piersond / mimics_stode Goto Github PK

Find it in RCrk_MIMICS, revamp it for here

Current parameterization is off now that we've removed fW influence on tau.

Also try parameterizing with simpler MAT dependent multiplier on V terms (i.e., equation = mx, instead of previous mx + b)

Currently these two versions live in separate branches on the repo. Discuss with @wwieder if its a) better for simplicity to keep these separate, or b) perhaps less work long-term if we combine the functionality into one version, where the N cycle can be turned on/off?

Check if R version MIMSOC values change if stode is run to 1e7 instead of 1e6.

Ensure that Tpars function works with CN-stode code.

The cumulative respiration rate changes by 10-20% depending on whether the incubation is initialized with 100 vs. 1000 units C. Appears to be less sensitive at higher values.

Need to identify the underlying cause. Perhaps the LITm pools is running out over 105 days at low I?

Best to avoid redundant code. There can be only one RXEQ.

Currently we're importing the MIMICS_STODE() function from MIMICS_steady_state_pools.R, which itself imports parameters and runs a simulation. Suggest we create a clean script for MIMICS_STODE.R to prevent errors and confusion.

Folder structure for this setup? Should we add a stode folder? Add a run folder? Both? Should we remove all scripts from the parent folder for better organization?

To set plausible range for MIMICS pools. Useful for parameter selection from MC output and for QC plots.

Functionality yet to be tested.

Try for a proper Metrop-Hastings method with convergence test.

• Check plan settings.
• Try removing unnecessary code, print statements, etc.

Worthwhile to use LiDET results to make ensure the litter decomp rates are in the right ballpark?

Best to avoid redundant code. There can be only one.

Loop d from 1 to nrow(daily input)

I might change the name of the LiDET_fMET to something like fixed_fMET to clarify what this logical flag is really doing.

https://github.com/piersond/MIMICS_STODE/blob/64bf9cc4ad51385f4b203d8f19c9864fc313cc93/calc_Tpars.R#L15C61-L15C61

Initial test shows 10-20% shift, but estimates stay on trendline. Perhaps MAT dependency results are the same either way?