Temporal variation in nitrogen fixation during encroachment of Prosopis glandulosa into grasslands of the Rio Grande Plains

Background/Question/Methods

Woody encroachment of nitrogen (N)-fixing trees has occurred extensively throughout semi-arid, subtropical grasslands and savannas over the last 150 years. In the Rio Grande Plains of Texas, encroachment of N-fixing Prosopis glandulosa (honey mesquite) is widespread and leads to soil N accretion, presumably from large inputs of fixed N. These N inputs have consequences for carbon storage and trace gas emissions over large areas. Predicting future N inputs on a regional scale requires identification of environmental and plant variables that influence N fixation.  Although variables including temperature, precipitation and plant phenology are hypothesized to influence fixation rates, these  relationships have  rarely been tested in the field. Current foliar δ¹⁵N techniques record an integrated signal of fixation over leaf lifetime and do not directly measure N source pools at specific time points. We have developed another δ¹⁵N-based technique to detect sub-leaf lifetime fixation patterns arising from to seasonal, phenological and plant age-related variation in N fixation rates. In a field study that included P. glandulosa trees ranging in age from 20-145 years, we sampled bulk soil, plant-available soluble soil N, xylem sap and foliage for δ¹⁵N analysis at three time points. 

Results/Conclusions

A large separation in xylem sap δ¹⁵N was observed between P. glandulosa and Zanthoxylum  fagara (lime prickly ash, an associated non-fixing species), with P. glandulosa xylem sap falling much closer to the expected range of atmospheric or fixed N across all age classes.  Z. fagara also showed an ~1:1 relationship between leaf and xylem sap δ¹⁵N , while in P. glandulosa the offset between the two was greater than 1 ‰. This likely suggests that P. glandulosa has large internal N fractionation or that N source use (including proportion of fixed versus soil-derived N) varied over the year. Concentrations of bioavailable soil N increased linearly with plant age, suggesting that increased soil N availability might provide a basis for down-regulation of N fixation during plant development in response to supply. However, we found no relationship between plant age and xylem or foliar δ¹⁵N.  If down-regulation is occurring, it may be obscured by changes in the isotopic composition of the soil bioavailable N pools as total soil N and cycling rates increase. Overall, data suggests that  isotopic analysis of a range of N pools, including xylem and plant-available soil N, can be used to infer patterns of N fixation on sub-annual scales.