Spatial Discretization in HYDRUS-1D | Meshing

[hmal]

Hi,

I’m a new user of HYDRUS and I have two questions to clarify.

1. I’m having difficulty understanding how the "Density" option in the Profile Information section affects the mesh when values are entered. I’ve attached a figure showing the mesh generated using the Density tab (Fig.1).

2. From what I’ve read so far, using a finer mesh generally improves the accuracy of the simulation. However, I’m modelling water vapour movement in an unbound granular material with a profile depth of 300 mm over 250 days, and I'm using a non-uniform mesh with finer elements near the surface (e.g., at z = 0, 0.8, 1.2, and 1.6 mm at top; this was created using above mentioned Density). Despite all time steps converging, the watbal% value of the Balance. out file exceeds 5%, which indicates a problem in water balance. I have attached a figure for your reference (Fig.2)

Could you please help me understand what's causing this and how to resolve it?

Thank you.

[Jirka]

The effects of nodal density on discretization is explained in detail in the Help.

Check out in the manual how the mass balance errors are calculated. They are related to cumulative boundary fluxes. Thus, if you have small cumulative boundary fluxes, you can get a large relative error, even when the absolute error is very small. In you case, you have an absolute error of 0.0919 while the total mass is 36.7. If you relate the absolute error to the total mass, you get a relative error of only about 0.25%. Thus, I would not worry about it too much.

J.

[hmal]

Thank you very much.