Service Team Activity Report (Healy 02-03)

15 July to 26 August 2002

Nome, Alaska to Nome, Alaska

 

 

CTD casts were performed with a rosette system consisting of a 12-place rosette frame with 30  liter  Niskin-type  bottles equipped with internal plastic coated springs and a 24-place SBE-32 Carousel pylon.  Underwater electronic components consisted of a

·        Sea-Bird Electronics, Inc. (SBE) 911plus CTD,

·        WetLabs C-Star transmissometer with a 25cm pathlength and 660nm wavelength,

·        Biospherical Instruments, Inc. Photosynthetically Active Radiation (PAR) sensor,

·        Chelsea MkIII Aquatracka fluorometer, and

·        Simrad, 5 volt - 500 meters altimeter.

Additionally, a Dr. Haardt fluorometer designed to detect colored organic matter (CDOM), and a Secchi disk  were mounted on the CTD package. The CTD, transmissometer and fluorometers were mounted horizontally along the bottom of the rosette frame. The PAR sensor was located at the top of the rosette. All sensors except the Secchi disk were interfaced with the CTD system. This instrument package provided pressure, dual temperature and dual conductivity channels as well as light transmissivity and fluorometric signals at a sample rate of 24 scans per second. 

The rosette system was suspended from a standard UNOLS 3 conductor 0.322” electromechanical cable.

The CTD used was serial number 09P12613-0474 and the sensor’s serial numbers are listed in Table 1.

 

TABLE 1. Instrument/Sensor Serial Numbers

 

Primary

Temperature

Primary

Conductivity

Secondary

Temperature

Secondary

Conductivity

Pressure

Transmissometer

SBE 3plus

SBE 4C

SBE 3plus

SBE 4C

401K-105

C-Star

03-2324

04-2112

03-2166

04-23193

69008

CST-479DR

 

Oxygen

Fluorometer

PAR

SBE 43

Aqua 3

QSP-2300

0060

88191

4644

 

The bottles on the rosette were General Oceanic 30 liter bottles. The bottles were equipped with nylon coated springs and silicone o-rings.  Bottle numbering is 1 to 12 with 12 as the deepest and 1 at the shallowest sampling level.

The distance of the mid-points of the 30 L Niskin bottles from the bottom-mounted sensors was ~1m .  The PAR sensor was ~ 0.6 m above the mid-point of the Niskin bottles, and the Secchi disk which is mounted on a rod was ~ 0.9 m above the mid-point of the 30 L Niskin bottles. The distance between the PAR sensor and the bottom mounted sensors was ~1.7 m.  The 30 Liter Niskin bottles are ~1.0 m long.

Most of the 30 liter Niskin-type bottles were manufactured by General  Oceanics, but some were Ocean Test Equipment bottles that had their external springs replaced with the same internal springs used in the General Oceanics bottles. Bottle 7 was replaced after station 010 cast 01 and again after station 24 cast 4.  Bottle 4 was replaced after station 14 cast 2. To minimize toxicity the bottles were equipped with silicone O-rings.  At times the plastic coating on the springs broke down and some rust was apparent. To minimize the occurrence of rust, the springs were inspected before the cruise and, as feasible during the cruise. During the mid-cruise servicing of the CTD/rosette system that occurred following station 024 cast 04, all springs were inspected and some were re-coated with plastic

 

CTD and Bottle Data Distribution

The CTD and bottle data can be obtained through the JOSS web-site, www.JOSS.ucar.edu/sbi. The data are reported using the WHP-Exchange format. In addition, the format can be obtained through the WOCE Hydrographic Program web-site, www.WHPO.ucsd.edu. Ascii files for each station were created with comments recorded on the CTD Station Logs during data acquisition. These ascii files include data processing comments noting any problems, the resolution, and footnoting that may have occurred.  A separate ascii file was also created with the comments from the Sample Log Sheets that include problems with the Niskin bottles that could compromise the samples. Comments arising from  inspection and checking of the data are also included in the ascii file. These comment files are also in the JOSS database.

CTD Data

CTD Laboratory Calibration Procedures

Pre-cruise laboratory calibrations of CTD pressure, temperature and conductivity sensors were used to generate coefficients for the calculation of these parameters from their respective sensor frequencies. The conductivity calibrations were performed at Sea-Bird Electronics, Inc. in Bellevue, Washington.  Calibration of the pressure and temperature sensors were performed by Shipboard Technical Support/Oceanographic Data Facility (STS/ODF) personnel. These laboratory temperature calibrations were referenced to the International Temperature Scale of 1990 (ITS-90).

CTD Data Acquisition

The CTD 911plus was operated generally as suggested in the Sea-Bird CTD Operating and Repair Manual, which contains a description of the system, its operation and functions (Sea-Bird Electronics, Inc., 2002). One difference from Sea-Bird’s operation is that data acquisition was started on deck. This procedure allows a check of the pressure offset and an unblocked reading of the transmissometer. The Seasoft acquisition program as described in the CTD Data Acquisition Software Manual (Sea-Bird Electronics, Inc., 2001) provided a real-time graphical display of selected parameters adequate to monitor CTD performance and information for the selection of bottle-tripping depths. Raw data from the CTD were archived on the PC’s hard disk at the full 24 Hz sampling rate.

A CTD Station Sheet form was filled in for each deployment, providing a record of times, positions, bottom depth, bottle sampling depths, and every attempt to trip a bottle, as well as any pertinent comments. When the equipment and personnel were ready, data acquisition was started.  The CTD operator pressed a control key (flag), which appends a summary line into one of the two files created for “inventory” files.  These files contain a summary of the time, ship’s position, and current scan number each time the control key is pressed.  They are used as a reference to mark important events during the cast, such as on deck pressure, when the lowering was initiated, when the package was at the bottom, when bottles were tripped and the on-deck pressure with ending position.  After the initial flag, the rosette/CTD system was lowered into the water and held at or near the surface for 3-5 minutes to permit activation of the CTD pumps and equilibration of the sensors.   Then, the operator again created a flag and simultaneously directed the winch operator to begin lowering.   The operator created a flag at the deepest point of the cast. Bottom depths were calculated by combining the distance above bottom, reported by the altimeter, and the maximum depth of the CTD package when bottom altimeter readings were available.  If there was no altimeter reading, then the bottom depth is reported from the ship’s Bathy 2000 or Knudsen model 320B/R depth recorder.  These data, corrected for the draft of the transducer, were logged in uncorrected meters (assuming a sound velocity of 1500 m/sec). If the altimeter and depth recorder data were unavailable, the final resort was to use depth data from the SeaBeam system (corrected sound velocities).

 

 The depth of each bottle trip was written on the station log and flagged in the data file.  The performance of all sensors was monitored during the cast.  After the rosette recovery, the operator created a final flag denoting the end of the cast. Any faulty equipment or exceptionally noisy data were noted on the log sheet.

CTD Data Processing

Pressure

CTD values determined on deck before and after each cast were compared to determine a pressure offset correction. The comparison suggested that no pressure offset was necessary. 

Temperature

The primary temperature sensor was calibrated just before the expedition.  The dual temperature sensors were monitored during the expedition and exhibited good agreement.  It appears that no additional corrections need to be applied. A post-cruise calibration will be performed.

Conductivity

Corrected CTD pressure and temperature values were used with bottle salinities to back-calculate bottle conductivities. Comparison of these bottle values with the CTD primary conductivity values indicated no additional offset needed to be applied to the data  

Transmissometer

A WetLab calibrated Transmissometer was utilized throughout the cruise.  An on deck calibration check was performed and even though there was little degradation from the last calibration the new coefficients were applied to the data set.

Oxygen, Fluorometer, and PAR

The CTD oxygen data are only intended for qualitative use.  Similarly, the fluorometric and PAR data are not calibrated.

 

CTD Data Processing

Sea-Bird Seasoft CTD processing software was employed. The processing programs are outlined below.  A more complete description may be found in the Sea-Bird Software Manual which is available from the Sea-Bird website (www.seabird.com).

The sequence of programs that were run in processing CTD data from this cruise are as follows:

·        DATCNV - Converts data from raw frequencies and voltages to corrected engineering units

·        WILDEDIT - Eliminates large spikes

·        CELLTM - Applies conductivity cell thermal mass correction

·        FILTER A low pass filter to smooth pressure for LOOPEDIT

·        LOOPEDIT - Marks scans where velocity is less than selected value to avoid pressure reversals from ship roll, or during bottle flushing.

·        DERIVE - Computes calculated parameters

·        BINAVG - Average data into desired pressure bins

The quality control steps included:

·        Sensor verification After the CTD was set up and sensor serial numbers and sensor location was entered into the computer, another check was made to verify that there were no tabulation errors.

·        Seasoft Configuration File was reviewed to verify that individual sensors were represented correctly, with the correct coefficients.

·        Temperature was verified by comparing primary and secondary sensor data.

·        Conductivity was checked by comparison of the two sensors with each other and with bottle salinity samples.

·        Position Check A chart of the ship’s track was produced and reviewed for any serious problems.  The positions were acquired from the ship’s Trimble P-code navigation system.

·        Visual Check Plots of each usable cast were produced and reviewed for any noise and spikes that may have been missed by the processing programs.

·        The density profile was checked for inversions that might have been produced by sensor noise or response mismatches. 

·        Additional Sea-Bird programs were run on all or some stations to maximize the data quality.

·        WFILTER - Provides a median filter for data smoothing of .CNV files

·        WFILTER was employed on selected stations where there were spikes in the data, specifically in the transmissometer data. This program was run after WILDEDIT

 

The winch grounding system was periodically cleaned to reduce modulo errors, which were not a significant problem during this leg (HLY0203).  During the previous leg (HLY0201) there were several modulo word errors at the beginning of the expedition. A check of all connections and cables was performed and it was found that a shielding around the winch motor eliminated the spiking and most of the noise in the CTD signal.

The CTD down trace is reported unless there was a problem then the up trace is reported.  A notation was made in the comments file if this was necessary. 

 

CTD Data Footnoting

WHP water sample quality flags were assigned to the CTDTMP (CTD temperature) and CTDSAL (CTD salinity) parameters as follows:

2                    Acceptable measurement.

3                    Questionable measurement. The data did not fit the bottle data, or there was a CTD conductivity calibration shift during the up-cast.

4                    Bad measurement. The CTD up-cast data were determined to be unusable for calculating a salinity.

7                    Despiked. The CTD data have been filtered to eliminate a spike or offset.

WHP water sample quality flags were assigned to the CTDOXY (CTD O2) parameter as follows:

1                    Not calibrated. Data are uncalibrated.

2                    Acceptable measurement.

3                    Questionable measurement.

4                    Bad measurement. The CTD data were determined to be unusable for calculating a dissolved oxygen concentration.

5                    Not reported. The CTD data could not be reported, typically when CTD salinity is coded 3 or 4.

7                    Despiked. The CTD data have been filtered to eliminate a spike or offset.

9                    Not sampled. No operational sensor was present on this cast. Either the sensor cover was left on or the depth rating necessitated removal

Data Comments

Fine structure including minor density inversions that may appear in the upper ~ 10 m of the profiles is most likely caused by ship discharges/turbulence.  A comparison cast made from a small boat at station 14 is included in the data on the JOSS web site.  These data revealed generally similar profiles, but minor density inversions occurred in the shipboard profile and were absent in the profile taken from the small boat. To minimize the ship effect, engine cooling water discharges were restricted to the port side of the Healy starting with station 002 during the previous leg (HLY0201) and continuing on this leg.  At about this time, a “yo yo” procedure was adopted  to induce bottle flushing whenever waves and ship motion were weak.  This procedure was employed for all bottle trips under quiescent conditions except for productivity casts, and for some thin low salinity surface lenses.  In the latter cases, the CTD was raised slowly so as not to disturb the thin low-salinity surface layers with the CTD wake, and the soak time was relied on  to flush the bottle.  Even though this procedure, may not have adequately flushed the surface bottle, it was sufficient to reveal some large salinity differences in the ~ 1m depth interval separating the CTD sensor from the bottle mid-point.  These situations occurred in melting ice under low winds and waves, and it is suspected that the water may have been stratified even within the surface Niskin bottle. Regardless of the procedure employed, the CTD operators were instructed to wait for at least 1 minute (typically > 1.5 minutes) before tripping the bottle.

 

All salinity, nutrient and dissolved oxygen data collected by the service team have gone through several stages of editing and are not likely to change significantly.  The chlorophyll observations reported are, however, preliminary and may undergo significant post-cruise editing.

Bottle Sampling

There were six generic types of casts performed with differing sampling protocols.  Generally speaking, the sampling during these casts were as follows, but there is some cast-to-cast variation.

·        Hydrographic

o       Oxygen,

o       Total CO2,

o       Total Alkalinity,

o       Nutrients

o       Chlorophyll

o       Salinity

o       O18/O16

o       Dissolved Organic Carbon

o       Dissolved Inorganic Carbon

o       Particulate Organic Matter

o       Benthic

o       Stable Isotopes

o       Radioisotopes

·        Productivity

o       Oxygen and/or Oxygen Respiration

o       Productivity

o       Nutrients

o       Chlorophyll

o       HPLC

o       Bacteria

o       Micro Zooplankton

o       Bio-Optics

·        Bio-Markers

o       Dissolved Organic Matter

o       Lignin

·        Radium

o       Nutrients

o       Radium

·        Zooplankton

o       Nutrients

 

 

The correspondence between individual sample containers and the rosette bottle from which the sample was drawn was recorded on the sample log for the cast. This log also included any comments or anomalous conditions noted about the rosette and bottles.

 

Normal sampling practice included opening the drain valve before the air vent on the bottle, to check for air leaks. This observation together with other diagnostic comments (e.g., "lanyard caught in lid", "valve left open") that might later prove useful in determining sample integrity were routinely noted on the sample log. Drawing oxygen samples also involved taking the sample draw temperature from the bottle.

Bottle Data Processing

After the samples were drawn and analyzed, the next stage of processing involved merging the different data streams into a common file. The rosette cast and bottle numbers were the primary identification for all ODF-analyzed samples taken from the bottle, and were used to merge the analytical results with the CTD data associated with the bottle.

Diagnostic comments from the sample log, and notes from analysts and/or bottle data processors were entered into a computer file associated with each station (the "quality" file) as part of the quality control procedure. Sample data from bottles suspected of leaking were checked to see if the properties were consistent with the profile for the cast, with adjacent stations, and, where applicable, with the CTD data. Direct inspection of the tabular data, property-property plots and vertical sections were all employed to check the data. Revisions were made whenever there was an objective reason to delete, annotate or re-calculate a datum. WHP water sample codes were selected to indicate the reliability of the individual parameters affected by the comments. WHP bottle codes were assigned where evidence showed the entire bottle was affected, as in the case of a leak, or a bottle trip at other than the intended depth.

Bottle Data Footnoting 

WHP water bottle quality codes were assigned as defined in the WOCE Operations Manual [Joyce] with the following additional interpretations:

2                    No problems noted.

3                    Leaking.  An air leak large enough to produce an observable effect on a sample is identified by a code of 3 on the bottle and a code of 4 on the oxygen.  (Small air leaks may have no observable effect, or may only affect gas samples.)

4                    Did not trip correctly.  Bottles tripped at other than the intended depth were assigned a code of 4.  There may be no problems with the associated water sample data.

5                    Not reported.  No water sample data reported.  This is a representative level derived from the CTD data for reporting purposes.  The sample number should be in the range of 80-99.

9                    The samples were not drawn from this bottle. 

WHP water sample quality flags were assigned using the following criteria:

1                    The sample for this measurement was drawn from the water bottle, but the results of the analysis were not (yet) received.

2                    Acceptable measurement.

3                    Questionable measurement. The data did not fit the station profile or adjacent station comparisons (or possibly CTD data comparisons). No notes from the analyst indicated a problem. The data could be acceptable, but are open to interpretation.

4                    Bad measurement. The data did not fit the station profile, adjacent stations or CTD data. There were analytical notes indicating a problem, but data values were reported. Sampling and analytical errors were also coded as 4.

5                    Not reported. There should always be a reason associated with a code of 5, usually that the sample was lost, contaminated or rendered unusable.

9                    The sample for this measurement was not drawn.

Not all of the quality codes are necessarily used on this data set.

Pressure and Temperatures

All pressures and temperatures for the bottle data tabulations were obtained by averaging CTD data for a brief interval at the time the bottle was closed and then applying the appropriate calibration data.

 

The temperatures are reported using the International Temperature Scale of 1990.

Salinity

Equipment and Techniques

Salinity samples were drawn into 200 ml high alumina borosilicate bottles, which were rinsed three times with sample prior to filling. The bottles were sealed with custom-made plastic insert thimbles and Nalgene screw caps This container provides very low container dissolution and sample evaporation. 

 

A Guildline Autosal 8400A #57-526, standardized with IAPSO Standard Seawater (SSW) batch P-140, was used to measure the salinities. Prior to the analyses, the samples were stored to permit equilibration to laboratory temperature, usually 8-20 hours.  The salinometer was modified by ODF and contained an interface for computer-aided measurement. The salinometer was standardized with a fresh vial of standard seawater at the beginning and end of the run.  The SSW vial at the end of the run was used as an unknown to check for drift. The salinometer cell was flushed until two successive readings met software criteria for consistency; these were then averaged for a final result. The estimated accuracy of bottle salinities run at sea is usually better than 0.002 PSU relative to the particular standard seawater batch used.

Laboratory Temperature

The temperature stability in the salinometer laboratory was fair, sometimes varying as much as 3.5ºC during a run of samples.  The laboratory temperature was generally 1-2ºC lower than the Autosal bath temperature.

Oxygen Analysis

Equipment and Techniques

Dissolved oxygen analyses were performed with an ODF-designed automated oxygen titrator using photometric end-point detection based on the absorption of 365nm wavelength ultra-violet light. The titration of the samples and the data logging were controlled by PC software. Thiosulfate was dispensed by a Dosimat 665 buret driver fitted with a 1.0 ml buret. The ODF method used a whole-bottle modified-Winkler titration following the technique of Carpenter (1965) with modifications by Culberson (1991), but with higher concentrations of potassium iodate standard (approximately 0.012N) and thiosulfate solution (55 g/l). Standard KIO3 solutions prepared ashore were run at the beginning of each run. Reagent and distilled water blanks were determined, to account for presence of oxidizing or reducing materials.

 

Sampling and Data Processing

 

Samples were collected for dissolved oxygen analyses soon after the rosette was brought on board. Using a Tygon drawing tube, nominal 125ml volume-calibrated iodine flasks were rinsed, then filled and allowed to overflow for at least 3 flask volumes. The sample draw temperature was measured with a small platinum resistance thermometer embedded in the drawing tube. Reagents were added to fix the oxygen before stoppering. The flasks were shaken twice to assure thorough dispersion of the precipitate, once immediately after drawing, and then again after about 20 minutes.  The samples were usually analyzed within a few hours of collection.

Thiosulfate normalities were calculated from each standardization and corrected to 20ºC. The 20ºC normalities and the blanks were plotted versus time and were reviewed for possible problems. New thiosulfate normalities were recalculated as a linear function of time, if warranted. The oxygen data were recalculated using the smoothed normality and an averaged reagent blank. Oxygens were converted from milliliters per liter to micromoles per kilogram using the sampling temperature.

 

Volumetric Calibration

 

Oxygen flask volumes were determined gravimetrically with degassed deionized water to determine flask volumes at ODF’s chemistry laboratory. This is done once before using flasks for the first time and periodically thereafter when a suspect bottle volume is detected. The volumetric flasks used in preparing standards were volume-calibrated by the same method, as was the 10 ml Dosimat buret used to dispense standard iodate solution.

Standards

 

Potassium iodate was obtained from Johnson Matthey Chemical Co. and was reported by the supplier to be >99.4% pure.

Nutrient Analysis

Equipment and Techniques

 

Nutrient analyses (phosphate, silicate, nitrate+nitrite, urea, ammonium, and nitrite) were performed on an ODF-modified 6-channel Technicon AutoAnalyzer II, generally within a few hours after sample collection.  Occasionally samples were refrigerated for longer periods and the data are annotated if it was felt that the storage time had a significant effect.  Frozen samples left for us to analyze from a previous group were also analyzed during this leg. The analog outputs from each of the six channels were digitized and logged automatically by computer (PC) at 2-second intervals.  A suite of frozen nutrient samples from the Bering Sea (~300) left for us to analyze by participants in HLY0202 (the cruise that separated the two Process cruises, HLY0201 and HLY0203) was also analyzed. Many of the frozen samples had extremely high silicate concentrations and only those values of  about 50 micromolar or less should be considered reliable.  This is because of the problem of silicate polymerization during freezing and because the silicate method employed was optimized for waters with silicate concentrations between 0 and 90 micromolar.  The samples were thawed in a warm tap water bath and analyzed as soon as possible in order to obtain the best possible data on the nutrients other than silicate.   For a discussion of the effects of freezing on silicate concentrations see Macdonald et al. (1986).

 

Silicate was analyzed using the technique of Armstrong et al., (Armstrong, 1967). The sample was passed through a 15mm flowcell and the absorbance measured at 660nm.

 

A modification of the Armstrong et al. (Armstrong 1967) procedure was used for the analysis of nitrate and nitrite. For the nitrate plus nitrite analysis, the seawater sample was passed through a cadmium reduction column where nitrate was quantitatively reduced to nitrite. The stream was then passed through a 15mm flowcell and the absorbance measured at 540nm.  The same technique was employed for nitrite analysis, except the cadmium column was bypassed, and a 50mm flowcell was used for measurement.  Periodic checks of the column efficiency were made by running alternate equal concentrations of NO2 and NO3 through the NO3 channel to ensure that column efficiencies were high (> 95%). Nitrite concentrations were subtracted from the nitrate+nitrite values to obtain nitrate concentrations.

Phosphate was analyzed using a modification of the Bernhardt and Wilhelms [ Bernhardt 1967.] technique. The reaction product was heated to ~55ºC to enhance color development, then passed through a 50mm flowcell and the absorbance measured at 820m.

Ammonium was determined by the Berthelot reaction (Patton and Crouch 1977) in which sodium hypochlorite and phenol react with ammonium ion to produce indophenol blue, a blue compound, with an absorption maximum at 637nm.  The solution was heated to 55°C and passed through a 50mm flowcell at 640nm.

Urea was analyzed via a modification of the method of Rahmatullah and Boyde (1980), which is based on the classic diacetyl monoxime method.  A solution of diacetyl monoxime, thiosemicarbizide and acetone is followed by the addition of ferric chloride, which acts as a catalyst.  The resultant solution is heated to 90°C and passed through a 50mm flowcell. The absorbance is measured at 520nm.

 

Sampling and Data Processing

 

Nutrient samples were drawn into 45 ml polypropylene, screw-capped “oak-ridge type” centrifuge tubes. The tubes were cleaned with 10% HCl and rinsed with sample three times before filling. Standardizations were performed at the beginning and end of each group of analyses (typically 6-24 samples) with an intermediate concentration mixed nutrient standard prepared prior to each run from a secondary standard in a low-nutrient seawater matrix. The secondary standards were prepared aboard ship by dilution from primary standard solutions.  Dry standards were pre-weighed at the laboratory at ODF, and transported to the vessel for dilution to the primary standard. Sets of 6-7 different standard concentrations covering the range of sample concentrations were analyzed periodically to determine the deviation from linearity, if any, as a function of concentration for each nutrient analysis.  A correction for non-linearity was applied to the final nutrient concentrations when necessary.

After each group of samples was analyzed, the raw data file was processed to produce another file of response factors, baseline values, and absorbances. Computer-produced absorbance readings were checked for accuracy against values taken from a strip chart recording. A stable deep seawater check sample was run frequently as a substandard check.

 

Nutrients, when reported in micromoles per kilogram, were converted from micromoles per liter by dividing by sample density calculated at 1 atm pressure (0 db), in situ salinity, and an assumed laboratory temperature of 25ºC.

Nutrient Standards

 

Na2SiF6, the silicate primary standard, was obtained from Johnson Matthey Company and Fisher Scientific and was reported by the suppliers to be >98% pure. Primary standards for nitrate (KNO3), nitrite (NaNO2), and phosphate (KH2PO4) were obtained from Johnson Matthey Chemical Company., and the supplier reported purities of 99.999%, 97%, and 99.999%, respectively. Ammonia, (NH4(SO4)2), and Urea primary standards were obtained from Fisher Scientific and reported to be >99% pure. In addition, during the HLY0201 leg, L. A. Codispoti supplied independent comparisons standards for all nutrients. All standard intercomparisons, produced agreement well within the precision of the employed methods.


 

References

Armstrong, F. A. J., Stearns, C. R., and Strickland, D. H., “The measurement of upwelling and subsequent biological processes by means of the Technicon Autoanalyzer and associated equipment,” Deep-Sea Research, 14, pp. 381-389, (1967).

Bernhardt, Wilhelms A., “The continuous determination of low level iron, soluble phosphate and total phosphate with the AutoAnalyzer”, Technicon Symposia, I,  pp. 385-389 (1967).

Carpenter, J. H., “The Chesapeake Bay Institute technique for the Winkler dissolved oxygen method,” Limnology and Oceanography, 10, pp. 141-143 (1965).

Culberson, C. H., Knapp, G., Stalcup, M., Williams, R.T., and Zemlyak, F., “A comparison of methods for the determination of dissolved oxygen in seawater,” Report WHPO 91-2, WOCE Hydrographic Programme Office (Aug 1991).

Gordon, L.I., Jennings, J.C., Ross, A.A. and J.M. Krest, “A Suggested Protocol for Continuous Flow Automated Analysis of Seawater Nutrients in the WOCE Hydrographic Program and the Joint Global Ocean Fluxes Study”. 1993. WOCE Hydrographic Programs Office, Methods Manual WHPO 91-1.

Intergovernmental Oceanographic Commission, Scientific Committee on Oceanic Research Manual and Guides 29 Protocols for the Joint Global Ocean Flux Study (JGOFS) Core Measurements.  UNESCO, 170pp., (1994).

Joyce, T. ed., and Corry, C. ed., “Requirements for WOCE Hydrographic Programme Data Reporting,” Report WHPO 90-1, WOCE Report No. 67/91 3.1, pp. 52-55, WOCE Hydrographic Programme Office, Woods Hole, MA, USA (May 1994, Rev. 2), UNPUBLISHED MANUSCRIPT

Macdonald, R.W., F. A. McLaughlin and C.S. Wong,  “The storage of reactive silicate samples by freezing,” Limnology and Oceanography, 31, pp. 1139-1142 (1986).

Patton, C.J. and Crouch, S.R., “Spectrophotometric and kinetics investigation of the Berthelot reaction for the determination of ammonia,” Analytical Chemistry, 49(3), pp.464-469 (1977).

Rahmatullah, Mohammed, and Boyde, T.R.C, “Improvements in the determination of urea using diacetyl monoxime; methods with and without deproteinisation,” Clinica Chimica Acta, 107, pp.3-9 1980.

Sea-Bird Electronics, Inc, CTD Data Acquisition Software Manual, March 2001

Sea-Bird Electronics, Inc., CTD Operating and Repair Manual, February 2002