eCLM-ParFlow: high ice concentrations

[kvrigor]

Migrated from https://gitlab.jsc.fz-juelich.de/HPSCTerrSys/tsmp-internal-development-tracking/-/issues/62

Originally reported by Mikael Kaandorp (Nov. 13, 2023).

---

To summarize; in the coupled simulations the volumetric soil moisture (H2OSOI) reaches unphysical values of much larger than 1 (I observed values larger than 10). When splitting the volumetric soil moisture in its liquid (SOILLIQ) and frozen (SOILICE) components, the liquid part looks reasonable (<1.0), the frozen part not so much. See the figures below, where I plotted the different variables at the end of february:

• Input files and namelists: /p/project/cjibg36/kaandorp2/eTSMP_setups/setup_etsmp_cordex_444x432_v2
• ERA5 forcing files: /p/largedata/jibg36/kaandorp2/setup_TSMP_DA/ERA5_EUR-11_CLM/

[kvrigor]

More from Haojin Zhao (reported June 7, 2024):

---

To investigate ice coupling behavior, I modified the start date in drv_in from the original 1 July, 2017 to 1 Jan, 2017, as this colder period is more suitable for evaluation. Simulations were run for both the no-ice and dev-soil-ice repositories, focusing on the surface layer (0.02m depth). Here's a summary of the findings:

1. The simulations from both runs give in general similar results for the variable H2OSOI.

2. Unrealistic values (>1) observed in the northern part appear to have vanished when compared with previous simulations by Mikael's, but still exists in central Europe. Note that the unrealistic soil moisture are only found in the first surface layer.

[kvrigor]

TODO

• Migrate EUR-11_eCLM-ParFlowv3.12_testcase to TSMP2_workflow-engine
• Replicate this issue with updated versions of parflow/dev-soil-ice and eclm/dev-soil-ice.
• Stefan K's suggestion: Plot SOILICE and SOILLIQ (kg/m²) instead of volumetric soil ice and volumetric soil water (mm³/mm³)
  • eCLM history field list