CRUISE REPORT:  AJAX
(Updated JAN 2011)



HIGHLIGHTS
                          Cruise Summary Information

        Section designation  AJAX leg 1 
                  ExpoCodes  316N19831007     
            Chief Scientist  Joseph L. Reid/SIO
                               Scripps Institution of Oceanography
                               9500 Gilman Drive, MS 0230
                               La Jolla, CA  92093-0230
                               Tel: 858-534-2055
                               Email: jreid@ucsd.edu     
                      Dates  7 OCT - 6 NOV 1983        
                       Ship  R/V KNORR                 
              Ports of call  Abidjan, Ivory Coast to Cape Town, South Africa 

                                          4°49'40.8"N
      Geographic boundaries  3°57'46.8"W              1°49'58.8"E
                                         41°21'10.8"S
                   Stations  49 
        Floats and drifters  0 
                   Moorings  0


        Section designation  AJAX leg 2
                  ExpoCodes  316N19840111
            Chief Scientist  Worth D. Nowlin, Jr./TAMU
                             Texas A&M University
                               O&M Building, Room 713 MS 3146
                               College Station, Texas 77843 
                               Tel: 979.845.3900  Fax: 979.847.8879
                               Email: wnowlin@tamu.edu
                      Dates  11 JAN - 19 FEB 1984
                       Ship  R/V KNORR
              Ports of call  Cape Town, South Africa to Punta Arenas, Chile

                                       44°0'57.6"S
                             0°25'12"W              1°28'12"E
                                       69°21'46.8"S
                   Stations  31
        Floats and drifters  0
                   Moorings  0







PHYSICAL, CHEMICAL AND IN-SITU CTD DATA

Sponsored by the National Science Foundation and the Office of Naval Research
SIO Reference 85-24, TAMU Reference 85-4-0
Approved for distribution: W.A. Nierenberg, Director, SIO
                           Worth D. Nowlin, Jr., Director, Div. Atmospheric & 
                           Marine Sciences, TAMU
15 December 1985


INTRODUCTION 

This data report presents hydrographic and CTD data collected aboard the R/V 
Knorr on Ajax Expedition Leg I, Abidjan, Ivory Coast to Cape Town, South 
Africa (7 October - 6 November, 1983), and on Leg II, Cape Town to Punta 
Arenas, Chile (11 January -19 February, 1984). Ajax Expedition was jointly 
funded by the National Science Foundation and the Office of Naval Research. 
The expedition objectives were to provide data on the general circulation of 
the South Atlantic, Weddell Gyre and the Scotia Sea. A line of hydrographic 
stations, with sampling from the surface to the bottom on most stations, was 
occupied along the Greenwich meridian from the coast of Africa at 5°N to the 
Antarctic ice edge at 69°S, and from there across the Scotia Ridge and 
through the Scotia Sea (Figure 1). In addition to the hydrographic 
measurements of temperature, salinity, oxygen and nutrients given in this 
report, water samples were also collected for analyses of other tracers such 
as Freons 11 and 12, tritium and helium-3. Table 1 summariz-es the various 
ancillary measurements and lists the principal investigators. 

The hydrographic data were, for the most part, collected and processed at sea 
by personnel of the Physical and Chemical Oceanographic Data Facility 
(PACODF) of Scripps Institution of Oceanography. Final adjustments to the 
hydrographic data and CTD processing were completed after the cruise. Tapes 
of the hydrographic data and CTD data at one-decibar intervals have been sent 
to the National Oceanographic Data Center. 


STANDARD PROCEDURES 

IN-SITU CONDUCTIVITY/TEMPERATURE/DEPTH (CTD) - ROSETTE CAST DATA 

Most stations consisted of two CTD-rosette casts to collect water samples 
from 36 or more levels from the sea surface to near the bottom. A rosette 
frame holding 24 10-liter PVC plastic bottles was lowered with the CTD probe, 
and the bottles were closed during the up cast at depths selected on the 
basis of temperature, salinity and oxygen features observed on the down cast 
CTD plots. The deep cast usually tripped 23 bottles from near the bottom up 
to less than 1000 meters. The 24th bottle was tripped at the surface to 
provide calibration checks for the CTD. The shallow cast typically overlapped 
one level of the deep cast and sampled an additional 12 levels up to the 
surface. Occasionally, more than one bottle was tripped at one level to 
provide additional water for other tracer analyses. Also, on some stations 
water was taken from the surface CTD check bottle of the deep cast for 
special chemical analyses. As the hydrographic data from these duplicate 
levels do not provide additional useful information on the water column 
characteristics, they have been omitted from this data report. Similarly, 
data from a few wire casts with special samplers for trace elements are not 
included. 

Pressure and temperature for the discrete levels are from the corrected CTD 
data taken at the time of the rosette bottle trip. The CTD pressure and 
temperature offsets were monitored by comparisons with deep-sea-reversing 
thermometers (DSRTs) mounted in at least three reversing racks per rosette 
cast. Two different CTDs were used on the expedition. The pressure sensor on 
one of the CTDs used on the first leg had a serious hysteresis problem 
between the down and up casts with surface offsets of 10 to 20 db different 
at the end of the cast compared to the beginning of the cast. The problem has 
been isolated to just the near-surface levels on the up cast and the data 
have been corrected. The pressure sensor was replaced at the beginning of the 
second leg. The new sensor had a much larger slope adjustment as function of 
pressure, but it behaved in a consistent manner. Corrections to the CTD 
pressures were based upon laboratory pressure calibrations and upon the CTD 
pressure offset observed as the CTD entered the water. Comparisons with the 
pressure obtained from unprotected DSRTs, though monitored, were not used 
because the CTD pressure sensor and the lab pressure calibration apparatus 
are considered to be more sensitive than unprotected mercury thermometers. 

Depth was calculated from pressure by Saunders' method (1981). Pressure is 
also given for reference. 

The CTD temperature corrections were based primarily upon comparisons with 
special deep low-range DSRTs. A two-point (0°C and at ambient temperature 
above 20°C) laboratory temperature calibration was performed on both CTDs 
prior to the cruise. However, the laboratory calibration was not very useful 
for two reasons: the ambient check differed by several milli-degrees on 
successive days, when the bath temperature differed by about one degree, and 
a two-point calibration is not adequate to determine the CTD temperature non-
linearity over the full temperature range of the instrument. The DSRT checks 
on CTD temperature sensor were not entirely satisfactory either. Because of 
the differing time responses of the two methods for measuring temperature and 
the depth separation between the CTD temperature probe and the DSRT rack 
mounted on the plastic bottles above the probe, the DSRTs do not provide very 
useful calibration data at shallower depths where the local temperature 
gradient exceeds .01°C per meter. The most useful in-situ CTD checks are in 
isothermal layers such as in the surface mixed layer, or in the low 
temperature gradients in the deep water. The surface DSRTs  were often 
higher-range, lower-precision thermometers which are read to the nearest 
0.01°C. The deeper DSRT racks contained low-range thermometers which were 
read to the nearest 0.001°C and they provided the data for the primary CTD 
temperature offset corrections. The temperature data in this report are 
listed to the nearest 0.001°C, but the uncertainty is several milli-degrees, 
especially in the warmer regions of the water column. 

Salinity samples were collected from each rosette bottle and analyzed usually 
within two days on an Autosal inductive salinometer. Wormley Standard 
Seawater (SSW) batch P92 was used to standardize the salinometer on stations 
60 through 116, batch P90 was used on the remaining stations. After the 
cruise, a problem was discovered on another salinometer: the machine had a 
jump in conductivity ratio when the suppression dial was changed from the 1.9 
(S < 35) to 2.0 (S > 35) setting. Closer examination of the Ajax salinity 
data revealed that one of the two salinometers used at sea definitely had a 
similar problem. The deep and bottom water salinities less than 35 listed in 
this report are correct relative to the SSW, but salinities greater than 35 
may have a systematic offset of .002 to .003 salinity. 

Comparison of the bottle salinities with CTD salinities was used to identify 
malfunctioning rosette bottles and to verify that water samples used for 
other chemical analyses were collected from the correct depth without 
contamination by leakage of the rosette bottle or from a mis-trip. The 
rosette sample bottles malfunctioned often, most frequently due to lanyards 
caught in the lids. In such cases, all water sample analyses were deleted and 
the erroneous salinity replaced with a corrected CTD value followed by the 
footnote letter "D".  All salinity values were calculated from the algorithms 
for the Practical Salinity Scale, 1978 (UNESCO, 1981) and are listed to three 
decimal places.

Dissolved oxygen was determined by the Winkler method as modified by 
Carpenter (1965), using the equipment and procedure outlined by Anderson 
(1971). In the hands of a skilled operator, this technique is capable of a 
precision of better than 0.01 ml/l O2. In practice, the error may be several 
times larger than that figure. In the early part of the first leg, the micro-
buret appears to have leaked, resulting in calculated oxygens that were too 
high; the deep cast oxygens on station 10 were lost.

Silicate, phosphate, nitrate and nitrite nutrients were determined at sea 
using an automated analyzer. The procedures used are similar to those 
described by Atlas et al. (1971). There were some problems in both the 
phosphate and nitrate analyses. A batch of artificial seawater used for 
standards on leg II was contaminated and resulted in inaccurate phosphate 
factors. Also, a change in cadmium reduction columns (at station 103) used in 
the nitrate analyses resulted in inconsistent nitrate reduction efficiency 
for several stations; nitrates were lost on three stations.  On several other 
stations, phosphate and/or nitrate was re-calculated on the basis of typical 
cruise NO3/PO4 ratios.  Most changes were less than 6%. Nitrite was not run on 
four stations because the colorimeter was used for another analyses. Nitrate 
could not be corrected for the presence of nitrite on those stations, so the 
results are listed as measured, NO3 + NO2. 


CONDUCTIVITY/TEMPERATURE/DEPTH (CTD) DATA 

Only the deep cast CTD lowering on each station was processed by PACODF. They 
provided tapes of the CTD temperature and salinity data at one-decibar 
intervals, usually for the down cast. At times, there were problems with the 
down cast and the up cast data are given instead. The oxygen probe data were 
not processed. The quality of the CTD data was degraded on some stations 
because of electrical noise from one of three different sources: the winch 
slip rings, the end cable termination and loose bulk head connectors to the 
CTD. At times the specific problem was quickly identified and fixed at sea, 
and at other times the problem persisted for several stations. The problem 
stations are readily apparent from inspection of the individual CTD station 
curves shown in the latter part of this report. The CTD data have been 
filtered somewhat, but no attempt has been made to improve the noisier 
stations. Some loss of CTD data is also apparent in the CTD plots in high 
gradient regions; the gradient filter used in processing the CTD data appears 
to have been too severe. After removal of some surface spikes and a few large 
spikes near 1000 db, standard depth data were extracted from the one-db 
interval tapes and are listed in this data report. The complete one-db 
interval tapes are available from NODC. 


TABULATED DATA 

The time given is the Greenwich Mean Time of the first bottle trip. For CTD 
casts, it is the start down time for down casts or the start up time for up 
casts. 

Station positions were derived from satellite fixes closest to, or bracketing 
the cast time. 

Bottom depths, determined acoustically, have been corrected using British 
Admiralty Tables (Carter, 1980). The bottom soundings were taken at the time 
that the rosette was near the bottom. 

Wind and wave directions are given to the nearest to degrees. Wind speed is 
given in knots and the wave height and period are given in feet and seconds, 
respectively. 

Weather conditions are coded using WMO code 4501 (Table 2). 

Barometer and air temperature are shown in millibars and degrees Celsius. 

The dominant cloud type is given in the standard two-letter code and the 
cloud amount is recorded in octos. 

Observed hydrographic data and interpolated standard level data have been 
interspersed and are presented together in depth sequence. Interpolated or 
extrapolated data are indicated by the footnote letters "ISL" listed after 
the depth. 

Potential temperatures have been calculated from the expressions given by 
Fofonoff (1977), based upon Bryden's (1973) results. 

Density-related parameters are calculated from the International Equation of 
State of Seawater, 1980 (UNESCO, 1981). Sigma-theta, sigma-2 and sigma-4 are 
the density anomalies for the sample moved adiabatically to the surface, 2000 
db and 4000 db using Fofonoff's (1977) procedure. SVA is the specific volume 
anomaly. 

Percent oxygen saturation was calculated from the equations of Weiss (1970). 


FOOTNOTES 

In addition to footnotes, special notations are used without footnotes 
because the meaning is always the same. 

ISL: After depth values indicates interpolated or extrapolated standard level. 
D:   CTD value listed in place of normal ship-board hydrographic measurement. 
H:   Ship-board hydrographic measurement listed in place of normal CTD values. 



LITERATURE CITED 

Anderson, G.C., compiler, 1971. "Oxygen Analysis," Marine Technician's 
    Handbook, SIO Ref. No. 71-8, Sea Grant Pub. No. 9, 29 pp. 

Atlas, E.L., J.C. Callaway, R.D. Tomlinson, L.I. Gordon, L. Barstow, and P.K. 
    Park, 1971. A Practical Manual for Use of the Technicon(r) 
    AutoAnalyzer(r) in Sea Water Nutrient Analysis; Revised. Oregon State 
    University Technical Report 215, Reference No. 71-22, 49 pp. 

Bryden, H.L., 1973. New polynomials for thermal expansion, adiabatic 
    temperature gradient and potential temperature of seawater. Deep-Sea 
    Res., 20: 401-408. 

Carpenter, J.H., 1965. The Chesapeake Bay Institute technique for the Winkler 
    dissolved oxygen method. Limnol. Oceanogr., 10: 141-143. 

Carter, D.J.T., 1980. Echo-sounding correction tables. Third Edition. 
    Hydrographic Department, Ministry of Defense, Taunton, U.K., NP 139: 150 pp. 

Fofonoff, N.P., 1977. Computation of potential temperature of seawater for an 
    arbitrary reference pressure. Deep-Sea Res., 24: 489-491. 

Saunders, P.M., 1981. Practical conversion of pressure to depth. J. Phys. 
    Oceanogr., 11: 573-574. 

UNESCO, 1981a. Background papers and supporting data on the Practical 
    Salinity Scale 1978. UNESCO Tech. Pap. in Mar. Sci., No. 37, 144 pp. 

UNESCO, 1981b. Background papers and supporting data on the International 
    Equation of State 1980. UNESCO Tech. Pap. in Mar. Sci., No. 38, 38 pp. 

Weiss, R.F., 1970. The solubility of nitrogen, oxygen and argon in water and 
    seawater.  Deep-Sea Res., 17: 721-735. 



TABLE 1: Ancillary Sampling on AJAX Expedition 

                Measurement:                Investigator:
                -------------------------   ------------------
                Underway pCO2, pCH4, pN20   R.F. Weiss, SIO
                Chlorofluoromethanes        R.F. Weiss, SIO
                Tritium                     W.J. Jenkins, WHOI
                3He                         W.J. Jenkins, WHOI
                Alkalinity, CO2, pCO2       T. Takahashi, LDGO
                                            D. Chipman, LDGO
                Pb                          E. Boyle, MIT
                Trace Metals                E. Boyle, MIT
                                            K. Bruland, UCSC
                                            J. Edmond, MIT
                Rare Earths                 G. Wasserburg, CIT
                                            D. Pipegras, CIT
                Nd Isotopes                 G. Wasserburg, CIT
                                            D. Pipegras, CIT
                Pb-21O                      K. Turekian, Yale
                Surface Ra-228/226          W. Moore, USC
                Photosynthetic Pigments     J.M. Brooks, TAMU


TABLE 2:  WMO Weather Code 4501 

                Code:  
                 0  Clear (no cloud at any level) 
                 1  Partly cloudy (clouds scattered or broken) 
                 2  Continuous layer(s) of c1oud(s) 
                 3  Sandstorm, dust storm, or blowing snow 
                 4  Fog, thick dust, or haze 
                 5  Drizzle 
                 6  Rain 
                 7  Snow, or rain and snow mixed 
                 8  Shower(s) 
                 9  Thunderstorm(s) 



AJAX LEG I PERSONNEL

        SHIP'S CAPTAIN 
           Casiles, David F., R/V Knorr 
          
        PERSONNEL PARTICIPATING IN THE COLLECTION OF DATA 
           Reid, Joseph L.         Chief Scientist, Professor, SIO
           Beaupre, Marie-Claude   Staff Research Associate, SIO 
           Chipman, David W.       Research Associate, LDGO 
           Conway, Carol B.        Staff Research Associate, SIO 
           Costello, James P.      Staff Research Associate, SIO 
           Field, Timothy J.       Marine Technician, SIO 
           Mantyla, Arnold W.      Specialist in Oceanography, SIO 
           Mattson, Carl W.        Electronics Technician, SIO 
           Muus, David A.          Staff Research Associate, SIO 
           Shen, Glen T.           Graduate Research Assistant, MIT 
           Van Woy, Frederick A.   Staff Research Associate, SIO 
           Warner, Mark J.         Research Assistant, SIO 
           Wells, James A.         Marine Technician, SIO 


AJAX LEG II PERSONNEL

        SHIP'S CAPTAIN 	
            Bowan, Richard, R/V Knorr 	
        	
        PERSONNEL PARTICIPATING IN THE COLLECTION OF DATA 
           Nowlin, Worth D., Jr.    Chief Scientist, Professor, TAMU 
           Bidigare, Robert R.      Assistant Research Scientist, TAMU 
           Bos, David L.            Staff Research Associate, SIO 
           Bullister, John L.       Research Assistant, SIO 
           Chipman, David W.        Research Associate, LDGO 
           Conway, Carol B.         Staff Research Associate, SIO 
           Costello, James P.       Staff Research Associate, SIO 
           Field, Timothy J.        Marine Technician, SIO 
           Frank, Tamara            Graduate Research Assistant, TAMU 
           Richter, Walter A.       Electronics Technician, SIO 
           Rintoul, Stephen R.      Graduate Research Assistant, WHOI/MIT 
           Trull, Thomas W.         Graduate Research Assistant, WHOI/MIT 
           Van Woy, Frederick A.    Research Associate, SIO 
           Wells, James A.          Marine Technician, SIO 
           Whitworth, Thomas, III   Associate Research Scientist, TAMU 
           Williams, Robert T.      Principle ADP Systems Analyst, SIO 
           Worley, Steven J.        Research Associate, TAMU 
           Zastrow, Colleen E.      Graduate Research Assistant, TAMU
           

Figure 1:  AJAX Expedition Station Positions



CARBON CHEMISTRY OF THE SOUTH ATLANTIC OCEAN AND THE WEDDELL SEA

David W. Chipman, Taro Takahashi, Steward C. Sutherland
Lamont-Doherty Geological Observatory
December, 1986


ACKNOWLEDGMENTS

We wish to express our thanks to Joseph L. Reid of Scripps Institution of 
Oceanography and Worth D. Nowlin, Jr., of Texas A&M University, for the 
opportunity to take part in the AJAX expedition, and to the officers and crew 
of the R/V Knorr and to other members of the scientific group on board, 
especially the group from the PACODF of SIO, for their generous assistance 
during the expedition. The carbonate chemistry measurements reported here 
were made possible through a grant from the National Science Foundation (OCE 
83-09987).




SECTION  I      EXPERIMENTAL PROCEDURES


INTRODUCTION

This report summarizes the seawater carbonate chemistry data which were 
obtained as part of the Long Lines (AJAX) Expedition in the South Atlantic 
Ocean and Weddell Sea, October-November 1983 and January-February 1984 on 
board the R/V Knorr. The partial pressure of CO2 (pCO2) and the concentration 
of total dissolved carbon dioxide (total CO2 or TCO2) were measured at sea on 
about 750 selected samples from most of the 138 hydrographic stations which 
were occupied during the two legs of the cruise. An additional 120 water 
samples were collected and analyzed for TCO2 in our land-based laboratory 
using the new technique of CO2 coulometry. All the analyses were performed at 
least in duplicates. In addition, the total alkalinity (TALK) has been 
computed for all of the samples for which the necessary pCO2, TCO2, salinity 
and nutrient concentration values have been determined. The experimental 
methods used for this study, the calibration techniques and precision of the 
measurements are discussed in detail.

Vertical sections for the pCO2, TCO2, TALK, and apparent oxygen utilization 
(AOU) along the Greenwich meridian from 50°N to 70°S are presented, for the 
depth ranges 0-1000 meters and 0-6000 meters (Figures 7-10).

The relationships between the various measured properties are indicated on 
property-property plots for samples throughout the water column (Figures 
11-20). 



EXPERIMENTAL METHODS

Gas Chromatograph System for pCO2 Analysis

The equilibrator-gas chromatograph system used during the expedition for the 
determination of partial pressure of CO2 was similar to the one which was used 
during the TTO-North Atlantic and TTO-Tropical Atlantic expeditions, and has 
been described elsewhere (Takahashi et al., 1982; Smethie et al., 1985) (see 
Figure 1). Briefly, water samples for analysis are drawn from the 10-liter 
Niskin samplers of a rosette cast directly into 500-ml narrow-necked Pyrex 
flasks which serve both as sample containers and equilibration vessels. The 
samples are poisoned with 1/4 ml of saturated mercuric chloride solution to 
prevent biological modification of the pCO2, and are stored in the dark until 
measurement, which normally was performed within 48 hours of sampling. A 
headspace of 3 to 5 ml was left above the water in the flasks to allow for 
thermal expansion during storage. The flasks are sealed air-tight using 
screw-caps with conical plastic liners.


Figure 1: Schematic diagram of gas chromatograph-based system for the 
          determination of CO2 partial pressure in seawater. Single solid 
          lines represent gas flow pathways, dotted lines represent data and 
          control signal paths, and solid double lines enclose thermally 
          isolated zones. The valves are shown in the orientation they would 
          have just after an equilibrated gas sample is injected into the 
          carrier gas stream of the gas chromatograph for analysis. After the 
          CO2 peak elutes from the precolumn into the analytical column, the 
          10-port valve will be returned to its normal position, backflusing 
          the water vapor and any hydrocarbons heavier than methane from the 
          system while connecting the sample loop to the equilibration 
          subsystem to prepare for the next sample.


Prior to analysis, the sample flasks are brought to 20.00 °C in a 
thermostatted water bath, and about 65 ml of the water is displaced with air 
of known CO2 concentration. The air in the flasks and in the tubing connecting 
the flasks to the gas chromatograph sample loop is recirculated continuously 
for approximately 20 minutes, with a gas disperser about 1 cm below the water 
surface providing large contact area between water end air bubbles. At the 
end of the equilibration period, the circulation pump is switched off and the 
air pressure throughout the system is allowed to equalize. A 1-ml aliquot of 
the equilibrated air is isolated from the equilibration subsystem and 
injected into the carrier gas stream of the gas chromatograph by cycling the 
gas sampling valve to which the sample loop is attached. The gas 
chromatograph, a Perkin-Elmer Model Sigma-4, uses hydrogen as the carrier gas 
and is equipped with a 2-meter column of Chromosorb 102 to separate the CO2 
from the other components of the air. After separation, the CO2 is converted 
into methane and water vapor by reaction with the hydrogen carrier in a 
catalytic converter of ruthenium operated at 380°C, in a manner similar to 
that described by Weiss (1981), but without the use of a palladium 
pre-catalyst. The methane produced by this reaction is then measured with a 
precision of ±0.05% (one standard deviation) using a flame ionization 
detector. The signal from the flame ionization detector is fed into a 
Perkin-Elmer Model Sigma-10 digital integrator, where the area of the CO2 peak 
is computed, and the concentration of CO2 in the sample is determined by 
comparison with the peak areas of known amounts of CO2 from reference gas 
mixtures. The GC detector response is calibrated at least once per hour by 
injecting, with the same sample loop, CO2-air mixtures which are calibrated 
against the World Meteorological Organization standards of C.D. Keeling.

The equilibrated air samples are saturated with water vapor at the 
temperature of equilibration and have the same pCO2 as the water sample. By 
injecting the air aliquot at the pressure of equilibration and without 
removing the water vapor, the partial pressure of CO2 is determined directly, 
without the need to know either the pressure of equilibration or the water 
vapor pressure (Takahashi et al., 1982). It is necessary to know the pressure 
of the calibration gas mixtures, which is done by venting the sample loop to 
atmospheric pressure after filling and measuring the atmospheric pressure by 
means of a high-accuracy electronic barometer (Setra Systems, Inc., Model 
270, accuracy ± 0.3 millibars, calibration traceable to NBS provided by 
manufacturer). Additional corrections are required to account for the change 
in pCO2 of the sample water due to the transfer of CO2 to or from the water 
during the equilibration with the recirculating air, and to account for the 
difference in pressure between the air in the equilibrator when the pump is 
running and that in the GC sample loop when the pump is off. The overall 
precision of the pCO2 measurement is estimated to be about ± 0.2%.


GC and Coulometric Systems for Total CO2 Measurements

Two independent methods were used for the measurement of total CO2 (TCO2) in 
seawater: determination at sea using the gas chromatograph described above 
with a separate CO2-extraction subsystem, and that in the shore-based 
laboratory of stored water samples using the new technique of CO-cou1ometry. 
Both techniques will be described below.

The gas-chromatograph system used on board the ship consisted of an extraction 
system for removing the CO2 from acidified 3-ml water samples linked to the 
same gas chromatograph that was used for the pCO2 analyses (see Figure 2). 
Water samples were drawn from the Niskin samplers into 125-ml glass bottles 
with ground-glass stoppers, which were greased with silicone stopcock grease. 
The bottles were filled with three rinses and at least one volume of 
overflow, 2 to 3 ml of the water was removed to provide a headspace for 
thermal expansion, and the stoppers were held in place with strong rubber 
bands. About 1/8 ml of saturated mercuric chloride solution was added to 
prevent biological alteration of the TCO2. An attempt was also made to draw 
samples using 60-ml plastic syringes equipped with plastic valves, but the 
samples were found to become contaminated with CO2 (presumably dissolved in 
the rubber end of the plunger) In a relatively short period of time, so that 
method was abandoned. For analysis, a metal (Hastelloy-C) sample loop of 
approximately 3 ml volume was filled with sample, with about 15-20 ml used for 
rinsing the tubing and loop, and the loop was connected to the recirculating 
carrier gas of the extraction system (CO2-free air). At the same time, a 1/3 
ml loop filled with 1N hydrochloric acid was similarly connected. The carrier 
gas forced the acid through the water-sample loop and the acidified water was 
then forced into a stripping column containing a coarse glass frit near the 
bottom. A small gas pump continually recirculated the gas through the 
acidified water and through a dilution volume of about 300 ml, until the CO2 
was thoroughly equilibrated between water and gas and the gas was well 
homogenized (approximately 6 minutes). During this period, the system was 
connected to the sample valve of the GC, so that the sample loop was 
continuously flushed and contained a representative aliquot of the gas at the 
end of the period. Since the air recirculates through the water, a small 
amount (approximately 1%) of the CO2 in the system remains dissolved in the 
acidified water. The maximum temperature variation due to changes in room 
temperature, 4°C, would cause a variation in the fraction of the CO2 in the 
water of approximately 10% of the amount retained, or about 2 umol/kg. The 
maxima salinity variation would cause a much smaller effect. Since the 
fraction remaining in the water is relatively small and nearly constant, the 
calibration procedure will allow variations in this amount of CO2 to be 
ignored. The circulation pump was then switched off and the gas sampling 
valve was switched to inject the aliquot into GC carrier gas stream. By 
selecting the appropriate volumes for the water sample and the dilution 
volume, the amount of CO2 introduced into the GC was kept close to the amount 
in one of the calibration gas mixtures (about 789 ppm).

In order to determine the accuracy of the shipboard TCO2 analyses, a number of 
samples were taken in 500-ml glass bottles with ground-glass stoppers in the 
same manner as the smaller samples taken for shipboard analysis. These 
samples were then shipped back to our shore-based laboratory for analysis 
using a totally independent method, CO2 coulometry. The basis of this method 
is described in a paper by Johnson et al. (1985). Briefly summarized, the 
method is as follows.

A sample of the seawater for analysis is measured by filling a calibrated 
sample pipet (of approximately 50 ml volume) with sufficient overflow to 
insure the thorough rinsing of the pipet and tubing, transfer-ring the water 
to an extraction tube, acidifying, and sweeping the evolved CO2 into the cell 
of a CO2 coulometer with a flow of CO2-free carrier gas. In the coulometer 
cell, the CO2 is quantitatively absorbed by a solution of ethanolamine in 
dimethylsulfoxide (DMSO). Reaction of the CO2 with the ethanolamine forms the 
weak acid hydroxyethylcarbamic acid. The pH change associated with the 
formation of this acid results in a color change of thymolphthalein in the 
solution. The color change, from deep blue to colorless, is detected by a 
photodiode, which continually monitors the transmissivity of the solution. 
The electronic circuitry of the coulometer, on detecting the change in the 
color of the pH indicator, causes a current to be passed through the cell, 
electro-generating hydoxyl (OH-) ions from a small amount of water in the 
solution. The OH-generated titrates the acid, returning the solution to its 
original pH (and hence color), at which point the circuitry interrupts the 
current flow. The product of current passed through the cell and time is 
related by the Faraday constant to the number of moles of OH-generated to 
titrate the acid and hence to the number of moles of CO2 absorbed to form the 
acid.


Figure 2: Schematic diagram of the CO2 extraction subsystem of the gas 
          chromatograph TCO2 analysis system. In operation, the lines shown 
          connected to the CC sample loop are attached to the 10-port valve 
          shown in Figure 1, in place of the pCO2, equilibration subsystem. 
          The valves on Figure 2 are shown in the position they would have 
          while CO2 is being stripped from the acidified seawater sample. 
          Valves V2 and V3, which allow the introduction of metered 
          quantities of hydrochloric acid and seawater respectively, have 
          been returned to the fill position to allow for the preparation of 
          the next sample. After cycling the 10-port valve (Figure 1) to 
          inject an aliquot of CO2-air mixture for analysis, V1 will be turned 
          to alternate position, which will cause CO2-free air to sweep the 
          system of CO2 and force the stripped acidified water in the 
          stripping-column out through V1 by way of V2 and V3 to a waste 
          reservoir.


We have calibrated the coulometer in four different ways: by injecting 
measured volumes of pure CO2 gas at known pressure and temperature, by 
analysis of the CO2 evolved from gravimetrically prepared solid calcium 
carbonate and sodium carbonate, and by injecting measured volumes of CO2-air 
mixtures (WMO-calibrated mixtures referred to above) at known temperature and 
pressure. With the exception of the last (which provides relatively small 
quantities of CO2 and hence is susceptible to small errors in the blank 
determination, and which differed by 0.2%), all of the calibration techniques 
agreed with the electrical calibration of the instrument (i.e. independently 
measuring current and time and comparing with the readout) to within 0.1%. We 
have also observed that the calibration of the coulometer has changed by no 
more than 0.1% over the 2 1/2 years we have used it. With care, the 
coulometer is capable of a precision of better than ± 1 uM/liter in samples 
of 2200 uM/liter.


Atmospheric CO2 measurements

Air samples were analyzed for CO2 concentration by filling the GC sample loop 
with air drawn from near either the ship's bow or stern (depending on the 
relative wind direction) using a metal-lined plastic sampling line and small 
bellows pump. The air sample was introduced directly into the sample loop 
without drying, and the excess pressure in the loop was allowed to vent to 
the atmosphere prior to injection into the GC for analysis. The calibration 
of the GC with standard gas mixtures at the same pressure allows the 
concentration of CO2 in the sample to be computed directly. Table 1 lists the 
concentration of CO2 in atmospheric samples taken during the two legs of the 
cruise. These values have been recalculated to give the concentration as a 
mole fraction of CO2 in dried air, the mole fraction in air which is saturated 
with water vapor at the temperature of the sea surface, and the partial 
pressure of CO2 in the air at the conditions of the sea surface (water 
saturated at sea surface temperature and at the ambient atmospheric 
pressure).


Table 1: Atmospheric concentration of CO2 observed during the two legs of AJAX 
         cruise. Samples were analyzed without removal of water vapor, and 
         the concentration in dry air (VCO2) have been computed using the 
         observed wet and dry bulb thermometer readings. The CO2 concentration 
         in air saturated with water vapor at the temperature of the sea surface 
         (VCO2*), and the partial pressure of CO2 in water-saturated air (pCO2) 
         at the barometric pressure which was observed at the of analysis (pCO2) 
         are also given.

     DATE      TIME      LAT         LONG     VCO2  VCO2*   PRESS   pCO2
          (GMT)             (DEG MIN)            (ppm)       (mb)  (uatm)
     --------  ----   ---------   ---------   ---   ----   ------   ----
     10/08/83  1800   02 59.6 N   03 46.2 W   345   333    1012.6   333
     10/09/83  0945   01 30.1 N   03 37.9 W   344   334    1012.8   334
     10/10/83  2200   01 30.3 S   03 20.0 W   345   335    1012.9   335
     10/11/83  1630   03 00.2 S   03 12.8 W   346   336    1014.1   336
     10/13/83  2200   07 33.0 S   01 07.8 W   342   333    1016.5   334
     10/15/83  0100   09 45.6 S   00 46.0 E   341   333    1015.5   333
     10/16/83  0300   11 59.9 S   00 51.7 E   341   333    1015.7   334
     10/16/83  2000   14 00.2 S   00 57.6 E   341   334    1017.0   335
     10/20/83  1830   21 01.4 S   01 19.2 E   343   335    1021.8   338
     10/23/83  1230   26 59.5 S   01 39.8 E   343   335    1026.2   339
     10/29/83  0930   39 00.5 S   00 59.2 E   342   337    1021.5   340
     11/01/83  2100   42 48.3 S   04 52.8 E   341   337    1014.8   338
     01/20/84  1900   51 50.2 S   01 11.6 E   342   339     990.0   332
     01/21/84  1530   53 49.0 S   01 20.2 E   343   340     967.6   325
     01/25/84  0300   61 00.2 S   00 52.7 E   343   341     993.2   334
     01/25/84  2130   62 00.4 S   00 44.9 E   342   339     996.8   334
     01/26/84  1500   64 00.1 S   00 20.4 E   342   340     993.9   333
     01/29/84  0300   69 21.8 S   00 19.1 W   342   340    1002.6   337
     02/01/84  1600   61 29.4 S   16 41.4 W   342   340     996.5   334
     02/05/84  2100   58 40.9 S   26 50.6 W   341   338    1005.7   336
     02/08/84  0100   56 48.5 S   34 17.8 W   341   338     983.2   328
     02/13/84  0100   59 45.7 S   48 55.7 W   342   339     997.4   334
     02/15/84  1300   60 48.9 5   55 38.6 W   342   340     993.9   334



CALIBRATION AND DATA REDUCTION METHODS

The methods used to calculate pCO2 or TCO2 from raw GC peak areas differed 
somewhat from the procedure which was followed in the past and requires a 
detailed discussion. The separate parts of the procedure are: 1) calibration 
of the GC against two or three standard gas mixtures at known pressure, to 
allow the pCO2 of the air in the sample loop to be determined from the CO2 
peak area; 2) for TCO2, the relative volumes of the water sample loop, air 
sample loop and CO2-extraction system are determined by measurement of the CO2 
evolved from samples of gravimetrically prepared sodium carbonate solution; 
and 3) also for TCO2 measurements, a correction for the extraction efficiency 
of the stripping system was applied using the comparison between the TCO2 
measured with the GC and that measured with CO2 coulometry.

Calibration of the GC

During Leg 1 of the expedition, the flame ionization detector (FID) response 
has been approximated by a straight line curve through the values of the two 
standard gas mixtures (294 and 789 ppm). During Leg 2, this same procedure 
was used for only the TCO2 analyses, while pCO2 measurements were calculated 
from a response curve which was parabolic and passed through the values of 
all three standard gas mixtures (CO2 concentrations of 294, 789 and 1388 ppm).

The response curve of a FID can be made to be very nearly linear over a wide 
range of concentrations. With the accumulation of deposits on the collector 
and with changes in the ratio of hydrogen to air feeding the flame, however, 
the response can become significantly non-linear. Figure 3 demonstrates the 
amount of error which would be introduced by the use of a linear rather than 
parabolic approximation to the actual detector response. The difference 
plotted is the concentration calculated from a parabolic curve less that 
calculated from a linear curve for a given peak area, divided by the "linear" 
concentration and multiplied by 100 to convert it into a percent difference. 
If the response were perfectly linear, the curve would fall, on the 0.0 
difference line. As can be seen from Figure 3a, which applies to Leg 1 data, 
the maximum error resulting from using the 2-standard linear approximation is 
no greater than 0.25%, or 4 ppm at a value of 1600 ppm. All of the TCO2 
analyses gave concentrations which were clustered closely around the 
concentration of the intermediate standard (789 ppm), and the maximum 
difference is consequently much smaller, less than 0.05 (equivalent to 1 
uM/kg) for all, samples. As shown in Figure 3b, the FID response curve was 
much more non-linear during Leg 2 compared to Leg 1 (note the change in scale 
of the vertical axis). During this leg, all pCO2 analyses, which range from 
approximately 300 to 1200 ppm, were computed using a 3-point response curve 
and consequently the large potential errors (up to l at 1600 ppm) do not 
apply. Only the central portion of the figure, representing the range of TCO2 
analyses, is meaningful, and again the possible errors in this range are 
small, being no greater than 0.08 (or less than 2 uM/kg).

In the short term, the detector response can change rapidly in response to 
changes in the laboratory temperature, and occasionally this change was as 
great as l% during the interval between successive calibration sequences 
(less than one hour). Rapid changes in the shipboard laboratory temperature 
were experienced as the outside door was often opened to reduce the level of 
freon contamination in the laboratory air. In order to reduce the effects of 
this change in detector response, we assume that the detector drift is linear 
with time and compute the detector response curve (either linear or 
parabolic) using the drift-corrected peak areas for the standard gas mixtures 
at the time of injection of the unknown by interpolating between the analyses 
of each standard from the calibration sequences preceding and following the 
unknown.


Figure 3: The difference between the CO2 concentrations estimated using 
          assumed linear and parabolic detector responses. The zero reference 
          line represents a linear fit to two calibration points, and the 
          curved lines represent the upper and lower limits of parabolic fits 
          to three calibration points, 
          a.) This panel applies to the measurements during Leg 1. The 
              vertical dashed and solid lines indicate the ranges of measured 
              pCO2 and TCO2 values respectively. It is seen that the 
              difference between the two-point linear calibration and the 
              three-point parabolic calibration is small: less than 0.02% for 
              TCO2 and less than 0.2% for pCO2. 
          b.) The curves represent the upper and lower limits observed during 
              Leg II, indicating that the detector response during Leg II was 
              more non-linear than that during Leg I. Therefore, for the pCO2 
              measurements during Leg II, parabolic fits to three calibration 
              points were used. For the TCO2 measurements demarked by solid 
              vertical lines, linear fits to two calibration points were used.



Calibration of TCO2 extraction system

The number of moles of CO2 in the GC sample loop at the time of injection of a 
TCO2 sample is a function of the volume of the GC sample loop, the volume of 
seawater metered into the extraction system by the seawater sample loop, and 
the temperature and volume of the gas volume of the extraction system, as 
well as the concentration of CO2 in the seawater sample. The temperature of 
the GC sample loop is constant, being located in the well-thermostatted GC 
column oven, and the remainder of the gas volume of the extraction system is 
submersed in a water bath, the temperature of which is monitored and recorded 
at the time of injection of each sample. Rather than determine the absolute 
volumes of the various parts of the extraction system, we have chosen to use 
solutions of known CO2 concentration to establish the relationship between CO2 
concentration of the sample and the number of moles of CO2 in the sample loop 
at the time of injection into the GC. Sodium carbonate powders (dried at 
180°C in air for about 16 hours) were weighed in our land-based laboratory, 
wrapped in aluminum foil, sealed in air-tight plastic vials and stored in 
silica gel desiccant for use in checking the extraction system during the 
cruise. At sea, solutions of known CO2 concentration were prepared by opening 
the foil packets and placing the entire packet in a volumetric flask filled 
to the reference line with low-CO2 water. During Leg 1, this water was 
prepared by stripping distilled water with CO2-free air overnight; during Leg 
2 It was found to be much easier to prepare the water as needed by passing 
water from the ship's evaporator through a pair of deionization columns 
(Cole-Parmer Research cartridge or equivalent). Regardless of the method used 
to produce the water, the concentration of CO2 was checked by running an 
aliquot as an unknown, and the final concentration of CO2 in the standard 
solutions was corrected for this residual CO2 . The standard solutions, once 
made up, could not be successfully stored for more than a few hours, and 
consequently only the initial calibration determined from a given solution 
was used. During the second leg of the expedition, drift in the calibration 
of the extraction system, possibly due to a progressive fouling of the 
seawater sample loop, amounted to a nearly linear decrease of about 0.75% 
over a period of 33 days. Due to difficulty experienced in preparing standard 
solutions on shipboard during the first leg, the drift during that leg is not 
as certain, but appears to have been of a similar magnitude as during the 
second leg. All TCO2 values have been corrected for this apparent change in 
the calibration factor of the extraction system by fitting the calibration 
factors computed from all calibration runs with a linear regression against 
cumulative run number, then using the equation of the resulting line to 
calculate system calibration factor to be applied to each analysis. Since all 
the calibration runs have been used to establish the calibration curve, only 
the effects of long-term variation in the volumes of the extraction system 
are removed by this procedure, while those due to short-term variations, such 
as in the extraction efficiency, may remain.


Extraction efficiency correction

A more serious problem than the long-term drift of the calibration factor of 
the extraction system was an occasional slight decrease in the extraction 
efficiency, apparently due to the accumulation of salt on the check valves of 
the circulation pump. Whenever the pumping rate was sufficiently reduced to 
be obvious, the pump was disassembled and cleaned, but before this stage was 
reached, the extraction efficiency appears to have been reduced to such an 
extent that the amount of time allowed for the extraction was insufficient, 
and the amount of CO2 in the recirculating gas was slightly low. This effect 
was not noticed during the cruise, but became evident when the TCO2 
measurements made by GC were compared with the coulometric analyses made on 
stored duplicate samples. There are three arguments which support the 
superiority of the coulometer data. First, as mentioned earlier, our 
coulometer calibration is consistent with four independent methods and does 
not appear to change with time. Secondly, the coulometer data are more 
precise (i.e. ±0.9 uM/kg, as shown in Figure 5). Thirdly, our long-term 
sample storage tests show that the total CO2 concentration in the poisoned 
samples is stably preserved for several months. For these reasons, we believe 
that the coulometric analyses of stored samples from the cruise give accurate 
values, and where these values do not agree with the shipboard GC values we 
feel justified in correcting the latter. We have attempted to correct the GC 
analyses for this effect by calculating the average difference between GC and 
coulometric analyses for blocks of GC analyses (all the analyses made during 
one analytical session, between periods when the GC was being used for pCO2 
analysis) and applying to the GC values as a multiplier the factor necessary 
to make this average difference equal zero. The largest correction required 
was less than 1.5%, end in general the correction was less than 0.3%.

Figure 4 shows a comparison of the coulometer TCO2 data with the GC TCO2 
values thus corrected. A 1:1 linear correlation is observed. The r.m.s. 
deviation of the data points about this trend line is approximately ± 5.4 
uM/kg. This represents a realistic estimate of the over-all precision 
including the inter-station variability. However, as shown in Figure 5, the 
intra-station precision of the GC analyses is about ± 3 uM/kg (root mean 
square deviation).


COMPARISON WITH THE RESULTS OF OTHER EXPEDITIONS

Figure 5 shows a plot of TCO2 versus depth for a location at approximately 
60°S, 1°E observed during the following three expeditions: Station 83 and 84 
of this cruise (AJAX, 24 January 1984), GEOSECS Station 89 (23 January 1973) 
and Station 241 of the 1986 Winter Weddell Sea Project (WWSP, 18 July 1986). 
The GEOSECS TCO2 values, as determined using the potentiometric alkalinity 
titrator, have been plotted as originally reported. However, a recent 
comparison of the GEOSECS TCO2 measurements in the North Atlantic with those 
computed from the alkalinity and pCO2 measurements made at the same location 
during the TTO/North Atlantic Study indicates that the Atlantic GEOSECS TCO2 
values are systematically in error and need to be corrected by subtracting a 
constant 14 micromoles/kg (Takahashi et al., 1985). Dashed curves on the 
figure represent the least squares parabolic fit to the GEOSECS TCO2 data 
points for all depths greater than 750 meters (short dashes for the original 
analyses, long dashes for the points corrected by -14 uM/kg). The WWSP TCO2 
values were obtained recently using shipboard CO2 coulometry (Chipman and 
Takahashi, 1986, unpublished data) and are completely independent of the 
GEOSECS and TTO/NAS data sets. For comparison, the seven WWSP data points 
from a similar depth range have likewise been fitted with a least squares 
parabola, shown on the figure as a solid curve. The superior quality of this 
data set is demonstrated by the r.m.s. deviation of ±0.9 uM/kg for this 
parabolic fit. The near coincidence of these two concentration profiles 
(average difference between the curves at depths greater than 700 meters is -
0.4 ± 3.7 uM/kg) gives confidence that the correction applied to the GEOSECS 
data is indeed justified. The stored TCO2 samples from AJAX Station 84, 
analyzed by coulometry, plot at slightly lower concentrations than the WWSP 
samples, with an average offset of about -1.7 ± 2.8 uM/kg for the six 
samples. The shipboard GC analyses from this station show very close 
agreement with the other analyses, with the average deviation from the WWSP 
trend being -1.7 ± 3.0 uM/kg. In addition, the analyses from Station 83, 
which is located within one degree of Station 84, similarly agree very well 
(average difference is -1.0 ± 3.4 uM/kg). Although the GC values for both of 
these stations have been plotted as corrected using the coulometric analyses, 
the correction increased the values by less than 1 uM/kg, and consequently 
the agreement with the other data sets is meaningful.


Figure 4: Comparison of the values of total CO2 concentration (TCO2) determined 
          on duplicate pairs of samples by means of gas chromatography (CC) 
          and shore-based coulometry. The coulometric measurements have been 
          used to provide the over-all calibration of the GC-based system; 
          consequently the points necessarily plot near the 1:1 line. Most of 
          the scatter of the data points about this line (i.e. a r.m.s. 
          deviation of 5.4 uM/kg) is attributed to random errors in the 
          GC-based measurements.

Figure 5: Comparison of the total CO2 concentration data obtained at GEOSECS 
          Station 89, Winter Weddell Sea Project (WWSP) Station 5 and Long 
          Lines (AJAX Station 83 and 84. These stations are located at about 
          60°S and 1°E. The GEOSECS data (+) were obtained by means of a 
          potentiometirc titration method; the WWSP data (•) :were obtained 
          by means of a coulometric method; and the AJAX data were obtained 
          by means of GC (∆ and solid ∆) and coulometric (upside-down ∆) 
          methods. The solid curve indicates a parabolic least squares fit to 
          the seven WWSP data points with a r.m.s. deviation of about 0.9 uM/kg. 
          The AJAX data, although scatter more widely, are consistent with the  
          WWSP data. The short-dashed curve represents a parabolic least squares 
          fit to the GEOSECS data with a r.m.s. deviation of about 4 uM/kg. When 
          the GEOSECS data are corrected by -14 uM/kg as suggested by Takahashi 
          et al. (1986), the corrected values are consistent with the WWSP data 
          as indicated by the long-dashed curve.



REFERENCES TO SECTION 1

Johnson, N.M., A.E. King and J.McN. Sieburth, 1985, Coulometric TCO2 analyses 
    for marine studies; an introduction. Marine Chem., 16, 61-82.

Smethie, W.M., Jr., T. Takahashi, D. Chipman and J. Ledwell, 1985, Gas 
    exchange and CO2 flux in the tropical Atlantic Ocean determined from 
    Rn-222 and pCO2 measurements. Jour. Geophys Res., 90, 7005-7022.

Takahashi, T., D. Chipman, N. Schechtman, J. Goddard and R. Wanninkhof, 1982, 
    Measurements of the partial pressure of CO2 in discrete water samples 
    during the North Atlantic Expedition, the Transient Tracers of Ocean 
    Project. Technical Report, Lamont-Doherty Geol. Obs., Palisades, N.Y., 
    268 pp.

Takahashi, T., W.S. Broecker and S. Langer, 1985, Redfield ratio based on 
    chemical data from isopycnal surfaces. Jour. Geophys Res., 6907-6924.

Weiss, R.F., 1981, Determinations of carbon dioxide and methane by dual 
    catalyst flame ionization chromatography and nitrous oxide by electron 
    capture chromatography. Jour. Chromatogr. Sci., 19, 611-616.





SECTION  II       MERIDIONAL PROFILES OF CARBON CHEMISTRY AND APPARENT OXYGEN 
                  UTILIZATION ALONG THE PRIME MERIDIAN, 4°N - 70°S


Figure 6:   Station locations for the Long Lines (AJAX) Expedition, 
            October, 1983 through February, 1984.

Figure 7a:  Meridional distribution of the total CO2 concentration (uM/kg) 
            along the prime meridian, 0-1000 meters

Figure 7b:  Meridional distribution of the total CO2 concentration (uM/kg) 
            along the prime meridian, 0-6000 meters

Figure 8a:  Meridional distribution of pCO2 (uatm) in seawater at 20°C along 
            the prime meridian, 0-1000 meters

Figure 8b:  Meridional distribution of pCO2 (uatm) in seawater at 20°C along 
            the prime meridian, 0-6000 meters

Figure 9a:  Meridional distribution of the total alkalinity (ueq/kg) along 
            the prime meridian, 0-1000 meters

Figure 9b:  Meridional distribution of the total alkalinity (ueq/kg) along 
            the prime meridian, 0-6000 meters

Figure 10a: Meridional distribution of the apparent oxygen utilization 
            (uM/kg) along the prime meridian, 0-1000 meters

Figure 10b: Meridional distribution of the apparent oxygen utilization 
            (uM/kg) along the prime meridian, 0-6000 meters






SECTION  III      PROPERTY-PROPERTY RELATIONSHIPS OF THE LONG LINES DATA


Figure 11:  Potential temperature versus salinity

Figure 12:  Total CO2 concentration versus potential temperature

Figure 13:  Total CO2 concentration versus salinity

Figure 14:  pCO2 (at in situ temperature) versus potential temperature

Figure 15:  pCO2 (at in situ temperature) versus salinity

Figure 16:  Total alkalinity versus potential temperature

Figure 17:  Total alkalinity versus salinity

Figure 18:  pCO2 (at 20°C) versus total CO2 concentration

Figure 19:  pCO2 (at 20°C) versus apparent oxygen utilization

Figure 20:  Apparent oxygen utilization versus total CO2 concentration

Figure 21:  Nitrate concentration versus total CO2 concentration

Figure 22:  Phosphate concentration versus total CO2 concentration

Figure 23:  Total alkalinity versus total CO2 concentration

Figure 24:  Silica concentration versus total alkalinity

Figure 25:  Nitrate concentration versus phosphate concentration

Figure 26:  Silica concentration versus nitrate concentration





SECTION  IV      LONG-LINES (AJAX) SOUTH ATLANTIC AND SOUTHERN OCEAN 
                 CARBONATE CHEMISTRY DATA

With the exception of the carbonate chemistry parameters (pCO2, TCO2 and 
TALK), all of the data were obtained by the Physical and Chemical 
Oceanographic Data Facility (PACODF) of the Scripps Institution of 
Oceanography (see "Physical, Chemical and in-situ CTD Data from the AJAX 
Expedition in the South Atlantic Ocean", SIO Reference 85-24 and TAMU 
Reference 85-4-D, 275 pp., 1985).

The carbonate chemistry data obtained by the Lamont-Doherty Geological 
Observatory group consist of the partial pressure of CO2 in seawater (pCO2) 
(measured at a constant temperature of 20.00 °C) and the total dissolved 
carbon dioxide concentration in seawater (TCO2). The TCO2 data are represented 
in micromoles per kilogram (uM/kg). The partial pressure of CO2 in seawater at 
20.00 °C (rather than at in situ or potential temperature) is expressed in 
microatmospheres (uatm). The partial pressure of CO2 at another temperature, 
T1, can be computed by using the following equation:

pCO2(T1) = pCO2(T20°C) exp ((T1-20) x 0.043)

In addition to the observed quantities pCO2 and TCO2, the table includes the 
computed quantity TALK (total alkalinity), in microequivalents per kilogram 
(uEq/kg), and apparent oxygen utilization (AOU), in micromoles per kilogram 
(uM/kg). TALK is computed from the observed pCO2 TCO2 (GC determination only), 
salinity, silica (SiO3) and phosphate (PO4) concentrations, using the computer 
subroutine given in Appendix A. In a few cases TALK appears in the table in 
spite of the absence of pCO2 and/or nutrient concentration values; in these 
cases a second Niskin sampler was tripped at the same depth and the values 
for the missing parameters are taken from the companion sampler. Data for the 
second sampler will normally be found on the immediately preceding or 
following line in the table. AOU has been computed by:

AOU = O2(sat.,8) - O2(obs.),

where O2(sat.,9) is the concentration of oxygen the water would have at its 
potential temperature if saturated with air, and O2(obs.) is the measured 
concentration of oxygen in the sample, both expressed in uM/kg. The value for 
O2(sat.,8) is computed using the solubility data of Weiss (1970).

The computed quantities "potential temperature", "sigma-theta", "sigma-2" and 
"sigma-4" have been calculated from the observed values of temperature, 
salinity and pressure using the UNESCO International Equation of State for 
Seawater (Millero et al., 1980), and the adiabatic temperature gradient 
expression of Fofonoff (1977), based on the results of Bryden (1973).





REFERENCES TO SECTION IV


Bryden, H.L., 1973, New po1ynoiials for thermal expansion, adiabatic 
    temperature gradient and potential temperature of seawater. Deep-Sea 
    Res., 20, 401-408.

Fofonoff, H.P., 1977, Computation of potential temperature of seawater for an 
    arbitrary reference pressure. Deep-Sea Res. 24, 489-491.

Millero, F.J., C.T. Chen, A.L. Bradshaw and K. Schleicher, 1980, A new high 
    pressure equation of state for seawater. Deep-Sea Res. 27, 255-264.

Physical, Chemical and in-situ CTD Data from the AJAX Expedition in the South 
    Atlantic Ocean, 1985, SIO Reference 85-24 and TAMU Reference 85-4-D, 275 pp.

Weiss, R.F., 1970, The solubility of nitrogen, oxygen and argon in water and 
    seawater. Deep-Sea Res. 17, 721-735.







APPENDIX A

C  **********  SUBROUTINE TALKX**********  

C  TO CALCULATE TOTAL ALKALINITY (TALK) FROM TOTAL CO2 AND PCO2

C This is a FORTRAN 77 program using the symbol @ in column C six as a 
continuation mark for a line of code longer than C 72 characters.

C   ********************************************************
C   * PROGRAMMER:  S. C. SUTHERLAND                        *
C   *              LAMONT-DOHERTY GEOLOGICAL OBSERVATORY   *
C   *              PALISADES, NY 10964                     *
C   *              U.S.A.                                  *
C   *              PHONE:  (914) 359 2900 EXTENSION 632    *
C   ********************************************************

SUBROUTINE TALKX (ITEMP,I5ALIM,1PO4,15103,IPCO2,1TCO2,TALKO1J)
REAL TEMP,SALIN.PO4,SI03,PCO2,TCO2,TALKOU,ITENP,ISALIN,1PO4
1,131O3,1PCO2,ITCO2,K1CO2,K2CO2,K6,KCO2,K1PO4,K2PO4,KH2O
2PKB,KSIO3,H2CO3,ASIO3,APO4,AH2O,ACO3,HCO3,CO3,ABORON.TBORON
3,CB,CU,HION
INTEGER I,J,K


C  The function "LOG" is the natural logarithm

C  UNITS:

C  TEMP IN DEG C.
C  SALINITY IN 0/00
C  PO4 IN UMOL/KG
C  SI03 IN UMOL/KG
C  PCO2 IN UATM
C  TCO2 IN UMOL/KG

C  VARIABLES WHICH BEGIN WITH I (IPCO2, ISI02. ETC.) ARE USED AS
C  INPUT VARIABLES, AND ARE MOVED TO WORKING VARIABLES WITHIN THE
C  SUBROUTINE, TO PROTECT THEIR SEPARATION FROM ANY CALLING ROUTINE

C  CONVERT UNITS TO MOLES/KG, ATMOSPHERES, AND DEG K.

     TEMP = ITEMP + 273.15
     SALIN = ISALIN
     PO4 = 1PO4 * 1E-6
     SI03 = 15103 * 1E-6
     PCO2 = IPCO2 * 1E-6
     TCO2 = ITCO2 * 1E-6

C  CALCULATE SOLUBILITY OF CO2 IN SEAWATER   KCO2

     KCO2 = EXP (-60.2409 + 9345.17 / TEMP  23.3585 *
     @  LOG (TEMP/100.) + SALIN * (0.023517 - 2.3656E-4
     * TEMP + 4.7036E-7 * TEMP * TEMP))
C  Calculate solubility constants K1 & K2 of carbonic acid
C  as determined by Mehrbach, et. al. (Mehrbach, C., C.H. Culberson,
C  J.E. Hawley and R.M. Pytkowicz 1973,Limnology and Oceanography,
C  18, 897-907.)

C  K1CO2, K2CO2

     KICO2 = 10. ** (13.7201 - 0.031334 * TEMP - 3235.76 / TEMP
     @ - 1.3E-5 * SALIN * TEMP + 0.1032 * SQRT(SALIN))

C  INTERMEDIATE VARIABLE K6

     K6 = ( -5371.9645 - 1.671221 * TEMP - 0.22913 * SALIN
     @  + 128375.28 / TEMP + 8.0944E-4 * SALIN * TEMP - 2.136
     @  * SALIN / TEMP * ( -18.3802 * LOG (SALIN) + 2194.3055
     @  * LOG (TEMP) + (5617.11 / TEMP) * LOG (SALIN)) / 2.302585)

     K2CO2 = EXP (2.302585 * K6)

C  Calculate dissociation constant of boric acid after Lyman (1956),
C  (Lyman, J. 1956, Buffer Mechanism of seawater,
C  ph.D. Thesis, Univ. of Calif., Los Angeles, 196 pp.)

   KB = 10. ** ( -9.26 + 0.00886 * SALIN + 0.01 * ITEMP)

C  Calculate dissociation constants for phosphoric acid after Kester
C  and Pytkowicz. 1967. (Kester. D.R. and R.N. Pytkowicz 1967,
C  Determination of the apparent dissociation constants of
C  phosphoric acid in seawater.
C  Limnology and Oceanograph, 12, 243-252)

   K1PO4 = EXP ( -9.039 -1450. / TEMP)

   K2PO4 = EXP ( 4.466 - 7276. / TEMP)

C  Calculate dissociation constant for silicic acid after Ingri (1959)
C  (Ingri, N., 1959, Equilibrium studies of polyanions IV. Silicate
C  ions in NaC1 medium, Acta Chemica Scandinavica, 13, 758-775.)

   KSI03 = 4.OE-1O

C  Calculate dissociation of water by the convention used by
C  Culberson and Pytkowicz (1973) (Culberson, C.H. and R.M. Pytkowicz,
C  1973, Ionization of water in seawater, Marine Chemistry,
C    1, 403-417.)

   KH2O = EXP (148.9082 - 13847.26 / TEMP - 23.6521 * LOG (TEMP)
   @  + ( -79.2447 + 3298.72 / TEMP + 12.0408 * LOG (TEMP))
   @  * SQRT(SALIN) - 1.9813E-2 * SALIN

C  Calculate activity of H+ ion by formulation of Takahashi, et.al.
C  (1982a) of Culberson and  Pytkowicz (1973) (Takahashi, T.,
C  R.T. Williams and D.L. Bos 1982a, Carbonate chemistry, in
C  "GEOSECS PACIFIC EXPEDITION", Vol. 3 Hydrographic Data,
C  1973-1974. W.S. Broecker, D.W. Spencer and H. Craig. U.S.
C  Government Printing Office, 78-82.)


   FH = 1.29 - 0.00204 * TEMP + 4.61E-4 * SALIN * SALIN
   @  - 1.48E-6 * SALIN * SALIN * TEMP

C  CALCULATE TOTAL BORON (BASED ON CULKIN, 1965)
C  (Culkin, F., 1965, The major constituents of sea water, in
C  CHEMICAL OCEANOGRAPHY, Vol. I, Chapter 4, First Edition,
C  J.P. Riley and G. Skirrow editors, Academic Press, London,
C  121-161.)

   TBORON = 4.106E-4 * SALIN / 35.

C  MAJOR CALCULATIONS BEGIN HERE:

   H2CO3 = KCO2 * PCO2

   HION = (K1CO2 + SQRT(K1CO2 * K1CO2 + 4. * K1CO2 * K2CO2
   @  * (TCO2 / H2CO3 - 1.))) / ( 2. * (TCO2 / H2CO3 * 1.))

   ACO2 = KCO2 * PCO2 * (K1CO2 / HION + 2. * K1CO2 * K2CO2
   @  / (HION * HION))

   ABORON = KB * TBORON  / (HION + KB)
   ASIO3 = KSIO3 * SI03 / (HION + KSIO3)

   APO4 = PO4 * (1 / (1 + K1PO4 / HION + K1PO4 * K2PO4
   @  /  HION**2) + 2 / (1 + HION / K1PO4 + K2PO4 / HION)
   @  + 3 / ( 1 + HION / K2PO4 + HION * HION / (K1PO4 * K2PO4)

   AH2O = KH2O * FH / HION - HION / FH

   TALKOU = (ACO2 + ABORON + ASIO3 + APO4 + AH2O)/1.E-6

   HCO3 = K1CO2 * H2CO3 / HION

   CO3 = K2CO2 * HCO3 / HION

   RETURN
   END




CHLOROFLUOROCARBON MEASUREMENTS

R.F. Weiss, J.L. Bullister, M.J. Warner, F.A. Van Woy, P.K. Salameh
June, 1990


INTRODUCTION

This report contains the results of measurements of the chlorofluorocarbons 
(CFCs) F-11 (trichlorofluoromethane) and F-12 (dichlorodifluoromethane) 
dissolved in seawater and in the atmosphere as measured during Legs I and H 
of Ajax expedition aboard R/V Knorr of the Woods Hole Oceanographic 
Institution. Leg I extended from Abidjan, Ivory Coast, to Cape Town, South 
Africa, between 7 October and 6 November, 1983, and included Stations 1 - 59. 
F.A. Van Woy and M.J. Warner were the CFC analysts on this leg. Leg II 
extended from Cape Town to Punta Arenas, Chile, between ii January and 19 
February, 1984, and included Stations 60 - 137. J.L. Bullister and F.A. Van 
Woy were the CFC analysts on this leg. Dissolved CFC concentrations were 
measured at most stations, with a total of 3276 analyses, and measurements of 
atmospheric CFC dry air mole fraction were made at regular intervals along 
the cruise track.

Hydrographic data were collected by the Oceanographic Data Facility of the 
Scripps Institution of Oceanography. This report includes hydrographic data 
for only those bottles from which CFC samples were analyzed. A complete 
listing of the hydrographic data is given in the expedition data report 
(Scripps Institution of Oceanography and Texas A & M University, 1985).


CFC MEASUREMENT TECHNIQUES

CFC measurements were carried out by shipboard electron-capture gas 
chromatography according to the methods described by Bullister and Weiss 
(1988). The results have been corrected for sampling and analysis blanks, the 
statistical variations of which can be responsible for occasional negative 
values near the detection limit. Sampling blanks generally decrease at the 
beginning of an expedition, as the equipment becomes cleaner with use. For 
Ajax expedition, the following median F-l1 and F-12 sampling blanks in 
picomoles per kilogram, as determined from analyses of deep waters at low 
latitudes which we believe to be CFC-free, were subtracted from all dissolved 
CFC measurements in the listed station intervals:


                            Stations   F-11    F-12
                            --------  ------  ------
                             1 -  11  0.0228  0.0146
                            12 -  22  0.0074  0.0097
                            23 - 137  0.0034  0.0040


It is important to emphasize that the data have been edited to remove serious 
"flyers" and contaminated samples, and to correct gross numerical errors. 
However, not all of the data have yet been subjected to the level of scrutiny 
associated with careful interpretive work. Readers are therefore requested to 
contact the authors for any revisions in the data which may post-date this 
report, and to draw to our attention any suspected inconsistencies. The 
results are reported on the SIO 1986 calibration scale. The precision (± one 
s.d.) of the seawater measurements is about 1% or about 0.005 pmol/kg, 
whichever is greater, for both CFCs. The precision of the atmospheric 
measurements is about 0.4% for both CFCs. The estimated systematic accuracy 
of the calibrations is about 1.3% for F-11 and 0.5% for F-12.

ATMOSPHERIC CFC DATA

Atmospheric CFC measurements were made at regular time intervals along the 
cruise track, using air pumped continuously from inlets located at the ship's 
bow or stern, depending upon the apparent wind direction. Measurements were 
made either on station or while the ship was underway. Typically a group of 3 
to 4 measurements were made consecutively, with a single geographic position 
being recorded for each group. Atmospheric F- 11 and F-12 concentrations are 
reported as dry air mole fractions.

Atmospheric CFC concentrations at each hydrographic station location were 
determined by averaging measurements taken within ±7 days of a station and 

within a radius starting at 60 km, and increased in steps of 60 km, until a 
minimum of 5 air values were found. These mean values are reported in the 
station listings.

SEAWATER CFC DATA

Dissolved CFC concentrations are reported in picomoles per kilogram. CFC 
saturation percentages relative to the mean atmospheric CFC concentrations 
reported at each station, and assuming a water-saturated atmosphere at the 
potential temperature of the sample and a barometric pressure of 1 
atmosphere, are calculated using the solubility functions of Warner and Weiss 
(1985). F-11/F-12 ratios are included in the station listings for all water 
samples in which the measured F-11 and F-12 concentrations are both greater 
than 0.01 picomoles per kilogram.

Replicate CFC seawater samples were analyzed at routine intervals. These 
generally consisted of two or more syringes drawn from the same Niskin 
bottle, although a single syringe sample may occasionally have been analyzed 
twice. The individual replicate analyses are listed in a separate table, and 
their mean values are reported in the main bottle data listings, annotated 
with an "R".

Potential temperature is calculated from the equations of Fofonoff (1977). 
Potential density is calculated from the International Equation of State of 
Seawater (UNESCO, 1981) using potential temperature in place of in situ 
temperature. Oxygen percentage saturation is calculated from the solubility 
function of Weiss (1970).

The following single-character footnotes appear in the data listings:


  CFC Footnotes                        Hydrography Footnotes
  ----------------------------------   ---------------------------------
  R = mean of replicate measurements   H = from thermometric data (value   
                                           normally from CTD)
  M = manual peak integration          D = from CTD (value normally from 
                                           discrete measurements)
  I = irregular digital integration    U = uncertain data


CFC PROFILE PLOTS

CFC profiles from 5 stations are plotted on each page along with a map 
showing the locations of the 5 stations along the cruise track. Each profile 
is plotted on two concentration scales, differing by a factor of ten, so that 
features of the high-CFC surface waters and the low-CFC deep waters can both 
be seen. AU profiles are plotted on the same depth scale. Note that the 
concentration scales for Stations 1 - 47 are different than for Stations 48 - 
137. A dashed vertical line indicates zero concentration for each scale. The 
bottom depth is indicated where it is known and is on scale. At stations 
where the reported bottom depth is shallower than the deepest reported 
bottle, the bottom has been drawn 10 meters below the deepest bottle.


CONTOUR SECTIONS

The F-11 contour sections were machine-generated using the optimal estimation 
method as adapted by Roemmich (1983). Contours are in units of picomoles per 
kilogram, and are spaced at approximately logarithmic intervals in which each 
successive contour above 0.05 pmol/kg represents a doubling in concentration. 
The bottom has been drawn using the bottom depths reported with the 
hydrographic data. At stations with missing bottom depths, values were 
interpolated from adjacent stations. At stations where the reported bottom 
depth is shallower than the deepest reported bottle, the bottom has been 
drawn 10 meters below the deepest bottle.


ACKNOWLEDGEMENTS

This work could not have been completed without the assistance of many 
colleagues. Chief Scientists J.L. Reid (Leg I) and W.D. Nowlin, Jr. (Leg II) 
organized and supervised the Ajax field program, and we thank them for the 
opportunity to participate. We also thank the Oceanographic Data Facility at 
Scripps for their assistance with all phases of the hydrographic work. We are 
grateful to the officers and crew of the R/V Knorr for their work during the 
expedition. This work was supported by a grant from the Ocean Sciences 
Section of the National Science Foundation.


REFERENCES


Bullister, J.L. and R.F. Weiss (1988) Determination of CCl3F and CC12F2 in 
    seawater and air. Deep-Sea Research 35, 839-853.

Fofonoff, N.P. (1977) Computation of potential temperature of seawater for an 
    arbitrary reference pressure. Deep-Sea Research 24, 489-491.

Roemmich, D. (1983) Optimal estimation of hydrographic station data and 
    derived fields. Journal of Physical Oceanography 13, 1544-1549.

Scripps Institution of Oceanography, The University of California at San 
    Diego and Department of Oceanography, Texas A&M University (1985) 
    Physical, Chemical and in-situ CTD Data from the Ajax Expedition in the 
    South Atlantic Ocean. 275 pp.

UNESCO (1981) Background papers and supporting data on the International 
    Equation of State of Seawater 1980. UNESCO Technical Papers in Marine 
    Science, 38, 192 pp.

Warner, M. J. and R. F. Weiss (1985) Solubilities of chlorofluorocarbons 11 
    and 12 in water and seawater. Deep-Sea Research 32, 1485-1497.



Figures:

AJAX Station Locations where CFCs were Measured

AJAX F-11 Section A Station Locations

AJAX F-11 Section A: F-11 concentrations in pMol/kg contoured as a function 
                     of depth and distance. Station numbers are indicated 
                     along the upper axis. Antarctica is on the left and 
                     Africa is on the right.

Ajax F-11 Section B Station Locations

AJAX F-11 Section B: F- 11 concentrations in pMol/kg contoured as a function 
                     of depth and distance. Station numbers are indicated 
                     along the upper axis.




DATA PROCESSING NOTES

Event Date  Contact            Data Type  Summary
----------  -----------------  ---------  --------------------------------------
2010-04-30  Bartolocci, Danie  BTL        Downloaded from CDIAC
            Grabbed the file below and crunched it through over here. One 
            thing- PCO2 has units of PPM (my brain reads part per million) 
            and not UATM, but I'm not a chemist. Are those units correct?
            Oh yeah, you and Alex both treat the AJAX expedition as one cruise 
            (Knorr 11/5), but we have it under two separate expos. I've just 
            asked Jim if we should be splitting, but was there a reason you 
            kept it as one file/cruise? Leg 1 ended in Nov and leg 2 doesn't 
            start till Jan.

            On Apr 29, 2010, at 2:38 PM, Robert Key wrote:
            > My version of the bottle data is posted at CDIAC
            > bob

2010-05-04  Diggs, Steve       Cruise ID  Split data files into 2 cruises 
            Just talked to Jim, and since there's a cruise in between the two 
            AJAX legs, it's a definite "split". So please split-up the one 
            file from CARINA into two files/cruises.

2010-05-06  Bartolocci, Danie  BTL        Data online 
            2010.05.05 DBK
            • As per Bob Key's email of 2010.04.30, PPM was changed to UATM.
            • Jim Swift confirmed split of original bottle file into two legs.
            • Generated netCDF files and checked in JOA. Zipped files into 
              ajax_316N19831007_nc_ hyd.zip.
            The notes below apply to all formatting and splitting of the 
            original ajax_316N19831007 bottle file.
            Leg one has been placed online. Notes file sent to Jerry. 
            2010.04.30 DBK

            Reformatting notes for the ajax_316N19831007 bottle file. This 
            file was obtained from CDIAC, originally submitted to CDIAC by 
            Bob Key. Original file named 316N19831007.exc.csv

            This file contained both legs and was originally reformatted as
            one file. However, since the cruise was marked by over a month 
            between legs, the files were split into two cruises.
            leg 1: ajax_316N19831007
            leg 2: ajax_316N19940111

            This directory contains leg 1, which was subsequently split off 
            of the originally formatted file containing both legs.

            At this time, it is unclear which file scheme should go online, 
            until resolved, the original file will remain (both legs).

            Following edits were made to parameter and units headers:
            • DBARS to DBAR
            • DEGC to DEG C
            • PCO2_TMP to PCO2TMP

            *NOTE: PCO2 currently has units of PPM. CCHDO accepted units are 
            UATM. Units will stand at this time but will be brought to Bob's 
            and Alex's attention for clarification.

            New parameter RA-8/6 and RA-8/6E were added to the parameter list 
            in order to format check and re-order the file. This parameter is 
            RA 228/226 Ratio and error, W. Moore is PI.

            File was re-ordered using copy_bottle_data.rb 
            Name/date stamp was added.
            Edited file named ajax_316N19831007_hy1.csv

            File was visually checked with JOA. No netCDF files were 
            generated at this time. WOCE formatted file was generated using 
            exchange_to_wocebot.rb. File was visual inspected for format 
            errors and placed online. WOCE file named ajax_316N19831007hy.txt 

