pressure head output exceeding upper limit of tension interval

A discussion forum for Hydrus-1D users.
Post Reply
tmcdowell2
Posts: 1
Joined: Sun Nov 17, 2019 10:12 pm
Location: USA

pressure head output exceeding upper limit of tension interval

Post by tmcdowell2 » Mon Mar 02, 2020 7:07 pm

Hey Everyone,

I am trying to simulate modern-day calcite dissolution-precipitation in a 453cm deep paleosol-loess sequence. My soil column has a node density of 905 which results in about 0.5cm spacing. I'm using UNSATCHEM with CO2 transport. The climate is semi-arid.

I have been able to run the model for 1 year with reasonable errors. However, I'm having a lot of difficulty extending the model to longer time periods. I have the model time units in days. I'm repeating the same 365 time variable conditions for multiple years.

The model stops converging after a few years and the model always stops converging around the same day range within the time variable conditions (171-182). Also at the point where the models stop converging, it is always the same observation node (node 4) where pressure head drops below the Upper limit of the tension interval. I believe this is why the model stops converging and this also makes me think that my model is unstable. However, I have no idea how to fix this. I was wondering if anyone has any advice on how to increase my model stability so the pressure head stays with in the limits of the tension intervals.

I've tried increasing the number of iterations, decreasing the upper limit of the tension interval to the wilting point, increasing the lower limit of the tension interval and increasing the initial pressure head in the soil profile.

Possible solutions that I have not tried for various reasons, but have considered
1. increasing the nodal density around the observational node where pressure head drops below the upper limit. I have not tried this yet because it would be a lot of work to adjust all the layer boundaries. I'm hoping there is a simpler fix.
2. Time units are too long--should be hourly. I have not tried this because I do not have access to hourly meteorological data at my site.
3. I'm pushing the model too far/it is too complex.


Thanks,
Taylor

Post Reply