Di-nitrogen (N2) fixation, the microbial conversion of N2 gas to biologically reactive ammonium, is a major component of the global nitrogen (N) cycle and has been extensively studied in open ocean and terrestrial ecosystems (Vitousek et al. 2013; Zehr and Capone 2021). Yet, rates and ecological dynamics remain poorly understood for the inland and coastal aquatic ecosystems (e.g., lakes, wetlands, rivers, streams, estuaries) that connect terrestrial and marine biomes (Fulweiler 2023; Marcarelli et al. 2022). Estimates of N inputs from N2 fixation to the global N cycle have not been updated in over 3 decades (Zehr and Capone 2021). Warming surface temperatures (Woolway et al. 2022; but see Reinl et al. 2023), changes in nutrient use and loading to aquatic ecosystems (Manning et al. 2020; Oelsner and Stets 2019), and increasing occurrence of harmful cyanobacterial blooms (Burford et al. 2020) make it essential to understand the spatial and temporal controls on this key biogeochemical process.

The Aquatic N2 Fixation Research Coordination Network (ANF RCN) was established in 2021 to cultivate a new paradigm of the fundamental, yet understudied, role of N2 fixation in ecosystems across the freshwater to marine continuum. To that end, we have identified the following objectives to propel our understanding of aquatic N2 fixation forward:

  • Synthesize the current state of our knowledge of the rates and biodiversity of diazotrophs (organisms that fix N2) for ecosystems along the freshwater-marine continuum.

  • Derive conceptual and quantitative models that characterize the role of biological stoichiometry in constraining N2 fixation at the population, community, and ecosystem scale, specifically comparing/contrasting the variable stoichiometries of diazotrophs with different metabolic strategies.

  • Develop a set of common mathematical and statistical tools enabling the upscaling of N2 fixation for comparison within and among diverse aquatic ecosystems and in regional and global N budgets.

  • Identify major gaps in knowledge that are prohibiting research advancement and generate a plan for coordinating N2 fixation research focused on these issues.

  • Create a community for sharing data, conceptual and quantitative models, and other aquatic N2 fixation information.

We are addressing the first 3 objectives through a series of workshops held between summer 2022 and spring 2025. The last two objectives are being met through grass-roots network building, including hosting a website (https://www.aquaticnfixation.com/) and email list for information sharing, and organizing a series of special sessions hosted at conferences, all of which are open to everyone interested in and actively working in the broad area of aquatic N2 fixation research. To date we have held two of these special sessions: “Exploring Rates and Biological Diversity of Nitrogen Fixation Across Land- and Aquascapes” at the AGU Fall Meeting in New Orleans, Louisiana, USA in December 2021, and “Integrating perspectives on nitrogen fixation across the aquascape” at the Joint Aquatic Sciences Meeting hosted by the Consortium of Aquatic Science Societies in Grand Rapids, Michigan, USA in May 2022. This special issue includes papers contributed by authors who were part of these special sessions, as well as others who are part of the extended RCN network. Here, we introduce these papers and their key contributions, organized around four main themes that point to important unanswered questions and potential areas for future study of N2 fixation across the aquascape.

Novel methods for determining the rates and fate of N2 fixation

A significant challenge facing researchers who study N2 fixation are limitations of the commonly-used methods for quantifying process rates (Fulweiler et al. 2015; White et al. 2020). One of the most common methodological applications for measuring N2 fixation is the acetylene reduction assay (ARA), which has been widely applied to diazotrophic symbionts with plants and free living diazotrophs in soils, sediments, biofilms, and waters. In this method, the nitrogenase enzyme reduces acetylene to ethylene proportionally to the rate by which N2 is reduced to ammonium. However, a major drawback is that the proportion of ethylene produced to ammonium fixed tends to vary across diazotrophic species and the environments in which they inhabit and many studies do not include study specific conversion factors. Kunza and Hall (2023) test the variability in acetylene reduction compared with 15N2 labeled water in benthic biofilms and Nostoc populations in four Wyoming streams. They found that the ratio of ethylene to ammonium produced in their assays varied substantially, and that ethylene production increased almost 2 times faster than 15N2 uptake as stream water temperature increased, providing the potential for temperature to bias N2 fixation rates determined using ARA.

Another increasingly used method for detecting N2 fixation are N2/Ar fluxes measured using membrane inlet mass spectrometry (MIMS). The idea is that N2/Ar ratios below or above expected equilibrium concentrations result from processes that consume (e.g., N2 fixation) or produce (e.g., denitrification) N2. A large challenge for quantifying rates of N2 fixation using this method lies in the fact that we are measuring changes in N2 fluxes against a background of high concentrations - we are essentially flooded in N2. Great care is taken in the collection and incubation of water and sediment samples so that we can mimic natural conditions to best describe environmentally relevant fluxes. Such incubations introduce logistical challenges and limits replication. One possible solution lies in in situ measurements of N2 fluxes, similar to the ecosystem metabolism techniques that use CO2 or oxygen changes through time to estimate gross primary production, ecosystem respiration, and net primary production. Recent efforts have adapted ecosystem metabolism approaches to use N2/Ar measurements to estimate daily denitrification (Reisinger et al. 2016) and N2 removal (Nifong et al. 2020) in streams. In this volume, Taylor et al. (2023) used ecosystem-level experiments to combine in situ N2/Ar ratios with other measurements of N2 fixation (whole-ecosystem 15N enrichments) and denitrification (intact sediment core incubations) in replicated (n = 12) independent mesocosms enriched with N and P. Overall they observed that mesocosms with NO3–N concentrations less 0.1 mg L−1 had N2/Ar saturation ratios below equilibrium indicative of net N2 fixation, while mesocosms with NO3-N concentrations greater than 1 mg L−1 had N2/Ar values above equilibrium indicative of net denitrification. This study is a major step forward as it highlights that N2/Ar grab samples are a helpful and complementary tool for a more complete understanding of N2 fluxes in aquatic systems.

In addition to the difficulties in quantifying ecosystem-scale rates of N2 fixation, it can be a challenge to identify which diazotrophs are contributing to those rates within and across environments. Here Geisler et al. (2023) apply a novel method which immunolabels nitrogenase (Geisler et al. 2019), and combine this with flow cytometry to quantify free-living unicellular diazototrophs. In this way they can determine what percentage of cells in a sample are actively synthesizing nitrogenase. Using this technique, Geisler et al. (2023) demonstrate that only 18% of the cells in a monoculture of Vibrio natriegens were synthesizing nitrogenase, and therefore that not all of the cells were fixing N2. When mixed in co-culture with non-diazotropic E. coli, the percentage of V. natriegnes cells fixing N2 stayed the same but overall N2 fixation rates increased, suggesting the potential for synergetic interactions when diazotrophs are living in mixed microbial communities. Finally, the authors demonstrated the utility of this technique by estimating that diazotrophs ranged from 0.1 to 4.7% of the total bacterial abundance in samples collected from natural marine and freshwater locations. The applications of this technique are potentially widespread, from identifying and quantifying unicellular diazotrophs in habitats where they’ve been rarely studied, to estimating cell-specific rates of N2 fixation for physiological studies of diazotrophs, to probing the community ecology of microbial communities where unicellular diazotrophs may be important keystone members.

Contributing new insight on process rates in understudied habitats

Despite decades of research on N2 fixation across the aquascape, there are still habitats and regions where we understand little about its ecological constraints and the relative rates compared to other biogeochemical processes (Marcarelli et al. 2022). This holds true even in marine ecosystems where most of the research on aquatic N2 fixation has been conducted. N2 fixation in N-replete marine environments such as upwelling regions has long been thought to be negligible. In this volume, Fernández Carrera et al. (2023) add to our limited knowledge on N2 fixation in upwelling regions by studying two locations in the eastern tropical Atlantic Ocean: the seasonal Equatorial upwelling and the Guinea Dome. It has previously been hypothesized that N2 fixation in the Equatorial upwelling is driven by the residual phosphorus left over after primary production (Subramaniam et al. 2013), while N2 fixation in the Guinea Dome is driven by atmospheric deposition of iron from the Sahara Desert. Fernández Carrera et al. (2023) found that large colonies Trichodesmium were patchily distributed in the euphotic zone of the Guinea Dome, but were found in high abundance towards the edges, which coincided with the highest rates of N2 fixation and a supply of phosphorus and iron. In contrast, Trichodesmium colonies were non-existent in the Equatorial upwelling study sites and N2 fixation rates were low. Overall, this study aids our knowledge by highlighting the variable distribution of Trichodesmium colonies—suggesting that lower resolution sampling, typical of many oceanographic studies, might not correctly capture the spatial distribution of these diazotrophs. Further, the lack of Trichodesmium in the Equatorial upwelling implies that other diazotrophs might be responsible for the low rates measured here.

Relative to other habitats along the aquascape, rivers have been relatively understudied, in part because rivers receive and transport reactive N inputs from adjacent terrestrial habitats that should meet the need of microbial assemblages living in these dynamic habitats (Marcarelli et al. 2008). Yet, we know that cyanobacteria are important members of the biofilm assemblages in well-lit streams (Scott and Marcarelli 2012), and riverine algal blooms are increasing in frequency and concern in rivers worldwide (Paerl et al. 2018). In this volume, Valett et al. (2023) leverage a 21-year database of water chemistry and benthic algae data from the Upper Clark Fork River in Montana, USA to show that riverine algal blooms are moderated by snowmelt disturbance, and that late-successional blooms of N2-fixing Nostoc cf. pruniforme are related to N:P ratios among sites. N2 fixation rates measured in one year at this site show that rates can be exceptionally high during the peak of the bloom formation, although the blooms are short-lived and subside with the end of the growing season in this strongly seasonal, montane river. This study joins a growing body of literature demonstrating that low-nutrient, high-light streams can host benthic cyanobacteria that can fix at very high rates for short periods of time when temperature and nutrient conditions are favorable (e.g., Grimm and Petrone 1997; Kunza and Hall 2023). Further, these ephemeral contributions may be important for supporting productivity and ecological assemblages in these habitats at key times and places.

Role of N2 fixation in biotic assemblages

How N2 fixation affects ecological interactions among species in aquatic ecosystems remains poorly understood, partially because of the incredible biodiversity of free-living aquatic diazotrophs (Fernandez et al. 2020) and also because of the incredibly complicated scales of potential interactions (Marcarelli et al. 2022). One possibility is that diazotrophs actively leak recently fixed N into their surrounding environment that may be used by other heterotrophic or autotrophic microorganisms for growth (Mulholland et al. 2006). In contrast, Kunza and Hall (2023, this issue) leveraged their measurements of 15N2 uptake to estimate that benthic biofilms in their study streams appeared to release less than 1% of their newly fixed N to the overlying water, although they acknowledge this could be because released N is quickly recycled into biofilms. Indeed, syntrophic relationships between diazotrophs and non-diazotrophs may be common and potentially explain observations of N2 fixation in unexpected times and places (Fulweiler 2023). Moreover, the cost of N2 fixation for some diazotrophs in terms of both energy and demography may make them susceptible to competition for other resources. For example, stoichiometric modeling has indicated that heterocytous cyanobacteria may be outcompeted by green algae for P due to their susceptibility to grazing while investing in heterocyte formation and nitrogenase production (Grover et al. 2022). Thus, the role of N2 fixation in community ecology has tremendous potential for new research explorations.

In this issue, Marino et al. (2023) found unexpected high rates of N2 fixation on epiphytes of seagrass (Zostera marina) in marine lagoons of the northeastern US. N2 fixation by these phyllosphere diazotrophs was light dependent, but even low light saturated the communities to produce maximum rates. This suggests that the epiphytes were likely cyanobacteria and the relative importance of epiphytic N2 fixation to the lagoon N budget suggested that these cyanobacteria provide a key ecosystem service to help maintain the health of seagrass communities. Similarly, Kohler et al. (2023) measured the importance of N2 fixation by Nostoc in the N budget of Antarctic streams. They identified three common microbial mat types in pristine streams of the of the McMurdo Dry Valley: black mats typically dominated by Nostoc, orange mats dominated by other cyanobacteria such as Leptolyngbya, Oscillatoria, and Phormidium, and green mats dominated by chlorophytes. They found that the distribution of these different microbial mats was spatially segregated both locally and regionally, with black Nostoc mats generally occupying stream margins and becoming more dominant further from glacial meltwater sources that can supply dissolved inorganic N to streams. These results point toward a previously unrecognized importance of N2 fixation as an N source for Antarctic stream communities. Collectively, these papers highlight the need for more research on the ecological role of N2 fixation whereby diazotoph and non-diazotroph populations interact both positively and negatively due to the novel introduction of ‘new’ reactive N into their environment.

Landscape heterogeneity and connectivity

Although we mostly think of N2 fixation as a source of N for specific diazotrophs or their close neighbors, fixed N2 eventually enters the environment either via leakage (Mulholland et al. 2006) or via death and decomposition (Kohler et al. 2018). How that fixed N moves across landscapes and contributes to processes in adjacent or downstream ecosystems has been a particular focus of attention for temperate and tropical forests that host trees with diazotrophic symbionts (Compton et al. 2003; Hedin et al. 2009). In this volume, Ardon et al. (2023) link long-term studies of forest productivity and stream nutrient concentrations from La Selva Biological Station in Costa Rica to demonstrate that in-stream nitrate concentrations were positively related to N2-fixing trees growing on slopes close to streams, but not to those growing elsewhere in the watershed. This suggests that landscape position and productivity of a few specific tree species may alter the N supply to downstream ecosystems, with consequences for processes like leaf litter decomposition in recipient streams. This suggests that understanding landscape heterogeneity will be important not only for understanding where and when N2 fixation occurs, but also for understanding the fate of fixed N2 in ecosystems.

Capturing the spatial variability in ecosystems for any process is complex, and this is especially true for quantifying N2 fixation and denitrification across ecotones where conditions change rapidly in space and time, such as at the interface between wetlands, streams and lakes. Yet Eberhard et al. (2023) endeavor to do just this as they quantify net N2 fluxes and environmental characteristics (e.g., temperature, nutrient concentrations, organic matter content) at five sites in Lake Huron and Lake Superior, USA. While they observed that median denitrification rates (914 µg N m−2 h−1) were higher than median N2 fixation rates (5.5 µg N m−2 h−1) – they found that the denitrification and N2 fixation were co-occurring in almost 50% of the measured sites across all five sampling locations and all habitats. Their predictive model suggests that organic matter was an important driver of both processes, and P concentrations could be important for N2 fixation, but that N2 fixation rates were not predicted by commonly assumed drivers like N availability, N:P ratios, or nutrient limitation status of the biofilms. This study highlights that the spatial scales at which we are able to measure environmental characteristics (e.g., nutrient limitation) may not be relevant to the spatial scales at which we are measuring rates of denitrification and N2 fixation—that is, what we measure may not be what the microbial community observes. This has important implications for how we design future studies and how we understand the complexities of N cycling across habitats.

Looking forward

The studies in this volume demonstrate the complicated nature of N2 fixation estimation in variable environments, illustrate the critical role of N2 fixation in ecological communities, and highlight that the fate of fixed N remains a critical unresolved puzzle for understanding these ecological and biogeochemical dynamics. Together, they point to a variety of areas at the frontier of N2 fixation research that can advance our understanding of this critical process in the coming decade, such as:

  • Development of new tools and refinement of existing protocols to better quantify N2 fixation rates in time and space, and to probe interactions between diazotrophs and non-diazotrophs in aquatic habitats.

  • Quantification of N2 fixation rates and diazotroph diversity in poorly studied habitats, particularly those that host non-cyanobacterial diazotrophs and those where study has been limited by geographic access, funding limitations and temperate bias.

  • Determining the short-term fate of fixed N and how that may vary among different diazotrophs depending on taxonomic identity, metabolic strategy, habitat, and syntrophies with other microbes.

  • Incorporating new N derived from N2 fixation into regional budgets over mid-to-long term time scales, including the transport and remineralization of N incorporated into diazotroph biomass and consequences for adjacent and/or downstream ecosystems.

Advancing knowledge at these frontiers will require a robust and collaborative global network of researchers who are sharing insights, techniques, and knowledge. This special issue is an important reflection of the community of aquatic N2 fixation researchers that exists, and which we hope will grow and thrive in the years to come.