Skip to content

eCLM-ParFlow: Bedrock depth treatment #130

Open

Description

@s-poll

Priority: low. This inconsistency issue is long known, however not systematically tackled yet.

Summary

In a ParFlow-coupled run there are two independent descriptions of where the permeable soil ends:

  • eCLM reads zbedrock from its surface dataset and converts it into a layer index col%nbedrock (src/clm5/main/initVerticalMod.F90:495-517). With use_bedrock = .true. everything below that index is "rock" for eCLM.
  • ParFlow gets its subsurface from its own indicator/permeability fields and simply solves Richards' equation over its full vertical grid.

eCLM static file generator should account for the bedrock layer and sync with ParF. In addition, the layers below nbedrock are treated in the coupled code path differently from how standalone eCLM treats them.

Issue 1: bedrock depth is not part of the preprocessing

Synchronisation of the eCLM zbedrock and the ParFlow indicator field (SoilGrids/BGR + alluvium). Currently, they come from separate chains. zbedrock on the eCLM surface dataset and the ParFlow indicator/permeability field are produced independently, so a column can be bedrock from ~2 m down in eCLM while ParFlow keeps transmitting water through the same depth range (and vice versa).

Issue 2: below-bedrock layers behave inconsistently in the coupled path

Done by AI, need to be checked in person.

quantity standalone eCLM (use_bedrock = .true.) eCLM-ParFlow
h2osoi_liq / h2osoi_ice, smp_l solved for 1..nbedrock only; below-bedrock layers start dry (h2osoi_vol = 0, src/clm5/biogeophys/WaterStateType.F90:875-885) and stay dry overwritten from ParFlow for all 1..nlevgrnd in soilwater_parflow (src/clm5/biogeophys/SoilWaterMovementMod.F90) — below-bedrock layers now carry water
hk_l computed for 1..nbedrock (compute_hydraulic_properties) same range (see #128) — layers nbedrock+1..nlevsoi keep their initial value
soil thermal properties thk = thk_bedrock, cv = csol_bedrock for j > nbedrock (src/clm5/biogeophys/SoilTemperatureMod.F90:701,819) unchanged — the layers are treated as dry rock although they now hold ParFlow water
water passed to the coupler — lnd2atmMod.F90:524 sums h2osoi_liq/ice only up to nbedrock
water table / baseflow referenced to zi(c,nbedrock) (SoilHydrologyMod.F90:1885-1914, 2095-2110) ParFlow owns this, eCLM's bedrock reference is bypassed
root fraction vegetated rootfr is not truncated at nbedrock (only pervious road is, SoilStateInitTimeConstMod.F90:218-223), while PhotosynthesisHydraulicStress loops j = 1, nlevsoi same

Two consequences worth checking explicitly:

  1. eCLM's bedrock is effectively bypassed for water storage (ParFlow fills the layers) but still active for the thermal state and for the aggregation in lnd2atmMod. So the same layer is "wet" for H2OSOI/SOILLIQ/TWS diagnostics and "dry rock" for soil temperature.
  2. Because rootfr is not truncated at bedrock but hk_l is only defined up to nbedrock, roots sitting in layers nbedrock+1..nlevsoi see soil_conductance = 0 in the plant-hydraulic-stress scheme (PhotosynthesisMod.F90:2841). This is inherited from standalone CLM5 rather than caused by the coupling, an IMO is fine.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    biogeophysicsIssues related to biogeophysical aspects of eCLM.coupled-parflowIssues related to eCLM-ParFlow coupling

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions