Last Update: 2000.06.26


C.1        CRUISE NARRATIVE


C.1.1      HIGHLIGHTS

Expedition Designation  74AB62_1
WOCE Line Designations  AR07E, AR12

Chief Scientist:        W. John Gould
                        Formerly IOS Wormley but now at
                        Southampton Oceanography Centre
                        Empress Dock
                        Southampton, SO14 3ZH, UK
                        Telephone  44 1703 596208
                        Fax        44 1703 596204
                        Email      wjg@soc.soton.ac.uk

Ship:                   RRS Charles Darwin

Ports of Call:          Barry, S.Wales to Troon, Scotland.

Cruise Dates:           1 August 1991 to 4 September 1991.


C.1.2  CRUISE SUMMARY

Cruise Track

The outbound cruise track extended from south of Ireland to Cape Farewell, 
Greenland, and the inboard track returned to the UK north of Ireland.  Two 
north-south tracks linked these lines.

Number of stations

A total of 96 CTD/Rosette stations were occupied using a General Oceanics 24 
bottle rosette equipped with 24 10-liter Niskin water sample bottles, and a NBIS 
Mk IIIa CTDO instrument with Sea Tech Inc. 1m path transmissometer and IOS DL 
10kHz pinger.  3 stns (No. 79-81) were occupied using a NBIS Mk V CTD.

Sampling

The following water sample measurements were made: salinity, oxygen, total 
nitrate, silicate and CFCs 11, 12, 113.  NOTE THAT NO PHOSPHATE MEASUREMENTS 
WERE MADE.

Float, Drifters and Moorings

None of the above items were launched on this cruise.


C.1.3  LIST OF PRINCIPAL INVESTIGATORS

Name                        Responsibility      Affiliation

P.Saunders/M.Brandon        CTD                 SOC/BAS
S.Bacon/M.Hartman           Salinity            SOC
S.Holley/R.Paylor           Nutrients           SOC
S.Holley/R.Paylor           Oxygen              SOC
D.Smythe-Wright/S.Boswell   CFCs                SOC
R.Tiddy                     Meteorology         SOC/Coventry Polytechnic
J.Gould/L.Povey             XBTs/Bathymetry     SOC
K.Heywood/M.Hartman         ADCP/Navigation     UEA/SOC
J.Gould/L.Povey             Thermosalinograph   SOC


C.1.4  SCIENTIFIC PROGRAMME

The WOCE Implementation Plan (World Climate Research Programme, WCRP-11 and 12, 
WMO/ICSU 1988) calls for the observation of the gyre scale circulation of the N 
Atlantic by means of the occupation of a series of CTD station grids that have 
come to be known as Control Volumes.  The rationale for this observational 
strategy has its origins in the observations of Armi and Stommel, 1977 (Journal 
of Physical Oceanography 13,828-857) who observed changes in the shape and 
strength of the eastern Atlantic subtropical over a period of several months.

The cruise described here is the first occupation of one such Control Volume.  
The data set collected will be used to:

a) determine the changes in distribution and physical and chemical 
  characteristics of water masses in the area from previous surveys, notably 
TTO.

b) to determine the "ages" of the water masses by the measurement of CFC 
   concentrations.

c) to collect a data set that would enable the absolute transport of water 
   masses through the area to be determined together with the fluxes of heat and 
   salt, across the area between the UK and Greenland.

d) to make high quality meteorological measurements from the vessel using the 
   IOSDL Multimet package.

e) to carry out biological sampling in order to determine the relationship 
   between ADCP backscatter strengths and biomass distribution.


C.1.5  REPORTS AND PUBLISHED RESULTS FROM THE CRUISE.

RRS Charles Darwin Cruise 62 01 Aug - 04 Sep 1991 CONVEX-WOCE Control Volume 
   AR12.  Cruise Report No 230 1992 IOS Deacon Laboratory, 60pp.

Hartman, M. 1992.  Shipboard ADCP observations during RRS Charles Darwin Cruise 
   62. IOS Deacon Laboratory, Report No. 298, 27pp.

Read, J.F. and W.J. Gould 1992. Cooling and freshening of the subpolar North 
   Atlantic since the 1960s. Nature, 360(6399), 55-57.

Bacon, S. 1997. Circulation and fluxes in the North Atlantic between Greenland 
   and Ireland. J. Phys. Ocenogr., 27(7), 1420-1435.


C.1.6  LIST OF PARTICIPANTS

GOULD  W John          SOC (Principal Scientist)
BACON Sheldon          SOC
BOSWELL Steve          SOC
BRANDON Mark           BAS
GWILLIAM Pat           SOC
HARTMAN Mark           SOC
HEYWOOD Karen          UEA
HILL Andy              RVS
HOLLEY Sue             SOC
JONES Doriel           RVS
PAULSON Chris          RVS
PAYLOR Rodger          SOC
POVEY Lara             SOC/Univ Surrey
READ Jane              SOC
SAUNDERS Peter         SOC
SMYTHE WRIGHT Denise   SOC
TIDDY Raoul            SOC/Coventry Polytechnic
WOODLEY Colin          RVS


C.2     UNDERWAY MEASUREMENTS

C.2.1   CTD MEASUREMENTS

Introduction

CTD profile and bottle data are presented from the Convex cruise Charles Darwin 
62, as reported by Gould et al. (1992).


Instrumentation

The CTD profiles were taken with a Neil Brown Systems MkIIIb CTD, mounted 
beneath a General Oceanics 24 by 10 litre bottle rosette. The CTD was fitted 
with a pressure sensor, conductivity cell, platinum resistance thermometer, a 
dissolved oxygen sensor, a Sea Tech 100cm path transmissometer and an IOS 10kHz 
pinger. At various times up to 8 SIS (Sensoren Instrumente Systeme) digital 
reversing thermometers (RVS: T213, T220, T254, T260, T262; IOSDL: T204, T399 and 
T400) were attached to the Niskin bottles. 2 SIS digital pressure meters (6132H 
and 6075S) were used. There was no CTD fluorometer.

For the bottle rosette system the bases and tops of the bottles were, 
respectively, 0.75m and 1.55m above the CTD pressure sensor. The digital 
reversing thermometers were mounted 1.38m above the CTD temperature sensor.


Data Acquisition

Lowering rates for the CTD package were generally in the range 0.5-1.0ms-1 but 
could be up to 1.5ms-1. CTD data were logged at 16 frames per second. The CTD 
deck unit passes raw data to a dedicated Level A microcomputer where 1 second 
averages are assembled. During this process the Level A calculates the rate of 
change of temperature and a median sorting routine detects and removes pressure 
spikes. This data is sent to the Level B for archival. The data are then passed 
to a Level C workstation for conversion to Pstar format and calibration.

Water samples were acquired on the up cast with the winch stopped. The CTD data 
acquisition system sends out a bottle firing code at the time of bottle firing. 
The code is logged as serial data by the Level A which timestamps its arrival.

A total of 96 stations (01 - 96) were occupied, of which 93 used the MkIIIb CTD. 
A Neil Brown MkV CTD was used at 3 of the stations. Data for stations 79-81 are 
from full depth casts using this instrument. Two trials had previously been 
conducted at stations 72 and 76.


Data Processing

CTD Data

The 1 second data passed to the Level C were converted to Pstar format and 
initially calibrated with coefficients from laboratory calibrations. The up cast 
data were extracted for merging with the bottle firing codes, on time, thus the 
CTD variables were reconciled with the bottle samples. Final calibrations were 
applied using the sample bottle data.

Pressure

A pre-cruise calibration, on 10th July 1991, using 14 points covering 0 to 
5500db was:

           p (dbar) = 0.997069 x praw + 4.091E-7 x praw2 - 7.31

at a temperature of 20 deg. C. The goodness of fit was 1.0db. The calibration was 
made under increasing pressure and therefore applies strictly to the down cast.

A post-cruise calibration, on 24th June 1992, shows a shift of order 4db over 
the year.

The following correction was applied for the effects of temperature on the 
pressure sensor:

                       pcor = p - 0.39 (Tlag - 9)

where Tlag is a lagged temperature, in deg. C, formed from the CTD temperature using 
a first order equation with a time constant of 400 seconds. The time constant 
and the temperature sensitivity of the pressure offset, 0.39 deg. C/db, were 
determined from laboratory measurements. The use of the default reference 
temperature (9 deg. C) was maintained from previous cruises.

On the up cast, a further correction is made for the hysteresis of the pressure 
sensor based on laboratory measurements. For a cast in which the maximum 
pressure reached is pmax dbar, the correction to the up cast CTD pressure (pin) 
is:

       pout = pin - ( dp5500(pin) - ( (pin/pmax) x dp5500(pmax) ) )

where dp5500(p) is the hysteresis after a cast to 5500db.

A significant difference was noticed between the digital reversing pressure 
meters and the CTD. The CTD pressure sensor readings at the bottom of the cast 
were converted to depth and added to the height above bottom measured by the 
10kHz pinger. These compared well with Carter corrected echo sounder 
measurements, confirming the calibration of the pressure sensor.

Temperature

The pre-cruise temperature calibration, obtained on 8th July 1991, was applied:

              T (deg. C) = 0.99865 x (0.0005 x Traw) - 0.0162

The goodness of fit (over a temperature range of 0.8 deg. C to 25 deg. C) was 0.4mK. 
Temperatures are given on the ITS90 scale. For the purpose of computing derived 
oceanographic variables, temperatures were converted to the 1968 scale (T68 = 
1.00024 T90).

The post-cruise calibration, on 10th January 1992, was:

              T (deg. C) = 0.99871 x (0.0005 x Traw) - 0.0156

In order to allow for the mismatch between the time constants of the temperature 
and conductivity sensors, the temperatures were further corrected according to 
the procedure described in the SCOR WG51 report (Crease et al. 1988). The time 
constant used was 0.25s giving a correction of:

                          T = T + (0.25 x delta-T)

where delta-T is the change in temperature over one second calculated by the Level A.

A comparison of the CTD temperatures with the SIS digital reversing thermometers 
gave the following differences:

Differences in temperature (corrected) between reversing thermometers and CTD
(Tctd-Tnnn)

Thermometer   T400  T399   T262   T260  T254  T220  T219  T213
Number          71    80     66     60    33    69    63    64
Mean diff mK  -5.5  -8.2  -10.9  -10.9  -0.7  -9.6  -8.2  -7.4
Std dev mK     2.5   1.7    8.1    4.5   1.5  10.0   3.2   9.9

The CTD temperatures were taken to be correct given their frequent calibration 
against triple point cells whereas the reversing thermometers use the original 
manufacturer's calibration. The reversing thermometers are also in shallower 
water than the CTD sensor.


Salinity

Raw conductivities were scaled to physical units using the relationship:

                  C (mmho/cm) = CR x (0.001 x Craw)

The choice of conductivity ratio (CR) was made to produce salinities that were 
approximately correct after comparison with bottle samples on the first few 
stations. CR was found to be 1.0000.

The conductivity is then corrected for the effects of pressure and temperature 
in the manner described in the SCOR WG51 report (Crease et al. 1988) with 
nominal values employed for the temperature expansion and pressure contraction 
coefficients of the material of the cell:

      Cnew = Cold x cfac x (1 + (Cpslope x p) + CTslope x (T - Torg)

where cfac (cell factor) = 0.99856, Cpslope = 1.5 x 10-8, CTslope = -6.5 x 10-6 
and Torg = 15 deg. C. Salinity for both the down and up casts was then 
calculated from the corrected temperature, pressure and conductivity.

For each station the differences between the bottle sample salinities (Sbot) and 
the up cast CTD salinity (Sctd) were calculated. A routine was used to derive 
coefficients for the relationship:

               Sbot - Sctd = a + (b x press) + (c x potemp)

The derived a, b, c coefficients were then used to correct all CTD salinities 
for both the up and down cast. Small residual errors remained between the 
corrected salinities and the bottle values that were a function of depth. These 
resulted in salinities that were high by of order 0.001 in depths greater than 
2500m. Similarly salinities were low by a similar amount between 1200 and 
1700m. The errors were more scattered above 500m due to the uncertainties 
created by the stronger salinity gradients.


Oxygen

The CTD oxygen sensor data were calibrated against bottle samples. No further 
details are available.


Transmittance

No calibration details are available for the transmissometer.


Sample Data

Chlorofluorocarbons

Samples for CFC compounds were always the first to be taken from the Niskin 
bottles. Analyses were conducted using 2 gas chromatographs, one belonging to 
JRC (for CFC-11 and CFC-12 measurements) and the other to PML (for additional 
CFC-113 analysis). Electron Capture Detection (ECD) was used to remove dissolved 
gases from the seawater sample, separate the compounds, and measure their 
concentration. The ECD was calibrated using a standard gas with a known fraction 
of each compound.

The CFC-113 data are considered suspect due to problems with the PML system. The 
data were not made available to BODC and are not presented here.


Oxygen

Samples for dissolved oxygen analysis were taken from the Niskin bottles after 
the CFCs. Samples were drawn into weight calibrated 120ml narrow necked glass 
bottles fitted with ground stoppers.  The pickling reagents were added on deck 
immediately after drawing and before capping.

Analysis was performed in a constant temperature laboratory at 21 deg. C using 
the Carpenter modifications (Carpenter, 1965) (with the exception of the 
thiosulphate titrant which was made up to a concentration of 0.2M) of the 
Winkler technique with an automated photometric endpoint detection system. This 
consisted of a solid state light source and photodiode detector peaked to the 
iodine signal, a Metrohm dossimat, and a PC and software supplied by Sensoren 
Instrumente Systeme. Blanking was performed using pure water and not seawater to 
avoid differences due to depth and position. Equations specified in the WOCE 
Operations and Methods manual (Culberson, 1991) were used to calculate the 
oxygen concentrations. A buoyancy factor was applied to the weight of the 
potassium iodate standard.

Precision was monitored by taking two or more duplicate samples from each cast. 
With the exception of stations 12 through 25, where a small air lock in the 
equipment caused reproducibility problems, the standard deviation of all 
duplicates was calculated to be 0.007ml/l which equates to 0.1% full scale. For 
stations 81 to 83 there was a periodic blockage in the aspirator which caused 
some of the titrations to fail.


Nutrients

Sampling for nutrients followed that for CFCs and dissolved oxygen. Samples were 
drawn directly into 30ml transparent high density polyethylene diluvials fitted 
with press on caps. Samples were analysed for nitrate and silicate using an 
Alpkem Corp. Rapid Flow Analyser, model RFA 300. The RFA system was sited in the 
constant temperature laboratory and the analysis temperature was 21 deg. C 
throughout the cruise.

Phosphate measurements were also planned, however, the capped head on the non-
toxic seawater supply in the constant temperature laboratory blew when the water 
supply was turned on. This resulted in the autoanalyser being flooded and 
damaged an electronic board in one of the photometer units and an air phase 
injection board on the pump assembly. This rendered one of the autoanalyser 
channels unserviceable and it was decided to sacrifice phosphate analysis.

The method for nitrate was that given in the Alpkem manual (Alpkem Corp., 1987) 
and involving the initial reduction of nitrate to nitrite by cadmium, in the 
form of an open tube cadmium reactor. Following diazotisation with 
sulphanilamide and coupling with naphthyl ethylene diamine dihydrochloride to 
form an azo dye, total nitrite was measured at an absorbance maximum of 
543nm. Interferences from metals were eliminated by the use of an imidazole 
buffer solution.

The method for silicate involved the addition of ammonium molybdate at acidic pH 
to produce b molybdosilicic acid, which was then reduced using ascorbic acid to 
form an intensely blue molybdenum complex with an absorbance maximum of 820nm. 
Interferences from phosphomolybdate and arsenomolybdate formed during the 
initial stages of the reaction were eliminated by decomposition with oxalic 
acid. The method employed was an adaptation of the Alpkem method given in the 
manual because it had been noted that slight temperature fluctuations 
markedly affect analytical precision and accuracy. To overcome this a 2ml mixing 
coil place in a heating bath at 37 deg. C was inserted before the flow cell of 
original cartridge. By inserting the heated coil the chemical reaction and hence 
colour development was allowed to proceed to completion before measurement.  

Listed below are the reagents used for analysis; stock 10mmol/l solutions of 
sodium hexafluorosilicate (0.96g) and potassium nitrate (0.505g) were used to 
standardise the analyses.

Nitrate + Nitrite:
Imidazole Buffer             6.81 g/l   pH 7.5
Combined reagent
    Sulphanilamide           5 g/l
    Hydrochloric Acid        100 ml/l
    Brij 35                  30% w/v    10 ml/l
N.E.D                        1 g/l

Silicate:
Combined reagent
    Ammonium Molybdate       10.8 g/l
    Sulphuric Acid           2.8 ml/l
    Sodium Dodecyl Sulphate  15% w/v    20 ml/l
Oxalic Acid                  10% w/v
Ascorbic Acid                18 g/l

The autoanalyser was calibrated using a six point standard curve. Mixed 
nitrate/silicate standards were prepared daily in artificial seawater from the 
stock 10mmol/l solutions. The standards ranged from 0 - 30mmol/l for nitrate and 
0 - 50mmol/l for silicate and were evenly spaced throughout the concentration 
range. Three individual stock standards of nitrate and silicate were prepared 
during the cruise.

Precision was assessed by analysing replicate samples taken from the same Niskin 
bottle. Two replicate samples were taken at random from each cast and analysed 
on the same analytical run.  Precisions based upon duplicate measurements were 
calculated by working out the standard deviation of duplicate pairs and 
expressing the figure as a percentage of the concentration of the highest 
standard. Precision for silicate was 0.15% and for nitrate 0.53%.

Accuracy was assessed on one occasion using commercially available standards 
supplied by the Sagami Chemical Company of Japan. The values obtained for these 
standards were found to be within 0.1% of the 10mmol/l silicate and 0.5% of the 
15mmol/l nitrate values quoted by the manufacturer. These standards were 
prepared in a similar matrix to the calibration standards.


Salinity

Salinity samples from the Niskin bottles were analysed using Guildline 8400 
bench salinometers (an old 8400 - first half of cruise, and a new 8400A - latter 
half) set to run at 24 deg. C in the temperature controlled laboratory (22 deg. 
C). Standardisation was done using IAPSO Standard Seawater batch P115.

The standard deviation of 200 pairs of duplicate samples was 0.001psu.


Oxygen and Hydrogen Isotope Ratios

Samples were collected into 250ml salinity bottles. The bottle necks were 
covered with a sheet of aluminium foil and the lid screwed on tight. Care was 
taken not to rip the aluminium foil because this acts as a water impermeable 
barrier to prevent contamination.

Delta oxygen-18 of water was measured using a mass spectrometer. No analyses 
were conducted for hydrogen isotope ratio.


Reversing Thermometers

8 SIS digital reversing thermometers were attached to the Niskin bottles.


BODC Data Processing

No further calibrations were applied to the data received by BODC. BODC were 
mainly concerned with the screening and banking of the data.


CTD Data

The CTD data were received as 2db averaged pressure sorted down cast data. The 
data were converted into the BODC internal format (PXF) to allow the use of in-
software tools, notably the graphics editor. Spikes in the data were manually 
flagged 'suspect' by modification of the associated quality control flag. In 
this way none of the original data values were edited or deleted during quality 
control. These data from cruise CD62 required little flagging and just a few 
points were set suspect.

Once screened, the CTD data were loaded into a database under the Oracle 
relational database management system. The start time stored in the database is 
the CTD deployment time, and the end time is the time the CTD was removed from 
the water. Actually these times are more precisely the start and end of data 
logging. Latitude and longitude are the mean positions between the start and 
end times calculated from the master navigation in the binary merged file. The 
longitude given to the end time for station 62 in the cruise report is incorrect 
and should be closer to -35_ 50.0.


Sample Data

BODC conducted extensive quality control to eliminate rosette misfiring and any 
incorrectly assigned flag codes. Before loading to the database the data were 
averaged if bottles fired within plus or minus 4.0db of each other. For stations 
19, 21 and 42 a bottle entry exists at a pressure greater than the maximum CTD 
pressure. This is most probably due to the CTD casts having been edited by the 
originating scientists. No bottle data were received for station 96.


References

Alpkem Corp. (1987). Operator's manual, RFA 300. Clackamas, Oregon. Unnumbered 
    loose leaf pages.

Carpenter, J.H. (1965). The Chesapeake Bay Institute technique for the Winkler d
    issolved oxygen method. Limnol. Oceanogr., 10, 141-143.

Crease, J. et al. (1988). The acquisition, calibration and analysis of CTD data. 
    UNESCO Technical Papers in Marine Science. No. 54, 96pp.

Culberson, C.H. (1991). Dissolved oxygen. WOCE operations manual WHPO 91-1. WOCE 
    Report No. 68/91.

Gould, W.J. et al. (1992). RRS Charles Darwin Cruise 62. Institute of 
    Oceanographic Sciences Deacon Laboratory, Cruise Report No. 230, 60pp.


C.2.2  SAMPLE SALINITY DETERMINATIONS (SB)

Both IOSDL Guildline Autosal salinometers, the old one (model 8400) and the new 
one (model 8400A) were carried on this cruise and operated at 24 degC in the 
Darwin constant temperature laboratory, which was maintained at 22 degC.  The 
former was operated for the first half of the cruise and the latter for the 
second half.  Both machines showed unusual stability of standardisation, neither 
deviating from their original values by more than 0.001 equivalent salinity.  
The old machine was initially preferred over the new one because the latter 
showed a large change in standardisation over the value at which it left IOSDL.  
This may be attributable to knocks suffered during transportation to Barry;  it 
did not affect the machine's subsequent performance.  Towards the end of the 
cruise, a fault developed on the FUNCTION switch on the new machine which 
resulted in display instability in the standby (SBY) position.  This will need 
to be rectified upon return.  Previous operational difficulties associated with 
the new machine, in particular its instability when run from a noisy power 
supply, have been rectified;  no difficulties on this count were experienced.

About 1800 CTD bottle samples and 200 samples drawn from the ship's non-toxic 
supply (for thermosalinograph calibration) were processed during the cruise.  
200 pairs of duplicate samples are included in the CTD sample figure, for which 
the standard deviation of the difference between salinities was 0.001.  Only one 
pair of duplicates was more than 0.002 different.  250 ampoules of standard 
seawater BATCH P115 used throughout were consumed.

The data processing route was changed during the cruise.  Initially, the Ocean 
Scientific International 'Salinity' program was used to generate files of 
salinity data on an IBM PS-2, which were transferred to an Apple Macintosh, 
edited as a Microsoft Word text file and then transferred to the shipboard SUN 
system.  Latterly, the data were entered directly into an Excel spreadsheet 
program on a Macintosh into which was pasted the approved conductivity ratio to 
salinity conversion formula.  After testing and comparison, the Excel system was 
used exclusively.  This was more efficient;  a text file saved from Excel could 
be transferred to the SUN with no requirement for intermediate stages of 
processing.

C.2.3  CHEMISTRY OVERVIEW (DS-W,RP,SMB,SH)

The primary aim of the chemistry work was to collect a narrowly spaced data set 
to the precision and accuracy required by the WOCE Hydrographic Programme (WHP) 
in order to chemically characterise and age the water masses of this area of the 
North Atlantic.  The data would subsequently be used for flux measurements and 
to calculate transport rates within the subpolar gyre.  Measurements of CFCs 
were planned at as many locations as possible (limited by the rate at which 
samples could be analysed) and of oxygen and nutrients at each station.  Samples 
were collected for oxygen and hydrogen isotope measurement by NERC Isotope 
Geosciences Laboratory (NIGL).


The CONVEX survey provided the first opportunity for comprehensive measurement 
of CFCs in both western and eastern basins of the subpolar gyre.  Because much 
of the water in the area is relatively new (i.e. was at the surface less than a 
decade ago) it was thought that the CFC-12/CFC-11 technique widely used within 
the WHP would not give sufficient information (the ratio of CFC-12/CFC-11 upon 
which the ageing technique is based has not changed much for more than a decade 
and so it can be difficult to age with any precision over this time period), and 
so measurements of the new CFC tracer, CFC-113 were planned.    In contrast the 
ratio of CFC-113 to either CFC-11 or CFC-12 has changed substantially in the 
past decade.  However, provided little or no mixing has taken place, it is 
possible to age core water mass with some reliability over the last 15 years 
using CFC-11 and CFC-12 data alone, by using a combination of the ratio with 
absolute concentrations.  Comparison with T, S and nutrient data facilitates 
this ageing technique.  Ageing from CFC-113 data is very new and requires some 
further development.  It also suffers from potentially severe CFC-113 
contamination problems; a problem which also hampers CFC-11 and CFC-12 
measurements but to a lesser extent provided the research vessel does not 
contain these substances within its air conditioning and refrigeration systems.  
Since two CFC systems were taken to sea, one belonging to the JCR (for CFC-11 
and 12 measurements) the other to ML (for additional CFC-113 analysis) the 
cruise provided an opportunity for intercalibration of the two systems.  It had 
been hoped that the JRC system would be upgraded for CFC-113 measurement prior 
to the cruise but due to late delivery of a number of components this was not 
possible.	Like the CFCs, oxygen measurements are indicative of water which has 
recently left the surface and are usually enriched in cold polar waters such as 
the overflow through the Denmark Strait.  The nutrients nitrate, phosphate and 
silicate compliment CFC and oxygen measurements in that water masses are often 
high in one and not in another.  A combination of all measurements results in 
distinct fingerprinting for many of the oceans water masses.

Oxygen and hydrogen isotope measurements can be used as oceanographic tracers 
and have been used successfully for tracing waters with a polar origin.  Both 
have heavy isotopes that are fractionated from their lighter more abundant 
counterparts during processes of evaporation and precipitation, where they are 
more concentrated in the liquid phase, and during freezing where the ice becomes 
enriched over the mother liquid.  On melting the high levels of O18 and H2 
remain and make them a useful tracer for assessing the amount of ice melt in a 
particular water mass.  Most of the tracer work to date has been at high 
latitudes, very little has been carried out in the subpolar area.  Furthermore 
the technique for oceanographic tracer use is new in the UK.  The cruise 
provided the opportunity to collect samples for a combined investigation with 
NIGL for assessing analytical techniques and usefulness of such data at lower 
latitudes.

The cruise also gave an opportunity to evaluate the workload, personnel 
requirement and sample handling and analysis techniques for the forthcoming WOCE 
WHP A-11.  

Sample collection

All samples were collected from depth using the IOSDL 10 litre Niskin bottles.  
These had been cleaned prior to the cruise using a high pressure water jet.  All 
O-ring seals and taps were removed, washed in decon solution and propan-2-ol 
then baked in a vacuum oven for 24 hours.  Cleaning and reassembling of the 
bottles was carried out just prior to the cruise to minimise contamination due 
to long storage.  54 bottles were taken to sea in case of contamination or 
damage, but only 24 where used.  All bottles, whether in use or not, remained on 
deck throughout the cruise.  None of the 24 working bottles showed a CFC 
contamination problem during the entire cruise.

C.2.4  CFCS

The bulk of the CFC measurements were made using the JRC CFC system.  Samples 
from 85 stations were analysed immediately after collection for CFC-11 and CFC-
12.  Due to the close station spacing and the required analysis time it was not  
possible to sample the full range of depth levels at all stations.  When turn-
over time was limited, sampling was concentrated on the deepest depth levels.  
(Data on number of samples taken on each station is given elsewhere in the CTD 
station list).

In general the JRC system worked extremely well.  Only two mishaps occurred, 
both due to the extreme weather.  First the cooling bath fell off the bench 
during a particularly bad roll and this fractured the flexible pipe.  It was 
quickly removed from the laboratory so as not to contaminate the CFC equipment 
and replaced with a spare.  No downtime resulted.  The second problem started 
with a noisy baseline from the ECD in the gas chromatograph.  The problem became 
substantially worse, rendering the system useless for measurement and eventually 
the signal died altogether. In the first instance, there appeared to be a 
correlation with the roll of the ship and so it was presumed that a wire had 
worked free due to the inclement weather.  It was not possible to verify this 
because of the need to open the radioactive detector and JRC/IOSDL do not have a 
licence for such work. Fortunately the JRC system is fitted with a spare ECD and 
so this was brought on line and the system functioned perfectly for the 
remainder of the cruise.  Approximately eight hours downtime resulted from this 
problem, with the loss of one station.  

The PML system seemed dogged with problems from the outset.  The computer 
running the calibration and data storage programs was not available due to 
failure on a previous cruise.  Backup software was supplied with the system but 
this was not comprehensive.  During set up one of the valve controllers failed 
and had to be rewired.  This was shortly followed by one of the valve actuators 
burning out, and since no spare was supplied this meant the valve V4 had to be 
turned manually.  Once set up the system developed a baseline oscillation that 
rendered data collection useless.  Although various theories where tested and 
advise sought from PML, the reason for the fault was not established.  After 
about 2 weeks at sea the system cured itself following particularly inclement 
weather and a time when the cryocool was not operational leaving the molecular 
sieves at ambient temperature for a few hours.  This suggested either a possible 
loose contact somewhere in the system or a pressure problem that was resolved 
when the flow rate in the sieves changed during heating and subsequent cooling.  
By this stage, however, it was decided to concentrate on the JRC system as this 
was giving good results.  Nevertheless work with the PML system did continue but 
the chromatography was poor, with considerable overlap of peaks.  Since time and 
personnel were at a premium it was not feasible to spend hours trying to obtain 
better separation when it was not obvious that this would have been achieved.  
We therefore concluded that there remain several question marks against the use 
of the PML system as a routine analytical tool.  Firstly, the chromatographic 
resolution needs improving.  Secondly, the system is very fragile and requires 
constant attention to maintain high quality results.  Thirdly, it is not robust 
enough to cope with the large sample turn around required by WHP cruises.

Both CFC systems were set up in the main laboratory of the ship.  From the CFC 
contamination aspect, the RRS Charles Darwin is a clean vessel with results from 
the JRC system suggesting no major contamination problems.  Some contamination 
was experienced with the PML system but this was traced to loose gas connections 
and it cleared once these had been tightened.  No contamination was found from 
the opening of the air conditioning unit door on the boat deck as had been 
suggested after Darwin 58 and 59.  

CFC concentration measurements suggest that Labrador seawater and Denmark Strait 
overflow water are of the order of 10 years old whereas the water at depth in 
the Eastern Basin is very much older corresponding to an age of nearer 50 years.

C.2.5  NUTRIENTS

Equipment and Technique

The nutrients nitrate and silicate were measured using an Alpkem Corp., Rapid 
Flow Analyser, model RFA 300 by a team of analysts from the James Rennell Centre 
for Ocean Circulation. The RFA system was sited in a constant temperature 
laboratory and the analysis temperature was 21C throughout the cruise.

The method for nitrate was that given in the Alpkem manual (Alpkem Corp., 1987) 
and involving the initial reduction of nitrate to nitrite by cadmium, in the 
form of an open tube cadmium reactor.  Following diazotisation with 
sulphanilamide and coupling with naphthyl ethylene diamine dihydrochloride to 
form an azo dye, total nitrite was measured at an absorbance maximum of 543 nm.  
Interferences from metals were eliminated by the use of an imidazole buffer 
solution.

The method for silicate involved the addition of ammonium molybdate at acidic pH 
to produce b molybdosilicic acid, which was then reduced using ascorbic acid to 
form an intensely blue molybdenum complex with an absorbance maximum of 820 nm.  
Interferences from phosphomolybdate and arsenomolybdate formed during the 
initial stages of the reaction were eliminated by decomposition with oxalic 
acid.  The method employed was an adaptation of the Alpkem method given in the 
manual because it had been noted that slight temperature fluctuations markedly 
affect analytical precision and accuracy.  To overcome this a 2 ml mixing coil 
place in a heating bath at 37C was inserted before the flow cell of the original 
cartridge.  By inserting the heated coil the chemical reaction and hence colour 
development was allowed to proceed to completion before measurement.  
Samples were not analysed for phosphate due to loss of the photometer at the 
start of the cruise.

Listed below are the reagents used for analysis; Stock 10 mmol/l solutions of 
Sodium Hexafluorosilicate (0.96 g) and Potassium Nitrate (0.505 g) were used to 
standardise the analyses.
Silicate.

Combined reagent
    Ammonium Molybdate        10.8 g /l
    Sulphuric Acid            2.8 ml /l
    Sodium Dodecyl Sulphate   15% w/v     20 ml /l
Oxalic Acid                   10% w/v
Ascorbic Acid                 18 g /l
Nitrate + Nitrite
Imidazole Buffer              6.81 g/l    pH 7.5
Combined reagent
    Sulphanilamide            5 g /l
    Hydrochloric Acid         100 ml /l
    Brij 35 30% w/v           10 ml /l
N.E.D                         1 g /l

Sampling Procedure

Sampling for nutrients followed that for CFCs and dissolved oxygen and was 
typically 45 minutes after the rosette was on deck.  Samples were drawn directly 
into 30 ml transparent high density polyethylene diluvials fitted with press on 
caps: each vial was flushed with three times its volume before filling.  Samples 
were immediately transferred to the constant temperature laboratory and all 
analyses completed within an hour and a half of collection.  An exception to 
this was on Stations 18 and 19 when due to a malfunction of the autoanalyser a 
delay was anticipated.  These samples were stored refrigerated at 4C for 
approximately one and a half hours before analysis.  For all samples the 2.5 ml 
autoanalyser polystyrene cups were rinsed with three times their volume before 
filling.

Calibration

The autoanalyser was calibrated using a six point standard curve.  Mixed 
nitrate/silicate standards were prepared daily in artificial sea-water from the 
stock 10 mmol/l solutions.  The standards ranged from 0 - 30 mmol/l for nitrate 
and 0 - 50 mmol/l for silicate and were evenly spaced throughout the 
concentration range.  Three individual stock standards of nitrate and silicate 
were prepared during the cruise.

Precision was assessed by analysing replicate samples taken from the same Niskin 
bottle.  Two replicate samples were taken at random from each cast and analysed 
on the same analytical run.  Precisions based upon duplicate measurements were 
calculated by working out the standard deviation of duplicate pairs and 
expressing the figure as a percentage of the concentration of the highest 
standard.  Precision for silicate was 0.15% and for nitrate 0.53%.

Accuracy was assessed on one occasion using commercially available standards 
supplied by the Sagami Chemical Company of Japan. The values obtained for these 
standards were found to be within 0.1% of the 10 mmol/l silicate and 0.5% of the 
15 mmol/l nitrate values quoted by the manufacturer.  These standards were 
prepared in a similar matrix to the calibration standards.

Units used

Nutrient data were converted from mmol/l to mmol/kg units by multiplication with 
the density calculated from sample salinity and a temperature of 21C; the 
temperature of the constant temperature laboratory where the analyses were 
performed.

Data quality checks

Nutrient data quality was checked for each station by comparison of plots of 
nutrient concentration versus potential temperature.  Plots for consecutive 
stations were overlaid to observe obvious errors such as large spikes in the 
data indicating contamination.  For example low silicate and nitrate values were 
observed for sample 23 station 41 and accidental dilution of the sample was 
suspected.  In such cases the data are not reported and a flag of 5 recorded.

Drift in the data was detected in a similar way.  For station 84 nitrate data 
shifted low but was not further queried since it correlated with subsequent 
stations.  Due to the wide range of silicate values, silicate data were checked 
on an expanded scale where necessary.

Where data appeared questionable analytical logsheets were scrutinised to ensure 
correct standardisation procedures and reagent composition. If no explanation 
was obvious further comparisons were made with nutrient concentration verses 
pressure plots, oxygen profiles and topography. When no obvious reason could be 
found to reject the data but it still looked unusual when plotted against 
potential temperature or pressure then a flag 3 was recorded.  For example 
nitrate data from stations 75 to 78 are high but substantial checking suggested 
no reason to consider the measurements as bad.  However questions still need to 
be addressed so the nitrate data for these stations have been flagged as 3.  
Data for station 69 are flagged 4 because a blockage in the sample probe was 
suspected.

A flag 9 is recorded for absent data.

Comparison with historical data sets

Data accuracy was checked by comparison with historical data sets from the same 
area: TTO (1981) and TYRO 90/3 (1991).  The TYRO 90/3 study concentrated on the 
WOCE AR7E line which also made up part of the CONVEX study area.  Data 
comparison was done in two ways:

Firstly, nutrient data from all of the CONVEX stations were plotted against 
potential temperature on a single graph.  Similar plots were prepared for the 
TTO and TYRO 90/3 data and the plots were overlaid.  This exercise was repeated 
using nutrient versus pressure plots.  The CONVEX data fell within the scatter 
of the TTO and TYRO 90/3 data.

Secondly, four TTO stations from the same location as four CONVEX stations, were 
selected and comparisons of data below 1000 m were made.  Comparisons in the 
surface waters were not valid due to seasonal differences and different 
biological activity effecting the surface nutrient concentrations.  The 
comparisons suggested that there was no detectable difference for silicate 
concentrations but nitrates differed from TTO by up to 0.8 mmol/kg.

Overall the silicate and nitrate chemistries worked well and the data from the 
repeat stations, cd62082 and cd62010, are consistent.

References

Alpkem Corp.,1987.  Operator's manual, RFA 300, Clackamas, Oregon. Unnumbered 
    loose leaf pages.


C.2.6  OXYGEN

Equipment and Technique

Samples from the Niskin bottles were analysed for oxygen using an automated 
Winkler titration unit supplied by Sensoren Instrumente Systeme.  It comprises a 
fully automated PC driven photometric endpoint detection method using a solid 
state light source and a photodiode detector peaked to the iodine signal.  The 
titrant is dispensed by a Metrohm Dosimat model 665 with a 1 ml exchange unit.  
All reagents were made up according to the procedures and concentrations 
specified by Carpenter (1965), with the exception of the thiosulphate titrant 
which was made up to a concentration of 0.2M.  This allowed the titration to be 
performed within one single delivery of the exchange unit.  The standard titre 
was typically 0.500 ml, and the blank titre 1 ml.  Details of equipment 
operation may be found in the SIS manual (1989).  All analyses were performed in 
a constant temperature laboratory at 21C 

Blanking and standardisation procedures were carried out at least once a day 
using deionised water.  All calculations were performed using the equations 
specified in the WOCE Operations and Methods manual.  A buoyancy factor was 
applied to the weight of the potassium iodate standard. 


Sampling Procedure

Samples for oxygen were taken, immediately after CFCs, directly into weight 
calibrated 120 ml narrow necked glass bottles fitted with ground stoppers.  
Sample flasks were flushed with three times their volume prior to filling and 
pickling reagents added immediately.  The samples were analysed about two hours 
after collection to allow them to equilibrate to the constant laboratory 
temperature.  All analysis was completed within three hours of the cast arriving 
on deck.

Calibration

Precision was monitored by taking two or more duplicate samples from each cast. 
Short term precision was monitored on 92 stations throughout the cruise. With 
the exception of stations 12 through 25, where a small air lock in the equipment 
caused reproducibility problems, the standard deviation of all duplicates was 
calculated to be 0.007 ml/l which equates to 0.1% full scale.  For stations 81 
to 83 there was a periodic blockage in the aspirator which caused some of the 
titrations to fail.

Units

The data were converted from ml/l to mmol/l using a factor of 44.66 to convert 
ml/l to mmol/l and to mmol/kg using seawater density calculated from the 
salinity and potential temperature of the sample.  Changes in bottle volume due 
to temperature effects were considered to be minimal.  Calculations on a 
previous cruise revealed that the change of volume resulted in less than 0.02% 
change for the entire temperature range encountered on CONVEX; for example a 
calculated oxygen value of 5.000 ml/l at 20C would alter by less than 0.001 ml/l 
if measured at 0C.

Data quality checks

Oxygen data were checked, in a similar way to the nutrients, by comparing plots 
of oxygen concentration versus potential temperature.  Plots for consecutive 
stations were overlaid to observe obvious errors and in such cases the data are 
not reported and a flag of 5 recorded.  When the titration failed a flag of 4 
(bad measurement) was recorded with the absent data.  Flag 9 was recorded when 
no sample was collected due to failure of the Niskin bottle.

Comparison with historical data sets

Data accuracy was checked in the same way as that for nutrients by comparison 
with TTO (1981) and TYRO 90/3 (1991) data.  The comparison was done in two ways; 

Firstly, oxygen data from all of the CONVEX stations were plotted against 
potential temperature on a single graph.  Similar plots were prepared for the 
TTO and TYRO 90/3 data and the plots overlaid.  This exercise was repeated using 
oxygen versus pressure plots.  The CONVEX data fell within the scatter of the 
TTO and TYRO 90/3 data.
Secondly, four TTO stations from the same location as four CONVEX stations, were 
selected and comparisons of data below 1000 m were made.  Comparisons in the 
surface waters were not valid due to seasonal differences and different 
biological activity effecting the concentrations. The comparisons suggested that 
there were no detectable differences.

Data from the repeat station, cd62082 and cd62010, were compared and found to be 
consistent.

References

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

SIS User Manual Version 2.01. Kiel, Germany.

WHP operations and Methods Manual, 1991.


C.2.7  OXYGEN AND HYDROGEN ISOTOPE RATIOS

A total of 165 samples were collected from 7 stations for isotope analysis.  
These included 6 duplicate samples for assessment by NIGL and a complete set of 
duplicated from station CD0072 for analysis at UEA.  The latter were requested 
by KH to evaluate the UEA analytical capability.  Samples were collected into 
250 ml salinity bottles following an initial rinse and two fillings and emptying 
method.  On the third fill the neck was covered with a sheet of aluminium foil 
and the lid screwed on tight.  Care was taken not to rip the aluminium foil 
because this acts as a water impermeable barrier to prevent contamination.

C.2.8  METEOROLOGY - THE MULTIMET SYSTEM (RJT)

Note the archived data is at the WOCE Meteorology DAC, Florida State University 
and can be retrieved via http://www.coaps.fsu.edu/WOCE or from the WOCE CD-ROMs 
v1.0 dataset on the Surface Meteorology Disc (Atlantic).  The Cruise is 
identified as AR7E/03.

Sensors

The sensors were situated on the foremast,  mainmast,  and on the port and 
starboard sides of the wheelhouse top.  The forward mast carried a propeller 
vane anemometer (R.M.Young serial number (S/N) 6992)  and a Gill sonic 
anemometer situated on the forward platform.  Two aspirated psychrometers 
(Vector instruments (VI) S/N 1066,  and S/N 1072) were situated port and 
starboard.  The short wave radiation sensors were situated to the far port and 
starboard side of the forward mast platform (Kipp and Zonnen S/N 11902837 and 
11871958).  The long wave radiation sensor  (Eppley S/N 27225F3) was situated at 
the top of the upper foremast.  On the wheelhousetop there was an aspirated 
psychrometer (VI S/N 1058) situated to the port side,  and also to the starboard 
side (VI S/N 1060),  in each case just aft of the ladder.  The main mast carried 
an anemometer (VI S/N 1895) and a wind direction sensor (VI S/N 2224) situated 
at the mast top.

Multimet

The IOSDL Multimet system recorded data once per minute throughout the cruise.  
All the data from all the sensors were recorded.  Wind speed data from the 
mainmast anemometer  stopped transmitting data on day 233, and only transmitted 
intermittently from then on.

Problems with the recorded data included a possible calibration error with the 
starboard dry bulb temperature on the wheelhouse top,  and the foremast 
anemometer data from day 237 onwards.

For each 24 hour segment, for head to wind (foremast anemometer direction 150 to 
210degT) values at night (20-24 and 0-8 hours) the mean starboard dry bulb 
temperature on the wheelhouse top read a maximum of 0.32C high compared to the 
mean starboard dry bulb temperature on the foremast;  and a maximum of 0.28C 
high compared to the mean port dry bulb temperature on the wheelhouse,  
(Possible calibration error).  For days 213 to 242.0 the mean starboard dry bulb 
temperature on the wheelhouse top read 0.21C high compared to the mean starboard 
dry bulb temperature on the forward mast;  and 0.13C high compared to the mean 
port dry bulb temperature on the wheelhouse.

The starboard wheelhouse top psychrometer was effected by heat from the funnel 
when the relative wind direction was between 72 degrees and 144 degrees (where 0 
degrees is the aft of the ship).  The port wheelhouse top psychrometer was 
effected by heat from the funnel when the relative wind direction was between 
288 and 360.

Each time the wheelhouse psychrometers were filled the water in the starboard 
psychrometer was considerably lower than the water in the port psychrometer.  
There was no leak in the bottle,  possible causes could be either funnel smoke 
or heat from bridge door being open.

The foremast anemometer was seen to be at an angle (instead of horizontal) on 
day 237.  This did not seem to have affected its ability to transmit data.  But 
on day 237,  the relative foremast wind direction seemed to be reading low by 
about 40-60 (should have been about 350 but it read about 290),  presumably due 
to the angle of the anemometer.  Calculation of the true wind speeds and 
directions from then on were incorrect,  in that when the ship changed 
directions a jump occurred in the true winds.

Shadowing by the foremast occasionally caused the port short wave solar 
radiation sensor to read lower than the starboard short wave solar radiation 
sensor.

Special interest should be taken of day 229 and 230,  where there was a very 
sudden increase in wind speed.

Processing the Multimet Data

The time series of each sensor shown on the BBC computer were checked daily,  as 
were the sensors.  The data was processed on the Sun in 24 hour segments.  The 
processing of the Multimet data used four execs (mmexec0 to mmexec3).
*  mmexec0 extracts the daily meteorological data from the RVS files and 
   applies calibrations to it.
*  mmexec1 plots the daily data.
*  mmexec2 calculates the specific humidities for each psychrometer, and then 
   calculates the temperature and specific humidity differences for each sensor,  
   it then calculates the head to wind values (using the foremast anemometer 
   data) at night to find any possible calibration errors in the psychrometers. 
*  mmexec3 calculates the true wind speed and direction using the foremast 
   anemometer data.

Problems with the Multimet Processing

Days 213 to 221 (inclusive) the meteorological calibration file was incorrect 
for the gyro (it read column one as 1,  it should have read 1.4117).  Thus for 
this period the archived data with file names:

mm$cruise$met,
mm$cruise$met.hour,
mm$cruise$met.q3,
mm$cruise$met.qdf

the gyro is wrong.  But it can be corrected by using a linear calibration of 
(gyro*1.4117) using pcalib.  The gyro plots for this period are therefore also 
incorrect,  the correct gyro plots can be seen in the true wind plots.  The 
appended mm$cruise file thus also had an incorrect gyro,  this was corrected 
using pcalib and this version archived.  The meteorological calibration file was 
changed for days 222 onwards.

Up to and including day 225,  there was a mistake in mmexec2,  and the archived 
files mm$cruise$met.q3 were really mm$cruise$met.q archived.  From day 226 
onwards the archived files are correct.

Sonic Anemometer

The data from the sonic anemometer on the foremast was logged constantly on a 
PC.  This data was backed up daily on to the Sun system,  when the ship was 
steaming so as not to loose good data.  The sonic stopped logging on day 238,  
but was restarted on day 239,  this data was therefore not logged. Data before 
and after seemed to be good.


SBWR

The data from the shipborne wave recorder was logged constantly on a PC.  This 
data was backed up daily on to the Sun system,  when the ship was steaming so as 
not to loose good data.  There were no problems.


C.2.9  XBT DATA (WJG,LP)

This data has been submitted to the NODC archive, Washington and can be found in 
the July-Sep 1991 BEST COPY GTSPP Archive.  This can be reached at the time of 
writing via http://www.nodc.noaa.gov/GTSPP/gtspp-home.html or via the WOCE CD-
ROM v1.0 dataset on the Upper Ocean Thermal Disc.

XBT data were obtained between the CTD stations and on some passage legs in 
order to provided improved spatial resolution of mesoscale features.  The 
shipboard system used was a Bathysystems SEAS system transmitting data via the 
GOES satellite.  The equipment was suppliedand installed by the MoD Hydrographic 
Department.

We were supplied with a new version of the Seas system software and were sent to 
sea with only very limited documentation. A fault developed in system on day 224 
that indicated that the GOES buffer was not being cleared and that the real time 
data were probably not being transmitted to the satellite.  The problem was not 
traced until day 234 when contact with MoD through RVS revealed that the power 
supply that was charging the batteries was not working.  The problem was 
rectified and thereafter no problems were encountered.   In future the vessel 
should carry a complete handbook of hardware and software documentation.

The new software has a number of annoying faults

a) the logged time for an XBT is the time at which the programme is set up.  
   Previous versions have used the much better system of attributing the drop    
   time to the time when the probe is launched, clearly much better and 
   sometimes up to 10 minutes later than the setup time with consequent errors 
   in position.

b) A most annoying feature is the format for entering positions.  A <return> 
   is required for the numerical value but not for the N/S E/W entry.  Similarly 
   a 3 digit degree entry is required for longitude and an entry of 2 or less 
   (e.g. 16 rather than 016) leads to errors.

c) When the water depth is less than the probe's maximum depth the programme 
   asks whether a shallower cutoff is required.  It would be much better to ask 
   the operator to specify a cutoff level after the profile is complete so that 
   a partial trace can be transmitted.


d) There is no way to dump the plot of the profile to the printer, a serious 
   omission, and no way to dump data except at the raw data rate.  You would 
   like to be able to list data every say 10 or 20 metres and to specify the 
   particular isotherms for which depths are required rather than just whole 
   degrees.

The probes used were Sippican T-7 and T-5 and Plessey T-7.  Few probes failed 
and these can be attributed largely to problems with the XBT software,  a 
leakage in the launcher cable and to the XBT wire fouling the ship in the often 
poor weather.  This latter problem was helped by the use of a length of plastic 
tubing that fitted over the probe canister.  In rough weather probes could not 
safely be launched from the stern and so they were often deployed near the hatch 
in a relatively sheltered position.  This seemed to work well.  Some T-5 probes 
were launched at too high a speed early in the cruise and these are noted in the 
XBT table.

Data were transferred from the PC floppies to the PSTAR system using a programme 
written on board by Doriel Jones.  No transfer software nor indeed any 
information on PC file formats was supplied.  The data were subsequently edited 
in PSTAR to remove spikes and near surface transients.

The temperatures matched well to the thermosalinograph values.


C.2.10  THERMOSALINOGRAPH (WJG, LP)

The thermosalinograph was based on Sea Bird temperature and conductivity sensors 
mounted in an enclosure in the Hydro Lab.  The data together with the sea water 
intake temperature were logged on the RVS computer system and later transferred 
to PSTAR.  Calibration samples were taken at two hourly intervals from the 
outflow from the thermosalinograph housing.  These were analysed and the data 
for the whole cruise plotted.  A rather constant offset was seen until day 230 
after which time the offset decreased with time.  There appeared to be marine 
growth in the output pipe from the thermosalinograph and presumably also 
therefore on the conductivity sensor.  This was not cleaned off since the sensor 
wads known to be fragile.  the logged TSG salinity values have yet to be 
corrected.

The offsets obtained in the very cold, low salinity water around green;land were 
very scattered but no cause for this scatter could be found.  It was not 
attributable to a time lag between the time the value was sensed and the time 
the sample was taken.

The temperature sensor values were compared with near surface CTD values and 
were found to be constant with time.


C.2.11  BATHYMETRY - SIMRAD ECHOSOUNDER (WJG)

Bathymetric data from the entire cruise at 1/2 minute intervals has been 
archived by the National Geophysical Data Centre and can be retrieved via 
http://www.ngdc.noaa.gov/mgg/mggd.html (Access Number 3CD6291) or via the WOCE 
CD-ROM v1.0 dataset on the DIU disc (Access AR12 CD6291).

The Simrad echosounder was used throughout the cruise and performed well.  It is 
pleasant not to have to fold Mufax rolls.  However we spent much time using it 
in pinger mode and a serious defect is the multiplicity of menu commands that 
need to be made in order to change from pinger to echosounder and back again.  
What is required is a limited number of preset menus that can be swapped between 
at a single keystroke.  We resolved the situation by running the Simrad and PES 
MkIII in parallel while on station.


C.2.12  BIOLOGICAL SAMPLING (KJH)

Zooplankton samples were taken at 35 stations (40 casts) using a vertically 
hauled 150 mm mesh net with opening area 0.25 m2, kindly loaned by Dr.R.Harris 
of PML. Details of the casts are given in the Table. During the first 2 weeks of 
the cruise, the kevlar winch on the CTD A frame was used, but this spool only 
held 200 m so net hauls were restricted to a maximum of 200 m, depth being 
determined by marks on the rope. However the Effer crane and small winch on the 
aft deck were attached to a 450 m spool of kevlar and this system was used 
thereafter so that deeper hauls could be achieved during the day. The 
disadvantage was that, given the close proximity of the winch to the ship's 
propeller, we were not allowed to do nets during strong winds. Wire out was 
determined using a meter sheave with electronic readout. Large wire angles 
during rough weather meant that actual cast depths were sometimes much less than 
the wire out. Although the 200 m spool should have been used for nighttime 
casts, the kevlar caught on the A frame during a CTD cast and snapped. 

Nets were hosed down when brought to the surface, and samples were sieved to 
remove excess water before being placed in 250 ml sample jars Samples were 
preserved in 10% formalin for later identification and analysis at UEA. In the 
east (legs A and B)  catches were primarily green and fluoresced when hosed - it 
is assumed that these were mainly phytoplankton. They were very difficult to 
sieve since the net clogged. The largest catches were obtained in the coldest 
waters near Greenland - these consisted of a dense dark pink mass of small 
animals (1-2 mm). The third type of catch was pale pink shrimplike animals 
(euphausids?) about 1cm in length.

C.2.13  ACOUSTIC DOPPLER CURRENT PROFILING (KJH, MH)

The ADCP recorded throughout the cruise using 64 8m bins (50 x 4m on the shelf) 
and a 2 minute sampling interval. It worked well on the whole and a good data 
set was collected. Given a transducer depth of 5m and blank beyond transmit 
length of 4m, the first bin was centred at a depth of 13m. Three zigzag 
calibration runs were conducted in the early part of the cruise, two in bottom 
tracking mode and one in deep water. One calibration run was conducted on the 
last day whilst steaming across the shelf.

We encountered problems with the RVS logging on the level ABC, which appeared to 
happen when we changed from bottom tracking to deep water mode, and again when 
we returned to bottom tracking mode. The RVS programme crashed both times and 
seemed unable to read the headers. When we changed into bottom tracking, bottom 
velocities and water depth were not being logged in the RVS data stream. On both 
occasions, the ADCP set up was investigated but no problems were found there - 
and when the RVS programme was restarted, it was again able to read the data 
stream. This problem was not resolved.

The hard disk on the PC crashed only 3 days after leaving port (Disk controller 
error), thereby losing some data which had not been logged by RVS while 
programmes to read the data stream had been sorted out. It is hoped that the 
pingdata files on the hard disk might be recoverable. Luckily a spare PC was 
available and the boards were changed around. ADCP DAS software version 2.48 was 
installed from floppy disk, and the logging configurations set up again. Initial 
problems with the data stream to the level ABC were due to the fact that the new 
PC has the level ABC attached to COMS1 not COMS2. Apart from 4 hours of lost 
data, the only repercussion was that logging of 'Ancillary Data' - ship's 
heading and water temperature - was not set to YES after the crash, and this was 
not noticed for a week. We would recommend in future that the spare PC is 
identical to that used for ADCP logging, since a great deal of time was spent 
swapping circuit boards around - the two PCs having different sets of boards. 

The ADCP is not sending its wakeup message correctly on startup. This is easily 
remedied by switching the ADCP off and then on again, then reinitialising 
(command RI). This is an irritating fault and appears to be getting worse.

During the ADCP calibration run at the end of the cruise, after changing to 
bottom tracking mode, the ADCP did not log ship's heading correctly - it got 
stuck at 360o. This is obvious on the ADCP screen. The ADCP was restarted using 
exactly the same configuration file and ship's heading was now fine. The cause 
of this problem was not found.

During the first few weeks of the cruise, good depth penetration was achieved - 
about 450 m on station and 350 m whilst steaming. However near the Greenland 
coast penetration was particularly poor and at times there were no bins with 
%good more than 25%, even on station. This may have been caused by bubbles since 
the sea was very rough. The ADCP Top Hat was bled but no air was found trapped. 
After leaving Greenland, the penetration never recovered and we obtained no data 
while steaming on leg G. This was probably because the weather was calm and we 
were steaming at 12 knots. It appears that 10 knots is the maximum for good ADCP 
data.

Logging of the acoustic backscattered signal strength from the 4 individual 
beams was undertaken throughout, for comparison with zooplankton abundance. A 
separate ADCP logging file for these data - adcpraw - was maintained by RVS on 
the level ABC. This was read and processed in the same way as the current data, 
and was then archived using 1Gb Exabyte tape. Some tests for acoustic 
backscatter noise level were conducted in Barry and in the middle of the cruise 
when the files from the PC hard disk were backed up. Temperature of the ADCP 
electronics was noted regularly and was found to vary by at least 5 C despite 
the air conditioning thermostat controls.

Data Processing route

The data from the PC running the ADCP software were read into the RVS level ABC 
logging system, retrieved from level B, and processed using a Sun Spark 
Workstation running SUNOS. The data were divided into two sets of files - 
stations and passage legs. The executives used in the processing route were : 
adpexec0, adpexec1, adpexec2, adpexec3,  adpexec4, preport and adppage1.

adpexec0: converts the data into pstar format and creates two files, one 
          bottom, one gridded.
adpexec1: corrects for ADCP clock drift.
adpexec2: applies the amplitude and pointing angle corrections. It also 
          averages the data to 15 minute intervals.
adpexec3: was used to plot the averaged ADCP data against the navigation 
          before merging the two.
adpexec4: merges the averaged ADCP data with the navigation data from which 
          the absolute currents are determined.
Preport:  uses the plotting programs ucontr and parrog to create postscript 
          files of percent good, relative amplitude of backscatter and absolute 
          current velocities.
adppage1: rearranges and does some editing on the postscript files created by 
          preport so that three plots are produced on the same page.

When the ship's speed was changing, spikes occurred in the absolute velocities, 
probably a function of the 15 minute averaging performed within adpexec2. 
Although the start time entered was the first datapoint with constant ship's 
velocity, the averaging program adpav2 appears to take this point as the 
midpoint of its 15 minute average, thus averaging data from up to 7.5 minutes 
before. The start times for files were delayed by 10 minutes to counter this 
Even so some very large vectors were still incurred, these were subsequently 
edited.

The program allav was run on each station and for the passages between these 
were appended for the whole cruise.
