Long-term phosphorus addition alleviates CO and NO emissions via altering soil microbial functions in secondary rather primary tropical forests.

Tropical forests, where the soils are nitrogen (N) rich but phosphorus (P) poor, have a disproportionate influence on global carbon (C) and N cycling. While N deposition substantially alters soil C and N retention in tropical forests, whether P input can alleviate these N-induced effects by regulating soil microbial functions remains unclear. We investigated soil microbial taxonomy and functional traits in response to 10-year independent and interactive effects of N and P additions in a primary and a secondary tropical forest in Hainan Island. In the primary forest, N addition boosted oligotrophic bacteria and phosphatase and enriched genes responsible for C-, P-mineralization, nitrification and denitrification, suggesting aggravated P limitation while N excess. This might stimulate P excavation via organic matter mineralization, and enhance N losses, thereby increasing soil CO and NO emissions by 86% and 110%, respectively. Phosphorus and NP additions elevated C-mining enzymes activity mainly due to intensified C limitation, causing 82% increase in CO emission. In secondary forest, P and NP additions reduced phosphatase activity, enriched fungal copiotrophs and increased microbial biomass, suggesting removal of nutrient deficiencies and stimulation of fungal growth. Meanwhile, soil CO emission decreased by 25% and NO emission declined by 52-82% due to alleviated P acquisition from organic matter decomposition and increased microbial C and N immobilization. Overall, N addition accelerates most microbial processes for C and N release in tropical forests. Long-term P addition increases C and N retention via reducing soil CO and NO emissions in the secondary but not primary forest because of strong C limitation to microbial N immobilization. Further, the seasonal and annual variations in CO and NO emissions should be considered in future studies to test the generalization of these findings and predict and model dynamics in greenhouse gas emissions and C and N cycling.