Transparent and reproducible crop model calibration using exclusively public data: Improving phenology and yield predictions in APSIMx

Reproducible APSIMx calibration studies using exclusively public data remain limited, yet such transparency is essential for reliable crop modeling. This study demonstrates that publicly available climate, soil, and observational datasets can improve yield and phenology predictions for winter wheat, silage maize, winter oilseed rape, and spring barley across two contrasting German pedoclimatic regions. We developed a stepwise calibration workflow using the croptimizR R package, optimizing phenology parameters before yield-related parameters. Calibration reduced phenology and yield RMSE by 1.9–14.3 days and 10–167 g m -2 across crops. However, improvements at one site occasionally coincided with deterioration at the other, reflecting inherent trade-offs when calibrating across contrasting environments. While public data introduce uncertainties limiting interannual variability capture, the resulting generic parameters provide robust regional estimates for Central European applications, demonstrating the potential of public datasets for transparent crop model calibration. All workflows are publicly available. • Crop calibration in APSIMx using public data sets yielded satisfactory results. • We present crop parameters and a calibration workflow for four key German crops. • Crop parameters were robust in contrasting pedoclimatic settings. • Data harmonization and limited model transparency can impede calibration access. • Critical assessment of public data uncertainties is essential.