Main controls on the denitrification rates during cropland revegetation in the southwest China Karst Critical Zone Observatory

The revegetation of karst ecosystems in subtropical southwest China that became severely degraded because of intensive agriculture relies on the persistence and recovery of soil functions such as nutrient cycling. Denitrification is a nitrogen (N) removal process that is carried out by nitrifying and denitrifying microorganisms. We still do not have a comprehensive understanding of denitrification and what drives it during revegetation of cropland in degraded karst ecosystems. We used the space-for-time chronosequence method to establish a restoration gradient (sloping cropland, abandoned sloping cropland, and secondary and primary forest) within the Karst Critical Zone Observatory in southwest China. We quantified the abundances of the nitrifier and denitrifier microbial communities and measured the potential denitrification rates (PDR) and basal denitrification rates (BDR) in incubated soils to investigate the denitrification activity in soils at different stages of restoration. The PDR increased through the cropland revegetation phases, and the BDR was lowest in the abandoned sloping cropland (p < 0.05). Both the PDR and BDR were positively correlated with the denitrifier abundances (i.e., nitrite reductase, nirK, and nirS), but were negatively correlated with nitrifier abundances (i.e., ammonia oxidation, AOA, and AOB). The BDR/PDR ratio was 84 % in the sloping cropland but was only 1% in the primary forest, which suggests that the N cycling was inefficient ('leaky') in actively managed farmland soils but was 'tight' in forest soils. The AOB and nirS abundances explained most of the PDR in the sloping cropland and abandoned sloping cropland, while the soil phosphorus (P) contents explained most of the variation in BDR and PDR in land that was forested in later phases of revegetation. Progressive increases in the PDR as the revegetation progressed was directly controlled by the soil available P and total P contents. Available P was also associated with increases in the carbon (C) and N contents, which influenced the BDR indirectly through increased abundances of nirK and nirS. We conclude that revegetation of cropland helped to facilitate 'tight' N-cycling in poor karst soils, and we suggest that judicious P fertilization on sloping cropland would reduce the denitrification activity in degraded ecosystems.