Table 5.

LUCC driver and its representation in (data = GPASOIL-v0) and (data = GPASOIL-v1).

Driver name	LUCC (land use and land cover change)
Description of its effect on the soil P	Decrease/increase of agricultural soil P resulting from land conversion within the same grid-cell (see Eq. 18).
Corresponding variables in our approach	Fractions of cropland ($frac\left(y,crop,g\right)$) and of grassland ($frac\left(y,grass,g\right)$) for any year y and grid-cell g as well as transitions between the 4 land-use categories considered in our study (e.g. ${\mathrm{\Delta }}_i\left(y,j,g\right)$ is the conversion from j to i with both i and j in {crop, grass, nonagri, urban}).
Dataset ref used in (data = GPASOIL-v0)	Fractions and transitions were provided by the Land Use Harmonization dataset75
Issues related to the use of this dataset in GPASOIL-v0
Representation in (data = GPASOIL-v1)	Fractions and transitions were provided by Chini et al.20, the update of the Land Use Harmonization dataset (the so-called LUH2-GCB2019 dataset). We made a corresponding between our land-use categories and Chini et al.20 categories. See Eqs. 59–62
Characteristics of the spatially explicit dataset used in (data = GPASOIL-v1)	Chini et al.20 provided LUH2-GCB2019 (10.3334/ORNLDAAC/185176), an update of the Land Use Harmonization dataset used in GPASOIL-v0. The LUH2-GCB2019 dataset provided fractions and transitions between crops (C3/C4, annual/perennial, C3-nitrogen fixing), grassland (managed pasture, rangeland), forest and urban land, at 0.25°x0.25° spatial resolution for any years between 850 and 2018.