SADORE PEARL MILLET: EXTRACTION AND MODEL TRANSFER
26 October 2026

Sources
Akponikpe, Gerard, Michels & Bielders (2010), European Journal of Agronomy32:144-154. DOI10.1016/j.eja.2009.09.005. Primary extraction: Tables1-2, pp146-148.
https://oar.icrisat.org/6294/1/EJA_32_2_144-154_2010.pdf
Akponikpe, Minet, Gerard, Defourny & Bielders (2011), Agricultural and Forest Meteorology151:215-227. DOI10.1016/j.agrformet.2010.10.007. Cross-check only; its farmer-density scenario is distinct.
https://oar.icrisat.org/10557/1/akponikp2011.pdf
The2008 thesis was located at https://dial.uclouvain.be/pr/boreal/object/boreal%3A19624/datastream/PDF_01/view but full text could not be retrieved in this session. No additional thesis coefficients are claimed.

Transfer approach
Use site coordinates, experimental sowing dates and final stand densities directly as management metadata. Thinning dates still need confirmation. Retain original soil layer boundaries. LL15, DUL and SAT are candidate WP, FC and saturation inputs, subject to different model definitions. These are experimental-field properties, not a soil map for all Sadore. DUL was calibrated; saturation calculated.
For the present homogeneous WOFOST balance, aggregate soil properties only across the selected root depth. For AquaCrop use explicit layers. Neither model's required conductivity can be inferred uniquely from APSIM SWCON. Root-zone PAW reported over1.5m is not sufficient evidence of cultivar maximum rooting depth.

AquaCrop
Fit emergence, flowering, senescence, maturity and yield-formation timing under its own GDD calculation. APSIM cardinal temperatures10/33/47C define a triangular response and do not translate mechanically into AquaCrop Tbase/Tupp. Fit green-canopy CC0/CCx/CGC/CDC, rooting and stress functions, Kcb, normalized WP and reference harvest index from observed well-watered growth then stressed treatments. APSIM RUE is not WP; grain number and grain growth are not HI0. Photoperiod-sensitive sowing-date effects need explicit evaluation with the four2005 sowings.
FAO formulation: https://www.fao.org/aquacrop/overview/calculation-scheme/

WOFOST
A generic pearl-millet file exists in the official WOFOST7.2 parameter collection (millet.yaml; keyword Pennisetum typhoides). It is a starting template, not a CIVT calibration. Calibrate TSUMEM/TSUM1/TSUM2 jointly with DTSMTB and photoperiod response, then leaf area, assimilation, partitioning, senescence, roots and water stress. The reported flowering-maturity380C days is a target in the APSIM thermal-time convention, not an immediately valid TSUM2. WOFOST assimilation cannot be set by copying APSIM RUE.
https://github.com/ajwdewit/WOFOST_crop_parameters/blob/wofost72/millet.yaml
https://pcse.readthedocs.io/en/develop/code.html

Implementation readiness
Keep extraction separate from runnable cases. Missing initial water, daily weather, actual planting depth, root parameters and model-specific physiology prevent a source-only complete calibrated configuration. Prefer station weather for reproducing1994/1995; GEE weather is an alternative forcing requiring separate evaluation. Current platform engines omit explicit fertility dynamics, central to the source experiments. A successful water-limited run will not reproduce N/P response or validate millet yield.

Files
reported_parameters.csv: numeric coefficients, units, location in paper and transfer notes.
soil_profile.csv: nine experimental soil layers; volumetric water contents are fractions.
source_case.json: site, cultivar, separate experiments, missing inputs and conflicts.