Discrete samples will be taken from all hydrographic casts
employing 24x5 1 Niskin bottles. Samples will be collected
immediately after the gas samples have been collected. 20 ml samples will
be taken into 25 ml high density linear polyethylene scintillation vials
fitted with conical polyethylene cap liners, after three rinses of
c.a. 5 ml each. Between sampling, the vials will contain the previous
nutrient sample. Samples will be stored refrigerated in the dark, and
will be analyzed directly from the vials as soon as possible after
collection, and no more than 24 hours thereafter.
Nutrient profiles will be made daily or as frequently as
possible to 100--125 m. We will deploy a 2" i.d. PVC hose connected
to a well de-watering pump delivering 500 l min into an on deck
190 l reservoir. The intake to the pumping system will be attached to a CTD/
BOPS package. A second pump in the reservoir will deliver 50 l min
through a garden hose to the continuous flow analyzer (CFA) in the
laboratory. The flowing seawater in the garden hose will be sub-sampled
through a solenoid valve alternating between sample and artificial seawater.
All nutrient analyses will be done using a six channel CFA to
measure ammonium, urea, nitrite, nitrate + nitrite, phosphate, and
silicate colorimetrically. Deep samples with nutrient concentrations
greater than the analytical range will be precisely diluted with low
nutrient artificial seawater (ASW) of known nutrient content to bring
their nutrient concentrations within the analytical range.
Ammonium (NH) will be measured using the indophenol
blue method (Berthelot's reaction). NH reacts with alkaline
hypochlorite in the presence of nitroprusside, and is complexed with
alkaline phenol. Precipitation of magnesium and calcium is prevented by
complexing with a strong citrate buffer. The reaction mixture is heated
to 80 C to accelerate the reaction, and absorbance is measured at
660 nm in a 50 mm flow cell.
Urea will be measured using a modification of the method
first described by DeManche et al. ( Limnol. Oceanog., 1973
(18) 686--689). The sample is made strongly acid with a mixture of
sulfuric and phosphoric acids containing trace amounts of ferric ion. A
pink chromophore is formed by reaction with a color reagent containing
diacetylmonoxime and thiosemicarbazide at 95 C, and absorbance is
measured at 520 nm in a 50 mm flow-cell.
Nitrite will be measured using the method of Bendschneider
and Robinson ( J. Mar. Res., 1952 (11) 87--96). Nitrite reacts
with sulfanilamide in strong acid medium to form a diazonium salt which is
then coupled with N-(1-napthyl)ethylenediamine dihydrochloride to form a
magenta diazo dye. Absorbance is measured at 540 nm in a 50 mm flow cell.
Nitrate plus Nitrite
Nitrate is first reduced to nitrite using a heterogeneous
reaction on a copperized cadmium column based on the method of Wood
et al. ( J. Mar. biol. Ass. U.S., (47) 23--31). Nitrite is then
determined as above.
Phosphate is determined using an automated version of the
method described by Murphy and Riley ( Anal. Chim. Acta, 1962 (12)
162--176). Phosphate reacts with acid molybdate to form phosphomolybdic
acid. The phosphomolybdic acid is reduced to a phosphomolybdenum blue
complex by ascorbic acid with mild heating (38 C), and the
resulting absorbance is measured at 880 nm in a 50 mm flow cell. Silicate
does not interfere because of the strongly acid conditions used for the reaction.
Molybdate reacts with silicate to form silicomolybdic acid
which is then reduced to a silicomolybdenum blue complex which is measured
at 660 nm in a 50 mm flow cell. The method differs from that described by
Armstrong et al. ( Deep Sea Res., 1967 (14) 381--389) in
that ascorbic acid is used as the reductant rather than stannous chloride.
Oxalic acid is used to prevent the interference of phosphate.
All samples will be run with an alternating ASW containing an
isotonic sodium chloride and magnesium sulfate mixture, and this will also
be the matrix for working standards. The ASW will be introduced to the
CFA after a DIW baseline on reagents has been established. This will
permit detection of any ASW contamination offsetting the baseline from
true zero (the ASW offset).
Standards will be prepared at 1 mM from pre-weighed salts or
compound. Primary standards will be prepared in freshly drawn de-ionized
water (DIW) using class A volumetric glassware. They will be stored in
high density polyethylene, refrigerated and in the dark.
A calibration working standard will be prepared in each new batch of ASW,
and will be monitored as described below (QA/QC). Analytical ranges
will be 5, 5, 5, 40, 5 and 40 µM respectively, with a precision of
±0.5 -- 1.0 % of full scale. Calibration will generally be made using
a single top working standard.
Computation of Sample Concentrations
Sample peak heights will be corrected for ASW offsets and for
interpolated baseline. Standards will be corrected for baseline only
since they are made in ASW. Mean calibration factors (concentration per
unit peak height) will be computed for each block of standards, and
interpolated calibration factors will be used to compute individual sample
Linearity will be checked periodically using a range of
working standards prepared by dilution of the calibration working standard
into the same ASW. Linearity will also be monitored by running one or
more standard dilutions as samples within the run from time to time.
New working standards will be made for each new batch of
ASW. Old working standards will be run as samples in runs calibrated
with new working standards. This procedure will maintain standard
continuity and provenance, and check for possible standard degradation
Analytical precision will be estimated from replicates of
standards and diluted standards during normal running.
Accuracy involves many more factors than precision, not all
of which can be tested. Replication of sampling into nutrient vials will
be examined, and time permitting, replicate bottles may be tripped at the
same depth to test replication of water sampling. There will be many
replicate casts at each station which will provide a measure of station
variance, but physical processes can change water column distributions, so
that this comparison is not a totally satisfactory measure of accuracy.
Comparison with historical data will be equally useful, but less than
perfect because of temporal change, although the time and space scales of
change of deep water nutrient properties are large and make these
comparisons more rigorous and they also will be part of the QA/QC procedure.