Theoretical evaluation of 15N isotopic methods for measuring symbiotic nitrogen fixation in the field

International audienceIsotopic methods for the measurement of symbiotic N2 fixation by leguminous plants in the field rely on the use of differences in 15N enrichment between the N sources potentially available for leguminous crops, soil mineral N and atmospheric N2 . This methodology has been fully documented, especially concerning limitations due to non uniform and non constant distribution of 15N and to the use of a reference plant to measure it. Although all authors recognise the necessity of isotopic methods for giving yield independent and time-integrated estimates of symbiotic fixation, they also agree that these methods intrinsically remain imperfect. Our aim in this chapter is (i) to briefly review the three major isotopic methods and recall the main assumptions they involve, (ii) to evaluate the theoretical precision of those methods by performing sensitivity analysis to all their parameters, in the perspective of precisely delimiting their validity domain and (iii) to quantify the error made when using the method with the largest spectrum of application and to propose solutions to minimise it. The natural abundance method (NA) is the simplest method as no added fertiliser is required. The method isotopic dilution (ID) requires 15N-labelled fertiliser application to increase 15N soil enrichment. The multi-enrichment technique (MET) relies on the use of several treatments receiving the same amount of fertiliser but labelled at variable 15N enrichments. Using an original mathematical analysis, we show that the precision of NA and MET is likely to be low if the difference in 15N abundance between the soil and atmosphere is low (lower than 8 ‰). Otherwise, the use of NA or MET requires a very precise determination of the isotopic fractionation rate due to symbiotic fixation (εFix). The ID method circumvents this problem. It has the largest validity domain as it can be used in soils slightly enriched in 15Nand does not require such precise determination of εFix. . However, the main assumption of the ID method is that the relative uptake of soil and fertiliser N is identical for the fixing and the non fixing plants. If this assumption is not valid, large discrepancies can occur between the actual and calculated contributions of symbiotic fixation to the overall N acquisition by the plants (pA). This error was evaluated as a function of the level of pA, 15N enrichments of the soil and fertiliser and the proportion of fertiliser N retrieved by the legume to total mineral N. Recommendations for optimal application of 15N labelled fertiliser used to enrich the soil mineral N are given