Evaporation Inversion

A discussion forum for Hydrus-1D users.
Post Reply
mattpatt
Posts: 3
Joined: Mon Jan 13, 2014 8:46 pm
Location: USA

Evaporation Inversion

Post by mattpatt » Sat Mar 15, 2014 4:15 am

Hello all,

I'm using Hydrus-1D to calculate hydraulic parameters for an evaporation experiment. The experimental setup is similar to that from the example, however the outcome is not nearly as nice. I used the UMS HyProp system which uses a simplified evaporation method (as proposed by Uwe Schindler and Wolfgang Durner). This system gave me an evaporation rate which was calculated by weight changes as was done in the example. In this scenario the evaporation rate happened to be constant based on weight changes. Tension data is taken at two tensiometers located at two known, equidistant heights every 10 minutes. For optimization I am fitting multiple parameters through the van Genuchten bimodal model. The HyProp program can also calculate these parameters, and I wanted to use Hydrus-1D as a comparison. I used the parameters estimated through the HyProp program as the initial guesses for Hydrus. Ksat was calculated separately and is locked. L is also locked. When I run the Hydrus model it fits relatively well, however towards the end of the data, when the tensions start to spread apart, the model lines do not. I have seen this with various other sample models I have run as well. It would appear that that model line is plotted on top of itself for some reason, and does not spread out to better fit the tension data in the dryer regions/ near the end of the dataset. This is problematic for datasets where the difference in tension at certain times becomes larger. I was wondering if anyone had any suggestions as to some mistake I may be making that is not allowing it to split at the end.

What I've tried:
- Fitting L. This works in some cases, however the L value estimated becomes overly high in most cases
- Fitting Ks. I've measured this myself, and when it is allowed to fit, it also becomes overly high or low in comparison to the measured value.
- Trying unimodal van Genuchten. This just does not fit the data at all, as far as I have seen. This may just be due to the soils being bimodal by nature.

Any help is much appreciated. As I mentioned, my setup is actually very similar to that in the example and the data is good (I can't attach my .h1d file, but I've included a picture of the plot), so I'm not quite sure why it won't split the blue and black lines at the end. Thanks!

- Matt
Attachments
example.png
The evaporation example fit, for comparison
example.png (11.72 KiB) Viewed 3463 times
Fit.png
My evaporation inversion, following a similar method as the example
Fit.png (6.23 KiB) Viewed 3463 times

aashok_02
Posts: 1
Joined: Fri Oct 11, 2019 5:34 am
Location: India

Re: Evaporation Inversion

Post by aashok_02 » Sat Feb 15, 2020 10:35 am

hello mattpatt,
I am also working with Hyprop to get the unsaturated zone parametres, and I need to compare those hyprop fit values with HYDRUS 1D but iam unable to run the hydrus 1d
will u please tell me some inputs like
1.Boundary conditions
2.mainly in SOIL PROFILE- GRAPHICAL EDITOR

Thank you sir

Jirka
Posts: 4847
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Evaporation Inversion

Post by Jirka » Sat Feb 15, 2020 8:06 pm

For guidance on how to set up the evaporation experiment in HYDRUS-1D see the Evapor example in the Inverse Project Group. This example is also described in Section "11.8. Example 8 - Inverse Analysis of an Evaporation Experiment" and in:

Šimůnek, J., O. Wendroth, and M. Th. van Genuchten, Parameter estimation analysis of the evaporation method for determining soil hydraulic properties, Soil Science Society of America Journal, 62(4), 894-905, 1998.

J.

Post Reply