HONO Emissions from Soil Bacteria as a Major Source of Atmospheric Reactive Nitrogen

From Soil to Sky Trace gases emitted either through the activity of microbial communities or from abiotic reactions in the soil influence atmospheric chemistry. In laboratory column experiments using several soil types, Oswald et al. (p. 1233) showed that soils from arid regions and farmlands can produce substantial quantities of nitric oxide (NO) and nitrous acid (HONO). Ammonia-oxidizing bacteria are the primary source of HONO at comparable levels to NO, thus serving as an important source of reactive nitrogen to the atmosphere. HONO emissions from soil are comparable to those of NO in arid and arable regions. Abiotic release of nitrous acid (HONO) in equilibrium with soil nitrite (NO2–) was suggested as an important contributor to the missing source of atmospheric HONO and hydroxyl radicals (OH). The role of total soil-derived HONO in the biogeochemical and atmospheric nitrogen cycles, however, has remained unknown. In laboratory experiments, we found that for nonacidic soils from arid and arable areas, reactive nitrogen emitted as HONO is comparable with emissions of nitric oxide (NO). We show that ammonia-oxidizing bacteria can directly release HONO in quantities larger than expected from the acid-base and Henry’s law equilibria of the aqueous phase in soil. This component of the nitrogen cycle constitutes an additional loss term for fixed nitrogen in soils and a source for reactive nitrogen in the atmosphere.

S oil biogenic NO emissions account for~20% of the total NO sources to the atmosphere (1) and vary as a function of microbial activity and physicochemical soil properties. NO is produced during nitrification, in which soil microbes convert ammonium (NH 4 + ) via NO 2 to nitrate (NO 3 -), both of which can accumulate in soil (2,3). In addition, the reduction of NO 3 -, which is known as denitrification, can cause a release of NO. The two microbial processes are mainly influenced by temperature, soil water content, pH value, and mineral nitrogen availability in the soil (4)(5)(6). Previous studies have shown that HONO may also be emitted from soil; this release may originate from the transformation of soil NH 4 + to NO 2 -(7) or from soil NO 2 because of a chemical acid-base equilibrium (8).
To estimate the contribution of soil HONO emissions to the total reactive nitrogen flux (HONO + NO) from the soil to the atmosphere and to elucidate the major processes influencing HONO release from soil, we studied the relation of soil HONO emissions to biogenic soil NO emissions under controlled laboratory conditions using the dynamic chamber method (9,10). Earlier studies have shown that results from using this technique are consistent with those from field measurements (9,11,12). We investigated soils from various ecosystems around the world, covering a wide range of soil pH, organic matter, and soil nutrient contents (table S1). The soil samples were wetted in order to reach water holding capacity (WHC) (10) and placed into the chamber, which was then continuously flushed with purified air (free of HONO, NO x , O 3 , hydrocarbons, and water vapor), leading to a slow drying of the soil sample during the course of the experiment. The gas-phase mixing ratio of HONO released by the soil sample was measured at the chamber exit with a long path absorption photometer (LOPAP) (13). Mixing ratios of NO and water vapor were also measured (14).
The characteristic moisture dependency of HONO and NO fluxes that is known from previous studies of soil biogenic NO emissions is shown in Fig. 1 (4,9,15,16). We found that the maximal emission fluxes of HONO and NO [henceforth denoted as optimum fluxes; F N,opt (HONO) and F N,opt (NO)] are of comparable magnitude and occur at similar optimum soil water content (SWC) (10)-within 10% WHC of one another for all investigated samples.
Chemical acid-base equilibrium calculations predict that abiotic HONO emissions from soil nitrite should be largest for soils with low pH and high NO 2 content (8). The soil pH reflects a sum parameter, which depends on the amount of acidic and basic species in soil, and regulates the solubility of soil constituents and the protonation equilibria. These variables, however, also influence nitrifier and denitrifier activity in soil. In general, abundance and diversity of bacteria are positively correlated with pH (17), and individuals mostly possess a maximum activity at a certain pH (18). In contrast to expectations based on the acid-base equilibrium, the results from different soil samples presented in Fig. 2 do not show a decrease of HONO fluxes with increasing pH. In fact, the neutral soil sample S12, taken from a wheat field in Germany, features extremely high values for HONO and NO emissions (F N,opt : 257.5 T 0.1 ng m −2 s −1 HONO, 134.8 T 0.6 ng m −2 s −1 NO). The second highest emission of HONO and NO was found for the alkaline, sodic soil represented by sample S17. Comparison with soil NO 2 and NH 4 + concentrations ( Fig. 2) clearly demonstrates that high HONO and NO emissions are favored for soils with high nutrient content.
The ratio of F N,opt (HONO) to F N,opt (NO) was found to be higher for arid and arable soils (on  S1). For soil pH values higher than 7, the optimum HONO emission fluxes always exceed 5 ng m −2 s −1 (in terms of N) and even reached~258 ng m −2 s −1 (at 25°C). We anticipate that HONO emissions are particularly relevant for arid and arable areas with neutral or alkaline soil pH, where they may substantially influence tropospheric chemistry. Potential HONO soil emission hot spots comprise, for instance, large areas of northern Africa, central/southwestern Asia, and North America as well as some regions around the Mediterranean Sea ( fig. S2), covering in total~20% of the terrestrial surface (excluding Antarctica). Given the high spatial variability of soil properties (such as pH and nutrients) and our limited amount of soil samples, these hot spot areas may be even larger. This previously neglected ground source of reactive nitrogen may explain the unexpectedly high daytime HONO mixing ratios observed in many studies (19).
In addition, NO is produced on a time scale of 30 min from the photolysis of HONO during daytime. Hence, soil HONO emissions in the identified hot spot areas ( fig. S2) may account for the observed discrepancies between soil emissions of reactive nitrogen estimated with global models by using the Yienger and Levy algorithm and those derived from "top-down" approaches by using nitrogen dioxide (NO 2 ) columns measured by satellites over arid ecosystems (20,21). Biogenic NO emissions are known to depend strongly on soil temperature (22). We measured the temperature dependency of F N (HONO) and  S1  S2  S3  S4  S5  S6  S7  S8 S9 S10 S11 S12 S13 S14 S15 S16 S17  fig. S3). A temperature increase from 20 to 30°C yielded Q 10 values (averaged over the whole SWC range) of 3.7 (T1.4) for HONO and 2.1 (T0.2) for NO, which is typical for soil respiratory systems (16,23,24). From an Arrhenius plot (Fig. 3A), we obtained similar activation energies for HONO (80 kJ mol −1 ) and for NO (75 kJ mol −1 ). These values are much lower than the activation energies reported for denitrification (202 to 250 kJ mol −1 ) (25) but are within the range reported for nitrification by ammonia-oxidizing bacteria (AOB) (25 to 149 kJ mol −1 ) (3,25), suggesting that the latter process governs the observed co-emission of HONO and NO.
To test this hypothesis, we investigated a pure culture of Nitrosomonas europaea, a common and well-studied AOB (26). A suspension of the pure culture (buffered at pH = 8.2) was applied to glass beads serving as an inert soil-like matrix (16), and the model system was treated like a soil sample (10). F N,opt (HONO) and F N,opt (NO) of the N. europaea culture suspension are compared in Fig. 3B with the emissions by using a sterile AOB nutrient solution additionally containing 0.14 mmol l −1 NO 2 -, which equals 0.5 mg kg −1 of NO 2 -(in terms of N) in soil. The NO 2 added to the sterile solution equals the NO 2 that would have been produced by the bacteria during the experiment and reflects the chemical contribution to the HONO emission from the model system, whereas the observed difference in F N,opt (HONO) between the sterile solution and the culture suspension can be attributed to the direct emissions by the AOB. The N. europaea culture emits four times more HONO than does the sterile reference, demonstrating that AOB can indeed act as a strong direct source of HONO.
Measured adenosine 5´-triphosphate (ATP) concentrations during the dry-out of S12 (Fig. 3C) show that soil microbes are active also under relatively dry conditions (% WHC < 20%), where F N,opt (HONO) is observed. Because ATP is an indicator for microbial activity in general, the maximum activity of AOB might not coincide with the maximal ATP concentration. We applied methyl iodide-a strong sterilization agent for soil (27) also targeting nitrification (28)-to a subsample of S12 (Fig. 3C). Both HONO and NO emission fluxes were reduced by~75%, revealing a strong microbial source. This demonstrates that the findings from the model system shown in Fig. 3B are transferable to a real soil sample. The residual emissions can largely be attributed to the chemical source because the ATP content and, hence, the microbial activity was reduced by~92% at the HONO emission optimum. These results explain the high HONO emissions from nonacidic soil samples.
The conceptual model in Fig. 4 shows that F N,opt (HONO) and F N,opt (NO) occur in the lower SWC range (~0 to 40% WHC) (16,29), whereas at high SWC (~40 to 80% WHC), nitrogen is released from soil mainly as the greenhouse gas N 2 O. In general, substrate diffusion is limited at low SWC, and gas diffusion is limited at higher SWC (30). HONO is produced and emitted during nitrification, which predominates at low SWC (5). Samples from different soil and land-use types show their maximal release of the respective nitrogen compound at different optimum SWC (15). The magnitude of the maximal emission of each compound varies depending on, for example, nutrient availability and abundance of soil bacteria.
HONO emissions by AOB and possibly other types of bacteria represent an additional component for gaseous losses from the soil nitrogen pool to the atmosphere. Our survey of soils from different ecosystems indicates that HONO emissions may account for up to 50% of the reactive nitrogen release from soil. This contribution of soil HONO emissions is currently not considered in model estimates of global soil reactive nitrogen emissions (1) and may constitute one of the major uncertainties in this budget. Furthermore, these HONO emissions contribute to atmospheric chemistry by enhancing the oxidation capacity of the lower atmosphere.  www.sciencemag.org SCIENCE VOL 341 13 SEPTEMBER 2013