The trouble with the bubble


Past studies have suggested that the ocean's nitrogen budget has a deficit of fixed nitrogen. This view may now change, thanks to upward revisions of the rate of nitrogen input through biological activity. See Letter p.361

As is the case for gardens, forests and fields, the availability of fixed nitrogen (such as nitrate and ammonia) can limit the productivity of our seas. Oceanographers are therefore interested in the relative magnitudes of sources and sinks of fixed nitrogen as dynamic controls of ocean fertility. The trajectory of the oceanic nitrogen inventory has been a long-debated and contentious topic: direct rate measurements of biologically mediated nitrogen fluxes suggest that the ocean is currently being depleted of this resource, whereas geochemical evidence indicates a steady-state, balanced nitrogen budget1. On page 361 of this issue, Großkopf et al.2 report that applying a new twist to an old method may reconcile these views and lead to a revision of our understanding of the present-day oceanic fixed-nitrogen budgetFootnote 1.

Unlike other elemental cycles, which are predominantly influenced by riverine, atmospheric or sedimentary fluxes (such as the iron and phosphorus cycles), the oceanic inventory of fixed nitrogen is largely set by the push and pull of two biological processes. First, fixed nitrogen is added to the oceans by microorganisms known as diazotrophs, which convert the nearly limitless supply of atmospheric nitrogen gas (N2) dissolved in sea water to ammonia. Second, in sediments and in oxygen-depleted zones of the ocean, fixed nitrogen is chemically reduced to N2 by the microbial processes of denitrification and anaerobic ammonia oxidation. Understanding the relative balance of the fluxes of fixed nitrogen in the ocean requires the rate measurements of these competing processes to be accurate and well constrained.

On the source side of this nitrogen budget, the majority of published estimates of marine N2-fixation rates3 are based on a fairly straightforward protocol4: add a bubble of isotopically labelled nitrogen gas (15N2) to a sample of sea water; calculate the initial enrichment of 15N2 in the sample using the ideal-gas law; incubate the sample for a specified period of time; and then measure the fraction of the isotopic tracer that is incorporated into cellular material in the sample. The net rate of N2 fixation is then calculated by multiplying the fraction of biomass labelled per unit of time by the mass of nitrogen in particles suspended in the sample (which are assumed to be living material).

When these rates are extrapolated from localized study regions to the expanse of ocean thought to be habitable to diazotrophs, the resulting global N2-fixation rates are generally found to be far less than those inferred by geochemical approaches1. This potential discrepancy set scientists on a “relentless search”5 for potentially missing sources of fixed nitrogen. These efforts paid off — they led to the discovery of a wider cast of nitrogen-fixing 'characters' than had been known before6; an expanded view of the ocean habitats occupied by such organisms7; the elucidation of previously unknown metabolic strategies for diazotrophic growth8; and descriptions of newly recognized symbioses between diazotrophs and other microorganisms9. Despite these revelations, the available N2-fixation-rate measurements still indicated a deficit of oceanic fixed nitrogen. And so the question remained: if the nitrogen budget is in balance, then where have rate measurements of oceanic nitrogen fluxes gone wrong?

In 2010, a potential solution to this conundrum was reported10 with the demonstration that when 15N2 is added to a sample of sea water, it does not rapidly dissolve in the surrounding liquid. This was an unexpectedly simple solution — if a fraction of the added 15N2 remains in the gas phase during fixation-rate measurements, instead of dissolving in sea water as was assumed, then one would underestimate N2-fixation rates. To assess the effect of incomplete 15N2 dissolution, laboratory experiments were performed in which the classic method was revised10: instead of adding a bubble of 15N2 directly to a sample, the tracer was added to aliquots of filtered and degassed sea water, which were agitated until the bubble dissolved fully. The isotopic enrichment of the water was then verified by mass spectrometry10. By adding a known volume of this water to samples, rather than a bubble of15N2, the initial concentration of added tracer could be constrained, leading to more-accurate measurements. This methodological tweak resulted in significantly higher fixation rates being measured in the laboratory — a finding that, if applicable to in situ measurements, could substantially reduce the apparent imbalances previously observed for the marine fixed-nitrogen budget.

Großkopf and colleagues now report the first application of the modified 'dissolution' method to the analysis of natural populations of diazotrophs. During two research cruises in the Atlantic Ocean, the authors made 17 side-by-side comparisons of N2-fixation rates obtained using the conventional bubble-addition approach4 and the revised method10. Their major finding was that rates measured using the dissolution method were on average 1.7 times greater than those derived from the bubble method. Accordingly, the authors used this number in a first-pass calculation to adjust estimates of global nitrogen-fixation rates upward.

A secondary finding of this work2 was that there was a “poor overall correlation between the N2-fixation rates calculated using the two methods”. The magnitude of the difference between the methods varies, and in a few instances the methods actually generated similar values. Großkopf et al. discuss mechanisms that may explain this variation. Their explanations include the use of different time intervals between the injection of a 15N2 bubble into a sample and the onset of the daily period of biological N2 fixation, as well as the variable buoyancy of diazotrophs and hence their proximity to the 15N2 bubble in the experiments. Intriguingly, both of these mechanisms seem to reflect the structure of diazotroph communities, to the extent that the identity of the dominant diazotroph along any cruise transect will affect the N2-fixation rates measured using the bubble-addition technique (Fig. 1).

Figure 1: Marine nitrogen fixation.

A commonly used method for measuring the rate of nitrogen (N2) fixation by diazotrophic microorganisms in the ocean is to add isotopically labelled nitrogen gas, 15N2, to a sample of sea water, and then to measure the uptake of this nitrogen tracer by diazotrophs in the sample. The 15N2 has classically been added as a bubble of gas4, but Großkopf et al.2 have compared the results of this traditional approach with those of a modified, more reliable version of the assay10 — in which the tracer is added as a solution in filtered sea water — for measurements of the N2-fixation rates of diazotrophs in the Atlantic Ocean. They found that the rates obtained using the modified method were, on average, greater than those obtained using the traditional approach. The differences between rates may in part track the composition of the diazotroph communities sampled — the discrepancies were larger when unicellular cyanobacteria, such as Crocosphaera, were in the majority, and smaller when the colony-forming genus Trichodesmium predominated.

Großkopf and colleagues' findings will surely stimulate a spirited and necessary debate on how, and if, we can reinterpret the past decade or so of N2-fixation measurements. In the past, the bubble-addition assay has been implemented inconsistently — different researchers used a range of incubation durations; experiments were initiated at different times of day; varying volumes of 15N2 were injected; and experiments unavoidably targeted a spectrum of nitrogen-fixing communities. Given Großkopf and colleagues' results and this legacy of methodological inconsistency, some past measurements may need no revision, whereas others probably need large adjustments. Unfortunately, reliably identifying which measurements require revision may not be possible. The way forward is therefore clear: to understand the current state of the oceanic nitrogen cycle, researchers need to forge a path to a standardized methodology and to assess whether the database of historical N2-fixation rates3 can be adjusted, or whether field efforts will need to begin anew.


  1. 1.

    *This article and the paper under discussion2 were published online on 8 August 2012.


  1. 1

    Codispoti, L. A. Biogeosci. Discuss. 3, 1203–1246 (2006).

    ADS  Article  Google Scholar 

  2. 2

    Großkopf, T. et al. Nature 488, 361–364 10.1038/nature11338 (2012).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Luo, Y.-W. et al. Earth Syst. Sci. Data Discuss. 5, 47–106 (2012).

    ADS  Article  Google Scholar 

  4. 4

    Montoya, J. P., Voss, M., Kahler, P. & Capone, D. G. Appl. Environ. Microbiol. 62, 986–993 (1996).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. 5

    Mahaffey, C., Michaels, A. F. & Capone, D. G. Am. J. Sci. 305, 546–595 (2005).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Zehr, J. P., Mellon, M. T. & Zani, S. Appl. Environ. Microbiol. 64, 3444–3450 (1998).

    CAS  PubMed  PubMed Central  Google Scholar 

  7. 7

    Deutsch, C., Sarmiento, J. L., Sigman, D. M., Gruber, N. & Dunne, J. P. Nature 445, 163–167 (2007).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Zehr, J. P. et al. Science 322, 1110–1112 (2008).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Foster, R. A. & O'Mullen, G. D. in Nitrogen in the Marine Environment 2nd edn (eds Capone, D. G. et al.) 1197–1218 (Elsevier, 2008).

    Google Scholar 

  10. 10

    Mohr, W., Großkopf, T., Wallace, D. W. R. & LaRoche, J. PLoS ONE 5, e12583 (2010).

    ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Angelicque E. White.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

White, A. The trouble with the bubble. Nature 488, 290–291 (2012). https://doi.org/10.1038/nature11481

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.