Initial and lower boundary conditions for rice paddy simulations in HYDRUS 1D

[SHYMA M]

Sir
I am simulating an experimental study for lowland rice paddy in HYDRUS 1D.
Total soil depth under consideration = 130 cm
Groundwater table = 200 cm
initially there was a ponded water depth of 1 cm.
Modelling in Hydrus 1D:
Soil texture is Clay (7 layers)
For giving the lower BC as constant pressure head, I have considered total soil column as 200 cm.
Upper BC = atmospheric BC with surface layer (h=10 cm)
Initial condition = -10 kPa (near to field capacity)
I have simulated the experiment in oryza, dssat and hydrus model.
unfortunately the surface runoff generated by hydrus is not matching with oryza and dssat after some time step. Please tell me any suggestions to improve the model. I have tried playing with increasing soil depth, initial pressure heads, lower boundary conditions. But it is either not converging or giving bad results. If there is any option in hydrus to select the solver for convergence, please let me know that as well. I have attached the modelling results from all the models here with (for surface runoff and ponding water depth)

[SHYMA M]

In surface runoff plot grey line shows results from Hydrus.

[Jirka]

I’m not familiar with Oryza, and thus cannot comment on that.
DSSAT has a tipping bucket approach to flow, which is undoubtedly inferior to the numerical solution of the Richards equation used by HYDRUS.

In the graphs in your post, it seems that you limit the surface water layer to 10 cm in Oryza, while in HYDRUS you go over this limit, which may be one reason for the smaller runoff in HYDRUS.

J.

[SHYMA M]

Thanks for the reply sir.
In all the models I have given Bund height as 10 cm. so whenever the input water exceeds 10 cm height it will generate surface runoff. In ORYZA model, it will directly give the the depth of ponded water layer. whereas in HYDRUS it gives in terms of pressure heads at surface. I have given upper boundary condition as atmospheric layer with h= 10 cm (to represent actual bund height in the field). Am I right sir?
Since I have given initial condition as -10 kPa through out the soil profile (200 cm total) and lower boundary condition is kept as constant pressure head, will this affect my results ? because ground water level is 200 cm from top and pressure at that point is 0. I have tried giving 0 pressure head at 200 cm depth keeping -10 kPa for all other nodes, but the results are not converging.

[Jirka]

Using “the upper boundary condition as atmospheric layer with h= 10 cm” is correct.

HYDRUS uses the length units for the pressure head, not the pressure units. Thus, the initial condition should be given in cm, rather than kPa. The initial condition needs to be realistic, and thus you should have a hydrostatic equilibrium above the bottom BC until you hit your initial value (i.e., -10 kPa converted to cm), and then you can keep it constant in the rest of the profile. J.

[SHYMA M]

Sir
I have given -10 cm as initial condition (sorry that in earlier post I mentioned kPa units). Since the profile is nearly saturated initially I have given -10 cm as initial condition for the entire profile irrespective of the depth of ground water table (0 pressure head). Is this fine ?
One more thing, I have noticed there is a shift in the plots of surface runoff and ponded water in hydrus compared to oryza. Values of surface runoff and most of the surface pressure heads in hydrus is matching with oryza, but its occuring one timestep ahead. Under time information window of the pre processing, I have given initial time as 160 th day and final time as 295 th day. There are 135 time variable BCs and meteorological data (starting from 161 th day to 295 th day). So I am in doubt that when generating the results hydrus gives the data from 160 th day or 161 th day onwards. I attach some pictures for reference.

[SHYMA M]

This is the results after unchecking the "T level information-every time steps" under print information.

[Jirka]

You need to understand how the time stepping in these codes work. If you have initial time equal to zero, then the first boundary record is given for 1 d, and corresponding fluxes are applied between zero and 1 d, etc (since it is how we define the first day). In your case with initial time of 160 d, the record given for 161 d is applied during the interval 160-161 d. J.

[SHYMA M]

I understand sir. Thank You.