Shifts in priming partly explain impacts of long-term nitrogen input in different chemical forms on soil organic carbon storage.

Input of labile organic carbon can enhance decomposition of extant soil organic carbon (SOC) through priming. We hypothesized that long-term nitrogen (N) input in different chemical forms alters SOC pools by altering priming effects associated with N-mediated changes in plants and soil microbes. The hypothesis was tested by integrating field experimental data of plants, soil microbes and two incubation experiments with soils that had experienced 10years of N enrichment with three chemical forms (ammonium, nitrate and both ammonium and nitrate) in an alpine meadow on the Tibetan Plateau. Incubations with glucose-C-13 addition at three rates were used to quantify effects of exogenous organic carbon input on the priming of SOC. Incubations with microbial inocula extracted from soils that had experienced different long-term N treatments were conducted to detect effects of N-mediated changes in soil microbes on priming effects. We found strong evidence and a mechanistic explanation for alteration of SOC pools following 10years of N enrichment with different chemical forms. We detected significant negative priming effects both in soils collected from ammonium-addition plots and in sterilized soils inoculated with soil microbes extracted from ammonium-addition plots. In contrast, significant positive priming effects were found both in soils collected from nitrate-addition plots and in sterilized soils inoculated with soil microbes extracted from nitrate-addition plots. Meanwhile, the abundance and richness of graminoids were higher and the abundance of soil microbes was lower in ammonium-addition than in nitrate-addition plots. Our findings provide evidence that shifts toward higher graminoid abundance and changes in soil microbial abundance mediated by N chemical forms are key drivers for priming effects and SOC pool changes, thereby linking human interference with the N cycle to climate change.