ORCHIDEE TRUNK albedo optimization
Round C (15/10/2025)
Step 1 is skipped, parameters from Round B are used

Select range and step: nirvisstep2-bgstep2b-bgstep3-snowstep3b-snowstep3c-snow	Final results: run.def run2.def run3.def alb_bg_optv4c_smoothed.nc alb_bg_optv4c_nonsmoothed.nc alb_bg_optv4c_nonsmoothedvis.nc
Post-check: https://orchidas.lsce.ipsl.fr/mapper?set=rev.9014&mode=albedo

Model: ORCHIDEE TRUNK rev9014
Configuration: default FG_CRUJRA_HIST, but TEST_ALB_RT_VEGMAX = y, optimized LEAF_SSA/LEAF_PSD from Round B
Resolution: 0.5°
Period: 2016-2020, monthly
Observations: MODIS (quality flag <= 5, snow fraction == 0)
Optimized parameters: background albedo at single pixel level
Range of variation: [-0.1, +0.1] from initial value
Method: variational L-BFGS-B with approximate leaf-snow-background albedo superposition
Gradient: analytical derivative
Data selection: (frac_snow == 0) & (tot_bare_soil > 0.5) & (cont_frac > 0.99) & (inland_water < 0.1)
Post-processing: posterior background smoothing with 3x3-pixel running average window

ALB_NIR RESULTS