Towards enhanced sensitivity of the 15N gas flux method for quantifying denitrification in soil

Denitrification is the least studied process of the global N cycle mainly due to the sensitivity required to discriminate small fluxes of soil emitted N2 against the high atmospheric N2 background. We aimed to enhance the sensitivity of the 15N Gas Flux method to measure in situ denitrification rates by optimising the quantity of 15N–NO3 tracer applied and by using an artificial atmosphere (containing 5 % N2, 20 % O2, 75 % He and 0.11 ppm of N2O) during field incubation. We first conducted a dose-response laboratory study to assess the stimulation effect of nitrate tracer addition. Subsequently, we developed two novel approaches to measure in situ denitrification rates, using either modified static chambers or intact soil cores inside plastic liners; where in both cases the entire headspace was replaced by the artificial atmosphere prior to incubation. Furthermore, we compared the two models of calculations of the 15N Gas Flux method (the “Mulvaney & Boast” and “Arah” models) as well as the calculated 15N enrichment of the soil denitrifying pool based on either N2 or N2O isotopologue distribution data. The results showed that doubling the amount of ambient nitrate did not lead to a significant stimulation of denitrification activity in our case. However, excessive amendment of nitrate (e.g. 20 times the ambient levels) increased the denitrification product ratio by stimulating nitrous oxide emission. Our two novel field techniques were successful in measuring in situ denitrification rates, however, the liner method was preferred due to a higher success rate of N2 flux detection (up to 90 %), a higher throughput (up to 24 cores at a time) and improved spatial resolution. Under high-resolution instruments, our N2 limit of detection was 160 ppb, which is 5-fold better than the original method. The Mulvaney & Boast model performed better than the Arah one and consistently yielded higher fluxes (17 % at maximum), especially for low 15N enrichments of the soil denitrifying pool and short times of incubation. The 15N enrichments calculated with either N2 or N2O data differed statistically, but the magnitude of difference was small (4.6 % at maximum). Measuring in situ denitrification is imperative to quantify realistic fluxes and the liner method presented here is an inexpensive, reproducible and high-resolution candidate. For increased sensitivity, we recommend using the method of Mulvaney & Boast for N2O emissions and the resulting 15N enrichment in combination with 29N2 data (only) to determine N2 emissions.