BiomeEP: expose additional parameters to R

[fabern]

This PR exposes additional BiomeEP parameters to R, making them amenable to calibration.

Status

Some of the inputs parameters were previously ignored [1], while other parameters and initial conditions were hardcoded [2],[3],[4].
For the initial conditions, many of them were derived based on parameter values for convenience. While this may seem convenient and pragmatic, it does not allow to stop and restart a simulation, since the user is not able to restart it by prescribing all the state variables. The convenience function have now been moved to the R part of the code. They are only triggered if the corresponding initial conditions are not explicitly provided. This allows to recover the former behavior by default but enables the users to control the initial conditions with more granularity if desired.
Lastly, a bug was present since init_cohort_nsc was not taken into account.

[1]: in params_species: phenotype,Vmax,alphaBM,leafLS,lAImax,CNleaf0,gamma_L,Vannual, betaON,betaOFF,leaf_size
[2]:

rsofun/src/gpp_biomee.mod.f90

Lines 22 to 35 in 9ddabd9

	real :: beta = 146.0 ! Unit cost of carboxylation (dimensionless)
	! Apply identical temperature ramp parameter for all PFTs
	real :: tau_acclim = 30.0 ! acclimation time scale of photosynthesis (d)
	real :: soilm_thetastar = 0.6 * 250
	real :: soilm_betao = 0.0
	real :: rd_to_vcmax = 0.014 ! Ratio of Rdark to Vcmax25, number from Atkin et al., 2015 for C3 herbaceous
	real :: kc_jmax = 0.41 ! Jmax cost ratio
	! these should be species-specific, temporary solution to put them here
	real :: kphio = 0.05 ! quantum yield efficiency at optimal temperature, phi_0 (Stocker et al., 2020 GMD Eq. 10 and Eq. 20)
	real :: kphio_par_a = 0.0 ! shape parameter of temperature-dependency of quantum yield efficiency
	real :: kphio_par_b = 25.0 ! optimal temperature of quantum yield efficiency (deg C)

[3]: Initial states were not exposed to R in init_cohort: init_cohort_age and the nitrogen pools: init_cohort_bl_n14, init_cohort_br_n14, init_cohort_bsw_n14,init_cohort_bHW_n14,init_cohort_seedC_n14, init_cohort_nsc_n14
[4]: Initial states were not exposed to Rin init_soil: init_fast_soil_N, init_slow_soil_N, init_pmicr_C, init_pmicr_d13C, init_pmicr_N, init_wcl1, init_wcl2, init_wcl3, init_N0_ecosystem

Changes

Changes of this PR are summarized in NEWS.md.
They are breaking changes.

This PR reduces the number of parameters required in the drivers for BiomeEP:

• params_tile$par_mort_under and params_tile$par_mort are removed and have to be replaced by params_species%mortrate_d_u and params_species%mortrate_d_c
• previously ignored parameters [1] were removed from the default driver. If a user still provides them they have to be NA, otherwise an error is thrown.
  • in params_species this concerns: phenotype,Vmax,alphaBM,leafLS,lAImax,CNleaf0,gamma_L,Vannual, betaON,betaOFF,`leaf_size

This PR also allows to optionally provide additional parameters/initial values in the drivers for BiomeEP:

• previously hardcoded parameters [2,3,4] can now optionally be prescribed in the driver. If they are not given, the former default values are automatically set in run_biomee_f_by_site()
  • in params_species this concerns: kphio,beta,rd_to_vcmax,kc_jmax,kphio_par_a,kphio_par_b
  • in params_tile this concerns: tau_acclim
  • in init_cohort this concerns: init_cohort_age and the nitrogen pools: init_cohort_bl_n14, init_cohort_br_n14, init_cohort_bsw_n14,init_cohort_bHW_n14,init_cohort_seedC_n14, init_cohort_nsc_n14.
  • in init_soils this concerns: init_fast_soil_N, init_slow_soil_N, init_pmicr_C,
    init_pmicr_d13C, init_pmicr_N, init_wcl1, init_wcl2, init_wcl3, init_N0_ecosystem

Lastly, this PR fixes the bug that init_cohort_nsc was not taken into account.init_cohort_nsc was made optional. Not providing init_cohort_nsc also recovers former default value.

[fabern]

Strategy for calibration

(as discussed, from memory, feel free to modify)

There are three kinds of parameter with respect to calibration:

• measurable, species-dependent traits, relatively static -> species-specific priors specified from observational data compilations (TRY, BAAD, ...)

  • wood density
  • SLA = 1/LMA
  • allometric coefficients and exponents (alpha, theta, ...)

• Not directly measurable, but emergent properties of the ecosystem configuration, that we believe are implicitly calibratable at ecosystem-level through the used target data

  • P-model parameters (calibrated in Paredes et al., 2026)
  • respiration scaling parameter
  • mortality scaling parameter
  • root-shoot ratio (phiRL)
  • phenology model parameters
  • ...

• model-specific, hardcoded

  • seedling properties
  • internal gap fraction
  • shape parameters of various functions
  • ...

[fabern]

@stineb: could you please have a quick look at the summarized changes from the first message in this PR. Is this streamlining of input parameters in your intent? I will base the calibration on this PR.