Different ways of fitting Van-Genuchten model on some experimental data - Part1

And here is my first scientific post. Actually I'm really glad for doing this ...I mean, I'm sharing what I have been through once and I hope this all become helpful to somebody one day.

Ok, in the past month I needed to fit the Van-Genuchten model on some experimental data to use for further analysis (Which I will probably explain in later posts).

So ...., it sounds easy job to do at first.

You have for example 9 suction head and the corresponding water content to each suction head. Then you would probably give these data to the solver in excel to estimate Van-Genuchten coefficients for you. But there are some other more sophesticated methods out there, which they can help you to have more precise estimate of your coefficients.

In the following I will try to explain some of these more complicated ways in complexity order:

Let's first define a scenario. we have 3 treatments and 3 replications for each treatment. We are intersted to see the effect of each treatment on soil water retention curve.

1- The first and mybe simplest way is to use RETC or SWRC fit model (http://swrcfit.sourceforge.net). Both of these two models use nonlinear curve fitting appraoch to do this job. You need to just have this in the mind that, you have to feed these two models with all 3 replications data at the same time .

In the RETC model you would be able to change the intial values for starting the curve fitting procedure while for the other model you won't.

The other thing that I need to mention is that you would be able to check the other water retention models like dual prosity models on your data by using these two models.

Warning: Nonlinear curve fitting procedure is highly dependant to the intial guess and wrong or out of range guesses would lead you to hell :D just kidding ... very very out of range results like negative 'n' or negative theta_r or theta_s.