Dual Porosity with two-site sorption

[gurparshad]

Good Morning Sir,

I am using the dual-porosity with two-site sorption model in the inverse mode to obtain solute transport parameters for a solute. The parameters I am optimizing are: Frac, Frac_M, Omega and Alpha. I obtained a good fit between the observed and the simulated data (R2=0.93). However, as per the optimized parameters, the entire sorption is happening in the mobile phase as 'Frac' came out to be 1. Therefore, I am wondering if this could be a realistic scenario since, as per my understanding, most of the sorption should have been happening in the immobile zone as more area is available for the solute to get sorbed? Also, I am aware of a lot of macropore flow happening in my soil due to the presence of a lot of earthworms. I have attached my model with this query.
Please guide me on which parameters I could change if you think this scenario is something unrealistic, considering the presence of macropore flow.
Thank you in advance.
Best regards,
Gurparshad

[Jirka]

I will not have time to look at your project due to my current travels.

I would also expect that more sorption occurs in the soil matrix rather than in macropores. If the model fits parameters that indicate the opposite, this likely reflects the observed solute retardation and thus the need for sorption in the main flow path (I’m hypothesizing without looking at the data).

J.

[gurparshad]

Thank you sir. I am sorry, I am afraid I did not fully understand 'the observed solute retardation and thus the need for sorption in the main flow path'. Could you please expand on this?

Best regards,
Gurparshad

[Jirka]

If the model did not fit sorption parameters in the mobile phase, it would predict earlier solute arrival (no retardation) than is observed in the data. J.

[gurparshad]

Thank you sir for your response. Actually, the model fitted the sorption parameters. The optimized values obtained were: Frac= 1, Frac_M=0.0035, Omega=0.000026 and alpha=0.0039.
Regards,
Gurparshad

[Jirka]

As I wrote above, I do not have time to look at your project due to my current travels. Anyway, it is entirely possible that your data (information that you enter into the objective function) does not provide enough information to fit such a complex model. The fitted parameters are then likely not well defined (which would be indicated by very large confidence intervals, often extending to negative values). J.

[gurparshad]

That makes sense, thank you so much Sir.

Best Regards,
Gurparshad

[gurparshad]

Hello Sir,

I have one additional question.
1. I have added fertilizer on top of the soil. Therefore, I have cTop data, which I got from the fertilizer solute release test. So I have the time variable concentration of the incoming solution at the upper boundary of my soil throughout the experiment.
2. I have collected leachate from the bottom of the soil, and am using it as inverse data to optimize the solute reaction parameters.

My concern is that, what should be the value of cBot for my model? Should it be the same as my inverse data, as it is the concentration of solution at the lower boundary?
Thank you!

Best Regards,
Gurparshad

[Jirka]

You can select (in the Inverse Solution window) the concentration mode. You should select the “flux concentration”. J.

[gurparshad]

Hello Sir,
I already used 'flux concentrations' in the inverse window. I used type 4 as we are entering concentration versus time data. This is the effluent data that I am collecting at the bottom of the soil. My only concern is can we use this flux concentration data (that I used as inverse data) as 'cBot' as I am using time-variable boundary condition for solute? I have cTop data, which I got from the fertilizer applied on the top but not sure what to input for 'cBot'.
Thanks,
Gurparshad

[Jirka]

I do not understand why you would want to use simulated outflow concentrations as a bottom BC. I assumed that you are using a zero concentration gradient (which is common in such examples) as a bottom BC. If you use it as a bottom BC (you cannot define bottom BC using outflow concentration, only for inflow), how would you define the objective function? J.

[gurparshad]

Thank you Dr. Jirka for your reply.

Indeed I have used 'Zero gradient' as the lower boundary condition.
You are right, having inverse data and cBot with the same values doesn't sounds right. So, I wanted to confirm with you before going any further.
The effluent data I collected from the data, is being used as 'Inverse data' for parameter optimization.

So, Would my cBot values in the 'Time variable boundary condition' window be zero?


Thank You,
Gurparshad.

[Jirka]

Zero, whatever. This info is ignored when a zero gradient BC is used. J.

[gurparshad]

Thank you sir, for your reply.

[gurparshad]

Hello Sir,
I have used inverse simulations to determine the solute transport parameters for several soil columns in my field and have achieved a good fit in all, except one column. For this particular column, regardless of the initial values of optimized parameters, the simulated solute consistently exhibits an abrupt appearance and disappearance at the bottom of the column. Therefore, I am wondering if you can suggest adjusting any particular parameter so that the falling limb of the solute breakthrough curve has a longer plateau and the solute leachate concentrations do not reduce abruptly. I acknowledge that providing specific recommendation without the knowledge of other conditions of the model is challenging, but any suggestion or guidance from your experience would be really helpful to me as I am stuck on this column since last year. Few information regarding the parameters:
1. I am using time-variable boundary condition for the solute addition on the top.
2. I am using Dual-Porosity Physical Nonequilibrium model and optimizing 2 parameters: Frac and Alpha. The initial values are frac=0.001 and Alpha= 0.01. The model optimized parameters are: 0.27237E-02 and 0.29593E-01, respectively.
3. The fixed parameters are: Dispersivity= 7.37, ThetaIm=0.11, Kd=84 and Beta=0.27
4. I am collecting leachate at the bottom of the soil column and using that leachate data, I am optimizing Frac and Alpha using inverse modeling.
5. Solute UBC= concentration flux and LBC: zero gradient
Attached is the image of the HYDRUS output.
Thank you for your time and guidance!
Best Regards,
Gurparshad