Response of soil microbial communities and rice yield to nitrogen reduction with green manure application in karst paddy areas

Fertilizer application practices are one of the major challenges facing agroecology. The agrobenefits of combined application of green manure and chemical fertilizers, and the potential of green manure to replace chemical fertilizers are now well documented. However, little is known about the impact of fertilization practices on microbial communities and tice yield. In this study, the diversity of bacterial and fungal communities, symbiotic networks and their relationship with soil function were analyzed in five fertilization treatments (N: 100% nitrogen fertilizer alone; M: green manure alone; MN60: green manure couple with 60% nitrogen fertilizer, MN80: green manure couple with 80% nitrogen fertilizer; and MN100: green manure couple with 100% nitrogen fertilizer). First, early rice yield was significantly higher by 12.6% in MN100 treatment in 2021 compared with N. Secondly, soil bacterial diversity showed an increasing trend with increasing N fertilizer application after green manure input, however, the opposite was true for fungal diversity. Microbial interaction analysis showed that different fertilizer applications changed soil microbial network complexity and fertilizer-induced changes in soil microbial interactions were closely related to soil environmental changes. Random forest models further predicted the importance of soil environment, microorganisms and rice yield. Overall, nitrogen fertilizer green manure altered rice yield due to its effects on soil environment and microbial communities. In the case of combined green manure and N fertilizer application, bacteria and fungi showed different responses to fertilization method, and the full amount of N fertilizer in combination with green manure reduced the complexity of soil microbial network. In contrast, for more ecologically sensitive karst areas, we recommend fertilization practices with reduced N by 20-40% for rice production. Graphical Abstract.