MCI GPP: ensembling a global model- and climate-independent gross primary productivity for 2001–2023

The model- and climate-independent (MCI) gross primary productivity (GPP) dataset provides monthly global GPP estimates at 0.05° resolution (2001-2023), developed through a machine learning-based fusion of 12 model simulations. It integrates MOD17 with CFE, EC-LUE, and MPI-Jena models under production efficiency model (PEM) and two leaf model (TLM) frameworks, driven by GMAO MERRA2 and ECMWF ERA5 meteorological datasets. Leveraging FLUXNET2015 observations as ground truth, the Random Forest algorithm bridges model outputs with field-measured GPP to generate a robust product. Advanced spatiotemporal gap-filling techniques ensure continuous coverage while uncertainty-based filtering removes low-quality pixels, yielding a model- and climate-independent dataset with enhanced spatiotemporal consistency for carbon cycle analysis and ecosystem monitoring.