Climate variability and ballooning populations are putting unprecedented pressure on agricultural croplands and their water use, which are vital for ensuring global food and water security in the 21st century. In this context, there is a need to produce consistent and accurate global food security support-analysis data (GFSAD) at fine spatial resolution that are generated routinely (e.g., every year). Hence, the overarching goal of this project was to produce GFSAD models, maps, and monitoring tools using machine learning algorithms (MLAs) on cloud computing platforms with the ability to handle multi-sensor remotely sensed big data, leading to a comprehensive set of cropland, crop water use, crop productivity, and crop water productivity products that ensure food security for all peopleglobally . The outcome resulted in the first global 30-m cropland extent product (https://croplands.org/app/map) derived using Landsat 8 16-day time-series data for the nominal year 2015. Overall accuracies in about 100 segments of the world were consistently higher than 90%, with producer’s accuracies typically higher than 85% and user’s accuracies typically higher than 80%. The study estimated global net cropland area (GNCA) of 1.872 billion hectares (about 12.6% of the Earth’s land area). Four countries, India, United States, China, and Russia, together had 35% of GNCA. The methods described in this study produced 30-m derived cropland areas for each country of the world as well as for sub-national regions such as provinces, districts, and counties and were compared with the national agricultural statistics compiled by the United Nations Food and Agriculture Organization (FAO) and others.
Global cropland extent map at 30-m resolution derived using Landsat 8 16-day time-series data for the nominal year 2015 (https://croplands.org/app/map).