Long-term Net Transformation and Quantitative Molecular Mechanisms of Soil Nitrogen During Natural Vegetation Recovery of Abandoned Farmland on the Loess Plateau of China

The availability of nitrogen (N) can alter vegetation species composition and diversity in degraded ecosystems. A comprehensive understanding of the dynamic fate of ammonium (NH-N) and nitrate (NO-N) processing and the underlying mechanisms are still lacking, particularly in arid to semi-arid degraded ecosystems. We compared and quantified the changes in the rates of net ammonification (R), nitrification (R) and total mineralization (R) and the abundance of bacteria, archaea, and microbial genes related to N transformation on the northern Loess Plateau of China across a 40-year chronosequence of farmland undergoing spontaneous restoration. We found that R, R, and R decreased in grassland soils (0-30-y sites) of different ages and exhibited significant increases at the 40-y sites. The capabilities of the soil to deliver NH-N and NO-N were not a limiting factor during the growing season after 40years of vegetation recovery. Soil mineral nitrogen may be not suitable for predicting and assessing the long-term (approximately 40years) restoration success and progress. The abundance of functional N genes showed differences in sensitivity to natural vegetation recovery of abandoned farmland, which likely reflects the fact that the multi-pathways driven by N functional microbial communities had a large influence on the dynamic fate of NH-N and NO-N. Quantitative response relationships between net N transformation rates and microbial genes related to N transformation were established, and these relationships confirmed that different N transformation processes were strongly linked with certain N functional genes, and collaboratively contributed to N transformation as vegetation recovery progressed. Specifically, R was controlled by AOA-amoA, AOB-amoA, and nxrA; R was governed by napA, narG, nirK, nirS, and nosZ; and R was controlled by nifH, apr, AOA-amoA, AOB-amoA, nirS, and nirK.