Sources of nitrous oxide emissions from hydroponic tomato cultivation: Evidence from stable isotope analyses

Introduction: Hydroponic vegetable cultivation is characterized by high intensity and frequent nitrogen fertilizer application, which is related to greenhouse gas emissions, especially in the form of nitrous oxide (N₂O). So far, there is little knowledge about the sources of N₂O emissions from hydroponic systems, with the few studies indicating that denitrification could play a major role.

Methods: Here, we use evidence from an experiment with tomato plants (Solanum lycopersicum) grown in a hydroponic greenhouse setup to further shed light into the process of N₂O production based on the N₂O isotopocule method and the ¹⁵N tracing approach. Gas samples from the headspace of rock wool substrate were collected prior to and after ¹⁵N labeling at two occasions using the closed chamber method and analyzed by gas chromatography and stable isotope ratio mass spectrometry.

Results: The isotopocule analyses revealed that either heterotrophic bacterial denitrification (bD) or nitrifier denitrification (nD) was the major source of N₂O emissions, when a typical nutrient solution with a low ammonium concentration (1–6 mg L⁻¹) was applied. Furthermore, the isotopic shift in ¹⁵N site preference and in δ¹⁸O values indicated that approximately 80–90% of the N₂O produced were already reduced to N₂ by denitrifiers inside the rock wool substrate. Despite higher concentrations of ammonium present during the ¹⁵N labeling (30–60 mg L⁻¹), results from the ¹⁵N tracing approach showed that N₂O mainly originated from bD. Both, ¹⁵N label supplied in the form of ammonium and ¹⁵N label supplied in the form of nitrate, increased the ¹⁵N enrichment of N₂O. This pointed to the contribution of other processes than bD. Nitrification activity was indicated by the conversion of small amounts of ¹⁵N-labeled ammonium into nitrate.

Discussion/Conclusion: Comparing the results from N₂O isotopocule analyses and the ¹⁵N tracing approach, likely a combination of bD, nD, and coupled nitrification and denitrification (cND) was responsible for the vast part of N₂O emissions observed in this study. Overall, our findings help to better understand the processes underlying N₂O and N₂ emissions from hydroponic tomato cultivation, and thereby facilitate the development of targeted N₂O mitigation measures.

1. Introduction

Based on a variety of technical innovations in greenhouse vegetable production, the use of soilless culture systems (commonly referred to as “hydroponics”) has grown in importance during the last 30–40 years (Gruda, 2009; Savvas et al., 2013; Savvas and Gruda, 2018). Controlled environment systems are considered by some as key part of future food production (Lakhiar et al., 2018; Cowan et al., 2022). This is largely due to the possibility of operating hydroponic systems in greenhouses in regions with unfavorable climatic conditions and in urban areas (Sharma et al., 2018; Small et al., 2019). Closed hydroponic systems also allow the re-utilization of drained nutrient solution from the root zone by recirculating the collected drain after mixing with stock solution. The high water and nutrient efficiency of closed hydroponic systems as well as the reduction of soil-borne diseases are considered as major advantages compared to soil-based cultivation (Gruda, 2009; Savvas and Gruda, 2018). Besides, the high water and nutrient efficiency makes hydroponic systems also interesting for the production of supplemental fresh food during space missions (Wheeler, 2017). Nonetheless, there are still losses occurring in the form of gaseous nitrogen (N) emissions, which may sum up to more than 10% of the N applied in the nutrient solution (Daum and Schenk, 1996a). Due to the high N application rate and dosage frequency in hydroponics, there is also a high potential for gaseous N emissions, in particular nitrous oxide (N₂O) from microbial processes such as nitrification (Ni) and heterotrophic bacterial denitrification (bD; Daum and Schenk, 1996b; Lin et al., 2022). If bD is complete, N losses in the form of molecular nitrogen (N₂) due to N₂O reduction might also occur. So far, only a few studies investigated volatile N losses from hydroponic systems. Some of these studies found N₂O emission factors higher than the IPCC estimate of 1% N₂O-N for applied N fertilizer in soil cultivation (Daum and Schenk, 1996a; Hashida et al., 2014; Yoshihara et al., 2016), while others found lower N₂O emission factors (Llorach-Massana et al., 2017; Halbert-Howard et al., 2021; Karlowsky et al., 2021).

The specialty of hydroponic systems is that inert substrates such as sand, perlite, or rock wool can be used, which limits the availability of organic carbon for heterotrophic denitrifiers. In this case, the hydroponic growing medium consists only of the substrate matrix and the supplied nutrient solution, which is mostly composed of mineral fertilizers dissolved in water. Nevertheless, bD has been considered as the main source of gaseous N emissions from hydroponic systems with inert substrates (Daum and Schenk, 1996a, 1996b, 1998). Whereas a more recent study by Lin et al. (2022) with tomato plants cultivated on peat and coir substrates found also significant shares of N₂O produced by Ni, which depended on the substrate used. In hydroponic systems with inert growing media, various factors may favor bD over Ni activity, i.e., (i) frequent irrigation pulses, (ii) slightly acidic pH values (pH 5–6.5) in the nutrient solution, (iii) often high nitrate (NO₃⁻) to ammonium (NH₄⁺) ratios, and (iv) the presence of root exudates and debris. Yet, there is little knowledge on the processes underlying gaseous N emissions from hydroponic systems. In particular, it is unclear to which extend other processes such as fungal denitrification (fD), nitrifier denitrification (nD), or coupled nitrification and denitrification (cND) play a role in hydroponic systems. A study of functional microbial genes by Hashida et al. (2014) found 3–5 times higher gene copy numbers for denitrifiers than for nitrifiers, but the abundance of functional Ni and bD genes had no clear relationship with measured N₂O emissions. N₂ emissions from bD, which are more difficult to analyze due to the high atmospheric concentration of N₂, have only been researched by Daum and Schenk (1996a, 1996b, 1997, 1998) in hydroponic systems, using the acetylene inhibition method. However, today, it is known that this method is not suitable to quantify N₂ production, mainly due to catalytic decomposition of NO in presence of O₂ (Felber et al., 2012; Nadeem et al., 2013), which cannot be excluded in the setup used in the Daum and Schenk studies (ibid.).

Alternative methods for detecting N₂ emissions include (i) the use of closed chambers filled with other inert gases such as helium and the analysis of N₂ in gas samples on a gas chromatograph (helium incubation method) (Scholefield et al., 1997), (ii) the labeling with ¹⁵N supplied by the fertilizer and the measurement of ¹⁵N contents in N₂O and N₂ (¹⁵N tracing approach) (e.g., Stevens and Laughlin, 1998; Buchen et al., 2016), and (iii) the analysis of the isotopic composition (δ¹⁸O, δ¹⁵Nbulk value and the intramolecular distribution of ¹⁵N in N₂O) of the four most abundant N₂O isotopocules, which are indicative for N₂O production pathways, but also altered during the N₂O reduction process (N₂O isotopocule method) (e.g., Decock and Six, 2013; Lewicka-Szczebak et al., 2017). Unfortunately, the helium incubation method to directly measure N₂ emissions requires a high technical effort and is very prone to leakage and is therefore mainly used for the analysis of soil cores in the laboratory (Groffman et al., 2006). Both, the N₂O isotopocule method and the ¹⁵N tracing approach, require little technical effort in the field or greenhouse, can be combined with the usual chamber-based gas flux measurements for detecting N₂O emission rates, and are suitable to assess the microbial processes that drive the N₂O emission (Lewicka-Szczebak et al., 2020). The N₂ isotopocule method works well with natural abundance stable isotope ratios and only requires the capacity for stable isotope analyses. However, due to the multitude of possible N₂O processes (Butterbach-Bahl et al., 2013) and the variability found in isotope contents and fractionation factors, uncertainties of its results have to be taken into account (Wu et al., 2019). The ¹⁵N tracing approach allows to quantify the conversion of ¹⁵N-enriched substrates such as NO₃⁻ or NH₄⁺ to different products, including N₂O and N₂ (¹⁵N mass balance). Though to obtain sufficient ¹⁵N enrichment of N₂ for detection of N₂ production, high amounts of expensive ¹⁵N tracer have to be applied, limiting the use of the ¹⁵N tracing approach for detecting N₂ fluxes by the experimental budget. Moreover, under ambient atmosphere, its sensitivity is quite low (Zaman et al., 2021).

In this study, we used a combination of the N₂O isotopocule method and the ¹⁵N tracing approach to further shed light into the processes underlying gaseous N emissions from hydroponic systems. Analyzing the N₂O isotopocules and using the dual isotope plot (“isotopocule mapping approach”) is the most common interpretation strategy to estimate the fractions of N₂O produced by bD and/or nD, fD, and Ni (e.g., Lewicka-Szczebak et al., 2017). The results from N₂O isotopocule analysis were also recently found to be in good accordance with the analysis of functional nitrifier and denitrifier genes (Lin et al., 2022). In contrast to the isotopocule method, the ¹⁵N tracing approach allows to estimate the fraction of N₂O derived from bD, without overlapping nD (e.g., Deppe et al., 2017). Hence, by combining the N₂O isotopocule method and the ¹⁵N tracing approach, it is possible to assess potential contributions of not well-studied microbial processes such as nD or cND in N₂O formation. Furthermore, we used two types of ¹⁵N label, i.e., ¹⁵NH₄⁺ and ¹⁵NO₃⁻, to determine the contribution of each N form in the emitted N₂O and to gain additional insights into N transformation processes. In our study, we focused on rock wool hydroponics and used tomato plants as a model, as the use of rock wool substrate is widespread in modern production greenhouses (Dannehl et al., 2015; Savvas and Gruda, 2018) and tomato is the most important vegetable crop worldwide (Schwarz et al., 2014). We conducted two sampling campaigns: (i) at the beginning of flowering and (ii) during fruit ripening, at which we expected different N₂O emission rates. In previous studies with rock wool substrate, higher N₂O emissions were found during tomato fruit ripening compared to earlier plant stages (Hashida et al., 2014; Karlowsky et al., 2021), and were attributed to shifts in plant physiology.

Overall, our aim was to better understand which microbial processes contribute to N₂O emission from hydroponic systems to enable tailored mitigation measures. We hypothesized that bD is the main source of N₂O emissions from hydroponic tomato cultivation on rock wool, and that NO₃⁻ is contributing to a higher share to N₂O emissions than NH₄⁺. Furthermore, we assumed that most of the applied ¹⁵N tracer can be recovered in the labeled nutrient solution, plant biomass, and gaseous N emissions in a hydroponic system with inert rock wool substrate.

2. Materials and methods

2.1. Experimental setup and hydroponic tomato cultivation

The experiment took place in an experimental glasshouse consisting of multiple heated cabins, each with a size of 64 m² and a roof top height of 4 m. Two of these cabins were used for this study, cabin no. 7 for pre-cultivating tomato plants (Solanum lycopersicum cv. ‘Cheramy F1’) and cabin no. 5 for conducting the experiment. Temperature in the cabins was set to 20/18°C (day/night), and roof top ventilation was opened at temperatures above 23/20°C (day/night). Shading was done automatically at photosynthetically active radiation (PAR) values above 900 μmol m⁻² s⁻¹ and artificial lighting was applied between 5:00 and 12:00 CET, if PAR values were below 180 μmol m⁻² s⁻¹. Air temperature and humidity in the cabins as well as roof top PAR were continuously monitored by a climate computer (Supplementary Figure S1). Tomato plants were sown on 26th July 2021 and after germination in moistened sand, 64 seedlings were transplanted into pre-weighed rock wool cubes (10 × 10 × 6.5 cm; Grodan B.V., Roermond, Netherlands) for further cultivation. On 2nd September each two planted rock wool cubes were put on one rock wool slab (100 × 20 × 7.5 cm; Grodan Vital, Grodan B.V., Roermond, Netherlands) at a distance of 50 cm. One-half of the planted rock wool slabs were installed in eight hydroponic units with elevated gutters in cabin no. 5, which included separate fertigation systems and were later used for the ¹⁵N labeling. The other half was further cultivated in cabin no. 7 in four gutters on the ground, which shared one fertigation system. In both cases, the collected drain solution (i.e., leachate) was re-used and mixed with fresh nutrient solution in storage tanks as needed (closed hydroponic system with re-circulating nutrient solution). The nutrient solution from the storage tanks was supplied to plants via pumps, PE tubes, and drippers inserted into the rock wool cubes. The tomato plants were supplied with a custom-made nutrient solution modified after the recipe of de Kreij et al. (2003), which had a high NO₃⁻ to NH₄⁺ ratio (~20:1) that was found optimal for tomato cultivation. Macro and micro nutrients were dissolved in de-ionized water targeting a pH of 5.6 and an electrical conductivity (EC) of 2 mS cm⁻¹. The pH and EC values in the storage tanks were regularly monitored (Supplementary Figure S2). Tomato seedlings were supplied with an N concentration of 361 mg L⁻¹ at the beginning (starter solution; 338 mg L⁻¹ NO₃⁻-N and 23 mg L⁻¹ NH₄⁺-N). After the development of the 5th truss and the first green fruits on, from 4th October, the N concentration in the nutrient solution was reduced to 165 mg L⁻¹ (refill solution; 151 mg L⁻¹ NO₃⁻-N and 14 mg L⁻¹ NH₄⁺-N). The composition of the different nutrient solutions used in this study can be found in Supplementary Table S1. Each hydroponic unit in cabin no. 5 consisted of a 4 m gutter in which three rock wool slabs, two with plants and one unplanted, were placed and a nutrient solution storage tank filled up to approximately 40 L (Supplementary Figure S3). Two sampling periods were selected according to expected differences in plant N uptake and associated assimilate distribution in the root-shoot system, representing high growth and N uptake rates during early development and a more balanced assimilate distribution during fruit ripening. The first sampling and ¹⁵N labeling campaign were performed on 22nd and 23rd September, when the tomato plants developed the 3rd truss and first flowers. Subsequently, the 16 planted rock wool slabs (32 plants) in cabin no. 5 were completely removed (destructive sampling, described below) and replaced by the other 16 planted rock wool slabs pre-cultivated in cabin no. 7 on 24th September. The eight unplanted rock wool slabs were also exchanged with fresh rock wool slabs. To avoid carryover of ¹⁵N label, the hydroponic gutters were covered with plastic film below the rock wool slabs until 23rd September to reduce contact with the ¹⁵N-enriched nutrient solution. Both, the gutters and pumps for nutrient solution, were thoroughly cleaned with a detergent/disinfectant (MENNO Florades^®, MENNO CHEMIE-VERTRIEB GMBH, Langer Kamp, Germany) before installing the unlabeled plants and rock wool slabs. Furthermore, the storage tanks and the tubing as well as the drippers for nutrient solution were completely replaced with new material. To ensure the supply of further growing plants with water and nutrients, larger storage tanks were used (Supplementary Figure S4) and filled up to approximately 200 l. The experiment ended with the second sampling and ¹⁵N labeling campaign on 3rd and 4th November, when the tomato plants developed the 8th truss and the first fruits were ripe.

2.2. Gas flux measurements

For measuring the gas fluxes, the closed chamber method as described by Karlowsky et al. (2021) was used. Acrylic glass chambers with two small openings for plant stems were fitted around the rock wool slabs (planted and unplanted) and sealed with foam rubber to obtain a closed headspace with a volume of approximately 16 l (Supplementary Figure S5). Over a period of 1 hour after closing, four gas samples (each 30 ml) were taken in 20 min intervals with a 30 ml syringe through a sampling port on top of the chamber. The gas samples were transferred to 20 ml glass vials with silicone/PTFE septa (type N17, MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany) for transport and were analyzed on the same day by a gas chromatograph (GC 2010 Plus, Shimadzu Corporation, Kyoto, Japan) equipped with an electron capture detector (ECD) for N₂O. The measured concentrations in μmol mol⁻¹ were converted to μmol m⁻³ by applying the ideal gas law, including a correction for the temperature at the time of sampling. Afterward, gas fluxes were calculated using the R package “gasfluxes” [version 0.4–4; (Fuss et al., 2020)] by robust linear regression (except one case with only 3 time points, for which standard linear regression had to be used). Input variables used were gas concentration (μmol m⁻³), chamber volume (m³), time after closing the chamber (h), and area covered (m²). The latter was set to 1 m² assuming a typical density of greenhouse-cultivated tomato plants of 2 plants m⁻². The resulting gas fluxes in μmol m⁻² h⁻¹ were further converted to g ha⁻¹ d⁻¹ based on molar masses.

2.3. Sampling and ¹⁵N labeling

Natural abundance samples were taken on 22nd September and 3rd November shortly before the ¹⁵N labeling from each hydroponic unit in cabin no. 5 (from here on called “experimental unit”). These included plant samples, nutrient solution samples, and gas samples from planted rock wool slabs. For the latter, 140 ml of air was collected from the headspace of rock wool substrate with a syringe at the end of gas flux measurements after 1 h of N₂O enrichment in the closed chambers. The gas samples were transferred into 120 ml crimp-cap glass vials closed with gray butyl septa (type ND20, IVA Analysentechnik GmbH & Co. KG, Meerbusch, Germany) for later stable isotope analysis. To determine natural abundance δ¹⁵N values of plants, the tips (first three leaflets) of 2–3 fully developed leaves from one plant in each experimental unit were sampled and dried at 80°C for at least 48 h. Approximately 15 ml of nutrient solution (mixture with leachates) was sampled from the storage tank of each experimental unit and then stored at −20°C for later δ¹⁵N analyses. In addition, three samples of de-ionized water were taken to determine the natural abundance δ¹⁸O values of the nutrient solution water.

On both dates, the ¹⁵N labeling took place directly after the natural abundance sampling at approximately 12:00 pm CET. The remaining nutrient solution in the experimental units was removed as far as possible and 15 l of ¹⁵N-labeled nutrient solution was added in the storage tanks of each unit. In a randomized way, four units received a nutrient solution with ¹⁵N-enriched NH₄⁺ (¹⁵NH₄⁺) and four units received a nutrient solution with ¹⁵N-enriched NO₃⁻ (¹⁵NO₃⁻). This was done by adding ammonium nitrate (NH₄NO₃; SIGMA-ALDRICH, Saint Louis, MO, United States) with 10.5/11 atom-% ¹⁵N (¹⁵NH₄⁺/¹⁵NO₃⁻) as only N source. The composition of the nutrient solution used for the ¹⁵N labeling can also be found in Supplementary Table S1. In total, 115 mg of ¹⁵N was applied to each ¹⁵NH₄⁺ unit and 120 mg of ¹⁵N to each ¹⁵NO₃⁻ unit (3.1 g NH₄NO₃ per unit), yielding an N concentration of 146 mg L⁻¹ (comparable to the standard refill solution). To distribute the ¹⁵N label in the hydroponic system, drip fertigation was run continuously for 30 min after adding the ¹⁵N labeled nutrient solution to the experimental units. After 4 h, a first sampling to determine the ¹⁵N enrichment in plant, nutrient solution and gas samples took place. The sampling was done analogously to the natural abundance sampling, including the determination of gas flux rates and the collection of gas samples for isotopic analyses as well as leaf and nutrient solution samples. Following the same scheme, the last sampling took place 24 h after the labeling. This time, also samples from the tomato stems, roots and fruits were taken. From the middle of the tomato plant ca., 10 cm of the stem was cut. Around 0.5 g of fresh roots was sampled from the interface of rock wool cubes and rock wool slabs, where a dense root net allowed to obtain root material without rock wool fibers. Root samples were washed in de-ionized water and dried with lint-free cellulose wipes to remove the ¹⁵N label from adhering nutrient solution. During the second sampling campaign, each three green fruits from different positions (top, mid, and bottom) of one plant per experimental unit were sampled. All plant samples were dried for a minimum of 48 h at 80°C before later processing for analysis. Different plants were used for obtaining plant material before labeling, 4 h after labeling, and 24 h after labeling in order to minimize sampling effects on ¹⁵N uptake. Gas samplings for stable isotope analysis always took place on the rock wool slab in the middle of each experimental unit, from which plant samples were taken only after the last gas sampling (24 h after labeling). On the unplanted rock wool slabs, additional gas flux measurements took place shortly before the 24 h sampling to determine the N₂O emission potential from re-circulated nutrient solution with leachate and therein contained organic carbon.

2.4. Analyses on nutrient solution, plant, and gas samples

The concentrations of NO₃⁻ and NH₄⁺ [mg N L⁻¹] were determined using flow injection analysis with photometric detection (FIAmodula; MLE GmbH, Dresden, Germany). Measurements of δ¹⁸O values in water samples were done by TC/EA coupled to a Delta V plus IRMS (Thermo Finnigan, Bremen, Germany) via a ConFlo IV interface. The δ¹⁵N values of NH₄⁺ and NO₃⁻ were determined according to Dyckmans et al. (2021) using a sample preparation unit for inorganic nitrogen (SPIN) coupled to a membrane inlet isotope ratio mass spectrometer (MIRMS; Delta plus; Thermo Finnigan) via a ConFlo III interface. Additional nutrient solution samples taken one day after the labeling were analyzed for their dissolved organic carbon content (DOC) using a liquiTOC analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany). Dried plant samples were transferred into 20 ml HDPE vials (Zinsser Analytic GmbH, Eschborn, Germany) and ground to a fine powder using a steel ball mill (MM400; RETSCH GmbH, Haan, Germany). Plant samples were analyzed for total N content (N_t) and their δ¹⁵N values using an Elemental Analyzer (EA) Flash 2000 (Thermo Fisher Scientific, Bremen, Germany), coupled with a Delta V isotope ratio mass spectrometer via a ConFlo IV interface (Thermo Fisher Scientific, Bremen, Germany). Data were normalized to the international scale for atmospheric nitrogen, by analysis of the international standards USGS40 and USGS41 (L-glutamic acid). Gas samples were analyzed for N₂O isotopocules (δ¹⁵N_N2O, δ¹⁸O_N2O) using a Delta V Isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany), coupled to an automatic preparation system with Precon plus Trace GC Isolink (Thermo Scientific, Bremen, Germany). In this setup, N₂O was pre-concentrated, separated, and purified, and afterward m/z 44, 45, and 46 of the intact N₂O⁺ ions as well as m/z 30 and 31 of NO⁺ fragment ions were determined (Lewicka-Szczebak et al., 2014). All measured delta values (δ) were expressed in permil (‰) deviation from the ¹⁵N/¹⁴N and ¹⁸O/¹⁶O ratios of the international reference standards (i.e., atmospheric N₂ and Vienna Standard Mean Ocean Water (VSMOW), respectively).

2.5. Data processing and calculations

Data from the analysis of natural abundance gas samples were evaluated for δ¹⁵Nα (δ¹⁵N of the central N position of the N₂O molecule), δ¹⁵Nβ (δ¹⁵N of the peripheral N position of the N₂O), and δ¹⁸O according to Toyoda and Yoshida (1999) and Röckmann et al. (2003). The ¹⁵N site preference (δ¹⁵N^SP) was defined as the difference of δ¹⁵Nα and δ¹⁵Nβ. The δ¹⁸O values of N₂O depend on δ¹⁸O values of precursors, i.e., for denitrification to >80% on H₂O-O of the nutrient solution (Lewicka-Szczebak et al., 2016). Therefore, δ¹⁸O values of the emitted N₂O (δ¹⁸O_N2O) were corrected for the δ¹⁸O values measured in the de-ionized water (δ¹⁸O_H2O) and expressed as δ¹⁸O_N2O/H2O values:

$\begin{array}{ll}{\delta }^{18}{\mathrm{O}}_{\mathrm{N}2\mathrm{O}/\mathrm{H}2\mathrm{O}}={\delta }^{18}{\mathrm{O}}_{\mathrm{N}2\mathrm{O}}-{\delta }^{18}{\mathrm{O}}_{\mathrm{H}2\mathrm{O}}& (1)\end{array}$

In the case of nitrification, the δ¹⁸O_N2O values depend on atmospheric oxygen (O₂) as a precursor (Kool et al., 2007). In contrast to bulk δ¹⁵N_N2O, δ¹⁵N^SP is known to be independent from source processes. During chamber air sampling, the collected N₂O was a mixture of atmospheric and substrate-emitted N₂O. Thus, δ values of substrate-emitted N₂O were corrected using a basic isotope mixing model according to Well et al. (2006). To calculate the contribution of N₂O production pathways and N₂O reduction to N₂, the isotopocule mapping approach based on δ¹⁵N^SP_N2O and δ¹⁸O_N2O values was applied (Lewicka-Szczebak et al., 2017; Buchen et al., 2018). For the mapping approach, literature values for δ¹⁸O and δ¹⁵N^SP_N2O of bD, fD, nD, and Ni were used as proposed by Yu et al. (2020) and Lewicka-Szczebak et al. (2020). To account for differences in oxygen precursors between denitrification and Ni, the literature values for δ¹⁸O_N2O of bD, fD, and nD were adjusted by the addition of δ¹⁸O_H2O (Lewicka-Szczebak et al., 2020). Based on the sample position in the map, the contribution of bD and/or nD, Ni, and fD was calculated based on mixing equations, while the contribution of N₂O reduction to N₂ was calculated from the Rayleigh equation. All calculations were done as described in detail by Buchen et al. (2018) and Zaman et al. (2021) (Chapter 7: “Isotopic Techniques to Measure N₂O, N₂ and Their Sources). Two possible cases of N₂O mixing and reduction were assumed: (i) N₂O, which is produced by bD is first partially reduced to N₂, followed by mixing of the residual N₂O with N₂O from other pathways or (ii) N₂O produced by various pathways is first mixed and then reduced to N₂. A detailed description is given in the supplement of Wu et al. (2019). Five samples from sampling 1 and four samples from sampling 2 with a low fraction of substrate-derived N₂O were excluded from the data analyses because the uncertainty in substrate-derived δ values increases exponentially as sample and atmospheric N₂O concentrations converge. Similar to Buchen et al. (2018), a threshold was used for the minimum difference between sample and atmospheric N₂O concentrations, which was determined based on measured N₂O concentrations in ambient air during the sampling. For sampling 1, the threshold was 337 ppb and for sampling 2, it was 359 ppb (65 ppb above the ambient air N₂O concentration). This was supported by a Gaussian error propagation, with the threshold limiting the propagated errors of δ¹⁵N^SP_N2O and δ¹⁸O_N2O to <6‰ and < 5‰, respectively.

Data from the analysis of ¹⁵N-enriched gas samples were only evaluated for bulk δ¹⁵N_N2O. For further calculations, δ¹⁵N values were converted to atom-%_15N to express the ¹⁵N enrichment:

$\begin{array}{ll}\mathit{\text{atom-}}{\%}_{{}^{15}N}=\frac{100\%}{\frac{1}{\left(\frac{\delta {}^{15}N}{1000‰}+1\right)\times R_{STD}}+1}& (2)\end{array}$

with R_STD being the isotopic ratio (¹⁵N/¹⁴N = 0.0036765) of atmospheric nitrogen. Calculations of the contributions of N₂O originating from the labeled and non-labeled pools were based on the non-equilibrium distribution of N₂O isotopocules, as described by Spott et al. (2006) and Bergsma et al. (2001). For labeling with ¹⁵NO₃⁻, this approach directly determines the ¹⁵N enrichment of the labeled N pool producing N₂O (ap_N2O) and the fraction of N₂O derived from that pool. Considering, the fraction of atmospheric N₂O in the samples, the fraction of NO₃⁻-derived N₂O in the emitted N₂O (f_PN2O) can be calculated. A detailed procedure is given in Deppe et al. (2017). However, due to the experimental setup, labeled N₂O could originate from two pools (NO₃⁻, NH₄⁺, or a mixture of both pools). Thus, for labeling with ¹⁵NH₄⁺, f_PN2O was estimated based on the ¹⁵N atom fraction of emitted N₂O (¹⁵a_N2O) using a mixing equation:

$\begin{array}{ll}{f}_{PN2O}=\frac{{}^{15}a{}_{N2O}-{}^{15}aN{H}_4^{+}}{{}^{15}aN{O}_3^{-}-{}^{15}aN{H}_4^{+}}& (3)\end{array}$

with ¹⁵aNO₃⁻ being the ¹⁵N enrichment of the NO₃⁻ pool and ¹⁵aNH₄⁺ being the ¹⁵N enrichment of the NH₄⁺ pool (cf. Eq. 2). The N₂O flux from the NO₃⁻ pool (NO₃⁻-derived N₂O) was calculated from f_PN2O by ordinary linear regression using the measured N₂O concentrations at t0 and after 1 h of chamber closure to determine the total N₂O flux (total N₂O), assuming that the increase in the N₂O emitted from the ¹⁵N-labeled pool was also linear as shown for the emission of total N₂O (Buchen et al., 2016). The N₂O flux from the NH₄⁺ pool (NH₄⁺-derived N₂O) was calculated analogously based on the fraction of NH₄⁺-derived N₂O in the emitted N₂O (f_NH4), which was deduced from f_PN2O (f_NH4 = 1 – f_PN2O). Thus, the NH₄⁺-derived N₂O was calculated as the difference between total N₂O and NO₃⁻-derived N₂O.

2.6. Calculation of excess ¹⁵N and ¹⁵N mass balance

To determine the amount of ¹⁵N tracer, which was recovered in the different pools 4 and 24 h after the labeling (excess ¹⁵N), atom-%_15N values were used to calculate atom-% ¹⁵N excess (APE):

$\begin{array}{ll}APE=\mathit{\text{atom-}}{\%}_{{}^{15}N,\mathit{\text{labeled}}}-\mathit{\text{atom-}}{\%}_{{}^{15}N,\mathit{\text{natural abundanc}}e}& (4)\end{array}$

with atom-%_15N,labeled being the atom-%_15N values of labeled samples and atom-%_{15N,natural abundance} being the atom-%_15N values of natural abundance samples. Afterward, excess ¹⁵N [mg ¹⁵N unit⁻¹] for each pool was calculated:

$\begin{array}{ll}\mathit{\text{excess}}{}^{15}N=\frac{APE}{100\%}\times N_{\mathit{\text{pool}}}& (5)\end{array}$

with N_pool being the N amount in each pool [mg N unit⁻¹] at the time of sampling (4/24 h after labeling). The N_pool values for plant biomass were calculated by multiplying the measured dry weight [g] of shoots (leaves + stems), roots and fruits per unit with their N_t content [g N g_{dry weight}⁻¹]. The N_pool values for NO₃⁻-N and NH₄⁺-N from the nutrient solution were calculated by multiplying the measured N concentrations [mg N L⁻¹] with the total volume of nutrient solution per unit [L]. The latter was a mixture of nutrient solution added for the labeling and remaining (unlabeled) nutrient solution in the rock wool substrate. The total volume of the nutrient solution was estimated based on the dilution of NH₄⁺-N concentrations from the labeled nutrient solution (73 mg N L⁻¹ in 15 l) at the 4 h sampling point, assuming that NH₄⁺-N concentrations in the unlabeled nutrient solutions were negligible (measured concentrations in natural abundance samples <2.5 mg N L⁻¹ at first sampling campaign and <7 mg N L⁻¹ at second sampling campaign) and that the N_t content as well as composition in the mixed nutrient solution did not substantially change during the 4 h. For the calculation of excess ¹⁵N, two neighboring units were excluded from the second sampling campaign, because of a spillover of labeled nutrient solution between these units. The N_pool values for N₂O were calculated from the measured gas flux rates [mg N h⁻¹] of planted and unplanted rock wool slabs. For the planted rock wool slabs, cumulative N₂O emissions [mg N] were calculated by linear integration between the natural abundance (0 h), 4 h, and 24 h samplings, and summation of hourly gas fluxes. For unplanted rock wool slabs, constant N₂O emission rates were assumed and used to calculate cumulative N₂O emissions, as they were not affected by plant activity. For calculating the N_pool value per unit, cumulative N₂O emissions from planted rock wool slabs were multiplied by 2 (two planted slabs per unit) and the cumulative N₂O emissions from unplanted slabs (one per unit) were added. Finally, the excess ¹⁵N values from the different pools were summed up to obtain the total amount of ¹⁵N recovered from the labeling (¹⁵N_total) and the ¹⁵N recovery rate [%] was calculated:

$\begin{array}{ll}{}^{15}N\mathit{\text{recovery rate}}=\frac{{}^{15}N{}_{\mathit{\text{total}}}}{{}^{15}N{}_{\mathit{\text{label}}}}\times 100\%& (6)\end{array}$

with ¹⁵N_label being the amount of ¹⁵N tracer [mg ¹⁵N unit⁻¹] added during the labeling.

2.7. Statistical analyses

All statistical analyses were done using the R software (version 4.2.0). Linear mixed-effects models were done using the R package ‘lme4’ (version 1.1–29), including the effects of individual hydroponic units as random intercept. Post-hoc tests on linear mixed-effects models were done using the R package “emmeans” (version 1.7.4–1), applying the Holm-Bonferroni correction method for multiple comparisons. If necessary, data were log- or square root-transformed prior to analysis to fulfill the requirements of normality and variance homogeneity.

3. Results

3.1. N₂O flux, isotopocule, and ¹⁵N tracer analyses

The N₂O flux measurements from this study are summarized in Table 1. In general, all fluxes were in the same range, except for the measurement 24 h after labeling during the first sampling, which was significantly (p < 0.05) higher than the other measurements. There was no significant difference between planted and unplanted rock wool slabs from the same sampling campaign. The trend to higher N₂O emissions from unplanted substrate during sampling 2 was reflected by higher DOC contents in the nutrient solution compared to sampling 1 (Table 1).

TABLE 1

Table 1. N₂O fluxes (determined by gas chromatography) and dissolved organic carbon (DOC) concentrations at the two sampling campaigns (sampling 1, S1; sampling 2, S2).

Results from isotopic analyses of N₂O are shown in Figure 1 as a δ¹⁵N^SP_N2O/δ¹⁸O_N2O map. The δ values from both samplings clearly scatter around the reduction line of N₂O derived from bD, indicating that either bD or nD or a mixture of both was the main source of N₂O. Moreover, the increased δ¹⁵N^SP_N2O and δ¹⁸O_N2O values compared to the literature value for bD indicate that a high share of N₂O was reduced before emitted to the atmosphere. Altogether, the differences in isotopic results between the first and the second sampling campaign were negligible (Table 2). Depending on which scenario (mixing of bD and fD or bD and Ni) and case (first reduction than mixing or first mixing than reduction) was assumed, the fraction of bD varied between 0.85 and 0.90, while the N₂O/(N₂O + N₂) ratio of bD (r_N2O) varied between 0.08 and 0.14. In consequence, the calculated N₂ fluxes were between six to ten times higher than the measured N₂O fluxes.

FIGURE 1

Figure 1. Results from N₂O isotopocule analysis of natural abundance ¹⁵N gas samples illustrated as δ¹⁵N^SP_N2O/δ¹⁸O_N2O map. The vertical axis shows the ¹⁵N site preference of N₂O (δ¹⁵N^SP_N2O) and the horizontal axis the abundance of the ¹⁸O isotope in the N₂O molecules (δ¹⁸O_N2O). Sample δ¹⁸O_N2O values were corrected for the ¹⁸O composition of water from the nutrient solution (δ¹⁸O_N2O/H2O) as described in Eq. 1. Closed circles represent the measurement-derived values and the corresponding error bars the estimated uncertainty. Other symbols indicate literature values as compiled in Lewicka-Szczebak et al. (2020) for N₂O produced from different microbial processes and the surrounding boxes reflect their variation (based on SD): Ni, nitrification (Yoshida, 1988; Sutka et al., 2006; Mandernack et al., 2009; Frame and Casciotti, 2010); fD, fungal denitrification (Sutka et al., 2008; Rohe et al., 2014; Maeda et al., 2015; Rohe et al., 2017); nD, nitrifier denitrification (Sutka et al., 2006; Frame and Casciotti, 2010); and bD, bacterial denitrification (Barford et al., 1999; Toyoda et al., 2005; Sutka et al., 2006; Lewicka-Szczebak et al., 2014, 2016; Rohe et al., 2017). According to Lewicka-Szczebak et al. (2020), the literature values of bD, fD and nD were adjusted by addition of the δ¹⁸O of water (−8.5‰) measured in this study to display expected endmember ranges. The solid line indicates the isotopic shift of N₂O due to fractionation from the partial reduction of N₂O to N₂ by bD (Menyailo and Hungate, 2006; Ostrom et al., 2007; Jinuntuya-Nortman et al., 2008; Well and Flessa, 2009; Lewicka-Szczebak et al., 2014, 2015) and is shown for theoretical r_N2O values of 1 to 0.05. The dotted and the dashed lines represent expected values for different mixing ratios of N₂O from bD and fD (bD-fD line) and N₂O from bD and Ni (bD-Ni line), respectively.

TABLE 2

Table 2. Measured N₂O flux, estimated fraction of N₂O from bacterial denitrification (f_bD), estimated N₂O/(N₂O + N₂) ratio of denitrification (r_N2O), and estimated N₂ flux for different mixing scenarios (bacterial denitrification and fungal denitrification, bD-fD; bacterial denitrification and nitrification, bD-Ni) and cases (reduction of N₂O from denitrification followed by mixing with N₂O from other sources, red-mix; mixing of N₂O from denitrification and other source followed by N₂O reduction, mix-red).

Although the same amounts of NO₃⁻-N and NH₄⁺-N were added in the form of NH₄NO₃ during each ¹⁵N labeling, NO₃- concentrations were clearly higher than NH₄+ concentrations in the nutrient solution after labeling (Table 3). This indicated that a significant amount of unlabeled nutrient solution with a high NO₃⁻ to NH₄⁺ ratio was still present in the rock wool substrate during ¹⁵N labeling. Regardless of the higher dilution of ¹⁵NO₃⁻ label (Table 3; Supplementary Figure S6), the ¹⁵N tracer could be detected in the emitted N₂O independent of the applied form (¹⁵NH₄⁺ or ¹⁵NO₃⁻). The ¹⁵a_N2O values mirrored the ¹⁵N enrichments of the labeled NO₃⁻ and NH₄⁺ pools, with higher values in of ¹⁵NH₄⁺-labeled units compared to ¹⁵NO₃⁻-labeled units (Supplementary Figure S6). The label dilution was considered for calculating NO₃⁻-derived N₂O and NH₄⁺-derived N₂O. The NO₃⁻-derived N₂O (Figures 2A,B) reflected the N₂O emission rates measured by GC (Table 1), with highest values found 24 h after the first labeling. There was no clear difference in NO₃⁻-derived N₂O between the ¹⁵NH₄⁺ and ¹⁵NO₃⁻ labels. In general, the NH₄⁺-derived N₂O values (Figures 2C,D) were lower than the NO₃⁻-derived N₂O values, but also followed the dynamics of N₂O emission rates measured by GC. Notably, NH₄⁺-derived N₂O was higher for ¹⁵NO₃⁻-labeled units compared to ¹⁵NH₄⁺-labeled units during sampling 2. Consequently, the calculated average f_PN2O values varied from 0.4 to 0.9 between the applied label forms, sampling times, and sampling campaigns (Figures 2C,D). During both sampling campaigns, an increase of f_PN2O from 4 h to 24 h after labeling was present for the ¹⁵NO₃⁻-labeled units, while there was no effect of sampling time for the ¹⁵NH₄⁺-labeled units. The latter showed higher f_PN2O values during the second sampling campaign, which was also significantly higher than for the ¹⁵NO₃⁻-labeled units at 4 h after labeling.

TABLE 3

Table 3. Concentrations and ¹⁵N-enrichment of dissolved ammonium and nitrate in the nutrient solution during the two sampling campaigns, including samples taken before ¹⁵N labeling (T0) and 4/24 h afterward (T4/T24).

FIGURE 2

Figure 2. NO₃⁻-derived N₂O fluxes (A,B), NH₄⁺-derived N₂O fluxes (C,D), and the estimated share of NO₃⁻-derived N₂O fluxes [f_PN2O; (E,F)]. Bars show the mean of n = 4 replicates and error bars the corresponding SD. Small letters indicate levels of significance for differences between label and sampling with p < 0.05 from linear mixed-effects models and Tukey post-hoc tests.

3.2. Recovery of ¹⁵N tracer in different pools

The natural abundance δ¹⁵N values from both samplings were equal (leaves) or slightly lower (NH₄⁺, NO₃⁻ and N₂O) at the second sampling, indicating that no carryover of ¹⁵N label occurred from the first sampling. The amount of ¹⁵N tracer from the ¹⁵N-enriched NH₄NO₃ added during the labelings that was recovered in different pools (dissolved NH₄⁺ and NO₃⁻, N₂O, plant biomass) was calculated as excess ¹⁵N (¹⁵N_exc). At both samplings, most of the ¹⁵N label remained in its original form after 24 h, i.e., as dissolved NH₄⁺ and NO₃⁻ (Table 4). There was a notable increase of ¹⁵N_exc of dissolved NO₃⁻ in the ¹⁵NH₄⁺-labeled units, indicating the conversion of NH₄⁺ to NO₃⁻ by Ni (up to 2% of added label during sampling 1). On the other side, the ¹⁵N_exc of dissolved NH₄⁺ in the ¹⁵NO₃⁻-labeled units was comparably low (at maximum 0.3% of added label during sampling 1). The ¹⁵N_exc of N₂O strongly differed between the two samplings, with up to 20 times higher values at sampling 1, reflecting the APE values of N₂O (Supplementary Figure S7). Despite the higher dilution of ¹⁵N tracer in the NO₃⁻ pool (Table 3) and the resulting lower ¹⁵N enrichments in the ¹⁵NO₃⁻-labeled units compared to ¹⁵NH₄⁺-labeled units (Supplementary Figure S6), there were no significant differences between the label types regarding the amount of ¹⁵N tracer found in N₂O, as shown by the ¹⁵N_exc values (Table 4). In all cases, the ¹⁵N_exc of total plant biomass was higher than the ¹⁵N_exc of N₂O. The highest plant ¹⁵N uptake was observed during the second sampling in ¹⁵NH₄⁺-labeled units. Irrespective of the generally higher ¹⁵N-enrichment of roots (Supplementary Table S2), most ¹⁵N tracer was found in shoots (i.e., the sum of stem leaf biomass; Table 4), as a consequence of the biomass difference (root to shoot ratio of 0.23). Only marginal amounts of ¹⁵N tracer were found in tomato fruits during sampling 2. Overall, the majority of ¹⁵N added during labelings was recovered in the studied pools, with the calculated ¹⁵N recovery rates varying around 100%.

TABLE 4

Table 4. Excess ¹⁵N (¹⁵N_exc) found in different pools 24 h after labeling with ¹⁵NH₄⁺ and ¹⁵NO₃⁻, total recovered ¹⁵N and recovery rate of ¹⁵N tracer from the labeling.

4. Discussion

In this study, we applied the N₂O isotopocule and ¹⁵N tracing approaches to better understand the sources of N₂O emission from hydroponic vegetable production systems, using tomato cultivation on rock wool substrate as a model. Furthermore, in our study, we determined r_N2O using the isotopocule mapping method (Lewicka-Szczebak et al., 2017), which had been shown to be in good agreement with the ¹⁵N gas flux method (Buchen et al., 2018; Lewicka-Szczebak et al., 2020). Therefore, for hydroponic systems, we determined this ratio for the first using an appropriate method.

As we hypothesized, the results from both N₂O isotope analyses (non-labeled and ¹⁵N-labeled) point to bD as main source of N₂O emissions from the hydroponic units. The scattering of the values around the reduction line of bD in the mapping approach of the N₂O isotopocules (Figure 1) suggests that most of the N₂O was produced by bD. Unfortunately, nD cannot be clearly separated from bD by the N₂O isotopocule mapping approach (Lewicka-Szczebak et al., 2017), due to the overlap of endmember values (i.e., theoretical values determined from literature values of pure cultures and the isotopic composition of water and N substrates). Thus, the calculated f_bD could actually be a mixture of bD and nD. The same is true for the fraction of Ni in N₂O emission (f_Ni), which cannot be clearly separated from the fraction of fD (f_fD) in the mapping approach. However, a mixed fraction (f_Ni/fD = 1 – f_bD) can be calculated, as previously done by Buchen et al. (2018). Depending on the mapping scenario and sampling campaign, the f_Ni/fD values vary between 0.10 and 0.15 in our study. In consequence, the contribution of fD and/or Ni seems small under typical tomato growing conditions in rock wool hydroponics with low NH₄⁺ supply. For better distinction of bD, we used the ¹⁵N tracing approach to determine the fraction of NO₃⁻-derived N₂O fluxes, i.e., f_PN2O. While f_PN2O can principally also include contributions from fD, we assume its impact was minor as shown by the isotopocule map (Figure 2). Therefore we assume f_PN2O is equivalent to f_bD from the isotopocule mapping approach but does not include N₂O fluxes from nD. Although the f_PN2O values are relatively variable (Figures 2E,F), they generally show that bD was the main source of N₂O emissions, even under increased NH₄⁺ supply. Hence the results from N₂O isotope analysis and ¹⁵N tracing were in good accordance with each other. On the other hand, the results from the ¹⁵N-labeling also show that a large part of N₂O can be formed from NH₄⁺ (Figures 2C,D), suggesting processes other than denitrification of added NO₃⁻ (Firestone and Davidson, 1989). Possibly, the increase of the NH₄⁺ concentration in the nutrient solution used for ¹⁵N-labeling compared to the non-labeled nutrient solution could have increased Ni and the associated N₂O formation from NH₄⁺. This is supported by the slight ¹⁵N-enrichment of NO₃⁻ found in units labeled with ¹⁵NH₄⁺ (Table 4), indicating the presence of Ni. Notably, the average f_bD values of ~0.87 from N₂O isotopocule analysis (Table 2) were higher than the average f_PN2O values of ~0.68 from ¹⁵N tracing (Figure 2). Assuming that microbial activities did no significantly change after adding the NH₄⁺-rich ¹⁵N label, we hypothesize that the observed difference in f_bD and f_PN2O values is due to microbial processes other than Ni that are associated with the release of N₂O from NH₄⁺.

Besides the conversion of hydroxyl amine (NH₂OH) to N₂O during Ni, there are several known pathways that explain the production of N₂O derived from NH₄⁺, in particular nD and cND (Baggs, 2011). Wrage-Mönnig et al. (2018) argue in their review that nD can be the predominant source of N₂O emissions under certain conditions. For example, this includes “environments with fluctuating aerobic-anaerobic conditions”, which are likely to occur in hydroponic systems with regular irrigation intervals (Schröder and Lieth, 2002). In contrast, Bakken and Frostegard (2017) fundamentally disagree with the concept of nD, based on the preferential electron flow in nitrifiers, and rather suggest that it is cND that accounts for the observations after all. In this sense, the O₂ consumption by Ni could lead to anoxic conditions facilitating bD (Zhu et al., 2015). Additionally, a process that also needs to be taken into account is co-denitrification (coD), i.e., the formation of hybrid N₂O and N₂ molecules with each one N atom derived from the classical denitrification pathway (N species: nitrite, NO₂⁻; nitric oxide, NO) and one N atom from another N species such as NH₂OH or amino compounds (Spott et al., 2011). In our study, coD may have been stimulated by the increased NH₄⁺ availability after adding the nutrient solutions for ¹⁵N labeling. This is supported by the lower ap_N2O values compared to the ¹⁵aNO₃⁻ values found in ¹⁵NO₃⁻-labeled units (Supplementary Figures S6A,B,E,F; Spott and Stange, 2007), suggesting that part of the emitted N₂O was derived from non-labeled NH₄⁺. Albeit the use of NH₄⁺ in coD was found quite rarely and organic N sources are thus perceived as the main source for forming hybrid N₂O/N₂ molecules with NO₂⁻-N or NO-N (Spott et al., 2011). Therefore, the combined fraction of nD and cND (f_nD/cND) can be estimated from f_PN2O and f_bD as described by Deppe et al. (2017), i.e., by calculating the difference of f_bD and f_PN2O (f_nD/cND = f_bD – f_PN2O). Depending on the scenario for f_bD, the values of f_nD/cND vary between 0.40–0.48 at T4 and 0.09–0.24 at T24 for the ¹⁵NO₃⁻-labeled units during both sampling campaigns. For the ¹⁵NH₄⁺-labeled units, this comparison seems not appropriate because the estimated f_PN2O values were partially higher than f_bD values. This is probably due to the assumption used in Eq. 3, i.e., that the labeled pool (¹⁵NO₃⁻ and ¹⁵NH₄⁺) is the same as the active pool. In contrast, the f_PN2O values of ¹⁵NO₃⁻-labeled units were determined via the non-random distribution of N₂O isotopologues and delivered the fraction of the active labeled pool used for N₂O production, which is not necessarily identical to the bulk NO₃⁻ pool (Deppe et al., 2017; Zaman et al., 2021).

Notably, measured N₂O emissions from the experimental units we used were low compared to previous studies of hydroponic systems (Daum and Schenk, 1996a; Hashida et al., 2014; Karlowsky et al., 2021), which reported emission rates that were one to two orders of magnitude higher. The low N₂O emission rates could have been a result of unfavorable conditions for denitrifier activity, such as low organic carbon contents and/or high oxygen availability in the substrate (Morley and Baggs, 2010). The accumulation of organic carbon due to root exudation and root decay might be key to N₂O emissions from inert substrates like rock wool, as we found in a previous study a steep increase of N₂O emission rates after 5 months of tomato cultivation following a phase of low N₂O emission rates (Karlowsky et al., 2021). In this study, we found an increase of DOC in the re-circulating nutrient solution from sampling 1 to sampling 2, but this was not related to higher N₂O emissions. Here, the slightly acidic conditions (pH values <4.6; Supplementary Figure S2) during sampling 2 may have limited denitrification, considering that N emissions from denitrification typically decrease at low pH values (Daum and Schenk, 1998; Farquharson and Baldock, 2007), which is also associated with a higher r_N2O value (e.g., Liu et al., 2010), but this was only visible in trend (Table 2). In general, N₂O fluxes were highly variable (Table 1), with a trend to higher emissions from planted rock wool slabs compared to unplanted rock wool slabs, especially during sampling 1. Thus, our findings indicate that considerable N₂O emissions may also occur from re-circulated nutrient solution, e.g., in collection and storage tanks or bio-filtration/disinfection units. Although it is unclear to which extent the rock wool matrix with its high pore space volumes (Dannehl et al., 2015) and a large surface area for microbial biofilms (Brand and Wohanka, 2001) might have promoted N₂O emissions from the re-circulated nutrient solution.

In addition to the above-discussed findings, we performed a ¹⁵N mass balance to check the plausibility of r_N2O and the calculated N₂O and N₂ emissions from the mapping approach, and to gain more insights into N dynamics in the hydroponic units. Unfortunately, the proportion of applied ¹⁵N label recovered as N₂O strongly varied between the two samplings, which can be attributed to temporal fluctuations resulting in a peak of N₂O emission rates at 24 h after labeling during sampling 1. This peak probably led to an overestimation of cumulative N₂O fluxes, especially considering that N₂O emission rates are typically lower during nighttime when no fertigation is done (Daum and Schenk, 1998; Yoshihara et al., 2016; Karlowsky et al., 2021). Due to highly variable and generally very moderate N₂O emissions as well as the high variability of ¹⁵N excess in plant material, the ¹⁵N mass balance in our case proved to be too uncertain to validate the calculated gas fluxes from the isotopocule mapping approach. In general, the results of the ¹⁵N mass balance reflect the findings from the ¹⁵N tracing approach and show in addition that the majority of ¹⁵N tracer applied to the hydroponic units was recovered in the nutrient solution, plant biomass, and N₂O emissions after 24 h. However, since only short-term N dynamics are included in the ¹⁵N mass balance, N use efficiency cannot be calculated with these data.

5. Conclusion

The findings of our study clearly show that bD was the major source of N₂O emissions from hydroponic tomato cultivation on rock wool substrate, and that up to 90% of initially produced N₂O was reduced to N₂ before gas emission. The combined results of N₂O isotopocule analysis and ¹⁵N tracing suggest that other microbial processes related to N₂O formation from NH₄⁺ (i.e., Ni, nD, and cND) play only a moderate role. However, with the methods used, it was not possible to determine the individual contribution of each of these processes to the observed N₂O emissions. Furthermore, the involvement of fD and coD remains unclear, but seems less likely since organic matter is supplied only by plant roots in the rock wool substrate. Therefore, future studies are needed to better distinguish N₂O sources other than bD, possibly combing isotopic approaches with molecular genetic methods such as functional gene analysis. As we also found N₂O emissions from root-less rock wool substrate, potential N₂O emissions from drained nutrient solution should be further researched. Ultimately, on the basis of our study, measures to reduce denitrifier activity appear to be the most promising option to mitigate N₂O emissions and N losses from hydroponic cultivation.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

SK: conceptualization, investigation, formal analysis, and writing—original draft. CB-T: investigation, formal analysis, and writing—original draft. LO: investigation and formal analysis. DS: conceptualization and methodology. RW: methodology and writing—review and editing. All authors contributed to the article and approved the submitted and revised version.

Funding

This project is supported by the Federal Ministry of Food and Agriculture (BMEL) based on the decision of the Parliament of the Federal Republic of Germany via the Federal Office for Agriculture and Food (BLE) under the innovation support program (funding code 281B204116 for project “HydroN2O”).

Acknowledgments

We thank Gundula Aust and the gardener team at IGZ for setting up and maintaining the experiment in the greenhouse. Martina Heuer, Jennifer Gier und Ute Rieß at Thünen Institute for help during isotopic analyses. Jens Dyckmanns and his team at Göttingen University for SPIN-MIRMS analysis.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb.2022.1080847/full#supplementary-material

References

Baggs, E. M. (2011). Soil microbial sources of nitrous oxide: recent advances in knowledge, emerging challenges and future direction. Curr. Opin. Environ. Sustain. 3, 321–327. doi: 10.1016/j.cosust.2011.08.011

CrossRef Full Text | Google Scholar

Bakken, L. R., and Frostegard, A. (2017). Sources and sinks for N2O, can microbiologist help to mitigate N2O emissions? Environ. Microbiol. 19, 4801–4805. doi: 10.1111/1462-2920.13978

PubMed Abstract | CrossRef Full Text | Google Scholar

Barford, C. C., Montoya, J. P., Altabet, M. A., and Mitchell, R. (1999). Steady-state nitrogen isotope effects of N2 and N2O production in Paracoccus denitrificans. Appl. Environ. Microbiol. 65, 989–994. doi: 10.1128/AEM.65.3.989-994.1999

PubMed Abstract | CrossRef Full Text | Google Scholar

Bergsma, T., Ostrom, N., Emmons, M., and Robertson, G. (2001). Measuring simultaneous fluxes from soil of N₂O and N₂ in the field using the ¹⁵N-gas “nonequilibrium” technique. Environ. Sci. Technol. 35, 4307–4312. doi: 10.1021/es010885u

PubMed Abstract | CrossRef Full Text | Google Scholar

Brand, T., and Wohanka, W. (2001). Importance and characterization of the biological component in slow filters. Acta Hortic. 554, 313–322. doi: 10.17660/ActaHortic.2001.554.34

CrossRef Full Text | Google Scholar

Buchen, C., Lewicka-Szczebak, D., Flessa, H., and Well, R. (2018). Estimating N₂O processes during grassland renewal and grassland conversion to maize cropping using N₂O isotopocules. Rapid Commun. Mass Spectrom. 32, 1053–1067. doi: 10.1002/rcm.8132

PubMed Abstract | CrossRef Full Text | Google Scholar

Buchen, C., Lewicka-Szczebak, D., Fuß, R., Helfrich, M., Flessa, H., and Well, R. (2016). Fluxes of N₂ and N₂O and contributing processes in summer after grassland renewal and grassland conversion to maize cropping on a plaggic anthrosol and a histic gleysol. Soil Biol. Biochem. 101, 6–19. doi: 10.1016/j.soilbio.2016.06.028

CrossRef Full Text | Google Scholar

Butterbach-Bahl, K., Baggs, E. M., Dannenmann, M., Kiese, R., and Zechmeister-Boltenstern, S. (2013). Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Philos. Trans. R. Soc. B 368, 20130122–20130113. doi: 10.1098/rstb.2013.0122

PubMed Abstract | CrossRef Full Text | Google Scholar

Cowan, N., Ferrier, L., Spears, B., Drewer, J., Reay, D., and Skiba, U. (2022). CEA systems: the means to achieve future food security and environmental sustainability? Front. Sustain. Food Syst. 6:891256. doi: 10.3389/fsufs.2022.891256

CrossRef Full Text | Google Scholar

Dannehl, D., Suhl, J., Ulrichs, C., and Schmidt, U. (2015). Evaluation of substitutes for rock wool as growing substrate for hydroponic tomato production. J. Appl. Bot. Food Qual. 88, 68–77. doi: 10.5073/JABFQ.2015.088.010

CrossRef Full Text | Google Scholar

Daum, D., and Schenk, M. K. (1996a). Gaseous nitrogen losses from a soilless culture system in the greenhouse. Plant Soil 183, 69–78. doi: 10.1007/bf02185566

CrossRef Full Text | Google Scholar

Daum, D., and Schenk, M. K. (1996b). Influence of nitrogen concentration and form in the nutrient solution on N₂O and N₂ emissions from a soilless culture system. Plant Soil 203, 279–288. doi: 10.1023/a:1004350628266

CrossRef Full Text | Google Scholar

Daum, D., and Schenk, M. K. (1997). Evaluation of the acetylene inhibition method for measuring denitrification in soilless plant culture systems. Biol. Fertil. Soils 24, 111–117. doi: 10.1007/bf01420230

CrossRef Full Text | Google Scholar

Daum, D., and Schenk, M. K. (1998). Influence of nutrient solution pH on N2O and N-2 emissions from a soilless culture system. Plant Soil 203, 279–287. doi: 10.1023/A:1004350628266

CrossRef Full Text | Google Scholar

de Kreij, C., Voogt, W., and Baas, R. (2003). “Nutrient Solutions and Water Quality for Soilless Cultures ”. (Naaldwijk, The Netherlands: Applied Plant Research, Division Glasshouse).

Google Scholar

Decock, C., and Six, J. (2013). How reliable is the intramolecular distribution of 15N in N2O to source partition N2O emitted from soil? Soil Biol. Biochem. 65, 114–127. doi: 10.1016/j.soilbio.2013.05.012

CrossRef Full Text | Google Scholar

Deppe, M., Well, R., Giesemann, A., Spott, O., and Flessa, H. (2017). Soil N₂O fluxes and related processes in laboratory incubations simulating ammonium fertilizer depots. Soil Biol. Biochem. 104, 68–80. doi: 10.1016/j.soilbio.2016.10.005

CrossRef Full Text | Google Scholar

Dyckmans, J., Eschenbach, W., Langel, R., Szwec, L., and Well, R. (2021). Nitrogen isotope analysis of aqueous ammonium and nitrate by membrane inlet isotope ratio mass spectrometry (MIRMS) at natural abundance levels. Rapid Commun. Mass Spectrom. 35:e9077. doi: 10.1002/rcm.9077

PubMed Abstract | CrossRef Full Text | Google Scholar

Farquharson, R., and Baldock, J. (2007). Concepts in modelling N2O emissions from land use. Plant Soil 309, 147–167. doi: 10.1007/s11104-007-9485-0

CrossRef Full Text | Google Scholar

Felber, R., Conen, F., Flechard, C. R., and Neftel, A. (2012). Theoretical and practical limitations of the acetylene inhibition technique to determine total denitrification losses. Biogeosciences 9, 4125–4138. doi: 10.5194/bg-9-4125-2012

CrossRef Full Text | Google Scholar

Firestone, M. K., and Davidson, E. A. (1989). “Microbiological basis of NO and N2O production and consumption in soil” in Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere: Report of the Dahlem Workshop on Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere. eds. M. O. Andreae and D. S. Schimel (New York, NY, USA: Wiley), 7–21.

Google Scholar

Frame, C. H., and Casciotti, K. L. (2010). Biogeochemical controls and isotopic signatures of nitrous oxide production by a marine ammonia-oxidizing bacterium. Biogeosciences 7, 2695–2709. doi: 10.5194/bg-7-2695-2010

CrossRef Full Text | Google Scholar

Fuss, R., Hueppi, R., and Pedersen, A. R. (2020). gasfluxes: Greenhouse Gas Flux Calculation from Chamber Measurements: Functions for greenhouse gas flux calculation from chamber measurements; Version: 0.4-4, Depends: R (>= 3.5.0) [Software]. https://CRAN.R-project.org/package=gasfluxes

Google Scholar

Groffman, P. M., Altabet, M. A., Böhlke, J. K., Butterbach-Bahl, K., David, M. B., Firestone, M. K., et al. (2006). Methods for measuring denitrification: diverse approaches to a difficult problem. Ecol. Appl. 16, 2091–2122. doi: 10.1890/1051-0761(2006)016[2091:Mfmdda]2.0.Co;2

PubMed Abstract | CrossRef Full Text | Google Scholar

Gruda, N. (2009). Do soilless culture systems have an influence on product quality of vegetables? J. Appl. Bot. Food Qual. 82, 141–147. doi: 10.18452/9433

CrossRef Full Text | Google Scholar

Halbert-Howard, A., Hafner, F., Karlowsky, S., Schwarz, D., and Krause, A. (2021). Evaluating recycling fertilizers for tomato cultivation in hydroponics, and their impact on greenhouse gas emissions. Environ. Sci. Pollut. Res. 28, 59284–59303. doi: 10.1007/s11356-020-10461-4

PubMed Abstract | CrossRef Full Text | Google Scholar

Hashida, S.-N., Johkan, M., Kitazaki, K., Shoji, K., Goto, F., and Yoshihara, T. (2014). Management of nitrogen fertilizer application, rather than functional gene abundance, governs nitrous oxide fluxes in hydroponics with rockwool. Plant Soil 374, 715–725. doi: 10.1007/s11104-013-1917-4

CrossRef Full Text | Google Scholar

Jinuntuya-Nortman, M., Sutka, R. L., Ostrom, P. H., Gandhi, H., and Ostrom, N. E. (2008). Isotopologue fractionation during microbial reduction of N2O within soil mesocosms as a function of water-filled pore space. Soil Biol. Biochem. 40, 2273–2280. doi: 10.1016/j.soilbio.2008.05.016

CrossRef Full Text | Google Scholar

Karlowsky, S., Gläser, M., Henschel, K., and Schwarz, D. (2021). Seasonal nitrous oxide emissions from hydroponic tomato and cucumber cultivation in a commercial greenhouse company. Front. Sustain. Food Syst. 5:626053. doi: 10.3389/fsufs.2021.626053

CrossRef Full Text | Google Scholar

Kool, D. M., Wrage, N., Oenema, O., Dolfing, J., and Van Groenigen, J. W. (2007). Oxygen exchange between (de)nitrification intermediates and H2O and its implications for source determination of NO3- and N2O: a review. Rapid Commun. Mass Spectrom. 21, 3569–3578. doi: 10.1002/rcm.3249

PubMed Abstract | CrossRef Full Text | Google Scholar

Lakhiar, I. A., Gao, J., Syed, T. N., Chandio, F. A., and Buttar, N. A. (2018). Modern plant cultivation technologies in agriculture under controlled environment: a review on aeroponics. J. Plant Interact. 13, 338–352. doi: 10.1080/17429145.2018.1472308

CrossRef Full Text | Google Scholar

Lewicka-Szczebak, D., Augustin, J., Giesemann, A., and Well, R. (2017). Quantifying N₂O reduction to N₂ based on N₂O isotopocules – validation with independent methods (helium incubation and ¹⁵N gas flux method). Biogeosciences 14, 711–732. doi: 10.5194/bg-14-711-2017

CrossRef Full Text | Google Scholar

Lewicka-Szczebak, D., Dyckmans, J., Kaiser, J., Marca, A., Augustin, J., and Well, R. (2016). Oxygen isotope fractionation during N2O production by soil denitrification. Biogeosciences 13, 1129–1144. doi: 10.5194/bg-13-1129-2016

CrossRef Full Text | Google Scholar

Lewicka-Szczebak, D., Lewicki, M. P., and Well, R. (2020). N2O isotope approaches for source partitioning of N2O production and estimation of N2O reduction – validation with the 15N gas-flux method in laboratory and field studies. Biogeosciences 17, 5513–5537. doi: 10.5194/bg-17-5513-2020

CrossRef Full Text | Google Scholar

Lewicka-Szczebak, D., Well, R., Bol, R., Gregory, A. S., Matthews, G. P., Misselbrook, T., et al. (2015). Isotope fractionation factors controlling isotopocule signatures of soil-emitted N(2)O produced by denitrification processes of various rates. Rapid Commun. Mass Spectrom. 29, 269–282. doi: 10.1002/rcm.7102

PubMed Abstract | CrossRef Full Text | Google Scholar

Lewicka-Szczebak, D., Well, R., Köster, J., Fuß, R., Senbayram, M., Dittert, K., et al. (2014). Experimental determinations of isotopic fractionation factors associated with N₂O production and reduction during denitrification in soils. Geochim. Cosmochim. Acta 134, 55–73. doi: 10.1016/j.gca.2014.03.010

CrossRef Full Text | Google Scholar

Lin, W., Li, Q., Zhou, W., Yang, R., Zhang, D., Wang, H., et al. (2022). Insights into production and consumption processes of nitrous oxide emitted from soilless culture systems by dual isotopocule plot and functional genes. Sci. Total Environ. 856:159046. doi: 10.1016/j.scitotenv.2022.159046

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, B., Morkved, P. T., Frostegard, A., and Bakken, L. R. (2010). Denitrification gene pools, transcription and kinetics of NO, N2O and N2 production as affected by soil pH. FEMS Microbiol. Ecol. 72, 407–417. doi: 10.1111/j.1574-6941.2010.00856.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Llorach-Massana, P., Muñoz, P., Riera, M. R., Gabarrell, X., Rieradevall, J., Montero, J. I., et al. (2017). N2O emissions from protected soilless crops for more precise food and urban agriculture life cycle assessments. J. Clean. Prod. 149, 1118–1126. doi: 10.1016/j.jclepro.2017.02.191

CrossRef Full Text | Google Scholar

Maeda, K., Spor, A., Edel-Hermann, V., Heraud, C., Breuil, M. C., Bizouard, F., et al. (2015). N2O production, a widespread trait in fungi. Sci. Rep. 5:9697. doi: 10.1038/srep09697

PubMed Abstract | CrossRef Full Text | Google Scholar

Mandernack, K. W., Mills, C. T., Johnson, C. A., Rahn, T., and Kinney, C. (2009). The δ15N and δ18O values of N2O produced during the co-oxidation of ammonia by methanotrophic bacteria. Chem. Geol. 267, 96–107. doi: 10.1016/j.chemgeo.2009.06.008

CrossRef Full Text | Google Scholar

Menyailo, O. V., and Hungate, B. A. (2006). Stable isotope discrimination during soil denitrification: production and consumption of nitrous oxide. Glob. Biogeochem. Cycles 20:GB3025. doi: 10.1029/2005gb002527

CrossRef Full Text | Google Scholar

Morley, N., and Baggs, E. M. (2010). Carbon and oxygen controls on N2O and N2 production during nitrate reduction. Soil Biol. Biochem. 42, 1864–1871. doi: 10.1016/j.soilbio.2010.07.008

CrossRef Full Text | Google Scholar

Nadeem, S., Dörsch, P., and Bakken, L. R. (2013). Autoxidation and acetylene-accelerated oxidation of NO in a 2-phase system: implications for the expression of denitrification in ex situ experiments. Soil Biol. Biochem. 57, 606–614. doi: 10.1016/j.soilbio.2012.10.007

CrossRef Full Text | Google Scholar

Ostrom, N. E., Pitt, A., Sutka, R., Ostrom, P. H., Grandy, A. S., Huizinga, K. M., et al. (2007). Isotopologue effects during N2O reduction in soils and in pure cultures of denitrifiers. J. Geophys. Res. 112:287. doi: 10.1029/2006jg000287

CrossRef Full Text | Google Scholar

Röckmann, T., Kaiser, J., Brenninkmeijer, C., and Brand, W. (2003). Gas chromatography/isotope-ratio mass spectrometry method for high-precision position-dependent ¹⁵N and ¹⁸O measurements of atmospheric nitrous oxide. Rapid Commun. Mass Spectrom. 17, 1897–1908. doi: 10.1002/rcm.1132

PubMed Abstract | CrossRef Full Text | Google Scholar

Rohe, L., Anderson, T. H., Braker, G., Flessa, H., Giesemann, A., Lewicka-Szczebak, D., et al. (2014). Dual isotope and isotopomer signatures of nitrous oxide from fungal denitrification - a pure culture study. Rapid Commun. Mass Spectrom. 28, 1893–1903. doi: 10.1002/rcm.6975

CrossRef Full Text | Google Scholar

Rohe, L., Well, R., and Lewicka-Szczebak, D. (2017). Use of oxygen isotopes to differentiate between nitrous oxide produced by fungi or bacteria during denitrification. Rapid Commun. Mass Spectrom. 31, 1297–1312. doi: 10.1002/rcm.7909

CrossRef Full Text | Google Scholar

Savvas, D., Gianquinto, G., Tuzel, Y., and Gruda, N. (2013). “Good agricultural practices for greenhouse vegetable crops - principles for Mediterranean climate areas, 12: soilless culture” in FAO Plant Production and Protection Paper eds. W. Baudoin, R. Nono-Womdim, N. Lutaladio, A. Hodder, N. Castilla, C. Leonardi, S. D. Pascale, M. Qaryouti, and R. Duffy, et al. (Rome, Italy: Food and Agriculture Organization of the United Nations).

Google Scholar

Savvas, D., and Gruda, N. (2018). Application of soilless culture technologies in the modern greenhouse industry – a review. Eur. J. Hortic. Sci. 83, 280–293. doi: 10.17660/eJHS.2018/83.5.2

CrossRef Full Text | Google Scholar

Scholefield, D., Hawkins, J. M. B., and Jackson, S. M. (1997). Development of a helium atmosphere soil incubation technique for direct measurement of nitrous oxide and dinitrogen fluxes during denitrification. Soil Biol. Biochem. 29, 1345–1352. doi: 10.1016/s0038-0717(97)00021-7

CrossRef Full Text | Google Scholar

Schröder, F.-G., and Lieth, J. H. (2002). “Chapter 7 irrigation control in hydroponics” in Hydroponic Production of Vegetables and Ornamentals. eds. D. Savvas and H. Passam (Egaleo, Greece: Embryo Publications).

Google Scholar

Schwarz, D., Thompson, A. J., and Klaring, H. P. (2014). Guidelines to use tomato in experiments with a controlled environment. Front. Plant Sci. 5:625. doi: 10.3389/fpls.2014.00625

PubMed Abstract | CrossRef Full Text | Google Scholar

Sharma, N., Acharya, S., Kumar, K., Singh, N., and Chaurasia, O. P. (2018). Hydroponics as an advanced technique for vegetable production: an overview. J. Soil Water Conserv. 17:364. doi: 10.5958/2455-7145.2018.00056.5

CrossRef Full Text | Google Scholar

Small, G. E., McDougall, R., and Metson, G. S. (2019). Would a sustainable city be self-sufficient in food production? Int. J. Des. Nat. Ecodyn. 14, 178–194. doi: 10.2495/dne-v14-n3-178-194

CrossRef Full Text | Google Scholar

Spott, O., Russow, R., Apelt, B., and Stange, C. (2006). A ¹⁵N-aided artificial atmosphere gas flow technique for online determination of soil N₂ release using the zeolite Köstrolith SX6®. Rapid Commun. Mass Spectrom. 20, 3267–3274. doi: 10.1002/rcm.2722

PubMed Abstract | CrossRef Full Text | Google Scholar

Spott, O., Russow, R., and Stange, C. F. (2011). Formation of hybrid N2O and hybrid N2 due to codenitrification: first review of a barely considered process of microbially mediated N-nitrosation. Soil Biol. Biochem. 43, 1995–2011. doi: 10.1016/j.soilbio.2011.06.014

CrossRef Full Text | Google Scholar

Spott, O., and Stange, C. F. (2007). A new mathematical approach for calculating the contribution of anammox, denitrification and atmosphere to an N2 mixture based on a 15N tracer technique. Rapid Commun. Mass Spectrom. 21, 2398–2406. doi: 10.1002/rcm.3098

PubMed Abstract | CrossRef Full Text | Google Scholar

Stevens, R. J., and Laughlin, R. J. (1998). Measurement of nitrous oxide and di-nitrogen emissions from agricultural soils. Nutr. Cycl. Agroecosyst. 52, 131–139. doi: 10.1023/a:1009715807023

CrossRef Full Text | Google Scholar

Sutka, R. L., Adams, G. C., Ostrom, N. E., and Ostrom, P. H. (2008). Isotopologue fractionation during N(2)O production by fungal denitrification. Rapid Commun. Mass Spectrom. 22, 3989–3996. doi: 10.1002/rcm.3820

PubMed Abstract | CrossRef Full Text | Google Scholar

Sutka, R. L., Ostrom, N. E., Ostrom, P. H., Breznak, J. A., Gandhi, H., Pitt, A. J., et al. (2006). Distinguishing nitrous oxide production from nitrification and denitrification on the basis of isotopomer abundances. Appl. Environ. Microbiol. 72, 638–644. doi: 10.1128/AEM.72.1.638-644.2006

PubMed Abstract | CrossRef Full Text | Google Scholar

Toyoda, S., Mutobe, H., Yamagishi, H., Yoshida, N., and Tanji, Y. (2005). Fractionation of N2O isotopomers during production by denitrifier. Soil Biol. Biochem. 37, 1535–1545. doi: 10.1016/j.soilbio.2005.01.009

CrossRef Full Text | Google Scholar

Toyoda, S., and Yoshida, N. (1999). Determination of nitrogen isotopomers of nitrous oxide on a modified isotope ratio mass spectrometer. Anal. Chem. 71, 4711–4718. doi: 10.1021/ac9904563

CrossRef Full Text | Google Scholar

Well, R., and Flessa, H. (2009). Isotopologue enrichment factors of N(2)O reduction in soils. Rapid Commun. Mass Spectrom. 23, 2996–3002. doi: 10.1002/rcm.4216

PubMed Abstract | CrossRef Full Text | Google Scholar

Well, R., Kurganova, I., de Gerenyu, V., and Flessa, H. (2006). Isotopomer signatures of soil-emitted N₂O under different moisture conditions - a microcosm study with arable loess soil. Soil Biol. Biochem. 38, 2923–2933. doi: 10.1016/j.soilbio.2006.05.003

CrossRef Full Text | Google Scholar

Wheeler, R. M. (2017). Agriculture for space: people and places paving the way. Open Agric. 2, 14–32. doi: 10.1515/opag-2017-0002

CrossRef Full Text | Google Scholar

Wrage-Mönnig, N., Horn, M. A., Well, R., Müller, C., Velthof, G., and Oenema, O. (2018). The role of nitrifier denitrification in the production of nitrous oxide revisited. Soil Biol. Biochem. 123, A3–A16. doi: 10.1016/j.soilbio.2018.03.020

CrossRef Full Text | Google Scholar

Wu, D., Well, R., Cárdenas, L., Fuß, R., Lewicka-Szczebak, D., Köster, J., et al. (2019). Quantifying N₂O reduction to N₂ during denitrification in soils via isotopic mapping approach: model evaluation and uncertainty analysis. Environ. Res. 179:108806. doi: 10.1016/j.envres.2019.108806

PubMed Abstract | CrossRef Full Text | Google Scholar

Yoshida, N. (1988). 15N-depleted N2O as a product of nitrification. Nature 335, 528–529. doi: 10.1038/335528a0

CrossRef Full Text | Google Scholar

Yoshihara, T., Tokura, A., Hashida, S. N., Kitazaki, K., Asobe, M., Enbutsu, K., et al. (2016). N2O emission from a tomato rockwool culture is highly responsive to photoirradiation conditions. Sci. Hortic. 201, 318–328. doi: 10.1016/j.scienta.2016.02.014

CrossRef Full Text | Google Scholar

Yu, L., Harris, E., Lewicka-Szczebak, D., Barthel, M., Blomberg, M. R. A., Harris, S. J., et al. (2020). What can we learn from N₂O isotope data? - analytics, processes and modelling. Rapid Commun. Mass Spectrom. 34:e8858. doi: 10.1002/rcm.8858

PubMed Abstract | CrossRef Full Text | Google Scholar

Zaman, M., Kleineidam, K., Bakken, L., Berendt, J., Bracken, C., Butterbach-Bahl, K., et al. (2021). “Isotopic techniques to measure N₂O, N₂ and their sources” in Measuring emission of agricultural greenhouse gases and developing mitigation options using nuclear and related techniques: Applications of nuclear techniques for GHGs. eds. M. Zaman, L. Heng, and C. Müller (Cham: Springer International Publishing), 213–301. doi: 10.1007/978-3-030-55396-8_7

CrossRef Full Text | Google Scholar

Zhu, K., Bruun, S., Larsen, M., Glud, R. N., and Jensen, L. S. (2015). Heterogeneity of O₂ dynamics in soil amended with animal manure and implications for greenhouse gas emissions. Soil Biol. Biochem. 84, 96–106. doi: 10.1016/j.soilbio.2015.02.012

CrossRef Full Text | Google Scholar