Issue in Simulating Soil Moisture

[Janani Kandasamy]

Hello,

I am encountering an issue with soil moisture simulations in EPIC V 1102.

In my simulations, soil moisture appears to remain constrained within a narrow range, often saturating around a particular value. Additionally, I observe abrupt day-to-day fluctuations, where soil moisture drops sharply on some days and increases sharply on others. This behavior is consistent across the entire simulation period. I have also tried to change a few parameters - WTMN, WTMX, FFC, FC and so on, SM value have changed but the pattern remain the same.

To further investigate, I ran EPIC using the default input datasets provided on the EPIC website. However, I observed a very similar soil moisture pattern in those simulations as well, which makes me wonder whether this behavior is related to model structure, default parameterization, or specific soil/hydrologic settings.

Please let me know what might be causing this behavior and whether there are particular parameters or settings (e.g., soil hydraulic properties, percolation, evapotranspiration, or control file options) that I should review or adjust to obtain more realistic soil moisture dynamics.

Thank you very much for your time and support.

Best regards,
Janani

[Luca Doro]

Dear Janani,

I am interested in understanding the reason for this behavior (as time allows). In the plot SM_Italy, I guess you are comparing the soil moisture simulated with three different approaches used to simulate soil temperature. Am I correct?

A few things you can already check are:

• Approach used to drive the wilting point and field capacity (input vs estimated). Check variable ISW in the control table (file EPICCONT.DAT).
• Approach used for saturated conductivity. Check variable ISAT in the control table.
• Approach used to estimate water percolation. Check variable IPRK in control table.
• If you are using the VSHC approach for water percolation (IPRK=1), check the value of PARM(93) in the parm file.

Let me know what you find.

Best,

Luca

[Janani Kandasamy]

Dear Luca, 

Thank you for the response and clarification! I have simulated soil moisture as per your suggestions.

After making these changes, I do see a shift in the absolute soil moisture values across the simulations. However, the overall temporal pattern remains largely unchanged — in particular, the sharp drops and recoveries repeat consistently across years and across the different configurations. The differences are mainly in magnitude rather than in the pattern of the soil moisture dynamics.

Please find the attached plot of simulated SM.

Please let me know your comments.

Thanks again for your time.

Best regards,
Janani

[Luca Doro]

Dear Janani,

Thank you for the update on your test. The results are not completely clear to me, but I believe I understand your point. My guess is that the concern is the abrupt change in soil water content during the summer. What I would do to understand what is happening is to plot the soil water content along with evapotranspiration, percolation, and precipitation.

This can give you an idea on what is driving the quick reduction in soil water content.

After this, we can decide how to move forward.

Luca

[Janani Kandasamy]

Dear Luca,

Thank you for the clarification. I have plotted precipitation, percolation, and ET to investigate the sudden rise and fall in soil moisture during summer. I found that soil moisture generally follows percolation; however, both does not appear to replicate precipitation. In the attached plot, you can see that from June 2023 to early August 2023, there were multiple small rainfall events, but percolation is almost zero, and the soil remains very dry. Intuitively, the soil moisture should have increased with these continuous rainfall events, but that does not seem to be happening. Please let me know what might be causing this.

Thank you for your guidance.

Best regards,
Janani

[Luca Doro]

Jenani,

Are you getting the precipitation data from the model output or from the daily weather file? I am asking this because I agree that percolation and soil moisture seem to be disconnected from precipitation, therefore I am wondering if the model is generating the precipitation data instead of using the data provided in the daily weather file.

Let me know.

Luca

[Janani Kandasamy]

Dear Luca,

Thanks for the response. The precipitation is being read from the daily weather (.dly) file, not generated by the model. I prepared a site-specific .dly file and  assigned the model to use it as the weather input. I also verified this in the .OUT file, where the precipitation values match those in the daily weather file. I’m attaching the relevant .OUT file for your reference.

Please let me know if you’d like me to double-check any other settings or outputs.

Thanks again for your time.

Best,
Janani

[Luca Doro]

Thank you for sharing the OUT file.

I see that weather data are generated starting from September 22, 2024. I believe this is outside the period considered in your simulation, so it shouldn't be a problem.

One thing you can do is to pick a few days with rainfall and look at rainfall, evapotranspiration, and runoff just to have a general idea if everything looks correct. After this, I can try to look into it myself as time allows.

Luca

[Janani Kandasamy]

Dear Luca,

Thank you for your reply.

The weather data in my setup actually start in January 2019 and end on September 22, 2024. Do you think I need to extend the weather file through December 2024? However, I believe this should not affect the soil moisture simulation for the period prior to September 2024, which is the time window I am using for my analysis.

Please let me know your thoughts.

Best,

Janani