Retreating lakes cause local groundwater head increases

[rubencalje]

Is your feature request related to a problem? Please describe.
We use the Lake Package to simulate a shallow lake with vertical lake–aquifer connections. When the lake level drops and the lake shrinks or retreats, some cells that previously acted as sinks (draining the aquifer) become inactive or disconnected. This reduces the drawdown effect of the lake in those areas, allowing groundwater heads to rise (or rebound) due to surrounding recharge or lateral flow. However, I would expect groundwater heads to decrease everywhere as the lake retreats.

Describe the solution you'd like
It would be helpful if, at locations where the lake used to be, the Lake Package could maintain a draining boundary condition with the head set at the ground surface. The drained water would represent overland flow and should be added to the lake water balance. This feature could also improve model convergence.

Describe alternatives you've considered
I considered adding a Drain boundary at cells that have become dry and using the Mover Package to route the drained water back to the lake. However, this approach causes issues when the lake level rises again and those cells become re-flooded.

Additional context
I’m interested to know if others have encountered the same issue and, if so, whether they have found a workaround or would support adding an option in the Lake Package to include a surface-level draining boundary for inactive vertical lake connections.

[FTW12345]

Would it perhaps be an option to enlarge layer thickness to prevent that cells bordering the surface ('maaiveld') AND the present or former lake bottom can become dry / inactive? (Typing this without knowing yet how the lak package interacts with aquifer cells.)

[rubencalje]

The cells themselves do not become inactive (I was not clear about that), but the draining boundary condition becomes inactive. So I think increasing the thickness of the cells is not the solution for my problem.

Thank you for the response though!

[emorway-usgs]

Hello @rubencalje, if you're willing to put together a small FloPy script that creates a small model with the situation you describe that would be helpful in looking into supporting this. There might be other considerations that aren't readily coming to mind that make this more of a challenge than first anticipated. Have a small example will hopefully help illuminate all the potential issues.

[FTW12345]

In my comment, I assumed different MODFLOW behaviour. Being incorrect, I removed my comment.

[rubencalje]

Hello @rubencalje, if you're willing to put together a small FloPy script that creates a small model with the situation you describe that would be helpful in looking into supporting this. There might be other considerations that aren't readily coming to mind that make this more of a challenge than first anticipated. Have a small example will hopefully help illuminate all the potential issues.

Thank you for the response. I will try to generate a simple model using FloPy. This might take some time however.

Having looked further into this issue, I have planned for my case to use the packages RCH, EVT and LAK, and specify them as follows:

• RCH will be active for all top cells (first model layer). (This means that recharge will be only specified through this package and not through LAK.)

• EVT:

  • will also be active for all top cells;
  • will be corrected for vegetation;
  • (which is a function of the days in a year; season dependent):
  • (A).

• LAK:

  • LAK cells will be placed upon cells where small lakes (pools) emerge in the wet season and disappear in the dry season (cells with a relatively low altitude);
  • For those LAK cells, we will assign evapotranspiration for open water (B) *** minus *** the for vegetation corrected evapotranspiration values (A): (C).

For the pools, we assume that their bottoms do not act as barriers. Thus, 100% of the recharge enters the model.

Effect:

• When lak cell is wet: Evapotranspiration will be (A) + (C) = (B). (Because A+(B-A)=B).)
• When lak cell is dry: Evapotranspiration will be (A).

Independent of the lak conditions (pools present in the wet season; pools disappeared in the dry season):

• recharge will always be active;
• evapotranspiration will always be active and have representative values (dry: vegetation values; wet: open water values).

This behaviour should be reflected in the modelled heads, also near the border of the pools.

I am not sure if this is a solution to my problem. It might be better to create your own issue.

One comment that might help you though: MODFLOW 6 automatically sets the flux from RCH and EVT to 0 at cells that are connected to an active lake (where the lake stage is above the lake bottom). So you can and need to specify recharge both in the LAK- and RCH-packages. And you can specify evapotranspiration for open water (B) in the LAK-package, and the for vegetation corrected evapotranspiration values (A) for the EVT-package.