Global dataset on nutrient removal with crop residues
Data files
Dec 16, 2025 version files 3.94 MB
-
crop_residue_coefficients.csv
21.71 KB
-
Nutrient_removal_with_crop_residue.csv
3.91 MB
-
README.md
5.38 KB
Abstract
This dataset provides global, annual estimates of how much nitrogen, phosphorus, and potassium are removed from cropland through the off-field removal of crop residues. It covers 210 countries from 1961 to 2023, based on information from official statistics, expert knowledge, and published sources. The dataset has been integrated into the FAO Cropland Nutrient Balance Database as of the 2025 release, enabling more complete assessments of nutrient balances and nutrient use efficiency.
1. Background and methods
Crop residues are the parts of the plant that remain in the field after harvest. Previous versions of the FAO - Cropland Nutrient Balance Database (FAO-CNB) only provided estimates of nutrient outputs through the harvested organs, but not for nutrient outputs associated with crop residue removal. When residues are removed from the field for other uses, such as off-field livestock feeding, bedding, fuel or mushroom cultivation, considerable amounts of nitrogen (N), phosphorus (P), and potassium (K) are removed out of the system. Nutrients may also be lost due to in-field residue burning. In contrast, nutrients in crop residues are recycled in the agroecosystem when they are used for grazing within the field or left to decompose above or below ground.
The file Nutrient_removal_with_crop_residue.csv contains estimates of nutrient removal from crop residue removal for the 158 crops, 210 countries, and years 1961–2023 included in the FAO-CNB. To estimate the share of residues that are removed from the field, we compiled information from official statistics (5 countries), surveyed 402 experts in agronomy of 102 countries, and retrieved data from 285 scientific and gray literature publications. We combined these sources using a tier-based approach that prioritized official and crop-specific data of higher reliability. We collected information for four periods (around 1970, 1990, 2010, and the most recent year) and reached full temporal coverage by applying linear interpolation between two time period estimates to obtain annual estimates. Together, the crop-country combinations for which we gathered data covered approximately 75 % of global cropland area. Missing information for minor crops was filled with average data from similar crops in the same country and year. For countries without available data (roughly 6 % of cropland area), we estimated the fraction of residue removal through spatial extrapolation using machine-learning models that related crop residue management to climate, crop type, and socio-economic characteristics.
We calculated annual N, P, and K outputs from residue removal using yields reported in the FAOSTAT Crop and Livestock Production Database and crop-specific global-average harvest indexes and nutrient concentrations in crop residues from the literature (see crop_residue_coefficients.csv). Values were computed by crop and averaged at national levels by considering crop harvested areas. The corresponding equations are provided in Section 3.2.
2. Data use statement.
These data are released to support their integration into the FAO-CNB. The authors are preparing a companion analytical paper that will provide an in-depth analysis of global patterns of crop residue nutrient removal and their implications for nutrient balances and use efficiency. Users are welcome to contact the corresponding author before potentially undertaking derivative analyses that may substantially overlap with the planned publication.
3. Database description
The database follows a long-format structure consistent with other FAOSTAT databases, with following column names:
- Area Code: Country code used in FAOSTAT databases
- Area Code (M49): United Nations M49 country codes
- Area: Country name
- Year: Harvest year
- Element: Reported variable (see 3.1 Variable definition below)
- Value: Value for the reported variable
- Unit: Unit of the reported variable
- Flag: Code indicating source of reported value (see 3.3 Flag description below).
3.1 Variable definition
- Crop residue removal: Percentage of above-ground crop residue that is removed outside of the field for other uses. Note that residue burning or grazing within the field are not included in the removal estimates.
- N, P, and K removed with crop residue: Amount of elemental nutrient that is removed from the field with the crop residue removed from the field.
3.2 Equations
- TRijk (kg DM ha-1) = Yijk (kg ha-1) × (DMHOj (%) / 100) × (1/HIj – 1)
- RRijk (kg DM ha⁻¹) = TRijk (kg DM ha⁻¹) × (RRijk (%) / 100)
- N RRijk (kg ha⁻¹) = RRijk (kg DM ha⁻¹) × (N TRj (%) / 100)
- P RRijk (kg ha⁻¹) = RRijk (kg DM ha⁻¹) × (P TRj (%) / 100)
- K RRijk (kg ha⁻¹) = RRijk (kg DM ha⁻¹) × (K TRj (%) / 100)
TR is total residue, Y is crop yield, DMHO is the dry-matter content of the harvested organ, HI is the harvest index, RR is the crop residue removed from the field, and N, P, and K represent nitrogen, phosphorus, and potassium, respectively. The subscripts i, j, and k denote year, crop, and country, respectively. N TRj (%) is the nitrogen concentration in crop residue for crop j, expressed on an oven-dry mass basis. N RR (kg ha⁻¹) is the amount of nitrogen removed from the field through crop residue removal per hectare.
3.3 Flags:
- A: official figure (Tier 1)
- E (sl): value estimated with survey and literature review data (Tier 2)
- E (ml): value estimated with machine learning extrapolation (Tier 3)
- E (ti): value estimated by temporal interpolation
