WHP Cruise Summary Information

WOCE Line:                    SR01
EXPOCODE:                     74DI198_1
Chief Scientist/affiliation:  David Turner/PML
Dates:                        1992.11.11 - 1992.12.17
Ship:                         DISCOVERY
Ports of call:                Stanley, Falklands to Punta Arenas, Chile
Number of stations            39
                                      5129.45'S
Geographic boundaries:        881.82'W        5740.03'W
                                      690.35'S
Floats and drifters deployed:  none
Moorings deployed/recovered:   none

Contributing Authors:(in order of appearance)
                    G. Griffiths  A. Weeks        W. Broadgate
                    B. Miller     D. Turner       C. Attwood
                    P. Machin     I. Bellan       A. Murray
                    A. Morrison   A. Weeks        D. Bone
                    C. Day        G. Moore        G. Savidge
                    T. Poole      J. Robertson    H. Waldron
                    P. Taylor     S. Debney       M. Behrenfeld
                    R. Pollard    R. Bellerby     M. Hilles
                    J. Allen      S. Knox         R. Pearce
                    J. Read       B. Head         E Morozov (DQE)
                    P. Machin     P. Nightingale  J.C. Jennings, Jr (DQE)

CONTENTS
1. Overview
   1.1   Introduction
   1.2   Cruise objectives
   1.3   Cruise itinerary
   1.4   Scientific achievements
   1.5   Logistics and equipment

2. Cruise narrative
   2.1   Mobilisation
   2.2   Scientific log

3. Scientific Activities
   3.1   Navigation
   3.2   Acoustic Doppler Current Profiler
   3.3   Meterology
   3.4   Surface temperature and salinity
   3.5   Starboard winch
   3.6   CTD operations
   3.7   CTD data processing and calibration
   3.8   SeaSoar operations
   3.9   SeaSoar data processing and calibration
   3.10   Undulating Oceanographic Recorder
   3.11   Discrete salinity measurements
   3.12   Optics
   3.13   Expendable bathythermographs
   3.14   MacSat satellite receiver
   3.15   Non-toxic supply
   3.16   pCO2
   3.17   Total CO2
   3.18   pH
   3.19   Alkalinity
   3.20   On-line oxygen
   3.21   Oxygen titrations
   3.22   Nutrients
   3.23   On-line fluorescence, chlorophyll and phytoplankton sampling
   3.24   Chlorophyll calibrations
   3.25   Biogenic sulphur compounds
   3.26   Low molecular weight halocarbons
   3.27   Hydrocarbons
   3.28   Particulate sampling
   3.29   13-C sampling
   3.30   Krill acoustics
   3.31   Size-fractionated primary production
   3.32   Size-fractionated new and regenerated production
   3.33   UV-B radiation and primary production
   3.34   Instrumentation notes
   3.35   Computing

Appendices
   A.    Scientific party
   B.    Stations worked
   C.    SeaSoar tows
   D.    UOR tows
   E.    Lightfish tows
   F.    Acoustic fish tows
   G.    On-line water sampling periods
   H.    XBT casts
   I.    Day number/date interconversions

List of Tables
   3.1   Calibration coefficients for the met. sensors
   3.2   Salinity intercalibration on Discovery samples
   3.3   Salinity intercalibration on James Clark Ross samples
   3.4   Chlorophyll-reflectance relationships from SeaSoar data
   3.5   Diffuse attenuation values from CTD casts
   3.6   Archived MacSat images

List of Figures
   1.1   Cruise track
   1.2   First survey track
   1.3   Second survey track
   1.4   Chlorophyll and salinity sections (Leg A, first survey)
   2.1   Surface chlorophyll, both surveys
   3.1   GPS drift on Drake Passage section
   3.2   Diffuse attenuation, first survey
   3.3   Chlorophyll-reflectance ratio relationships
   3.4   Reflectance ratio, first survey
   3.5   Non-toxic water system
   3.6   pCO2, first survey
   3.7   pH offsets
   3.8   Krill distributions
   3.9   New and regenerated productivity stations

DQE Reports
    CTD DQE
    Nutrients DQE

1.   OVERVIEW
1.1  Introduction

This cruise formed part of the Sterna 92 expedition, carried out in 
conjunction with RRS James Clark Ross.  It is noteworthy that this, the final 
expedition of the BOFS programme, saw new and productive collaboration with 
other groups of oceanographers.  The collaboration with the British Antarctic 
Survey was most marked on the James Clark Ross, although we were pleased to 
welcome two BAS scientists to BOFS aboard the Discovery.  The major new 
collaboration aboard the Discovery was with physical oceanographers from the 
WOCE community, whose enthusiastic participation helped to make this the most 
thoroughly interdisciplinary cruise of the BOFS programme.  Equally welcome 
and enthusiastic was the contribution of visiting American and South African 
scientists to the primary production studies.  The full scientific party in 
listed in Appendix A.

1.2  Cruise objectives

The overall objective of the Sterna 92 expedition, as set out in the cruise 
proposal, was "To evaluate the magnitude and variability of biogeochemical 
fluxes (particularly carbon and nitrogen), during early summer in the South 
East Pacific Sector of the Southern Ocean, with emphasis on rates and 
processes in the marginal ice zone".

Four specific objectives were identified:

1.  To determine ocean-atmosphere exchanges of radiatively active gases, and 
    the factors influencing such fluxes, over a wide latitudinal range.
2.  To investigate the interactions between the biological, chemical and 
    physical processes that control carbon fluxes in the euphotic zone.
3.  To assess the impact of sea-ice on biogeochemical fluxes, in order to 
    estimate the importance of climatic feedback effects.
4.  To determine the export of biogenic material from the upper ocean and 
    its subsequent fate.

The Discovery programme, with a strong emphasis on underway survey 
measurements, concentrated on the first two of these components, while the 
James Clark Ross concentrated on components 3 and 4 with a programme 
emphasising station work.

1.3  Cruise itinerary

 5/11/92  Discovery arrives in Stanley
10/11/92  Discovery departs Stanley for Berkeley Sound
11/11/92  Bunkering complete: Discovery departs Berkeley Sound
17/12/92  Discovery arrives in Punta Arenas

1.4  Scientific achievements

The cruise track is shown in Figure 1.1*.  The first objective listed above 
was addressed by two transects, one across the Drake Passage and one along 
longitude 88W.  Measurements of a wide range of gases in surface water 
(oxygen, carbon dioxide, sulphur gases, halocarbons and hydrocarbons) were 
complemented by biological, physical, and meteorological measurements.  The 
line of the Drake Passage transect was chosen to coincide with a WOCE repeat 
section, thus allowing the physical oceanographers to address WOCE objectives 
simultaneously with the BOFS studies.  Similarly, the 88W transect coincides 
with a WOCE Hydrographic Programme line to be worked during 1993, with the 
hope that the BOFS data will complement that of WOCE.  Both transects crossed 
the Subantarctic and Polar Fronts and will, when the data are worked up, 
provide valuable information on gas concentrations and fluxes in those areas.  
The 88W transect is particularly valuable since it is an area where no 
previous pCO2 measurements are available.

The second objective was addressed by two intensive grid surveys, together 
with a CTD section, centred around longitude 85W (see Figures 1.2* and 1.3* for 
the survey tracks).  The surveys were designed to map an intense phytoplankton 
bloom which extended southwards for approximately 70 miles from a sharp 
northern boundary close to 67S.  The first survey covered the northern edge 
of the bloom, and the second the southern edge.  The two surveys overlap, thus 
providing information on the temporal evolution of the bloom.  A wide range of 
physical, chemical, and biological measurements were carried out during the 
survey, as can be seen from the detailed scientific reports in section 3.

Many features of this bloom remain unexplained at the present time.  We had 
envisaged in the proposal that blooms close to the ice edge would be formed in 
shallow mixed layers of lower salinity water, i.e. stabilised by fresher water 
from melting ice: this expectation was based on (scanty) published data from 
the Ross and Weddell Seas, there being no previous work from the 
Bellingshausen Sea to guide us.  The bloom we found, in waters which had been 
largely covered by brash ice only a few weeks before, was actually in water 
more saline than the surface water immediately to the north (which contained 
very little chlorophyll; see Figure 1.4*).  Our colleagues on the James Clark 
Ross reported that even at the ice edge, which was to the south of our working 
area, there was no sign of a mixed layer stabilised by melt water.  Nutrients 
were plentiful both within and outside the bloom.  It is clear that this bloom 
is unlike those encountered in the North Atlantic in previous BOFS cruises.

In scientific terms, the cruise can be judged successful.  We have 
successfully mapped a large phytoplankton bloom off the ice edge.  In 
addition, station work by the James Clark Ross within our survey area has 
complemented our underway survey measurements with a wide range of in situ 
rate and particle flux measurements.

Figure 1.1*  Cruise track

Figure 1.2*  First survey track. Capital letters show the names given to the 
             survey legs (section 3.8; Appendix C).

Figure 1.3*  Second survey track. Capital letters show the names given to the 
             survey legs (sections 3.8, 3.10; Appendices C, D).

Figure 1.4*  Chlorophyll and salinity sections (leg A, first survey). Letters 
             above the upper panel show CTD section locations: A - deep CTD; B 
             - 1000m CTD; C - productivity station.

1.5  Logistics and equipment

A number of problems in this area added to the difficulties of the cruise, and 
limited the amount of scientific work which could be carried out, although the 
major cruise objectives were achieved as reported above.  The major problem 
areas are summarised here; a full report has been submitted separately to RVS.

First, the ship.  This was the first scientific cruise on the Discovery 
following a major rebuild which added 11m to the ship's length.  Prior to the 
cruise, a significant time had been spent on winch trials, but very little 
time had been available for scientific trials.  This resulted in some delays 
and slow progress, particularly in the first few days of the cruise, since 
many systems were being used or deployed for the first time.

Second, the ship's chef became seriously ill in the early part of the cruise, 
and had to be evacuated to land by the James Clark Ross.  This was a major 
loss of science time to Sterna 92 as a whole.

Third, the loss of the SeaSoar halfway through the second survey was a major 
blow to the scientific programme.  We were fortunate to be able to borrow a 
UOR from the James Clark Ross to enable us to collect at least some data on 
mixed layer structure during the final part of that survey.

2.   CRUISE NARRATIVE
2.1  Mobilisation

Following cancellation of our original flights (2/11/92) by the RAF, the 
scientific party flew to Stanley in two groups: an advance party of 6 scientists 
and 2 RVS technicians (arrived in Stanley 30/10/92), and a main party (arrived 
in Stanley on 7/11/92 after a 30 hour delay in their RAF flight).  Discovery 
arrived in Stanley on 5/11/92, 48 hours later than planned, having been delayed 
60 hours in Capetown awaiting the arrival of spares by airfreight.  Departure 
from Stanley on the planned date of 8/11/92 was impossible because (i) the late 
arrival of Discovery and of the main scientific party meant that preparation of 
scientific equipment for sailing was far from complete, and (ii) it was 
necessary to await further airfreighted ship's spares, which arrived on the 
evening of 10/11/92.  Discovery then left Stanley for Berkeley Sound for 
bunkering: this was completed on the morning of 11/11/92.  The scientific log 
begins on departure from Berkeley Sound.

With the benefit of hindsight, it is apparent that Stanley is an inconvenient 
port for RVS ships.  Port calls at Stanley leave scientists and RVS staff 
reliant on the RAF, whose Tristars provide the sole air link between the UK 
and the Falklands.  Seats on the RAF flights are available only when not 
required by the military, and are not confirmed until a few weeks before 
departure: we were given 4 weeks notice that our request for seats on 2/11/92 
could not be met.  It also appears that delays to the flights are not 
uncommon: our main party were delayed 30 hours at Brize Norton, while the 
scientific party returning from the James Clark Ross BOFS cruise were delayed 
24 hours at Ascension Island.  It is recommended that future Southern Ocean 
cruises should use ports served by commercial airlines wherever possible.

2.2  Scientific log

Note: all times are in GMT and all dates as day number: see Appendix I for day 
      number to date conversions.

Drake Passage transect

Discovery got under way at 316/1400 following completion of bunkering.  Since 
there had been no scientific trials prior to this cruise, the first requirement 
was to test all equipment to be deployed overside.  We steamed east in order to 
reach 300m depth for a shallow CTD cast. On reaching this depth (1615), further 
work was required before the CTD was deployed, so we steamed south to remain at 
this water depth, heaving to at 2230 for a shallow CTD cast.  This was 
completed successfully, together with firing of Niskin bottles.  This was 
Discovery station 12198, the first scientific station for over two years.  On 
the same station, a 30L GoFlo cast was also successfully completed, using a 
kevlar line on the starboard gantry.  The BAS acoustic fish was given a brief 
test, revealing some problems in the deck unit.  Deck lab watches, including 
regular water sampling, started at 317/0200.  Instruments on line at that stage 
were pCO2, oxygen, and the autoanalyser.  The RVS surface system had also begun 
logging. ADCP calibration was carried out between 317/0030 and 317/0353.  The 
James Clark Ross had turned south at that time, having reached 65S 84W.

SeaSoar was launched at 317/1355, and was soon undulating well to 380m.  It 
was decided to proceed with the Drake Passage SeaSoar and ADCP transect for UK 
WOCE.  The SeaSoar cable was shortened to 90m for passage over Burdwood Bank 
(317/1826 to 320/0502).  The Turner Designs fluorometer was by this time on-
line in the deck lab.  Lightfish was deployed at 318/1458 using the starboard 
stern crane instead of the Schatt Davit, which towed the fish too close to the 
ship.  Even on the crane, the lightfish appeared to tow close to the SeaSoar, 
which was itself to starboard of the centre line.  The potentiometric pH 
system was on-line at this stage, and tests of the various batches of TCO2 
chemicals continued.  Major problems were encountered with the oxygen 
titrator.  The size-fractionated productivity groups began a 3-4 day 
experimental sequence.

At 319/0940 the SeaSoar cable was shortened to 200m due to fog, increasing 
again to 600m at 319/1400 in clearer weather.  At 319/1652 SeaSoar cable was 
again shortened (to 300m) and speed reduced to 6 knots in poor visibility.  
The first icebergs were sighted on the radar south of the Polar Front in very 
cold surface water(-1C).  The remainder of the SeaSoar transect was completed 
with 300m cable out at 6 knots: SeaSoar undulating to 200m but not reaching 
the surface.  SeaSoar was recovered at 320/0745.  Even in calm water the 
SeaSoar swung to within 18" of the ship's stern on leaving the water, showing 
the need for improved control of the fish on recovery.  Discovery then 
continued on passage to Potter Cove (King George Island) for calibration of 
the acoustic fish.  At 320/1100 the fog cleared, giving a stunning view of 
Elephant Island as the ship steamed into the Bransfield Strait.  At 320/1330, 
the wire jumped off the PES winch sheave during recovery of the acoustic fish.  
Fast reactions by Colin Day, and skilful work by RVS and ship's staff ensured 
a safe recovery.  Deck lab watches stood down at 320/1900 following completion 
of the transect into the Bransfield Strait.

Potter Cove

Discovery anchored in Potter Cove at 321/1233. Another stunning view! The 
initial anchorage dragged, with re-anchoring complete by 1420.  The acoustic 
fish calibration began at 1436, and was complete at 2040.  Radio contact was 
made with the Argentine research base Jubany, which is situated on the shore 
of Potter Cove: Jane Robertson communicating very effectively in Spanish.  At 
1700, when the weather had abated a little, a small party (David Turner, Jane 
Robertson and Anne Morrison) went ashore to the base, bearing traditional 
gifts (whisky and wall shield) which were well received.  We were given a 
friendly reception, although there were few English speakers on the base.  
Plans to entertain some of the base staff to dinner on Discovery, and for more 
scientists to visit the base, were curtailed by worsening weather.  Getting 
the ship's boat off a lee shore for return to the Discovery could be described 
as challenging, but was safely accomplished thanks to the efforts of Andy 
Adams, Phil Gauld and Ian Slater.  All were back on board at 2050.

Bellingshausen Sea transect

With the Drake Passage transect and acoustic calibration completed, Discovery 
headed for the 85W longitude along which the James Clark Ross had made their 
transect into the ice.  We headed initially for 65S 85W, following which the 
intention was to carry out an exploratory transect southwards.  SeaSoar was not 
be towed on this westward transect - the physicists had discovered that the CTD 
level A had been incorrectly programmed, and had to reprocess the raw data 
files from the Drake Passage transect.  Changes to the code could not be made 
on board (development software not licensed for this ship), so revised code was 
awaited from Barry in time for the next Seasoar deployment.  On day 321 one of 
the thermosalinograph temperature sensors failed, leaving us with no high 
quality sea surface temperature measurement (although in the event the ADCP 
sensor gave good data, see section 3.4).

Productivity stations (300m CTD followed by GoFlo casts on kevlar) were carried 
out at 1100 on days 322 and 323, complete within 2 hours each time.  The 
acoustic fish was recovered at 323/1300 following completion of a shelf break 
transect to allow the ship to proceed quickly to 85W (the acoustic fish cannot 
be towed safely above 10 knots).  Most of the rest of day 323 was spent doing 3 
knots in thick fog!  The fog cleared at 2000.

A third daily productivity station at 324/1100 resulted in the loss of a 30L 
GoFlo bottle when the kevlar line broke, for reasons which were not clear.  
The weather then steadily worsened, and by 1540 the lightfish was recovered 
and the ship battened down in 40-50 knot winds.  At 1718 the roller door to 
the water bottle annex was stove in by a wave, which proceeded to flood the 
deck lab to a depth of several inches, causing wet feet but no other damage.  
The ship then hove to for repairs, which were complete at 325/0030.  During 
day 324 the Master became increasingly concerned about the health of Chef Glyn 
Davies, and spoke at length to the doctor on board the James Clark Ross.  As a 
result, it was agreed that the two ships should rendezvous as soon as possible 
to allow the doctor to examine Glyn.  On-line measurements and sampling 
continued in the deck lab during passage to the rendezvous.

The two ships met at 325/1600: the James Clark Ross doctor came over to 
Discovery, examined Glyn Davies, and advised that he required urgent 
evacuation to shore on medical grounds.  The use of Discovery for this 
evacuation was offered, but the doctor was clear that medical considerations, 
which were paramount in this situation, favoured the use of the James Clark 
Ross.  This was a major blow to the scientific programme on the James Clark 
Ross, which had already suffered one medical evacuation during their cruise.  
Glyn was duly transferred and evacuated to Stanley, receiving a blood 
transfusion on the ship.  He was later reported to be recovering well in the 
Falklands.  During the rendezvous, Nick Owens visited Discovery to discuss 
plans for the two cruises, and intercalibration CTD casts were carried out.  
The ship remained hove to for the evening.  On day 326, following a 
productivity station at 1100, the ship remained on station until 1500 to allow 
completion of welding work to make good storm damage.  At the same time, the 
oxygen and alkalinity photometers were moved to the cooler climes of the water 
bottle annex and the hangar respectively to try to overcome the problems of 
intense degassing of samples during titration.  At 326/1500 the weather was 
too poor to launch SeaSoar as hoped, so we carried out instead a section of 
1000m CTD stations every 20 miles from 82W to 85W along 65S.  The surface 
water was much warmer at 85W, probably on the edge of the polar front.  The 
westerly section was completed at 328/0030, with weather still too poor for 
launching SeaSoar, so a southerly CTD section at 20 mile spacings was begun 
along 85W.  On the southerly course there was a heavy swell on the starboard 
quarter, but there seemed to be no danger of being pooped at 10 knots, much to 
the relief of the productivity groups with their incubators on the aft deck.

First survey

The CTD station at 6540'S, together with productivity casts, was completed by 
328/1324, and with an improvement in the weather SeaSoar was launched and a 
southerly transect begun. We crossed the Antarctic Circle at 2134.  By 329/0200 
we had reached 6709'S, and headed northwest away from the ice during the hours 
of darkness to avoid the need to recover SeaSoar: the ice edge at this time was 
believed to be between 6730'S and 68S.  A southerly course was resumed at 
329/0800 in improved light and weather: a further northwesterly diversion was 
caused by a blizzard between 1100 and 1150.  When the southerly course was 
resumed we were at longitude 8620'W. South of 67 we encountered high 
chlorophyll (seen briefly also at the southern end of the earlier transect on 
85W).  The James Clark Ross had earlier encountered high chlorophyll at this 
latitude when the area was covered with brash ice.  It was decided to survey 
this chlorophyll patch with north-south legs at 8 mile (20' longitude) spacings, 
with northern and southern limits of 6630'S and 68S.  In order to minimise 
speed reductions due to poor visibility, the survey was arranged so that the 
northern end was carried out at night and the southern end during the day.  The 
current tow along 8620'W was declared to be the first leg of the survey.  The 
intention was to work to 84W, and to ask the James Clark Ross to work from 68S 
into the ice on their return.

The first 6 legs of the survey worked according to plan (see Figure 1.2*), 
although leg Z (8520'W) was terminated early in order to keep the correct 
day/night timing.  Some deviations were required to avoid icebergs, and 
occasional slowing was required in fog.  On day 331, Nick Owens confirmed that 
the James Clark Ross would work southwards down 85W on their arrival 
(expected day 332).  The timetable for their ice station work had been reduced 
from 4 days to 2 days per station in view of the time lost. On day 332, it was 
impossible to keep to the southerly course planned for leg C (8420'W) owing 
to 50 knot winds and heavy seas; the track drifted steadily to the east.  At 
332/1943 the survey was abandoned, and the ship headed northwest into the 
weather.  A further problem concerned the pendulum arm attached to the stern 
gantry: it could be seen that the securing bolts had worked loose, rendering 
the arm liable to come away entirely if the remaining supports (hydraulic ram 
and weld) gave way.  Since the arm was above the SeaSoar recovery position, 
SeaSoar could not be recovered either.

On day 333 the weather had moderated: at 0940 the lightfish was recovered and 
the SeaSoar cable shortened.  The weather was calm enough to allow access to 
the top of the stern gantry, and temporary repairs to the pendulum arm were 
completed by 1727, following which SeaSoar was recovered using a drogue to 
keep the fish away from the stern: this worked very successfully.

Intensive CTD section

The next scientific activity was a CTD section crossing the northern edge of 
the chlorophyll patch: the water to the north is less saline, flowing 
easterly, and has very low chlorophyll concentrations (Figure 2.1).  The 
survey had revealed a large bloom, but nothing like the meltwater-stabilised 
mixed layer envisaged in the proposal!  The plan was to carry out casts every 
5 miles from 6645'S to 6730'S along 85W, with 4000m casts at the ends, 
1000m elsewhere (Figure 1.4).  The stations at 6705'S and 6715'S 
incorporated productivity casts in low and high chlorophyll water 
respectively.  The first deep cast was very slow (333/2227 to 334/0543) due to 
spooling problems on recovery, although this problem was successfully overcome 
by Colin Day and Tony Poole, and deep casts ran very smoothly thereafter.  
When the first deep cast was eventually completed, the multisampler was found 
to have fired all the bottles at the same depth.  A replacement multisampler 
was installed ready for the next cast at 334/1625.  The remainder of the 
section proceeded without incident, and bottle samples eventually obtained for 
all stations.  The section was completed at 335/2125.  Despite the large 
number of samples, the deck lab crew managed to cope with full sets of oxygen 
and nutrient samples from this section.

Second survey

The initial plan for the second survey was a repeat of the first, but moving 30' 
further south to map the southern edge of the chlorophyll bloom, and to study 
the evolution of the bloom itself.  Contacts with the James Clark Ross revealed 
that the chlorophyll decreased to a low level at about 6815'S, and remained low 
all the way to the ice edge at 69S: this indicated that there was no value in 
the Discovery working all the way to the ice edge.  The SeaSoar was launched at 
335/2230, and the ship steamed towards the survey start point at 67S 84W (the 
survey worked east to west this time, again synchronising the north/south cycle 
with night/day in the same way).  However, SeaSoar was recovered at 336/0300 
before the start of the survey itself, due to incipient bad weather.  Recovery, 
in poor conditions, using the drogue once again, plus lateral lines, kept the 
fish well under control.  When the weather had abated, the SeaSoar was launched 
once again at the start point (67S 84W) at 337/0200 and the survey restarted.

On the first leg, there was a sharp change in chemistry at about 6825'S, with 
pCO2 increasing rapidly to equilibrium with the atmosphere, and nutrients also 
increasing sharply.  The leg was continued to 6835'S to sample these new 
conditions fully: however, when the SeaSoar data were processed it was clear 
that this had not been far enough south to define the physical boundary, so 
with effect from leg B (8440'W) the survey was run from 6730'W to 69S in 
order to map the southern boundary fully (see Figure 1.3*).  The northern 
boundary of the chlorophyll bloom was abandoned in order to keep the survey 
legs the correct length for a repeating day/night cycle.  All went well until 
SeaSoar was lost at 338/1037 just after completing leg A.  Separate reports 
have been submitted on this event.

In order to obtain mixed layer information for the western half of this second 
survey, we requested loan of the UOR, plus driver Ian Bellan, from the James 
Clark Ross.  To our pleasure they readily agreed.  The James Clark Ross was 
then on station at 6815'S  85W, so we steamed along leg Z (8520'W) to 
collect surface data before rendezvous at 339/2045.  The transfers were 
completed successfully, and a new UOR cable was wound onto the SeaSoar winch.

In order to re-establish the day/night cycle correctly, we steamed to 6740'S 
8520'W, and launched the UOR at 340/0518.  Ian Bellan had advised that tows 
of no more than 8-9 hours should be attempted due to limited battery life in 
these cold waters.  The UOR was recovered at 1327, and launched again at 1700 
for leg Y (8540'S).  However, shortly after commencing this leg, examination 
of the first leg data revealed a problem with the temperature sensor.  The UOR 
was recovered and the sensor replaced, launching again at 2300.  This leg was 
completed successfully, but a fault was found with the UOR conductivity sensor 
on recovery.  Further attention by Ian Bellan solved this problem, and two 
further tows were carried out along legs X and YZ (86W and 8530'W), these 
being chosen to cover the western half of the survey as effectively as 
possible.  All survey work was completed by 342/0527.  Discovery hove to close 
to the James Clark Ross at 342/1230, but no transfer was possible due to poor 
weather.  A productivity station was carried out while hove to.  The Master of 
the James Clark Ross preferred to wait until day 343 for transfer, and 
transfers were finally completed by 343/1250, although the weather was by then 
poorer than on the afternoon of day 342.  Figure 2.1* shows surface chlorophyll 
from this survey.


Figure 2.1*  Surface chlorophyll, both surveys


88W transect

This transect had originally been envisaged as a combined SeaSoar/surface 
chemistry/CTD transect.  With the loss of the SeaSoar, it was decided to carry 
out CTD stations at 2.5' (150 mile) intervals, with XBT casts every 20 miles, 
and full surface chemistry and biology sampling, with the addition of 
particulate sampling (section 3.28).  A limited number of productivity 
stations were also worked.  The transect was started at 69S (344/0024) in 
order to assess the extent of the chlorophyll bloom at this longitude.  The 
final station at 5130'S was completed at 349/1315.


3.   SCIENTIFIC ACTIVITIES

Note: all times are in GMT and all dates as day number: see Appendix I for day 
      number to date conversions.

3.1  Navigation 
     (Gwyn Griffiths)

Discovery is presently fitted with five satellite navigation receivers; four 
operate with the Global Positioning System (GPS), of which one is the radio-
code clock, and the fifth operates on the Transit system.  Of the GPS 
receivers, the bridge tend to prefer the Racal/Decca receiver for navigation, 
with the older Trimble receiver as a back-up.  Scientific navigation 
information comes from the Trimble.  The other GPS receiver is a 24 channel 
Ashtech instrument, used primarily for attitude measurements - see below on 
measuring the ship's heading.

Dead reckoning

Unlike the Trimble receiver on Charles Darwin, the one on Discovery is not 
supplemented by a rubidium clock that allows fixes when only two satellites 
are visible.  The periods of two satellites amounted to about 45 minutes a day 
in the working area.  Therefore, the lack of an external clock compromised the 
navigation for a significant period, for both deck officers and scientists.  
The traditional method of coping with an absence of navigation fixes is to use 
dead reckoning, using ship's heading and speed.  However, the electromagnetic 
speed log had been unusable since Cape Town.  Phil Taylor and Bill Miller 
traced the problem to the sensor unit itself, and searched in vain for a spare 
unit on board.  There was no spare - this is unacceptable for such an 
essential navigation tool.  There were occasions on which it was necessary to 
keep the bridge informed by telephone from the main lab every 2 minutes of the 
ship's speed through the water from the ADCP.

GPS availability

The availability of GPS coverage, irrespective of quality, during the cruise 
was generally acceptable, indeed had the rubidium clock been available, it 
would have approached 100%.  Typically, there were 30 gaps a day of over 1 
minute, with about 6 greater than 10 minutes.  However, many fixes were of poor 
precision due to satellite geometry, the cumulative time of poor fixes (pdop>5) 
amounting to about three hours a day.  The 'mission planning' software supplied 
with the Ashtech GPS receiver allows graphical and tabular displays of GPS 
availability and it was invaluable in planning the time of the adcp 
calibration, when good GPS coverage is essential.  As the almanac data are 
obtained from the satellites each pass, the information is up-to-date and 
accurate.

GPS accuracy

The quoted accuracy of GPS with Selective Availability (SA) in operation is 
that a position is within 100m of the true position for 95% of the time (SA is 
the deliberate degradation of the GPS signal for commercial users).  The error 
can be estimated from a stationary ship, and this was done whilst tied up to 
the pontoon at Port Stanley.  Over an 18 hour period, the standard deviation of 
the longitude was 25.8m, and that of the latitude was 32.5m from the Trimble 
receiver.  The comparable figures for the Ashtech receiver were 27.4m and 
37.9m.  This accuracy is acceptable for normal navigation requirements, but 
when deriving ship's speed over the ground for the calculation of absolute 
velocities from the ADCP the degradation introduced by SA is very noticeable.

Navigation procedure

The raw one second GPS position data from the Trimble were read from the level 
C into PStar.  Fixes with a pdop of greater then 5 were rejected, and then 
averaged to provide positions one minute apart.  Gaps in the position data 
were filled by linear interpolation between good fixes as no DR was possible.  
The speed of the ship over the ground, and the distance run were then 
calculated from the one minute data.  The ship velocity over the ground needed 
for obtaining absolute currents from the adcp were obtained by taking the one 
minute velocities, median filtering and then averaging to 15 minutes, to match 
the adcp averaging interval.  Archive copies in PStar format of the raw GPS 
and Transit navigation data were made at the end of the cruise.

Measuring the ship's heading - gyro and 3D GPS

The ship's heading is traditionally measured using a gyrocompass, the new 
Discovery is fitted with two SG Brown Mk1000 instruments, known as No.1 and 
No.2.  Heading input to the shipboard computer system was from No.1 gyro 
throughout the cruise.  The No.1 gyro was also used as the navigating gyro 
from leaving Port Stanley until 334/1600, when a power failure caused both 
gyros to shut down.  As No.2 settled quicker, it was switched to act as the 
navigation gyro.

This cruise saw the first operational use on an NERC research vessel of a new 
system to determine heading based on differential GPS measurements.  By 
measuring the relative positions of an array of four antennas to a precision 
of 1-2mm over baselines of 6-10m, it can provide heading measurements to an 
accuracy of 0.05.  The technique is free from drift and heading dependent 
errors that can reach 2-3 with marine gyrocompasses.  The primary reason for 
installing the system is to provide accurate heading for underway ADCP current 
measurements.  However, the GPS heading data are not continuously available, 
as each of four antennas needs to be locked on to four satellites to acquire 
attitude information.  With the present incomplete GPS constellation, this 
requirement is only met for part of the day, such that only some 40% of 15 
minute intervals through a day had attitude data from GPS.  The system also 
suffers from short-term drop-outs, even in the presence of good satellite 
coverage, and no data are available for tens of seconds whilst the system re-
acquires lock.  This may well be due to two of the antennas being located next 
to the lifeboats, and it is suggested that they be raised whilst maintaining 
their rigidity.

Figure 3.1*  Ashtech GPS heading (hdg), gyrocompass heading, and difference (a-
             ghdg) during the 3 day crossing of the Drake Passage. The slow 
             drift amounts to 1.5

The new heading data were used to correct the gyro headings, the gyro then 
acting as an interpolation method between 3D GPS data.  A series of four C 
shell scripts running PStar programs (including two new ones) were written to 
handle the attitude data and to apply the corrections to the adcp velocity 
data.  An example of where the GPS system showed a significant drift in the 
gyrocompass was along the Drake Passage section.  A drift of 2 occurred over 
the three days of the passage, as shown in Figure 3.1*.  The effect of 
correcting the adcp for the heading error was to reduce the mean east 
component of current from 11cm.s^-1 to 7.1cm.s^-1, the difference, 3.9cm.s^-1, 
being due to a spurious component to the left of the ship's track caused by 
gyro error.

3.2  Acoustic Doppler Current Profiler 
     (Gwyn Griffiths, Bill Miller)

Instrumentation

The RD Instruments 150kHz ADCP was in operation throughout the cruise.  
Neither the instrument or the IBM data acquisition system (DAS) gave any 
hardware problems.  The firmware version was 17.10, released on 5 March 1992, 
and the DAS software was version 2.48.  The system PC should not be run in 
'turbo' mode as this will cause gpib comms errors.  VMtest was run whilst at 
Port Stanley, with satisfactory results.  The firmware revision incorporates a 
new 'fish rejection' algorithm and minor improvements to bottom tracking.

Air was bled from the transducer space several times during the cruise, 
probably the air was getting trapped during the frequent periods of heavy 
seas.  It is recommended that a more automated method of bleeding the system 
should be devised.

Whilst alongside the pontoon at Port Stanley, the ADCP heading was checked 
against gyro compass No.1, and was found to agree to 0.1.  The ADCP 
temperature sensor was also checked against the hull temperature readout on 
the bridge - the ADCP showing 0.9 low.  The configuration was 64 by 8m cells, 
the ensemble average period was two minutes, and the auto scaling mode was 
selected for sound velocity correction using the ADCP sea surface temperature.

Calibration

The zig zag course technique was used to calibrate the ADCP whilst steaming at 
8 knots on the continental shelf on passage from the Falklands to the first 
station, each leg of the zig zag was 20 minutes.  The scaling factor (A) and 
the misalignment angle (phi) were calculated following the procedure developed 
on previous cruises.  However, after the Drake Passage section, it became clear 
that the calibration was incorrect.  There was a bias towards the left of the 
ship's track, and currents approaching and leaving a station were inconsistent 
with those on station.  The source of the problem, and the solution, took a few 
days to puzzle out.  The culprit was found to be the method of calculating the 
ship's velocity from successive 1 minute GPS fixes averaged over the steady 
part of the zig-zags.  This method had proved itself using GPS without 
degradation due to Selective Availability (SA).  Evidently, with SA in full 
effect, the method did not work, as it introduced a bias into the ship speed 
estimate because of the 'bad' fixes that were thrown out before forming the one 
minute averages.  The effect of this editing was to bias the ship velocity.  
For example, setting an acceptance window of 500cm.s^-1 for ship speed rejects 
spikes above 500cm.s^-1 but does not reject a spike to -499cm.s^-1 from a true 
value of +400cm.s^-1.  The solution was to calculate the mean ship speed during 
the zig zag from the positions at the start and end of the steady section only.  
This revised procedure gave an amplitude scaling factor (A) of 0.997 with a 
standard deviation of 0.016  The alignment angle error (phi) was -1.55 with a 
standard deviation of 0.65.

Data processing

The majority of the data processing for the ADCP was done using PStar command 
scripts as developed on Vivaldi and previous cruises.  The ADCP data were 
processed every 24 hours as one set.  After all calibration, editing and 
merging with navigation, it was split into two series: 'on station' covering 
periods of minimal ship speed whilst on CTD and productivity stations, and 
'underway' covering SeaSoar runs and all associated course changes and 
manoeuvres.  Obtainable ranges varied from in excess of 450m down to 100m, 
primarily related to the amount of scattering organisms present in the water 
column and the ship's speed.  The greatest ranges were achieved in the 
plankton bloom and when the wind and sea were low - the lowest ranges were on 
courses steaming into moderate to high winds (>25 knots) and at ship speeds in 
excess of 10 knots.  At such times there were periods of total data loss, 
where the % good did not exceed 25% at any depth, and there were periods where 
there was an indication of a spurious current along the ship's track.  
However, the installation is a great improvement on the original Discovery 
installation with the transducer in the asdic pod, and subjectively, the 
performance in poor weather was better than, or certainly  no worse than on 
the Charles Darwin.

Results

Colour contoured velocities were plotted for each section of the SeaSoar 
surveys.  They showed clear correlation between the ADCP cross track currents 
and the density structure.  In general, the currents showed little shear, the 
features extending to beyond the depth range of the instrument.  Maps of the 
currents at 101m along the survey tracks were also produced.  Whilst on 
station the ADCP velocity measurements showed much lower scatter than whilst 
underway.
 
The ADCP backscatter strength forms a broad-brush estimate of zooplankton 
abundance.  The two minute ensemble averages for all beams clearly showed 
diurnal migration, and enhanced scattering near fronts and eddies.  The 
contrast in scattering across the polar front in the Drake Passage was 
particularly remarkable.  Single ping data were also collected from the 
individual beams.  At times, these data showed the presence of krill swarms, 
with spatial dimensions such that only some of the beams were affected, and 
that the 2 minute average (covering typically 500m) hardly showed a signature.  
The backscatter intensity of such swarms was at times over 50dB above the 
background level, and the new RD fish rejection algorithm was turned off to 
avoid data rejection.

3.3  Meteorology 
     (Polly Machin, Gwyn Griffiths, Bill Miller)

Instrumentation

The meteorological monitoring system used on Discovery is a modified version 
of the metlogger installed on James Clark Ross in 1991 and comprises the 
following instruments:

(i)   an R.M.Young Instruments 05103 wind monitor (sn 11277), including wind 
      vane and anemometer - this is situated on the foremast
(ii)  two Vector Instruments psychrometers, located port and starboard on the 
      foremast (sn 1072 and 1073 respectively), measuring wet and dry bulb air 
      temperatures
(iii) two Didcot cosine collector PAR sensors (spectral range 400-700nm)  
      located port and starboard on the foremast (sn 27150 and 27151
      respectively)
(iv)  two Kipp and Zonen total irradiance sensors located port and starboard 
      (sn 92015 and 92016 respectively)
(v)   a longwave pyrgeometer, fitted by Keith Birch prior to Discovery's 
      departure from UK waters, also on the foremast;
(vi)  a hull-mounted RS Components platinum resistance thermometer, recording 
      sea surface temperatures
(vii) a Vaisala DPA21 aneroid barometer (sn 465569), located in the main lab

(ii)  and (v) are IOSDL instruments replacing components of the standard RVS 
      system. The metlogger system was designed as a general purpose 
      meteorological data package for use on cruises not requiring 
      meteorological research standard instrumentation and therefore complements 
      the IOSDL designed Multimet system which is geared towards a higher 
      standard of performance. The Discovery metlogger system was developed as 
      part of a joint project between BAS and the Instrument and Sensors Group 
      at RVS and uses modular sensor packages and signal conditioning. For all 
      sensors apart from the barometer, the conversion to digital data takes 
      place at the module (by Rhopoint) and data are transmitted to the logger 
      by an RS485 link (data from the barometer are transmitted by an RS232C 
      link). Pre-cruise maintenance and inspection of the foremast sensors was 
      carried out in Port Stanley, but no maintenance was done at sea.

Data processing

As this cruise was the first time the metlogger system had been used on 
Discovery, considerable thought was given on how best to process the data from 
the various instruments.  Unlike most shipboard instruments that have a 
dedicated Level A interface, the metlogger PC emulates a standard Level A 
interface and transmits the data directly to the Level B in Ship Message 
Protocol (SMP).  The data are transferred to the Level C and then reformatted 
from Level C to PStar format to allow processing under Unix, using a series of 
pexec scripts based on the set of scripts used for the IOSDL Multimet system.  
However, given the new instrument configuration on this cruise and a number of 
errors and inconsistencies in the Multimet scripts, a considerable degree of 
rewriting was necessary.  Additionally, since the emlog was unavailable for 
the cruise, the speed of the ship over ground was calculated from GPS position 
fixes, averaged over one minute.  Note that all wind velocities in this report 
are given in meteorological convention, i.e. where the wind is coming from - 
this differs from the standard oceanographic convention for currents etc which 
are defined as where they are going to.

The Unix shell script metexec0 was used to retrieve 24 hour sections of data 
from the Level C and convert them into PStar format.  Metexec1 was used to 
calibrate all instruments apart from the aneroid barometer and wind monitor, 
and histograms of the calibrated output were produced for all sensors as a 
range check, to allow editing of obvious spikes.  Ship's navigation data were 
merged with the met file by metexec2.  Metexec3 would normally be used to 
merge the emlog file and was therefore not used for this cruise.  Ship's 
heading (gyro) was merged by metexec4 and a combination of the ship's velocity 
components and heading was used in metexec5 for the conversion from relative 
to absolute wind velocities.  Metexec6, an appending script was used to 
generate a full time series from the individual 24 hour files and 10 minute 
averaged data files (vector averaging for the absolute wind velocities) were 
created by metexec7, both for general use by interested scientists and to 
allow the production of time series and vector summary plots.

Calibration

With the exception of the aneroid barometer and wind monitor, which were 
calibrated by the manufacturers prior to installation and were therefore 
logged through to the Level B as calibrated output, all instruments were 
calibrated during PStar processing of the met. data.  The calibration 
algorithms applied were derived either from manufacturers calibration 
certificates or from calibrations undertaken by RVS and IOSDL prior to the 
cruise.  Details are given in Table 3.1.

Table 3.1  Calibration coefficients for the met. sensors

Measurement         Calibration         Source        Comments
--------------------------------------------------------------------------------
wind speed          y = 0.1x            manufacturer  dm.s-1 to ms-1
wind dir            none                manufacturer  calibrated on 
                                                        installation
port wet bulb temp  a = -23.71101       IOSDL         equation takes the form:
                    b = 6.84806E-3                      y=a+x(b+x(c+dx))
                    c = 5.626587E-6    
                    d = 1.077627E-9    
port dry bulb temp  a = -23.84735       IOSDL         as above
                    b = 5.788879E-3    
                    c = 5.648462E-6    
                    d = 9.076649E-10    
starboard wet bulb  a = -21.63646       IOSDL         as above
  temp              b = 2.580562E-3    
                    c = 7.893778E-6    
                    d = 6.608683E-10    
starboard dry bulb  a = -20.18834       IOSDL         as above
  temp              b = 9.73387E-4    
                    c = 7.835114E-6    
                    d = 5.250384E-10    
sea surface temp    a = 2.9755E-4       RVS (range    equation takes the form: 
                    b = 0.99189         +5 to +25C)    y=ax2+bx+c
                    c = 0.26705    
port PAR            y = x/(5*12.86E-6)  manufacturer  5 is the signal 
                                                        amplification factor
starboard PAR       y = x/(5*12.87E-6)  manufacturer  as above
port total          y = x/(2*48.49E-3)  manufacturer  2 is the signal
  irradiance                                            amplification factor
                                                        amplification factor
starboard total     y = x/(2*43.63E-3)  manufacturer  as above
  irradiance      
longwave radiation  y = 0.23364486x     IOSDL         includes a *5 signal 
                                                        amplification factor
barometric          none                manufacturer  calibrated prior to 
  pressure                                              installation


Performance

Generally, the instruments performed very well, particularly for a first 
cruise.  The PC display was very useful and frequently consulted, particularly 
when crossing fronts, during storms and generally by those who resented being 
cut off from reality by dogged portholes!  There were, inevitably, a couple of 
problems which have been detailed below, but these are minor details in a 
system which has generally proven reliable and readily usable.

The port dry bulb sensor was very noisy throughout the majority of the cruise, 
particularly at near-zero and sub-zero temperatures; the reason for this is 
not clear.  However; the remaining signal is strongly correlated with the port 
wet bulb temperature and is almost certainly worth keeping - the noise can 
easily be flagged out at BODC using an interactive graphical editor. 

The longwave radiometer channel gave intermittent communication errors, which 
most likely arise from the Rhopoint module.  Both PAR sensors recorded 
negative irradiances during dark periods.  At present it is not clear whether 
this is a calibration offset, in which case the entire PAR profiles will be 
consistently underestimated by approximately 5Wm^-2 or whether it is a 
nonlinearity near zero, as suggested by comparison at low-zero light levels 
with the output from the PML PAR sensor.  However, this comparison is of 
limited use due to the different configurations of the two types of sensor.  
(The Didcot instruments are cosine collectors; the PML sensor is hemispherical 
- i.e. 2-pi).  Recalibration of the sensors would be the only unequivocal way to 
resolve this problem.

The conversion from relative to absolute wind velocities is dependent upon a 
continuous record of heading from the ship's gyro.  The current PStar 
processing script merges the ship's navigation file to the met file by linear 
interpolation: this is not suitable for a parameter such as ship's heading 
which "wraps around" at 360.  The problem was circumvented on this cruise by 
going back to the original 1 second gyro file, but this is, at best, a 
temporary solution.

Redrawing of the LabTech display on the PC causes an unacceptable loss of data 
acquisition.  This is exacerbated by users' impatience with the slow screen 
redraw, sometimes resulting in several redraw commands being issued.  
Separation of the screen handling and data acquisition functions of the system 
would be a welcome bonus - the display is an excellent one, but need not 
compromise data acquisition.

Results

Winds were predominantly northeasterlies and northwesterlies, and were fairly 
stable in terms of direction (although a reversal did occur during leg C of 
the second SeaSoar survey).  The second survey was generally blessed with 
lighter winds than the first, whereas the westward transect was more prone to 
strong winds (and greater swell) than the Drake Passage run, with the final 
part of the westward transect/start of the first SeaSoar survey being 
particularly hard hit by winds consistently in the range 10-20 ms^-1.  The 88W 
transect was notable for the consistently westerly winds, mostly in the range 
8-15 ms^-1.

The solar radiation data, although noisy, are encouraging - where there is a 
drop in longwave radiation signalling breaks in cloud cover, total irradiance 
increases relative to PAR (clouds are less "transparent" to PAR).  From the 
longwave radiation plots it is apparent that cloud cover is intermittent, but 
generally quite high - there are very few cloud-free periods greater than 12 
hours.

The wind speed and barometric pressure data show that the operational area was 
swept by rapidly changing weather systems, where wind speed was not closely 
related to pressure.  Even when the pressure dropped to almost 950mb, the wind 
speed did not exceed 20ms^-1, while the period of 15-20ms^-1 winds persisted for 
less than 12 hours.  There were very few occasions when the wind speed dropped 
below 5ms^-1 and these also tended to be of short duration (6 - 12 hours).

Future work

Several of the data channels as they stand are fairly noisy and need to be 
rigorously screened on a datacycle-by-datacycle basis; this can probably be 
best achieved at BODC by making use of the available interactive graphical 
editing facilities.  The port dry bulb temperature in particular needs careful 
attention.  The calibration uncertainties  of the port and starboard PAR would 
ideally be resolved by a recalibration of the instruments.  Failing that, 
detailed comparisons of all the available, fully screened PAR and total 
irradiance channels will be necessary to determine and correct the calibration 
offsets.  Given a relatively cloud-free day, it should be possible to predict 
the ratio of PAR to total irradiance, given the solar constant and knowledge of 
the appropriate geometries of the hemispherical and cosine collector 
instruments.


3.4  Surface temperature and salinity 
     (Anne Morrison, Bill Miller)

Instrumentation

The thermosalinograph (TSG) system consists of a Falmouth Scientific (FSI) 
Ocean Temperature Module (OTM) mounted in the hull, at the non-toxic water 
supply intake, and an Ocean Conductivity Module (OCM) mounted, along with a 
second TM, within a polythene tube, through which the non-toxic supply passes.  
The OTM contains an FSI reference grade PRT.  The hull-mounted OTM (sn 1339) is 
located in the forward hold at a depth of approximately 4-4.5m below the water 
surface, 2.2m to the starboard of the centreline: the datastream from this 
module is known as the "remote temperature".  The OCM (sn 1331) and second OTM 
(sn 1340) are located on the starboard side of the hangar; they are mounted in 
a polythene tube through which the non-toxic water supply is pumped: this OTM 
measures the "housing temperature".  A header tank is located approximately 
2.5m above (by the winch control room) to supply enough water pressure for 
adequate rate of flow, free of bubbles; the volume flow rate through the tube 
is approximately 20L.min-1, well within the manufacturer's recommendations.  
The flow through the tube is upwards, passing the OTM before the OCM.  The 
measurements of temperature and conductivity from all the modules are logged 
through a PC, emulating a level A interface, and on to the level B at 15s 
intervals.  The sampling frequency is adjustable, through the computer.  The 
temperature resolution of the FSI OTM is 0.0001C with manufacturer's quoted 
accuracy being 0.003C in the range -2 to 32C.  Its specifications state that 
is should be stable to 0.0005C per month. 

During the first few days of the cruise it was apparent that the remote 
temperature module had noise spikes in the data (0.01 to 0.015C).  On day 
321 whilst investigating this problem by swopping with the housing temperature 
module, the latter module stopped functioning properly.  For the rest of the 
cruise the original remote temperature module was used in the housing in order 
to get salinity data, relying on the met. and ADCP sensors for sea surface 
temperature.  The problems with the FSI temp modules have been reported 
separately to RVS and to the manufacturers, who are in the process of 
providing modifications.

Sea surface temperature

Near-surface temperature measurements were obtained from four different sensors: 
(i) SeaSoar; (ii) met. package sensor; (iii) ADCP sensor; (iv) thermosalinograph 
(FSI) sensor.

SeaSoar temperature sensor.

This is believed to be the most accurately calibrated and most stable of the 
temperature sensors; its resolution is 0.0005C and is accurate to 0.001C.  
In the following sections, the results from the other sensors are compared 
with those from SeaSoar to assess their stability and accuracy.  The SeaSoar 
was deployed in the Drake Passage and in the two ice edge surveys.  Data from 
all of these surveys are used in this assessment of the quality and 
calibration of temperature measurements from each instrument.

Met. package sensor.

This is a hull-mounted sea temperature module consisting of an RS components 
platinum resistance thermometer (PRT) mounted on the port side (around 4m from 
the centreline) of the hull in the forward hold at approximately 3m below the 
sea surface.  30s averages of 1s data are transferred to the level B system.  
After extraction to the level C computer, the sensor output is corrected 
according to Table 3.1; it should be noted that the calibration was valid only 
down to 5C, while the temperatures in our survey regions were within the 
range -2 to 8C.

SeaSoar temperatures at 3m depth were extracted and combined with the 
meteorological measurement of sea temperature.  The correlation between the 
two temperatures is quite good at low temperatures, but poorer at the higher 
temperatures in our range of measurements.  The met-measured sea temperature 
is steadily higher than the SeaSoar temperatures by about 0.15C, with 
standard deviation of 0.175C, which can be applied as a shift to correct the 
met. measurement.  The Rhopoint module for this sensor limits its resolution 
to 0.1C, which makes it of very limited value for sea surface temperature 
measurements;  0.05C or better would be more useful.  However, it is worth 
noting that this instrument gives stable results.

ADCP temperature sensor

This is a PRT mounted in the transducer unit at the forward end of the winch 
room, about 1.2m to port of the ship's centreline, at approximately 5m below 
the sea surface.  Temperature is sampled every ping, approximately once per 
second, then averaged over each 2 minute ensemble and logged through a PC 
which emulates a level A interface unit.  The 2 minute measurement is then 
transferred to the level B system.  No laboratory calculated calibration 
coefficients are available for this device.  The temperature resolution of the 
PRT is 0.012C, and the manufacturer's specification accuracy is 0.2C within 
the range -5 to 45C.

SeaSoar data between 4.5 and 5.5m below the sea surface were extracted from 
the survey datasets, and merged with the ADCP temperature.  Over the Burdwood 
Bank region, massive spikes occur, as great as 0.7C in magnitude; these are 
believed to occur because there is much structure in the near-surface 
temperature, and if the SeaSoar depth and ADCP measurement depth are different 
by even a small amount, a large temperature difference will be recorded.  
During the first of the two ice edge surveys, a shift in differences from day 
330 to day 332 cannot be explained.  Excluding the large differences at the 
higher temperatures, a temperature dependent relationship is evident between 
the SeaSoar/ADCP temperature difference and ADCP temperature.  A linear 
regression gives

                      delta-T = (-0.014 * T(ADCP)) + 0.011

After applying the correction delta-T to ADCP temperature, the ADCP 
temperature to SeaSoar temperature difference has a standard deviation of 
0.031C.

Thermosalinograph

During the transect of the Drake Passage, both temperature sensors were 
available on the TSG.  Comparisons of its temperature measurements with those 
of SeaSoar were made with extracts from SeaSoar data at depths between 3.5 and 
4.5m, because the TSG is taking the temperature of water at around 4m.  In 
regions where the SeaSoar was towed at shallow depths (less than 200m of 
cable), for example, over Burdwood Bank, the agreement is poor because there 
was a great deal of structure in the near-surface temperature, which means 
that even a small error in matching SeaSoar depth to thermosalinograph intake 
depth, leads to large differences in temperature.  In the region from 55 to 
58S, the agreement between the remote temperature and SeaSoar temperature is 
excellent.  Within this region, the offset has a mean value of -0.008C and 
standard deviation of 0.008C, which can be applied as a correction to the 
remote temperature.  At 58S the SeaSoar was again towed at shallow depth (fog 
and the risk of icebergs slowed the ship down) causing large differences 
between the remote and SeaSoar temperatures.  On returning to deeper towing 
(59S).  The same offset as before is seen in remote temperature, but the data 
is much spikier, having a standard deviation of 0.035C, which was why the 
remote module was removed for investigation.  During the two ice edge surveys, 
only housing temperature was available.  This does not show a steady 
difference from SeaSoar temperatures, not surprisingly, because heating of the 
non-toxic supply on its way to the hangar unit may vary.  Variations in the 
mean of the difference between the housing temperature and SeaSoar temperature 
are of the order of 0.1C.

Salinity

The output from the conductivity and housing temperature sensors of the TSG 
were processed using PStar program sal83 to give salinity measurements, taking 
the pressure to be 0db at all times.  Water bottle samples were taken from the 
non-toxic system at hourly intervals throughout the duration of the cruise, 
and their salinities determined using a salinometer (section 3.11).  These 
were combined and compared with the TSG salinity data.  This showed that the 
TSG salinities differed from the bottle salinities by a varying amount up to 
0.1psu It was decided that the best way to calibrate the TSG was to first of 
all ignore any bottle salinities which differed greatly from the TSG (probably 
due to errors in recording the time of drawing the sample).  Then, the bottle 
samples being sparser in time than the TSG data, the differences were 
interpolated in time to assign a correction factor to each TSG value.  This 
correction was then added to the TSG data.  The mean difference was -0.091psu, 
with standard deviation of 0.019.  Therefore, corrected salinities should be 
correct to within about 0.02psu.


3.5  Starboard winch 
     (Colin Day, Tony Poole)

A total of 39 CTD operations were required during this cruise.  The depths of 
deployment ranged between 150m and 5200m.  Considering that this was the first 
fully operational scientific cruise since the Discovery conversion, the 
operation of both the winch and starboard gantry systems proved to be reliable 
and effective.  Several problems were encountered on the first deep cast 
(station 12212), as was only to be expected on the first operational use of 
the new Discovery winch to full depth.  A slack turn on the cable drum at the 
start of the haul caused cable lay problems which were not spotted until 2500m 
of wire had been hauled in.  Several hours were lost while the cable was paid 
out and the drum relaid.  This teething problem was resolved thereafter by 
careful setting of the back-tension on the inboard accumulator, and all 
subsequent CTD operations were successfully completed.  The winch/gantry 
system demonstrated the potential to significantly reduce the deployment and 
recovery time, the limiting factors in this case being the operational speed 
of the CTD package through the water.  The operation of the starboard gantry 
provided the additional benefit of allowing CTD operations to proceed safely 
and unhindered in sea conditions which would have made deployment hazardous on 
other NERC ships.

3.6  CTD operations 
     (Bill Miller, Phil Taylor, Raymond Pollard)

Instrumentation

The main sensor package included an EG&G Neil Brown MK IIIb CTD (sn 01-1195) 
fitted with a new oxygen sensor (sn 2 - 6-20).  Also fitted were a Chelsea 
Instruments Aquatracka MK II fluorometer (sn SA226), a Sea Tech 25D 
transmissometer (sn 79D), and two PAR sensors measuring upwelling light (sn 8) 
and downwelling light (sn 10).  In addition a General Oceanics 1015 rosette 
sampler (sn RVS-02) was used for bottle sampling; this comprised 12 x 10L 
Niskin bottles fitted with stainless springs.  Finally 2 x SIS 4002 (sn 220 
and 238) deep sea reversing thermometers were fitted to one of the bottles.

Some initial problems were encountered with connectors: the sea cable tail and 
its Y-cable mating connector gave intermittent operation as did the rosette 
bulkhead connector and its mating connector.  The sea cable connection was 
bypassed (spliced) and the rosette bulkhead connector replaced.  The 
transmissometer mating connector had to be 'adjusted' after one noisy cast.

During the first deep cast (station 12212) the rosette failed (a manufacturing 
fault on the pylon top plate assembly), this was replaced by the spare rosette 
unit (sn RVS-04) which, apart from not confirming, worked fine.  Bottle depths 
were confirmed from salinity measurements.  The original unit functioned 
properly in the lab after the damaged top plate had been replaced with a new 
assembly and the unit refilled with oil.

CTD casts (see Appendix B for details) can be divided into three kinds:

(i)   shallow (150-300m) casts to determine near-surface productivity and light 
      levels for productivity casts using 30L GoFlos that immediately followed
(ii)  intermediate depth casts to 1000m to determine the upper ocean water 
      masses, often used also for light and productivity information
(iii) full depth casts to determine water masses and geostrophic transport 
      through the whole water column

After an initial trial cast (station 12198) shortly after sailing from Port 
Stanley, the next five casts (12200-04) were shallow productivity casts usually 
in the morning.  One of these (12203) was occupied during a rendezvous with the 
James Clark Ross, who made a CTD cast simultaneously.  Duplicate salinity 
samples were drawn and exchanged between ships for calibration and 
intercalibration (see section 3.7 for results).  Five casts to 1000m (12205-09) 
were then made along 65S at 20 mile intervals.

On reaching the working longitude of 85W, the weather was too poor to launch 
SeaSoar, so the CTD section was continued down 85W for two more casts (12210-
11) at 20 mile intervals.  The SeaSoar was then deployed for a 4-day survey, 
following which 12 more CTD stations (12212-23) were occupied along 85W.  These 
casts were designed to span the west to east front which had been repeatedly 
crossed during the SeaSoar survey.  The casts were to 1000m at 5 mile intervals, 
to resolve deeper structure than the 400m SeaSoar could observe.  The northern- 
and southern-most casts (12212 and 12223) were to full-depth to allow the 
geostrophic transport to be calculated.  As noted above, the multi-sampler 
failed on the first deep station (12212): on analysis of the bottle samples it 
was realised that all had been tripped at the same depth.  Indeed, the multi-
sampler could not be cocked at the start of 12213, so the cast was done without 
bottle samples to obtain light levels so that productivity work could proceed, 
and the station was reworked with multi-sampler later as 12214.  Similarly, the 
top 1000m of the first deep cast (12212) were reworked as station 12218 once the 
multi-sampler had been replaced.  At the beginning of 12213 it was also found 
that the conductivity cell had frozen, and the cast was restarted once it had 
defrosted.

After 12223, the SeaSoar was deployed for the second ice-edge survey, which 
terminated prematurely with the loss of the SeaSoar.  Stations 12224-6 were 
occupied for productivity during the subsequent UOR survey.  Finally, stations 
12227-37 were worked along 88W from 69S to 5130'S.  Deep stations were 
occupied every 2.5 of latitude, occasionally preceded by a shallow cast for 
productivity work.

3.7  CTD data processing and calibration 
     (Raymond Pollard, John Allen, Anne Morrison, Jane Read)

Processing

CTD data were logged in the usual way through the RVS Level A/B/C computer 
system, then transferred into the IOSDL PStar system on another workstation.  
Each cast was split into down and up casts, the former being fully processed, 
the latter used only to obtain calibration values.  In processing the SeaSoar 
data (section 3.9), it was realised that the Level A software, which had been 
rewritten, no longer obeyed the correct algorithm for matching the time 
constant of the temperature sensor to that of the conductivity sensor in order 
to calculate salinity without bias and with minimal spiking.  This problem was 
present in the CTD Level A also, but to a lesser extent because of its slower 
profiling rate, particularly during the first 6 casts, all of which were 
shallow.  Revised Level A software was received and tested in time for use in 
all the 1000m casts, starting with station 12205.  Final calibrations were 
derived and applied as described below.

Calibration

Initial calibration was done by PStar program 'ctdcal'.  Further calibrations 
were applied later, as described below.

Pressure

The calibration used was:

                       P = -7.1 + 0.1 * 0.9987653 * P(raw)

The constant offset supplied by RVS was -5.13287, but this was changed to -7.1 
after noting the value of pressure when the CTD entered the water.

Temperature

The calibration supplied by RVS was used throughout:

                  T = 0.0044057 + T(raw) * 0.0005 * 0.9999902

The reversing thermometers (section 3.6) were used to record temperature on 27 
casts.  After correction, T220 read higher than T238 by between 0 and 2mK, 
giving a difference of (0.85  0.82mK) (mean  standard deviation).  Thus the 
thermometers agreed to better than 0.001C in the mean.  The comparable CTD 
temperature was between 0 and 8mK higher than the mean of the two 
thermometers, giving a correction to the CTD temperatures (if the SIS 
thermometers are absolutely correct) of (-4.48  1.74mK).  However, experience 
has shown that the SIS reversing thermometer calibrations are no more reliable 
than the CTD calibration, so the CTD temperatures have not been adjusted.  We 
conclude that the CTD calibration was stable throughout the cruise, and that 
temperatures are absolutely correct to within 0.004 - 0.005C, and are 
possibly high by that amount.

Salinity

CTD salinity was first calculated using the default conductivity ratio 

                   conductivity = 0.001 * raw conductivity.

To match the differing time constant of the temperature and conductivity 
cells, the temperature response was speeded up by 0.5s.  During the first part 
of the cruise salinity samples were drawn from all 12 Niskin bottles, fired at 
various depths.  The samples were analysed by three analysts (section 3.11).  
Once it was established that the conductivity cell was stable, the number of 
samples drawn was reduced, first to 6 and later to 4 samples per cast.  The 
exceptions were the two full depth casts (12212 and 12223) made on either side 
of a front in the Bellingshausen Sea.  The first of these had problems and the 
bottles were all tripped together at an unknown depth.  Samples from the 
second cast gave greater confidence to the calibration.  The difference 
between discrete bottle samples and CTD values was found to vary with salinity 
so a linear correction was calculated.  Where the difference between bottle 
and CTD values was less than 0.025 and greater than 0.045 the values were 
ignored.  The calibration coefficients calculated were:

                       S(true) = S(CTD) * 0.99347 + 0.257

with a standard deviation of  0.002

Oxygen

Oxygen is calculated for the Beckmann oxygen sensor attached to the CTD by

           [O2]/mL.L-1 = rho * C(oxy) * exp ( -alpha * T + beta * P )
where
                 T = a * T(CTD) + b * T(oxy)	(with a + b = 1)

and C(oxy) and T(oxy) are the oxygen sensor current and temperature readings 
respectively and P is the CTD pressure.  The CTD oxygen sensor was calibrated 
from 126 discrete bottle samples drawn from various depths over 13 different 
casts up to 12223.  The data were combined to produce the coefficients rho 
=1.276292, alpha =-0.02666 and beta =0.0001494 with a standard deviation of 
0.12 mL.L-1.  The residuals from this least squares regression varied from cast 
to cast, averaging from -0.14 to +0.15, and over depth, with the greatest 
residuals in the thermocline.  Oxygen temperature was weighted to CTD 
temperature in the ratio b:a = 0.25:0.75 for the calculations.

Considerable problems were experienced with the oxygen titration unit (section 
3.21), so the bottle values may require further reworking and culling to 
improve the calibration.  However, it was noted that the surface values of 
percent oxygen saturation for stations 12209-23 using the above calibrations 
were well correlated with the patchy blooms that were present, ranging from 
104% in the strong bloom to 98% where the bloom was virtually absent.  This 
suggests that the oxygens are correct within about 2%.

Chlorophyll

This calibration is reported separately (section 3.24)

Transmissometer

The A/D converter in the transmissometer scales 4096 counts to 0 to 10V.  The 
instrument converts 100% transmittance to 5V.  The basic conversion equation 
is thus

                         %T = 0.002442 * 20.0 * count.

Ageing of the light source may change this calibration, so the highest deck 
reading after thorough cleaning of the glass (4.763) was used to correct 
transmittance by comparison with the manufacturer's value (4.744) to give

                    corrected %T = (4.744 / 4.763) * raw %T

Potential transmittance (potran) is then calculated by correcting for the 
index of refraction and compressibility of seawater using in situ values of 
pressure, temperature and salinity.  The RVS transmissometer has a 0.25m path 
length.  Thus, the beam attenuance per metre, independent of instrument, is 
given by the attenuance divided by path length:

               beam attenuance (m^-1) = - ln (potran/100) / 0.25

Light

The CTD was fitted with upward and downward looking PAR irradiance sensors, 
with calibrations supplied as follows:

               dwirr (downward) = 6.6470 - 0.001 * 12.353 * count
               uwirr (upward)   = 6.5746 - 0.001 * 12.427 * count.

CTD Intercalibration between Discovery and James Clark Ross

During the first rendezvous between the Discovery and the James Clark Ross an 
intercalibration of CTD and discrete samples was undertaken.  This was 
Discovery station 12203 (Appendix B) and James Clark Ross station 27.  Two 
sets of samples were drawn on board each ship and duplicates exchanged almost 
immediately.  The results are shown in Tables 3.2 and 3.3.  The duplicate 
samples from each cast are within 0.004 of each other with a mean difference 
of better than 0.002 (0.0017) and standard deviation of 0.0008.  The results 
from the Discovery analysis are all consistently higher than those from the 
James Clark Ross, which should be born in mind if making comparisons between 
the data from each cruise.  Comparison between the samples from the different 
casts reflects the variability in the upper ocean structure, and the only real 
comparison that can be made is within the surface layer.  This was well mixed 
down to 50m and it can be seen that the samples from the two casts were 
generally within 0.001psu.  CTD data in the tables is uncalibrated and so 
cannot usefully be compared.

Table 3.2  Salinity intercalibration on Discovery samples (Station 12203)

      | CTD readings      | Analysis (Disco)  | Analysis (JCR)   | Difference
------|-------------------|-------------------|------------------|----------
Wire  | Temp/C  Salinity | Bottle   Salinity | Bottle  Salinity |   /psu
out/m |           /psu    |  no       /psu    |  no      /psu    |
------|-------------------|-------------------|------------------|----------
250   | 1.492    34.415   |  B31      34.446  | S1      34.445   |   0.001
200   | 0.860    34.280   |  B32      34.314  | S3      34.313   |   0.001
150   | 0.710    34.177   |  B35      34.187  | S5      34.186   |   0.001
100   |-1.382    33.941   |  B37      33.976  | S6      33.976   |   0.000
 80   |-1.372    33.782   |  B39      33.819  | S8      33.818   |   0.001
 60   |-1.173    33.765   |  B41      33.801  | S10     33.800   |   0.001
 40   |-1.099    33.763   |  B43      33.800  | S15     33.799   |   0.001
 30   |-1.084    33.763   |  B45      33.799  | S13     33.798   |   0.001
 20   |-1.064    33.761   |  B47      33.800  | S17     33.799   |   0.001
 10   |-1.045    33.761   |  B49      33.800  | S18     33.798   |   0.002
  2   |-1.038    33.761   |  B51      33.799  | S19     33.797   |   0.002
      |                   |  B52      33.799  | S22     33.797   |   0.002
  2   |-1.038    33.761   |  B53      33.799  | S23     33.797   |   0.002


Table 3.3  Salinity intercalibration on James Clark Ross samples (Station 27)

      | CTD readings      | Analysis (Disco)  | Analysis (JCR)   | Difference
------|-------------------|-------------------|------------------|----------
Wire  | Temp/C  Salinity | Bottle   Salinity | Bottle  Salinity |   /psu
out/m |           /psu    |  no       /psu    |  no      /psu    |
------|-------------------|-------------------|------------------|----------
250   | 1.391    34.345   |  14       34.430  |   3     34.432   |   0.002
200   | 0.903    34.240   |  15       34.328  |   4     34.324   |   0.004
150   | 0.294    34.095   |  16       34.171  |   5     34.173   |   0.002
100   |-1.360    33.873   |  17       33.966  |   6     33.967   |   0.001
 80   |-1.384    33.729   |  18       33.818  |   7     33.820   |   0.002
 60   |-1.102    33.710   |  19       33.798  |   8     33.801   |   0.003
 40   |-1.108    33.711   |  20       33.798  |   9     33.800   |   0.002
 30   |-1.052    33.710   |  21       33.798  |  10     33.800   |   0.002
 20   |-1.052    33.710   |  22       33.798  |  11     33.800   |   0.002
 10   |-1.042    33.709   |  23       33.798  |  12     33.799   |   0.002
  2   |-1.032    33.708   |  24       33.797  |  13     33.799   |   0.002


3.8  SeaSoar operations 
     (Bill Miller, Phil Taylor, Raymond Pollard)

Instrumentation

The RVS SeaSoar (manufactured by Chelsea Instruments) was fitted with a 
Chelsea Instruments Aquatracka MK II fluorometer (sn SA 240) together with an 
EG&G Neil Brown MK IIIb CTD (sn 01-1181) fitted in a lightweight pressure 
case.  The CTD was also fitted with oxygen and light sensors.  The oxygen 
sensor (sn 1-8-27) was a new unit fitted prior to deployment.  An independent 
pressure sensor, used for control purposes only was also included.

The performance of SeaSoar was good maintaining a yoyo from surface to 400m 
plus at ship speeds between 7.5 and 9 knots, sometimes in heavy swells.  In 
periods of bad weather the ship's speed reduced to 5-6 knots and SeaSoar would 
not reach the surface.  Data quality was good with only the occasional fouling 
of the conductivity cell.

At 339/1034 the towing cable parted some 20m from the SeaSoar, with the 
ensuing loss of all instrumentation.  No warning of impending failure was 
given, good data being collected until 2s before.  The details of the loss 
have been reported separately.

SeaSoar surveys

The SeaSoar tows undertaken are summarised in Appendix C.  One day after 
sailing from Port Stanley, Falkland Islands, the SeaSoar was deployed on day 
317 for a trial run in the passage south of the Falklands and north of 
Burdwood Bank.  However, as it functioned perfectly, the first survey line, 
across Drake Passage (Figure 1.1*), was run without a break.  The cable length 
outboard was shortened from 600m to 200m during the transit across Burdwood 
Bank from 317/1830 to 318/0440, and was again shortened to increase the ship's 
manoeuvrability in poor visibility and in the vicinity of ice from 319/0940 to 
319/1040 and from 319/1630 to 320/0707.

The second deployment began as an exploratory run from 6540'S at 85W south 
towards the ice edge.  It was soon decided to convert the run without a break 
into the first ice edge survey (Figure 1.2*).  After a dogleg to move the 
southward track westward to 8620'W, a survey of 8 legs at 8 mile spacing (20' 
longitude) from 6630'S to 68S was planned.  The central leg down 85W is 
labelled A.  The first six legs (W, X, Y, Z, A and B) were completed without 
major incident, although there were detours to avoid icebergs and it was 
occasionally necessary to drop the speed below the optimum 8 knots in poor 
visibility.

Weather conditions were poor throughout the first ice edge survey, and on the 
7th leg (C), running south along 8420'W it was impossible to keep to the 
track (Figure 1.2*).  It was also found that a boom attached to the stern 
gantry had sheared several bolts and was in danger of falling onto the SeaSoar 
cable.  Because the SeaSoar could not safely be recovered until the boom was 
secured, and the boom could not be secured until the weather abated, the ship 
steamed slowly into the wind on a northwestward track from 332/2000 and later 
east of north until repairs could begin at 333/1100.  Although the survey was 
officially abandoned at the end of leg C, the last run across the middle of 
the survey area has been processed as 'stormleg'.

The first attempt at a second survey (335/2230 to 336/0300) was aborted because 
of forecast severe weather, and the ship remained hove to for 24 hours until 
337/0200, when SeaSoar was again successfully deployed.  Thus the second survey 
(Figure 1.3*) began nearly 4 days after the first had ended.  The intention was 
to run north-south at the same longitudes as in the first ice edge survey, but 
from 67S to 6830'S, later extended to 69S.  Also, the lines were worked 
upstream from east to west, as the first ice edge survey had repeatedly crossed 
a strong front with an eastward flowing frontal jet.  The first four legs (D, B, 
C and A, using the same labels as for the first survey) were completed without 
incident. Shortly after the end of leg A along 85W, the SeaSoar was lost.


3.9  SeaSoar data processing and calibration 
     (Raymond Pollard, John Allen, Polly Machin, Anne Morrison, Jane Read)

Processing

SeaSoar data were logged in the usual way through the RVS Level ABC computer 
system, then transferred into the IOSDL PStar system on another workstation.  
Data were split into 4-hourly sections, for the convenience of watchkeepers, 
who processed, applied initial calibrations, plotted and edited the data.  The 
major task, as always, was to examine the data for offsets caused by 
biological fouling, usually requiring an offset in salinity to recover to a 
good calibration relative to data before and after the fouling event.  While 
this procedure is time-consuming, the great merit of the Neil Brown 
conductivity cell is that, once the fouling clears, it recovers its 
calibration reliably, and calibration drift is rare.

Towards the end of the Drake Passage survey, it was realised that the Level A 
software, which had been rewritten, no longer obeyed the correct algorithm for 
matching the time constant of the temperature sensor to that of the conductivity 
sensor in order to calculate salinity without bias and with minimal spiking.  
The value allowed for pressure spiking had also been set too small, resulting in 
good data being discarded whenever the climb or dive rate exceeded 2ms^-1, which 
commonly occurs shortly after each turn. The software could only be corrected by 
RVS at Barry.  While this was being done, the entire Drake Passage survey was 
replayed from the backups of raw 8Hz data maintained by the RVS electronics 
division.  A PStar program to reduce the data from 8Hz to approximately one 
sample per second using the correct algorithm was revived, and very clean data 
with the 16 light channels correctly demultiplexed were retrieved.  The time 
base was retrieved by setting the start time, which was logged on the backup 
tape, and calculating the time between samples so that the resultant plots 
matched those originally processed through the Level A.  This revealed that the 
CTD sampling rate was 8 samples in 1.02413s.  Revised Level A software was 
received and tested in time for use in the ice edge surveys.  It was not 
perfect, particularly in regard to demultiplexing of the light channels, but the 
time constants of temperature and conductivity could be correctly matched, so 
the normal Level A route was used for the remainder of the cruise.  The 4-hourly 
sections were appended and contoured either every 12 hours (Drake Passage) or 
for each leg (ice edge surveys).  Final calibrations were derived and applied as 
described below.

Calibration

Initial calibration was done by PStar program 'ctdcal', and further 
calibrations were applied later, as described below.

Temperature

The calibration supplied by RVS was used throughout:

                   T = 0.00274 + T(raw) * 0.0005 * 0.9996194

Experience has shown that the temperature calibration is precise and very 
stable, and cannot be easily checked, as it is more accurate than any other 
sensor onboard ship.  Any shift would probably manifest itself as a severe 
offset in salinity, which was not observed.  No bias could be detected when 
our T/S plots were compared with those of other investigators who had recently 
worked in similar areas (R Peterson, Burdwood Bank, September 1992, personal 
communication; J Swift, Bellingshausen Sea along 67S, February 1992, personal 
communication).  We therefore take the SeaSoar temperatures to be absolutely 
correct to within perhaps 0.003C, with relative drift during the survey no 
more than that.

Salinity

The SeaSoar salinity was first given an approximate calibration by use of a 
conductivity ratio that gave reasonable answers for the area being surveyed.  
Bias between down and up profiles caused by the different time constants of 
the temperature and conductivity cells was minimised by choosing a time 
constant by which the temperature was speeded up (for the calculation of 
salinity only) to minimise the hysteresis between down and up T/S profiles.  
The value chosen was 0.35s, on the large side for the Neil Brown platinum 
resistance thermometer.

Relative calibration was maintained by comparing T/S profiles four at a time 
with a master T/S plot which was gradually developed for each survey area.  
Fouling of up to 0.2 or larger in salinity could occur, with frequent offsets 
of order 0.010 to 0.050.  These were not always easy to spot, but the contour 
plots proved to be sensitive indicators of offsets (maintained for a profile 
or more) of as little as 0.010.  Despite the care that was taken, there were 
occasions, in particular when the T/S relation changed rapidly across the 
front that lay across the ice edge survey area, when determination of the 
timing and magnitude of an offset proved almost impossible to correct with 
100% confidence.  Thus there may be occasions when the salinity is in error by 
as much as 0.030, but we expect the calibration to be within 0.010 more than 
99% of the time.

Absolute calibration was subsequently achieved by comparison with hourly 
samples drawn from the sea-water supply to the thermosalinograph.  The taking 
of the samples was timed to coincide with the SeaSoar surface turn, however, 
later comparisons between the time the bottle sample was drawn and the 
corrected SeaSoar time showed that the two rarely coincided.  However, the 
seasonal mixed layer was sufficiently horizontally homogeneous that 
interpolating between SeaSoar profiles gave a good comparison with the bottle 
samples, except across fronts where the surface gradients were higher.  These 
latter data were ignored in assessing the salinity correction.

The SeaSoar conductivity cell proved very stable for most of the cruise and 
straight offsets were applied for the first two deployments to bring the 
surface data within 0.010 of the salinity samples.  Across the Drake Passage 
salinity was corrected by +0.015 and for the first ice edge survey the 
correction increased to +0.024.  On the second ice edge survey the conductivity 
cell displayed a previously unobserved, aberrant behaviour, oscillating wildly 
over a 0.04 salinity range.  The cause of this is unknown but it was most 
likely some kind of fouling, possibly by krill, as it recovered eventually.  
The oscillation in salinity made it very difficult to identify offsets due to 
fouling, or to ascertain the relative calibration.  Comparison with discrete 
salinity samples showed that the deployment began with a +0.024 offset, but 
after the conductivity cell started oscillating this was masked by a larger and 
more variable offset which appeared to jump randomly between two values about 
0.030 apart.  Together with the 0.024 offset, we estimate the correction to be 
0.0440.015. About half way through the deployment the conductivity cell 
appeared to recover and relative calibration of the potential 
temperature/salinity plots left a correction of only +0.010 to be made.

Oxygen

Oxygen is calculated for the Beckmann oxygen sensor attached to the CTD by

           [O2]/mL.L^-1 = rho * Coxy * exp ( - alpha * T + beta * P )
where
                 T = a * T(CTD) + b * T(oxy)         (with a + b = 1)

where C(oxy) and T(oxy) are the oxygen current and temperature values, and P 
is the CTD pressure.  For deep CTDs, normal practice has been to choose 'a', 
the ratio between the CTD temperature (which has a short time constant) and 
the internal Beckmann unit temperature (T(oxy)) to reduce hysteresis between 
down and up casts.  The constants rho, alpha and beta are then chosen by a 
least squares fit to all available oxygen data from one or many CTD casts.  
This procedure did not significantly reduce the hysteresis when tried on the 
sensor in the SeaSoar, because (a) there was very little temperature variation 
with depth south of the Polar Front and (b) the SeaSoar cycles through the 
oxygen gradient much more rapidly than a CTD.

A new procedure was therefore adopted on this cruise. The oxygen current was 
speeded up by the formula

C(oxy) (t0) = rawC(oxy) (t0) + tau [rawC(oxy) (t1) - rawC(oxy) (t-1)]/(t1 - t-1)

where t-1, t0 and t1 were successive 1-second values, and the oxygen 
temperature was ignored (b = 0).  Several values of tau were tried until 
hysteresis was minimised.  It was found that the value of tau depended on the 
depth of profiling, and two values which improved the fit were chosen as 15s 
for the Drake Passage section, and 10s for the ice edge surveys.  The least 
squares fit to determine the remaining constants also could not be applied 
because (a) oxygen samples obviously cannot be drawn at the SeaSoar position, 
(b) the least squares fit equations tend to be ill-conditioned.  This is so 
because both pressure and temperature tend to vary monotonically with depth, 
so the fit is sensitive to errors in bottle values caused by sampling problems 
or samples not carefully matched to the SeaSoar values in space and time.

Oxygen samples available consisted of (a) those drawn from CTD casts along 
85W between the first and second ice edge surveys and (b) two-hourly surface 
samples during SeaSoar runs.  The latter were used to choose rho with a typical 
value of alpha (-0.036), but exhibited rather wide scatter with some obviously 
bad sample values possibly caused by difficulties in sampling off the non-
toxic supply.  However, it had been noted that the percent saturation of the 
calibrated CTD oxygens correlated well with patches of high and low 
productivity, ranging from 1.04 (104%) in the major bloom, through 1.00 in 
smaller blooms to 0.98 where the bloom was absent.  By choosing rho = 1.472, 
alpha = - 0.036, these percent saturations were closely matched by the SeaSoar 
oxygens in and out of the bloom areas.  The pressure coefficient beta does not 
affect the fit at zero pressure.

To approximate the vertical profiles of oxygen, we made use of the samples 
drawn on CTD casts at 300m.  T/S profiles from the CTD casts along 85W were 
compared with SeaSoar T/S profiles from the 4-km gridded file (averaged over 
down and up casts) for the 85W run during the first ice edge survey.  As this 
survey crossed and recrossed the front, where the T/S relation changed 
rapidly, profiles could be matched within a few km of similar features.  
Applying rho and alpha given above, with a nominal value of beta = 0.00014, it 
was found that the SeaSoar values were too low.  By adjusting the value of beta 
to 0.0003965, the 300m values were brought in line with the sample values to 
within 0.02 + 0.10 mL.L^-1, where 0.10 is the standard deviation over 10 
samples.  This is equivalent to a near-linear stretching of the oxygen values 
with pressure, because

                         exp (beta * P) = 1 + beta * P
for beta * P << 1.

In the thermocline, accurate calibration is impossible because of the oxygen 
sensor hysteresis.  In summary, we expect that our calibrated oxygen values 
will be correct within about 2%, and are more likely to be low than high (if 
the 98% surface saturation values should be increased to 100%, say).

Chlorophyll

See section 3.24

Light

See section 3.12


3.10  Undulating Oceanographic Recorder 
      (David Turner, Ian Bellan)

After the loss of the SeaSoar, the second ice edge survey was completed using 
the UOR loaned by the James Clark Ross.  Five runs were completed, four of 
them of about 8 hours duration, which was the maximum battery life at this 
water temperature; the UOR tows are summarised in Appendix D.  The second tow, 
along leg Y, was terminated after less than an hour, since examination of the 
data from the first tow revealed a problem with the temperature sensor.  The 
UOR was recovered, the sensor replaced, and the tow restarted.  On completion, 
a fault was found with the UOR conductivity sensor, although this did not 
affect the other sensors on this tow.  Further attention by Ian Bellan solved 
this problem, and two further tows were carried out along legs X and YZ (86W 
and 8530'W), these being chosen to cover the western half of the survey as 
effectively as possible.  After transfer to PStar, the UOR data was processed 
in a similar way to that from SeaSoar.


3.11  Discrete salinity measurements 
      (Anne Morrison, Jane Read, John Allen)

The RVS Autosal salinometer (Model 8400A, sn 52395) was used routinely 
throughout the cruise to measure the salinity of samples drawn hourly from the 
thermosalinograph and from most CTD casts.  The salinometer was operated at 
24C in a controlled environment of 22C, and proved to be fairly stable.  
However, on day 345 it was extremely unstable, apparently a result of power 
surges in the 'clean' electricity supply, which cut out completely at 0840.  
Subsequently an uninterruptable power supply unit was used to ensure steady 
power to the Autosal, and very stable results were obtained.  We suspect that 
some air was being drawn into the cell, leading to slow filling and the odd 
bubble, but because the coils were completely covered, the results were not 
adversely affected.  Standard sea water ampoules (batch P120) were used to 
standardise at the beginning and end of each session, which generally 
consisted of 24 samples: this standardising was less frequent than is usually 
recommended, because we had a limited supply of standard sea water.  We used 
the "Salinity Master" Excel spreadsheet to generate salinities from hand-
recorded conductivity ratios.

We also experimented with the SIS Softsal (Version 1) data management package.  
This system acquires the conductivity ratio data directly from the salinometer 
to a personal computer, where they are converted to salinities by an 
undocumented method.  The user can select the allowed standard deviation and 
standardise using standard sea water and substandards.  The results are stored 
to datafiles with headers containing information on the status of the Autosal.  
In automatic mode the system requires that three consecutive measurements are 
within the selected standard deviation, with each single measurement being the 
average of ten consecutive readings which must be within another selected 
standard deviation.  We found that this can sometimes mean that many flushings 
are required if the standard deviation allowed is small, and the occasional 
rogue value is occurring.  This can be overcome by operating in manual mode, 
where the user selects which measurements to accept.  The package also 
includes a post-processing facility which uses the control data (standard sea 
water, sub-standards) to remove zero order drift from the results.  We found a 
few bugs in operating the package, but nothing insurmountable.  The salinities 
achieved from Softsal were compared with the results from "Salinity Master" 
for the same data set, and found to be very close, within 0.0005psu.

3.12  Optics 
      (Alison Weeks, Gerald Moore)

Seasoar

Optical sensors were installed on Seasoar to provide data for the 
interpretation of images from SeaWiFS (Sea viewing Wide Field of view Sensor), 
due to be launched in mid to late 1993; the spectral reflectance from surface 
Seasoar data will be used for early development of Southern Ocean bio-optical 
algorithms.  In addition, the relationship between the optical properties of 
the ocean and the mixed layer depth (from spectral diffuse attenuation 
coefficients, K(d), K(u), derived from Seasoar data) will be investigated.

The data from the upwelling and downwelling irradiance sensors on Seasoar have 
been processed for the first time on this cruise.  The sensors measured 
upwelling and downwelling radiance at seven wavelengths centred on the SeaWiFS 
bands (410, 443, 490, 520, 550, 635 and 665nm); additionally downwelling PAR 
and upwelling chlorophyll natural fluorescence (683nm) were recorded.  The 
data from the seven SeaWiFs bands were used to derive reflectance, K(d) & 
K(u), and the data from PAR to derive K(d).  The data was processed separately 
from the other Seasoar sensors because it was necessary to pair adjacent up 
and down casts to eliminate the attitude change between up and down casts.

The data obtained from Seasoar were processed to obtain optical profiles from 
the upper 180m of the water column, the exact extent of the profile depending 
on light levels and water constituents.  The Drake passage transect produced 
287 profiles, and the first and second surveys produced 441 and 255 
respectively.  Of the sensors, good data were obtained from all but the 665nm 
sensors and the 683nm upwelling sensor; these sensors proved insufficiently 
sensitive.  A number of problems were encountered in processing the data: 
first, that the sign bit was not passed to the Level-A, causing light levels 
at the surface to be apparently lower than those at depth; and second, the 
optical data were multiplexed whereas the CTD data were not.  The CTD data 
comes from Seasoar in 1 second averages, but the light data comes in discrete 
samples within that second.  Furthermore, the upwelling and downwelling lights 
are sampled out of phase.  The maximum rate of climb or descent from Seasoar 
can be up to 4ms^-1.  The result of this is that reflectance is changed by 10% 
in waters with high K(d) when Seasoar has a high vertical velocity, and that 
K(d)/K(u) is changed when SeaSoar decelerates or accelerates.  With some 
software effort these problems were resolved, and the entire SeaSoar data set 
was processed to produce reflectance, K(d) and K(u).

Preliminary analysis has begun on the first SeaSoar survey.  Near surface 
reflectance has been related to chlorophyll fluorescence, which had a simple 
linear quench function applied (section 3.24), removing some of the variance 
due to quenching, but not all.  The relationship between reflectance and 
chlorophyll fluorescence (Table 3.4) shows the best agreement at 443nm, or the 
peak chlorophyll a in vivo absorbance, indicating a species assemblage with 
low accessory pigment levels for the levels of chlorophyll encountered.

Table 3.4  Chlorophyll-reflectance relationships from SeaSoar data-*a 

                        Wavelength/nm  r^2   alpha   beta
                        ----------------------------------
                        410      0.48  1.31   -28.64
                        443      0.66  1.42   -19.16
                        490      0.53  1.31   -31.83
                        520      0.40  1.51   -34.41
                        550      0.01    -        -
                        632      0.04    -        -
                        665        -     -        - 
      *a- parameters quoted for the regression ln[chl] = alpha + beta * R


A comparison of surface K(d) from the SeaSoar (Figure 3.2*) with chlorophyll 
from the calibrated underway fluorometer (Turner Designs) from the first 
survey (Figure 2.1*) shows that the parameters are highly correlated.  The 
average optical depth north of the front is 14m, whereas to the south it is 
only 6m.


Figure 3.2*  Diffuse attenuation, first survey


Chlorophyll and reflectance data were compared using the NASA CZCS (Coastal 
Zone Color Scanner) algorithm i.e. chl = A.(Ri/Rj)^B, where the wavelengths i, 
j are usually 443nm and 550nm for oceanic waters.  The initial fit for the 
data with A = 2.63 and B = -0.88 (r^2 = 0.42) was promising; however the 
exponent B was higher than that expected value of -1.27 (Figure 3.3*).  A 
comparison against solar altitude showed little improvement.  When the data 
was restricted to those occasions where there was vector illumination (i.e. 
clear skies), as determined from long wave radiation, and the solar altitude 
was greater than 70 the resulting exponent was -1.18 (A = 3.11, r^2 = 0.79) 
within error limits of the standard result (Figure 3.3*).  The scale factor is 
much higher than the expected 0.51, indicating a possible underestimate of 
chlorophyll, and differences in phytoplankton biology.  The use of the 
standard NASA retrieval algorithm would result in an underestimate of 
chlorophyll in remote sensed images.


Figure 3.3*  Chlorophyll-reflectance ratio relationships


Work is in progress relating the above water light field to the data, by 
comparing data from the hemispherical and cosine par detectors.  Fully 
calibrated chlorophyll which will be needed to complete the algorithms is in 
preparation (section 3.24).  The results to date indicate that the SeaSoar 
light sensors have provided a valuable data set to aid interpretation of 
Antarctic remote sensing.

Lightfish

Lightfish was deployed to provide sub-surface irradiance measurements at 6 
wavelengths for the development of multi-spectral reflectance algorithms for 
phytoplankton biomass for SeaWiFS.  It was also used to compare the horizontal 
scales of variation of ocean colour, phytoplankton biomass and physical 
structure.  By conducting a spatial survey in an area of 200km^2 where the 
phytoplankton biomass is heterogeneous, it will be possible to calculate the 
loss of accuracy of SeaWiFS by the averaging effect of the pixel dimensions 
(1.1km local coverage or 4km global coverage).

The data from the upwelling and downwelling irradiance sensors have been 
logged for the first time via a Level A, along with the other underway 
parameters.  The parameters derived from these sensors are reflectance at 6 
wavelengths (410, 443, 490, 520, 550, and 670 nm).  Lightfish was deployed for 
21 days during the cruise (Appendix E).

The NASA CZCS algorithm for phytoplankton retrieval was applied to 
uncalibrated fluorescence (Turner) data during the cruise, and showed 
encouraging results, with typical r^2 values from .5 to.75.  Surface plots of 
reflectance ratios log10(R443/R550) are shown in Figure 3.4, and clearly 
correlate with the areas of high and low chlorophyll (high chl = low 
log10(R443/R550)).  We anticipate some exciting work comparing the reflectance 
ratio values from Lightfish with other underway parameters sampled on the 
cruise, such as pCO2, TCO2, and HPLC pigment measurements.

A number of problems were encountered with deployment and data acquisition of 
Lightfish.  Firstly, as the portable winch which was situated on the starboard 
side of the aft deck had no slip rings (despite timely requests to RVS) it was 
necessary to disconnect Lightfish on every deployment, and every time the 
depth was adjusted.  When Lightfish was disconnected, the Level A had to be 
reset, sometimes repeatedly.  A slip ring winch would have resulted in an 
uninterrupted data flow between deployments.  A second problem was that the 
ships wiring in the four junction boxes used needed checking and tightening.  
Other problems with Lightfish concerned the design of the casing around the 
optical filters, and with the sensitivity of the sensors: these will be 
addressed at Southampton.

It would be useful to be able to check the linearity and calibration of both 
the SeaSoar and the Lightfish sensors on board, by using a calibrated light 
source and a series of ND filters that could be mounted in a Gershun tube.


Figure 3.4*  Reflectance ratio, first survey


Diffuse Attenuation (K(d)) from CTD casts

K(d)(PAR) was calculated on CTD profiles to enable optical depths of 100, 50, 
25,10,1 and 0.1% to be determined for obtaining water for primary production 
experiments.  The resulting values are shown in Table 3.5 (for full details of 
CTD stations see Appendix B).


Table 3.5  Diffuse attenuation values from CTD casts

                  Station-*a   K(d)   r(K(d))    K(u)   r (K(u))
                  ----------------------------------------------
                  12198p    -0.08857  -0.636     *b       *b
                  12200p    -0.07393  -0.975  -0.08039  -0.885
                  12201p    -0.04908  -0.995  -0.05612  -0.987
                  12202p    -0.04646  -0.992  -0.04835  -0.977
                  12203p    -0.12164  -0.995  -0.13197  -0.993
                  12204p    -0.07612  -0.993  -0.08583  -0.994
                  12205     -0.06934  -0.984  -0.06648  -0.845
                  12206     -0.09345  -0.424     *b       *b
                  12207p    -0.06709  -0.990  -0.07531  -0.982
                  12208     -0.07451  -0.980  -0.09006  -0.983
                  12209     -0.04122  -0.994  -0.04175  -0.971
                  12210     -0.08549  -0.751     *b       *b
                  12211p    -0.09685  -0.993  -0.10855  -0.995
                  12213p    -0.05118  -0.983  -0.05422  -0.917
                  12214     -0.04607  -0.975  -0.05802  -0.961
                  12215     -0.04016  -0.993  -0.05438  -0.991
                  12216     -0.03985  -0.991  -0.06344  -0.828
                  12217     -0.05448  -0.993  -0.06336  -0.955
                  12218     -0.06541  -0.951  -0.06022  -0.612
                  12219     -0.06105  -0.923  -0.07438  -0.978
                  12220p    -0.11298  -0.968  -0.09884  -0.997
                  12221     -0.15623  -0.988  -0.19077  -0.978
                  12222     -0.15401  -0.966  -0.16003  -0.984
                  12224p    -0.15625  -0.99   -0.17855  -0.990
                  12225p    -0.12049  -0.995  -0.16295  -0.897
                  12226p    -0.1192   -0.998  -0.12471  -0.997
                  12227p    -0.04602  -0.998  -0.05386  -0.997
                  12230p    -0.04779  -0.999  -0.05045  -0.999
                  12234p    -0.06275  -0.999  -0.07290  -0.994
                  
            *a productivity stations are indicated by the suffix p
            *b insufficient upwelling light for determination of K(u)


3.13  Expendable bathythermographs 
      (John Allen, Bill Miller)

The Bathy Systems Inc. XBT program version 1.1 was used on a S.E.S.U.  
(Hydrographic Department) supplied deck unit to record XBT launches.  The XBT 
probes were launched from a Sippican Corporation hand launcher belonging to 
RVS.  The preferred launch position was the stern quarter more in the lee of 
the wind.  The connection point located in the Bosun's store was used since it 
provided the best shelter whilst preparing the launcher.  The XBT program on 
the deck unit converted the voltage drop across the probe into resistance and 
thence to temperature using the manufacturers algorithms.  In addition the XBT 
program identified critical turning points on the temperature v depth trace 
and transmitted the information in batches to the GOES satellite (part of the 
GTS network).  The XBT data was transferred to the RVS level A by a 'walknet' 
3.5" disk as no hard wiring existed between the XBT deck unit and the RVS 
firmware.  The transferred data took the form of a string of ASCII temperature 
values recorded at 10Hz.  This system is both cumbersome and time consuming, 
and it is suggested that a level A be programmed to receive ASCII coded hex 
data direct from the XBT controller in real time.

The probe depth (D) was calculated from a 2nd order time of flight (t) 
polynomial supplied by the manufacturers of the probes, Sippican Corporation.

                 For T7  (760m) probes  D = 6.472t - 0.00216t^2
                 For T5 (1800m) probes  D = 6.828t - 0.00182t^2

Several authors have suggested that these polynomials are poor and have put 
forward alternatives.  Seven XBT's on the northward transect along 88W were 
coincident with CTD stations, and further work will include the assessment of 
alternative fall rate equations.  However, initial comparison of CTD and XBT 
temperature traces did not suggest any problem existed with the manufacturer's 
equation.

During the bulk of the cruise five T7 and three of the deeper T5 XBT probes 
were launched at various points of interest.  These were:

xp198012 - in 415m of water off the Falkland Islands shelf, partly as an 
           equipment test and partly to obtain a temperature profile during the 
           ADCP calibration tracks.
xp198016,  017, 019 - deep XBT T5 drops during the SeaSoar transect across 
           Drake Passage to examine the deeper temperature structure.  Copper 
           wire found around the SeaSoar cable discouraged further XBT launching 
           during SeaSoar tows.
xp198020,  021, 022, 023 - XBT's dropped alternately with CTD casts on the east 
         - west transect along 65 S.

The other 59 XBT probes, all T7s, were launched on the transect up 88 W at the 
end of the cruise.  Details of all successful XBT launches are given in 
Appendix H.  Gaps in the sequential numbering indicate probe failures for one 
or more of the following reasons:

(i)   Insulation leakage - During periods of poor weather, when launching from 
      the rear of the afterdeck was felt to be unsafe, the probes were launched 
      from just outside the Bosun's workshop.  Under such circumstances the ship 
      was generally turned into the wind when launching, however the shape of 
      the superstructure was such that a considerable quantity of the copper 
      wire was blown over the afterdeck as it spooled from the launcher 
      resulting in short 
      circuiting of the copper wire.
(ii)  Premature and delayed launching of the probe - This generally resulted 
      from poor use of radios when communicating between the main lab and the 
      afterdeck.
(iii) Thermal shock - The first box of T7 XBTs were brought out of the hold 
      and stored at too high a temperature before being launched into sub-zero 
      water temperatures.  Two probes are believed to have failed for this 
      reason.

A fourth, and more serious problem encountered with the XBT probes was that of 
wire stretch.  The thin, two core, copper can become stretched during the 
descent of a probe due to snagging at either the spool in the launching 
canister or that in the probe itself.  This has the effect of increasing the 
resistance measured at the deck unit and hence recording erroneously high 
temperatures.  Once the wire has stretched it can of course never recover and 
if detected on a temperature trace, any record below that point of detection 
should be disregarded.  Occasionally wire stretch was obvious, however, more 
often the wire stretch would be very gradual and a temperature would look 
good: in such cases detailed comparison with known T/S curves for the area was 
required in order to identify erroneous temperature profiles.

3.14  MacSat satellite receiver 
      (John Allen)

This cruise saw the first use of this latest version of Newcastle Computer 
Services satellite image grabbing software and hardware on an RVS ship.  As 
with the previous version, it can only be used on board ship with an omni-
directional aerial which can receive infrared (IR) and visible spectrum images 
from polar orbiting satellites.  The major change with the new version 
concerned hardware as virtually all the receiver switching was software driven 
via a Dartcom micro-controlled interface box and an Apple Macintosh standard 
SCSI connection.  The requirement for a special interface board to be inserted 
into the Macintosh has disappeared and an LC or more powerful Macintosh could 
be used as a platform for the software.  In addition, the new software 
contained several useful new image processing tools; of particular merit were 
the image resampling and the auto contrast options.  The most welcome change 
was the auto-save option, not for the reason that its name suggests, but that 
with this option switched on both the IR and the visible images were recorded 
simultaneously.  A major disappointment, however, was the new geographical 
overlay tool, which, despite great care being taken to input our position 
correctly, failed to be of much assistance at these high latitudes.  The 
suggested overlay was simply not believable and our doubts were generally 
confirmed when on odd occasions land was identified on an image.

Unfortunately cloud cover, as expected at these latitudes, prevented any 
useful thermal images of fronts being made during the cruise.  However, ice 
extent could be identified at times and the tracks of particular storms were 
followed.  Only the long near overhead passes were worth receiving as others 
were not geographically relevant and gave poor reception.

Little attempt was made to pick up other than NOAA satellites as it was not 
clear how old our prediction data was for other polar orbiting platforms.  
NOAA's 10, 11 and 12 were successfully received during the cruise.  Table 3.6 
lists all images that have been archived on 3.5" disk.  The image format still 
takes up considerable space and compression routines do not help very much.  A 
shortage of disks forced the making of only one copy of each image and only 
images that could be located geographically and showed significant cloud free 
areas were archived.


Table 3.6  Archived MacSat images

                    Date      Time    IR  Visible  Disk no.
                    --------------------------------------
                    11/11/92  23:26z  yes  yes     198001
                    14/11/92  20:33z  yes  yes     198002
                    15/11/92  11:31z  yes  yes     198003
                    17/11/92  04:28z  yes  yes     198004
                    17/11/92  06:08z  yes  yes     198006
                    18/11/92  03:52z  yes  yes     198005
                    18/11/92  05:56z  yes  yes     198007
                    19/11/92  10:05z  yes  yes     198008
                    23/11/92  06:36z  yes  no      198009
                    26/11/92  12:38z  yes  no      198009
                     1/12/92  12:00z  yes  no      198010
                     1/12/92  12:34z  yes  no      198010
                     2/12/92  01:37z  yes  yes     198011
                     2/12/92  08:13z  yes  yes     198012
                     2/12/92  03:58z  yes  no      198013
                     3/12/92  16:58z  yes  yes     198014
                     5/12/92  09:27z  yes  yes     198015
                     5/12/92  12:04z  yes  yes     198016
                     6/12/92  22:53z  yes  yes     198017
                     7/12/92  23:15z  yes  yes     198018
                     8/12/92  11:42z  yes  yes     198019
                     8/12/92  13:25z  yes  yes     198020
                     9/12/92  11:20z  yes  yes     198021
                     9/12/92  20:34z  yes  yes     198022
                    10/12/92  10:57z  yes  yes     198023
                    10/12/92  11:40z  yes  yes     198024
                    10/12/92  12:39z  yes  yes     198025
                    11/12/92  10:37z  yes  yes     198026
                    11/12/92  12:17z  yes  yes     198027
                    12/12/92  07:52z  yes  yes     198028
                    12/12/92  11:54z  yes  yes     198029
                    

3.15  Non-toxic supply 
      (David Turner)

The non-toxic intake is drawn from approximately 4m below the waterline, using 
a pump with a nominal throughput of 200L.min^-1.  This supply is fed to a 
header tank on the hangar top through 2" stainless steel piping, and thence to 
the non-toxic taps in the laboratories.  Since a debubbled non-toxic supply 
was required for this cruise, the normal non-toxic taps were not used: instead 
the non-toxic supply on the hangar top was fed directly into a black plastic 
header tank (nominal volume 50L) with a headspace connected to the atmosphere.  
The (debubbled) outlet from this tank was then piped directly to the 
laboratories: hangar (3 taps); water bottle annex (3 taps); deck laboratory (3 
taps).  We are grateful to Tony Poole for installing the necessary pipework in 
time for the cruise.  The RVS surface system in the hangar (thermosalinograph, 
transmissometer, fluorometer) was fed by two further debubbling header tanks 
connected in parallel.  The arrangement of the supply to the various on-line 
instruments is shown in Figure 3.5*.



Figure 3.5*  Non-toxic water system

During most of the cruise, a regular sampling schedule from the debubbled 
supply was maintained.  The sampling periods are detailed in Appendix G, where 
it can be seen that sampling was suspended on days 320-321 (Potter Cove), days 
326-327 (first James Clark Ross rendezvous), and days 342-343 (final James 
Clark Ross rendezvous).  Samples were taken for chlorophyll (section 3.23), 
phytoplankton species (section 3.23), oxygen (section 3.21), alkalinity 
(section 3.19), and ammonium (section 3.22).  Salinity samples (section 3.11) 
were drawn from close to the thermosalinograph.  From flow rate measurements, 
the water was estimated to take 2 minutes on average to travel from the hull 
inlet point to the deck lab sampling position: this delay was taken into 
account in synchronising water sampling with SeaSoar surfacing.  Salinity 
samples (section 3.11) were drawn from close to the thermosalinograph.

3.16  pCO2
      (Jane Robertson)

pCO2 was measured using a gas chromatograph and showerhead equilibrator.  Air 
for the GC system was supplied using a compressor in the lab and hydrogen 
through the piped gas supply into the deck lab.  The hydrogen line was 
pressure tested before leaving Stanley and was found to be leak free, but 
subsequently several leaks were found on the valve and tap fittings in the gas 
bottle store.  It is not apparent whether this was the result of vibration 
from the ship or the change in temperature experienced as the ship went south.  
The leaks did not set off the gas alarm in the gas bottle store, which could 
have been due to the considerable draught in the store as both doors were 
commonly kept open and/or due to a fault in the alarm itself.  Subsequent to 
this event both doors were kept closed and a faulty electrical connection on 
the alarm repaired.  It is recommended that checks are made periodically on 
both the gas supply lines and the respective alarms.

The equilibrator was placed in the water bottle annex as this kept it close to 
both the de-aerated non-toxic supply and the GC in the deck lab.  The water 
bottle annex was kept cold, with auxiliary heating switched off enabling us to 
keep the temperature in the equilibrator as close as possible to the in situ 
temperature.  The pCO2 system required a few modifications as originally it was 
designed to operate on one continuous run of lab bench.  After some soldering 
by the PSO the equipment was joined up around a corner.  The programme for the 
system was written on board with an original cycle time of 17 mins which was 
reduced to 13 mins following the first survey.  During each cycle two 
measurements of the pCO2 in the water were made bracketed by standard and 
marine air measurements.  Marine air was approx. 355 ppm  2 ppm as measured 
throughout the cruise with no discernible change with latitude, which is as 
expected in the Southern Ocean.  The equilibrator caused some problems by 
filling up with water, which was not draining away fast enough to cope with the 
flow.  This was partially solved by re-machining parts of it, although it still 
required occasional attention.  There was an average increase in temperature of 
1.5 in the equilibrator from in situ temperature, which is four times that 
experienced on previous (temperate water) BOFS cruises.  After about one week 
of constant use the two motors on the system sheared and had to be replaced 
with spares.  The spares worked successfully with no further problems.  The 
system was working almost continuously on surveys and the final transect, 
giving a pCO2 value approximately every 6 minutes.

Preliminary calibration with a single standard value was done on board to 
provide data for the first survey to be contoured.  The contour map (Figure 
3.6*) shows considerable structure similar to chlorophyll and pH distributions, 
and related to the front that was part of the survey area.  This data is only 
preliminary and further calibration using a running standard and temperature 
corrections will be necessary before the data is ready to go onto the Bidston 
database.  These corrections and changes should be made in the next few months 
and the data will be sent to Bidston by the end of April along with data from 
a further cruise in the Southern Ocean in Feb/March 1993 (Discovery 200).
 

Figure 3.6*  pCO2, first survey


3.17  Total CO2 
      (Sean Debney, Jane Robertson)

Total CO2 is a measurement of dissolved inorganic carbon (DIC) and represents 
carbonate, bicarbonate and un-ionised species of CO2.  A thermodynamic 
relationship exists between TCO2, pCO2, pH and alkalinity and it was an aim of 
this cruise to determine all four parameters sequentially at sea for the first 
time.

The analytical system consists of two main components, the extractor unit and 
coulometric detector.  A sea water sample is filled from the non toxic supply 
and is fed under gravity to a calibrated pipette.  This is discharged into a 
stripping chamber where orthophosphoric acid quantitatively converts the DIC 
to CO2.  The CO2 is purged by a nitrogen carrier gas flow into the reaction 
cell where it is coulometrically titrated to an end point.

It was originally intended to measure TCO2 in continuous underway mode to 
parallel pCO2 and pH but several major problems made this impossible.  The 
coulometer chemicals and the WOCE TCO2 standards of known carbon content were 
stored in the deck lab chill store during transit from the UK.  At some time 
during this period the temperature fell from the nominal 10C to below 0C as 
was evident from the shattered remains of the standard storage bottles.  The 
standards were stored in 500mL Pyrex bottles and well insulated with 
polystyrene suggesting that the temperature remained below 0C for a 
significant period.  Apart from the financial loss, which was in excess of 
500, the loss of standards meant that the quantity and quality of the science 
originally proposed for this cruise was completely disrupted.  The coulometer 
chemicals would also have been subjected to freezing temperatures and it is 
not certain what effect this had on the solutions.  The chemical suppliers 
were contacted once freezing was suspected, and they explained that the 
chemistry of the solutions may be sensitive to such a temperature change.  
These factors forced a change to discrete sampling from the non toxic supply, 
and the analysis of up to 5 replicates of each sample to give a measure of 
confidence in analytical precision of the technique.  There were additional 
periods of system downtime due to various hardware problems in the extractor 
unit.  These included the failure of two valves, contaminated sample lines, 
defective glassware and degassing of the water in the sample bottle, the 
latter being solved by the employment of an insulation jacket.

A continual problem throughout the duration of the cruise was the intermittent 
electrical spikes and surges from the 'clean' supply.  This was also 
experienced by other operators in the deck lab, main lab and winch control.  
It is difficult to determine the absolute effect of this variability in supply 
on the performance of the coulometer, though there were definite periods when 
large surges adversely affected several other pieces of instrumentation 
simultaneously.  The synergistic effect of substandard chemical performance 
and variability in power supply undoubtedly reduced the sensitivity and 
stability of the coulometer.  This was manifested in poor intra sample 
reproducibility, high unstable blank values and premature demise of the 
reaction cell.

In discrete mode the system allowed, at best, a sample throughput of twenty 
samples a day, which seriously reduced the spatial resolution of surface 
mapping compared with that achievable in underway mode.  Despite this, the 
initial intercalibration of raw TCO2, pCO2 and pH data from survey one shows 
no major offset, giving some confidence in the accuracy of all three 
measurements made on this cruise.  The forced employment of a discrete 
sampling regime, although vital for quality control under the imposed 
conditions, did not provide a large number of data points for accurate 
contouring.  The plot included, however does indicate the delineation of TCO2 
across the front which compares favourably with contouring plots of pCO2 and 
pH.  (TCO2 data points used in contouring are marked on the plot).  All TCO2 
analyses have been entered on the 'underway' spreadsheet, however the data 
have to be recalculated to take into account calibration of the pipette volume 
and corrected thermosalinograph data.  These data will therefore not be ready 
for inclusion on the BOFS database at BODC until the summer of 1993.


3.18  pH 
      (Richard Bellerby)

As a master variable, pH exerts an important control on geochemical and 
biological processes in aquatic systems.  Seawater pH reflects the status of 
the carbonate system which provides the major short term CO2 buffer.  The 
carbonate system can be quantified by measuring any two of the parameters pH, 
pCO2, TCO2 and alkalinity.  It was the aim of this work to accurately measure 
pH and in so doing validate a novel spectrophotometric method for the on-line, 
real-time measurement of seawater pH.  Seawater pH, on the total hydrogen 
scale, was measured using a continuous, on-line multi-wavelength 
spectrophotometric method which utilises the acid-base absorption properties 
of the sulfonaphthalein indicator phenol red.  Simultaneously, an on-line 
potentiometric system was available to compare methods on the same seawater.

pH was measured at temperatures some 20-25C above in situ, and require 
correction for in situ salinity.  Preliminary calculations have been performed 
at selected times in the first survey: coincident pCO2 and TCO2 measurements 
were used to calculate pH, and the result compared with the measured 
spectrophotometric pH.  The results should be treated with some caution, since 
all three parameters require further salinity and/or temperature corrections.  
The residual pH, the difference between measured and calculated pH, is 
presented in Figure 3.7.  The mean residual for all analyses was 0.01pH units 
with a standard deviation of 0.007pH (n=48).  This reduces to 0.0070.005pH 
when only those values with absorption standard deviation less than 0.001 are 
used (n=24).  Continual analysis of discrete samples have shown a precision of 
about 0.005 pH units throughout the cruise.  Contour plots of raw pH data for 
surveys 1 and 2 show distinct areas of high and low pH which correlate well 
with features on the survey plots of pCO2 and temperature.

Data from the traditional potentiometric method experienced large jumps in 
electrode potential observed on other cruises.  This has previously been 
attributed to electrical noise from the ship's motion through the water, 
passing through the on-line system, and the ship's electrical supply.  
Maintaining a constant temperature between the pH buffers and the seawater 
also proved difficult.  Spectrophotometric measurements in areas with high 
chlorophyll levels were affected by the particulate load of the seawater.  
High standard deviations were observed in these regions possibly due to 
settling of detritus whilst the sample is held in the flowcell for analysis.  
Increasingly throughout the northward transect, small bubbles were seen to 
become entrapped in the flowcell where previously they had passed through.  
This was thought to be due to organics coating the inner walls.  The system 
was flushed with Decon to no avail, and then with nitric acid, which worked 
only in the short term.  It is suggested that the on-line seawater should be 
filtered prior to analysis to reduce the particulate load of the supply and 
that regular flushing with hydrogen peroxide may alleviate the problem of 
bubble trapping.


Figure 3.7*  pH offsets


3.19  Alkalinity 
      (Susan Knox)

Alkalinity was measured by photometric titration of seawater with hydrochloric 
acid in a procedure developed at Plymouth Marine Laboratory.  The indicator 
was bromophenol blue, chosen because its pK value lies beyond the bicarbonate 
equivalence point, the area of the titration curve of interest.  The 
absorbance was measured at 582  2nm in a 100mL sample bottle: the hardware is 
based on a standard oxygen endpoint detector, modified for measurement at the 
appropriate wavelength.

The alkalinity procedure was fully automated until the Metrohm burette 
irretrievably broke down.  The only other burette on board was attached to the 
oxygen titration system and so had to be shared between the two analyses.  
Unfortunately the replacement burette had different external control 
connections, with the result that alkalinity titrations had to be run by a 
semi-manual procedure.  As with the oxygen titration method, a rapid and 
precise photometric measurement was necessary, this time at each addition of 
acid, in order to obtain a good titration curve.  Initially all the problems 
that beset the oxygen analysis (section 3.21) applied to the alkalinity 
titration, especially degassing which was exacerbated by the release of carbon 
dioxide as the sample became more acidic.  By moving the equipment to a cold 
area (the hangar) the degassing was effectively subdued, as was the analyst in 
these spartan conditions.  Thereafter it became apparent that the movement of 
the ship, either itself or in its effect on the water in the water jacket was 
causing the light signal to vary significantly so that no reliable alkalinity 
data was forthcoming.  Several series of samples were analysed when the ship 
was hove to and at quiet intervals but the amount of successful data was very 
scanty.

The alkalinity procedure itself has been shown to work but more modification 
is necessary before it can be used at sea.

3.20  On-line oxygen 
      (Susan Knox)

On-line oxygen was measured using an Endeco type 1125 pulsed dissolved oxygen 
system loaned by Southampton University Department of Oceanography.  The 
oxygen sensor comprises a Clarke type polarographic cell containing a silver 
anode and a gold cathode surrounded by electrolyte and covered by a Teflon 
membrane.  Two such sensors were mounted in a perspex container through which 
the bubble free non-toxic supply flowed at a rate of 2 litres per minute.  The 
sampling rate was micro-processor controlled and was set to once every 5 
minutes.  Four sensors were provided and each had been pre-calibrated in low 
temperature seawater at Southampton University.  To provide extra calibration 
points, samples were taken from the sensor container outlet every 2 hours and 
analysed by Winkler titration.  A more detailed analysis of the calibration 
data is necessary before absolute oxygen values can be extracted from the 
Endeco output.

At the start of the cruise the pair of sensors did not track one another, with 
one sensor steadily drifting downwards and the other appearing to respond as 
expected.  A replacement electrode with fresh electrolyte and a new membrane 
also showed the same downward drift.  At first this was thought to be due to 
the low temperature of the seawater, about 1C, but the same drift was also 
seen in distilled water held at ambient laboratory temperature.  A closer 
inspection of the sensors revealed the gold cathode to be heavily tarnished.  
A light application of crocus paper removed the film and thereafter the 
electrodes performed satisfactorily.  As a result of the delay in getting the 
sensors to work continuous oxygen data is available from one electrode for the 
second box survey and from two sensors for the 88W transect north.


3.21  Oxygen titrations 
      (Susan Knox)

Discrete oxygen samples were taken from the non-toxic supply at 2 hourly 
intervals throughout the cruise, and from Niskin bottles on selected CTD 
casts.  These samples were analysed by Winkler titration with a photometric 
endpoint detection using equipment loaned from Queens University, Belfast.

At first the system was unworkable because of air leaks in the Metrohm 
burette.  These were traced to the washers which proved to be incompatible 
with the tube terminations.  Once the washers were removed the system became 
airtight and usable.  The next area of concern was the endpoint detection.  It 
was often very difficult and at times impossible to get an endpoint due to the 
degassing of the sample as it warmed up from its sampling temperature of 1C 
or less to the laboratory temperature of 23C.  Minute gas bubbles coated the 
sides of the sample bottle and floated free in the solution causing 
interference with the light path.  Neither allowing the sample to warm to 
ambient temperature before analysis, nor filling the water jacket with ice 
cold water helped in any way.  The eventual solution was to move the equipment 
to a cold environment (the water bottle annex) with an ambient temperature 
near that of the sample.  This solved the gas bubble problem, and in doing 
this the water jacketed measuring cell was sited fore and aft rather than 
athwartships as it had been in the laboratory.  This reduced the slopping of 
the water across the light path whenever Discovery rolled, which the new ship 
has a marked tendency to do, and reduced the noise on the photometer output.

The oxygen titration data available at the end of the cruise require further 
correction for (i) the standardised thiosulphate concentration, and (ii) the 
correct calibrated bottle volumes, which were not available from Belfast 
during the cruise.  The Antarctic surface waters were generally saturated with 
respect to dissolved oxygen and the CTD casts in this area showed the presence 
of a very well marked oxygen depletion layer below about 200m.  Along the 88W 
transect going north the oxygen minimum layer occurred at 1500m.

3.22  Nutrients 
      (Bob Head)

A Chemlab based autoanalyser system (ex North Sea Community project and Charles 
Darwin cruises 46/90 and 47/90 of the BOFS programme) was used to measure 
concentrations of nitrite, nitrate, phosphate, silicate and ammonium.  The 
methods used were the standard wet chemical methods developed at PML.  The 
colorimeter outputs were logged to a Siemens chart recorder and a level A 
interface.  The autoanalyser system was used in an online mode with the 
seawater input being taken from the Turner designs fluorometer outflow (section 
3.23): the fluorometer output was logged to the same chart recorder to enable 
visual comparison and fluorescence and nutrient changes).  Change-over to 
discrete sampling mode was effected by a three way valve in the autoanalyser 
sample input line to enable the input of both calibration standards and 
discrete samples from CTD casts or productivity experiments (sections 3.31 and 
3.32).  Nutrient calibration was effected by daily running of standards 
dissolved in low nutrient seawater (collected at Berkeley Sound, Falkland 
Islands) until a constant plateau was attained.  Milli-Q water was used as a 
reagent blank.

Discrete sample analysis required human autosamplers, running each sample for 
4 minutes interspersed with a 1 minute wash in low nutrient seawater.  
Calibration was effected by running standards in the same way.  Ammonium 
samples were run manually using 10mL seawater samples in test tubes, addition 
of reagents and colour development in the dark for 3 hours.  The sample was 
then introduced into the flowcell (minus the debubbler) with a 20mL syringe.  
Optical densities were recorded as chart units.  The main bulk of samples was 
composed of the CTD section across the working area on 85W and samples from 
productivity experiments (sections 3.31 and 3.32).  Workup of the discrete 
data should be completed in approximately 6 months.  Work up of the on-line 
measurements will be undertaken by BODC.

In the working area typical maximum nutrient values were 35m nitrate, >50m 
silicate and 2m phosphate.  Ammonium values were low with maximum values 
being of the order of 0.6-0.7m.  In the high bloom areas during the two 
surveys nutrient reduction could be observed corresponding to changes in 
chlorophyll fluorescence.  Some small surface ammonium increases could be 
detected in areas of intense krill swarms.  The high values of nitrate and 
silicate were an analytical problem which needed to be resolved as the system 
was set up initially for nutrient levels encountered in the North Atlantic.  
Nitrate and silicate concentrations were physically reduced by incorporating a 
Milli-Q diluent line in a ratio of 1:4.  Due to the set up of the analyser 
with 50mm flowcells it was observed that for both these analytes that there 
was a non-linear relationship between concentration and output.  
Concentrations of nitrate and silicate will therefore have to be calculated 
using a nonlinear curve fit.  The main reason for non linearity at high 
concentrations was assumed to be the provision of a 50mm flowcell, use of a 
15mm cell would have improved the precision of calibration values.  During the 
cruise the ammonium channel gave trouble due to precipitation of reagents in 
the analytical line.  The reason for this was unclear, but it was certainly 
due in part to the low water temperature.  There was also contamination of the 
reagents in the deck lab which manifested itself as a steady upward drift.  
Due to all these factors, ammonium was measured manually as described above, 
with samples taken hourly from the non-toxic supply.

The Chemlab system has now been around for a number of years and is really 
overdue for refurbishment and updating.  As the autoanalyser is a multi-user 
piece of equipment, there is a need for a manual together with parts lists and 
suppliers to be written and supplied with the equipment as not all users are 
expert in automated analysis.  Hopefully in the near future some funding can 
be made available to update a number of the components to enable continued use 
and better reproducibility.


3.23  On-line fluorescence, chlorophyll, and phytoplankton sampling 
      (Bob Head, Bill Miller)

On-line chlorophyll fluorescence was measured with a Turner Designs model 10-
005 rack mounted fluorometer (kindly loaned by M J Fasham of the Rennell 
Centre) fitted with a 10-020 high volume flowcell with a debubbled seawater 
input at a flowrate of 3 litres per minute.  The fluorometer output was 
recorded on a Siemens chart recorder and also logged to a level A interface.  
In addition, a fluorometer (sn 229) and transmissometer (sn 99D) were 
incorporated in the RVS 'surface' system in the hangar (see sections 3.4, 
3.15) These were logged by the same system as the thermosalinograph (section 
3.4), and ran virtually continuously through the cruise, outputting their data 
in SMP format to the level B system.  There appeared to be no problems 
communicating directly with level B.  During mobilisation a communication 
error with the transmissometer and fluorometer Rhopoint modules was indicated.  
This was found to be a software error and was soon rectified.  At present, the 
FSI and Rhopoint modules communicate through separate ports on the PC.  Future 
modifications to the FSI modules should allow the system to be configured 
using a single port as was originally planned.

More than 800 discrete samples were taken for chlorophyll 'a' and phaeophytin 
determinations.  Samples were taken both from the fluorometer output at hourly 
or 2 hourly intervals, and from CTD and GoFlo casts.  At the same time 100mL 
preserved samples for phytoplankton species identification were taken into 
both Lugols iodine and borax-buffered formaldehyde.  Analysis of these samples 
will be used not only to identify the distribution of the different species of 
phytoplankton in the cruise area, but also to estimate levels of phytoplankton 
carbon.

Analysis of chlorophyll 'a' and phaeopigments was carried out by filtering 
100mL aliquots of seawater onto 25mm GFF micropore filters and extraction by 
10mL 90% acetone for a minimum of 15 hours.  Prior to analysis a further 
addition of 10mL acetone was made.  Analysis was carried out on a Turner 
Designs 10AU fluorometer (QUB Belfast) fitted with a 25mm cuvette system: the 
chlorophyll fluorescence was measured before and after the addition of 10% 
hydrochloric acid.  Fluorometer calibration was carried out with pure 
chlorophyll 'a' standard (Sigma C5753 from spinach).  Calibration checks were 
made at regular intervals during the cruise.  The results are reported as 
mg.m^-3 chlorophyll 'a' and phaeopigments.  The underway discrete chlorophyll 
'a' analyses have been used to calibrate the Turner Designs on-line 
fluorometer, the Chelsea Instruments Aquatracka fluorometer mounted in the 
tank in the hanger area, and the SeaSoar fluorometer.  Similarly, the 
chlorophyll 'a' data from CTD bottle casts have been used to calibrate the CTD 
sensors (see section 3.24 for details of these calibrations).

All chlorophyll analyses were completed during the cruise.  The time scale for 
work up of the Lugols and formaldehyde preserved samples is uncertain but it 
is hoped that a proportion of the samples may be completed before the end of 
1993.  In the survey area, chlorophyll 'a' values were low both to the north 
and south of the bloom (<1mg.m^-3) whilst in the main bloom values greater than 
7mg.m^-3 were observed.  Phaeopigment values were generally low, both in the 
bloom and non-bloom areas.


3.24  Chlorophyll calibrations 
      (Gerald Moore, Alison Weeks)

CTD

The fluorescence from the CTD sensor was quenched at the surface.  Data for an 
initial calibration was selected where PAR was below 2Wm^-2.  On this basis two 
distinct calibrations were determined, for stations 12198 and 12200 combined, 
and for stations 12201 and beyond:

  ln [chl] = -8.0345 + 6.313V - 1.2020V^2	(stations 12198, 12200; r^2=0.95)
  ln [chl] = -7.5940 + 4.603V - 0.5721V^2	(stations 12201 et seq; r^2=0.88)

where V (volts) = fluorescence raw counts * 0.002441

The difference in calibration is due to a change in species assemblage.  Both 
calibrations were found to be non-linear at low chlorophylls or depths greater 
than 50m.  The calibrations are given below.  There were insufficient data to 
determine a reliable quench correction factor.  When the individual calibration 
curves are combined the overall variance explained was 95%, which gives an 
upper bound on the variability of the chlorophyll analysis.

On-line

The initial intention was to use the suite of programs developed during the 
Vivaldi cruise by John Hemmings (Rennell Centre).  However, there were problems 
in using these programs due to instrument errors, the light regime, and the 
species assemblages encountered during the cruise.  John Hemmings' methodology 
was used whenever possible, as discussed below.

The initial problem was caused by the Turner Designs fluorometer.  The first 
stage was to establish an instrument offset, which proved to be difficult due 
to range switching of the fluorometer.  The new level A software automatically 
adjusts the recorded data for the instrument range setting.  This assumes that 
the gain for each range is calibrated exactly, and that no offsets occur 
between ranges.  Without processing a set of standards it was not possible to 
determine if the gains were correctly calibrated, however a simple manual range 
check showed that there was a change in offset between ranges.  The range is 
not recorded by the level A, so the range change was determined by examination 
of a probability plot of the output from the fluorometer.  This showed a range 
change at 6.5 and 12.7V rather than the expected 5 and 15.8V.  For a 
preliminary calibration, the ranges identified were adjusted by inter-
calibration with the Chelsea Instruments fluorometer, a second underway sensor.  
A full check will involve a test of the instrument base and manual extraction 
of the range changes from the chart recorder output.

With the Turner and Chelsea datastreams adjusted it was possible to determine 
an offset and an initial calibration.  From this the chlorophyll yield was 
determined at the light minimum (PAR less than 5Wm^-2 as logged by the met. 
system, section 3.3).  This "night" yield was adjusted for changes in region, 
by using cluster analysis.  Three regions were determined: the Drake Passage 
transect; Potter Cove to the survey grid; and the survey grid.  These areas 
provided a base to adjust the "dark" chlorophyll for quenching.  When John 
Hemmings' model was used it was not possible to get a reliable fit between 
fluorescence yield and PAR.  Six other theoretical and empirical models were 
tested on the data.  None of these proved able to remove the effect of light in 
a reliable manner.  The cause of this problem is twofold: first the particular 
characteristics of the area; and second the statistical bias produced by the 
operational need to survey the high chlorophyll area in the daytime (see 
section 2).  There are two possible methods to correct the data: first is to 
use an empirical fit of the daily yield to a smooth function and correct the 
chlorophyll on a sample to sample basis; the second is to develop a better 
model of the quenching, to account for latitude and variation in 
photoadaptation.  For preliminary data analysis, the calibrations developed for 
the Chelsea and Turner fluorometers for night time are detailed below.

Turner Designs: (output is V volts)
          Instrument offset 5.9240; calibration after offset correction is:
          [chl] = 0.009853V + 0.004236V^2 (r^2=0.86)
          Region adjustments (scale factor) 6.0849 (Drake Passage)
                                            4.2519 (from Potter Cove)
                                            0.9622 (survey area)

Aquatracka (output X is exp(volts))
          Instrument offset 2.8368; calibration after offset correction is: 
          [chl] = 0.02772X + 0.001753X^2 (r^2=0.91)
          Region adjustments (scale factor) 3.4571 (Drake Passage)
                                            1.4859 (from Potter Cove)
                                            0.9416 (survey area)

Seasoar

The Chelsea Instruments fluorometer on the Seasoar showed a good relationship 
with the underway samples when data were selected for PAR less than 2Wm^-2.  Two 
linear calibrations were obtained, one for the Drake Passage transect and one 
for the survey area.  One of the main problems with the calibration was the 
poor time recording for the samples, which resulted in about 5% of the values 
being unusable because it was ambiguous as to which undulation they 
corresponded to.  The SeaSoar chlorophyll was subject to the same problems in 
determining a quench function as discussed above.  The calibrations are given 
below.

       Drake Passage transect:  ln[chl] = -2.5067 + 1.1071V (r^2 = 0.81)
       Survey area:             ln[chl] = -4.8823 + 1.8957V (r^2 =0. 92)

where V (volts) = fluorescence raw counts * 0.002441

Discussion

The major problem was caused by the range changes in the Turner Designs 
fluorometer.  This could have been solved if the ranges were available on the 
level A datastream, for incorporation into a calibration model, and this is 
strongly advised.  Another option is to keep the instrument on a single range; 
this would give a resolution of 0.2% for the selected range, which is 
sufficient on a 0-10 mg.m^-3 nominal range given the accuracy of the 
instrument.  Although the nature of the present survey caused special 
problems, it is doubtful that full calibration of chlorophyll fluorescence 
data can be achieved in near real time except on cruises which return to a 
previously studied area.


3.25  Biogenic sulphur compounds 
      (Phil Nightingale, Wendy Broadgate)

The main aim of this work was to determine the surface water concentration 
field for dimethyl sulphide (DMS) in order to estimate sea to air fluxes.  Of 
particular interest was the opportunity to investigate the influence of the 
large seasonal algal blooms in the Antarctic ocean on the global flux of 
marine biogenic sulphur.  A secondary aim was to determine the inter-
relationships between DMS, its precursor dimethyl sulphoniopropionate (DMSP) 
in both the particulate and dissolved phases and chlorophyll, temperature, 
salinity, phytoplankton species and other parameters such as primary 
productivity.

Methodology

Samples were predominantly collected from the ship's non-toxic seawater 
supply, sampling being coordinated with discrete measurements of oxygen, 
alkalinity, chlorophyll and preserved phytoplankton.  In addition, comparative 
samples of surface seawater were collected from CTD casts.  All samples were 
processed within 10 minutes of collection (except for deep water casts where 
samples were stored underwater at in situ temperatures for a maximum of 6 
hours).  Volumes of seawater used for DMS analyses were typically between 100 
and 200mL reflecting the low levels observed in the study area.  All samples 
were filtered using AP25 depth filters, the filter then being used for DMSP 
particulate analysis.  These were stored in the dark in a NaOH solution 
(~0.3M) for a minimum of six hours.  Dissolved levels of DMSP were determined 
by addition of 35mL of seawater to 1mL of 10M NaOH and samples stored as for 
DMSP particulate.  DMS was extracted from samples and pre-concentrated using 
purge and cryo-trap techniques and subsequently analysed in situ by dual 
channel flame photometric gas chromatographic techniques.

Results

Over 150 samples were successfully analysed for DMS and DMSP in both the 
particulate and dissolved phases.  These were collected with a 4 hourly 
frequency on the transects south from the Falklands and west from Elephant 
Island to the Bellingshausen Sea.  Extremely low levels of these compounds 
were encountered, concentrations were typically between 5 and 25ngSl^-1 and 150 
to 1000ngSl^-1 for DMS and DMSP (particulate) respectively.  Levels of DMSP 
(dissolved) were regularly below the detection limit, estimated to be 300ngSl^-1.  
The sampling frequency was doubled during the first bloom survey and 
considerably elevated levels of DMS and DMSP were observed in this region.  
Preliminary data analysis suggests the maximum DMS concentration to be 
300ngSl^-1, over an order of magnitude lower than levels typically encountered 
in algal blooms in the North Sea and North Atlantic.

Unfortunately, equipment problems encountered close to the end of the first 
survey lead to the cessation of both DMS and DMSP determinations, permanently 
in the latter case.  The fault appears to be electronic and the GC will 
require specialised attention back in the UK.  Total loss of detector 
sensitivity coincided with other equipment failures and the problem may well 
be related to surges and/or spikes in the ship's electrical supply.  Damage to 
one channel of the FPD GC initially appeared to be of a temporary nature only 
and DMS analyses were resumed after a loss of nearly four days.  However, 
detector sensitivity was unusually unstable and data collected during the 
second bloom survey will require careful screening in the UK to assess it's 
validity.  This second channel ceased to work permanently during the 88W 
transect, and DMS determinations were reluctantly abandoned.


3.26  Low molecular weight halocarbons 
      (Phil Nightingale)

A combination of electron capture detection and megabore capillary analytical 
techniques enables the concentrations of a wide range of halogenated compounds 
in seawater to be determined.  The sources of these compounds can be 
predominantly natural (eg methyl iodide), purely anthropogenic (eg carbon 
tetrachloride) or may be a combination of both (eg bromoform).  One of the 
main aims of this cruise was to obtain data on the spatial variation in levels 
of halocarbons in areas where no data have been reported.  Data obtained will 
be used to obtain estimates for sea to air fluxes.  In addition, this cruise 
offered a unique opportunity to investigate the marine source of biogenic 
halocarbons in particular, methyl iodide, chloroiodomethane, dibromomethane, 
bromoform, dibromochloromethane, dichlorobromomethane and chloroform.  
Previously the only field data to provide firm evidence for natural production 
of these compounds has come from coastal areas (such as the North Sea ) where 
macroalgae known to release halocarbons are present.  Data from a bloom area 
where both anthropogenic and macroalgal sources are absent will be invaluable 
in determining the role of phytoplankton in the global cycling of bromine and 
iodine.  Deep water casts are also useful in this respect as the presence of 
these compounds in waters where carbon tetrachloride and freon-11 are absent 
indicates that they must have substantial natural sources.

Methodology

Underway samples were collected from the non-toxic supply and coincided with 
sampling for DMS, chlorophyll, oxygen and other parameters also measured 
discretely.  Samples were collected from CTD rosette casts, using steel-sprung 
Niskins, varying in depth from 2 to 4500m.  Initial comparison of surface 
waters collected from the non-toxic supply and from the CTD indicate that 
although the former is heavily contaminated with F11 it is suitable for 
determination of most low molecular weight halocarbons.  Samples were analysed 
within 10 minutes with the exception of deep water casts where samples were 
stored under water at ambient seawater temperatures for up to a maximum of ten 
hours.  Halocarbons were extracted from seawater by purging with nitrogen 
doped with hydrogen (1%) that had been precleaned by passage over a palladium 
catalyst.  Samples were preconcentrated by trapping above liquid nitrogen and 
separation achieved using a DB 624 megabore column.  Detection was by electron 
capture and chromatograms were collected and archived using a specialised data 
software package.

Results

No results are available for the transects south from the Falklands to 
Elephant Island and west from there to the Bellingshausen Sea.  The extraction 
system was badly contaminated and a new line consisting of entirely new 
valves, tubing and fittings had to be constructed.  In addition, peak 
separation by the DB624 column originally used was found to be unacceptably 
poor and a new column plumbed in.  Problems were also encountered with one of 
the PCB's in the GC and with spiking affecting the data capture unit.  
However, all of the aims of the cruise were achieved.  Samples were collected 
at frequencies of 1hr and 2hr during both bloom surveys, from deep CTD casts 
and on the 88W transect.  During the latter period, samples were taken to 
coincide with Si, TCO2, POC, C/N ratio determinations in an effort to gain 
more insight into possible production of methyl iodide by diatoms and other 
species of phytoplankton.  Transects into and out of the bloom area should 
also give unique information on production of halocarbons by phytoplankton.  
Samples of marine air have also been collected for analysis using a GCMS that 
should enable some of the unknown peaks routinely seen on chromatograms to be 
identified.  Data will, however, require considerable and careful analysis 
before any conclusions can be reached.


3.27  Hydrocarbons 
      (Wendy Broadgate)

Due to their high reactivity, particularly with the OH radical, light non-
methane hydrocarbons play an important role in tropospheric chemistry, 
especially in the global budget of carbon monoxide and ozone.  Anthropogenic 
emissions (fossil fuel burning, natural oil and gas excavation, and biomass 
burning) are the major source of these compounds.  However, little data has 
been published on the marine production of non-methane hydrocarbons, although 
it's importance has been recognised.  Measurements in both air and surface 
seawater in unpolluted regions, especially during periods of high biological 
activity, are necessary to determine the fluxes of hydrocarbons between the 
sea and air and therefore the global marine source.  Samples were analyzed for 
the following non-methane hydrocarbons using gas chromatography and flame 
ionisation detection (GC-FID): ethane, ethene, propane, propene, acetylene, n- 
and i-butane, butene, pentane, pentene, hexane, hexene, heptane, heptene, 
benzene, 2-methyl butane, 2,2-dimethyl propane, 2-methyl pentane, 3-methyl 
pentane and 2,2-dimethyl butane.

Methods

Samples of air and seawater were preconcentrated and analyzed in situ by GC-
FID employing a 50m 0.53mm i.d. Al2O3 capillary column.  Air samples (2L) were 
cryoconcentrated on a 1/8" o.d stainless steel trap packed with 60 mesh glass 
beads.  Seawater samples (1.4L) were purged with nitrogen gas (CP grade) at 60 
mL/min for 30 minutes to remove the volatile gases which were carried in the 
gas stream through several water traps and condensers and concentrated in the 
same way as the air samples.  Calibration of the system was carried out by the 
injection of gaseous standards into a nitrogen gas stream and concentration as 
above.  The sample bottles were flushed six times to 60 p.s.i. with nitrogen 
gas between each analysis and a blank was run.  After each seawater analysis 
the sample was drained from the purge tube under nitrogen gas and a blank was 
obtained by purging the empty vessel.  Because the temperature programme of 
the GC took one hour the rate of analysis was reduced to approx. one sample 
every 3 hours.  As the analysis required constant attention it was not run 24 
hours a day.

None of this work could have been carried out without a reliable supply of 
liquid nitrogen.  In general this is a problem at sea because loss rates from 
storage dewars are such that 100% loss is observed within 30 days.  However, 
the use of an "Iwatani" liquid nitrogen plant on this cruise has proved to be 
very successful.  It consists of a nitrogen gas generator, a cooling unit and 
a compressor.  The 40L dewar which the unit uses was maintained at a level 
greater than 30L despite the removal of around 10L per day for use in 
experiments.

Sampling and storage

Air samples were obtained by pumping air into stainless steel electropolished 
canisters using a battery operated metal bellows pump.  Due to potential 
problems with navigation and communication it was not permitted to run a 
stainless steel tube from the mast on the monkey island to the lab.  Two tubes 
(one nylon and one teflon, 1/4" o.d.) were installed along this route and 
tested for contamination by comparison with samples taken on the monkey island 
using a 3m stainless steel tube (1/4" o.d.) but both were found to be 
unsatisfactory.  All air samples were therefore taken through a 3m stainless 
steel tube either on the monkey island or through a window on the bridge with 
the tube protruding over a metre from the bridge to windward (14m above sea 
level).  To eliminate the possibility of contamination from the ship's funnel 
no samples were taken when the wind direction was astern.  Samples obtained in 
this way were analyzed within 2 hours of collection.

High pressure air samples were taken at five sites.  These samples are stable 
for several years and can be taken back to the laboratory for reanalysis by 
gas chromatography-mass spectrometry.  Bottles were flushed with the sample as 
above, then partially immersed in liquid nitrogen for 5 minutes.  This 
liquefies air and a pressure of 200 p.s.i. can be obtained once the bottle 
warms to room temperature.

Water samples (1.6L) were taken from the non-toxic supply.  The bottles were 
stored in the dark under flowing seawater for up to 4 hours but generally 
analyzed immediately.  Samples at depth were taken from CTD and GoFlo 
deployments.  Underway samples were routinely collected over the whole cruise 
track, although analytical problems resulted in few successful measurements 
early in the cruise.  Seven CTD casts were analyzed each at two depths 
(combinations of 2m, 25m, 40m, 150m and 4000m).  Surface CTD samples were 
compared with non-toxic seawater samples at the same site.

Problems encountered

Initial problems with the air sampling lines were time consuming.  Once these 
had been overcome, severe problems with the seawater blanks were observed.  
Eventually it was found that the prep. line was contaminated, and it was 
rebuilt with new valves and tubing.  Further contamination resulted from two 
O-rings which were replaced with teflon fittings.  There were also ongoing 
problems with the sensitivity of the detector.  These were thought to be due 
to changing flow rates of the air and hydrogen gas supply lines which were 
shared with two other GC's.  However it may also be due to changing voltages 
and spikes produced by the "clean" electricity supply which was found to be 
"dirty".  Confusing behaviour from the data acquisition system attached to the 
computer was also due to spikes in the electricity supply.  The removal of a 
high current-drawing air compressor from the circuit corrected this 
malfunction.

Results

Despite the problems encountered during the first half of the cruise, a varied 
raw data set has been acquired from regions of diverse biological 
productivity.  However, the data requires processing before the results are 
available.


3.28  Particulate sampling 
      (Jane Robertson)

Along the 88W transect, samples were taken and filtered on an hourly basis 
for approximately 20 hours per day.  The samples were frozen and will be taken 
back to the UK at the end of Discovery 200 (March 1993).  They will be 
analysed by Dr. H. Kennedy (UCNW) for particulate and dissolved 13-C, and for 
particulate C, N, and Si.The resulting data will be analysed in conjunction 
with concomitant measurements of pCO2, TCO2 and pH.  The analyses will 
probably be complete by the end of 1993.


3.29  13-C sampling 
      (Colin Attwood)

Stable carbon isotope ratios are useful indicators of fluxes of carbon around 
the globe and through trophic webs.  The primary variation of the ratio of 
inorganic 13-C:12-C occurs latitudinally on a macro-scale in the atmosphere 
and in the ocean, and between water masses of different origin.  Both air and 
sea-water samples were taken for later mass-spectral analysis in Cape Town to 
measure the carbon isotope ratios in the Bellingshausen Sea and Antarctic 
Peninsula regions.

Air sampling

A polycarbonate hose of 8mm internal diameter was secured to an upright 
structure on the ship's monkey island, 16m above the water line and well 
forward of the ship's exhaust outlets.  The hose was run down to the water 
bottle annex and connected to the sampling apparatus.  This consisted of a 
water trap and a glass sampling bottle (500mL) with a silicon sealed plastic 
screw-cap through which an inlet and an outlet 6mm tube were inserted.  Air 
was sucked from the supply hose through the water trap and through the sample 
bottle with a suction pump.  After twenty minutes of suction, the pump and 
water trap were disconnected and the air sample was sealed in the bottle.  
Initially a tubular drying chamber was attached at the terminal end of the 
supply hose on the monkey island.  This chamber was later inserted adjacent to 
the water trap inside the deck laboratory as sub-zero temperatures had blocked 
the chamber with ice.  In total, forty air samples were taken, seventeen of 
which were spaced one degree of latitude apart along the northward transect of 
88W.

Seawater sampling

While underway, seawater samples were taken from the ship's non-toxic sea-
water supply.  When the CTD was deployed, water was taken from selected 
depths.  The sampling procedure was the same in both cases.  350mL plastic 
sampling bottles with a press on cap were used to store the samples.  The 
bottles were filled to overflowing from a Teflon tube.  0.15mL of 50% 
saturated HgCl2 solution was pipetted into the sample to poison metabolic 
activity.  The bottles were slightly squeezed when sealed, which allowed for 
the expansion of sea-water as it warmed without the necessity of leaving an 
air space.  Thirty-three underway samples were collected covering a range of 
latitudes between 53 and 69S.  141 samples were taken from Niskin bottles 
deployed with the CTD.  From shallow casts samples were taken from the 
following depths: 2, 40, 100, 200, 300, and 500m; from the deeper casts 
samples were taken from 2, 500, 1500, 2500, 3500m and the bottom.

Sample processing

The samples will be processed at the University of Cape Town upon the ship's 
arrival (February 1993).  From air and water samples, CO2 gas will be 
extracted and then injected into a mass-spectrometer for 13-C analysis.  
Details of procedures to be carried out in Cape Town will be made available 
together with the results.


3.30  Krill acoustics 
      (Alistair Murray and Doug Bone)

Objectives

The study aims to provide an acoustic data set at two frequencies (38 and 
120kHz) in the marginal ice edge zone (MIZ) in the Bellingshausen Sea, an area 
not previously surveyed by BAS.  We hope this will lead to an increased 
understanding of the physical and biological factors controlling the 
distribution and aggregation of krill.

Equipment and deployment

The equipment used was a Simrad EK400 scientific sounder operating in 
multiplex mode at 38 and 120kHz.  The data were logged to a PC using a custom 
data acquisition card and software - the Biosonics Echo Signal Processor 
(ESP).  The PC and sounder were set up in the main lab.  Some problems were 
experienced with the EK400 and PC in the early stages of the cruise.  It 
proved impossible to achieve multiplex operation although everything appeared 
satisfactory when logging either frequency on its own.  After a visit from the 
BAS technician from James Clark Ross the problem was resolved by reverting to 
the earlier version of the Biosonics software running under Windows 3.00.

The towfish was installed whilst the ship was alongside at Port Stanley.  A 
test deployment of the fish was made not long after departing Berkeley sound.  
Unfortunately, whilst lowering the fish to the water, it swung as the ship 
rolled and the tail fins struck the side of the ship distorting them 
sufficiently to cause problems later in the cruise when attempting to tow at 
more than about 9 knots.  This problem was cured temporarily by fitting thin 
wire stays to realign the fins, unfortunately these soon broke and thicker 
replacements had sufficient drag to cause the fish to veer under the ship.  
This problem prevented us using the system on the final transect up 88W where 
it was necessary to maintain 10 knots steaming speed.

Calibration

This was carried out on day 321 with the ship anchored in Potter Cove, King 
George Island in a depth of about 30m.  The calibration rig was assembled on 
the starboard deck and craned over the side with the standard target (a 38.1mm 
tungsten carbide sphere) suspended about 5m below it.  The calibration of the 
38kHz sounder was straightforward and the results appear satisfactory.  
However, the 120kHz calibration was very difficult and the results are 
suspect.

Operation

Details of all acoustic tows are given in Appendix F.  On the transect from 
the Falklands to Elephant Island, several acoustic runs were made at both 38 
and 120kHz.  After the calibration at Potter Cove, the passage down the 
Bransfield Straight was run at 38kHz.  The fish was recovered during the 
passage to the marginal ice edge zone study area as the ship's speed was at 
times too fast for the towfish.  The towfish was deployed during the first and 
second grid surveys and much useful data collected.

Results

Some targets were detected on the runs just to the south of the Antarctic 
polar front.  A few scattered marks were seen on the 38kHz run from Potter 
Cove down the Bransfield Strait.  On arrival at the MIZ study area, no 
significant targets were detected for some two days, until the ship steamed 
south of about 6730'S.  The general pattern of target distribution is shown 
in Figure 3.8*.  Most echo traces were characteristic of krill swarms in 
appearance.  No large layers or concentrations were found, and only a few 
large swarms.  Most swarms were small and many were fairly shallow (in the top 
80m or so of the water column).  There was a total absence of the diffuse 
layer type targets that are common around South Georgia (for instance) and 
which usually turn out to be copepods or small species of euphausid.  When 
observed on the scope display most swarms gave a very similar response on both 
38 and 120kHz.  Some swarms were clearly associated with chlorophyll patches 
and some were being preyed on by Minke whales.  The area where krill swarms 
were found was south of a front found during the first grid survey.  Thus 
there was a clear association of krill with watermass.  Some swarms at the 
southern end of the second survey were in an area of low chlorophyll and high 
nutrients where the CO2 was in equilibrium with the atmosphere - suggesting 
that there had not been any bloom in this area.  When the physics and biology 
of the survey area are further analyzed and interpreted it may be possible to 
make some inferences about the processes controlling the distribution of krill 
swarms.


3.31  Size-fractionated primary production 
      (Graham Savidge)

This work was designed to address two linked objectives based on separate but 
complementary experimental approaches.  The foremost aim of the programme was 
to employ the 14-C uptake technique to establish the spatial variability of 
size-fractionated (SF) primary production within the survey zone and to relate 
this to the prevailing conditions.  The secondary objective was to obtain 
estimates of the growth potential of SF phytoplankton populations maintained 
under a range of ambient irradiances using batch cultures of natural 
populations obtained from the survey area.

With the cruise programme being based on survey mode, it was not possible to 
obtain estimates of primary production from in situ incubations.  The 
considerable variability expected in the ambient environmental conditions for 
the area suggested that the optimal approach to determining primary production 
should be based on estimates derived from alpha and P(max) values obtained from 
P:I curves generated under defined artificial irradiance conditions.  During 
the cruise, the P:I characteristics of 40 samples were determined with 11 
samples being assayed during the transect to the main survey area, 21 assayed 
during the study in the main survey area and the remaining 8 samples being 
collected from the final northwards transect.  Comparisons between productivity 
determined from in situ observations and derived values of alpha and P(max) 
have been made during the concurrent James Clark Ross cruise.

Samples were collected once or twice per day either from the surface using a 
bucket or from selected depths using clean GoFlo bottles mounted on Kevlar 
line and incubated for 4h at 24 constant irradiances in a light gradient 
incubator.  The samples were cooled by running seawater from the non-toxic 
supply, with temperatures generally being held within 0.5C of ambient.  
Following incubation, samples were filtered under controlled vacuum through a 
cascade of defined pore size filters allowing separation of the phytoplankton 
population into fractions >18m (microphytoplankton), 2-18m 
(nanophytoplankton) and 0.2-2m (picophytoplankton). 14-C uptake by the SF 
populations was assayed using standard liquid scintillation techniques with 
the counts being corrected for dark uptake based on samples to which 100L of 
saturated DCMU had been added.  No problems were encountered with the 14-C 
uptake technique.  alpha and P(max) values were estimated from the uptake data 
using a curve fitting routine based on a standard P:I curve.  For each 14-C 
uptake determination, complementary samples were also taken for the 
determination of SF chlorophyll concentrations in the initial sample and also 
for taxonomic assay.  Primary processing of the 14-C and associated data has 
been virtually completed.
 

Figure 3.8*  Observed krill distributions, plotted on second survey track


Seven growth experiments were carried out during the cruise with the initial 
three sample populations being taken during the transit passage to the main 
survey area and the remaining four samples being taken from a range of sites 
in the survey area itself.  The experiments were set up using surface samples 
which were incubated for 4 days in 1l polycarbonate bottles under 100, 70, 53, 
34, 14 and 3% ambient irradiance in an incubator mounted on the aft deck and 
cooled from the non-toxic seawater supply.  Sub-samples were taken at 
approximately 24h intervals for the determination of SF chlorophyll 
concentrations, using size divisions as previously, and also for the assay of 
nutrient concentrations in the culture bottles.  These data were available 
from all experiments at the end of the cruise.

Preliminary assessment of the SF chlorophyll data has confirmed the generally 
low chlorophyll concentrations throughout much of the cruise area indicated 
from other approaches employed with the exception of the well-defined band of 
high chlorophyll located well offshore of the ice edge in the main survey 
zone.  In the extensive low chlorophyll areas, the nano- and picophytoplankton 
constituted the greater proportion of the total phytoplankton population but 
with the microphytoplankton fraction responsible for the major increases in 
the zone of higher chlorophyll.  Large diatom cells were clearly visible in 
several samples from this area.

The 14-C productivity data indicated particularly low values of alpha and 
P(max) expressed on a per unit chlorophyll to be predominant throughout the 
cruise area with values for alpha generally ranging from 0.001-0.002gC[Em^
-2s^-1]^-1 [gchl]^(-1)h^-1 and P(max) values ranging from <0.5 to 1 
gC[gchl]^(-1)h^-1.  Higher values of both parameters were typically associated 
with the two smaller size fractions.  On a per unit volume basis, values of 
P(max) tended to reflect the local chlorophyll concentration.  Within the 
detailed survey area variations in the values of alpha and P(max) on a per unit 
chlorophyll basis were, however, observed and these data will be analysed in 
relation to the local phytoplankton population characteristics and hydrographic 
structure.

A significant and unexpected observation from the cruise was the relative 
constancy of the mixed layer depth (MLD), as indicated by the distribution of 
chlorophyll fluorescence, at approximately 70m across much of the cruise area, 
including the detailed survey zone.  The transcending of this common depth 
across hydrographic structures characterised by contrasting phytoplankton 
biomass levels suggests that the MLD may not be significant in controlling the 
growth of phytoplankton in this region.  A very marked feature of the region 
was the extremely short time scale variability in the meteorological 
conditions and a major objective in data work-up will be to model the 
phytoplankton response to this variability using observed alpha and P(max) 
values together with representative data on ambient irradiance input, MLDs, 
PAR vertical attenuation coefficients and vertical distributions of 
chlorophyll as obtained during the cruise.  The basic format for an 
appropriate model is in place.  The model will also provide predictions of 
integrated column productivity.

Phytoplankton population increases were recorded in all growth experiments 
with greatest increases generally being observed for the microphytoplankton 
fraction.  In cultures with relatively large inocula nutrient limitation was 
recorded after 2-3 days.  A clear pattern to emerge was for maximum growth to 
be associated with the two lowest irradiances employed and for growth to 
commence after a 24h lag period.  Data on the growth;irradiance 
characteristics will be referred to ambient irradiance levels recorded over 
the period of the experiment and analysed in conjunction with both the depth 
distribution of total irradiance and the MLD.

This report would be incomplete without an especial thanks to Mike Behrenfeld 
and Mike Hilles for their help in sample collection, to Bob Head for carrying 
out the nutrient analyses and to Alastair Murray for his willingness to assist 
with the chlorophyll analyses.


3.32  Size-fractionated new and regenerated production 
      (Howard Waldron, Colin Attwood)

The primary objective of this study was to investigate the nitrogen dynamics 
of the phytoplankton community in an open ocean area of the Bellingshausen 
Sea, close to the retreating ice-edge during the austral Spring, and along a 
S-N transect from the main study area to around 50S.  Additional work of a 
similar nature was also undertaken opportunistically across the Drake Passage 
and in the north eastern sector of the Bellingshausen Sea whilst en route from 
the Falkland Islands.  The sampling programme consisted mainly of two types of 
experiment:

1. Size-fractionation. 
   This work was carried out while underway using bucket samples of surface 
   water. Grazers were excluded by pre-screening (<200m) and two 5L sub-samples 
   (for NO3 and urea uptake) and one 6L (for NH4 uptake and regeneration) were 
   supplemented with the appropriate concentration of 15N salt solution. 1L of 
   the NH4 sub-sample was drawn off and treated for later particulate N and 
   isotope dilution analyses at time zero (R0). The three 5L bottles were then 
   incubated at ambient sea surface temperature for 24h on deck at the 100% 
   light level. The samples were size-fractionated, post-incubation (<200m, 2L; 
   <20m, 2L; <2m, 1L) and the experiment terminated by filtration onto pre-
   ashed 47mm GFF filters. 900mL of the NH4 filtrate was retained as previously 
   for isotope dilution measurements at time Rt. 
2. Productivity Stations. A light profile was obtained from the CTD deployment 
   and bulk GoFlo water samples were subsequently obtained from the 100, 50, 25, 
   10, 1 and usually 0.1% light levels. For each of these depths two 2L 
   subsamples of water were supplemented with the appropriate concentration of 
   Na 15-NO3 and CO(15-NH2)2 respectively. For NH4-N uptake, a 3L sample was 
   spiked with 15-NH4Cl. 2L of this were drawn off and incubated with the NO3 
   and urea samples for 24h on deck at the appropriate light level 
   (cooled/warmed by surface water supply). Note that all samples were pre-
   screened. As in the case of size-fractionated work, 1L of the NH4 subsample 
   was used at time R0 and time Rt for a combination of particulate N and 
   isotope dilution analyses.  Experiments were terminated by filtration. 

In addition to the above, a time-series experiment was conducted over 3 days 
during the early part of the cruise and one of the productivity stations 
consisted of an inter-calibration exercise with the James Clark Ross.  Figure 
3.9* shows the generalized location of stations and detail with respect to the 
study area.  In summary, 14 productivity stations were completed (including 
one inter-calibration) and 8 underway experiments (one of which was a time-
series).  The identification of a density (haline-dominated) and fluorescence 
front during the SeaSoar survey made for a particularly rewarding choice of 
station positions.

Results from this work will not be available until some time after Discovery 
returns to Cape Town in early February 1993.  It is anticipated, however, that 
when post-cruise analyses have been completed, figures will be published of 
new and regenerated production (and ammonium regeneration) over the depth 
range of the euphotic zone and between different size classes of the 
phytoplankton community.  This has important implications for the biological 
viability of the early spring bloom and the extent to which these waters act 
as a sink for carbon.

Figure 3.9*  New and regenerated productivity stations. Legend: O size 
             fractionation experiment, surface water; O time series experiment 
             surface water; X water column production station


3.33  UV-B radiation and primary production 
      (Mike Behrenfeld, Mike Hilles)

Man-made chemicals released at the earth's surface are resulting in global 
decreases in stratospheric ozone concentrations.  The direct effect of 
stratospheric ozone depletion is an increase in surface intensities of 
ultraviolet-B radiation (UVBR: 290-320nm).  The short wavelengths of UVBR are 
biologically damaging and, therefore, increases in UVBR represent a 
significant threat to both terrestrial and marine organisms.  Currently, the 
largest decreases in stratospheric ozone have occurred over Antarctica each 
austral spring since 1978.  This cruise provided a unique opportunity to 
measure UVBR effects on marine organisms in this region of large stratospheric 
ozone depletions.

Measurements were made on the photoinhibitory potential of UVBR on Antarctic 
marine phytoplankton productivity, as measured by radiolabelled carbon uptake 
(NaH14CO3).  UVBR effects on phytoplankton photosynthesis were determined for 
both ambient and artificially enhanced UVBR doses.  Enhanced UVBR doses were 
both quantitatively and spectrally realistic.  Phytoplankton samples were 
collected from 100%, 10%, and 1% light levels during cruise stations using 
GoFlo bottles and incubated at surface intensities.  Three sample depths were 
used to allow determination of UVBR photoadaptation as a function of light 
history.  In addition to productivity experiments, measurements were also made 
of ambient solar radiation.  These measurements included broadband 
photosynthetically active radiation (PAR) doses, narrow band UVBR doses (at 
290, 300, 310, 320nm), and per-nm spectral measurements for 280-330nm.  Light 
data were collected as mean intensity over 1-2 minute intervals.

Carbon fixation rate, as measured during the UVBR studies, ranged from 0.02 to 
3.05 mgCm^-3hr^-1 in phytoplankton samples exposed to surface intensities of PAR 
and shielded from all UVBR.  Carbon fixation in the ambient UVBR treatment 
samples were generally not significantly different than those in the UVBR 
excluded treatment.  The lack of an ambient UVBR effect was not surprising, 
however, since weather conditions during almost the entire cruise were such 
that solar UVBR intensities were negligible.  Carbon fixation rates in the 
enhanced UVBR treatment were significantly depressed compared to the UVBR 
excluded treatment.  Decreases in carbon uptake from UVBR enhancement ranged 
from <10% to 50%, with an average decrease of 15%.  All productivity values 
stated above are preliminary and will require additional analysis before being 
publishable.

Solar broadband PAR measurements were collected for each day from 316 to 347 
(excluding 320), have been condensed onto Lotus spreadsheets, and copies 
rendered for the BOFS database.  Spectral and narrow band measurements of UVBR 
were also completed each day during the cruise (excluding days 334 and 335), 
but will require substantial time for analysis since these data have to be 
carefully checked for wavelength offset and calibrated for intensity.  At 
best, this data may be available for the BOFS database at BODC by June/July, 
1993.


3.34  Instrumentation notes 
      (Bill Miller, Phil Taylor)

Wave Recorder

The system was run only during deep CTD cast when the ship was hove to.  No 
problems were observed; the data will be analysed at the Rennell Centre to 
assess system performance.

Millipore Water Purification System

The system provided good quality water throughout the cruise; many thanks to 
Tony Poole for connecting it the ship supply prior to arrival at Port Stanley.

Test equipment

It is strongly recommended that a few thousand pounds be invested in new 
equipment for use by RVS personnel on the ship.  There is not a decent 
oscilloscope on board, the equipment looks like that which came off the old 
Discovery over two years ago.

Clean AC Supply

This has already been reported to RVS and actions proposed.  The salinometer 
had periods of instability until it was supplied through an UPS (borrowed from 
BAS).  The clean supply is going to cause problems until an adequate dirty 
distribution is put in so as to keep the clean supply clean.  This is 
particularly so in the deck lab.

30L GoFlo bottles

All bottles were serviced on board before deployment and worked well.  One 
bottle was lost when the kevlar line parted as the bottle was being hauled to 
the surface.  It is suspected that a slack turn may have fouled a roller on 
the rex roth winch and jammed; as a precaution the roller was removed.

EM log

No EM log data was provided during the cruise owing to a faulty sensor and a 
lack of spares on board.  It was confirmed soon after leaving Port Stanley that 
the fore/aft sensor on the EM log was u/s, one of the electrodes showing a 32-omega 
resistance path to ground.  The log was subsequently removed and a blanking 
plug inserted in its place.

SIMRAD EA 500 echo sounder

The EA 500 echo sounder had to be run on the hull transducer for the duration 
of the cruise because the PES winch was being used for the deployment of the 
BAS acoustic fish (section 3.30).  Performance was adversely affected (no 
bottom trace!) when the ship was pitching even in moderate weather, however no 
such deterioration in performance was present when the ship was rolling.  
Although much improved, the same characteristics were displayed when the ship 
was on station.


3.35  Computing 
      (Rod Pearce, Raymond Pollard)

ABC system

Discovery was equipped with an RVS ABC data logging and processing system.  
Scientific, navigational and operation data were logged in 'real-time' by 13 
level A units and a further 3 level A emulating PC's.  Details of the 
instruments and equipment logged by level A are given below:

Navigational       Trimble GPS receiver, Ashtech GPS receiver, Transit 
                   satellite receiver, Chernikeef log, gyrocompass.
Underway sampling  RVS surface sampling system, nutrient analyzer, Turner 
                   Designs fluorometer, pCO2 analyzer, pH analyzer.
Deployed packages  CTD, SeaSoar, Lightfish.
Other instruments  RVS metlogger, PML PAR light meter, wave recorder, 
                   echo sounder, RVS winch monitoring system.

The CTD and SeaSoar instrument packages were logged using an VME architecture, 
OS-9 based Mk 2 level A.  After completing the first SeaSoar transect it was 
identified by Raymond Pollard that the Mk 2 level A software derived the 
delta-T variable incorrectly.  Gerald Moore also identified that the level A 
was misinterpreting some of the sign information in the raw data frames.  It 
was not possible to rectify these faults on board as no code development 
facilities for the Mk 2 level A's had been provided.  Details of the problems 
and the associated solutions had to be sent to RVS Barry for compilation, 
copies of the corrected Level A executables were then returned five days later 
via the ship-shore communications link.  The raw data from the first SeaSoar 
transect was retrieved from the PC archive files and converted into level C 
format for reprocessing by the PStar team.

Both the Trimble and the Ashtech GPS receivers were logged by Syntel/OS-9 
based Mk 2 level A units.  There were problems associated with the logging of 
data from both receivers; the Trimble level A frequently failed to log all the 
data contained in the fix messages output by the receiver; the Ashtech level A 
regularly stopped logging data and had to be restarted manually.  Both of 
these problems have been reported to RVS Barry for further investigation.

The level B unit received over 1.2 Gigabytes of serial data and system 
messages from the various level A sources during the cruise.  The messages 
were logged to tape and forwarded via an ethernet link to the level C system.

The level C system was based around a single SUN IPC workstation with 400 
Mbytes of hard disk capacity for holding raw and processed data files.  In 
addition to the data received from the level B the level C system also logged 
data directly from the PC controlling the ADCP (section 3.2).  Data from other 
off-line sources (including the towed UOR package, the acoustic fish, the XBT 
system and some surface sample analyzers) were also incorporated into level C 
system as individual data files.

Level C data processing was limited to navigation, depth and the contouring of 
some surface sample files.  As the ship's log was inoperative, the processed 
navigation data was derived purely from averaged GPS fixes.  The SeaSoar/CTD 
processing and the surface sample calibration work was carried out by the 
PStar team.  A further three SUN workstations were provided for use by the 
PStar team.  All four workstations were connected to the ship LAN (Local Area 
Network) allowing PStar users direct access to the raw data files held on the 
level C workstation.

PStar

Three SUN workstations were used for PStar processing throughout the cruise, 
with several Macs used as extra terminals.  All saw heavy use.  Most of the 
system and data were held on a 1.4Gbyte drive attached to Discovery2, but 
three 0.3Gbyte drives attached to Discovery3 and 4 were also filled.  One of 
the disks associated with PStar workstations developed a fault early in the 
cruise, it is unclear whether this fault was in any way linked with the 
electrical supply problems experienced by other scientific equipment during 
the cruise.  The large space allocation proved most valuable in allowing data 
to be reworked and scientifically analysed during the cruise.  Archiving 
utilised 150mbyte cartridges, and an optical drive was a new innovation.  
Files were backed up on both media before being deleted from the system.  
Heavy use was made of an HP Laserjet colour printer and the Nicolet drum 
plotter.  While there were a few system or computer crashes, most often 
through lack of swop space, the system proved reliable, allowing all data from 
the CTD, SeaSoar, ADCP and navigational instruments to be processed to near-
final state during the cruise.

--------------------------------------------------------------------------------



APPENDIX A.  SCIENTIFIC PARTY

David Turner      Plymouth Marine Laboratory          (Principal Scientist)
John Allen        James Rennell Centre
Colin Attwood     University of Capetown
Mike Behrenfeld   Oregon State University
Richard Bellerby  Plymouth Marine Laboratory/University of Plymouth
Doug Bone         British Antarctic Survey
Wendy Broadgate   University of East Anglia
Sean Debney       University College of North Wales/Europa Scientific
Gwyn Griffiths    James Rennell Centre
Bob Head          Plymouth Marine Laboratory
Mike Hilles       Western Washington University
Susan Knox        Plymouth Marine Laboratory
Polly Machin      British Oceanographic Data Centre
Gerald Moore      Southampton University/Plymouth Marine Laboratory
Anne Morrison     NERC Remote Sensing Applications Development Unit
Alistair Murray   British Antarctic Survey
Phil Nightingale  University of East Anglia/Plymouth Marine Laboratory
Raymond Pollard   James Rennell Centre
Jane Read         Institute of Oceanographic Sciences Deacon Laboratory
Jane Robertson    Plymouth Marine Laboratory/University College of North Wales
Graham Savidge    Queen's University, Belfast
Howard Waldron    University of Capetown
Alison Weeks      Southampton University
Bill Miller       NERC Research Vessel Services       (Senior Technical Officer)
Colin Day         NERC Research Vessel Services
Rod Pearce        NERC Research Vessel Services
Tony Poole        NERC Research Vessel Services
Phil Taylor       NERC Research Vessel Services





APPENDIX B.  STATIONS WORKED

Day  Time  Station  Lat. S    Long. W   Cast     Water    Gear
                                        depth/m  depth/m  
------------------------------------------------------------------
316  2255  12198#1  5229.2'  5738.7'   349      361      CTD
317  0009  12198#2  5230.7'  5737.7'                   30L GoFlo
320  0805  12199#1  6049.5'  5445.6'    70             30L GoFlo
320  0824  12199#2  6049.5'  5445.5'    30             30L GoFlo
320  0830  12199#3  6049.5'  5445.5'     5             30L GoFlo
322  1138  12200#1  6249.9'  6034.8'   248      293      CTD
322  1215  12200#2  6250.1'  6034.9'   124             30L GoFlo
322  1224  12200#3  6250.2'  6035.0'    62             30L GoFlo
322  1230  12200#4  6250.2'  6035.0'    37             30L GoFlo
322  1236  12200#5  6250.2'  6035.0'    23             30L GoFlo
322  1242  12200#6  6250.2'  6035.0'     2             30L GoFlo
323  1126  12201#1  6327.5'  6617.5'   299      3246       CTD
323  1210  12201#2  6326.7'  6617.5'   182             30L GoFlo
323  1226  12201#3  6326.8'  6617.5'    97             30L GoFlo
323  1235  12201#4  6326.6'  6617.5'    54             30L GoFlo
323  1244  12201#5  6326.4'  6617.4'    28             30L GoFlo
324  1130  12202#1  6405.8'  7328.9'   298      3748     CTD
324  1207  12202#2  6406.1'  7330.0'    60             30L GoFlo
324  1215  12202#3  6406.3'  7330.3'   100             30L GoFlo
324  1225  12202#4  6406.5'  7330.6'    80             30L GoFlo
324  1233  12202#5  6406.6'  7330.9'    40             30L GoFlo
324  1336  12202#6  6406.6'  7331.9'    12             30L GoFlo
325  1835  12203#1  6501.1'  7922.6'   246      4109       CTD
325  1930  12203#2  6500.6'  7920.3'    27             30L GoFlo
325  1940  12203#3  6500.6'  7920.3'     2             30L GoFlo
326  1120  12204#1  6454.4'  7928.5'   297      4175     CTD
326  1204  12204#2  6455.6'  7925.7'     2             30L Go Flo
326  1211  12204#3  6455.6'  7925.7'   100             30L Go Flo
326  1221  12204#4  6455.7'  7925.7'    76             30L Go Flo
326  1228  12204#5  6455.9'  7925.7'    46             30L Go Flo
326  1235  12204#6  6456.0'  7925.6'    23             30L Go Flo
326  2342  12205    6500.7'  8139.3'   993      4299     CTD
327  0601  12206    6459.7'  8230.8'   994      4478     CTD
327  1100  12207#1  6500.4'  8322.1'  1000      4540     CTD
327  1254  12207#2  6500.7'  8322.7'     2             30L GoFlo
327  1302  12207#3  6500.7'  8322.8'    99             30L GoFlo
327  1312  12207#4  6500.8'  8322.8'    66             30L GoFlo
327  1320  12207#5  6500.9'  8322.8'    33             30L GoFlo
327  1324  12207#6  6500.9'  8322.9'    20             30L GoFlo
327  1327  12207#7  6501.0'  8322.9'    10             30L GoFlo
327  1736  12208    6500.2'  8409.8'   996      4547     CTD
327  2358  12209    6459.2'  8500.3'   993      4591     CTD
328  0421  12210    6520.0'  84'59.9'   995      4537     CTD
328  1220  12211#1  6540.4'  8459.9'   993      4533     CTD
328  1309  12211#2  6540.6'  8459.6'    60             30L GoFlo
328  1317  12211#3  6540.7'  8500.1'    30             30L GoFlo
328  1323  12211#4  6540.7'  8500.0'     2             30L GoFlo
334  0012  12212    6644.7'  8500.5'  4349      4389     CTD
334  0953  12213#1  6704.3'  84'55.8'   999      4310     CTD
334  1210  12213#2  6705.9'  8451.1'     5             30L GoFlo
334  1220  12213#3  6706.0'  8450.9'    38             30L GoFlo
334  1226  12213#4  6706.0'  8450.9'    19             30L GoFlo
334  1237  12213#5  6705.9'  8450.4'     6             30L GoFlo
334  1707  12214    6705.3'  8457.2'  1000      4310       CTD
334  1946  12215    6659.4'  84'59.7'  1001      4329     CTD
334  2145  12216    6654.9'  8459.0'   999      4346     CTD
334  2334  12217    6649.9'  8459.6'   998      4356     CTD
335  0143  12218    6645.0'  8459.9'   998      4386     CTD
335  0714  12219    6709.8'  8459.1'   998      4291     CTD
335  1123  12220#1  6716.1'  8500.8'   998      4281     CTD
335  1229  12220#2  6715.6'  8459.6'    60             30L GoFlo
335  1238  12220#3  6715.5'  8459.3'    40             30L GoFlo
335  1244  12220#4  6715.5'  8459.1'    20             30L GoFlo
335  1249  12220#5  6715.5'  8459.1'    12             30L GoFlo
335  1254  12220#6  6715.4'  8459.0'     2             30L GoFlo
335  1431  12221    6720.2'  8459.4'   999      4295     CTD
335  1642  12222    6725.4'  8459.6'  1001      4243     CTD
335  1957  12223    6729.8'  8459.2'  4185      4207     CTD
339  1322  12224#1  6725.6'  8518.3'   301      4220     CTD
339  1351  12224#2  6725.7'  8518.6'    43             30L GoFlo
339  1357  12224#3  6725.7'  8518.7'    28             30L GoFlo
339  1403  12224#4  6725.8'  8518.8'    14             30L GoFlo
339  1408  12224#5  6725.8'  8518.8'     9             30L GoFlo
339  1412  12224#6  6725.8'  8518.8'     4             30L GoFlo
341  0748  12225#1  6727.6'  8600.1'   301      4159     CTD
341  0808  12225#2  6727.5'  8559.9'    46             30L GoFlo
341  0820  12225#3  6727.5'  8559.9'    30             30L GoFlo
341  0830  12225#4  6727.5'  8559.9'    15             30L GoFlo
341  0840  12225#5  6727.5'  8559.9'     9             30L GoFlo
341  0845  12225#6  6727.5'  8559.9'     5             30L GoFlo
342  1310  12226#1  6736.0'  8503.7'   300      4146     CTD
342  1336  12226#2  6735.9'  8503.9'    57             30L GoFlo
342  1343  12226#3  6735.8'  8504.0'    38             30L GoFlo
342  1352  12226#4  6735.8'  8504.1'    19             30L GoFlo
342  1356  12226#5  6735.7'  8504.2'    12             30L GoFlo
342  1401  12226#6  6735.7'  8504.2'     6             30L GoFlo
344  0039  12227    6900.0'  8800.3'   154      3440     CTD
344  0235  12228    6900.5'  8802.1'  3423      3445     CTD
344  2129  12229    6630.6'  8800.1'  4459      4470     CTD
345  1411  12230    6359.9'  8759.4'   295      4761     CTD
345  1745  12231    6358.6'  8759.8'  4710      4763     CTD
346  1110  12232    6129.9'  8759.8'  4781      4868     CTD
347  0550  12233    5900.4'  8800.2'  5012      5023     CTD
347  2143  12234    5630.7'  8759.4'   204      5476     CTD
347  2359  12235    5631.8'  8758.0'  5071      5482     CTD
348  1827  12236    5359.9'  8801.0'  5025      5060     CTD
349  1137  12237    5129.4'  8759.2'  4700      4759     CTD




APPENDIX C.  SEASOAR TOWS


1. Drake Passage transect (66 hours)
                                     Day   Time   Lat. S     Long. W
---------------------------------------------------------------------
Launch                               317   1330   5308.0'   5911.0'
Recovery                             320   0745   6044.5'   5446.6'


2. First survey (125 hours)
                                     Day   Time   Lat. S     Long. W
Launch, northern end of leg A        328   1355   6540.7'   8501.2'
Alter course to NW in heavy weather  329   0200   6709.2'   8459.4'
Begin leg W                          329   0820   6643.1'   8620.4'
Alter course to NW again             329   1059   6703.6'   8614.1'
Resume leg W                         329   1148   6659.9'   8623.8'
Turn to leg X                        329   1924   6758.7'   8626.0'
Begin leg X                          329   2024   6759.7'   8604.8'
Turn to leg Y                        330   0730   6632.5'   8605.3'
Begin leg Y                          330   0843   6631.7'   8542.6'
Turn to leg Z                        330   2018   6756.4'   8541.7'
Begin leg Z                          330   2142   6759.0'   8519.0'
Turn to leg A                        331   0930   6641.9'   8529.4'
Begin leg A                          331   1130   6650.3'   8459.6'
Turn to leg B                        331   2020   6759.5'   8459.0'
Begin leg B                          331   2122   6758.9'   8440.0'
Turn to leg C                        332   0845   6636.4'   8446.0'
Leg C, forced off course by weather  332   1045   6639.2'   8418.7'
'Stormleg'                           332   2007   6747.1'   8339.4'
Recovery                             333   1906   6623.7'   8551.2'


3. Passage tow (5 hours)
                                     Day   Time   Lat. S     Long. W
---------------------------------------------------------------------
Launch                               335   2232   6729.4'   8458.2'
Recovery                             336   0356   6704.4'   8405.5'


4. Second survey (56 hours)
                                     Day   Time   Lat. S     Long. W
---------------------------------------------------------------------
Launch, northern end of leg D        337   0245   6659.7'   8358.5'
Turn to leg C                        337   1651   6835.2'   8400.2'
Begin leg C                          337   1751   6835.9'   8417.2'
Turn to leg B                        338   0620   6700.0'   8420.0'
Begin leg B                          338   0730   6701.4'   8438.4'
Turn to leg A                        338   2132   6859.6'   8441.0'
Begin leg A                          338   2245   6858.2'   8500.1'
Turn to leg Z                        339   1010   6729.3'   8459.6'
SeaSoar lost                         339   1037   6729.0'   8507.9'





APPENDIX D.  UOR TOWS
                                     Day   Time   Lat. S     Long. W
---------------------------------------------------------------------
Launch, leg Z                        340   0518   6737.2'   8520.4'
Recover, leg Z                       340   1327   6822.6'   8519.7'
Launch, leg Y                        340   1700   6900.4'   8532.9'
Recover to fix temp. sensor          340   1754   6900.5'   8552.7'
Relaunch, leg Y                      340   2312   6837.3'   8539.6'
Recover, leg Y                       341   0700   6728.7'   8559.2'
Launch, leg X                        341   0906   6727.9'   8600.0'
Recover, leg X                       341   1733   6848.8'   8600.3'
Launch, leg YZ                       341   2012   6859.0'   8559.9'
Recover, leg YZ                      342   0503   6752.2'   8535.5'




APPENDIX E.  LIGHTFISH TOWS

File              Start              |            End  
       Day  Time  Lat.S     Long.W   | Day  Time  Lat.S     Long.W
-------------------------------------|------------------------------
lf318  318  1233  5551.5'  5752.6' | 319  0000  5718.8'  5701.3'
lf319  319  0000  5718.8'  5701.3' | 320  0000  5954.0'  5522.0'
lf320  320  0000  5954.0'  5522.0' | 320  1300  6113.0'  5441.8'
lf322  322  1503  6255.1'  6110.0' | 323  0000  6315.4'  6354.0'
lf323  323  0000  6315.4'  6354.0' | 323  2331  6342.7'  6906.3'
lf324  324  1105  6405.6'  7328.2' | 324  1457  6409.2'  7356.1'
lf328  328  1711  6558.2'  8500.9' | 329  0000  6652.7'  8502.8'
lf329  329  0000  6652.7'  8502.8' | 330  0000  6730.8'  8603.4'
lf330  330  0000  6730.8'  8603.4' | 331  0000  6744.4'  8519.0'
lf331  331  0000  6744.4'  8519.0' | 332  0000  6739.6'  8437.8'
lf332  332  0000  6739.6'  8437.8' | 333  0000  6737.1'  8428.4'
lf333  333  0000  6737.1'  8428.4' | 333  0926  6633.1'  8454.0'
lf335  335  2159  6725.5'  8456.5' | 336  0000  6720.9'  8447.3'
lf336  336  0000  6720.9'  8447.3' | 336  0300  6705.1'  8410.8'
lf337  337  0257  6700.9'  8400.0' | 338  0000  6748.7'  8421.3'
lf338  338  0000  6748.7'  8421.3' | 339  0000  6847.5'  8500.1'
lf339  339  0000  6847.5'  8500.1' | 339  2142  6819.2'  8456.4'
lf340  340  0000  6817.3'  8508.7' | 340  2351  6830.9'  8539.3'
lf341  341  0000  6830.9'  8539.2' | 341  1258  6803.1'  8600.3'
lf343  343  1859  6821.4'  8628.5' | 344  0000  6859.4'  8757.4'
lf344  344  0000  6859.4'  8757.4' | 344  1530  6711.2'  8805.7'




APPENDIX F.  ACOUSTIC FISH TOWS

  Run             Start              |            End  
       Day  Time  Lat.S     Long.W   | Day  Time  Lat.S     Long.W
-------------------------------------|------------------------------
    1  319  1143  5834.2'  5610.0' | 319  1300  5842.4'  5605.6'
    2  319  1301  5842.4'  5605.6' | 319  1400  5848.0'  5603.8'
    3  319  1400  5848.0'  5603.8' | 320  0700  6041.6'  5448.3'
    4  320  1110  6054.0'  5435.0' | 320  1253  6113.0'  5441.8'
    5  322  0200  6222.5'  5842.3' | 322  1055  6250.1'  6034.5'
    6  322  1340  6251.3'  6049.2' | 323  1040  6327.8'  6615.6'
    7  327  0734  6454.7'  8255,0' | 327  1100  6500.0'  8320.6'
    8  327  1100  6500.0'  8320.6' | 327  1202  6500.4'  8322.1'
    9  327  1400  6500.9'  8328.2' | 328  0100  6500.3'  8500.6'
   10  328  0130  6500.3'  8500.6' | 328  0324  6520.2'  8459.8'
   11  328  0909  6517.7'  8510.8' | 328  1125  6540.2'  8459.5'
   12  328  1400  6540.7'  8501.2' | 330  0608  6642.2'  8603.5'
   13  330  0844  6631.7'  8542.6' | 331  0930  6641.9'  8529.4'
   14  331  1130  6650.3'  8459.6' | 332  0845  6634.0'  8446.0'
   15  332  1043  6639.2'  8418.7' | 333  0922  6635.1'  8452.6'
   16  333  1920  6624.9'  8548.0' | 333  2215  6644.7'  8501.0'
   17  334  0645  6644.1'  8458.7' | 334  0852  6705.1'  8457.8'
   18  334  1757  6705.3'  8455.4' | 334  1857  6659.7'  8500.1'
   19  334  2026  6659.5'  8459.4' | 334  2057  6655.0'  8459.8'
   20  334  2228  6654.9'  8458.4' | 334  2304  6649.9'  8459.8'
   21  335  0224  6644.8'  8500.3' | 335  0618  6710.0'  8500.1'
   22  335  1318  6715.3'  8458.5' | 335  1821  6730.2'  8459.8'
   23  335  2320  6724.0'  8453.4' | 336  0245  6705.1'  8410.8'
   24  336  2315  6638.0'  8334.9' | 337  0130  6657.4'  8357.0'
   25  337  0300  6700.9'  8400.0' | 337  1700  6835.8'  8401.7'
   26  337  1700  6835.8'  8401.7' | 338  0620  6700.0'  8420.0'
   27  338  0632  6659.2'  8422.6' | 338  2230  6859.5'  8458.0'
   28  338  2230  6859.5'  8458.0' | 339  1010  6729.3'  8459.6'
   29  339  1430  6725.9'  8519.0' | 339  2051  6819.2'  8457.1'
   30  339  2345  6818.2'  8502.5' | 340  0500  6737.8'  8520.5'
   31  340  1400  6842.2'  8520.2' | 340  1820  6859.0'  8552.1'
   32  340  2315  6837.3'  8539.6' | 341  0615  6729.5'  8541.4'
   33  341  0916  6727.9'  8600.0' | 341  2000  6900.4'  8530.0'
   34  341  2015  6859.0'  8529.9' | 342  0500  6752.3'  8531.6'
   35  343  1525  6757.0'  8529.2' | 343  2354  6859.4'  8757.4'
   36  344  0406  6859.4'  8804.9' | 344  1212  6739.9'  8759.8'
   37  344  1445  6719.0'  8800.9' | 344  1620  6202.0'  8803.1'
   38  344  2335  6629.8'  8802.1' | 345  0245  6559.8'  8800.7'



APPENDIX G. NON-TOXIC SAMPLING PERIODS

                  Start              |            End  
       Day  Time  Lat.S     Long.W   | Day  Time  Lat.S     Long.W
-------------------------------------|------------------------------
       317  0200  5241.8'  5739.4' | 320  1900  6140.3'  5546.8'
       322  0400  6226.0'  5901.7' | 325  1400  6444.1'  7856.9'
       328  0000  6459.2'  8500.3' | 342  0300  6804.7'  8531.3'
       344  0000  6859.4'  8757.4' | 349  0900  5135.0'  8758.5'




APPENDIX H.  XBT CASTS

XBT no.   Day  Time  Lat.S     Long. W  Depth  Probe 
                                          (m)  type
---------------------------------------------------
xp198012  317  0007  5230.7'  5737.7'   415  T7
xp198016  318  1512  5613.2'  5743.9'  3761  T5
xp198017  319  0203  5734.0'  5653.2'  3620  T5
xp198019  320  0815  6044.5'  5445.5'  3167  T5
xp198020  327  0011  6500.8'  8140.1'  4333  T7
xp198021  327  0252  6459.9'  8207.1'  4313  T7
xp198022  327  0910  6500.0'  8256.8'  3866  T7
xp198023  327  1520  6500.0'  8345.7'  3887  T7
xp198026  344  0053  6900.2'  8800.6'  3482  T7
xp198028  344  0354  6900.1'  8803.4'  3488  T7
xp198029  344  0607  6840.0'  8805.0'  3880  T7
xp198030  344  0759  6819.8'  8803.6'  4450  T7
xp198031  344  1002  6758.9'  8800.8'  4024  T7
xp198032  344  1212  6739.9'  8759.8'  4060  T7
xp198033  344  1418  6720.1'  8759.5'  4390  T7
xp198034  344  1654  6659.6'  8804.0'  4446  T7
xp198035  344  1848  6640.0'  8800.0'  4324  T7
xp198036  344  1959  6630.1'  8759.7'  4500  T7
xp198037  345  0040  6620.0'  8800.6'  4534  T7
xp198038  345  0240  6600.0'  8800.4'  4061  T7
xp198039  345  0436  6539.9'  8759.8'  4616  T7
xp198040  345  0618  6520.0'  8800.0'  4643  T7
xp198041  345  0759  6459.8'  8800.3'  4676  T7
xp198043  345  1039  6439.4'  8804.3'  4700  T7
xp198044  345  1238  6420.0'  8801.4'  4813  T7
xp198045  345  1533  6359.8'  8759.3'  4779  T7
xp198046  345  2138  6338.5'  8800.9'  4800  T7
xp198047  345  2315  6319.7'  8801.0'  4810  T7
xp198048  346  0055  6300.0'  8800.1'  4820  T7
xp198049  346  0248  6239.7'  8801.4'  4829  T7
xp198050  346  0434  6219.9'  8800.5'  4842  T7
xp198051  346  0620  6200.0'  8759.9'  4850  T7
xp198053  346  0818  6139.0'  8800.6'  4600  T7
xp198054  346  1247  6130.0'  8759.7'  4880  T7
xp198055  346  1440  6119.3'  8801.6'  4880  T7
xp198056  346  1642  6100.0'  8800.6'  3920  T7
xp198057  346  1845  6039.1'  8800.1'  4371  T7
xp198058  346  2038  6019.8'  8800.0'  4446  T7
xp198059  346  2232  5959.8'  8758.2'  4900  T7
xp198060  347  0029  5939.5'  8759.4'  5056  T7
xp198061  347  0215  5919.9'  8759.1'  4607  T7
xp198062  347  0717  5900.5'  8800.4'  5041  T7
xp198063  347  0922  5829.2'  8759.1'  5232  T7
xp198064  347  1101  5820.3'  8800.0'  5107  T7
xp198066  347  1300  5759.3'  8800.4'  5175  T7
xp198068  347  1458  5737.8'  8800.6'  5240  T7
xp198069  347  1636  5719.9'  8800.0'  4984  T7
xp198070  347  1824  5659.8'  8759.3'  4081  T7
xp198071  347  2013  5639.9'  8800.8'  4000  T7
xp198072  348  0156  5632.6'  8756.9'  5440  T7
xp198074  348  0335  5619.9'  8800.2'  5000  T7
xp198075  348  0528  5600.0'  8800.0'  4743  T7
xp198076  348  0720  5540.1'  8802.4'  4000  T7
xp198077  348  0911  5520.0'  8759.9'  4613  T7
xp198078  348  1104  5459.5'  8759.3'  4000  T7
xp198079  348  1305  5440.0'  8800.4'  4000  T7
xp198080  348  1459  5419.7'  8759.9'  4601  T7
xp198081  348  2002  5400.1'  8801.4'  5150  T7
xp198082  348  2156  5339.8'  8801.5'  5016  T7
xp198083  348  2335  5320.1'  8801.7'  4996  T7
xp198084  349  0118  5300.0'  8800.2'  4944  T7
xp198085  349  0302  5240.0'  8800.5'  4850  T7
xp198086  349  0449  5219.3'  8800.8'  4680  T7
xp198087  349  0638  5159.9'  8800.5'  5000  T7
xp198088  349  0830  5140.0'  8759.0'  4697  T7
xp198089  349  0945  5129.6'  8759.9'  4500  T7




APPENDIX I.  DAY NUMBER/DATE INTERCONVERSION

Date       Day     Date      Day
--------------------------------
11/11/92   316     1/12/92   336
12/11/92   317     2/12/92   337
13/11/92   318     3/12/92   338
14/11/92   319     4/12/92   339
15/11/92   320     5/12/92   340
16/11/92   321     6/12/92   341
17/11/92   322     7/12/92   341
18/11/92   323     8/12/92   342
19/11/92   324     9/12/92   343
20/11/92   325    10/12/92   344
21/11/92   326    11/12/92   345
22/11/92   327    12/12/92   346
23/11/92   328    13/12/92   347
24/11/92   329    14/12/92   348
25/11/92   330      
26/11/92   331      
27/11/92   332      
28/11/92   333      
29/11/92   334      
30/11/92   335      






DATA QUALITY EVALUATIONS


CTD DQE: 198th cruise of the r/v "Discovery",
         WOCE section SR01 in the South East Pacific
         (Eugene Morozov)

Data quality of 2-db CTD temperature, salinity and oxygen profiles and 
reference rosette samples were examined.  Vertical distributions and 
theta-salinity curves were compared for individual stations using the 
data of up and down CTD casts and rosette probes.  Data of several 
neighboring stations were compared.

The CTD and bottle temperature and salinity measurements is a high-
quality data set with not so much work for DQE.  I would mention only a 
bad SALINITY bottle on station 29660 at 504 db (34.763), I flag it 4.

The oxygen data needs much work.  I find bad calibration for the 
majority of the stations and several bad bottles.  It is very difficult 
to make a quality evaluation when both measurements do not satisfy WOCE 
requirements.

CTDOXY for stations 29632 and 29634 are higher than bottle OXYGEN by 10 
mol/kg in the entire depth range.  CTDOXY for station 2936 is high by 15 
mol/kg in the upper 100 db.  

There were no reference OXYGEN measurements for stations 29638 through 
29658 thus it is difficult to say when the calibration changed for 
better in the upper layer.  I find it acceptable for the upper layer on 
station 29660.  The calibration for stations 29660-29668 is better for 
the upper layer but below 500 db it is higher than norm. 

I flag bad OXYGEN measurements:

              Station 29660  1014 db  high OXYGEN (194.2),  flag 4
              Station 29662    66 db  high OXYGEN (362.4),  flag 4

The calibration for stations 29670, 29672, and 29674 is good.


At the same time there are two  bottle OXYGENs that are not good:

              Station 29670   507 db  high OXYGEN (217.9),  flag 4
              Station 29676  1015 db  high OXYGEN (195.1),  flag 3

The CTDOXY calibration is high for measurements below 2000 db for 
stations beginning with 29678 and to the end of the section.  There is a 
mulfunction of CTDOXY sensor on station 29678 between 850 and 900 db 
wich gave higher values.  These measurements should be flagged as bad or 
Qble.

The bottle OXYGEN (264.6) is bad on station 29694 at 505 db, I flag it -4.

STNNBR  CASTNO  SAMPNO  CTDPRS  CTDSAL  CTDOXY  SALNTY  OXYGEN  QUALT1  QUALT2
------------------------------------------------------------------------------
29660     1       2      504.8                  34.7630          ~~2~    ~~4~
29660     1       1     1014.8                          194.2    ~~~2    ~~~4
29662     1       9       66.1                          362.4    ~~~2    ~~~4
29670     1       2      507.0                          217.9    ~~~2    ~~~4
29676     1       1     1015.1                          195.1    ~~~2    ~~~3
29694     1       5      505.7                          264.6    ~~~2    ~~~4





NUTRIENTS DQE:					
(J.C. Jennings, Jr.)
11 June, 1997

Only a small number of the Discovery cruise 198 stations have nutrient and 
dissolved oxygen data.  It generally does not meet WOCE quality standards.  The 
DI198 data were compared to data from the WOCE S4 and the 1975 ISOS hydrographic 
program to aid in the DQE effort.  The phosphate data are extremely scattered 
and many of the concentrations appear to be too low for this region.  Almost all 
of the nitrate and silicate data were anomalously low for the Southern Ocean and 
we considered it to be of doubtful quality.  Although some of the surface layer 
nitrate and silicate concentrations are certainly possible, the reported 
concentrations below the thermocline are ca. half of what would be expected.  
Many of the nitrate concentrations were evaluated as questionable by the data 
originators and it seems likely that there was a problem with standardization.  
Although there are no accuracy specifications for  nutrient data in the WHP, 
these data are of questionable accuracy.  The precision of all the nutrient data 
save nitrite does not meet WHP specifications.

The dissolved oxygen concentrations are for the most part quite plausible and in 
agreement with oxygen data from the same region reported in the WOCE S4 data 
set.

The following specific problems were identified:

1.) Low phosphate:  All of stations 29632, 29660, 29662, 29664, 29666.
                    Station 29634, bottles 1 and 12
                    Station 29636, bottles 9 and 12
                    Station 29638, bottles 8, 9, 10, and 11
                    Station 29668, all bottles except # 1
2.) High phosphate: Station 29634, bottle  6
                    Station 29674, bottles 1 - 6
3.) Low nitrate:    All of stations 29660, 29662, 29664, 29666, 29668, 29670, 
                                    29672, 29674, 29676, 29678
                    Station 29634, bottles 1 and 3
                    Station 29636, bottles 1, 2 and 3
                    Station 29638, bottles 2, 3, and 4
4.) Low silicate:   All of stations 29660, 29662, 29664, 29666, 29668, 29670, 
                                    29672,29674, 29676, 29678
                    Station 29634, bottles 1 and 3
                    Station 29636, bottles 1 - 4
                    Station 29638, bottles 1, 2, and 3
5.) Low oxygen:     Station 29660, bottle  3
6.) High oxygen:    Station 29660, bottle  3
                    Station 29662, bottle  1
                    Station 29670, bottles 2 and 6
                    Station 29668, bottle  8
                    Station 29674, bottle 10
                    Station 29676, bottles 7 and 8
                    Station 29678, bottle  2

* All figures shown in PDF file.
