Issue modeling a soil with zone of unsaturated vapor

Hi,

I am using EOS3 to model a evaporating soil and have encountered some behavior with soil air relative humidity that I think is unrealistic.

My 1D setup has a static block at the bottom to act as the water table (persistent water source), a set of soil blocks above it, and another static block at the top to act as the atmosphere (with relative humidity, RH, set by air mass fraction in single-phase conditions). I spin up the model with a flow-only scenario of continuous rainfall (injection into the top soil block).

Then, I use the flow-only output as initial conditions for simulating an evaporating soil by setting water injection to zero. This setup produces expected results when the entire soil column is vapor-saturated (i.e., soil air RH = 100% in all blocks) in that liquid saturation (Sl) reaches a minimum residual value (Slr) set by the Van Genutchen parameterization for relative permeability.

However, I encounter unexpected behavior when I simulate scenarios of more intense soil drying, where soil air RH becomes less than 100% (see picture, black line and x's). Initially, drying proceeds as in the vapor-saturated scenario. However, as soon as a soil block dries to vapor-unsaturated (i.e., liquid saturation goes to zero and soil air RH becomes <100%), its RH immediately drops to the atmospheric value (40% in this case). In other words, at the interface between vapor-saturated (100% RH) and vapor-unsaturated (<100% RH) soil air, there is an abrupt change.

I don’t expect this abrupt change. Instead, soil air RH should vary smoothly from 100% at the vapor-saturated/vapor-unsaturated interface to atmospheric conditions at the soil-atmosphere interface (see picture, red dashed line). In essence, the soil should act as an extended boundary layer over the 10s cm from the soil surface to where soil air becomes saturated.

I’m wondering if there is some glaring issue with my model setup causing this behavior, but I have played around with the mesh spacing, maximum time step, initial conditions, and the functions for relative permeability and capillary pressure (Van Genutchen setup) without success.

I have also attached input files for the flow-only spin up and the soil evaporation simulation.

Thanks,

Tyler