SUMMARY: Seawater is collected from known depths using CTD-rosette sampling protocols. Subsamples are carefully drawn and stored in acid-washed polyethylene bottles. Nitrate/nitrite is measured with an azo dye either before (nitrite) or after (nitrite plus nitrate) subsamples are passed through a cadmium reduction column. Dissolved organic nitrogen is determined after quantitative conversion to inorganic N by exposure to UV radiation.

1. Principle

In seawater the forms of dissolved nitrogen of greatest interest are, in order of decreasing oxidation state: nitrate, nitrite, ammonium and organic nitrogen. All these forms of nitrogen, as well as nitrogen gas (N2), are biochemically interconvertible and are components of the biological nitrogen cycle.

In this method nitrate is quantitatively reduced to nitrite in a copperized cadmium reduction column. The nitrite thus produced, along with any nitrite present in the original sample, is coupled with an aromatic amine, which in turn is reacted with a second aromatic amine to produce an azo dye. The extinction due to the dye is then read spectrophotometrically. A second subsample is analyzed without prior reduction in order to determine the nitrite level. Nitrate is calculated by difference between the [nitrate+nitrite] and nitrite concentrations, using standard solutions. For surface water samples (<100 m) where the [nitrate+nitrite] concentration is generally <0.05 µM, we have employed a low-level assay procedure which is based on the production and detection of nitrous oxides.

Total dissolved nitrogen (TDN) is determined by UV oxidation of the sample and subsequent analysis for dissolved inorganic nitrogen (DIN = nitrite + nitrate + ammonia). Dissolved organic nitrogen (DON) is computed from the relationship DON = TDN - DIN, where TDN is total dissolved nitrogen after UV oxidation and DIN is the sum of the dissolved inorganic nitrogen species before UV oxidation. As an alternative to the UV oxidation method, Walsh (1989) has described a high-temperature (1100°C) combustion method which has been applied to open ocean samples collected in the North Pacific Ocean. No significant differences were observed between these two procedures (Walsh, 1989).

2. Precautions

Contamination is the primary concern with these samples. This is particularly true with samples collected from the euphotic zone, where inorganic nutrient concentrations are extremely low (<0.2 µM). In order to avoid contamination, sample bottles must be meticulously cleaned with dilute HCl and rinsed with deionized distilled water (DDW) before use. Samples are stored frozen until analysis, generally within 1-2 weeks of sample collection.

3. Sampling Collection and Storage

NOTE: The currently held "dogma" in the oceanographic literature is that seawater samples must be processed fresh and on board ship for high-precision, low-level inorganic nutrient analyses (Morse et al., 1982; Venrick and Hayward, 1985). However, extensive results from automated analyses of nutrients in tropical seawaters (Ryle et al., 1981) and the VERTEX program (D. Karl and S. Moore, unpubl. results) which included direct comparisons of [NO3+NO2], PO4 and SiO3 determinations in fresh vs. frozen samples would suggest otherwise. Provided that caution is taken to collect and store the samples in an environment free of potential contamination, we found no significant treatment effect. A similar conclusion was presented by Walsh et al. (manuscript) following the analysis of a wide range of seawater samples that were either analyzed fresh or frozen and stored for varying periods of time. They conclude that, "Despite published and voiced opinions to the contrary, there appears to be no adequate basis either from the literature or from our experiments for across-the-board dismissal of high-precision nutrient analyses undertaken on properly stored seawater samples." As we are not able to take our autoanalyzer to sea on the HOT program cruises, we have focussed our attention on maintaining a contamination-free environment during collection and storage of nutrient samples.

3.1. Sample collection
3.1.1. Rinse the nutrient sample bottle (acid-washed, 125 ml polyethylene bottle) 3 times before filling. Fill to approximately 2/3 full, tighten cap and freeze.
3.1.2. Record cruise, cast and Niskin bottle number on the bottle and data sheet.

4. Sample Analysis

4.1. Standard procedure

Currently, GOFS and WOCE nutrients collected during the Hawaii Ocean Time Series cruises are analyzed by the Hawaii Institute of Marine Biology Analytical Facility. Mr. Ted Walsh has provided us with the following procedure for the analysis of dissolved N.

4.1.1. Nitrate (NO3-) plus Nitrite (NO2-)

[NO3-+NO2-] analyses are performed on a four-channel Technicon Autoanalyzer IIR continuous flow system. The automated wet chemistries generally follow the standard methods of seawater analysis as given by Technicon (1977), which involve: (1) reduction of nitrate to nitrite using a copperized cadmium reduction column, (2) reaction of nitrite with sulfanilamide for diazotization and (3) coupling with N-1-napthylethylenediamine dihydrochloride (NED) forming a purple azo dye (Armstrong et al., 1967). The dye absorbance is read through a 15 mm pathlength flowcell at 550 nm. The reduction column is looped in line using a Hamilton 4-way valve and can be by-passed for nitrite analysis only. Both nitrate and nitrite standards are run to check column efficiency. If speciation is desired, nitrite is determined separately by omitting the reduction step. Nitrate is calculated by difference.

4.1.2. Dissolved Organic Nitrogen (DON)

The method for DON involves initial UV digestion of a seawater sample followed by autoanalysis of the digestion products for [nitrate+nitrite], as above, and ammonium using the Berthelot (indophenol) method. The modified photooxidation technique (Armstrong et al., 1966) utilizes a 24 hour irradiation. Details are given in Walsh (1989). Periodic calibration checks of the UV lamp efficiency are made using a dissolved organic nitrogen (2,2-bypyridyl) standard. As a general rule, the UV lamp is replaced after approximately 700-800 hr of use. DON is calculated by difference between the sum of [nitrate+ nitrite+ammonium] before and after UV treatment.

4.2. Low-level procedures for nitrate (NO3-) and nitrite (NO2-) by chemiluminescence

Nanomolar quantitites of nitrate and nitrite are routinely analyzed using the chemiluminescence techniques of Cox (1980) and Garside (1982). This technique relies on the wet chemical reduction of nitrate and nitrite to nitric oxide in a highly acidic solution of sulfuric acid, ferrous ammonium sulfate, and ammonium molybdate. The reduced nitric oxide is carried by an inert carrier gas (argon), scrubbed of acid and water vapors in a cold finger filled with 6 M sodium hydroxide solution followed by an anhydrous sodium carbonate filled drying tube. The gas stream is then routed through a membrane dryer and the nitric oxide is combined with ozone and simultaneously exposed to a photomultiplier tube. The nitric oxide is further oxidized to excited nitrogen dioxide emitting a photon as it returns to ground state, and the emitted light is detected by the photomultiplier.

4.2.1. [NO3- + NO2-] determinations

[NO3- + NO2-] analyses are performed on an Antek model 720 nitrogen oxide analyzer. Ten ml of concentrated sulfuric acid (36 N) plus 2 ml each of ferrous ammonium sulfate (4%, w/v) followed by ammonium molybdate (2%, w/v) are dispensed into the reaction tube. The reaction tube is inserted into the carrier gas line and the reagents degassed. A 10 ml water sample or standard solution is introduced into the reaction tube with a syringe through a septum fitted to the side arm of the reaction tube. Total reaction light emission is recorded using an automated integrator.

4.2.2. NO3- determinations

Sample analysis is conducted, as above (Chapter 7, section 4.2.1), except that the sample (10 ml) is pretreated with 0.2 ml of sulfanilamide prior to injection into the reaction tube. Since sulfanilamide quantitatively binds nitrite, the integral is the result of nitrate only. Therefore, nitrite can be obtained from the difference of the two analyses.

5. Calibration, Data Reduction and Calculations

5.1. Calibration stocks and regression standards

The calibration of dissolved inorganic nutrients in the autoanalysis of seawater samples is performed using standard solutions containing N, P and Si. A nutrient stock solution is prepared by dissolving dried (65°C, 72 hours) analytical grade reagent chemicals with distilled-deionized water in 1 liter glass volumetric flasks containing 1 ml of chloroform. Once dissolved, this stock solution is immediately transferred into 1 liter amber polypropylene bottles and stored at 4°C. The reagent chemicals and concentrations are: phosphate (KH2PO4, 1 mM), nitrate (KNO3, 5 mM) and silica (Na2SiF6, 4 mM).

Working standards are prepared daily by volumetric dilutions of the stock using glass pipettes and a plastic (polymethylpentene; PMP) volumetric flask. All pipettes and PMP flasks are acid-washed (1 M HCl) and gravimetrically calibrated prior to use. The daily regression standards are prepared by diluting the working standard with low nutrient natural seawater diluent (SWDIL). The SWDIL is filtered open ocean surface seawater that is stored in a carboy at room temperature. By using this technique all standards are matrix-matched with the seawater samples and any cross-nutrient interference effect should be accounted for.

Cross-nutrient interference and reagent contamination was evaluated by preparing separate solutions, as above, but with one of the three standards omitted. Only phosphate showed a slightly measurable increase (+0.014 µM) in the presence of 40 µM-NO3 and 160 µM-Si. The linear regressions of the standards were applied to all seawater sample peaks for calculating each batch of cruise samples. Typical correlations produced r2 values that were between 0.9999 and 0.99999.

5.2. Blank corrections

All seawater standard absorbance peaks were corrected for the absorbance of the seawater diluent (SWDIL). All seawater sample peaks were corrected for the refractive index absorbance for each unique nutrient detection system. The refractive index corrections (in apparent uM units) ranged from approximately 0.13 (for P), 0.23 (for N) to 2.41 (for Si), and represent the increase in absorbance that is due strictly to the presence of dissolved salts in seawater when compared to the distilled-deionized water baseline. These corrections are made running seawater (35 ‰ salinity) through the autoanalyzer with DDW only in reagent lines. The Levor surfactant used routinely in the phosphate channel was omitted from the DDW lines during the refractive index measurement because Levor reacts erratically with seawater in the absence of the acidic color reagent.

6. Accuracy and Precision

The detection limit for nitrate plus nitrite is approximately 0.03 uM with a coefficient of variation for field-collected replicates of 0.3%. The detection limit for DON is 0.05 with a coefficient of variation of 4%.

7. Equipment/Supplies

8. Reagents

9. References

POSTSCRIPT: During year 1 of the HOT program, we routinely measured NH4+ concentrations using the standard Berthelot (indophenol) method. Concentrations of NH4+ were consistently at, or below, our detection limits (<0.05 µM) throughout the water column. Although a new method has been described for low-level determinations of NH4+ in seawater (Brzezinski, 1988, Limnology and Oceanography, 33, 1176-1182), we have not yet successfully adapted this procedure for our routine determinations. Until that method, or a suitable alternative, is available we have decided to delete NH4+ measurements from our list of "standard procedures."