Add new feature: microbially explicit soil decomposition module

ecaas commented 3 years ago

$HOME = path/to/ctsm

Purpose with this issue

Implementing a module for soil decomposition. This is intended to work as an alternative to the CENTURY-based decomposition cascade at $(HOME)/src/soilbiogeochem/SoilBiogeochemDecompCascadeBGCMod.F90, and will therefore interact with several files in the soilbiogeochem-folder.

The purposed module includes mycorrhizal activity, and will therefore also interact with the FUN module $(HOME)/src/biogeochem/CNFUNMod.F90.

Brief science information

We propose a module that explicitly represents the activity of decomposing saprotrophs and mycorrhizal fungi in the soil. By doing this, we want to better understand how these microorganisms react to climate change, and how that may affect the carbon climate feedback.

Brief model information:

We have built a model framework inspired by the MIMICS model Wieder et al. 2015 and Baskaran et al. 2017. The model is vertically resolved, and includes separate carbon and nitrogen reservoirs for saprotrophs (two pools) and mycorrhizal fungi (3 pools), the plant (one pool), plant litter (two pools), and soil organic matter (3 pools) in each soil layer. In addition, one inorganic N pool is included. An overview of the represented processes is given below.

We intend to refine model parameterizations based on observations/experiments in Boreal/Arctic environments, and mainly use them for site and regional simulations in these areas.

Plan for the implementation of the soil decomposition functionality:

Until this point, the soil decomposition model has been developed as a stand-alone module. Although further refinements are required, it has proven itself to be able to produce steady state results that are comparable with observations.

Some key points on the present state of the model:

It is written in Fortran90
C and N concentration in each pool and layer is stored in arrays that are updated every timestep. Fluxes between pools are stored as single values, and also updated every timestep. We plan to replace this setup to use types instead.
C enters the system through the plant pool, and leaves through heterotrophic respiration.
N enters the system through deposition, and leaves through leaching.
The module needs information about soil temperature, soil moisture, and clay content.

We would like the further development of the module to be in line with the CTSM structure and framework, so a merge into the framework is possible. Thus, we would appreciate any thoughts and inputs, especially considering:

an object-oriented vs procedural based structure for the model components. For us, it seems natural to chose an object-oriented approach for dealing with the flow of data.
- Would you prefer us to introduce our own objects, or is it natural to work with a type already present in CTSM?
- How could the described model parameters be adapted to the used variables in CTSM?
At which point of the function calls of CTSM should the purposed model be included?
How would the purposed model relate to FATES?

Long term goal

Add the above mentioned module as an alternative soil decomposition routine in CTSM.
Couple with MOSART-DOC (issue #1216)

People involved: @sunnivin

wwieder commented 3 years ago

Super exciting to hear about his work.

We should coordinate before too long because @slevisconsulting will be starting to work on the basic MIMICS implementation in CTSM soon. It would be nice to try and share as much code / infrastructure as possible for these parallel projects, if possible.

ecaas commented 3 years ago

Great, I look forward to continuing this work! Yes, to share code and structure with the MIMICS implementation sound like a good idea. Maybe a zoom-meeting in the near future would be helpful to coordinate?

wwieder commented 1 month ago

I'm going to close this issue since MIMICS is already included in CLM and now MIMICS+ is being worked on by @mvdebolskiy and @elisacw are working on this now

ESCOMP / CTSM