Scaling terrestrial denitrification: Natural N isotopes and nirS gene abundance in temperate biomes

Background/Question/Methods

Denitrification is a complex, microbially-driven process representing a significant nitrogen (N) loss pathway from the biosphere. Stable isotope analyses help examine complex N cycling dynamics, but current models lack sufficient N isotope data for denitrification across global ecosystems – particularly in dry to mesic sites. Additionally, there has never been simultaneous examination of isotope effect and quantifiable soil microbial DNA in any natural system. Our study addresses gaps in current isotopic datasets: (1) What is the isotope effect of denitrification across diverse ecosystems, including deserts; and (2) Can we explicitly link isotopic data to the presence of denitrifying functional gene nirS (nitrite reduction) in soil? Study sites across California represent a wide range of natural ecosystems: oak woodlands (control, grazed, burn treatments), redwood forest (control, clearcut), chaparral (control, burn), and desert (under dominant Larrea tridentata, interstitial space). We analyzed bulk δ¹⁵N in soil pits (50cm depth, subsampling at each horizon) and in foliage from dominant vegetation. We also measured δ¹⁵N and δ¹⁸O in surface soil NO₃^-(12cm depth) both in-situ and before and after week-long, in-situ incubations. Soil microbial DNA was extracted at all sites, and qPCR was used to quantify copies of nirS gene per gram of soil.

Results/Conclusions

In-situ estimates of isotope effects suggest that denitrification occurs across most temperate biomes and conditions examined. Denitrifiers preferentially utilize lighter (14N and 16O) isotopes in NO₃^-, such that the loss of NO₃^- in the soil is expected to correlate negatively with δ¹⁵N and δ¹⁸O. This negative relationship was observed more: in the undisturbed (-3.85) than clearcut redwood site (-1.96); at the burn chaparral (-4.09) than the control chaparral (-2.56); and at the burn ungrazed oak woodland (-7.51) than the burn grazed (-2.07). In contrast, the interstitial desert soil (i.e. soil between vegetation islands) exhibited a decrease in δ¹⁵N of NO₃^- as ln[NO₃^-] decreased (7.28); and desert soils under L.tridentata exhibited no pattern. All soil bulk δ¹⁵N increased with depth and then steadied off, suggesting N transformations occur primarily near 10cm depth. Similarly, nirS abundances peaked near 10cm depth then decreased.  Interstitial desert exhibited a higher peak nirS copies/g soil (x10⁸) than oak woodland sites (x10⁷). Combined data suggests that denitrification does not dominate desert surface soil biogeochemistry, although there is high potential sub-surface activity. Other – likely non-fractionating – processes may further consume the N pool, such as plant uptake.