Description and methods
These data cover major agricultural N and P flows for 160 crop types plus pastures and meadows across 187 regions for the period 2010-2022. The dataset includes flows of inorganic fertilizer application, manure, atmospheric deposition, nutrient removal, leaching and runoff for both N and P. For N, we additionally estimate biological fixation and gaseous emissions (N2O, NH3 and N2), while weathering is included for P. The dataset was constructed by collating, gap-filling and harmonizing national-level FAOSTAT statistics with geospatial soil, crop and fertilization maps (HWSD, CROPGRIDS, NPKGRIDS), crop-specific fertilization data (HFUBC) and environmental modeling (IPCC Tier 1, IMAGE-DPPS).
Resolution
Spatial: national level
Temporal: annual (2010-2022)
Crop: 160 crops plus pastures/meadows
File descriptions
- Read_ME.txt: This description.
- variable_overview.csv: Descriptions, format and units of variables included in results.
- results_csv.zip: Contains "N_flows.csv" and "P_flows.csv", as well as "N_quality_flags_n.csv" and "P_quality_flags.csv", as standard comma-separated values files. These are the core datasets plus information on data quality for each datapoint. They each contain 391069 rows, organized by country, year, and crop item. Each observation includes the calculated nutrient flows and the associated harvested area. The variables included in the dataset are defined in "variable_overview.csv". Recommended for use in Python, MATLAB, spreadsheet software, etc.
- results_rds.zip: Contains "N_flows.rds" and "P_flows.rds", as well as "N_quality_flags_n.rds" and "P_quality_flags.rds", containing information on the construction of each datapoint as .rds files. They each contain 391069 rows, organized by country, year, and crop item. Each observation includes the calculated nutrient flows and the associated harvested area. The variables included in the dataset are defined in "variable_overview.csv". Recommended for use in R.
- scripts_inputs.zip: Contains the modularized structure of all scripts needed for the creation of the dataset and its technical validation. The folders "inst" and "scripts" contain data and scripts originally created by the authors. "data" is an empty folder structure that will be filled automatically when running the scripts. The folder "inputs" contains several subfolders. In each of these a ReadMe.md file provides instructions on how to attain the input data required for executing the scripts. The scripts must be sourced in order for the modular workflow to function correctly.
How to cite
Hinz, O., Einarsson, R., Mogollón, J. M. & Bruckner, M. Global agricultural nitrogen and phosphorus flows for 160 crops and pastures for 187 regions in 2010–2022. version 1.0 (Zenodo, July 2026).
doi.org
Dependencies
R version 4.4.2 (2024-10-31), cowplot, data.table, dplyr, ggplot2, openxlsx, rnaturalearth, rnaturalearthdata, rnaturalearthhires, scales, sf, terra, tidyr, tidyverse