Fertilization is an effective measure to decrease soil erosion by promoting vegetation restoration, while it affects soil respiration (Rs) greatly through influencing soil properties and plant root growth. A factorial combination of N addition (0, 50 and 100 kg N ha^-1 yr^-1 defined as N0, N50, and N100) and P addition (0, 40 and 80 kg P₂O₅ ha^-1 yr^-1 defined as P0, P40, and P80) experiment was conducted in a farming-withdrawn grassland to investigate the Rs and its components (heterotrophic respiration, Rh; autotrophic respiration, Ra) in responses to different nitrogen (N) and phosphorus (P) addition levels from May 2018 to October 2019 on the semiarid Loess Plateau. Results showed that Rs, Rh and Ra presented single-peak curves both in diurnal and seasonal variations across all N and P addition treatments. The highest Rs, Rh and Ra values during the day occurred at 12:00–14:00 and the highest values during the year appeared in July, which were consistent with soil temperature changes. Rs, Rh and Ra increased significantly under N and/or P addition from May to September. Compared with no fertilization, Rh and Ra significantly increased by 4.2–5.1% and 43.9–46.4%, and by 13.2–15.8% and 23.0–34.2% under N or P addition alone, respectively. N and P combined addition increased Rh and Ra by 25.4–30.7% and 62.0–111.4% than N or P addition alone. Rs and Ra reached the highest values under N100 combined with P80 addition, but Rh reached the highest values under N50 combined with P80 addition treatment. Annual cumulative CO₂ emission increased by 24.4%−26.5% under N addition alone, 17.0–27.4% under P addition alone and 42.2–68.7% under N and P combined addition. Compared with no fertilization, root biomass and soil microbial biomass carbon significantly increased by 15.8–122% and 5.2–92.5% after N and P additions, respectively. However, the positive effect of N and P addition on Rh was weaker than that of Ra, which decreased the contribution of Rh to Rs. Our findings point to the need of rational N and P combination to tradeoff biomass production improvement and soil carbon emission when applying fertilization to promote grassland restoration in semiarid Loess Plateau.

施肥是通过促进植被恢复来减少土壤侵蚀的有效措施，而施肥通过影响土壤特性和植物根系生长而极大地影响土壤呼吸（Rs）。N添加（0、50和100 kg N ha ^-1 yr ^-1定义为N0，N50和N100）和P添加（0、40和80 kg P ₂ O ₅ ha ^-1 yr ^-1）的阶乘组合在退耕的草地上进行了定义为P0，P40和P80的实验，以研究Rs及其组成（异养呼吸，Rh；自养呼吸，Ra）对不同氮（N）和磷（P）添加的响应半干旱黄土高原地区的水位从2018年5月到2019年10月。结果表明，在所有氮和磷添加处理中，Rs，Rh和Ra均表现出昼夜和季节变化的单峰曲线。一天中最高的Rs，Rh和Ra值出现在12：00-14：00，而一年中的最高值出现在7月，这与土壤温度的变化一致。从5月到9月，在添加氮和/或磷的情况下，Rs，Rh和Ra显着增加。与不施肥相比，Rh和Ra分别显着增加了4.2–5.1％和43.9–46.4％，以及13.2–15.8％和23。单独添加N或P时分别为0–34.2％。氮和磷的联合添加比单独添加氮或磷的Rh和Ra分别提高了25.4–30.7％和62.0–111.4％。Rs和Ra在N100结合P80的添加下达到最高值，而Rh在N50结合P80的处理下达到最高值。年度累计CO在单独添加氮的条件下，_2的排放量增加了24.4％-26.5％，仅在添加磷的情况下增加了17.0-27.4％，在单独添加氮和磷的条件下增加了42.2-68.7％。与不施肥相比，添加氮和磷后，根系生物量和土壤微生物量碳分别显着增加了15.8–122％和5.2–92.5％。然而，氮和磷的添加对Rh的积极作用要弱于Ra，从而降低了Rh对Rs的贡献。我们的研究结果表明，在施肥以促进半干旱黄土高原的草地恢复时，需要合理的氮和磷组合权衡生物量生产的改善和土壤碳排放的权衡。