Effects of 18 years repeated N fertilizer applications on gross N transformation rates in a subtropical rain-fed purple soil

In this study, a 15N tracing experiment was conducted to determine gross N transformation rates in soils with 18 years of repeated N fertilizer applications and to quantify the underlying mechanisms to explain the variability of crop yield in a subtropical rain-fed purple soil. Six N fertilizer treatments were evaluated: unfertilized control (CK), mineral N fertilizer (NPK), pig manure (OM), mineral N fertilizer plus pig manure (OM-NPK), crop residue (RSD) and mineral N fertilizer plus crop residue (RSD-NPK). The results showed that the soil organic carbon (SOC) concentration in OM (average 13.79 g kg-1) and RSD-NPK (average 13.15 g kg-1) were significantly higher than the other treatments (varied from 5.91 to 11.23 g kg -1) after 18 years. Higher maize yields were observed in NPK, OM, OM-NPK, and RSD-NPK, and especially in RSD-NPK and OM-NPK after 18 years, compared with the data measured in the 11th years. Soil gross N mineralization rates (M) were significantly higher in all N fertilizer treatments (varied from 13.00 to 31.17 mg N kg -1 soil d-1) compared to CK (average 5.53 mg N kg -1 soil d-1). After 18 years of repeated N fertilizer applications, M were significantly higher than those in the 11th years for the same N fertilizer treatment, indicating that long-term repeated N fertilizer application, whether mineral N or organic fertilizer, continuously enhances M. M was positively correlated with SOC (p < 0.05) and total N (TN) (p < 0.05), suggesting that soil organic substrates are key factors regulating M. Maize yield was positively correlated with M (p < 0.05), suggesting that available N produced through soil N mineralization played the important role in yield formation. All N fertilizer treatments continuously stimulated gross autotro-phic nitrification rates (ONH4). ONH4 increased with increasing M (p < 0.01) and ammonia-oxidizing bacteria (AOB) abundance (p < 0.05). However, compared to the 11th year, rates of dissimilatory nitrate reduction to ammonium (DNRA) and immobilization of NH4+ (except RSD) and NO3- (except OM-NPK) decreased significantly in the N fertilizer treatments in the 18th year, indicating that N retention capacity is reduced over time and the risk of N losses increased. Considering the best balance among food security, soil C sequestration, and reduction of N leaching, mineral N in combination with straw return (RSD-NPK) can be recommended for regions with similar soils and climates.