Abstract
Nitrogen-fixing microorganisms (NFMs) are important components of soil N sinks and are influenced by multiple environmental factors. We established a random forest model optimized by the distributed delayed particle swarm optimization (RODDPSO) algorithm to analyze the global NFM data. Soil pH, organic carbon (OC), mean annual precipitation (MAP), altitude, and total phosphorus (TP) are factors with contributions greater than 10% to NFMs. pH, OC, and MAP are the top three factors at the global scale. The tipping points of pH and OC for the NFMs were 7.84 and 2.71%, respectively. The contribution of MAP first increased but then decreased with peak value at 1,265.65 mm. Under the scenario SSP 8.5, 12% of the NFMs increase occur in Africa in 2100; 16% and 36% of the NFMs decrease in North America and Oceania in 2100, respectively. Our work created a global NFMs map and identified the critical tipping points.