Seasonal and individual variation in leguminous tree nitrogen fixation in a natural ecosystem

Background/Question/Methods

Predicting the magnitude of symbiotic nitrogen (N) fixation by woody plants has traditionally been impeded by the difficulty of measuring fixation rates in the field. Foliar δ¹⁵N measurement is the most commonly applied method for estimating fixation in woody perennials, especially when root nodules cannot be reliably recovered. However foliar δ¹⁵N values record an integrated signal of fixation over leaf lifetime and must rely on interspecies comparisons. The influence of factors such as climatic variation or phenology-driven N demand on N fixation is therefore difficult to quantify, because these processes operate on sub-annual scales. Furthermore, intrinsic differences in uptake and tissue fractionation among species are a source of inherent error associated with the foliar method.

We attempted to improve the temporal resolution of δ¹⁵N data by sampling tree xylem sap (a N pool with a much shorter turnover time than foliage) at multiple time points within 12 months.  We compared these data to foliar and rooting-zone soil soluble δ¹⁵N samples and tracked individual trees over time. The method was applied to N-fixing tree Prosopis glandulosa (honey mesquite), a widespread facilitator of woody encroachment in south Texas. N-fixing trees are common facilitators of encroachment worldwide and drive significant ecosystem N inputs. 

Results/Conclusions

We observed significant seasonal variation in P. glandulosa xylem sap δ¹⁵N that could not be explained by variation in the soil inorganic N source, suggesting that sub-annual environmental or phenological processes may influence plant N fixation rates. Furthermore, we observed that the difference between xylem sap and foliar δ¹⁵N within P. glandulosa was substantially different to that within the best available woody non-fixing reference species in this system. This was sufficient to cause a two-fold difference in the calculated %NDFA (percent nitrogen derived from fixation) derived from foliar values compared with those derived from xylem sap δ¹⁵N.

 Between individuals, the range of variation in xylem sap δ¹⁵N at a single time point was 2.5 ‰ and within individuals, this value varied by up to 2 ‰ across seasons. Significant linear relationships between the sap δ¹⁵N values of individual trees over time suggest that some trees consistently fix more N or access soil N pools that are consistently isotopically distinct. The variation in potential fixation between trees could not be explained by tree age, soil total N or inorganic N concentrations; instead we propose that genotype interactions or extent of rhizobial colonization may account for this observation.