LPF rewetting not working

[Laura Rod]

Hello

I have a model with three layers. I'm using the LPF package where I set the three layers as convertible and wettable. However, after running the model for over a hundred stress periods i realized that the dry cells remained inactive even though the water table was rising and the dry cells remained wet for several stress periods (please see figure) . Right now I'm not sure why this is happening, and I don't think it's a display issue. Has anyone had this issue before? am I missing any step?

[Sean Czarniecki]

Check your wetting criteria.  There are different ways that a cell can become active again.  One is related to neighboring cells also being wet.  One is related just to the water level in the cell below.  ....and then there is the threshold criteria.  If you want it to definitely rewet, make the threshold criteria very small (the water level being just above the bottom elevation of the cell).  Just note that a smaller criteria can also result in a cell turning on and off every other time step, which could hurt model convergence.  Personally, I do usually go with smaller values unless I see that I'm not converging.

[Laura Rod]

Thanks a lot Sean!

On this last run I activated the "allow wetting of cells" option, wetting factor of 1, wetting iteration interval of 4, and the wetting eqn h=BOT+WETFCT(Hn-BOT).

You mentioned on your reply that you usually go with a smaller value for wetting criteria, how small would you recommend? I reduced the wetting criteria of 0.0001, the wetting eqn h=BOT+WETFCT(THRESH), and it ran without issues; the wetting is working on the cells that were wet at the beginning of the simulation, got dry and wetted again. However, I saw that there are some cells that were dry at the beginning of the simulation and get wet later due to less pumping, but the dry flag doesn't disappear, and instead it is showing the underlying cell as flooded. Do you know what would be causing this?

[Sean Czarniecki]

Without taking too much time on this....my answer to your last question is "time."  If your time steps take too big of a jump, the cell below may show that the one above is ready to get wet, but if the overall convergence criteria have been achieved, your model either comes to conclusion or moves on to the next time step and things change again.  Smaller time steps usually resolves the problem....BUT now that I see your parameters, it looks like you are having it rewet every 4 iterations.....well, that means that there are 3 iterations where that top cell *should/could* have rewet, but didn't.

Rewetting can definitely cause model instability, but I think you are trying the right things.  Your threshold is definitely small enough.

Looking at your figure, I would say that having one or two cells on a slope that don't get wet might be considered acceptable, depending on your needs, although if you can fix it, of course that would be better.

[Laura Rod]

Thanks Sean

Knowing that the rewetting parameters are not crazy is a relief.

The area is very flat, I used an exaggeration of 50 to see how the water table was moving with time. My stress periods are every month, 10 time steps each stress period. I'll make the time steps smaller and hopefully that will fix my problem.

Again, thanks for your time.