep. The river cross sections in the MIKE-Hydro model were derived automatically from the DEM that was used in the MIKE-SHE model. The MIKE-SHE model has an unsaturated zone (UZ) and saturated zone (SZ), the latter set-up to use the ‘Linear Reservoir’ calculation option. After playing around with increasing the MIKE-Hydro calculation timestep from 10 mins to 5 seconds, and changing the UZ and SZ layers back to global settings, I managed to get the model to run for about 51% after which it crashed, showing a similar error message about river water levels. The message is as follows:

‘ERROR: Water level exceeded maximum on 1 grid point(s). Largest exceedance on reach 'Berg', chainage: 15500.00 m. Water depth is 0.4886 m which is more than 4.0000 times the cross section height. Water level is 245.44 m. (HD_ERR_WLAboveMaxOnXS) Simulation time = 2013-08-27 11:07:30 (time step 6764489). Exception: Water level exceeded maximum on 1 grid point(s). Largest exceedance on reach 'Berg', chainage: 15500.00 m. Water depth is 0.4886 m which is more than 4.0000 times the cross section height. Water level is 245.44 m.’

Answer 1

The error comes from the water depth at a channel cross section being simulated as 4 times the height of that cross section. The “cross section height” here is the elevation difference between the outer edges and the deepest point.

Something to note from the error message is that the 'water level' isn’t 245.44 m deep, it is 0.4886 m deep. The elevation of the simulated river water surface is 245.44 m above sea level (mamsl). This is not very clear from the phrasing of the error text, but helps to understand when looking at your cross section input to fix the issue.

The hydraulic model engine (MIKE-1D) will assume a vertical wall on either side of a cross section when the cross section is totally flooded, up until the water is simulated as being 4 times the cross section height. At this point it flags that there is likely an unreasonable amount of water.

You will want to check your cross sections: did they realistically capture the channel depth? If you automatically generated your cross sections from a DEM, depending on the resolution of that DEM, you will likely need to manually adjust (i.e. deepen) parts of the cross sections to have a reasonable channel thalweg depth versus the flanking terrain (banks, floodplain). Most DEMs don’t adequately capture the incision of river channels, especially in rugged areas and if the DEM is quite coarse. Getting this error for a water depth of about 0.4885 m suggests that the topography entered for this cross section (Berg reach, chainage: 15500.00 m) may be almost flat: 0.488 m / 4 = 0.122 m channel height (elevation range)?

If deepening your cross sections doesn’t help, or if you think your cross sections are how they should be, also check the longitudinal profile of your cross section thalwegs, or lowest points, moving down the reach. If there is a big bump somewhere water may be pooling upstream of this, contributing to the excessive depth simulated.

Answer 2

I actually found another reason that this error occurs: processing timestep. The timestep control in MIKE-hydro. If fixing cross sections doesn't work, then adjust the timestep length to something that works for your computer. The default is 10 minutes. Mine runs on 30 seconds, where I was previously getting errors about cross sections and exceeding water level maxima.