Elusive marine nitrogen fixation

Every year, the ocean is losing more than 200 Tg N from its bioavailable nitrogen pool through denitrification processes: that is, all of the processes that convert fixed nitrogen to nitrogen gas (1). If this loss wasn't compensated by N-inputs, the ocean would quickly loose its fertility, because fixed nitrogen is the most important nutrient fueling phytoplankton growth in the ocean (2). Thanks to the N₂-fixers, a special group of marine plankton that live in the sunlit upper ocean and that are able to convert dissolved nitrogen gas into bioavailable nitrogen such as ammonium and organic nitrogen, a substantial part of this loss is compensated for (3). A question that has fascinated oceanographers for decades (4, 5) is how denitrification and N₂ fixation are coupled to each other; that is, how the ocean is ensuring that these two major sink and source processes are “dancing” with each other in step. Without such synchronization, the ocean would have experienced large swings in its fixed nitrogen inventory, potentially causing major disruptions to ocean life, something that Earth has not seen in its recent history. One hypothesis is that marine N₂ fixation occurs in close spatial association with the major regions of marine denitrification, such as the eastern tropical North and South Pacific (6) (Fig. 1). However, until recently, no one has looked for N₂ fixation in these regions (7), largely because this was against the prevailing view that N₂-fixers thrive mostly in regions where their high iron demand is met by sufficient Fe deposition from the atmosphere, such as the subtropical North Atlantic (8). Knapp et al. (9) now report on the results from their journey to one of these major denitrification regions. Their goal was to find the purported high levels of N₂ fixation associated with these regions, and to test the hypothesis of the close spatial association between N₂ fixation and denitrification.

Fig. 1.

Map of vertically integrated N₂-fixation rates in the subtropical and tropical Pacific and western Atlantic (µmol m⁻² d⁻¹). The color shading in the background are N₂-fixation estimates inferred from surface nutrients through an inverse modeling approach (6). The rectangles show a gridded synthesis of directly measured N₂-fixation rates (7), whereas the circles indicate the new measurements added by Knapp et al. (9). The green shaded areas denote the oxygen-deficient zones of the eastern tropical North and South Pacific, where the majority of the global ocean water column denitrification is taking place. Although the inverse modeling results would have suggested a close spatial association between N₂ fixation and denitrification in the eastern tropical Pacific, the new results by Knapp et al. (9) do not support this conclusion.

In 2010 and 2011, Knapp et al. (9) undertook two cruises to the eastern tropical South Pacific, one of the most intense oxygen-deficient zones of the world’s ocean, which alone is responsible for about 40% of global water column denitrification (10). In the eastern tropical South Pacific, oxygen drops to very low oxygen levels (less than 1% of saturation) already at depths as shallow as 100 m, and the ocean remains virtually oxygen-free down to several hundred meters depth, thus creating the conditions under which the denitrifiers thrive. Most of these organisms are heterotrophs that have the ability to switch from oxygen to nitrate as the oxidant to remineralize the organic matter that is being been transported there from the surface ocean (1). Additional denitrification can occur through anaerobic ammonium oxidation (anammox) (1). Both processes lead to waters that are highly deficient in the concentration of nitrate relative to that of phosphate, although both nutrients remain abundant. When these waters come to the surface, they first induce a bloom of regular phytoplankton that consume nitrate and phosphate roughly with the Redfield ratio of 16:1 (3). This usually continues until all nitrate is exhausted, leaving an excess of phosphate behind. Knapp et al. (9) targeted these nitrate-poor but phosphate-rich surface waters because they should provide the ideal niche for N₂-fixers. Under these conditions, the N₂-fixers have ample light, which they need for photosynthesis, and thanks to regular phytoplankton being hampered in their growth—because of their being limited by nitrate availability—the N₂-fixers have limited competition for the phosphate (and other nutrients, such as iron) they still need to take up from the water to fuel their growth.

During their two cruises, Knapp et al. (9) measured at a total of six stations the rates of N₂ fixation, as well as the contribution of the newly fixed nitrogen to the total export of organic nitrogen. The authors used a broad set of methods, including geochemical mass balances on the basis of the natural abundance of ¹⁵N, direct incubation experiments with ¹⁵N-labeled N₂, and genomic analyses, with an emphasis on the quantification of the major genes encoding nitrogen fixation. Each of these methods has its strengths and weaknesses. The ¹⁵N mass balance approach integrates N₂ fixation in time and space, and is therefore well suited to smooth out any short-term spatial and temporal variability. However, it estimates only the fraction of the exported organic nitrogen stemming from N₂ fixation, requiring an independent estimate of the magnitude of export production to convert this fraction into a N₂-fixation estimate. Its application requires many assumptions: for example, steady-state, limited horizontal mixing and low export of dissolved organic nitrogen, as well as a good estimate of the ¹⁵N content of the nitrate that is being supplied into the region of interest. As a result of these uncertainties, the ¹⁵N mass balance approach has a detection limit that is quite high, preventing it from detecting pervasive, but low levels of N₂ fixation. Direct estimates provide relatively accurate point measurements of the biomass and rates of N₂ fixation, but due to their very limited number, the conclusions drawn from them are highly sensitive to unrecognized spatial and temporal variations.

Irrespective of the methods, Knapp et al. (9) found at all stations undetectable-to-low rates of N₂ fixation. With the isotopic mass balance approach, they were able to detect N₂ fixation at only two of the six stations and the rates ranged from 0 to 23 µmol N m⁻² d⁻¹ when they used the sediment trap fluxes to multiply the fractional estimates of nitrogen fixation to obtain rates of N₂ fixation. Higher rates would be obtained if alternative export estimates were used, pushing the upper range to values of 200 µmol N m⁻² d⁻¹ or more. However, these alternative export estimates are fraught with high uncertainties. Nevertheless, the direct incubation experiments also tended to give higher N₂-fixation rates than the mass balance values suggest. In particular, the direct measurements showed measurable rates at all stations, yielding on average rates of about 50 µmol N m⁻² d⁻¹. Although sizeable, all of these rates are much lower than what is generally observed in the subtropical North Atlantic (order of 200 µmol N m⁻² d⁻¹ and more); that is, in a region far away from the place that generates the nitrate deficits thought to stimulate N₂ fixation (Fig. 1). The estimates are also much smaller than N₂-fixation rates inferred from surface nutrients through an inverse modeling approach (6): that is, from the study that gave rise to the spatial-coupling hypothesis. Thus, the new data by Knapp et al. (9) put the hypothesis of a close spatial coupling between N₂ fixation and denitrification into serious question.

However, before fully rejecting this hypothesis, it is worthwhile to consider potential reasons for this discrepancy. The high level of variability in space and time that characterizes the ocean represents a challenge for the interpretation of any observation, but may be particularly critical here given the small number of stations. Although the isotopic mass balance is less prone to biases in such situations, this method still integrates only over a period of a few days and over spatial scales of a few tens of kilometers: that is, smaller than the dominant scales of variability in the ocean, which occurs on mesoscales (tens to hundreds of kilometers) and lasts days to weeks. Thus, with only six stations there is a good chance of missing potential hotspots of N₂ fixation, such as has been identified for denitrification (11). In contrast, the inverse modeling method (6) implicitly integrates over many years, if not decades, and over scales of 100–1,000 km; that is, it is likely not biased by variability. Furthermore, there might be strong spatial gradients in N₂ fixation, and Knapp et al. (9) just may have missed the high N₂-fixation regions. However, the waters targeted by Knapp et al. had the ingredients for fostering the growth of the N₂-fixers, so if these hotspots exist, they would have had a good chance to sample it. Yet, whereas the results of Knapp et al. present a serious challenge to the spatial-coupling hypothesis, six measurements from the edge of the purported high N₂-fixation area are clearly not enough to refute the hypothesis.

On the other hand, the inverse-model approach that underlies the spatial-coupling hypothesis (6) might have diagnosed the spatial distribution of the N₂ fixation incorrectly, as a result of its neglecting variable phytoplankton stoichiometry, for example. Furthermore, the detailed spatial distribution diagnosed by this approach is rather sensitive to the treatment of dissolved organic phosphorus, as this permits some spatial decoupling of N₂ fixation and denitrification in their approach. In addition, potential biases in the nutrient database could have caused overly large diagnosed N₂-fixation rates in the eastern tropical Pacific. Notwithstanding these potential errors, support for the spatial proximity of N₂ fixation and denitrification comes from the finding that subsurface oxygen is, next to sea-surface temperature, the best predictor for the global spatial distribution of the measured rates of N₂ fixation (7).

Accepting for the moment that N₂ fixation is low in the nitrate-poor and phosphate-rich waters of the eastern tropical South Pacific, the question emerges: what prevents the N₂-fixers from excelling there? Knapp et al. (9) suggest that the N₂-fixers are limited by iron, which is scarce in this part of the ocean as a result of very low atmospheric input. This is a very feasible explanation, given the substantial iron requirements of the N₂-fixers, but Knapp et al. did not make the necessary measurements and incubation experiments to support this statement. If iron was indeed the missing ingredient in the eastern tropical and subtropical Pacific, then one expects the western subtropical and tropical Pacific (i.e., the regions closer to the continental sources of iron) to host the majority of the N₂-fixers in the Pacific. More recent model simulations and the available N₂-fixation rate observations indeed support such a distribution, but the available data are too sparse to constrain this statement well.

Thus, if N₂ fixation and denitrification are, after all, spatially much more separated than hypothesized, then our quest to understand the synchronization between the two dance partners (i.e., how N₂ fixation and denitrification are coupled to each other to ensure a global homoestasis of the marine nitrogen cycle) remains unsolved. Although laboratory studies are very important to better understand the factors that control the growth of individual N₂-fixers, it is only a vast increase in the coverage of marine N₂-fixation measurements that will permit us to better determine how nitrogen fixation is controlled in the environment. Thus, the search for the elusive N₂-fixers needs to continue. The journey remains exciting, though, because the range of marine organisms that have the capability to fix nitrogen is rapidly expanding and already spans several clades, including symbiotic and parasitic organisms, as well as heterotrophs (12). At the same time, the marine nitrogen cycle is becoming an ever faster moving target of study. Ocean warming, ocean acidification, and deoxygenation, together with a massive increase in the input of fixed nitrogen through rivers and the atmosphere, are pushing the marine nitrogen cycle and its organisms strongly out of balance (5).