UZ water balance error

I have a coupled MIKE SHE/MIKE 11 model set up for a small ag watershed; UZ, SZ, FDOF components. My watershed is about 3 km2 size and cell size is 45x45m. My current model set up is giving me an average annual water balance error of about 3% of annual rainfall. Is it normal or little excessive? Most of this error is due to water balance error in UZ; i am using Richards equation. For example, for a single day rainfall of 130 mm the UZ error is -34 mm (it is the maximum rainfall during my 6 yrs simulation period). The WB tool have shown that daily UZ errors are more on the days of heavy rainfall.
Which options should i try to reduce this error in UZ simulations?
Thank you,

Hi,
Based on your description you should try and reduce the WB error. An error of 3% on average might be ok, but the error during the high intensity rain events is too high.

Some of the issues to look at:
- Look into the WM_Print.log file to evaluate if the maximum iteration is exceeded. If so try to adjust the SZ and UZ max time step so they are closer. The SZ max time step should not be more than 4 times the UZ time step.
- Evaluate the precipitation time step control, and try reducing the values as this will force additional time steps during rain fall events – this will increase the total computational time but might reduce the WB error
- Evaluate the spatial extent of the error, and if it’s mainly on a single soil type or if it’s distributed across all soil types. This might lead to an evaluation of the soil parameters if it’s mainly on a single soil type.

Oluf (DHI)