Wrong concentrations solute transport

A discussion forum for users of the new HYDRUS 2D/3D. HYDRUS is a software package for simulating water, heat and solute movement in two- and three-dimensional variably saturated media. Happy Posting!
Post Reply
emiel
Posts: 4
Joined: Mon Mar 06, 2017 12:58 pm
Location: Netherlands

Wrong concentrations solute transport

Post by emiel » Wed Mar 08, 2017 5:59 pm

Dear,

I am trying to simulate pesticide transport: from the moment the pesticides are contributed to the land, till the outflow of the pesticides through drains. I have made a model (800 cm x 235 cm) with on top a Atmospheric boundary condition (precipitation and evaporation), no flux boundary on the sites, a constant head on drain level and a variable flux on the drain (one drain) in the center of the model.

I am using Standard solute transport to simulate pesticide transport and I started modelling with a conservative compound (no sorption (Kd=0, Beta=1) and decay).

Parameters:
- 732 days, timestep = 1 day
- Time Weighting Scheme: Crank-Nicholson Scheme
- Space Weighting Scheme: Upstream Weighting FE
-Pulse duration (day): 1
- Stability criterion 2
- Absolute Concentration Tolerane: 0
- Relative Concentration Tolerance: 0
- Cvalue 1= 1 ug/cm3 on the third day, all the other days Cvalue1=0 (day three: Prec=0.83 cm/day, Evap=0.01 cm/day)

If I run the model with these parameters, and a targeted FE size of 21 cm, the concentrations of the compound show a strange trend. Really high (max. 5e+18 ug/cm3) and really low concentrations (min. -4e+18 ug/cm3) enter the model, especially after +-150 seconds. The peclet number is in between 52-64. It seems if there is not only infiltration of the solute on the third day, but also on the other days (but not sure, the problem could also be numerical).

If I change the targeted FE size to 10 cm, the same strange trend is seen, just as the really high (max. 6e+24) and low (min. -5e+24) concentrations.

What am I doing wrong? Or what can I change in the model to get better results?

Thanks!

Jirka
Posts: 4762
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Wrong concentrations solute transport

Post by Jirka » Wed Mar 08, 2017 6:30 pm

If there is evaporation (on other days than three) which is larger than precipitation, you will likely need finer vertical discretization close to the top boundary. Note that you can make discretization larger in the horizontal direction than in the vertical direction using FE Mesh Stretching. If you do not have sufficiently fine vertical discretization at the soil surface, you may get numerical problems. You should also increase dispersivity and define diffusion so that you lower the Peclet number to 2-5. Make sure also that your hCritA is not excessively high (see the guidance in the help). J.



emiel
Posts: 4
Joined: Mon Mar 06, 2017 12:58 pm
Location: Netherlands

Re: Wrong concentrations solute transport

Post by emiel » Wed Apr 26, 2017 3:51 pm

Thanks a lot for the helpful information, I solved the problem. I have a different problem now regarding the units of solute transport:

I set a solute concentration of 2 ug/cm3 at Cvalue1 at the third day of the model, the solute concentration takes a day. If I open, from the results, the text file for the solute flux through the atmospheric boundary, it shows a solute flux of -1.328 e3 ug/cm/days for the whole day.

In the help function I found the units of concentration and solute flux:
- Concentration: mass of solute/ volume of water
- Solute flux: mass of solute/ volume of water * length of soil/ time unit (days)

If I calculate the solute flux the answer differs: 2 ug/cm3*(800 (width of model)/1)= 1600 ug/cm/days. What is the reason for this? Is the model still not stable enough (peclet number max. 5.6)?

Thanks!

Jirka
Posts: 4762
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Wrong concentrations solute transport

Post by Jirka » Thu Apr 27, 2017 11:13 am

There is not enough information in the question to be able to answer it. J.

emiel
Posts: 4
Joined: Mon Mar 06, 2017 12:58 pm
Location: Netherlands

Re: Wrong concentrations solute transport

Post by emiel » Fri Apr 28, 2017 11:45 am

Sorry, some more information:

I have made a model (800 cm x 235 cm) with on top a Atmospheric boundary condition (precipitation and evaporation), no flux boundary on the sites, a constant head at the bottom of the model with the constant head on drain level (at - 80 cm - surface level) and a seepage face boundary on the drain (one drain) in the center of the model.

The model is build up out of sand, 2 surfaces with the same characteristics, with a small poorly permeable clay layer at the bottom of the model to simulate the seepage flow. The FE-mesh has a targeted FE size of 21 cm with a FE-MESH refinement in the top surface with a element size of 5 cm.

I am using Standard solute transport to simulate pesticide transport and I started modelling with a conservative compound (no sorption (Kd=0, Beta=1) and decay). The solute transport is time-variable and the only solute infiltration occurs on the third day.

Parameters:
- 732 days, timestep = 1 day
- Time Weighting Scheme: Crank-Nicholson Scheme
- Space Weighting Scheme: Upstream Weighting FE
-Pulse duration (day): 1
- Stability criterion 2
- Absolute Concentration Tolerane: 0.01
- Relative Concentration Tolerance: 0.001
- Maximum number of iterations: 10
- Cvalue 1= 2 ug/cm3 on the third day, all the other days Cvalue1=0 (day three: Prec=0.83 cm/day, Evap=0.01 cm/day)

I hope the information above is enough for an understanding of the model!

Thanks.

Jirka
Posts: 4762
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Wrong concentrations solute transport

Post by Jirka » Fri Apr 28, 2017 4:29 pm

Check the relative mass balance error (in Mass Balance Information). If the mass balance error is larger than 1% for water flow or 2-3% for solute transport, then your spatial discretization is not adequate. Read the general rules about spatial and temporal discretization (FAQ 8):
http://www.pc-progress.com/Documents/No ... zation.pdf

J.

Post Reply