Potential denitrification activity response to long-term nitrogen fertilization - A global meta-analysis

Abstract

Denitrification is one of the most important biological process by which reactive N is removed in the biosphere, but the responses of potential denitrification activity (PDA) and associated gene abundances to long-term N fertilization are poorly understood. We compiled data from 1041 observations across 62 studies and quantified the effects of long-term N input on PDA, the denitrification N2O/(N2O+N2) product ratio, and abundances of denitrifying communities using meta-analysis. Boosted regression tree (BRT) analysis showed that soil pH was the most important explanatory variables among which explained the change of PDA, followed by soil organic carbon (SOC) and duration of fertilization. Long-term N fertilization significantly increased potential denitrification activity by 75.9% and increased the relative abundances of nirK, nirS, and nosZ gene copies by 60.4%, 77.0%, and 25.1%, respectively. Further, long-term N loading increased the denitrification N2O/(N2O+N2) ratio by 22.1% and nir(K+S)/nosZ ratios by 27.7%, respectively. The effect of long-term N fertilization on potential denitrification activity was positively correlated with SOC (R2 = 0.11; n = 132; P < 0.001), while long-term N loading increased the SOC by 29.0% compared to the unfertilized control. The responses of nirK, nirS, and nosZ abundances to long-term N fertilization were positively correlated with soil pH, whereas N loading decreased the soil pH by 4.6%. Thus, we postulate that long-term N fertilization might have a double effect on the activities and populations of denitrifying communities due to the increased SOC and decreased soil pH.