ORCHIDEE-TRUNK OPTIMIZATION
ROUND 7

Select: config PFT2 PFT3 PFT3b PFT4 PFT5 PFT5b PFT6 PFT7 PFT7b PFT8 PFT9

Modifications for setups 3b, 5b, 7b:

1. BA variable is added to control variables with ranges of variation:

   Variable	PFT2-6	PFT7	PFT8-9
   BA [m2/ha]	25 - 40	20 - 30	10 - 25



2. ALPHA_SELF_THINNING is added to optimized parameters with range of variation [500 - 4000].


3. LAI_MEAN_GS control variable data are used from the full length of time-series (all other variables data are used from the 10Y prior to the latest BA maximum).

General setup description:

1. Model version: ORCHIDEE_trunk revision 8237.


2. Control variables with prescribed ranges of variation:

   Variable	PFT2	PFT3	PFT4-6	PFT7-9
   GPP [gC/m2/y]	2000 - 4000	2000 - 4000	800 - 2200	400 - 1000
   LAI_MEAN_GS	4 - 8	3 - 8	2 - 6	1.5 - 5
   HEIGHT [m]	20 - 50	20 - 50	10 - 30	5 - 20
   ng_ratio	0.25 - 0.37	0.4 - 0.5	0.45 - 0.55	0.4 - 0.5
   lr_ratio	0.5 - 2
   bm_ratio	0 - 0.1
   ge_ratio	1 - 3

   ng_ratio = NPP / GPP
   lr_ratio = LEAF_M_c / ROOT_M_c
   bm_ratio = (LABILE_M_c + RESERVE_M_c) / (SAP_M_BE_c + SAP_M_AB_c + HEART_M_BE_c + HEART_M_AB_c)
   ge_ratio = GPP / EVAP


3. Auxiliary variable: BA [m2/ha] (basal area). The BA variable is used to select the slice of data to be used in optimization. It corresponds to the 10Y of data prior to the latest maximum in the BA time-series.


4. Visualized variable: SUGAR_LOAD.


5. Optimized parameters with default values and ranges of variation
   (vmin = 0.5*vdef and vmax = 2*vdef for all parameters except K_LATOSA_FRAC, TREE_FF; for VMAX_UPTAKE_XXX vmin=0.25*vdef):

   Parameter	PFT2	PFT3	PFT4	PFT5	PFT6	PFT7	PFT8	PFT9
   K_LATOSA_MIN	7500 [3750 - 15000]	6000 [3000 - 12000]	5600 [2800 - 11200]	2989 [1494.5 - 5978]	15310 [7655 - 30620]	9120 [4560 - 18240]	15310 [7655 - 30620]	15310 [7655 - 30620]
   K_LATOSA_FRAC	2 [1 - 3]	2 [1 - 3]	1.964 [1 - 3]	2.007 [1 - 3]	1.9595 [1 - 3]	1.9737 [1 - 3]	1.9595 [1 - 3]	1.9595 [1 - 3]
   K_BELOWGROUND	8.944E-8 [4.472E-8 - 1.7888E-7]	6.6E-8 [3.3E-8 - 1.32E-7]	1.291E-8 [6.455E-9 - 2.582E-8]	4.04E-8 [2.02E-8 - 8.08E-8]	3.287E-7 [1.6435E-7 - 6.574E-7]	5.254E-9 [2.627E-9 - 1.0508E-8]	3.287E-7 [1.6435E-7 - 6.574E-7]	3.287E-8 [1.6435E-8 - 6.574E-8]
   VMAX_UPTAKE_NH4	13.6 [3.4 - 27.2]	12 [3 - 24]	9.241 [2.31025 - 18.482]	18 [4.5 - 36]	11.81 [2.9525 - 23.62]	8.217 [2.05425 - 16.434]	11.15 [2.7875 - 22.3]	11.15 [2.7875 - 22.3]
   VMAX_UPTAKE_NO3	13.13 [3.2825 - 26.26]	12 [3 - 24]	13.03 [3.2575 - 26.06]	5.996 [1.499 - 11.992]	17.31 [4.3275 - 34.62]	11.92 [2.98 - 23.84]	10.6 [2.65 - 21.2]	10.6 [2.65 - 21.2]
   COEFF_MAINT_INIT	0.04531 [0.022655 - 0.09062]	0.04 [0.02 - 0.08]	0.07758 [0.03879 - 0.15516]	0.02938 [0.01469 - 0.05876]	0.361 [0.1805 - 0.722]	0.0318 [0.0159 - 0.0636]	0.1072 [0.0536 - 0.2144]	0.1072 [0.0536 - 0.2144]
   SLAINIT	0.03963 [0.019815 - 0.07926]	0.044 [0.022 - 0.088]	0.01365 [0.006825 - 0.0273]	0.02703 [0.013515 - 0.05406]	0.05749 [0.028745 - 0.11498]	0.01886 [0.00943 - 0.03772]	0.04164 [0.02082 - 0.08328]	0.04164 [0.02082 - 0.08328]
   NUE_OPT	14.08 [7.04 - 28.16]	30 [15 - 60]	20.79 [10.395 - 41.58]	19.95 [9.975 - 39.9]	56.61 [28.305 - 113.22]	20.59 [10.295 - 41.18]	27.75 [13.875 - 55.5]	27.75 [13.875 - 55.5]
   TREE_FF	0.4681 [0.23405 - 1]	0.4234 [0.2117 - 1]	0.6 [0.3 - 1]	0.6 [0.3 - 1]	0.703 [0.3515 - 1]	0.7501 [0.37505 - 1]	0.7 [0.35 - 1]	0.8187 [0.40935 - 1]
   LONGEVITY_SAP	7300 [3650 - 14600]	3766 [1883 - 7532]	3600 [1800 - 7200]	6643 [3321.5 - 13286]	2140 [1070 - 4280]	4046 [2023 - 8092]	2140 [1070 - 4280]	2140 [1070 - 4280]

   K_LATOSA_MAX = K_LATOSA_MIN * K_LATOSA_FRAC
   


6. Observation error (R) is set as RMSD for simulations time-series vs the mean between min/max values of range of variation.


7. Parameter error (B) is set as 15% of the parameter range of variation.


8. Cost-function is calculated with the standard formula, but the obs-sim difference is calculated as deviation from the prescribed range of variation and set to 0 for the time-series part falling inside the range.


9. Five representative pixels are selected over the globe for each PFT, simulation at each site is done in two modes (unmanaged/managed).


10. Atmospheric forcing: CRUJRA v2.4 global dataset with the single pixel data extraction corresponding to each site.


11. The spinup over sites is done in 1 step, recycling 1901-1910 forcing for 340Y with ATM_CO2 = 350 ppm, using flags:
    FOREST_MANAGED_FORCED = 1
    RECRUITMENT_PFT__XXXXX = y


12. Optimization simulations are done in two forest managements modes:
    FOREST_MANAGED_FORCED = 1, RECRUITMENT_PFT__XXXXX = y
    FOREST_MANAGED_FORCED = 2, RECRUITMENT_PFT__XXXXX = n


13. Optimization simulations are restarted from the spinup with a clear cut event (FORCED_CLEAR_CUT = y is added to run.def for the first simulation year).


14. Optimization simulations are done for 150Y (1871-2020), during the first 30Y recycling 1901-1910 forcing.


15. Minimization method: Genetic Algorithm (population = 30, number of iterations = 15).


16. Computing platform: LSCE obelix with 15 processors.