Abstract: Nitrogen fertilization is the primary cause of nitrous oxide (N₂O) emission from cropland soils. Straw returning may further accelerate the emission. However, the combined effect of these two factors on N₂O emission from paddy soils and the microbial mechanisms remain unclear. We conducted a pot experiment on typical paddy soil with three treatments: rice straw + no nitrogen application (SF₀), rice straw + reduced nitrogen application (SF₇₀, 70% of conventional nitrogen fertilizer application), and rice straw + conventional nitrogen application (SF₁₀₀), to analyze soil N₂O flux during the growth period of rice and the functional genes (amoA, nirS, nirK, and nosZ). The yield of rice was measured after harvest. The results showed that: 1) The cumulative N₂O emissions during the rice growth period were 67.77, 92.91, and 130.43 mg·m^-2 for SF₀, SF₇₀, and SF₁₀₀ treatments, respectively. 2) The effect of different treatments on cumulative N₂O emissions varied at different growth stages of rice. Compared to SF₀, the SF₇₀ treatment had no significant effect on cumulative N₂O emission during tillering and heading stages but significantly increased the emission during jointing and mature stages; the SF₁₀₀ treatment significantly increased the cumulative N₂O emission throughout the whole growth period. 3) At the tillering and heading stages, the copy abundance of amoA gene for ammonia-oxidizing bacteria, nirK, and nosZ genes significantly increased after nitrogen application, and the increase was relatively consistent between SF₇₀ and SF₁₀₀ in most cases. At the jointing and mature stages, the copy abundance of amoA gene for ammonia-oxidizing bacteria in the SF₇₀ and SF₁₀₀ treatments increased, while that of amoA gene for ammonia-oxidizing archaea significantly increased with the amount of nitrogen application. There was no significant difference in the copy numbers of nirK gene among different treatments at these two growth stages. The abundance of nirS gene in the SF₁₀₀ treatment was significantly higher than that in the SF₀ treatment. The abundance of nosZ gene in the SF₇₀ and SF₁₀₀ treatments were significantly higher than in the SF₀ treatment during the jointing stage. 4) Compared to SF₀, rice yields under SF₇₀ and SF₁₀₀ treatments increased by 49.2% and 57.8%, respectively, without significant difference between them. Our results indicated that reduced nitrogen application under straw return conditions has a smaller impact on rice yield compared to conventional nitrogen application but can significantly reduce cumulative N₂O emissions, with ammonia-oxidizing archaea being the main contributors to such changes.

Key words: nitrous oxide, functional gene, nitrogen fertilizer reduction, straw return, paddy soil