Objectives This study explored soil greenhouse gas emissions and mechanism of carbon and nitrogen transformation under the co-incorporation of wheat straw and leguminous green manure residues in Qinghai Plateau of China, which can provide a theoretical basis for the scientific utilization of straw and green manure in dryland agricultural ecosystems.

Methods An incubation experiment was carried out using wheat straw and hairy vetch (Vicia villosa Roth L.). There were four treatments: no straw addition control (CK), hairy vetch alone (VS), wheat straw alone (WS), and wheat straw and hairy vetch together (VWS). Greenhouse gas emissions, soil active carbon (C) and nitrogen (N) components, soil enzyme activities and the abundances of bacteria, archaea, fungi, amoA, nirK and narG were determined.

Results Compared with VS, VWS treatment decreased of cumulative emissions of CO₂ and N₂O by 24.8% and 74.6%, increased cumulative absorption of CH₄ by 9.1%, and significantly reduced global warming potential (GWP) by 76.1% (P<0.05). Compared with WS treatment, VWS increased cumulative CO₂ emissions by 33.7%, but decreased cumulative CH₄ absorption by 12.0%, and the cumulative N₂O emissions by 43.1%, GWP by 49.4%. VWS treatment significantly increased soil pH compared with CK and VS. Compared with VS treatment, VWS increased the contents of the dissolved organic C (DOC) and microbial biomass C (SMBC) by 21.6% and 4.9%, but decreased the mineral N (Nmin), dissolved organic N (DON) and microbial biomass N (SMBN) by 77.3%, 59.5%, and 6.3%. Compared with WS treatment, VWS increased Nmin, DOC, DON, SMBC and SMBN by 64.0%, 22.5%, 56.5%, 23.2%, and 27.8%, respectively. Plant residues significantly increased enzyme activities, and VWS treatment recorded higher α-glucosidase (AG) and β-glucosidase (BG) activities, and lower nitrite reductase (NIR) activity than the other treatments. The abundances of fungi and archaea in VWS treatment were 83.8% and 69.8% higher than those in WS treatment, but 62.6% and 20.3% lower than those in VS treatment, respectively. The bacterial abundance of VWS treatment was 33.4% lower than that of VS treatment. In VWS, the abundances of AOB amoA, nirK and narG genes were decreased by 56.6%, 41.4%, and 16.3%, than those in VS treatment, and by 30.3%, 25.9%, and 12.0% compared with WS treatment, respectively. The results of correlation analysis showed that CO₂ and CH₄ emissions were positively correlated with soil organic C, total N, DOC, SMBC, and SMBN, AG, NAG and NR activities, and the abundances fungi, bacteria, nirK and narG. N₂O emission was positively correlated with soil total N, DON, SMBC and SMBN, the AOB amoA and nirK abundances, and was negatively correlated with soil pH and BG activity. Partial Least Squares Path Model analysis (PLS-PM) showed that AOB amoA was the main functional bacteria regulating soil N₂O emission.

Conclusions Both the single and co-application of wheat straw and hairy vetch increases soil CO₂ emissions and reduces CH₄ absorption. Co-incorporating of wheat straw and hairy vetch is more effective in reducing N₂O emission and GWP through the reduced abundance of AOB, so is an effective measure for greenhouse gas emission reduction in dryland soil of Qinghai Plateau in China.