MAPS: A new model using data fusion to enhance the accuracy of high-resolution mapping for livestock production systems

To meet growing demand for animal-based foods, livestock production has intensified to maximize output with limited resources and space. This has increased the spatial heterogeneity of livestock distribution, which has in turn caused severe nutrient loss and increased risk of antimicrobial resistance, zoonotic disease, and human exposure to disease and pollution. There is an urgent need for spatially explicit impact assessments, but current methods lack the resolution needed to accurately map fine-scale livestock distribution. Here we developed a mapping agricultural production systems (MAPS) model by fusing enterprise registration information (ERI), which can directly represent agricultural activities, with other currently available data to generate high-resolution mapping. Using an example of pig production in China, the global leading pig producer, we show that MAPS improves the accuracy of location/size estimates of livestock production systems by 12%-84%, illustrating a 44% underestimation of pig numbers in dense farming areas (>1,000 pigs/ km(2)) by existing methods. MAPS also reveals a spatial transfer of pig production from rural to peri-urban areas, implying more decoupled pig-crop systems in China. MAPS enables spatially explicit impact assessments to support sustainable planning of intensive livestock production systems that can alleviate nutrient loss and health risks.