Global Crop-Specific Fertilization Dataset from 1961-2019

TL;DR

This paper develops a high-resolution, long-term dataset of fertilizer application rates for major crops worldwide from 1961 to 2019 using machine learning models, aiding research in food security and climate change.

Contribution

It introduces a novel dataset of crop-specific fertilizer application rates at 5-arcmin resolution, generated through machine learning predictions validated against existing data.

Findings

Created a comprehensive 1961-2019 fertilizer dataset for 13 crop groups.

Validated predictions with existing databases and driver analysis.

Provides a valuable resource for environmental and socioeconomic studies.

Abstract

As global fertilizer application rates increase, high-quality datasets are paramount for comprehensive analyses to support informed decision-making and policy formulation in crucial areas such as food security or climate change. This study aims to fill existing data gaps by employing two machine learning models, eXtreme Gradient Boosting and HistGradientBoosting algorithms to produce precise country-level predictions of nitrogen ($N$), phosphorus pentoxide ($P_2O_5$), and potassium oxide ($K_2O$) application rates. Subsequently, we created a comprehensive dataset of 5-arcmin resolution maps depicting the application rates of each fertilizer for 13 major crop groups from 1961 to 2019. The predictions were validated by both comparing with existing databases and by assessing the drivers of fertilizer application rates using the model's SHapley Additive exPlanations. This extensive dataset is poised to be a valuable resource for assessing fertilization trends, identifying the socioeconomic, agricultural, and environmental drivers of fertilizer application rates, and serving as an input for various applications, including environmental modeling, causal analysis, fertilizer price predictions, and forecasting.