INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion Appendix References

Denitrification and nitrogen fixation are important counteractive processes affecting nitrogen dynamics in coastal sediments. Phytoplankton production can be limited by nitrogen availability (35), and sediments often serve as an important source of dissolved inorganic N (calculated as NH₄⁺ + NO₂ + NO₃). Denitrification is the only biological process that transforms combined N to gaseous forms (N₂ or N₂O) (36, 37). These gaseous end products are unavailable to most producers (e.g., phytoplankton and bacteria) unless N₂ is transformed into organic N during nitrogen fixation (12). Benthic denitrification is a significant sink for combined N in systems and may drive systems toward N limitation (37). Nitrogen fixation increases the amount of biologically available N (7, 38, 42).

Despite the important role of denitrification in coastal and open ocean systems, accurate measurement is hindered by high background levels of N₂ gas in the atmosphere and water column (1, 8, 9, 18, 20, 24, 30, 32, 39, 40). Isotope pairing and membrane inlet mass spectrometry (MIMS) techniques have improved the accuracy and precision of denitrification measurements (4, 15, 16, 25, 28, 29, 33).

The isotope-pairing technique developed by Nielsen (29) estimates denitrification by monitoring changes in nitrogen gas with different isotope compositions (²⁹N₂ = ¹⁴N + ¹⁵N, ³⁰N₂ = ¹⁵N + ¹⁵N) after enriching the overlying water with ¹⁵NO₃. The ²⁸N₂ (¹⁴N + ¹⁴N) production rate was calculated from the relative ²⁹N₂ and ³⁰N₂ production rates (4, 28, 29, 33). Avoiding the measurement of ²⁸N₂ reduced the possibility of contamination. A model simulation showed that the modified gradient created by ¹⁵NO₃ addition has a minimal effect on in situ denitrification rates (26).

Quadruple mass spectrometers have been linked with a MIMS to enhance measurement of dissolved gases (15). The MIMS improved the accuracy of dissolved gas measurement and decreased chances of contamination. Observing the change in ratio between N₂ and Ar measured with the MIMS provides a sensitive and convenient determination of net denitrification rates (8, 15), but measured N₂ flux is a net result of production (denitrification) and consumption (nitrogen fixation).

Nitrogen fixation is a process mediated by microbes that convert N₂ to organic nitrogen. Benthic nitrogen fixation can increase nitrogen availability for biological production in coastal regions (7, 38, 42). Nitrogen fixation rates are low in most coastal sediments, except for areas covered by microbial mats or sea grass beds. Simultaneous measurements of denitrification and nitrogen fixation are desirable, but such data are not common. A method to measure both processes in the same sample is needed to quantify the two processes in situations where both may be important.

Available methods for nitrogen fixation measurement require certain assumptions and have shortcomings (25, 38). A common method for nitrogen fixation measurement is an acetylene (C₂H₂) reduction assay (10). Nitrogen fixers do not discriminate between C₂H₂ and N₂ as substrates during nitrogen fixation. Although simple, inexpensive, and sensitive, this technique requires a conversion constant relating the ethylene production rate to the N₂ reduction rate (theoretical ratio = 3 mol of acetylene per 1 mol of N₂ reduction) (38). The ratio varies and is affected by environmental conditions (38).

We expanded the capability of the MIMS system (15) to measure different isotopic forms of N₂ gas (²⁹N₂ and ³⁰N₂) relative to Ar (S. An and W. S. Gardner, submitted for publication), for isotope-pairing experiments with the MIMS system. By comparing fluxes of three N₂ gas species (²⁸N₂, ²⁹N₂, and ³⁰N₂), production (denitrification) and consumption (nitrogen fixation) rates for N₂ gas could be measured at the same time. Here, we report the methodology and formulas developed for the simultaneous estimation of the two processes. Procedures for determining the different isotopic forms of N₂ gas (²⁹N₂ and ³⁰N₂) using the MIMS and testing of the method are described . In a "potential-denitrification" experiment, conditions were optimized for denitrification measurements and fluxes for the three forms of N₂ gas were observed. The sensitivities of the MIMS system for three isotopic forms of N₂ were evaluated by comparing measured fluxes with expected production rates for each gas. A second experiment measured rates in algal mat sediments where nitrogen fixation was expected. The method differentiated and quantified both denitrification and total nitrogen fixation (net N₂ change + denitrification) rates. Finally, rates of denitrification and nitrogen fixation were measured on intact sediment cores from Laguna Madre, Texas. Assumptions and problems associated with the method are discussed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion Appendix References

(i) Formulas. The ²⁸N₂ production rate estimated from the isotope-pairing technique should match the ²⁸N₂ production rate estimated from direct methods (for example, MIMS) in cores where no nitrogen fixation occurs. The proportional changes in three N₂ species with time after ¹⁵NO₃ addition can be predicted under such conditions. If nitrogen fixation occurs, however, predicted changes in the proportion of the three N₂ species with time will not match the measured changes. The difference between predicted and measured proportional changes of the three N₂ species can reflect nitrogen fixation and denitrification rates.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(ii) Dissolved ²⁹N₂ and ³⁰N₂ measurements with the MIMS system. We modified the method of Kana et al. (15) to measure dissolved ²⁹N₂ and ³⁰N₂ in addition to Ar, O₂, and ²⁸N₂. The ²⁹N₂ and ³⁰N₂ concentrations were obtained from the "excess" atomic mass unit (AMU) 29 and 30 signals, respectively, caused by the conversion of added ¹⁵NO₃. The ratio between ²⁹N₂ and ²⁸N₂ is 0.00732 (0.00366 × 2) in natural samples, since the natural abundance of ¹⁵N is 0.366% (23). The relationship between ²⁹N₂ and ²⁸N₂ in samples was obtained from 30-ppt artificial seawater (standard water) held at different temperatures (21 and 30°C) (Fig. 1a). The ratio in our MIMS system was close to the theoretical value (Fig. 1a), but it was calculated in each measurement to account for small variability between measurements. This relationship was used to determine the excess AMU 29 signal. The feed water in a flowthrough incubation experiment does not have excess AMU 29 signal resulting from ²⁹N₂, while outflow water has excess AMU 29 signal, indicating that ²⁹N₂ was produced during denitrification (Fig. 1a). The excess AMU 29 signal was converted to excess ²⁹N₂ concentration by comparing the results with those from standard water.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Relationships between different AMU signals in liquid samples measured with the quadruple mass spectrometer. Solid symbols, standard water (, 30 ppt, 21°C; , 30 ppt, 30°C); open symbols, samples (, outflow water; , inflow water). The line represents the regression among the standard water samples. The signal represents partial pressure of gases (in pascals [102]) inside the mass spectrometer. Data are from the work of An and Gardner (submitted).

(iii) Sediment incubation experiments. The MIMS system was evaluated for measuring different N₂ species in three sediment incubation experiments. The developed formulas were applied to data from each experiment to obtain denitrification and nitrogen fixation rates under different conditions.

(a) Potential denitrification bottle experiment. A potential denitrification experiment was performed in closed bottles. Sediment slurries (20 ml) from a fresh water pond (Research Park, College Station, Tex.) were placed in serum bottles (120 ml). The bottles were filled with distilled water saturated with air at 21°C before the bottles were sealed with gas tight caps. The distilled water was enriched with glucose (final concentration = 2 mM) to provide a carbon source. Five treatments with different proportions of ¹⁴NO₃ and ¹⁵NO₃ (0 and 100%, 25 and 75%, 50 and 50%, 75 and 25%, and 100 and 0%) and a control (no NO₃ added) were prepared. Total concentrations of nitrate in each treatment bottle were the same (32 µg-atoms of N bottle¹). Duplicate bottles were prepared for each treatment to give a total of 12 bottles. Each bottle was sealed so as to be gas tight with no headspace. Sediment slurries in the bottles were mixed well and then incubated at room temperature (21°C). After 24 h of incubation, dissolved gases from one bottle of each treatment (total of six) were measured with the MIMS system. Turbid sediment slurry samples were drawn into the membrane inlet tubes without filtration. Before measurements, the sediment slurry was mixed vigorously and large particles were allowed to settle out. The concentration of each nitrogen gas species was calculated as described above. After 48 h of incubation, the N₂ forms in the remaining six bottles were measured. The total production of each N₂ gas species was estimated by the concentration difference between each treatment and control.

(b) Algal mat sediment incubation. Undisturbed sediment cores (12-cm diameter, 30-cm length) with bottom water were collected in a shallow salt marsh area with algal mats near Port Aransas, Tex., in April 2000. The cores were transported to a temperature-controlled incubation room (21°C), and a flowthrough plunger with Teflon inlet and outlet tubes was installed over each sediment core. The flowthrough chamber setup consisted of an intake water vessel, Teflon flow tubes, a peristaltic pump, and a sample collection vessel (21). The bottom water collected from the site was passed over the core surface at a rate of 1.2 ml min¹. Duplicate sediment cores were incubated under four different treatment conditions (total of eight). For the GL treatment, flowthrough water was enriched with glucose (2 mM) and incubated in the light (~300 microeinsteins m² s¹) to maximize N₂ fixation rates. GD cores were enriched with glucose and incubated without light. NL cores were incubated in the light (~300 microeinsteins m² s¹) without glucose enrichment. ND cores were incubated in the dark without glucose enrichment. Water column depth over the sediment was maintained at about 5 cm to give a water volume of ca. 570 ml in each core. Triplicate samples of feed water and outlet water were collected at intervals for dissolved-gas analysis after an initial incubation period of 1 day, to allow steady-state conditions to develop (21). Water samples were collected for analysis of dissolved inorganic nitrogen compounds (NH₄⁺, NO₃, and NO₂).

(c) Sediment core incubations with shallow estuarine sediments. The flowthrough incubation experiment was repeated with sediment cores from Laguna Madre, a shallow, semienclosed estuary in the southeastern part of the Texas coast. Laguna Madre is a negative estuary where freshwater input is less than evaporation (water residence time = 1 year). The salinity is often more than 40 ppt and can vary annually up to 60 ppt (6). Laguna Madre stations had depths of 0.8 to 0.9 m and were populated with Thalassia testudinum (300 to 600 shoots m²) (22). During our sampling period, salinity was lower than in earlier observations (6) and did not exceed 40 ppt. Bottom water was oxygenated at most stations due to wind-driven mixing. Sediment cores (four in each station) were collected from two sites (L155 and L189). Station details are described elsewhere (An and Gardner, submitted). One half of the cores were incubated under dim light (~30 microeinsteins m² s¹), and the others were covered with aluminum foil. The dim-light condition did not cause significant light effects, so data from the two treatments were combined. The ¹⁵NO₃ was added after the day 1 samples were taken.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion Appendix References

(i) Potential denitrification experiment. Most of the added nitrate was converted to N₂ gas within 24 h. The average production rate of total N₂ (²⁸N₂ + ²⁹N₂ + ³⁰N₂) was 28.6 (standard error [SE] = 1.5) µg-atoms of N bottle¹ day¹, accounting for 90% of the added nitrate (32 µg-atoms of N bottle¹). The concentrations of the three N₂ species were similar after 48 h of incubation (27.9 ± 1.2 µg-atoms of N bottle¹). For calculations, the N₂ species production was assumed to be completed at 24 h for both the 24- and 48-h-incubation samples. The production rates may have been higher than calculated because conversion of added nitrate to N₂ gas may have occurred before 24 h.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Production rates of N2 isotope species in a potential-denitrification experiment with sediments enriched with different proportions of 14NO3 and 15NO3. The line represents the expected production rate of the three N2 isotope species based on the proportion of each NO3 species. Symbols represent measured production rates of three N2 gases (solid symbols, day 1; open symbols, day 2). Total amounts of NO3 were constant in each treatment (32 µmol bottle1). About 90% of added NO3 was recovered as N2 gas in this experiment.

1

1

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Denitrification based on 14NO3 (D14') and 15NO3 (D15') during potential denitrification experiment. Error bar represents 1 SE for two incubation bottles.

(ii) Sediment core incubations with algal mats. An experiment was performed on sediments covered with an algal mat to evaluate the method in a nitrogen fixation environment (14, 43). Potential enhancers for the nitrogen fixation process, such as high light and a usable carbon source (11, 14, 25), were provided to some core samples. Before the ¹⁵NO₃ addition (days 1 and 2), ²⁸N₂ fluxes into the sediment were observed with all treatments (Table 1), suggesting high nitrogen fixation rates. The lighted core without glucose (NL) had a higher negative flux than dark cores without glucose (ND). The glucose-enriched cores (GD and GL) had higher negative N₂ fluxes than treatments without glucose enrichments (ND and NL). The light effect was not obvious among glucose-enriched treatments (GD and GL). Sediment oxygen demand (SOD) was higher in dark (ND and GD) than light (NL and GL) treatments, suggesting active photosynthesis. The differences in SOD between dark and light cores were about 1,000 µmol of O₂ m² h¹ and resemble typical benthic primary production rates reported for shallow coastal environments (31). The glucose effect on SOD was not observed before the ¹⁵NO₃ addition.

2

1

2

1

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Total denitrification (DNF) and nitrogen fixation (NF) rates under different light and carbon enrichment conditions. Data are averages ± SE for two sediment cores. Duplicate measurements were not available for day 4 and 7 of treatment GD and day 3 of treatment GL.

2

1

2

1

2

1

2

1

2

1

2

1

(iii) Sediment core incubation with shallow estuarine sediment. The cores were maintained at near natural conditions except that the illuminated ones were maintained under continuous laboratory fluorescent lighting rather than natural outdoor light conditions. Although the water depths were shallow at the sampling stations, in situ light levels at the sediment-water interface were about 10% of the surface values due to high turbidity (S. An and W. S. Gardner, Nitrogen Cycling in Laguna Madre and Baffin Bay, final report to the Texas Water Development Board). The dim light treatment during our laboratory incubation tended to lower SOD compared to dark treatments, suggesting that some benthic primary production activity occurred at the sites. However, data from the two treatments were combined because light effects were not obvious in other processes.

2

1

2

1

2

1

2

1

2

1

View larger version (22K): [in this window] [in a new window]   FIG. 5.   Nitrogen gas flux changes during sediment incubation experiments. The arrow shows the time at which 15NO3 was added. Total denitrification is calculated as 28N2 + 29N2 + 30N2 flux. Data are averages ± 1 SE for four replicate cores. Note the scale change in panel C.

2

1

2

1

2

1

2

1

(iv) Sensitivity test. To evaluate measurement errors associated with ²⁹N₂ and ³⁰N₂ measurement, dissolved-gas data for seawater (30 ppt) at different temperatures (30 and 21°C) measured over 4 days were combined (total n = 32). The average excesses of AMU 29 signal and AMU 30 signal were 0.00007 (SE = 0.0002) and 0.001 (SE = 0.001) µM, respectively. The amount of dissolved gases interfering with the ²⁹N₂ and ³⁰N₂ measurement is different in experimental samples than in standard water, and the measurement error is larger. For example, nitric oxide (NO) produced during denitrification would interfere with the AMU 30 signal measurement and cause the ³⁰N₂ flux to be overestimated (41). The formation of ¹³CO⁺ from CO₂ also may interfere with ²⁹N₂ measurements (5). However, the excess AMU 29 signal and its variability in water samples prepared in different ways (different salinity, sulfite addition, or helium bubbling) were similar to those of standard water, suggesting that the interference of CO₂ in ²⁹N₂ measurement was small.