WHP CRUISE SUMMARY INFORMATION WOCE section designation A11 Expedition designation (EXPOCODE) 74DI199_1 Chief Scientist(s) and their affiliation Peter Saunders, IOSDL Dates 1992.12.22 - 1993.02.01 Ship DISCOVERY Ports of call Punta Arenas, Chile to Cape Town, South Africa Number of stations 91 Geographic boundaries of the stations 30°13.50''S 00°09.35''W 17°50.72''E 45°04.62''S Floats and drifters deployed none Moorings deployed or recovered none Contributing Authors (In order of appearance) B. A. King S. Bacon P. Chapman S.E. Holley D.J. Hydes D. Smythe-Wright S.M. Boswell D. Price S. Jordan R. Phipps S. Whittle T.J.P. Gwilliam S.R. Thompson R. Marsh M.G. Beney A.J. Taylor K.J. Heywood P.K. Smith S. Cunningham M.P. Meredith V.C. Cornell INSTITUTE OF OCEANOGRAPHIC SCIENCES DEACON LABORATORY CRUISE REPORT NO. 234 RRS Discovery Cruise 199 22 DEC 1992 - 01 FEB 1993 WOCE A11 IN THE SOUTH ATLANTIC Principal Scientist P M Saunders Institute of Oceanographic Sciences Deacon Laboratory, Brook Road, Wormley, Godalming, Surrey, GU8 5UB, UK. Version 2 June 1994 ABSTRACT RRS Discovery cruise 199 was a UK contribution to the World Ocean Circulation Experiment (WOCE) one-time survey, its designation A11. The cruise ports were Punta Arenas, Chile to Cape Town, S. Africa. 91 full- depth stations were worked with a NBIS Mk3b CTD and a GO 24x10 liter rosette water sampler. Salinity, oxygen, silicate, nitrate, phosphate were measured on each station, CFC-11, CFC-12, and CFC-113 measured on every other station and XBT drops (mostly T7) made between stations. Meteorological parameters, sea-surface temperature and salinity, and current profiles to 300m (from a hull-mounted RDI 150 kHz ADCP) were measured throughout the cruise. To improve estimates of the ship's heading (and hence currents) a 3-dimensional gps receiver from Ashtech was employed. Provisional examination of the data indicates that it is of sufficient quality to meet the principal aim of the cruise, namely to determine the exchange of physical and chemical properties between the S. Atlantic and Southern Ocean. Electronic versions of the text of this document, plus hard copy figures are lodged with the WOCE Hydrographic Planning Office, Woods Hole, Mass and with the British Oceanographic Data Centre at Bidston, Merseyside. Keywords ACOUSTIC DOPPLER CURRENT PROFILER (ADCP) A11 WOCE ONE-TIME SURVEY CFC 11,12,113 CORE PROJECT 1 CTD OBSERVATIONS "DISCOVERY"/RRS - CRUISE (1992-3) 199 NUTRIENTS OXYGEN WOCE WHP Cruise and Data Information CONTENTS 1 CRUISE NARRATIVE 1.1 Highlights 1.2 Cruise Summary 1.3 List of Principle Investigators 1.4.1 Scientific Programme and Methods 1.4.2 Preliminary Results 1.5 Major Problems Encountered on the Cruise 1.6 Other Observations of Note 1.7 List of Cruise Participants 2 MEASUREMENT TECHNIQUES AND CALIBRATIONS A general note on data quality checking 2.1 Sample salinity measurements 2.2 Sample oxygen measurements 2.3 Nutrients 2.4 CFC-11, CFC-12, and CFC-113 2.5 Samples taken for other chemical measurements a) Oxygen and Hydrogen isotope ratios b) Iodine 2.6 CTD Measurements a) Gantry and Winch Arrangements b) Equipment, calibrations and standards c) CTD Data Collection and Processing 2.7 XBTs 2.8 Acoustic Doppler Current Profiler (ADCP) 2.9 Navigation a) GPS-Trimble b) Electromagnetic log and gyrocompass c) Ashtech GPS3DF Instrument 2.10 Underway Observations a) Echosounding b) Meteorological Measurements c) Thermosalinograph measurements d) Satellite Image Acquisition and Processing 2.11 Shipboard computing 2.12 Cruise diary COMMENCEMENT OF THE A11 SECTION (45°S, 60°W) THE TURNING POINT ON THE A11 SECTION (45°S, 15°W). END OF A11 SECTION Acknowledgements CTD STATION LIST XBT STATION LIST FIGURE LEGENDS FIGURES 1-20 DQE Reports CTD Nutrients Data Status Notes 1. CRUISE NARRATIVE 1.1 Highlights Expedition Designation: WHP One-time Survey, A11 Chief Scientist: Peter M Saunders, IOSDL Ship: RRS Discovery, newly lengthened to 90.2m Ports of Call: Punta Arenas, Chile to Cape Town, S. Africa Cruise Dates: December 22, 1992 to February 1, 1993 1.2 Cruise Summary Cruise Track The cruise track and station locations are shown in Figure 1: only small volume samples were taken. Sampling The following water sample measurements were made:- salinity, oxygen, total nitrate, phosphate, silicate and CFCs 11,12 and 113, the freons on alternate stations. CTD salinity and oxygen were also measured. The depths in m sampled were:- 5(10), 50, 100, 150, 200, 250, 350, 500, 750, 1000, 1250, 1500, 1750, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 6000 meters. Number of Stations A total of 91 CTD/rosette stations were occupied using a General Oceanics 24 bottle rosette equipped with 24 10-litre Niskin water sample bottles, and a NBIS Mk IIIb CTD equipped with a SensorMedic oxygen sensor, Sea Tech Inc 1 m path transmissometer, Simrad altimeter model 807-200m, and IOSDL 10 kHz pinger. Floats, Drifters, and Moorings No floats, drifters, or moorings were deployed on this cruise. Reporting Electronic versions of the text of this document, plus hard copy figures are lodged with the WOCE Hydrographic planning office, Woods Hole, Mass and with the British Oceanographic Data Centre at Bidston, Merseyside. We plan to lodge electronic copies of most of the data from the cruise at these same sites by the end of 1993. 1.3 List of Principle Investigators The principal investigators responsible for the major parameters measured on the cruise are listed in Table 1. The responsibility for all tasks undertaken on the cruise will be found in table 2. TABLE 1: PRINCIPAL INVESTIGATORS NAME RESPONSIBILITY AFFILIATION B. King CTD IOSDL S. Bacon Salinity JRC D. Hydes Nutrients IOSDL P. Chapman Oxygen Texas A & M D. Smythe-Wright CFC JRC P. Saunders ADCP IOSDL P. Smith Meteorology IOSDL S. Thompson XBTs IOSDL M. Meredith Satellite imagery UEA (MACSAT) and thermosalinograph 1.4.1 Scientific Programme and Methods The principal objectives of the cruise were:- a) To estimate the exchange of heat, freshwater, nutrients and freons across the section, i.e. between the Southern Ocean and the South Atlantic b) To determine the water mass characteristics on the section and to determine whether and where secular changes are found, and c) To submit to the WHPO a data set, a fit companion to other WHP one time survey cruises, and thereby contribute to the global measurements necessary to meet the objectives of WOCE. The principal instruments employed in the measurement programme consisted of a NBIS Mk IIIa CTD and General Oceanics rosette mounted within a tubular aluminum frame of dimensions 1.8m height x 1.5m diameter. The package was weighted to give a free fall speed in excess of 2 ms-1. Subsidiary instrumentation consisted of a 1m transmissometer, altimeter (with 200m range for bottom finding) and 10 kHz location pinger. Four of the rosette bottles were fitted with SIS digital reversing thermometers (6) and pressure meters (2). The wire was a single conductor 10mm steel rope manufactured by Rochester Cables, and the winch was of traction winch design built by Kley France. A complex folding gantry of RVS Barry design ensured the virtually automatic launching and recovery of the CTD/rosette package in all conditions within which the ship could be safely operated. After a cast the rosette was placed on deck and secured, the rosette pylon was drenched in fresh water and the CTD sensors covered with protective housings. Subsequently digital instrumentation was read and freon samples were drawn followed in order by samples for oxygen, nutrient and salinity analysis. The rosette was stored on deck throughout the cruise and all sampling was performed there. In moderate weather the rosette would be pushed forward on a railway about 3 m to obtain further shelter. In rain umbrellas could be clamped to the rosette frame in order to protect the samples and in rough seas the ship remained on station until sampling was completed. Other and, in some cases, crucial additional measurements were made throughout the cruise. XBTs were launched between CTD stations and more frequently in the slope regions at each end of the cruise section. Acoustic Doppler Current Profiler (ADCP) measurements were made continuously employing a hull mounted 150 kHz unit manufactured by RDI. In support of the ADCP measurements a GPS3DF receiver manufactured by Ashtech, Inc provided heading information superior to that of the ship's gyro. Underway measurements of surface temperature and salinity were made by a FSI thermosalinograph and a Simrad 500 Echosounder provided continuous water depth measurements. Other navigation information was supplied by a Trimble GPS receiver and all data were logged by networked SUN workstations with terminals widely available in the main and computer labs. A description of the methods of measurement, calibration and analysis of the data received from these various sources will be found in section 2 of this report. 1.4.2 Preliminary Results Figure 2 shows the distribution of sample observations made on the A11 section. Since data from the South Atlantic Ventilation Experiment (SAVE) were available on the ship (thanks to WHPO), we were able to compare A11 and SAVE sample data. The property distributions were very similar, but small differences were noted in the deep water which became evident with potential temperature < 1.0°C or salinity in the range 34.66 - 34.72. A11 salinity measurements agreed well with the SAVE 5 leg data, but were more saline by 0.002 than adjacent SAVE 4 data: the differences amongst the SAVE data were not previously known to us. Nitrates showed agreement with both SAVE 4 and 5 measurements, but at the deepest levels silicates and oxygens were slightly lower by 2.5 µmol/kg (Figure 3) and 2.5 µmol/kg (Figure 4) respectively; phosphates were lower by about 0.08 µmol/kg. These preliminary results, whose magnitude but not sign depends on which historic set is compared, apply principally within the Argentine Basin, and possible causes of the differences are under investigation. A more unexpected result, which owed nothing to the accuracy of the measurements, was the extreme northern position of the Subtropical Convergence on the NE leg of the track (Figure 1). Although the water became progressively warmer along this leg, the surface salinity remained below 35 until a ring was encountered centered on 36°20'S and 4°00'E. The ring had a thermostad of temperature 13.5°C, salinity 35.2 and a maximum depth of 600m. An anticyclonic circulation of 30 cms-1 was observed by the ADCP. It may have been an Agulhas ring which had over-wintered south of the convergence, or a Brazil Current ring shed in the WBC retro-flexion zone which had migrated eastward. Opinions in the scientific party were split about equally, but a closer post-cruise examination of the data may well resolve the question. Beyond its NE edge, near 35°40'S and 5°00'E we encountered the subtropical gyre, with a surface salinity exceeding 36 and temperature of 20°C. This observation appears to confirm Deacon's (1937) assertion of the northward migration of the convergence in summer in this region. Within the subtropical gyre a second hydrographic feature was encountered. This was defined by two hydrographic casts and 5 XBTs and was centered at 33°30'S, 9°45'E and extended for 300 km along the track. Within it, the 15°C isotherm plunged to a depth of 250m, while outside it the same isotherm was nearer a depth of 100m. An anticyclonic circulation was measured by the ADCP with currents approaching 75 cms-1. This was undoubtedly a recent Agulhas ring. The ADCP instrumentation furnished, we believe, important new data on the cruise: it functioned incomparably better than when installed on the previous 10m-shorter version of the ship. The most important results derived from it were found in the western boundary region. On the Argentine Slope, on two crossings of the Falklands Current, large and persistent northward velocities were found at 100m depth (30 - 50 cms- 1). These were considerably in excess of those predicted by the geostrophic shear (relative to the bottom), and consequently bottom velocities of 15 - 30 cms-1 are inferred. The consequences for transport in the WBC and exchange across the section are considerable. On the South African slope, along-slope velocities were also observed on a crossing of the Benguela Current. However these were quite small and variable in direction and a preliminary analysis suggested they were dominated by transient (tidal or inertial) components. Also of note were ADCP observations made in a storm near 45°S 21°W: winds approached 30 ms-1 for a brief period, and striking inertial oscillations (circa 40 cms-1) were recorded. Since meteorological measurements were made aboard the ship, it is hoped that given the high quality of the ADCP data, it may prove possible to deduce the integrated Ekman drift on this cruise. 1.5 Major Problems Encountered on the Cruise Two GO rosettes were available and both were utilized. Misfiring and double tripping were initially widespread, but when their sensitivity to the lanyard tension was recognized it became possible to reduce them to acceptable levels. Nevertheless a post-cruise review estimates the overall number of double trips as nearly 10% of the total number of samples. Thus a larger than expected number of duplicate samples was achieved. It is our recommendation and intention for the future that lanyard tensions be measured, monitored and set to a value which allows a properly reliable operation of the unit. As mentioned in Section 1.4.1 the winch was of complex traction winch design; it was put to use only on the previous cruise and because of its newness, inevitably there were difficulties. On the 1st of January at 0600, control failure occurred: it was approximately 36 hours before the fault was identified, the electronic component replaced and control settings optimized to allow station work to proceed. The efforts of all involved deserve recognition and thanks. Although we believe this was a unique situation, a different problem occurred twice and was potentially liable to occur anytime there was a large swell. Because the CTD/rosette takes time to shed air from all its component parts, very close to the surface it is vulnerable to heavy swell: it may 'float'. In such circumstances the wire goes slack, and on both occasions the wire jumped out of a sheave pair at the foot of the gantry (where the wire direction changed from horizontal to vertical). Even in the short term this is probably a rectifiable fault, but on the cruise it cost us 4 hours both times it occurred. Concerning the instrumentation for analysis, two problems were noted. Early on, the SIS unit for determination of oxygen concentration became unreliable: the photometric end point detection system was no longer stable. Fortunately a backup amperometric system, the Metrohm 686 titroprocessor, was available, and this was used for the bulk of the cruise measurements. The CFC measurements also experienced difficulties which led to the loss of some data. Shortly after the start of the cruise the CFC-12 measurements exhibited severe contamination which was believed to be due to the accidental release of oil from the ship and its capture in the non-toxic seawater system used to store the sample syringes. To bypass this problem, syringes were stored in surplus sample water, a practice however, which did not eliminate the contamination. Early CFC-12 measurements may be expected to be of lower quality than expected on the cruise, but the CFC-11 and CFC-113 measurements should be unaffected. 1.6 Other Observations of Note On the 16th January, a large iceberg was sighted: its location was determined as 44°50'S 14°22'W. In view of a much more southerly position and crossing of the Falkland Current three weeks earlier in the cruise, this was an odd location to observe one for the first time. On the 19th January in about 3700m of water, RRS Discovery passed over a flat-topped seamount near 40°48'S 5°40'W: it is not recorded on the GEBCO chart and its minimum depth was near 750m. We propose the name New Discovery Seamount for this 3000 m high feature. 1.7 List of Cruise Participants The members of the scientific party are listed in Table 2, along with their responsibilities. TABLE 2: CRUISE PARTICIPANTS NAME RESPONSIBILITIES AFFILIATION S. Bacon Salinity JRC M. Beney Data acquisition RVS S. Boswell CFCs JRC P. Chapman Oxygens, nutrients Texas A & M V. Cornell Data archiving, Macsat JRC N. Crisp CTD operations IOSDL S. Cunningham CTD/sample analysis JRC P. Gwilliam CTD operations (IC) IOSDL V. Gouretski ADCP/historical hydrography UEA K. Heywood CTD/sample analysis UEA S. Holley Oxygens, nutrients JRC D. Hydes Nutrients, oxygens IOSDL S. Jordan Mech. Eng (IC) RVS B. King CTD/sample analysis IOSDL R. Marsh ADCP JRC M. Meredith Thermosalinograph, Macsat UEA D. Price CFCs JRC R. Phipps Mechanical Engineer RVS P. Saunders PSO, ADCP IOSDL P. Smith CTD operations, Meteorology IOSDL D. Smythe-Wright CFCs (IC) JRC A. Taylor Electrical Engineer RVS S. Thompson GPS, XBTs IOSDL S. Whittle Mechanical Engineer IOSDL Abbreviations IOSDL Institute of Oceanographic Sciences, Deacon Laboratory - Wormley JRC James Rennell Centre - Southampton RVS Research Vessel Services - Barry UEA University of East Anglia - Norwich IC In charge of 2 MEASUREMENT TECHNIQUES AND CALIBRATIONS A general note on data quality checking (Oct 93) by: B. A. King Note that a number of sections on data quality checking have been added to this report (the .DOC file kept by the WHPO) since the submission to the WHPO of the initial cruise report in February 1993. Such additions are identified with dates in the subheadings. The consequence of maintaining a single report file is that some figures are introduced out of order, and some information may appear more than once in the text. One problem when looking for small differences between two profiles of sample data for example between adjacent stations in a single data set or a comparison of data from different cruises, is that the size of any difference is likely to be smaller than the variation of the property over a few hundred meters in the vertical. This combines with the fact that the samples are not necessarily collected at the same vertical coordinate (usually pressure or potential temperature) to create something of a difficulty. However, the following procedure has been found to be a useful way round this problem, both for checking the internal consistency of the data set and in the comparison with historical data. (i) The deep data are plotted in a theta-property plot, and a fraction of the data selected which are closely described by a linear regression of the sample value on potential temperature. This invariably led to different regressions for the western and the eastern basin. Typically, the western basin regression would be calculated from data with theta < 1.0 degree, and the eastern basin regression from data with theta < 1.2 degrees. (ii) For each sample value, the chosen regression is used to compute a 'predicted' value of the sample, and the anomaly between the observed value and this predicted value is calculated. If the data are well described by a linear fit with theta, these anomalies should be small, probably an order of magnitude smaller than the variation in the vertical of the fitted data. (iii) There are now a number advantages: first, it is now straightforward to compare samples collected at different depths, by comparing their anomalies; second, any offset between profiles of a magnitude greater than the normal scatter in the anomalies is immediately apparent; third, the mean value of the anomalies for a station provides a simple and objective way to summarize the property value for that station in a single number. The key to this technique is to use the same prediction for every station being considered for inter-comparison. For comparisons between cruises it is not particularly important which data set is used to determine the fitting equation, so long as it removes the background distribution in each data set. We have used linear fits based on the present data. Comparison with historical data (Oct 93) In the course of assessing the quality of the present data, comparisons have been made with data from the following cruises. Station positions are shown in Figure 8 using these symbols: Present Cruise, WHP A11: 'pluses' SAVE leg 4: 'crosses' AJAX (N-S section on 1 east): 'inverted triangles' Atlantis II cruise 107 (W-E section on 46 south): 'triangles' All SAVE 4 data have been considered, and only extracts from the AJAX and Atlantis II-107 data. Analysis of the deep data from SAVE 4 shows gaps for the central stations; these were shallower stations while crossing the Mid-Atlantic Ridge. Data from the western basin have been compared where potential temperature is cooler than 1.0°, and eastern basin data when potential temperature is cooler than 1.2°. Duplicate analyses from multiple trips of Niskin bottles (Oct 93) From time to time throughout the cruise, there were casts on which the multi-sampler had problems in tripping Niskin bottles correctly. This could result in either zero or two bottle closures for one trigger signal. While this unreliability was a nuisance in some respects, and led to quite a lot of careful scrutiny of sample analyses to sort out the depths at which bottles had closed, it had the advantage of providing a number of duplicate samples for all the tracer analyses. While these are not quite independent duplicate samples, in the sense that they were generally analyzed in the same run by the same analyst, they were more independent than replicate samples drawn from the same Niskin bottle. Furthermore, the fact that they were duplicates will have been unknown to the analyst at the time the analysis was performed. The total number of such duplicates for which the salinity, oxygen and three nutrients are all good is 198 (out of 1642 samples with all tracers good); i.e. about 12% of the total number of samples. Out of these 198, 87 are from depths greater than 3000 meters. The mean and standard deviations of these five tracers (198 samples) is as follows (units are µmol/kg except for salinity, percentages of full-scale in brackets): standard deviation salinity 0.0017 (0.0009 for pressures > 3000) oxygen 0.86 (0.3%) nitrate 0.15 (0.4%) phosphate 0.026 (1%) silicate 0.30 (0.2%) For the 87 samples from pressures greater than 3000 decibars, the statistic for salinity is better than for the full set; this is a reflection of the greater homogeneity of the water column there. The statistics for the other tracers are not significantly different. 2.1 Sample salinity measurements by: S. Bacon On RRS Discovery cruise 199 the salinity analysis of samples was carried out exclusively on the IOSDL Guildline Autosal salinometer model 8400, modified by addition of an Ocean Scientific International peristaltic- type sample intake pump. The instrument was operated in the ship's constant temperature laboratory at a bath temperature of 24°C with the laboratory set to 20.5°C. This difference in temperature was larger than normally employed and only arose through a misunderstanding, but was allowed to remain rather than disturb the salinometer again when it became clear that the machine was quite 'happy' operating thus. Standardization was effected by use of IAPSO Standard Seawater batch P120, of which 110 ampoules were consumed. Two of these were imperfectly sealed, and were discarded; two were evidently of incorrect (too high) salinity, and one more was thought dubious. These latter three were not used as standards. The standardization history of the salinometer has been constructed, in which standardization drift is represented as equivalent salinity (ES) change referenced to the first standard measurement of the cruise. The instrument was remarkably stable, not changing from its initial standardization by more than 0.001 ES until the last ten days of the cruise, when the seas generally were calmer and the outside temperature increased, although it is difficult to associate such changes in external conditions with the observed behavior of the salinometer, unless the ship's power supply is implicated in some way. Excluding the two bad standards, the mean standardization drift was 0.0007 ES, with a standard deviation of 0.0007 ES, for 108 standards. There were 46 pairs of replicate (i.e. from the same rosette bottle) samples drawn; and 210 pairs of duplicate (i.e. from different rosette bottles fired at the same depth) samples. Of the duplicate pairs, 87 were from below 3000 m. The standard deviations of the three groups of sample pairs are given in table S1 below. TABLE S1 Salinity replicate and duplicate statistics QUANTITY STANDARD NUMBER DEVIATION OF PAIRS Duplicates 0.0019 208 Duplicates 0.0009 87 (from >3000m) Replicates 0.0008 46 See text above table for the distinction between replicates and duplicates. Reconciliation with CTD data, and data quality control (Oct 93) by: B. A. King Salinity samples values reported by the analyst were considered for data quality flagging according to three criteria: a) The analyst may have marked the sample as suspect or bad if the analysis was unsatisfactory in some way. b) Sample values were compared with those from neighboring stations in property-property plots. It was found that the salinity samples could be described by S = 34.6760 + 0.04746 x theta for theta < 1.0 in the western basin, and by S = 34.6762 + 0.08052 x theta for theta < 1.2 in the eastern basin. Note in passing that the regressions for the two basins intersect at a salinity of 34.676 and at a potential temperature indistinguishable from zero degrees. The sample salinity anomalies (for theta < 1.0 and theta < 1.2 in the two basins) have been calculated relative to these regressions and averaged for each station. The result is shown in Figure 9. Station 12296 appears to be somewhat different from the others, but was the last station occupied in the western basin before encountering the mid- atlantic ridge. Although the deep water at 12296 is slightly more saline than the preceding stations, it is still much fresher (order 0.04) than the eastern basin stations. c) Having established the station-to-station consistency, individual bad samples were sought by comparing sample values with calibrated CTD salinity values. Note that samples with large residuals had already been rejected from the CTD calibration procedure, but not yet flagged as suspect. The rms of the residuals was 0.001 for 430 samples at depths greater than 3000 meters. Of these, 407 samples had residuals smaller than 0.002. All samples with residuals greater than 0.005 were then inspected on an individual basis, and a reason sought for the large residual. Mostly these were traced to regions where there is a strong vertical gradient in salinity. Many cases were found where the sample salinity corresponded to the CTD salinity measured a few meters deeper than where the winch was stopped and the Niskin bottle closed. It is therefore concluded that the 'flushing distance' for the Niskin Bottle is of the order of five meters. Commonly, the residual was 2 meters times dS/dz, the vertical Salinity gradient per meter. dS/dz could be up to 0.005 per meter; some residuals were as large as 0.020. In these cases, the sample salinity flag was left as 2, there being no reason to doubt either the correctness of the drawing of the sample, nor the accuracy of the analysis. Examples of large residuals are sample numbers 26622, 27823 The majority of other cases of large residuals occurred when the upcast CTD salinity was noisy for some reason: for example, when the ship was rolling and the CTD was in a significant salinity gradient. Again, in such cases the sample flag was left as 2 so long as there was no other reason to flag the sample as suspect. In some cases, where the CTD salinity seemed to be good, and no reason could be found for there to be a large residual, the sample was flagged as suspect or bad. The residuals for all samples flagged as good are plotted against pressure in Figure 10. (Stations 12251-12255 and 12325 are excluded from this figure. This is because of particular uncertainties in the CTD data for those stations; this is discussed in detail in the section on CTD data.) Note the quite large residuals in the upper 500 m which arise mainly from the Niskin flushing problem. Note also that there is a small but perceptible systematic variation in residuals. This is of order 0.001 or less at depths greater than 1500 meters. This could arise from the flushing problem, or some residual behavior of the CTD salinities. It is considered to be sufficiently small that it can be ignored, so it remains uncorrected in the CTD data. Comparison with historical data (Oct 93) Figure 11 shows the anomaly of the SAVE leg 4 salinities (station averages) with respect to the standard fit; SAVE leg 4 data are seen to be generally fresher, on average by 0.0015 to 0.002. However, at the intersection of our cruise with SAVE leg 5, the deep salinity data were found to be in agreement. Figure 12 shows the anomaly of the Atlantis II salinities, which are slightly higher than ours. However, the discrepancy is not quite as high as it appears from the figure, which shows station averages and is therefore susceptible to individual large anomalies: the mean anomaly for 69 deep samples is 0.0025. Note in passing that Figure 9 also shows the trend in the deep theta-S relation across the western basin as observed on the present cruise: 0.0035 in salinity across 40 stations. The rms of the station averages about the trend is 0.0009. Conclusion The salinity sample data are believed to be of a high standard, with good precision and internal consistency. Although there are biases with respect to some other fairly recent historical data, we see no reason to doubt the absolute accuracy of our data. We note for emphasis that all our samples were calibrated with respect to batch P120 of Standard Seawater. 2.2 Sample oxygen measurements by: P. Chapman, S.E. Holley and D.J. Hydes Equipment and techniques Bottle oxygen samples were taken in calibrated clear glass bottles immediately following the drawing of samples for CFCs. The temperature of the water at the time of chemical fixation was measured to allow corrections to be made for the change in density of the sample between the closure of the rosette bottle and the fixing of the dissolved oxygen. Analysis followed the Winkler whole bottle method. The thiosulphate titration was carried out in a controlled environment laboratory maintained at temperatures between 21 and 22°C. Thiosulphate normality was determined on a daily basis and whenever new reagents were made up. Duplicate samples were taken on every cast; usually these were from the deepest four bottles. For the early stations, the end point was determined by an automatic photometric method manufactured by SIS (Germany). After station 12253, however, the instrument began giving erroneous endpoint readings since a distinct yellow colour was sometimes still visible in the titration flasks. This was not consistent, and some analyses within each run appeared to titrate correctly; however, all samples from stations 12253, 12254, 12255, and 12257 have been flagged as suspect. For stations 12258 to 12337, i.e. the bulk of the cruise, an "amperometric titration to a dead stop" following the method of Culberson and Huang (1987) was used. A Metrohm Titrator and a Dosimat 665 (10 ml) automatic burette was employed. Titration volumes in deep waters were approximately 5 ml and the smallest increment from the burette was 2 microlitres. The volume of oxygen dissolved in the water was converted to mass fraction by use of the factor 44.66 and an appropriate value of the density; corrections for the volume of oxygen added with the reagents and for impurities in the manganese chloride were also made as described in the WOCE Manual of Operations and Methods (Culberson, 1991). Reproducibility of measurements Approximately 1900 samples were taken during the cruise; in addition, a large number of duplicates were analyzed. Statistics on the duplicates are given in Table O1. These include both duplicates taken from the same bottle (replicates) and those taken from different bottles fired at the same depth and invariably unknown to the analysts. While the photometric method was being used, 22 samples were taken from separate bottles all fired at a depth of 2500 m at station 12240 (Table O1). The data gave a standard deviation of 0.63 µmol, or 0.3%. However, 12 pairs of duplicates taken from the same bottle for stations 12250-12256 gave a mean difference of 1.2 µmol with a standard deviation of 1.29 µmol (approximately 0.56%, Table O1). Duplicates from 223 pairs of samples taken from the same bottle later in the cruise while the amperometric method was in use had a mean difference of 0.64 µmol, and standard deviation of 0.85 µmol, while 13 samples from 5455m from station 12277 gave a standard deviation of 0.35 µmol (0.15%, Table O1). A further series of multiple samples was taken from different bottles fired at the same depth as a result of double trips by the rosette. The results of these are also given in Table O1. The mean difference for 166 sets taken over all depths and analyzed by the amperometric method was 0.57 µmol; the standard deviation of the differences was 0.65 µmol. These figures are not significantly different from duplicates taken from the same bottle (replicates). Comparisons with historical data Data taken at on this cruise on stations 12271-12274, 12282-12286, and 12296-12299 were compared SAVE stations 289-293, 260-264, and 200-203 respectively. Additionally, stations 12313-12316 were compared with data obtained at AJAX stations 46 and 47 near the Greenwich meridian. Some of this is shown in Figs. 3 and 4. Apart from difference in the near surface data resulting from changes in water masses in the area, there is a large measure of agreement. However, at the deepest levels the present cruise data at a given potential temperature (or salinity) shows an offset of between 2 and 6 µmol kg-1, in all cases less than the historic data. We are currently investigating the cause of these offsets. References CULBERSON, C.H. and S. HUANG, 1987. Automated amperometric oxygen titration. Deep-Sea Research, 34, 875-880. CULBERSON, C.H. 1991. 15 pp in the WOCE Operations Manual (WHP Operations and Methods) WHPO 91/1, Woods Hole. TABLE O1 Statistics of duplicates and replicates obtained by both the photometric and amperometric methods. Sample depths are given where appropriate. stn(s) number depth(s) oxygen concentration µM/kg m mean (diff) std dev %mean Photometric method 12240 22 2500 208.5 0.63 0.3 12250-56 12 all 1.2 1.29 0.56 Amperometric method 12277 13 234 230.1 0.35 0.15 12258-337 223 all 0.64 0.85 0.40 12258-337 166 all 0.57 0.65 0.30 Reconciliation with CTD data, and further data quality control (Oct 93) by: B. A. King Oxygen samples were assessed for data quality and data quality flagging in the following manner: a) The analyst may have flagged the sample as suspect or bad. b) The data were plotted in station groups, with both pressure and potential temperature as the vertical coordinate. This enabled outliers to be identified and investigated. Very commonly, some other evidence was found which resulted in a flag of suspect or bad. However, samples were not flagged as suspect solely because they were outliers. c) Sample values believed to be good were used for calibration of CTD oxygens, as described elsewhere. Residuals between sample oxygens and CTD oxygens were then calculated and inspected on a sample by sample, station by station, basis. On the basis of this inspection, a small number of samples previously marked as suspect were promoted to good. More commonly, samples were downgraded from good to suspect, or suspect to bad. It was recognized that in certain parts of the water column, particularly where vertical gradients were strong, quite large residuals could genuinely arise. These could arise from a number of sources, including the following i) the Niskin Bottle flushing length, discussed in the salinity section ii) the relatively slow response of the CTD sensors iii) mismatch between oxygen samples collected on the upcast, and CTD oxygen values collected on the downcast (see the discussion in the CTD section) Samples with large residuals (>5 µmol/kg) were permitted to retain a good flag if it was believed that one of these effects was responsible for the size of residual. d) Sample numbers for which other tracers had been found to be suspect (especially nutrients) were given special scrutiny in oxygen, and vice-versa, and flags adjusted where necessary. Final reconciliation with CTD data (Oct 93) After the data quality procedures had been completed, the CTD oxygens were re-calibrated using, in general, only data flagged as good. However, there were some exceptions. For stations 12253-12257, there were not enough good data (see the analysts' discussion above); accordingly those stations were calibrated using data flagged as suspect. The list of suspect (flag 3) sample numbers used in CTD calibration is as follows: 25301, 25302, 25303, 25304, 25305, 25307, 25308, 25309, 25310, 25312, 25313, 25316, 25317, 25318, 25319 25401, 25402, 25403, 25404, 25406, 25407, 25408, 25410, 25411, 25412, 25413, 25416, 25417, 25419 25501, 25502, 25503, 25504, 25505, 25506, 25507, 25508, 25509, 25510, 25511, 25512, 25513, 25514, 25515, 25516, 25517, 25518, 25519, 25603 25701, 25702, 25703, 25704, 25706, 25707, 25708, 25710, 25711, 25712, 25713, 25714, 25715, 25716, 25717, 25718, 25719, 25720, 25721, 25722 Similarly, there are sample data believed to be good, which were unsuitable for use as CTD calibration samples, mainly because of the reasons given in (c) above. The following good (flag 2) samples were excluded from the CTD calibration: 25824 25914, 25915, 25924 26622 26720 27230 27736 27921 29428 30119, 30120 30213, 30218, 30219 30322 30520, 30521 30614, 30615, 30619 30720 30820 31119, 31120 32117 33210, 33214 33315 Finally, the CTD calibration sometimes lacked a good sample near the surface (for example on stations 12269 and 12270, where there were multi-sampler problems). In these cases, plausible near-surface sample values were 'invented', solely for the purpose of CTD calibration, and based either on neighboring stations or slight over-saturation (2%) of near-surface water. The list of sample numbers for which this was done is as follows: 25108, 25109 26010, 26011, 26012 26914, 26915, 26916 27013, 27014, 27015 Summary of sample minus CTD residuals (Oct 93) The residuals between all samples eventually flagged as good, and the CTD oxygens, are summarized in Table O2: TABLE O2 Residuals of sample-CTD oxygens, averaged into 500 meter depth bins. PRESSURE MEAN STD DEV # IN SAMPLE >6000 -1.41 0.49 4 5500-6000 -2.34 1.12 19 5000-5500 -0.73 1.16 93 4500-5000 -0.11 1.65 70 4000-4500 0.67 1.55 72 3500-4000 0.54 1.83 79 3000-3500 1.14 1.46 83 2500-3000 -0.01 1.80 75 2000-2500 0.71 2.00 147 1500-2000 0.65 1.94 165 1000-1500 -1.10 1.73 165 500-1000 -0.98 2.60 175 0- 500 0.28 3.30 532 All 0.03 2.66 1686 All>3000 0.14 1.73 420 Note that 1679 out of 1686 samples have a residual smaller than 10 µmol/kg. Temperature used for converting µmol/l to µmol/kg (Oct 93) Requirement: Oxygen concentrations were reported by the analysts in µmol/l, and need to be converted to µmol/kg by introducing the density of the water at the time when the oxygen fixing reagents were added on deck. The density is computed from the sample salinity and an estimate of the temperature at time of fixing. Note that for a salinity of 35, 0.1% in density is equivalent to 4° at 20°C and 8° at 2°C. We should therefore aim to get the temperature at time of fixing correct to about 2° or 4°. An attempt was therefore made to measure the temperature of the oxygen sample at the time that the oxygen fixing reagents were added on deck. This was done by flushing a spare sample bottle with water from the Niskin Bottle, and measuring the temperature of the sample with a PRT; temperatures were recorded for 80% of the oxygen samples drawn. These temperatures are reported as OXYTMP in the .SEA file. For deep samples, OXYTMP is always warmer than THETA, the CTD potential temperature measured at the time the Niskin Bottle is closed. This is what would be expected. However, it was found that for many shallow samples, especially in the eastern basin where sea surface temperatures could be as high as 20 degrees, OXYTMP was cooler than THETA. On some occasions, this could be traced to night-time stations where the air temperature was up to 4 or 5 degrees cooler than SST; on other occasions there was no apparent reason why OXYTMP should be any cooler than THETA, so the observations remain as a mystery. We therefore conclude that these apparently improbable values result from inconsistent or otherwise inadequate procedure for measuring OXYTMP. For example, the probe may have been permitted to be subject to evaporation, or incomplete temperature equilibration. This procedure will be investigated further on subsequent cruises. Note in passing that during the cruise, the probe used to measure OXYTMP failed. After repair, it was calibrated against a SIS digital reversing thermometer at 20 points between zero and 30°. The resulting linear calibration had residuals of no greater than 0.1°. In reaching a final decision on which temperature to use for converting volume to mass units, there are thus two main considerations: a) OXYTMP is unavailable for about 20% of samples. This includes a series of stations in mid-cruise (12272-12277) between the failure of the probe and the introduction of the repaired probe. It is necessary to use some method for creating OXYTMP for samples where it was not measured. b) We have some reservations about the reliability of individual OXYTMP measurements. It was therefore decided to use a simple function of THETA to predict the OXYTMP used for data conversion, this function being based on the observed OXYTMP values. This has the advantages of providing a complete set of OXYTMPs, and removes the vulnerability to a single poor temperature determination on deck. The chosen fit was THETA > 12 : OXYTMP = THETA THETA < 12 : OXYTMP = 3.612 + 0.699 x THETA The coefficients in the regression equation are the least squares fit to 1296 samples with THETA < 12, constrained to pass through OXYTMP=THETA=12 degrees. Thus OXYTMP was found to be about 3.5 degrees warmer than THETA when THETA was near zero. The residuals of 'measured' OXYTMP about 'predicted' OXYTMP are shown in Figure 13 (measured minus predicted), where they are plotted against THETA. We are satisfied that the resulting predictions are adequate for converting the oxygen units. For THETA cooler than 12 degrees, the residuals have zero mean, standard deviation 0.9 and all but one residual is smaller than 4 degrees. For THETA warmer than 12, the mean residual is -0.9, standard deviation 1.3 and 153 out of 156 residuals are within 4 degrees of the mean. We repeat for clarity and emphasis, that the OXYTMP reported in the .SEA file is the observed value, when present. However, the value used for conversion of oxygen concentration units was calculated from THETA according to the above formulae. These formulae are not expected to be definitive for all ocean basins. The amount of warming expected as a Niskin Bottle is hauled through, say 3000 meters of the water column will clearly depend on the temperature profile. However, we believe our present prescription to be amply adequate for the present purpose. Further comparisons with historical data (Oct 93) Further comparisons of sample data with historical data have been undertaken using anomalies with respect to average conditions, as introduced in the discussion of salinity. The standard fits were defined using least-squares fits to the data from A11, using data where theta < 1.0 in the western basin, and theta < 1.2 in the eastern basin. The resulting theta-oxygen relations were then (in µmol/l) western basin: O2 = 223.90 - 17.53 x theta eastern basin: O2 = 216.14 + 4.57 x theta Using a density of 1.028 kg/l, these are equivalent to (in µmol/kg) western basin: O2 = 217.80 - 17.05 x theta eastern basin: O2 = 210.25 + 4.45 x theta Note that not only are the deep oxygen values somewhat different between the two basins, but that the vertical gradients are of opposite signs. The intersection of the regressions is at a potential temperature of 0.35, where the oxygen value is 212 µmol/kg. The A11 data may now be compared with other data and inspected for bias by comparing the anomalies with respect to these standard fits, illustrated in Figures 14 to 17. Relative to A11 data (Figure 14), the following represent the median offsets: Figure 15 SAVE leg 4 + 4.0 (± 1.9) µmol/kg Figure 16 AtlantisII-107 + 1.0 (± 1.7) µmol/kg Figure 17 AJAX + 7.0 (± 0.75) µmol/kg Our data seem to be quite clearly lower in oxygen than the AJAX and SAVE leg 4 data; the comparison with Atlantis II data is somewhat inconclusive. The reason for the biases between the data sets is something of a mystery; we merely note them here. 2.3 Nutrients by: D.J. Hydes, P. Chapman and S.E. Holley Equipment and techniques The nutrient analyses were performed on an Alpkem Corporation Rapid Flow Analyzer, Model RFA-300. The methods used were: - Silicate: the standard AAII molybdate-ascorbic acid method with the addition of a 37°C heating bath (Hydes 1984) to reduce the reproducibility problems encountered when analyzing samples of different temperatures, noted on an earlier cruise when the standard Alpkem method was used (Saunders et al 1991, c.f. Joyce et al 1991). Phosphate used the standard (Murphy and Riley 1962) reagents and reagent to sea water ratios but with separate additions of ascorbic acid and mixed molybdate - sulphuric acid - tartrate to overcome the problem of the instability of a mixed reagent including ascorbic acid. Nitrate was determined using the standard Alpkem method. Previous experience has shown that better reproducibilities are achieved when the instrument is run in a laboratory with a stable temperature. The Alpkem was located in the new constant temperature laboratory on Discovery. The temperature was maintained between 21 and 22°C. A drawback of this location was that the large air circulation in the laboratory leads to enhanced evaporation of samples in the open cups sitting in the analyzer tray, and possibly to some contamination due to dust circulating in the air-stream. This was ameliorated by fitting a cardboard skirt round the sample tray lid. Sampling Procedures Sampling of nutrients followed that for trace gases (CFCs on this cruise) and oxygen. Samples were drawn into virgin polystyrene 30ml Coulter Counter Vials (ElKay). These were rinsed three times before filling. Samples were then analyzed as rapidly as possible after collection to avoid build up of a sample back log. Samples cups of 2.0 ml capacity were used. These were rinsed once by filling completely before filling with analyte. Tests carried out on the cruise showed that samples from all depths stored for a week in a refrigerator at 4°C were not significantly effected by storage. Calibration and Standards The calibrations of all the volumetric flasks used on the cruise were checked before packing and these were re-calibrated if necessary. Calibrations of the three Finn pipettes used on the cruise were checked before packing. The six Eppendorf fixed volume pipettes were delivered too late to be calibrated before the cruise. However in use no difference was detectable between the results achieved with the Finn pipettes and Eppendorfs. Nutrient standards Nutrient primary standards were prepared from salts dried at 110°C for two hours and cooled over silica gel in a dessicator before weighing. Precision of weighing was to better than 1 part per thousand. Nitrate 0.510g of potassium nitrate was dissolved in 500 ml of distilled water in a calibrated volumetric PP flask at a temperature of 21-22°C. Nitrite 0.345g of sodium nitrite was dissolved in 500 ml of distilled water in a calibrated volumetric PP flask at a temperature of 21-22°C. Phosphate 0.681g of potassium dihydrogen phosphate was dissolved in 500 ml of distilled water in a calibrated volumetric PP flask at a temperature of 21-22°C. Working standards were prepared from a secondary standard made by diluting 5.00 ml of the primary standard measured using a Finn pipette Digital 1.00 to 5.00 ml adjustable volume, in a 100 ml calibrated glass volumetric flask. Silicate 0.960g of sodium silica fluoride was dissolved in 500 ml of distilled water in a calibrated volumetric PP flask at a temperature of 21-22°C. Dissolution was started by grinding the fluoride powder to a paste with a few drops of water in 30 ml polythene beaker using a plastic rod for three to four minutes. Secondary calibration standards. A uniform set of six mixed secondary standards were prepared in artificial seawater, Concentrations (µM) were Nitrate 40, 30, 20, 10, and 0; Phosphate 2.5, 2.0, 1.5, 1.0, 0.5 and 0, Silicate 150, 100, 75, 50, 25 and 0 up to station 12288 and 150, 120, 90, 60, 30 and 0 thereafter. The artificial seawater was a 40ppt solution of Analar grade Sodium Chloride. Nutrients were undetectable in these solutions relative to Ocean Science International (OSI) Low Nutrient Sea Water which contains 0.7µM Si, 0.0µM NO3 and 0.0 µM PO4. On one occasion the solution was found to contain 0.6µM PO4 and consequently was not used. Establishment of a Quality Control QC Sample At a test station 12240 occupied on 26 December a large volume of deep water was collected with the idea of using this as a quality control standard when its stability had been verified. Samples of this water where run at intervals over the next two weeks. From station 12291 onwards a sample of 12240 water was measured as a "QC" sample on each analyzer run. The scatter of the data are shown in Fig 5. Silicate returned a consistent result with occasional flyers. The phosphate results suggest that the first (up to 12301) and second (up to 12319) one liter sub-sample were unstable but the third sample was stable. This may be due to the surface of the polythene bottle storage equilibriating with the sample. The sharp shift in the apparent nitrate concentration in the QC between stations 12311 and 12312 is currently inexplicable. It does not correspond to a change in primary standard concentration. It was difficult to detect in the contour plots, but does appear to be present when concentrations were compared along isopycnal surfaces. Reproducibility For the QC standard 189 measurements were made. The means were Silicate 78.85, Nitrate 28.85, Phosphate 1.79, percent standard deviations Silicate 1.05, Nitrate 2.45, Phosphate 2.35. For 10 replicates of the top standard run after station 12337 the percent standard deviations were Silicate 0.22, Nitrate 0.25, Phosphate 1.1. Reference HYDES, D.J. 1984 A manual of methods for the continuous flow determination of ammonia, nitrate-nitrite, phosphate and silicate in seawater. Institute of Oceanographic Sciences Report No 177, 40pp. JOYCE, T., CORRY, C. and STALCUP, M. 1991 Editors of WOCE operations manual, part 3.1.2 Requirements for WOCE hydrographic programme data reporting. US WOCE WHP Office 90-1, 71pp. MURPHY, J and RILEY, J.P. 1962 A modified single solution method for the determination of phosphate in natural waters. Anal. Chem. Acta, 27,31 36. SAUNDERS, P.M., GOULD, W.J., HYDES, D.J. and BRANDON, M. 1991 CTDO and nutrient data from Charles Darwin cruise 50 in the Iceland Faroes region. Institute of Oceanographic Sciences Deacon Laboratory, Report No 282, 74pp Further data quality control of nutrient samples (Oct 93) by: B. A. King Data quality control was tackled in a similar way as for salinity and oxygen, but of course there is no CTD sensor to assist in the rejection of poor sample values. Initially therefore, property-property plots were used to identify the sample numbers of outliers. These were mainly with theta or pressure as one coordinate, but plots of pairs of nutrients were also used. Outliers identified by this means were then inspected individually, and reasons sought for why they might have occurred. Suspect or bad flags were assigned to some or all of the nutrients in a total of 18 samples. Conversion between mass and volume units (Oct 93) The appropriate density for converting volume to mass units of nutrient analyses is the density in the lab where known volumes of sample were measured. Using a lab temperature of 21° and a mean salinity of 35, gives a density of 1.025 kg/l; density changes due to salinity variation amount to about 0.1%, and have been ignored. A density of 1.025 kg/l has been used to convert the data reported in the .SEA file. Internal consistency and comparison with historical data(Oct 93) As with the other tracers, standard regressions of the deep data onto potential temperature were defined in each basin, and used for comparing station data within and outside the cruise. The standard fits were as follows (µmol/l): western basin: NO2+3 = 33.88 - 1.42 x theta phspht = 2.228 - 0.121 x theta silcat = 126.90 - 17.85 x theta eastern basin: NO2+3 = 33.523 - 3.91 x theta phspht = 2.319 - 0.303 x theta silcat = 134.12 - 35.58 x theta At a density of 1.025 kg/l, these are equivalent to (in µmol/kg) western basin: NO2+3 = 33.05 - 1.385 x theta phspht = 2.173 - 0.118 x theta silcat = 123.80 - 17.41 x theta eastern basin: NO2+3 = 32.705 - 3.815 x theta phspht = 2.262 - 0.296 x theta silcat = 130.85 - 34.71 x theta Using the anomalies relative to these fits, it was possible to monitor the variation in the deep properties of the calibrated nutrient data. Note in passing that the eastern basin nitrate data fell in two families, offset from one another (discussed below). The regression was determined from just one family of data. Nitrates (Oct 93) A plot of the station average anomaly against station number made it immediately apparent that there was a problem (of the order of 1 µmol/l) in the consistency of standardization between groups of stations. Furthermore, abrupt changes in the deep nitrate values corresponded to changes in the nitrate value in the QC sample shown in Figure 5. Further investigation showed that all the significant changes in the apparent deep nitrate values occurred at stations where some adjustment had been made to the auto-analyzer. For example, adjusting the sensitivity to keep the instrument response to the top standard near the top of the scale, or a reactivation of the cadmium column. That such adjustments should lead to changes in the calibrated sample data is clearly not entirely satisfactory. After all, the whole point of standardization is that the concentration in the sample is being determined relative to that of the standard, and should be independent of the instrument settings used. Clearly the adjustments that were made had different affects on the standards and on the samples. The reason for this is not known. The cadmium column was reactivated before the analysis runs for stations 12284, 12312 and 12322. The first two of these were marked by a fall in the apparent concentration of deep sample nitrates. Calibration of the deep samples appeared unchanged after the third event. As part of the investigation of the standardization of the auto- analyzer, the instrument peak heights for the various standard concentrations came under renewed scrutiny. Time series plots of these peak heights were found to be a useful way of monitoring the performance of the instrument, and led to the identification of some hitherto unnoticed poor standard values. Joint inspection of the peak heights for the standards with the calibrated sample values was found to be illuminating. For example, it enabled a poorly determined baseline to be identified and corrected, which led to adjustment of some sample values. It also facilitated the correlation of instrument changes with apparent, but what we now know to be spurious, changes in deep sample values. It is our intention that on future cruises we will maintain this practice of carrying the information about instrument standardization and adjustment through to the inspection of sample data. Another result of the scrutiny of the standard peak heights was some investigation of the appropriate order of polynomial that should be used in the calibration. Unfortunately, the SOFTPAC software used to apply the calibration and drift corrections does not seem to have a facility for displaying the residuals between the standard concentrations and the fitted polynomial. Instead, the standard concentrations and the fitted polynomial are displayed on a graph, which ranges over the full scale of the variable. This makes it very difficult to determine the relative merits of one polynomial compared with another, and also makes it difficult to identify poor values that should be discarded from that particular set of calibration data. For example, a standard which has a lack of fit of 0.5 µmol/l should probably be discarded from the fit, but is hard to detect in the graphical display. Accordingly, the standard peak heights were reanalyzed in Excel spreadsheets, and the following conclusions drawn: a) The instrument peak heights should be calibrated using a second order polynomial fit. The coefficient of the quadratic term is positive. After fitting the polynomial to six standard concentrations, the rms error is of the order of 0.1 µmol/l. b) In a number of stations, poor peak heights for individual standards had been retained in the ship board calibration of the data, which should have been discarded. This was made apparent by inspection of the residuals after fitting the quadratic polynomial. Although for future cruises errors of this size should be eliminated, they were not considered to have had sufficient impact to make it worthwhile re-calibrating the data. Fixing the offsets arising from instrumental adjustment: As described earlier there are spurious changes in the deep sample values, associated with auto-analyzer adjustments. These have been fixed as follows: a) Stations 12284 to 12287: This group of stations, immediately after a reactivation of the cadmium column, were low relative to adjacent stations. The jump to lower values was clearly associated with the change to the column, but it is not clear why the values increase again. The average anomaly of deep nitrates for these four stations were compared with the average for four stations on either side (12279-12283 and 12288-12291) and found to be 1.56 µmol/l low. Using a mean deep nitrate value of 33.5 µmol/l, it was decided to scale all the sample nitrates for those four stations by a factor of 1.046. b) Stations 12312 to 12337: This group again follows a reactivation of the column, which was combined with an adjustment to the sensitivity of the instrument, and has lower values than preceding stations; however the nitrates do not appear to return to a higher value. The nitrate value in the QC sample shows the same behaviour. There was sufficient difference between the stations before and after 12312 that the standard regression for nitrate on potential temperature in this basin was determined from one group only, stations after 12312 being chosen. It was decided that one group of eastern basin stations should be adjusted relative to the other to bring them into agreement. There being no absolute means of deciding which were superior, the adjustment was applied to stations 12312 and following. Comparison of the deep nitrate anomaly for 12312-12337 with 12302-12311 indicated that a correction of 1.46 µmol/l was required. With a mean concentration of 30 µmol/l, this led to a scaling by a factor of 1.048 for all nitrate data for station 12312 to the end of the cruise. Note that since the standard regression had been calculated on data from these stations, all the deep eastern basin data are now about 1.5 µmol/l higher than the standard fit. Silicates (Oct 93) A plot of deep silicate anomaly against station number showed that as with nitrates there were some stations which were offset compared with adjacent stations. Unlike the nitrates, however, the silicate values did not seem to be so susceptible to adjustments of the instrument. Five stations stood out in particular, and these were examined and adjusted as follows: a) Station 12287: Examination of the calibration peak heights showed that they were about 10% low compared with preceding stations; there had clearly been a loss of sensitivity in the instrument for the analysis of this station. Accordingly, silicates for this station were scaled by a factor of 0.989 (-1.4 µmol/l at a concentration of 125 µmol/l) to bring the deep values into agreement with stations 12284-12290. b) Stations 12318, 12319, 12323, 12325. These stations all had unusually high anomalies for the deep silicate. 12318, 12323, 12325 all show up as spuriously high in the QC values of silicate shown in Figure 5. 12318 and 12319 also had lower than usual peak heights for the standardization. We therefore decided to reduce all four stations by a uniform factor, to bring their mean anomaly into agreement with the average for stations 12320, 12321, 12322, 12324, 12326. The required adjustment was -2.092 µmol/l at a mean value of 108 µmol/l, so a scaling factor of 0.981 was applied. Phosphates (Oct 93) No special adjustments were considered necessary for the phosphate data. The relatively greater uncertainty in the phosphate measurements means that the kind of corrections identified for nitrate and silicate are either unnecessary or undetected. Comparison with historical data (Oct 93) The internal consistency of the nutrient data (albeit after corrections to some stations) and comparison with other cruises is summarized in Figures 18 (nitrate), 19 (silicate) and 20 (phosphate); each figure has three parts (a) is this cruise, (b) is SAVE leg 4 data and (c) is AJAX data. These figures enable offsets to be identified, as well as showing the degree of scatter in each data set. The symbols show station averages of the deep sample anomalies. The relative offsets are further summarized in Table N1. The data were sorted into bins of size 0.25, 0.5, 0.025 µmol/l for nitrate, silicate and phosphate, and the center value of the bin containing the median is shown. Standard deviations of the station average anomalies are given in brackets. The standard error of the estimate of the mean/median is somewhat smaller than the standard deviation. TABLE N1 Medians of station-average offsets between sample data and standard regressions, for various data sets. Units are µmol/l. Values in brackets are standard deviations of the station average anomalies around the mean. A11 SAVE AJAX nitrate (west) 0.25(0.39) -0.25 (0.73) none nitrate (east) 1.5 (0.22) 0.75 (0.24) 0.5 (0.09) silicate -0.5 (1.33) 2.5 (1.53) 0.5 (0.37) phosphate 0 (0.025) 0.125(0.06) 0.05(0.015) Compared with SAVE, our nitrates are seen to be about 0.5 µmol/l (1.5%) high, silicates 2.5 µmol/l (2%) low and phosphates 0.125 µmol/l (5%) low. These differences are all significantly more than the internal uncertainty in the data. This demonstrates that our ability to maintain reproducibility over the period of a cruise is rather better than our confidence in the absolute accuracy of the data. The upper limits for accuracy given in the WOCE requirements are 1% for nitrate, 3% for silicate and 2% for phosphate. 2.4 CFC-11, CFC-12, and CFC-113 by: D. Smythe-Wright, S.M. Boswell and D. Price Sample collection All samples were collected from depth using 10 liter Niskin bottles. These had been cleaned prior to the cruise using a high-pressure water jet. All 'O' rings, seals and taps were removed, washed in Decon solution and propanol then baked in a vacuum oven for 24 hours. Cleaning and reassembling of the bottles was carried out at the commencement of the cruise to minimize contamination due to long storage. Of the 24 bottles initially assembled three had to be replaced due to leakage. None of the 27 working bottles showed a CFC contamination problem during the entire cruise. All bottles in use remained outside on deck throughout the cruise, those not in use were stored in aluminum boxes inside the hanger where there was a free flow of air to minimize contamination. Equipment and technique Chlorofluorocarbons CFC-11, CFC-12 and CFC-113 were measured on a total of 46 stations. The analytical measuring technique was a modification of that described in Smythe-Wright (1991a & b). In the modified system trapping was achieved using a 10 cm Poracil B trap cooled to below - 45°C. Subsequent de-sorption was by means of a water bath at 100°C. The trap was positioned on the exterior of the GC oven and not on the extraction board as in the original system. Valves V6 and V7 were replaced respectively with automated 8 port and 6 port Valco valves sited inside the GC oven to give better chromatographic resolution. Gases were forward flushed off the trap into a 3 m pre-column and subsequently chromatographcally separated using a 75 m long DB 624 megabore column. The pre-column was of the same material as the main column. Samples for analysis were drawn first from the Niskin bottles and stored under clean seawater. The analysis was completed mostly within 12 hours of the samples coming on board. Duplicate samples were run on most but not all casts due to the long analytical turn over time. Air samples were run daily from an air intake high up on the foremast. Air was pumped from this location through a single length of Dekoron tubing using a metal bellow pump. Calibration All CFC-11 and CFC-12 analyses were calibrated using 12 point calibration curves constructed from a gas standard calibrated by Weiss at SIO. This standard was contained in an Airco spectra seal cylinder as recommended in WHP, 1991. CFC-113 analyses were calibrated in a similar fashion using a compressed air standard prepared at the JRC and calibrated by Haine at PML. Contamination Because of a delay in customs clearance of the airfreight, the CFC equipment was delivered to the ship less than 24 hours before departure. This delay had a knock-on effect and compounded a number of teething problems, mainly due to two blocked valves and a contamination problem which masked the CFC-12 chromatographic peak. This resulted in the loss of data from a number of stations at the beginning of the cruise. The nature and source of the contamination problems was never totally discovered. It seemed to be related to the aquarium baths and the nontoxic seawater supply used for storing the syringes prior to analysis. The problem appeared some days after sailing and was overcome chromatographically by reducing the carrier gas flow and thereby separating the contamination from the CFC-12 peak. This meant that the overall analysis time was lengthened to 25 minutes and consequently restricted CFC analysis to every other CTD cast. Comparison with historical data Data accuracy was checked by comparison with SAVE leg 4 and 5 data and with data from the Ajax experiment. Some comparisons are given in Figure 6. Since four years has elapsed since these programmes some deviation in the data was expected particularly in the surface and deepest waters. In all cases deviations were consistent with the increase in atmospheric concentrations over the four-year period. Reference SMYTHE-WRIGHT, D., 1990a. Chemical Tracer Studies at IOSDL I. The design and construction of analytical equipment for the measurement of Chlorofluorocarbons in seawater and air. Institute of Oceanographic Sciences Deacon Laboratory Report No 274, 78 pp. SMYTHE-WRIGHT, D., 1990b. Chemical Tracer Studies at IOSDL II. Method manual for the routine shipboard measurement of Chlorofluorocarbons in seawater and air. Institute of Oceanographic Sciences Deacon Laboratory Report No 275, 64 pp. WHPO, 1991 WOCE Operations Manual. WHP Office Report WHPO 91-1 WOCE Report No 68/91. Woods Hole Mass, USA. 2.5 Samples taken for other chemical measurements a) Oxygen and Hydrogen isotope ratios by: S.M. Boswell A total of 241 samples were collected from 12 stations for isotope analysis by UEA. These included 18 duplicate samples from station 12333. Samples were collected directly into 50 ml glass vials following an initial rinse and two filling/emptying method. The caps were then sealed using parafilm and stored in the refrigerator. A total of 8 samples from the first three stations were lost when the fridge opened in rough weather. Samples thereafter were stored in the cold store. b) Iodine by: P. Chapman A total of 78 samples were collected from full water depth casts at Stations 12255, 12288, 12305 and 12335. These will be analyzed by Dr G Luther, University of Delaware USA. 2.6 CTD Measurements a) Gantry and Winch Arrangements by: S. Jordan, R. Phipps, S. Whittle Midships Gantry This gantry is of a novel design, and basically acts in the manner of a parallelogram-lifting table. While the gantry is moving from the inboard to outboard positions, the block from which the package is suspended describes an arc of a circle; due to the lifting action of the gantry, no winch movement is normally necessary while the package is being lifted outboard. Various loads, in our case the CTD package, can be safely deployed in virtually any sea state in which the ship can keep station. The performance of the gantry surpassed expectations. One reservation of note concerns the leading of the wire around a number of sheaves required to make the wire follow the parallelogram shape of the gantry. On two occasions, during deployment and with the CTD package at the sea surface, there became sufficient slack in the wire for it to jump off one of the sheaves. 10 Ton Traction Winch The CTD package was deployed using the 10T Traction Winch. The maximum descent/ascent rate required was 60m/min, therefore only one boost and two main pumps were required for successful operation (two boost and four main pumps being available). The following problems were noted:- a) A bearing on the scrolling gear was found to be excessively worn. This was replaced with a minimal loss of scientific cruise time (25/12/92). Inspection of the bearing showed it to be incorrectly designed or assembled. b) The 37kW storage system hydraulic power packs failed to provide power, a fault which persisted after various valves were stripped, cleaned and reassembled (1/1/93). The fault was eventually traced to an erratically operating potentiometer (by P.Gwilliam and A.Taylor). Approximately 36 Hours of scientific time was lost. c) Inboard compensator and back tension adjustments were needed more or less continuously. Although these were carried out with no loss of scientific time, a satisfactory solution was not found on the cruise. With known limitations the winch worked reasonably well and appears to have future expansion potential. It must be noted that the manufacturers intend to modify some of this system during the next ship refit, which should eliminate the problems encountered. The mechanical technicians are gaining more knowledge and confidence of the traction winch system and are especially pleased to have managed to repair/maintain the system with minimal down time. b) Equipment, calibrations and standards by: T.J.P. Gwilliam The CTD equipment used on this cruise was the property of IOS. The following equipment was deployed on the CTD/multi-sampler underwater frame:- 1. Neil Brown MK. 3 CTD complete with Sensormedics oxygen cell. IOS identification: DEEP01 2. Sea Tech. 100cm folded path transmissometer. Serial No.: 35. 3. General Oceanics 10 liter 24 bottle rosette. Model 1015. IOS identification: 01. 4. Six SIS (Sensoren Instrumente Systeme) digital reversing thermometers and two SIS digital reversing pressure meters. Serial numbers are detailed elsewhere in the report. 5. Simrad Altimeter, Model 807-200M 6. IOSDL 10 kHz. pinger. Backup equipment consisted of spare CTD, transmissometer, rosette, Niskin bottles, pinger and underwater frame. The shipboard equipment consisted of two complete integral systems for demodulating and displaying the CTD data as well as controlling the rosette multi-sampler. Each system included the following major units:- 1. EG&G demodulator. Model 1401. 2. IBM PS2 PC system with 80Mbyte tape system for archiving the data. 3. EG&G non-data interrupt rosette firing module. Calibration of the MK3 CTD temperature and pressure sensors was carried out at the IOSDL calibration facility. Conductivity and oxygen cell calibration was carried out at sea by reconciliation with sample values. Reversing thermometers were also calibrated in the lab, three at IOSDL and four at the Research Vessel Base. CTD temperature calibration - IOSDL DEEP01 - 19 June 92 was calibrated in degrees centigrade in the ITS-90 scale at six temperatures ranging from 0.19 to 25.3°. The transfer standard had been calibrated on 25 March 92 at the triple points of Mercury and water, and at the melting point of Gallium. The following linear fit for CTD temperature was found, with a rms error of 0.4 millidegrees. T = 0.9986622 x Traw - 0.01282084 No post-cruise laboratory calibration is available at present (March 1993). The CTD equipment is required on Discovery for two subsequent cruises, and will not be returned to IOSDL until at least June 1993. Stability of temperature calibration during the cruise was monitored by comparison with reversing thermometers, and this is discussed in the description of reversing thermometer data. CTD pressure calibration - IOSDL DEEP01 - 24 June 92 was calibrated by comparison with a Paroscientific Digiquartz model 240 portable transfer standard, in series with a deadweight tester; the Digiquartz was used as the pressure standard. The following quadratic fit for CTD pressure was found at an ambient temperature of 20°C, with a rms error of 1.8 dbar. P = 3.066286E-07 x Praw**2 + 0.9978454 x Praw - 12.6 Further corrections were applied during data processing for variation of offset with temperature, and up/down hysteresis. Equipment performance General With deployments at approximately four hourly intervals, power to the CTD was maintained throughout the cruise to minimize interruption problems. For satisfactory operation the optimum sea cable input voltage and current levels were 80 volts at 640 milliamps. Power distribution for the CTD, rosette and altimeter was controlled by a simple circuit in a separate 6 inch diameter pressure case mounted on the frame. The sea cable was terminated before sailing and a further three times during the cruise when cable damage occurred on deployment in heavy swell conditions. In two of the instances, the slack was sufficient to bounce the cable from the winch gantry pulleys, resulting in the instrument package free falling through the water for several meters. Approximately 30/40 meters of cable had to be discarded when this occurred. CTD As usual at the start of a cruise, the oxygen sensor was renewed before installing the system into the underwater frame. The first cast, to test the winch and CTD system, highlighted a wiring fault with the conductivity electronics which was quickly identified and corrected. Before station 12287 (near mid-cruise) the conductivity cell was flushed out with 10% hydrochloric acid as data from the previous two stations had indicated contamination. 24 Bottle Rosette System. It was this system that gave the most problems, non-closing of bottles and double bottle closing producing a lack operational confidence. Cures seemed, at times, to be the result of a "black art" rather than engineering expertise. The pylon was washed down immediately after each recovery with hot fresh water and the mechanical switching mechanism lubricated with silicon oil before the next deployment. Several times during the cruise the operational rosette pylon (01) was serviced on the frame and also interchanged with the backup unit (IOS identification 02) for a more detailed mechanical inspection and overhaul. The present system of codes, indicating bottle-firing information, is not satisfactory. Misfire codes transmitted when one or more bottles had in fact closed, multiple trips that could not be identified, and a lack of cam position information are just a few of the problems that need to be resolved. In one instance seawater ingress via the camshaft, on pylon 01, caused corrosion damage to the 24-way rotary code switch which had to be replaced. Perhaps there would have been greater protection had the switch been mounted on the shaft beneath the motor. Prior to the cruise the springs in all the bottles had been changed for ones of a different type at the request of the CFC analysts: these alternative springs had a different length and tension from the originals. Unfortunately, during the cruise the spring fastenings on the bottle end caps were mechanically breaking down to such an extent that the original springs were restored. During the cruise, three bottles were changed as suspected "leakers". Transmissometer. The transmissometer worked well throughout most of the cruise, but there were times when noise on the data, although not at an unacceptable level, proved difficult to trace and eliminate. The voltage in air was 4.310 volts, and the blackout offset was 16 millivolts. Towards the end of the cruise a slight leak in the prism pressure balancing mechanism was observed, which will require attention back at the laboratory. SIS Thermometers and Pressure Meters. Apart from routine battery replacements, one unit, T228, was removed after station 12248; the temperature readings were found to be in error by several hundred millidegrees. Comparison studies with the CTD data to check stability and accuracy were carried out and the results are shown elsewhere in this report. Altimeter and 10 kHz Pinger. This was the first IOSDL cruise where "depth off bottom" information was included into the CTD data stream and digitally displayed on the CTD monitor: the results were very satisfactory. The unit invariably locked onto the bottom from a range of 200 meters and tracked to the depth required with no problems. The 10 kHz. pinger, working in conjunction with the ship's Echosounder had in the past been the only way of obtaining this information. As the cruise progressed, and confidence increased with the altimeter, the 10 kHz. system was used more in a backup role. Apart from requiring battery changes the pingers themselves were totally reliable. Shipboard Equipment Overall the deck equipment worked satisfactorily with only one minor problem on one of the 1401 deck units. The acquisition software worked well and 12 tapes of 80 Mbytes of backup CTD data were archived. c) CTD Data Collection and Processing (updated June 94) by: B.A. King Data Capture and Reporting CTD data are passed from the CTD Deck Unit to a small dedicated microcomputer ('Level A') where one-second averages of all the raw values are assembled. This process includes checking for pressure jumps exceeding 100 raw units (10db for the pressure transducer on the CTD) and discarding of spikes detected by a median-sorting routine. The rate of change of temperature is also estimated. A fuller account of this procedure is given by Pollard et al. (1987). The one-second data are passed to a SUN workstation and archived. Calibration algorithms are then applied (as will be described) along with further editing procedures. Partially processed data are archived after various stages of processing. CTD salinity and dissolved oxygen concentrations are reconciled with sample values, and any necessary adjustments made. CTD temperatures and pressures are compared with reversing measurements. The downcast data are extracted, sorted on pressure and averaged to 2db intervals: any gaps in the averaged data are filled by linear interpolation. Information concerning all the CTD stations, is shown in the accompanying station list (either at the end of this report or in the accompanying .SUM file). With reference to the stated requirements for WHPO data reporting, note in passing: (a) The number of frames of data averaged into the 2db intervals is not reported. The IOSDL data processing path does not keep track of this information. (b) Approximately half the stations had the 1 db level missing from the averaged 2db files; i.e. the shallowest level was the 3db level. This situation would arise on stations where poor weather did not allow the CTD package to be brought close to the surface for the start of the downcast after the 'soaking' period at 10 meters depth. On such stations, the data have been extrapolated to the surface by replicating the T, S and O data from the shallowest available level (usually 3db, occasionally 5db), to provide a complete profile commencing with a 1 decibar data cycle. Such extrapolated data have been assigned a data quality flag of 2. Station 12286 In general downcast CTD data are reported. One exception is station 12286, where upcast data are reported. The conductivity had a number of fouling events on the downcast, identified by a number of jumps of order 0.002 to low values in the T/S relation. The upcast data appear to be satisfactory. The sorted, averaged 2db file was therefore compiled from the upcast data for all variables. After this station the conductivity cell was cleaned with dilute acid. After this the quality of the salinity data considerably improved, and the required cell offset changed by about 0.006 in salinity, suggesting that an accumulation of contamination had also been cleared away. Temperature calibration The following calibration was applied to the CTD temperature data:- T = Traw x 0.998662 - 0.01282 This calibration was in degrees C on the ITS-90 scale, which was used for all temperature data reported from this cruise. It was determined from a six-point calibration on 19 June 1992. A post-cruise temperature calibration was determined from a 12-point calibration on 8 July 1993 as follows: T = Traw x 0.998559 - 0.01409. This being sufficiently close to the initial calibration (a change in offset of about 1.3 millidegrees during the intervening 12 months), no changes were made to the temperature data. For the purpose of computing derived oceanographic variables, temperatures were converted to the 1968 scale, using T68 = 1.00024 T90 as suggested by Saunders (1990). However, all reported temperatures are in the ITS-90 scale. In order to allow for the mismatch between the time constants of the temperature and conductivity sensors, the temperatures were corrected according to the procedure described in the SCOR WG 51 report (Crease et al., 1988). The time constant used was 0.20 seconds. Thus a time rate of change of temperature (called deltaT) was computed, from 16Hz data in the level A, for each one-second data ensemble. Temperature T was then replaced by T + 0.2 x deltaT. Pressure calibration The following calibration was applied to the CTD pressure data, based on the 24 June 1992 calibration:- P = Praw **2 x 3.066286E-7 + Praw x 0.997845 - 9 The calibration applied to the data included an offset different from that found in the lab calibration and given in section 2.5b. The chosen offset gave correct pressures on deck and over the top few meters of the cast. A post-cruise pressure calibration at IOSDL on 7 July 1993 provided a laboratory calibration of P = Praw **2 x 4.172168E-7 + Praw x 0.996952 - 9 which differs from the pre-cruise calibration by less than 2 decibars over the range 0-6000. The data from the pre-cruise calibration were therefore accepted unchanged. A further correction was made for the effect of temperature on the CTD pressure offset:- Pnew= Pold - 0.4 (Tlag - 20) . Here Tlag is a lagged temperature, in degrees C, constructed from the CTD temperatures. The time constant for the lagged temperature was 400 seconds. Lagged temperature is updated in the following manner. If T is the CTD temperature, tdel the time interval in seconds over which Tlag is being updated, and tconst the time constant, then W = exp (- tdel/tconst) Tlag(t=t0+tdel) = W x Tlag(t=t0) + (1 - W) x T(t=t0+tdel). The values of 400 seconds for tconst and the sensitivity of 0.4 db per °C are based on laboratory tests. During the cruise, the variation of deck pressure value with ambient temperature was monitored. A least squares linear fit to the set of 73 deck pressure/temperature pairs collected had a slope of 0.49 and an offset of 5.4 db at 10°: this agrees with the applied correction to within 1.5 dbar over the range 0 to 20°C. A final adjustment to pressure is to make a correction to upcast pressures for hysteresis in the sensor. This is calculated on the basis of laboratory measurements of the hysteresis. The hysteresis after a cast to 5500m (denoted by dp5500(p)) is given in Table H1a for pressures at 500db intervals. Intermediate values are found by linear interpolation. If the observed pressure lies outside the range defined by the table, dp5500(p) is set to zero. For a cast in which the maximum pressure reached is pmax dbar, the correction applied to the upcast CTD pressure (pin) is pout = pin - (dp5500 (pin) - ((pin/pmax) * dp5500 (pmax))) Two thirds of the way through the cruise, at station 12303, a slight hysteresis between the up and down theta-S relationship was noted; upcast salinity was lower than the down. The size of the difference was small near the bottom of the cast, growing to a maximum of about 0.002 at about 3000 meters. At shallower depths the shape of the theta-S curve made it impossible to determine differences to the required accuracy. After some consideration, it was felt that the most likely cause of this was the CTD pressure (after the above correction for hysteresis) still reading slightly too high on the upcast. Accordingly the size of the hysteresis correction was increased, so that upcast pressures read slightly lower, and Table H1b was used. TABLE H1 (a) Laboratory measurements of hysteresis in pressure sensor dp5500(p) = (upcast - downcast) pressure at various pressures, p, in a simulated 5500m cast. (b) revised form of hysteresis used for stations 12303- 12337 (a) (b) p dp5500(p) dp5500(p) db db db 5500 0.0 0.0 5000 1.0 0.0 4500 1.2 1.2 4000 1.8 2.8 3500 2.4 4.4 3000 3.0 6.0 2500 3.4 6.8 2000 4.8 6.6 1500 5.6 6.5 1000 6.0 6.4 500 6.3 6.3 0 0.0 0.0 Extraction of upcast data for calibration Following procedures developed on previous cruises, CTD data were extracted for salinity and oxygen calibration as follows: The Niskin bottle firing events were logged using a level A microprocessor dedicated to that purpose. This provided accurate times of the bottle closures. The CTD data after nominal calibration were averaged into 10-second bins, and merged onto the firing events using linear interpolation on time; the time for both the CTD data and the firing events were provided by the ship's master clock, and were therefore reliable. The 10-second averages were believed to be representative of the CTD data for the water sampled. After coefficients for calibration of the CTD oxygen or salinity had been calculated and applied to the 1 Hz data, the averaging and merging procedure was repeated as often as necessary, until the calibration was finalized. In this way, residuals were always calculated between the sample values and the latest estimate of the calibrated CTD data. Salinity calibration Salinity was calibrated during the course of the cruise, by comparison with upcast sample salinities. This was done on a station by station basis. A cell conductivity ratio of 0.996683 was estimated from early stations, and this was applied to all station data as an initial calibration. The initial calibration was followed by the correction to conductivity ratio:- Cnew = Cold x (1 - 6.5E-6 x (T-15) + 1.5E-8 x P) After reconciliation with sample salinities, vertical profiles of residuals showed a systematic depth dependence. A final salinity calibration on a station by station basis was made by fitting the residuals with the form a + b * T + c * P. The need for this procedure is not understood. We do not necessarily believe that this correction represents some physical response of the cell to temperature and pressure. Rather, it is simply a convenient way of fitting the salinity residuals with two variables which have different variation over the water column; however, since it successfully removes most of the systematic part of the salinity residuals, it is considered to be a satisfactory tool for the correction of the CTD salinity data. The offset at the bottom of each station introduced by the expression above, which may be used as a description of the drift of the cell, was monitored and varied between -0.008 and +0.008 (but not monotonically). A full list of the coefficients appears in Table H4, which is located at the end of this section. Unlike the oxygen calibration procedure (q.v.), the agreement between upcast and downcast T/S profiles was good. It was therefore decided that the calibration of upcast CTD salinities by comparison with sample salinities would provide adequately calibrated downcast CTD salinity data. Stations 12251-12255: These stations required special attention for salinity calibration. An extra temperature sensor (FSI) had been introduced on the rosette and interfaced to the CTD for evaluation purposes. This extra power demand on the CTD meant that the conductivity cell did not return to satisfactory values for some while after firing a bottle, while the rosette pylon was recharging. Once the problem had been properly identified the power supply to the CTD was increased, and the problem solved. In the mean time, however several profiles of data were collected for which the upcast salinities were suspect or useless. Accordingly, straightforward comparison of upcast CTD salinities with sample salinity could not be used for CTD calibration. The CTD data were therefore scrutinized to ensure that bad data cycles (sometimes several hundred meters worth) were excluded from the calibration. Some salinity sample values were not used, if it was not possible to find a suitable CTD value for comparison. Sometimes a matching downcast CTD salinity would be used, in the manner employed for oxygen calibration. The final downcast CTD salinity values are believed to be satisfactory. However, there remain a number of sample minus CTD residuals which are quite large, mainly associated with poor upcast CTD salinities. The residuals for these five stations are therefore omitted entirely from Figure 10. Station 12325: Two casts were required to complete station 12325. The rosette jammed part of the way through the upcast, so no samples were collected in the upper 1500 meters of cast 1. A second cast to 1500 meters was carried out to obtain a complete sample profile (sample numbers 32525-32538). The CTD data reported are the downcast of cast number 1. Having applied a single CTD salinity calibration to the two casts together, the salinity residuals for cast number 2 are rather large; basically, the CTD salinities are 0.003 to 0.005 higher than the bottle values. Two further pieces of evidence are available: (a) The CTD calibration was offset by about 0.002 for station 12326 & 12327 (see Table H4). (b) The FSI conductivity cell, described elsewhere, was in use on this station. Inspection of the conductivity data from the two sensors supports the suggestion that the NBIS salinity data did indeed drift to higher values on cast number 2. We therefore conclude that the cast 1 downcast salinities which form the cruise data set are satisfactory and that the cast 2 upcast CTD salinities, which appear in the sample .SEA file, are questionable. Oxygen calibration CTD oxygens were calibrated by fitting to sample values using the following formula:- O2 = oxsat(T, S) x rho x (oxyc + c) x exp (a x (W x ctdT + (1-W) x oxyT) + b x P) where the coefficients rho, a, b, the oxyc offset c and the weight W were chosen on a station by station basis to minimise the rms residual. W is forced to lie in the range 0 to 1. The fitting of oxygen data at sea did not allow for an offset to the oxygen current, and required the weight W to be specified by the user. The resulting fits were not entirely satisfactory: rms errors were about 3-4 µmol/kg, and there was a tendency for the calibrated CTD data to produce the wrong oxygen gradient in the deep water. Introducing the time rate of change of oxyc had little effect but, in contrast, an offset in oxyc (of the order of -0.07 µA) produced a significant improvement. IOSDL has not previously found it necessary to introduce an offset in oxyc in order to achieve satisfactory oxygen fits, and the value required is rather greater than suggested in the WOCE Manual of operations and methods. With hindsight, we suspect that this offset indicates an unusual oxygen cell, which should probably have been replaced. However, having introduced the offset, there is no reason to doubt the quality of the derived CTD oxygen data. Table H5, located at the end of the section, gives oxygen fitting coefficients and residuals station by station. For a few stations, where there were insufficient sample values to fit all five coefficients sensibly, b and/or c were chosen from values on nearby stations. For some stations, several passes through the fitting procedure were used to arrive at the final coefficients. After an initial fit, outliers were identified, and excluded from subsequent fits. In this way the CTD data were used to help identify sample values requiring 'suspect' or 'bad' flags. There is further discussion of this in the section describing the oxygen sample data. In general, samples believed to be suspect for any reason, were excluded from the CTD fitting. However, it was sometimes necessary to include them (stations 12253 to 12257, for example, where all samples were suspect), and such included samples are listed in the sample oxygen discussion. Furthermore some 'good' samples could not be fitted properly with the CTD data - typically in regions of strong vertical gradient. These samples were also excluded from the fit if their exclusion resulted in significantly improved residuals over the rest of the profile. Numbers of samples excluded for this reason are also listed elsewhere. The residuals between CTD and sample oxygens are summarized in a table in Section 2.2, where they are averaged into 500 meter depth bins. The errors appear to have a systematic form. However, the rms difference of all samples is 2.66 µmol/kg, and 1.73 µmol/kg for samples from deeper than 3000 dbar. We therefore consider the CTD data to be acceptable in their present form. Calibration of downcast CTD oxygen data using upcast samples: The calibration algorithm for the CTD oxygen data generally produced up and down profiles which did not match particularly well, either as pressure/oxygen profiles or as potemp/oxygen profiles. This is believed to be a widespread problem, arising from the calibration algorithm not being a sufficiently good model of the true response of the sensor. However, we know that some investigators find that they can get consistently good up/down matching. Whether this varies from cell to cell, is a subtle function of the electronics of the CTD, or a function of the way the algorithm is applied, we do not know. In the present data, up/down agreement varied from very good to appalling, with no apparent reason or change of procedure. The fact remains, therefore, that we require to bring the downcast CTD oxygens (which we report for all but station 12286) into agreement with the upcast samples. For each sample, we thus need to extract a downcast CTD data cycle (press, temp, oxyc, oxyt) for calibration against sample oxygen. Again following procedures developed on previous cruises, we extracted a downcast data cycle of CTD data as follows: a) the pressure, potential temperature and potential density (referenced to the nearest round multiple of 500db) at the bottle closing time were extracted. This provided a choice of three parameters which could be used to find a suitable matching downcast data cycle. No one parameter was considered to be universally the best. Matching on pressure was not considered to be ideal, because of vertical motion of water during the elapsed time between down and up cast passing through the same water mass. In general, because of up/down salinity biases on some stations, potential temperature (supposed conserved while internal waves pass through) would seem to be the best, and preferable to potential density. However, the profiles encountered on this cruise included ones where, because of the salinity gradient, there were reversals in potential temperature, or regions of very weak potential temperature gradient. In these cases, potential temperature was not suitable for matching, and potential density was used. Potential density could not be used throughout, however, because apart from vulnerability to poor salinity values there were also regions where potential temperature and salinity had reasonable gradients but potential density had only very weak gradients. The matching procedure therefore usually employed potential temperature at pressures greater than 3000db, and potential density at pressures less than 3000db. Matched data cycles where the up/down pressure difference was greater than 10% were flagged and received special attention. b) the CTD downcast was scanned for pairs of data cycles which bracketed the chosen parameter, and closer of the pair listed. c) where step (b) produced more than one candidate data cycle (arising from potemp reversals, for instance), the one with the nearest pressure was chosen. d) thus far, the procedure was entirely automated. Every matching data cycle was then examined for plausibility, by (subjective) consideration of agreement of pressure, temperature and salinity between up and down data cycles. If agreement was poor, or if the automatic procedure (ie choose the one with the nearest pressure from two or more possibles ) had apparently chosen the wrong data cycle, a different data cycle was specified to be the matching one. This quite commonly occurred for the shallowest sample, when the data cycle with matching pressure might be specified instead. e) the CTD values from the resulting set of up to 24 downcast data cycles were employed in the oxygen fitting algorithm. Conversion from µmol/l to µmol/kg: Because of the sequence of events, and the careful thought that went into the conversion of oxygen units, the CTD oxygen data were fitted to sample data measured in µmol/l, that had not yet been converted to µmol/kg. Accordingly the CTD data also require conversion. Since the requirement is for the converted CTD data to fit the converted sample data, the CTD data throughout the cruise have been scaled using density calculated as for the sample data (see the discussion of sample oxygens), namely one calculated from measured salinity and a temperature which is a piecewise linear function of measured potential temperature. Transmissometer data Transmittance data from 1 one meter folded path transmissometer were routinely collected throughout the cruise. At present (June 1994) station to station inconsistencies in the calibration of these data mean that they are not ready for submission to the WHPO with the bulk of the CTD data. They will be submitted in due course, after completing best efforts at their calibration. SIS thermometer data, and the stability of the CTD temperature sensor Six SIS digital temperature meters and two digital pressure meters were used throughout the cruise. These, along with salinity and chemical data from the rosette water samples, were used to determine the depth of bottle firings. Digital Reversing Temperature Meters (RTM) The digital temperature meters were calibrated using the linear fits given in Table H2. In addition to these another sensor, T228, was discarded after the first station of the A11 cruise. A comparison of CTD and RTM temperatures is given in Table H3 below. The table has four parts. Parts (a) and (b) present data from the entire section, with part (b) for temperature colder than 2°; as expected, the latter have generally smaller standard deviations. Parts (c) and (d) show the data colder than 2° further subdivided about station 12293, which is one of the stations over the mid-Atlantic Ridge. Three numbers of observations are given in each part, corresponding to the number of differences greater than 10 millidegrees, considered as outliers and discarded, the number less than 10 millidegrees, from which mean and standard deviation are calculated, and the number within two standard deviations of the mean. The most significant feature of these tables is the change in mean value of ctd-T399 and ctd-T400 between the two halves of the cruise, the mean difference changing by 1.3 millidegrees. This is rather more than the standard deviation of the measurement, and much more than the standard error of the estimate of the mean for each group. Although this might be thought to indicate an offset in CTD temperature calibration (there being no change in the T400-T399 difference), there is no evidence for this in the ctd-T401 and ctd-T219 pairs. Our tentative conclusion is that the difference arises because the temperature observed at rosette position 1 is generally warmer in the eastern basin than in the western basin. Note the mean temperature of the observations, which is shown in the last column of Table H3 (c) and (d). We suppose that non-linearity in the response of either CTD or RTM temperature near zero may be the cause of the change in CTD-RTM difference. If it is the behavior of the RTM thermometers that is nonlinear, then it must be very similar in the two thermometers; this is not unreasonable for two instruments of the same type. On the other hand, we do not exclude the possibility of nonlinear behavior in the CTD temperature. When the CTD is re- calibrated on return to IOSDL, careful attention will be paid to establishing the linearity or otherwise of the calibration near zero. (Note added, May 1994: This effect was examined by careful calibration of the CTD near zero degrees in late 1993. Although some other CTD instruments have been found by IOSDL to have nonlinear errors of several millidegrees, the instrument used during A11 had errors of no more than 0.5 millidegrees, and then only within 0.2 degrees of zero. CTD non- linearity near zero is therefore unable to account for the observed change in CTD-RTM difference.) In any case the overall consistency of the CTD and RTM comparisons and the magnitude of the change in differences amongst them strongly imply that there was no significant change in the CTD calibration between the start and the end of the cruise. Digital Reversing Pressure Meters (RPM) Two reversing pressure meters were used :- Rosette Pressure position meter 1 P6132H 8 P6075S Despite the shortcomings in the RPM performances, which are described below, their data were very useful in confirming or identifying the depth of bottle closures. Calibration of P6075S were carried out by the manufacturer on both 13.2.88 and 27 3 90 the latter at temperatures of both 3 and 20°C. These indicated that corrections of between -7 and +3 dbar were required over the range 0 to 5400 dbar. However residuals between the calibrated RPM and the CTD were found on cruise 199 to exceed 30 dbar at pressures greater than 3000 dbar. P6132H was calibrated by the manufacturer on 22.2.90. Linear interpolation was used to correct the RPM between the following calibration values in dbars:- (P6132H pressure, correction applied), (0006,-6), (0975,+6), (1949,+12), (2930,+12), (3915,+8), (4907,-4), (5405,-11), (6022,-22). The last pair was not supplied by the manufacturer, but was an extrapolation of the manufacturer's information. In general, after applying the above calibration, P6132H shows a consistent offset compared with the CTD of about 14 dbars over the range 1800 - 6000 dbar. Discrepancies of similar magnitude between RPM and CTD pressures have been noted on a number of previous IOS cruises, see for example the CONVEX cruise report (Gould et al, 1992). On cruise 199 the CTD bottom pressures were converted to depth and were compared with corrected Echosounder depths minus depth of CTD off bottom: the differences had a mean value of 3 meters and 75 percent were smaller than 12 meters. On the CONVEX cruise an even smaller mean for nearly 100 stations was found. We are therefore quite confident of the CTD pressure calibration and in the near future plan to carry out calibration and other tests of the RPM instruments at IOSDL. Reference CREASE, J. et al. 1988 The acquisition, calibration and analysis of CTD data. Unesco Technical Papers in Marine Science, No 54, 96pp. GOULD, W.J. et al. 1992 RRS Charles Darwin Cruise 62, 01 Aug-04 Sep 1991. CONVEX-WOCE Control Volume AR12. IOSDL, Institute of Oceanographic Sciences Deacon Laboratory Cruise Report, No 230, 60pp. POLLARD, R.T., READ, J.F. and SMITHERS, J. 1987 CTD sections across the southwest Indian Ocean and Antarctic Circumpolar Current in southern summer 1986/7.Institute of Oceanographic Sciences Deacon Laboratory Report No 243, 161pp. SAUNDERS, P.M. 1990 The International Temperature Scale 1990, ITS-90. WOCE Newsletter No 10, p10. (Unpublished manuscript). TABLE H2 Digital RTM calibrations. Tcal = b x Traw + a Position Thermometer b a Date off Source on rosette calibration 1 T399 1.00031 -0.00331 20/7/92 IOSDL 1 T400 1.00006 0.00146 20/7/92 IOSDL 4 T401 1.00016 -0.01002 20/7/92 IOSDL 4 T219 0.99992 -0.01250 18/8/92 RVS 8 T238 0.99992 0.00175 18/8/92 RVS 12 T220 0.99999 -0.00570 18/8/92 RVS TABLE H3 Summary of RTM data (a) (b) All Data T < 2° Pair n n n mean sd n n n mean sd >10 <10 <2sd mdeg mdeg >10 <10 <2sd mdeg mdeg mdeg mdeg mdeg mdeg ctd-T399 1 92 90 1.0 1.6 0 75 72 1.1 1.5 ctd-T400 2 91 88 0.8 1.3 1 74 70 0.9 1.1 ctd-T401 3 90 84 2.1 2.2 2 60 56 2.0 1.7 ctd-T219 5 82 76 -6.7 2.3 2 56 52 -6.8 1.7 ctd-T238 9 80 75 1.6 2.5 0 17 16 0.7 3.0 ctd-T220 9 69 65 2.0 2.6 0 1 1 1.9 - T400-T399 0 93 89 0.4 0.9 0 75 72 0.4 0.9 T401-T219 4 82 79 -8.7 1.8 2 56 55 -8.6 1.4 (c) (d) stnnbr < 12293 stnnbr > 12293 Pair n n n mean sd mean n n n mean sd mean >10 <10 <2sd mdeg mdeg temp >10 <10 <2sd mdeg mdeg temp mdeg temp degC mdeg temp degC ctd-T399 0 38 36 0.6 0.9 0.33 0 36 35 1.9 0.8 1.19 ctd-T400 0 38 37 0.3 1.0 0.33 0 36 35 1.6 0.9 1.19 ctd-T401 1 39 37 2.0 1.5 0.60 1 20 19 2.2 1.8 1.46 ctd-T219 0 36 34 -6.8 1.7 0.48 2 19 18 -6.6 1.5 1.46 T400-T399 0 38 36 0.3 0.9 0.33 0 36 34 0.3 0.5 1.19 T401-T219 0 36 35 -8.9 1.5 0.48 2 19 18 -8.9 1.1 1.46 TABLE H4 Final CTD salinity adjustments. S = S + (a + b * T + c * P)/1000 The number in the deep offset column was the offset applied to the CTD salinities at the bottom of the cast as a result of the residual fitting procedure. Station a b c Deep Comments number offset 12247 -4.95 0.19 -0.00597 -0.0054 12248 -2.80 0.00 0.00000 -0.0028 12249 -2.00 0.00 0.00000 -0.0020 12250 -2.00 0.00 0.00000 -0.0020 12251 -2.00 0.00 0.00000 -0.0020 12252 -1.87 0.09 -0.00141 -0.0054 12253 -7.09 0.38 0.00128 -0.0025 12254 -4.00 0.00 0.00000 -0.0040 12255 -2.35 -0.15 0.00002 -0.0023 12256 -2.20 0.11 -0.00058 -0.0050 12257 1.58 0.10 0.00000 0.0016 12258 -1.34 0.50 0.00015 -0.0004 12259 3.35 -0.27 -0.00031 0.0015 12260 3.50 0.00 0.00000 0.0035 12261 3.86 -0.62 -0.00030 0.0018 12262 0.04 0.28 0.00034 0.0022 12263 3.60 -0.21 -0.00005 0.0032 12264 4.70 -0.64 -0.00069 0.0006 12265 3.90 -0.17 -0.00020 0.0028 12266 2.90 0.00 -0.00013 0.0022 12267 3.80 -0.37 -0.00029 0.0022 12268 5.30 -0.25 -0.00045 0.0029 12269 7.60 -0.78 -0.00151 -0.0004 12270 2.40 0.00 0.00000 0.0024 12271 2.60 0.02 -0.00029 0.0011 12272 0.50 0.18 0.00052 0.0031 12273 3.30 0.04 -0.00030 0.0018 12274 1.10 0.16 -0.00032 -0.0005 12275 2.10 -0.12 -0.00048 -0.0003 12276 -1.50 0.35 0.00014 -0.0007 12277 0.10 0.02 -0.00004 -0.0001 12278 1.00 0.13 -0.00035 -0.0009 12279 4.80 -0.30 -0.00068 0.0011 12280 -0.60 0.30 -0.00032 -0.0022 12281 2.00 -0.10 -0.00048 -0.0005 12282 4.10 0.16 -0.00012 0.0035 12283 0.20 0.13 -0.00048 -0.0023 12284 0.80 0.06 -0.00018 -0.0001 12285 6.40 0.27 0.00029 0.0078 12286 4.90 -0.37 -0.00048 0.0026 12287 -0.80 -0.45 -0.00063 -0.0040 Note 1 12288 -3.60 0.11 -0.00027 -0.0048 12289 -8.00 0.58 0.00072 -0.0045 12290 -3.60 -0.03 -0.00038 -0.0052 12291 -2.20 -0.01 -0.00064 -0.0051 12292 -0.90 -0.25 -0.00086 -0.0046 12293 -0.50 -0.27 -0.00091 -0.0041 12294 -6.89 0.39 0.00057 -0.0044 12295 -8.45 0.55 0.00102 -0.0042 12296 -1.02 -0.40 -0.00120 -0.0059 12297 -5.67 0.05 0.00012 -0.0052 12298 -6.96 0.31 0.00045 -0.0052 12299 -4.61 0.01 0.00029 -0.0036 12300 1.24 -0.39 -0.00039 -0.0008 12301 1.53 0.07 -0.00056 -0.0005 12302 1.62 0.21 -0.00063 -0.0006 12303 1.90 -0.32 -0.00084 -0.0018 12304 3.35 -0.43 -0.00096 -0.0009 12305 -0.04 0.04 -0.00024 -0.0010 12306 0.99 -0.18 -0.00047 -0.0010 12307 -0.63 -0.02 -0.00011 -0.0011 12308 2.03 -0.24 -0.00083 -0.0014 12309 1.37 -0.22 -0.00031 -0.0001 12310 -2.34 0.09 -0.00023 -0.0033 12311 -0.51 -0.03 -0.00031 -0.0020 12312 -0.03 -0.06 -0.00016 -0.0008 12313 1.24 -0.14 -0.00011 0.0005 12314 0.79 -0.05 -0.00031 -0.0008 12315 0.99 -0.13 -0.00045 -0.0015 12316 -0.55 0.11 -0.00032 -0.0021 12317 -1.64 0.03 -0.00072 -0.0053 12318 -1.07 -0.11 -0.00108 -0.0068 12319 -2.59 -0.09 -0.00064 -0.0061 12320 -1.56 -0.06 -0.00069 -0.0053 12321 -2.92 0.31 -0.00050 -0.0052 12322 -2.06 -0.27 -0.00058 -0.0055 12323 -2.90 -0.10 -0.00041 -0.0052 12324 -2.18 0.00 -0.00067 -0.0056 12325 -2.64 -0.09 -0.00062 -0.0059 12326 -5.69 0.05 -0.00036 -0.0075 12327 -12.42 0.36 0.00659 -0.0078 12328 -11.91 0.36 0.01548 -0.0025 12329 -7.69 0.40 0.00176 -0.0046 12330 -3.45 0.10 0.00079 -0.0020 12331 -2.96 0.12 0.00048 -0.0016 12332 -1.92 0.01 -0.00017 -0.0023 12333 -0.83 -0.05 -0.00057 -0.0027 12334 0.66 -0.13 -0.00110 -0.0035 12335 -1.58 -0.17 -0.00083 -0.0052 12336 -1.37 -0.03 -0.00033 -0.0029 12337 1.62 -0.20 -0.00080 -0.0025 Notes 1) The conductivity cell was cleaned prior to station 12287 with dilute acid. TABLE H5 CTD oxygen fitting coefficients and residuals O2 = oxsat(T,S) * rho * (oxyc + c) * exp(a * (W*ctdT + (1-W)*oxyT) + b*P) Station rho a b c W rms No. of number residual samples µmol/kg in fit 12247 1.3509 -0.04531 0.0002200* -0.0500* 0.5232 4.50 5 12248 1.4430 -0.04985 0.0002200* -0.0500* 0.3548 5.81 8 12249 1.5021 -0.05487 0.0001365 -0.0500* 0.4718 2.54 10 12250 1.3023 -0.04065 0.0002307 -0.0500* 0.4500* 10.31 10 12251 1.3957 -0.04580 0.0002102 -0.0564 0.4625 0.97 9 12252 1.3773 -0.04171 0.0001875 -0.0364 0.3563 3.75 18 12253 1.3662 -0.03939 0.0002661 -0.0798 0.5888 4.87 15 12254 1.4825 -0.04791 0.0002278 -0.0792 0.5610 5.06 14 12255 1.4115 -0.04461 0.0001792 -0.0342 0.4481 4.16 19 12256 1.5397 -0.03968 0.0002144 -0.0799 0.0288 4.60 22 12257 1.5118 -0.05284 0.0003048 -0.1309 0.4454 3.40 20 12258 1.3285 -0.03882 0.0001408 0.0145 0.4457 2.53 23 12259 1.4914 -0.03994 0.0002219 -0.0783 0.2443 3.49 21 12260 1.4591 -0.04499 0.0001978 -0.0549 0.6260 2.68 12 12261 1.5485 -0.04432 0.0002515 -0.1069 0.4804 3.57 24 12262 1.5619 -0.04913 0.0002569 -0.1128 0.3995 3.98 23 12263 1.2720 -0.03901 0.0001499 -0.0179 0.5207 3.92 24 12264 1.4108 -0.04008 0.0002459 -0.1271 0.3886 3.52 23 12265 1.3179 -0.03916 0.0002020 -0.0811 0.4963 3.38 23 12266 1.3719 -0.03900 0.0002317 -0.1109 0.4432 2.07 22 12267 1.3776 -0.03755 0.0002525 -0.1244 0.5492 1.55 17 12268 1.3353 -0.03949 0.0002305 -0.1048 0.6906 2.30 23 12269 1.3596 -0.04107 0.0002267 -0.1036 0.5512 1.32 16 12270 1.2812 -0.03574 0.0002142 -0.0823 0.4933 1.96 15 12271 1.3696 -0.03904 0.0002298 -0.1067 0.6887 3.02 24 12272 1.3317 -0.03626 0.0002229 -0.0944 0.8601 2.89 20 12273 1.3189 -0.03277 0.0002282 -0.1049 0.4223 2.87 21 12274 1.3318 -0.03581 0.0002334 -0.1088 0.4931 2.27 23 12275 1.2802 -0.03730 0.0002004 -0.0718 0.6542 3.22 23 12276 1.3504 -0.04087 0.0002080 -0.0862 0.8183 4.58 23 12277 1.3447 -0.03611 0.0002030 -0.0817 0.4094 1.90 35 12278 1.3839 -0.03719 0.0002150 -0.0940 0.4477 2.11 24 12279 1.3907 -0.04412 0.0002012 -0.0813 0.7213 3.04 22 12280 1.3703 -0.04452 0.0002014 -0.0771 0.6437 2.57 24 12281 1.3132 -0.03819 0.0001860 -0.0599 0.4888 3.25 22 12282 1.3118 -0.03885 0.0001818 -0.0516 0.8816 3.88 23 12283 1.3900 -0.04221 0.0002096 -0.0952 0.5435 2.68 23 12284 1.4284 -0.04711 0.0001855 -0.0750 0.8929 3.56 22 12285 1.3718 -0.04497 0.0002173 -0.1035 0.4343 2.74 22 12286 1.4744 -0.05322 0.0001871 -0.0730 0.0000 2.33 21 12287 1.2674 -0.03658 0.0001526 -0.0035 0.6970 1.88 23 12288 1.3757 -0.04111 0.0001603 -0.0315 0.5249 3.25 21 12289 1.4411 -0.04461 0.0001840 -0.0607 0.4252 2.29 21 12290 1.3827 -0.04342 0.0001806 -0.0481 0.6347 3.20 20 12291 1.4099 -0.04423 0.0001728 -0.0461 0.6092 2.39 22 12292 1.3167 -0.03853 0.0001678 -0.0255 0.7202 1.91 19 12293 1.3517 -0.03807 0.0001819 -0.0454 0.5418 3.33 22 12294 1.2880 -0.03851 0.0002598 -0.1160 0.8752 2.95 20 12295 1.4685 -0.04358 0.0002230 -0.0930 0.6440 4.69 19 12296 1.3764 -0.03828 0.0002070 -0.0653 0.6122 1.76 21 12297 1.4189 -0.04084 0.0002023 -0.0662 0.4213 1.35 17 12298 1.3558 -0.04550 0.0001894 -0.0379 0.8781 3.33 17 12299 1.4762 -0.05084 0.0001891 -0.0593 0.6076 4.23 17 12300 1.3929 -0.04131 0.0001923 -0.0554 0.7530 3.63 19 12301 1.3777 -0.03845 0.0001937 -0.0532 0.4853 1.93 18 12302 1.2843 -0.03212 0.0001847 -0.0310 0.4087 1.52 16 12303 1.3022 -0.03259 0.0001876 -0.0385 0.4467 2.01 19 12304 1.3744 -0.03734 0.0001998 -0.0604 0.5594 2.55 19 12305 1.2735 -0.03122 0.0001826 -0.0271 0.4398 2.59 19 12306 1.2605 -0.03233 0.0001727 -0.0158 0.4381 3.27 14 12307 1.3993 -0.03702 0.0001913 -0.0598 0.4224 2.63 18 12308 1.3340 -0.03127 0.0001973 -0.0548 0.1266 2.75 19 12309 1.3658 -0.03775 0.0001957 -0.0576 0.6492 2.40 19 12310 1.3309 -0.03648 0.0002067 -0.0774 0.3781 2.64 22 12311 1.4089 -0.03736 0.0002030 -0.0756 0.2534 2.04 22 12312 1.3595 -0.03842 0.0002131 -0.0828 0.5908 1.22 22 12313 1.4339 -0.03848 0.0002101 -0.0861 0.4055 2.40 23 12314 1.3870 -0.03763 0.0001923 -0.0628 0.5510 2.40 22 12315 1.4477 -0.04060 0.0002107 -0.0863 0.6599 2.24 23 12316 1.4004 -0.03675 0.0001983 -0.0709 0.3691 2.63 22 12317 1.3552 -0.03659 0.0002223 -0.0919 0.5974 2.50 22 12318 1.4037 -0.03577 0.0002085 -0.0817 0.2703 2.46 23 12319 1.3553 -0.03552 0.0002037 -0.0761 0.4580 2.49 24 12320 1.3557 -0.03484 0.0002043 -0.0762 0.3342 3.41 23 12321 1.3599 -0.03461 0.0001987 -0.0719 0.2483 1.95 21 12322 1.3645 -0.03356 0.0001958 -0.0723 0.1471 2.48 24 12323 1.3633 -0.03291 0.0001991 -0.0754 0.0926 2.95 23 12324 1.3647 -0.03653 0.0002016 -0.0756 0.4403 3.43 22 12325 1.3672 -0.03429 0.0001998 -0.0745 0.4250 3.41 22 12326 1.3245 -0.03249 0.0002033 -0.0732 0.2570 3.19 22 12327 1.2291 -0.03339 0.0001850 0.0055 0.2482 2.73 6 12328 1.2354 -0.03276 0.0002498 -0.0500* 0.6402 3.91 9 12329 1.1587 -0.02998 0.0002850 -0.0500* 0.8344 3.39 12 12330 1.2193 -0.03124 0.0003188 -0.0879 0.6940 3.60 12 12331 1.2067 -0.03070 0.0002420 -0.0538 0.5210 1.76 14 12332 1.2640 -0.03385 0.0002108 -0.0500* 0.5344 2.03 14 12333 1.3681 -0.03430 0.0002046 -0.0730 0.1478 2.33 17 12334 1.2027 -0.02930 0.0001942 -0.0238 0.2846 3.34 18 12335 1.3430 -0.03520 0.0002112 -0.0754 0.3253 4.67 21 12336 1.3180 -0.03468 0.0002247 -0.0888 0.3926 4.09 22 12337 1.3640 -0.03421 0.0002080 -0.0771 0.3187 3.31 21 Notes 1) Coefficients marked with an asterisk (*) were specified rather than fitted. 2) The rms residual is found from the sum of the squared residuals, divided by the number of samples used in the fit minus the number of fitted coefficients. 2.7 XBTs by: S.R. Thompson XBT profiles during Discovery cruise 199 were collected using the Bathy Systems Inc. XBT program version 1.1 and SA-810 XBT controller, with the probes launched from a Sippican Corporation hand-held launcher. The inflection points calculated by the program were transmitted to the GTS network after each launch via the GOES satellite. ASCII versions of the raw data were transferred to the RVS level A using a diskette. An inter-comparison was carried out by comparing profiles made in a marked mixed layer with the surface temperature measured on the thermosalinograph in regions of low horizontal temperature gradient. Linear regression of TSG onto XBT temperature gave a slope of 0.99 and an uncertainty of 0.01, with an offset of 0.2° at 10°C. Launch 107 was a calibration run using the test probe. This yielded 14.85° for a resistor chosen to give a value of 15.0. Two problems were noted with the software:- 1) The bucket temperature information in the header does not appear to be saved. This means that if a file is not transmitted to the satellite immediately after the launch then the temperature must be re-entered in the header. 2) The column indicating whether the file has been transmitted sometimes fails to show a 'Y' after transmission. Information concerning all the successful launches is shown in the accompanying XBT station list (end of the report). All launches were T7 probes unless marked otherwise and breaks in the launch numbers indicate probe failures, of which there were nine (eight T7 and one T5). Launches 101 to 125 did not form part of the A11 section 2.8 Acoustic Doppler Current Profiler (ADCP) by: P.M. Saunders and R. Marsh The instrument used was a RDI 150 kHz unit, hull-mounted approximately 2m to port of the keel of the ship and approximately 33m aft of the bow at the waterline. On this cruise the firmware version was 17.10 and the data acquisition software was 2.48. For most of its operation the instrument was used in the water-tracking mode, recording 2 minute averaged data in 64 x 8m bins from 8m to 512m. On the shelf at the start and end of the cruise, the instrument was put into a mode in which both water and the bottom are tracked. Here 2 minute averaged data was collected in 50 x 4m bins from 6m to 200 m depth. The performance of the instrument was excellent throughout the cruise: on station, profiles were almost always recorded to 300m depth, and whilst steaming, except in the heaviest weather, profiles in excess of 200m were the norm. Data were passed in real time from the deck unit to a SUN workstation acquisition area: once a day, 24 hours of the data were read into the processing area. Our processing has much in common with that of Griffiths (1992) except in one or two important respects, but for completeness will be outlined here. Stage 0 was to capture the 24 hours of data and write it into an appropriate format. Stage 1 consisted of correcting the time base for instrument clock drift and changing the time stamp from end of data period to center of data period. Stage 2 consisted of applying misalignment corrections (to be described below), averaging data into 10 minute periods, merging with the ship's motion over the earth from GPS navigation and thereby deriving, by algebraic addition, current components averaged over the same interval. At this stage error velocities were displayed as time series to identify both depths of good data and periods of poor data: there were remarkably few of the latter. Stages 3 and 4 of the processing were novel: average profiles were constructed in approximate 4 hour chunks whose boundaries were selected by inspection and corresponded to 'on station' and 'steaming' activities. Data for maneuvering periods were excluded. The average profiles were identified by the station number, with the addition of the letter A to indicate the steaming period after the station. A cruise data set was constructed by appending the files together and we expect to employ this modest body of data in a combined analysis with the hydrographic data. For more detailed studies of the Ekman layer, for example, and the response of the upper ocean to storm force winds, the 10 minute data set will be utilized. As is well known, a key element in the determination of currents (water motion over the Earth's surface) from the ADCP is the ship's gyro. This allows the fore and aft and athwartships components of flow determined from the RDI instrument to be resolved into east and north components and so added to the ship's motion determined by navigation (GPS). The results are sensitive to gyro error, gyro drift, and the alignment of the transducers on the hull. In order to evaluate these errors, zigzag calibration exercises (Pollard and Read, 1989) were carried out on 4 occasions:- 24 December (courses 0°, 090°), 8 January (courses 045°, 135°), 21 January (courses 015°, 105°), and 31 January (courses 015°, 105°). The results from the first 3 calibration exercises showed a small increase in the misalignment angle from 0.5° to 1.0° to the right of the apparent gyro direction. On board the initial value of 0.55° was used in the preliminary analysis of the data. Ashore considerable post processing will be undertaken to correct for both directional and gyro errors (see the section 2.9c). References GRIFFITHS, G.1992. Handbook for VM-ADCP-PSTAR system as used on RRS Charles Darwin and RRS Discovery. James Rennell Centre for Ocean Circulation Internal document No.4, 24pp. POLLARD, R.T. and J.F.READ, 1989. A method for calibrating ship-mounted acoustic doppler profiles and the limitation of gyro compasses. Journal of Atmospheric and Oceanic Technology, 6, 859-865. 2.9 Navigation a) GPS-Trimble by: P.M. Saunders and M.G. Beney Navigation, i.e. ship position and velocity over the ground, was provided throughout the cruise by a Trimble GPS receiver. No rubidium clock was available so at least 3 satellites were required for a fix. The observations are interfaced via a level A microprocessor (see section 2.11 on computing) into the SUN acquisition system. In order to prevent hanging or crashing of the level A, which was of new design, the sample rate was set to 0 and data was logged at approximately 1 Hz. Editing of this data was carried out to exclude a small but tiresome number of zero times, zero latitudes, zero longitudes, northern hemisphere positions (!) or otherwise suspect data and sub-sampled at 30 second intervals. This data known as 'gps' was archived and provided coverage for approximately 95 % of the cruise. In order to complete the navigation data set for 100 % of the time, during periods of absent or inaccurate GPS fixes the ship's gyro and Emlog data were combined to give a dead reckoning position. Such data is flagged and the data is known as 'bestnav'. Transit satellite data were not used on the cruise. Positions were logged in port at the start of the cruise and a rms position error of approximately 30 m was found. Evidently selective availability was in operation at this time. Underway errors are known to be larger. b) Electromagnetic log and gyrocompass by: A.J. Taylor Ship speed is determined by a Chernikeeff log with sensor head approximately 0.25 m beyond the hull of the ship. Because of a sensor failure on the previous cruise a new unit was installed in Punta Arenas and zeroed whilst at the dock. Initially when underway a nominal calibration was applied, but at 11.0 kt smg as determined by a navigation unit (decca Mk52), the indicated speed was 12.24 kt, so a scaling was introduced to bring the two into agreement. The same adjustment was made to the port/starboard component. On January 8 the sensor head was rotated approximately 5° anti-clockwise to reduce a spurious athwartship drift of about 1.3 kt at full speed. Improved log calibrations will be obtained by comparison with ADCP data (including the zig-zags) but because this will have a minor impact on 'bestnav' calculations we do not anticipate recalculating navigation for this reason. Two S.G.Brown gyrocompass units (SGB1000) are installed on the Bridge. Because of a long lag noted with unit 1 on the previous cruise, unit 2 was employed for primary navigation throughout cruise 199. The output was logged via a level A microprocessor at 1 Hz and was free of gaps. The accuracy of heading is discussed in the following section. c) Ashtech GPS3DF Instrument by: S.R. Thompson This instrument, newly acquired for the cruise, measures not only the position but also the three dimensional attitude of the ship from the GPS system, i.e. ship's roll, pitch and, most significantly for the ADCP work, heading. The determination of attitude is performed by an array of four antennas approximately in the form of a square of side 8m. Data were logged in the deck unit of the receiver at 0.2 Hz frequency (because the level A failed to work reliably) and down loaded to the SUN workstations twice per day. King and Cooper (1993) have described details of the instrument, its installation and preliminary results on a 7 day trial cruise of RRS Discovery. They demonstrated that the gyro error is a function of ship's heading and also that it changes with time after a ship maneuver: in port they confirm the accuracy claimed by the manufacturer of 0.05°. On cruise 199 we elected to use the second of the two ship's Gyro compass units, (i.e. a different one from King and Cooper), and our preliminary results show that this instrument also experiences gyro error related to the ship's heading and time-dependent errors after maneuvering. Also long term drift of the gyro is apparent. For both instruments, these variations are of the order of 1°. Data quality control was implemented in the manner described by King and Cooper (loc cit). For reasons not currently understood only approximately one third of one minute averages of the difference between Ashtech and gyro headings contain data, far less than they encountered at the same latitude in the North Atlantic. Ten minute average differences have also been constructed and assembled in 5 day summaries. These will be used in post processing of the ADCP data and are expected to bring significant changes especially for underway estimates of currents. References KING, B.A. and E.B. COOPER, 1993. Comparison of ship's heading determined from an array of GPS antennas with heading from conventional gyrocompass measurements. Submitted to Deep-Sea Research. 2.10 Underway Observations a) Echosounding by: A.J. Taylor Equipment The bathymetry equipment installed on RRS Discovery consists of:- Hull mounted transducer, Precision Echosounding (PES) 'fish' transducer, and Simrad EA500 Hydrographic Echosounder. Operation The Simrad Echosounder was used during the cruise for bottom detection and determining the height of the CTD off the bottom during casts. While in bottom detection mode the depth values were passed via a RVS level A interface to the level C system for processing. Data were logged at a 30 second interval. The transducers were connected to the Simrad equipment via an external switch. A uniform sound velocity of 1500 meters/sec was used during the cruise. A visual display of the return echo was displayed on the Simrad VDU. Hardcopy output was produced on a color inkjet printer and a Waverley thermal line-scan recorder. Performance While on station and steaming during the initial few weeks of the cruise, the PES fish transducer was used. This gave good return signals on station and adequate return signals whilst steaming at 10 knots. After the second week the return signal when steaming deteriorated rapidly and the hull transducer was used whilst underway. Upon recovery of the fish on day 025 prior to steaming for Capetown, it was found that the lowest section of fairing was split in two. This was probably hitting the fish and the cause of noise whilst steaming. The fairing was replaced before being re-deployed on day 028, and a good signals were obtained whilst underway for the remainder of the cruise. When coming on station the PES fish sank considerably from its steaming depth: this resulted in a 17m offset between the PES fish and the hull transducer on the graphic display. The fish returned a lower depth than the hull transducer. The amount of cable submerged whilst on station was measured to be approximately 22m, thereby accounting for the offset. The Hewlett Packard inkjet printer developed a fault after one week and was replaced by the Waverley line-scan recorder. This was quite unreliable and was itself replaced, when a new inkjet printer was delivered by the Capetown pilot on 27 January. As is well known the automatic depth finder performance is adversely affected when the signal to noise ratio is small. In these circumstances the digitally recorded data is frequently unreliable. Given strip-chart records the situation can be recognized and rectified. Except for the first few and the last few days, such records are unavailable on cruise 199. Consequently the overall quality of the depth measurements is very disappointing. (Note added by P.M.Saunders, 9 Feb '93). b) Meteorological Measurements by: K.J. Heywood and P.K. Smith The meteorological monitoring system used on RRS Discovery comprises the following instruments:- * an R.M. Young Instruments Type 05103 wind velocity propeller - vane sensor, located on the foremast to port. * two Vector Instruments psychrometers, located on the foremast to starboard (serial numbers 1072 and 1073). * (1073 was replaced by 1071 during the cruise). * two Didcot cosine collector PAR sensors (spectral range 400-700nm) located port and starboard on the foremast (serial numbers 0150 and 0151 respectively). * two Kipp and Zonen total irradiance sensors located on the foremast to port and starboard (serial numbers 92015 and 92016 respectively). * an Eppley longwave pyrogeometer located on the foremast top pole (serial number 26207F3). * a hull-mounted RVS/RS Components platinum resistance thermometer, recording sea surface temperatures. * a V”is”l” DPA21 aneroid barometer, located in the main lab. * a Gill sonic anemometer located on the foremast to starboard. * a ship borne wave recorder. 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. Data were recorded as 1 minute averages. Processing The Unix shell script metexec0 was used to retrieve 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 direction output from the wind velocity sensor. Ship's navigation data including gyro heading (bestnav, derived from GPS and dead-reckoning) were merged with the met file by metexec2. Metexec3 and metexec4 were not normally used for this cruise. 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 files, metexecp was used to produce plots, and the Pstar program metflx was used to derive wind stress and heat fluxes. Calibration With the exception of the aneroid barometer and wind direction output from the wind velocity sensor where any conversion or calibration is performed by the metlogger PC 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 M1. Problems encountered Air temperatures The RVS PC display system showed slightly higher readings than expected. This was due to the calibration coefficients being only nominal values. Also the calibration file used a 2nd order polynomial, whereas the IOS calibration uses a 3rd order polynomial. Using the calibration data for each psychrometer, new values were calculated and entered into the calibration file. These gave good readings on the display. The correct 3 order coefficients were in the Pstar calibration file. On 29/12/92 (day 364) the port psychrometer data became very noisy. It was replaced and new calibration coefficients entered into the calibration file (/pstar/src/extras/cal/met 199. cal). There is a gap in the port data between 1600 hrs and 1845 hrs. No further problems occurred during the cruise. Long Wave Radiometer This gave good readings at the start of the cruise, but began giving some low readings during 1st January (day 367). The signal slowly deteriorated becoming more erratic. The battery was replaced on 16th January (day 382) and good readings were obtained for the rest of the cruise. Sonic Anemometer The Asymmetric Sonic Anemometer was mounted on the foremast with North facing forward. The system gave good readings. The system stores processed data on both hard disk and floppy disk. To store the raw data an optical disk was installed with a capacity of 20 days' data. There was some difficulty in setting up the software but eventually the optical disk recorded raw data. There was some complex interaction between the system clock and the optical disk software. As the software needs the time and date information in the data files and in naming the files, the software halts if the internal clock is in error. This error occurred between once in 3 days to 3 times in a day. Re-booting and resetting the time and date resumed normal operation. Ship Borne Wave Recorder The computer and associated software worked well during the cruise with very few errors. The signal amplification/conditioning unit showed a large d.c. offset and low amplitude signal for the Port Pressure Transducer. This transducer was flushed, which considerably reduced the d.c. offset and increased the signal amplitude. Further flushing produced a further improvement but there was still a small d.c. offset and the amplitude remained slightly smaller than the starboard pressure transducer. The last calibration was at the refit and a d.c. offset was noted then. Met Observations during the cruise Weather conditions during the cruise were remarkably clement, with the exception of a storm in mid January. The maximum wind speed observed was 28 ms-1 on 13th January, producing the largest waveheights. TABLE M1: Calibration coefficients for the met. sensors Measurement Calibration coeffs source if not IOS y=a+bx+cx2+dx3 a b c d Wind speed 0 0.1 0 0 mfr Wind dirn 0 1.0 0 0 mfr swet -21.63646 2.580562e-3 7.893778e-6 0.660868e-9 sdry -20.18834 9.733870e-4 7.835114e-6 0.525038e-9 up to day 364 pwet -23.71101 6.848060e-3 5.626587e-6 1.077627e-9 pdry -23.84735 5.788879e-3 5.648462e-6 0.907665e-9 after day 364 pwet -24.38268 6.720888e-3 5.840227e-6 0.969597e-9 pdry -23.36777 5.245053e-3 5.784058e-6 0.882978e-9 sea 0.26705 0.99189 2.9755e-4 0 RVS longwave 0 0.23364486 0 0 y=x/(ab) pPAR 5 12.86e-6 sPAR 5 12.87e-6 pirr 2 48.49e-3 sirr 2 43.63e-3 c) Thermosalinograph measurements by: S. Cunningham Instrument and Technique Continuous underway measurements of surface salinity and temperature were made with a Falmouth Scientific Inc. (FSI) shipboard mounted thermosalinograph (TSG). Salinity samples were drawn from the non-toxic sea water supply at four hourly intervals, and used to calibrate conductivities obtained from the TSG. The instrument was run continuously throughout the cruise. The TSG comprises of two FSI sensor 'modules', an Ocean Conductivity Module (OCM) and an Ocean Temperature Module (OTM) both fitted within the same laboratory housing. Sea surface temperature is measured by a second OTM situated on the suction side of the non-toxic supply in the forward hold. The non-toxic intake is 5 m below the sea surface. Data from the OCM and OTM modules are passed to a personal computer (pc). The pc imitates the traditional Level A system, passing it to Level B at 30 second intervals. Sensor Calibrations The temperature modules are installed pre-calibrated to a laboratory standard and laboratory calibration data are used to obtain four polynomial coefficients. A similar procedure is employed for the conductivity module. Underway Salinity Sampling Salinity samples were drawn from the non-toxic supply at four hourly intervals. These samples were then analyzed on a Guildline 8400 using standard sea water batch P120. Calibration of TSG Salinities against Underway Salinity Samples TSG conductivity measurements at 30 second interval were median de- spiked, discarding data more than 0.01 mmho/cm from a mean computed over 5 adjacent data values. Conductivity of the bottle samples was calculated at a pressure of 0 dbar and at the temperatures of the TSG OTM. The TSG data were merged onto the bottle data and the conductivity difference between the bottles and TSG calculated. After excluding outliers, a linear regression between the conductivities was determined and applied to the TSG values. TSG salinities were computed along with the difference from the bottle salinities. This difference was filtered with a Gaussian filter of half width 12 hours and normalized peak height of 0.38. TSG salinities were then corrected by adding the filtered difference. A plot of the corrected salinity and temperature at the surface for the entire cruise is shown in Figure 7. Estimate of the TSG accuracy and salinity residuals Due to particular difficulties with the instrument, the estimate of salinity residuals has been split into two portions. For the period day of year=359 to day=23 (389) the mean difference between the bottle and TSG salinities was -0.0009 with a standard deviation of 0.0145. For the period day=23 to day=32 the mean salinity difference was 0.0005 with a standard deviation of 0.02. Over the period from 23 0000Z to 27 0825Z the housing temperature sensor produced unreliable results. A current leakage was found between the platinum resistance thermometer and the surrounding seawater. This caused the probe to oxidize and eventually fail. At about the same time the pumps for the non-toxic supply failed and an alternative set were switched on. This caused a decrease in the flow rate and a corresponding increase in lag time for water from the non-toxic intake to reach the TSG, from approximately 5 to 10 minutes. Degradation of the conductivity results is likely. On day=26 at 0555Z the housing OTM was replaced. For the period 23 0000Z to 26 0555Z a reconstructed housing temperature was derived from the remote temperatures. Given the uncertainties in lag time and the alternative heating and cooling of the non-toxic supply through the ship (during this period for surface temperatures less than 20.2°C the supply is warmed and above that cooled) the reconstructed temperatures are not likely to be better than 0.2°C. The uncertainty probably accounts for most of the spread in the salinity residuals over this latter period. d) Satellite Image Acquisition and Processing by: M.P. Meredith and V.C. Cornell Equipment and function On this cruise equipment was installed for the capture, display and processing of polar-orbiting weather satellite imagery. This consisted of an omni-directional VHF antenna mounted on the main mast, a pre- amplifier to compensate for feeder cable losses of up to 10db, a Dartcom system II receiver, an 8-bit 15MHz microcontrolled interface to control the frequency and mode of the receiver, and an Apple Macintosh IIsi computer with the MacSat 2.1 software supplied jointly by Dartcom and Newcastle Computer Services. The equipment was used to receive data sent from the NOAA satellites 10, 11 and 12 via the Automatic Picture Transmission (APT) system at 137.50 and 137.62 MHz. Although the software allows the capture of geostationary weather satellite images, the hardware necessary for this was not present. No attempt was made to capture images from polar- orbiting satellites other than the NOAA series. The data collected were from the Advanced Very High Resolution Radiometer (AVHRR), a five-channel radiometer featuring one visible, two near-infra red and two thermal infra-red channels, though the APT system only allows for the visible channel plus one infra-red channel to be received. The APT system also reduces the spatial resolution of the data from its maximum of 1.1 km square at nadir to approximately 4 km square. Data from almost all the radiometers' swath width is captured with MacSat; an 800 x 800 pixel image covers approximately 3000 km square, and has a maximum of 256 digitization levels per pixel. Procedure During the cruise, most of the longer satellite passes (>12 minutes) were captured. Shorter passes generally did not contain enough noise- free data to warrant their capture. The vast majority of images were from the infra-red channel, since the previous cruise experienced serial error problems with the Auto Save function (the function enabling both channels to be acquired simultaneously), which led to the loss of the images. Thus only one of the two channels was available, and the infrared data were deemed more useful than the visible for our purposes. Once captured, the time/date, ship's position, and whether the satellite was in an ascending or descending pass was recorded, and a geographical overlay created for the image. This shows lines of latitude and longitude, ship's position at time of acquisition, and, if relevant, a coarse coastline. Three standard color palettes were created to enable depiction of sea brightness temperature. One would not suffice since the manual contrast stretch facilities of MacSat (adjusting the RGB response curves for the image) were found to be very cumbersome, and the Auto Contrast function is only useful for gray scale images. Color hardcopies were produced for each image by using the Mac's screen- dump tool. This creates a TeachText picture of the screen, which can then be printed to a postscript file, transferred to the Sun workstations using ftp, converted to a PCL file and outputted to the HP Paintjet printer. This was considered a better procedure than using MacSat's print option, since not only can the whole image be displayed on one A4 sheet, but the geographical overlay can be also be printed on the image. Some images were transferred to more sophisticated image processing software on the Suns; this, along with the image file format and file archiving, is discussed elsewhere. Problems Difficulties encountered on the previous cruise concerning the gross inaccuracy of the geographical overlay were to a large extent resolved. Updated files containing the Keplerian orbital elements for the satellites were obtained by fax from Newcastle Computer Services on two occasions as a matter of course, and on a third (1st Jan), when an error in the orbital element calculations became apparent. Also, the Mac's internal clock was corrected each day, since it gains approximately one second per day on GMT. Such an error is not insignificant for satellites travelling at 27,000 km/h, and would greatly affect the positioning of the overlay if left unaltered for a number of days. However, even with these measures being taken, the overlay could still be as much as a degree or two out, and the uncertainty should be borne in mind when considering images without coastline in them. Noise contamination of images was a frequent problem, and although MacSat has a noise reduction filter, this is of use only for presentation purposes and obviously cannot replace missing data values. Whether the problem was caused by atmospheric conditions, insufficient signal amplification or faulty hardware remains unknown. A further unsolved problem is the overlay tool's failure to plot lines of latitude for descending satellite passes. We think this can only be attributable to a bug in the program. Initially, difficulties were encountered with the loss of images due to serial errors during acquisition. This was caused by a slowing of the Mac to the point where it could not keep up with the incoming data stream, and was solved by ensuring that there were no telnet connections active, no print jobs queued and no Appleshare volumes present on the workspace at the time of capturing an image. Observations Several significant oceanographic features were observed in the satellite imagery captured during the course of the cruise. The retroflexion of the Falkland Current at the Brazil Current was clearly visible, and when the thermosalinograph (TSG) showed an increase in temperature, the MacSat image revealed a warm ring shed from the conflict of the two currents. Many of the images showed the position of the Subtropical Front to the north of the cruise track, and, towards the end of the cruise, the coastal upwelling region associated with the Benguela Current is clearly visible. An Agulhas ring was possibly observed, but not certainly, since cloud contamination partially obscures the feature. The cloud images also proved illuminating, especially during the severe storm encountered on the 13/14th January 1993. 2.11 Shipboard computing by: M.G. Beney and V.C. Cornell RVS logging System 'ABC' The RVS logging system comprises of 3 distinguishable parts or levels. Each level is referred to by one of the following letters A, B or C, and the whole system is called the 'ABC' system. A Level A consists of a microprocessor based intelligent interface with firmware which collects data from a piece of scientific equipment, checks and filters it, and outputs it as SMP (ship message protocol) formatted messages. There are two versions of dedicated Level A's, a MkI based on a 8085 processor using CEXEC as the operating system, and a MkII based on a 68000 processor running OS9 as the operating system. In addition there are pseudo Level A's which are PC's around which a piece of equipment it based, which are also capable of generating SMP messages. The Level B collects each of the Level A SMP messages and writes them to disk and backup cartridge tape. The Level B monitors the frequency of these messages, and besides providing a central display for the data messages also warns the operator when messages fail to appear. The Level B, which is based on a 68030 processor using OS9 as the operating system, collates the data and outputs it to the network. The Level C, which is a SUN IPC (4/40), takes this data and parses it into RVS data files. These data files are constructed on a RVS styled database for speed of access. The following list shows the instrument Level As and the variables which were logged by the Level C. The first column shows the name used by the Level A. Brackets after the Level A name indicate whether it was a MkI (1), MkII (2) or IBM compatible PC (PC), based Level A. The "adcp" data was collected directly by the Level C through one of its serial ports (ttya). The data was written to the data file named in column 2 with the variable names shown in column 3. Level A Datafile Variables BOTTLES(1) bottles code CTD_17C(2) ctd_17 press temp cond trans alt oxyc oxyt temp2 cond2 deltat nframs GPS_ATT(2) gps_att hdg pitch roll mrms brms attf sec GPS_TRIM(2) gps_trim lat lon pdop hvel hdg svc s1-s5 GYRO_RVS(2) gyro_rvs heading LOG_CHF(2) log_chf speedfa speedps METLOGGR(PC) metloggr winspd windir pwettemp pdrytemp swettemp sdrytemp seatemp ppar ptir spar stir lwave baro MX1107(1) mx1107 lat lon slt sln el it ct dist dir sat r status SIM500(2) sim500 uncdepth rpow angfa angps SURFLOG(PC) surflog temp_h temp_m cond WAVE(1) wave height WINCH(PC) winch cabltype cablout rate tension btension comp angle The following list shows data files which contained data directly collected by the Level C adcp_raw rawampl beamno bindepth adcp bindepth heading temp velew velns velvert velerr ampl good bottomew bottomns depth xbt depth temp The following datafiles were archived: relmov gps mx1107 bestnav bestdrf winch wave metloggr surflog adcp adcp_raw ctd and xbt. These RVS archives have only limited life and are only intended as (fall-) backups. Processing of data Virtually all of the data processing was performed using the interactive "pstar" suite of about 300 documented programs (Alderson et al,1991). This continuously updated system is installed on RVS ships as well as at labs ashore. RVS data files were converted to "pstar" data files using the program 'datapup'. Archiving of pstar files Archiving took place on a daily basis. Copies were made of all processed files on Sony erasable magneto-optical disks. These were mounted as standard unix file systems. In addition files were copied to Quarter Inch Cartridge (QIC) tape in both raw sequential and unix tar format. Six sides of optical disk data were taken ashore at the end of the cruise, totaling about 1.5 Gigabytes. Equipment available on cruise 199:- Personal Computers (Operating under Apple system 7.01) 3 Apple Macintosh Classics (40 Mb Hard Disc, 4Mb RAM) 1 Apple Macintosh ClassicII (40 Mb Hard Disc, 4Mb RAM) 1 Apple Macintosh II si (80 Mb Hard Disc, 5Mb RAM) The last was connected to a Dartcom System II satellite image receiver. Sun Workstations (Operating under Sunsoft's version 4.1.1) Node name Type Ram Hard Disc Peripherals (Mb) (Mb) discovery1 IPC 12 2x327 Exabyte drive 1x207 QIC 150 tape discovery2 IPC 12 1x207 Magneto/optic 1x1200 QIC 150 tape discovery3 Sparc stn 8 2x327 discovery4 Sparc stn 8 2x237 Output devices:- * Apple LaserWriter II (Mono Laser Printer). * Hewlett Packard Paintjet XL (InkJet Colour Plotter). * Tektronix 4693RGB (Thermal transfer plotter). * Hewlett Packard LaserJet III (Mono Laser Printer). * NEC Pinwriter P5 (Dot Matrix line printer). * Bruning Drum-type Pen Plotter. Networking All PCs, workstations and a number of output devices were connected to a thin Ethernet (10Base2) local area network. The Sun workstations have integral Ethernet interfaces, the Apple Macintoshes were connected via external SCSI Ethernet interfaces. References ALDERSON, S.G., GRIFFITHS, M.J., READ, J.F. and R.T. POLLARD, 1991. PEXEX PROCESSING SYSTEM, Internal document, Institute of Oceanographic Sciences Deacon Laboratory, about 450 pp. 2.12 Cruise diary by: P.M. Saunders 22 December Day 357/1992 RRS Discovery left Punta Arenas at 1700P (1400Z) with a pilot aboard, about 9 hours later than planned. All times are given as ship's time and the relation of ship's time to GMT stated whenever the relationship is altered. The delay was occasioned by the late arrival of the customs paperwork for the various items of airfreight. Amongst these was the CFC equipment which came on board, late on the 20 December. A new emlog was installed and an arbitrary calibration applied to yield reasonable ship's speed. The navigation and the Acoustic Doppler current Profiler (ADCP) were logged from departure. 23 December Day 358 Calm seas, some pitching motion as course is set 050° across the Argentine shelf 0430 (0730Z). At 0900 the first officer gave a safety briefing and this was followed by a science briefing by the PSO. At 1030 there was fire and boat drill, followed by a tour of the ship pointing out escape routes etc. Around 0130P (0430Z) the thermosalinograph was started up. At 0200 the Echosounder fish was streamed and after repairs to the fairing clips RRS Discovery resumed speed. 24 December Day 359 Given continuing fair weather it was decided to undertake an ADCP calibration exercise; this was performed between 1300 and 1600P. The results were satisfactory. See the ADCP account in this report. 25 December Day 360 A trial of the mid-ship winch was undertaken as station 12238 between 0628 and 0729P. A depth of 500m was reached and, after recovery, repairs were made to the winch scrolling gear and to the CTD, so that the exercise proved fruitful. Whilst the RRS Discovery continued northwards towards the latitude 45°S, crew and scientific party celebrated the festive occasion. 26 December Day 361 A test station 12239 was started in approximately 4000m of water at 0830P and was concluded about 1130P. The Rosette jammed after 5 firings and the ctd display was very noisy. The new altimeter unit worked well. Lanyard tensions were reduced and some cables replaced. An XBT was launched. A second station at the same location (45° 00'S 47° 30'W) 12240 to a depth of 2500m was more successful. With the wind 25-30kts samples were drawn on station, and at 1810P the ship turned west into an ADCP/XBT section. Some light rolling ensued. 27 December Day 362 A murky drizzly foggy morning turned into a bright sunny afternoon as 8 XBT/ADCP stations (12240-12247) were occupied in all. The wind died away and during passage, a tongue of cool surface water circa 8.5°C was encountered with warmer water 11.5°C to both west and east; it was the Falklands current. COMMENCEMENT OF THE A11 SECTION (45°S, 60°W) At 2000P station 122047, the first in the transoceanic section was begun: the water depth was about 250m and the initial objective was the Mid Atlantic Ridge nearly 1900 miles away. In order to assure good ADCP data, stations in the western boundary current were assigned a minimum duration of 2 hours. 28 December Day 363 Overnight stations in 500m, 1000m, and 1500m were occupied in calm seas the last within and close to the western edge of the Falklands current. Stations continued at 500m spacing down the slope, with spacing that varied between 3 and 30 n-miles. 29 December Day 364 Overnight the wind increased sharply and reached 35kts but by 0800P it decreased to 20kts under cloudless skies. Approximately 125kg of lead was removed from the rosette to reduce wire tension for the deeper casts. On station 12256 started near noon and completed at 1600P the deep western boundary current was detected; a nepheloid layer of thickness 400m defined it, at a depth below 4350m. The weather was fine enough for maintenance of the psychrometers on the foremast to be carried out. 30 December Day 365 Overnight the wind increased from the west to 30kts and the sea began to build. Station 12258 was begun at 0250P and after the cast had reached 2500m the ship's bow-thruster malfunctioned and the CTD/Rosette were recovered by 0450P. After repairs a second cast to the station was begun; it reached 5500m and was completed by 1240. (Subsequently it was learned that one of the motors that rotated the thruster needed parts which were not available on board.) Use was made of the railway to move the rosette to a protected position for sampling. This proved helpful. Station 12259 was carried out between 1650. and 2150 to a depth of 5630m; by now seas had built and some difficulties were encountered in hauling at the bottom of the cast. Fire and boat drill engaged those not involved directly in station work. 31 December Day 366 At 0000 the ship's master-clock decided it had started a new year and clock day was reset to 0. Some difficulties are to be expected in the subsequent processing of the data!! At this same time a station was started in about 5770m of water with strong SW squalls and high seas. This proved unwise. About two hours later it became quite evident that coupled with a strong current shear, the wire could no longer be controlled. Accordingly stn 12260 was abandoned at a depth of about 2800m. During the day wind and sea subsided and soon after midday stn 12261 was begun in 5900m of water. The station reached within 20m of the bottom where a very strong nepheloid layer was encountered and all gear was recovered by 1700. The performance of the winch in these circumstances was very satisfactory. At 2200 station 12262 was begun, again in nearly 5900m of water; the maximum expected water depth for the section was found between these latter casts. 1 January Day 001/1993 The New Year was welcomed whilst completing the station. Again a very strong nepheloid layer was seen. Unfortunately apart from this success little else went right on the day. At 0600 the ship hove to on station; RRS Discovery remained in this vicinity for the remainder of the day as both the engineers on board and those at the RVS base, over 6000n-mi away, attempted to diagnose and repair a defunct winch. The timing was inopportune, occurring on a bank holiday followed by a weekend. To ensure a quiet night, there was no work programme. 2 January Day 002 After considerable effort overnight the problem was identified. A faulty electrical component in the control logic circuit was found and replaced with an identical unit from the main winch, which was unserviceable. At about 0630 a series of shallow lowerings was begun: these were employed to fix the winch control settings, which were quite different from those prior to the breakdown. At 1620 station 12263 was begun in approximately 5750m of water, in the location arrived at approximately 36 hours earlier. The weather for this entire period had been (gallingly) fine. Immediately after launch the transmissometer failed, due to a cable connection adrift, but the cast was continued to full depth. On subsequent stations the transmissometer performed well. 3 January Day 003 The normal routine of station work was resumed with XBTs at a location midway between CTD casts. Mud waves were spotted and the chart recorder of the Echosounder which had been malfunctioning repaired and activated. At 1204 the level B system stopped logging and approximately 8 minutes of data was lost. This was during station 12265. The CTD data was recovered from the deck unit PS2 , but other data was lost. On the following station 12266 a strong nepheloid layer was again seen, suggesting strong currents on the abyssal plain. 4 January Day 004 Fine weather and a flat calm prevailed and the depth of the abyssal plain continued to shallow. A large school of pilot whales investigated RRS Discovery on station 12269, which was also noteworthy because the rosette jammed in position 13 and all shallow samples were lost. As the ship steamed away from the station the flanks of the Zapiola ridge were encountered at 2000P. The action of the Echosounder chart recorder continued erratically. CFC measurements were halted because of contamination. 5 January Day 005 On station 12270, 0100 - 0500P, the Rosette jammed at or near position 13 and samples were not collected at shallow depths. Since the previous station had experienced a similar sample loss, the failure to add a second shallow cast was unfortunate. Samples were collected in the rain but the protection of umbrellas was deemed unnecessary. After station 12271 a NEly wind came up and the ships progress was hindered. The ADCP lost penetration and subsequent analysis revealed the presence of the bogus "current following the ship" of 50-80 cms-1 always(?) seen when heading into a sea. On station 12272 the rosette again jammed at mid bottle so a second cast was made to 1500m depth. 6 January Day 006 On the overnight station 12273 bottles 9 10 11 were not cocked but the Rosette again malfunctioned so that after the samples were drawn the Rosette was stripped of all equipment for an overhaul. The spare Rosette (No 2) was mobilized and functioned satisfactorily for the next station. The wind and sea were subsiding but low temperatures prevailed as the RRS Discovery re-entered the sub-Antarctic zone. 7 January Day 007 On the overnight and morning stations the rosette performed satisfactorily but on station 12277 all bottles were closed below 1500m so a second cast was undertaken. Together the casts lasted from 1115 to 1745. After the Zapiola ridge with crests near 4900m, stations were now on the abyssal plain with depths over 5300m. At 1615 there was fire and boat drill. The performance of the ADCP continued poor and air was bled from the sensor pod without significant improvement. 8 January Day 008 The clocks were advanced 1 hour at 0001P so that ship time was now GMT- 2. Station 12278 at 3545W which was completed at 0200 in a flat calm had a depth of 5470m and was the maximum reached between the Zapiola ridge and the mid-Atlantic ridge; on this and subsequent stations the measurements differed substantially from the GEBCO chart. Mud waves continued to be seen. At 1530 in continuing flat calm seas the emlog, which had shown a cross track drift of about 1.3 kts, was rotated anti-clockwise about 5° to a more nearly correct direction. On the completion of station 12280 at 1740 a second ADCP zig-zag calibration exercise was begun to attempt to verify the gyro drift measured by the Ashtech GPS receiver. The experiment concluded at 2100 still in very calm seas. 9 January Day 009 In the early hours of the following morning a seal was spotted close to the ship and the barometer began to fall. At 0900 ships time the wind began to freshen from the Southeast and the barometer fell precipitately. At the start of station 12283 the wind was 45 kts from the south; almost immediately it began to diminish and by the end of the station it was only 25 kts. The lowering and handling of the ctd was straightforward despite the conditions. A comparison was made between measurements made on leg5 of SAVE near 45S and 41W (stns 290-293) and those on this cruise (12269-74). The salts and nitrates were in good agreement, the oxygens about 1.5% low and the silicates 3% low. 10 January Day 010 During the night the wind continued to come westerly and the considerable swell caused heavy rolling. This was uncomfortable for the ship's complement and on station led to very heavy snatch loadings. For the first time significant irregularities arose in the lay of the wire on the storage drum. At about 0745 station 12285 was commenced. About 4m down a high swell caught the Rosette and the wire was instantaneously so slack that it jumped off the sheave pair at the foot of the gantry. The wire was stopped off on the top of the Gantry, and inboard the wire was paid out, correctly rerouted and the load taken up again. The package was recovered on deck and a large kink located; about 20m of wire was cut off and the end reterminated. At the same time the Rosette No 1 was restored since No 2 had starting registering numerous misfires. The station was then restarted at 1000 after a delay of 2hours 15 minutes, and proceeded normally until about 3500m on recovery when attempts were made to improve the lay of the wire on the drum. Eventually the station was completed at 1500. Meanwhile the sea was subsiding. BAK reported a green flash at sunset. 11 January Day 011 The day started fair and concern for the CTD performance proved unnecessary. The regulation 3 stations were performed and the first colored Macsat images with a grid of lat and lon lines and the position of the ship were printed. Some but not all of these features had been available previously. Prior to station 12289 two lead weights (125kg) were restored to the Rosette in order to improve the shallow descent rate on the down cast. 12 January Day 012 A stiff northerly blew up during stn 12291 (1040- 1400) now in only 4400m of water. The next stations were accompanied by increasing rigor of the conditions. On both of them the Rosette was moved forward on the railway and sampling was undertaken on station. The wind and sea increased although during the evening the sky cleared. 13 January Day 013 At midnight the ship's clock, on which time this log is based, was advanced one hour to become GMT-1. On station 12293 in 3500m of water (0130-0430) conditions deteriorated markedly and by recovery the wind was blowing 45kts gusting to 55. The wind was now from west-northwest and despite clear skies continued to blow a gale; the seas were the largest seen on the cruise so far. We remained jogging, i.e. going slowly upwind, for the rest of the day. The ADCP functioned well and remarkable inertial oscillations were seen with an amplitude exceeding 50 cms-1. 14 January Day 014 After midnight the wind began to build again and by 0400 reached 50-60 kts, slowly backing to the south of west. The seas were, without exaggeration, mountainous with continuous spume blown from the crests. The pitching of the ship was severe but tolerable but the occasional heavy rolling was very uncomfortable. Not surprisingly the ADCP functioned only poorly. During daylight hours wind and seas moderated only very slowly and not until 2000 was the ship able to run before the seas towards the next station position. 15 January Day 015 At 0320 RRS Discovery arrived on station and the work programme was resumed. The seas were moderate - as was the performance of the Rosette. A second cast was made to 1000m to collect samples in the upper ocean. The decision was made to increase station spacing to 50 nautical miles for the foreseeable future. 16 January Day 016 A series of routine stations were made in shallow water depths, until on station 12298 (1050-1300) in 2500m of water the crest of the Mid Atlantic Ridge was reached. A mid-cruise break and PES survey had been planned but in view of the recent enforced delay this was no longer possible. By now the sea had quieted down and the skies were clear. THE TURNING POINT ON THE A11 SECTION (45°S, 15°W). At 1300 RRS Discovery steamed away on a course 059° towards the coast of South Africa and the conclusion of the section just over 1700 miles away. Within a short time a large iceberg was sighted (!) and at 1500 was passed at a range of 6 miles. 17 January Day 017 A day of calm seas and routine station work. Having crossed the ridge warmer water is encountered at all levels. A new inductive FSI conductivity cell is fitted to the CTD and yields encouraging results. A substitute Echosounder chart recorder is in action at last . Light rain fell about 1930. 18 January Day 018 At 0000 ships time the clocks are advanced 1hour so that ships time and GMT now agree. A sunny morning gives way to a rainy cloudy afternoon; by 1900 the wind is northerly blowing 25-30 kts. The umbrellas and their clamps on the Rosette frame are in use for the first time. The transmissometer develops intermittent and persistent noise; it is not clear whether the noise is oceanic or instrumental. Casts continue at a 50 mile spacing up to station 12306 (2015 - 2345). The surface temperature remains near 13 -14°C. I had expected it would rise before now. 19 January Day 019 An eventful day. After the station it was decided to resume a 42 mile spacing which had been characteristic of the leg on 45S. During steaming between stations 12307 and -8 two remarkable topographic features were encountered. The first of these was seen at 0930 (XBT 72) at location 40 58S and 6 01W; a seamount was detected rising to about 2300m from a sea floor near 3700m. This was tentatively identified as the flanks of the Admiral Zenker seamount. As the proposed site of the CTD station was neared, a second seamount was observed. This rose to a depth of 750m at 1054 at which point XBT 73 was dropped, 40° 48'S 5° 40'W. The seamount was flat topped (a Guyot) and for a distance of about 6 miles the depth was less than 1000m. A further 8 miles on, station 12308 was completed in 3700m of water. There is no indication of the seamount on any charts available to us; the name New Discovery Seamount is proposed. An overcast morning gave way to a sunny day although a brisk NW'ly wind persisted. 20 January Day 020 Station 12310 started in conventional fashion just before 0100, but as the Rosette was raised towards the surface a wave carried it upwards, the wire went slack and jumped off the sheave pair at the foot of the Gantry. This was a repeat of the event of station 12285 on the 10th of January. Eventually the package was recovered, 35m of wire removed, a new termination made and the cast restarted about 0300. For much of the day a moderate Westerly swell persisted and made the station work slightly difficult for the winch drivers. At the end of station 12311 when the package was recovered a kink was found in the wire which required cutting off about 10m of wire and a retermination - for the second time in the day. 21 January Day 021 During the night the swell diminished and station 12313 in over 5000m of water allowed the wire lay on the drum to be improved substantially. Surface water temperatures have now risen to 16°C but the absence of a marked subtropical convergence (RRS Discovery at 0700 is at 38.7S) has surprised a number on board. After station number 12315 we crossed the Greenwich Meridian at 2025 , a minor milestone. The crossing was made at the start of the third ADCP calibration exercise 2020 -2300 in which alternate courses were 015° and 105°. 22 January Day 022 The station work continues. After station 12316 maintenance work was carried out on the rosette and CTD cabling was replaced. Nevertheless a noisy transmissometer record was obtained. Shortly before station 12318 the surface salinity exceeded 35 for the first time (near 37°S 2°E). 23 January Day 023 Calm seas continue but the station spacing is augmented to 60 miles in order to anticipate a potential medical emergency and permit a dash to Cape Town if required. During the course of the day a remarkable lens of cool saline water is seen by XBTs 85-88 and CTD station 12320 and approximately 100 miles across. This takes the form of a 600m deep thermostad of temperature 13.5°C and salinity 35.2 which is capped by warmer fresher water. There is speculation that this is the remnant of an Agulhas ring, shed in the retroflection zone which has overwintered south of the convergence. But it is much cooler and fresher than any observed before. After passage through the ring the water freshens to 34.95 and temperature 18.5°C; perhaps Deacon's assertion (1937) that the seasonal migration of the sub-tropical convergence is large in this area with a maximum northwards location in summer is being verified on the cruise. At about 1930 there is an abrupt jump on the thermosalinograph. The salinity rises to 36 and the temperature to 20°C. Hallelujah! The latitude is 35° 40'S and the longitude 5° 00'E. 24 January Day 024 The routine continues in calm clear subtropical weather with 60 mile spacing of the stations. Even underway the ADCP penetration is 300m. For the past few days the winch operation under light loads has been erratic; lets hope it lasts to Cape Town. In the late afternoon a Barbecue on the after deck whist the ship was on station 12324 was a pleasant social occasion. 25 January Day 025 Today we passed through what is certainly an Agulhas ring. It took 20 hours and involved stations 12325 and 12326 and XBTs 097 - 102. The 15°C isotherm went from a depth of 100m or less outside the ring to over 350m within the ring. The extreme locations were 33° 49'S, 8 48°E to 33° 07'S, 10° 44'E, a distance of 105 n-miles. Both of the stations involved had problems. On station 12325 the rosette firing hung up at bottle 11; there were no samples above 1500m. Consequently a second cast was made to 1500m. On station 12326 a number of the hydraulic units shut down after start-up, attributed to a frozen cable-hauler, and the cast was delayed 30 minutes. During this cast the decision was made to proceed to Cape Town to put ashore the PSO whose medical condition was causing him and others concern. Prior to departure the Echosounding fish was brought on board. It was decided to launch XBTs every 2 hours on the way in. Dr. King agreed to act as PSO. 26 January Day 026 XBTs continued as RRS Discovery steamed into a stiff SE'ly wind, and later a 3 kt current. For only the second time in the cruise the ADCP data return was zero. The thermosalinograph failed due to a defective temperature element, which was replaced. It had given poor data for 3 days. 27 January Day 027 Clocks were advanced one hour to bring ships time to GMT + 2. Just outside Cape Town harbour, 2 miles off Green Point and at about 1200, the ship was met by a small boat and the PSO was put aboard. By 1330 the ship was underway and a speedy northward passage at an average speed of 13.5 kt was then made, with XBTs at 4 hourly intervals, to reach the eastern of the line and work south-westward from there toward station 12326. The ADCP housing was bled of air and repairs were made to clips on the Echosounder fairing. A new printer was installed for the Echosounder and for the first time excellent records were obtained 28 January Day 028 Station 12327 was commenced in 230m of water at 0730 and by 0800 was completed. Because of the pressure of time the decision was made NOT to remain for a minimum of two hours on station to obtain good ADCP records as we had in the WBC. On the following station in just under 500m of water, at 1045 the Echosounding fish was deployed. Thereafter stations were occupied at water-depth increments of 500m and at distances of between 10 and 40 n-miles, down the slope. The last station occupied on this day was 12332 in 2500m of water. 29 January Day 029 A superb calm day with a green flash at sunrise. Four stations were occupied today with the last, 12336, in a depth of water just under 4000m. 30 January Day 030 Today between 0530 and 0900 the last station 12337 was occupied at a distance of only 43 n-miles from station 12326. Thus the line is satisfactorily completed - a tribute not only to the entire scientific party but also to the entire ship's complement. END OF A11 SECTION Between 1200 and 1530 winch trials were carried out with the aim of improving the performance of the inboard compensation unit, and also to test the performance of the Mk 5 CTD. Unfortunately neither was successful, and at the end of them the Echosounding fish was recovered and course was set for Cape Town. Underway data logging was concluded at 2400 and watches were stood down at 1600. 31 January Day 031 In light winds the ship made good speed so that by 1720 it was possible to start the fourth ADCP calibration exercise of the cruise. The zig- zags were conducted on courses 105° and 015° respectively and ended at 2040. 1 February Day 032 The ship docked in Cape Town at 0830, concluding a most successful cruise. Acknowledgements This cruise, a UK contribution to the World Ocean Circulation Experiment (WOCE), was made possible by the parent body of (almost) all of the participants, namely The Natural Environment Research Council. Substantial support was also furnished by Ministry of Defense through the MoD/Research Council's joint scheme and by the generous provision of XBTs from DNOM, Taunton. The scientific party is grateful to the professional dedication of the Master, Captain Keith Avery, the officers and entire crew of RRS Discovery, - especially for the smooth running of a long cruise encompassing both Xmas and the New Year. We also wish to acknowledge the support of the shore-side staff of the Research Vessels Base (Barry) and Mr. R. Bonner (IOSDL) for their expertise in the mobilization and demobilization of the cruise in distant ports. CTD STATION LIST date time, gmt depth, m Samples Stn Cast mmddyy start bottom end latitude longitude uncwtr ht off wire max p no notes 12247 1 122792 2300 2312 2340 44 58.99 S 60 00.12 W 237 5 233 235 6 CFC 12248 1 122892 0143 0206 0229 44 59.66 S 59 56.35 W 476 8 455 461 9 CFC 12249 1 122892 0424 0458 0532 44 58.98 S 59 46.96 W 1007 9 963 971 10 12250 1 122892 0936 1018 1109 44 58.85 S 59 07.14 W 1504 11 1480 1481 11 12251 1 122892 1429 1516 1622 44 59.38 S 58 33.03 W 1908 10 1860 1891 7 CFC 12252 1 122892 1739 1831 1948 44 59.28 S 58 24.85 W 2603 10 2563 2613 18 12253 1 122892 2112 2216 2338 44 59.93 S 58 21.71 W 3139 14 3080 3137 19 12254 1 122992 0216 0325 0511 44 59.68 S 57 49.19 W 3447 9 3399 3455 19 12255 1 122992 0915 1050 1258 45 00.59 S 57 24.73 W 4011 7 3965 4049 19 I 12256 1 122992 1509 1652 1844 45 01.16 S 56 59.79 W 4773 11 4755 4841 23 CFC 12257 1 123092 2209 0012 0240 45 01.26 S 56 29.90 W 5304 11 5313 5399 22 CFC 12258 2 123092 0817 1018 1239 45 01.47 S 55 45.28 W 5497 14 5555 5609 24 12259 1 123092 1659 1904 2144 44 58.85 S 54 47.41 W 5648 28 5836 5765 24 CFC 12260 1 123192 0323 0501 0630 45 01.23 S 53 50.68 W 5784 -99 3121 2763 9 12261 1 123192 1529 1739 1955 44 56.33 S 52 49.00 W 5901 21 5977 6037 24 CFC 12262 1 010193 0108 0320 0528 45 02.14 S 51 44.47 W 5916 18 5950 6051 23 12263 1 010293 1919 2118 2331 45 02.34 S 50 44.74 W 5763 19 5795 5893 24 CFC 12264 1 010393 0324 0512 0710 45 01.47 S 49 45.10 W 5562 11 5547 5685 24 12265 1 010393 1108 1259 1510 45 00.23 S 48 46.16 W 5390 10 5373 5499 24 CFC 12266 1 010393 1857 2047 2246 44 59.84 S 47 45.75 W 5271 18 5245 5367 23 12267 1 010493 0248 0436 0639 45 00.58 S 46 45.22 W 5206 4 5185 5311 20 12268 1 010493 1027 1208 1402 45 00.17 S 45 45.26 W 5127 12 5100 5223 23 12269 1 010493 1750 1931 2124 44 59.33 S 44 44.39 W 5088 6 5065 5179 13 12270 1 010593 0110 0256 0452 44 59.89 S 43 45.12 W 4899 12 4866 4977 15 12271 1 010593 0855 1050 1253 44 59.98 S 42 45.23 W 5201 5 5205 5305 24 CFC 12272 1 010593 1808 2001 2208 44 59.62 S 41 45.07 W 4964 16 4946 5019 9 12272 2 010593 2235 2315 0006 44 57.87 S 41 45.72 W 4924 -99 1500 1513 16 12273 1 010693 0541 0740 0948 45 00.10 S 40 45.44 W 4980 13 4960 5055 24 CFC 12274 1 010693 1427 1606 1801 44 58.87 S 39 45.79 W 4990 6 4996 5075 23 12275 1 010793 2229 0010 0202 44 58.90 S 38 43.82 W 4866 8 4842 4941 23 CFC 12276 1 010793 0612 0805 1011 45 00.55 S 37 42.91 W 5123 12 5155 5215 23 12277 1 010793 1419 1611 1821 44 59.23 S 36 44.71 W 5328 6 5334 5457 24 12277 2 010793 1925 2002 2037 44 59.72 S 36 45.03 W 5328 -99 1250 1264 12 12278 1 010893 0027 0214 0417 44 59.43 S 35 45.06 W 5490 11 5470 5607 24 12279 1 010893 0824 1013 1209 44 59.82 S 34 45.83 W 5311 8 5290 5413 23 CFC 12280 1 010893 1557 1742 1936 44 59.47 S 33 46.29 W 5172 4 5155 5267 24 12281 1 010993 0115 0258 0450 45 00.88 S 32 46.44 W 5124 9 5092 5209 23 CFC 12282 1 010993 0850 1035 1232 45 00.11 S 31 43.23 W 5187 8 5170 5281 23 12283 1 010993 1649 1847 2043 44 59.33 S 30 45.53 W 5096 12 5070 5185 24 CFC 12284 1 011093 0140 0346 0545 44 59.93 S 29 46.54 W 4933 15 4938 5009 22 12285 1 011093 1209 1410 1659 44 59.87 S 28 44.49 W 4726 8 4695 4795 22 CFC 12286 1 011093 2048 2234 0021 45 00.32 S 27 46.34 W 4532 10 4510 4587 21 12287 1 011193 0420 0603 0758 45 00.49 S 26 44.13 W 4764 9 4755 4853 23 CFC 12288 1 011193 1133 1326 1515 44 59.58 S 25 43.88 W 4648 8 4612 4717 21 I 12289 1 011193 1904 2040 2225 45 01.11 S 24 44.97 W 4523 11 4570 4605 21 CFC 12290 1 011293 0201 0334 0513 44 59.25 S 23 43.44 W 4221 12 4200 4259 20 12291 1 011293 0848 1019 1158 44 59.65 S 22 45.72 W 4438 14 4405 4471 22 CFC 12292 1 011293 1608 1743 1918 45 00.89 S 21 44.24 W 3981 7 4108 4185 20 12293 1 011393 0028 0156 0321 44 59.69 S 20 42.84 W 3527 17 3534 3599 22 CFC 12294 1 011593 0425 0554 0727 44 59.28 S 19 44.55 W 3830 15 3850 3895 10 12294 2 011593 0810 0844 0917 44 58.70 S 19 45.83 W 3931 -99 1000 1008 12 12295 1 011593 1338 1522 1656 45 00.91 S 18 34.32 W 3760 17 3906 3929 20 CFC 12296 1 011593 2106 2229 2353 45 00.03 S 17 23.82 W -999 23 3710 3787 21 12297 1 011693 0419 0547 0708 44 59.30 S 16 13.50 W 3322 10 3290 3327 19 CFC 12298 1 011693 1145 1253 1357 45 00.39 S 15 00.50 W 2757 23 2715 2747 17 12299 1 011693 1824 1953 2114 44 32.66 S 14 00.18 W 3300 10 3335 3399 18 CFC 12300 1 011793 0133 0258 0421 44 06.73 S 13 00.66 W 3727 12 3658 3733 19 12301 1 011793 0823 0946 1115 43 40.01 S 12 01.24 W 3612 12 3650 3709 20 CFC 12302 1 011793 1518 1644 1815 43 14.00 S 11 02.67 W 4006 10 3970 4049 19 CFC 12303 1 011793 2233 2355 0124 42 49.20 S 10 05.30 W 3840 11 3810 3877 22 CFC 12304 1 011893 0547 0712 0847 42 22.27 S 9 05.17 W 3841 7 3765 3843 19 12305 1 011893 1302 1430 1558 41 55.79 S 8 10.09 W 3991 10 3964 4043 21 CFC,I 12306 1 011893 2023 2151 2320 41 31.40 S 7 12.31 W 3715 68 3730 3769 19 12307 1 011993 0311 0449 0624 41 09.32 S 6 24.00 W 4066 2 4065 4123 20 CFC 12308 1 011993 1218 1343 1512 40 42.26 S 5 26.50 W 3718 10 3680 3755 20 12309 1 011993 1809 1938 2102 40 25.38 S 4 49.82 W 3979 10 3964 4023 19 CFC 12310 1 012093 0252 0449 0638 40 04.98 S 4 02.44 W 4627 6 4605 4709 22 12311 1 012093 1030 1229 1414 39 43.22 S 3 15.24 W 4488 11 4507 4593 24 CFC 12312 1 012093 1758 1941 2133 39 21.31 S 2 28.67 W 4561 12 4535 4631 23 12313 1 012193 0107 0255 0443 38 58.22 S 1 41.36 W 4924 11 5038 5151 23 12314 1 012193 0814 1012 1207 38 36.46 S 0 56.54 W 4983 11 4996 5107 22 12315 1 012193 1539 1733 1933 38 14.88 S 0 09.74 W 5184 10 5180 5299 23 CFC 12316 1 012293 0013 0202 0359 37 52.89 S 0 35.51 E 5013 10 4997 5107 22 12317 1 012293 0733 0918 1107 37 31.26 S 1 21.00 E 5074 10 5080 5193 22 CFC 12318 1 012293 1435 1627 1825 37 09.08 S 2 06.23 E 5106 5 5135 5209 23 12319 1 012293 2146 2335 0127 36 47.42 S 2 51.51 E 5154 10 5155 5267 24 CFC 12320 1 012393 0623 0815 1019 36 16.60 S 3 55.53 E 5190 12 5195 5305 23 CFC 12321 1 012393 1522 1715 1903 35 44.79 S 5 00.15 E 5201 12 5220 5307 23 CFC 12322 1 012493 2357 0200 0404 35 12.76 S 6 02.17 E 5257 10 5270 5369 24 12323 1 012493 0915 1058 1247 34 42.33 S 7 04.20 E 5219 10 5220 5337 24 CFC 12324 1 012493 1751 1943 2159 34 09.16 S 8 07.24 E 4989 9 5035 5103 22 12325 1 012593 0301 0457 0638 33 40.82 S 9 08.81 E 5017 5 5065 5135 23 CFC 12325 2 012593 0820 0903 0955 33 40.25 S 9 07.97 E 5019 -99 1500 1502 23 CFC 12326 1 012593 1547 1750 1944 33 05.25 S 10 08.97 E 4999 4 5075 5103 23 CFC 12327 1 012893 0533 0550 0604 30 13.59 S 15 37.16 E 238 4 230 233 6 CFC 12328 1 012893 0849 0915 0939 30 28.27 S 15 08.12 E 471 8 465 465 9 CFC 12329 1 012893 1101 1138 1209 30 34.23 S 14 58.84 E 999 9 992 997 12 CFC 12330 1 012893 1325 1411 1454 30 40.29 S 14 47.44 E 1497 10 1505 1495 12 12331 1 012893 1612 1703 1759 30 45.31 S 14 38.59 E 1986 10 1965 1985 14 CFC 12332 1 012893 1948 2047 2145 30 53.45 S 14 22.05 E 2498 10 2460 2503 16 12333 1 012993 0046 0151 0301 31 14.14 S 13 44.98 E 3044 9 3012 3063 18 CFC 12334 1 012993 0611 0733 0856 31 32.54 S 13 09.65 E 3497 10 3475 3531 18 12335 1 012993 1303 1427 1553 31 57.82 S 12 21.75 E 4002 7 3980 4055 21 CFC,I 12336 1 012993 1955 2127 2305 32 19.16 S 11 38.47 E 4412 9 4400 4483 22 12337 1 013093 0330 0521 0700 32 42.47 S 10 53.87 E 4809 9 4795 4901 22 CFC Notes 1) Position is reported for the time at the bottom of the cast 2) Salinity, oxygen, silicate, phosphate, nitrate+nitrite were sampled for all bottles 3) CFC denotes CFC-11, 12 and 113 and I denotes Iodine XBT STATION LIST date time, gmt depth, m Samples Stn Cast mmddyy start bottom end latitude longitude uncwtr ht off wire max p no notes 1 1 122692 1444 45 01.15 S 57 30.29 W 3872 -99 -999 1833 T5 2 1 122792 0025 45 00.03 S 57 59.88 W 3206 -99 -999 1833 T5 3 1 122792 0501 44 59.52 S 58 29.94 W 2059 -99 -999 1833 T5 4 1 122792 0924 44 59.76 S 58 59.97 W 1558 -99 -999 1793 T5 5 1 122792 1304 44 59.81 S 59 14.85 W 1412 -99 -999 1431 T5 6 1 122792 1617 44 59.68 S 59 30.39 W 1222 -99 -999 1241 T5 7 1 122792 1937 44 59.22 S 59 45.30 W 1047 -99 -999 1063 T5 8 1 122792 2248 44 59.28 S 59 59.68 W 238 -99 -999 237 9 1 122892 0719 44 57.28 S 59 33.73 W 1200 -99 -999 763 10 1 122892 0826 44 58.87 S 59 18.64 W 1389 -99 -999 763 11 1 122892 1229 44 57.48 S 58 57.00 W 1610 -99 -999 763 14 1 122892 2011 44 59.49 S 58 23.29 W 2705 -99 -999 763 15 1 122992 0032 44 59.40 S 58 10.75 W 3376 -99 -999 763 16 1 122992 0111 44 59.72 S 58 01.56 W 3110 -99 -999 763 17 1 122992 0753 44 58.68 S 57 36.32 W 4200 -99 -999 763 18 1 122992 1404 45 00.70 S 57 11.82 W 4333 -99 -999 763 19 1 122992 2011 45 00.97 S 56 44.38 W 5073 -99 -999 763 20 1 123092 0408 45 01.40 S 56 07.95 W 5380 -99 -999 763 21 1 123092 1510 45 00.75 S 55 07.73 W 5550 -99 -999 923 22 1 123192 0020 44 58.77 S 54 14.43 W 5731 -99 -999 763 23 1 123192 1236 44 59.45 S 53 14.63 W 5846 -99 -999 763 24 1 123192 2244 44 54.19 S 52 15.82 W 5954 -99 -999 763 25 1 010193 0727 45 02.18 S 51 15.42 W 5844 -99 -999 763 26 1 010393 0119 45 02.76 S 50 17.40 W 5685 -99 -999 763 27 1 010393 0909 45 00.68 S 49 15.04 W 5440 -99 -999 763 28 1 010393 1717 45 00.19 S 48 12.44 W 5300 -99 -999 763 29 1 010493 0049 44 59.50 S 47 14.55 W 5572 -99 -999 763 30 1 010493 0833 44 59.57 S 46 14.03 W 5860 -99 -999 763 31 1 010493 1600 45 00.22 S 45 14.28 W 5026 -99 -999 763 32 1 010493 2313 44 59.68 S 44 15.05 W 4946 -99 -999 763 33 1 010593 0626 45 00.00 S 43 15.00 W 4980 -99 -999 763 34 1 010593 1551 44 59.83 S 42 11.74 W 5060 -99 -999 763 35 1 010693 0325 44 58.86 S 41 12.86 W 4919 -99 -999 763 36 1 010693 1238 44 58.62 S 40 05.32 W 4804 -99 -999 763 37 1 010693 2024 44 59.05 S 39 14.19 W 5000 -99 -999 763 38 1 010793 0424 44 59.00 S 38 11.16 W 4752 -99 -999 763 39 1 010793 1207 45 01.85 S 37 13.04 W 5151 -99 -999 763 40 1 010793 2233 44 59.11 S 36 15.31 W 5460 -99 -999 763 41 1 010893 0616 44 59.75 S 35 14.73 W 5475 -99 -999 763 42 1 010893 1406 44 59.89 S 34 16.15 W 5220 -99 -999 763 43 1 010893 2302 44 56.96 S 33 15.18 W 5051 -99 -999 763 44 1 010993 0657 45 00.48 S 32 14.31 W 5050 -99 -999 763 45 1 010993 1424 44 59.78 S 31 13.73 W 5170 -99 -999 763 46 1 010993 2321 44 56.24 S 30 15.39 W 4945 -99 -999 763 49 1 011093 0801 45 00.27 S 29 10.79 W 4638 -99 -999 763 50 1 011093 1847 44 59.53 S 28 14.82 W 4650 -99 -999 763 51 1 011193 0236 45 00.69 S 27 12.41 W 4684 -99 -999 763 52 1 011193 0936 44 58.39 S 26 14.51 W 4780 -99 -999 763 53 1 011193 1657 44 59.70 S 25 17.14 W -999 -99 -999 763 54 1 011293 0013 45 01.56 S 24 14.14 W 4620 -99 -999 763 55 1 011293 0649 44 58.98 S 23 15.42 W 4550 -99 -999 763 56 1 011293 1410 45 00.38 S 22 13.35 W 3638 -99 -999 763 57 1 011293 2207 45 00.06 S 21 15.61 W 4134 -99 -999 763 58 1 011593 0237 44 59.22 S 20 13.85 W 3003 -99 -999 763 59 1 011593 1136 44 59.68 S 19 08.56 W 3756 -99 -999 763 60 1 011593 1851 45 00.33 S 18 00.20 W 3500 -99 -999 763 61 1 011693 0215 44 59.82 S 16 45.68 W 3800 -99 -999 763 62 1 011693 0934 44 58.69 S 15 36.62 W 1863 -99 -999 763 63 1 011693 1601 44 47.01 S 14 31.47 W 3500 -99 -999 763 64 1 011693 2318 44 20.51 S 13 30.92 W 3237 -99 -999 763 65 1 011793 0614 43 54.69 S 12 32.06 W 3456 -99 -999 763 66 1 011793 1308 43 27.77 S 11 33.83 W 3675 -99 -999 763 67 1 011793 2022 43 01.46 S 10 34.51 W 3740 -99 -999 763 68 1 011893 0333 42 35.99 S 9 36.56 W 3670 -99 -999 763 69 1 011893 1056 42 08.83 S 8 37.80 W 3888 -99 -999 763 70 1 011893 1808 41 42.77 S 7 42.57 W 3748 -99 -999 763 71 1 011993 0115 41 20.75 S 6 49.00 W 3822 -99 -999 763 72 1 011993 0920 40 57.97 S 6 00.83 W 3507 -99 -999 763 73 1 011993 1056 40 48.49 S 5 39.99 W 800 -99 -999 763 74 1 011993 1620 40 36.56 S 5 13.18 W 3804 -99 -999 763 75 1 011993 2254 40 14.11 S 4 26.07 W 3559 -99 -999 763 76 1 012093 0843 39 53.47 S 3 38.92 W 4327 -99 -999 763 77 1 012093 1607 39 31.75 S 2 49.93 W 4400 -99 -999 763 78 1 012093 2320 39 09.62 S 2 05.28 W 4520 -99 -999 763 79 1 012193 0627 38 46.25 S 1 19.64 W 5306 -99 -999 763 80 1 012193 1351 38 25.54 S 0 33.41 W 5800 -99 -999 763 81 1 012193 2151 38 05.02 S 0 12.46 E 5147 -99 -999 763 82 1 012293 0544 37 40.57 S 0 58.21 E 5250 -99 -999 763 83 1 012293 1252 37 20.15 S 1 42.27 E 5050 -99 -999 763 84 1 012293 2005 36 57.74 S 2 27.23 E 4901 -99 -999 763 85 1 012393 0304 36 35.92 S 3 12.92 E 5200 -99 -999 763 86 1 012393 0448 36 24.64 S 3 35.94 E 5185 -99 -999 811 T5 87 1 012393 0456 36 23.71 S 3 37.77 E 5185 -99 -999 763 88 1 012393 1152 36 05.99 S 4 14.75 E 5086 -99 -999 763 89 1 012393 1344 35 54.63 S 4 38.80 E 5086 -99 -999 763 90 1 012393 2039 35 34.53 S 5 18.52 E 5196 -99 -999 763 91 1 012393 2216 35 23.98 S 5 40.47 E 5218 -99 -999 763 92 1 012493 0543 35 01.66 S 6 20.62 E 5350 -99 -999 763 93 1 012493 0732 34 51.83 S 6 43.58 E 5230 -99 -999 763 94 1 012493 1429 34 31.35 S 7 24.46 E 5128 -99 -999 763 95 1 012493 1603 34 20.72 S 7 44.96 E 5197 -99 -999 763 96 1 012493 2336 33 59.47 S 8 26.69 E -999 -99 -999 763 97 1 012593 0123 33 48.81 S 8 48.91 E 5047 -99 -999 763 98 1 012593 1144 33 28.67 S 9 27.81 E 4968 -99 -999 763 99 1 012593 1325 33 17.37 S 9 48.52 E 4994 -99 -999 1127 T5 100 1 012593 1335 33 16.45 S 9 50.40 E 4988 -99 -999 1832 T5 101 1 012593 2104 33 05.15 S 10 25.95 E 4800 -99 -999 1832 T5 102 1 012593 2240 33 07.38 S 10 44.54 E 4752 -99 -999 763 103 1 012693 0002 33 09.10 S 11 03.09 E 4850 -99 -999 763 104 1 012693 0203 33 11.80 S 11 30.40 E 4800 -99 -999 763 105 1 012693 0400 33 14.49 S 11 54.79 E -999 -99 -999 763 106 1 012693 0559 33 16.76 S 12 18.73 E -999 -99 -999 763 108 1 012693 0802 33 18.47 S 12 43.31 E 4695 -99 -999 763 109 1 012693 0957 33 20.03 S 13 06.57 E -999 -99 -999 763 110 1 012693 1204 33 22.71 S 13 32.09 E 5000 -99 -999 763 111 1 012693 1400 33 25.48 S 13 56.13 E -999 -99 -999 763 112 1 012693 1559 33 27.22 S 14 21.13 E 4350 -99 -999 763 113 1 012693 1800 33 29.98 S 14 46.22 E 4350 -99 -999 763 114 1 012693 1959 33 32.10 S 15 11.23 E -999 -99 -999 763 115 1 012693 2204 33 34.89 S 15 36.97 E 3460 -99 -999 763 116 1 012793 0006 33 37.37 S 16 00.34 E 3407 -99 -999 763 117 1 012793 0200 33 40.10 S 16 25.13 E 3000 -99 -999 763 118 1 012793 0400 33 42.98 S 16 54.44 E 1800 -99 -999 763 119 1 012793 0600 33 45.64 S 17 25.31 E 530 -99 -999 599 120 1 012793 1309 33 22.96 S 17 50.72 E 163 -99 -999 204 121 1 012793 1500 33 01.59 S 17 35.71 E 250 -99 -999 305 123 1 012793 1908 32 13.98 S 16 59.59 E 281 -99 -999 443 124 1 012793 2259 31 28.25 S 16 28.93 E 363 -99 -999 405 125 1 012893 0300 30 39.45 S 15 54.50 E 190 -99 -999 255 126 1 012893 0732 30 21.31 S 15 21.89 E 280 -99 -999 340 127 1 012893 0814 30 25.43 S 15 13.43 E 353 -99 -999 405 128 1 012893 1036 30 32.45 S 15 01.53 E 750 -99 -999 763 129 1 012893 1248 30 37.04 S 14 52.74 E 1288 -99 -999 1195 T5 130 1 012893 1552 30 44.33 S 14 39.77 E 1890 -99 -999 1216 T5 131 1 012893 1837 30 48.51 S 14 33.04 E 2250 -99 -999 1832 T5 132 1 012893 2308 31 02.28 S 14 05.94 E 2750 -99 -999 1260 T5 133 1 012993 0417 31 22.17 S 13 31.87 E 3250 -99 -999 1832 T5 134 1 012993 1104 31 46.48 S 12 44.51 E 3800 -99 -999 1207 T5 135 1 012993 1755 32 08.56 S 12 00.10 E 4160 -99 -999 1832 T5 136 1 013093 0112 32 30.27 S 11 16.38 E 4620 -99 -999 1832 T5 --------------------------------------------------------------------------------------- FIGURE LEGENDS *All figures shown in PDF file Figure 1. The A11 cruise track defined by CTD/Rosette stations. Isobaths of 200m and 3000m are superimposed Figure 2. The location of 10 l water samples collected on cruise A11. Depth is in dbar. Figure 3. Silicate concentration versus potential temperature for A11 (*) and SAVE 4 data: both are in the Argentine basin and for the whole water column. The inset, for the deepest levels, shows the small discrepancy between the data sets. Figure 4. Dissolved oxygen concentration versus salinity for A11 (*) and SAVE 4 data: both are in the Argentine basin and for the whole water column. The inset, where the deepest levels form the left branch of the Y, shows the small discrepancy between the data sets. Figure 5. Deep water collected on station 12240 from 2500m was used as quality control for the nutrient measurements: results are shown for the last 50 stations of the cruise. Figure 6. A comparison of CFC-11 and CFC-12 data from (a) SAVE station 291 and A11 station 12273 and (b) SAVE station 200 and A11 station 12295. Figure 7. Surface salinity (bold) and temperature (broken) on cruise A11. The cruise begins on the Argentine shelf, passes through the Falkland current (day363), the Brazil current retroflection (day 365),traverses the Subantarctic Zone until somewhere between day 386 and 390 it enters the subtropical gyre. The cruise ends in S.Africa Figure 8. Location of A11 and historical data. Pluses, this cruise. Crosses, SAVE leg 4. Triangles, Atlantis II Cruise 107. Inverted triangles, AJAX. Figure 9. This cruise: station averages of anomaly of salinity relative to standard fits. Horizontal axis: station number. Vertical axis: salinity anomaly. Figure 10. This cruise: sample minus CTD salinity residuals for all samples flagged as good. Horizontal axis: pressure. Vertical axis: salinity residual. Figure 11. SAVE leg 4: station averages of anomaly of salinity relative to standard fits. Horizontal axis: station number. Vertical axis: salinity anomaly. Figure 12. Atlantis II Cruise 107: station averages of anomaly of salinity relative to standard fits. Horizontal axis: longitude. Vertical axis: salinity anomaly. Figure 13. This cruise: comparison of measured with predicted OXYTMP. Horizontal axis: THETA. Vertical axis: measured minus predicted OXYTMP. Figure 14. This cruise: station averages of anomaly of oxygen relative to standard fits. Horizontal axis: station number. Vertical axis: oxygen anomaly (µmol/l). Figure 15. SAVE leg 4: station averages of anomaly of oxygen relative to standard fits. Horizontal axis: station number. Vertical axis: oxygen anomaly (µmol/kg). Figure 16. Atlantis II Cruise 107: station averages of anomaly of oxygen relative to standard fits. Horizontal axis: longitude. Vertical axis: oxygen anomaly (µmol/l). Figure 17. AJAX: station averages of anomaly of oxygen relative to standard fits. Horizontal axis: latitude. Vertical axis: oxygen anomaly (µmol/l). Intersects with A11 at 38 degrees south. Figure 18. Station averages of nitrate anomaly relative to standard fits. Horizontal axis: station number or latitude. Vertical axis: nitrate anomaly (µmol/l). (18a) - This cruise, (18b) - SAVE, (18c) - AJAX. Figure 19. As Figure 18, but silicate. Figure 20. As Figure 18, but phosphate. ------------------------------------------------------------------------------------------ May 3, 1996 Bob Millard Data Quality Control Report for WOCE cruise A11 The overall potential temperature versus salinity plot of figure 1a shows a range of variation of potential temperature from slightly less than zero to 22 C while the salinity varies from 33.75 to 35.65 psu. Figure 1b expands scales for lower layer and shows the two deep water masses, the colder and fresher Argentine Basin and the slightly warmer Cape Basin. A few noisy salinities are apparent in figure 1b. The oxygens values range from 155 to 330 Umol/kg, as the potential temperature versus oxygen plots of figure 2 show. Figures 1 and 2 contain all of the two decibar observations plus the water sample salinities and oxygens. To the resolution of these plots the temperature, salinity, and oxygen appear to be well behaved, except for a few noisy deep salinities. The water sample file salinity and oxygen data for both the CTD and bottle data are examined and the DQE quality word for these four parameters set in the second quality word of file A11.RCM. The CTD oxygens in the bottle file were found on average to be 6.0 Umol/kg higher than the bottle oxygens and all CTD oxygens were flagged as questionable. I agree with most of the other salinity and oxygen quality word assessments of the PI. A summary of the modified quality words (except for CTD oxygen) is given in Appendix II. A total of 83 bottle observations had salinity or oxygen quality words adjusted. Most of these occurred in the station group 12251 to 12255 where the CTD salinity was originally flagged as questionable by the PI but I found the CTD salinity observation differed from the bottle data by less than 3 standard deviations and in some cases by less than 0.001 psu (see Ds (ctd-ws) in Appendix II). The evaluation of the CTD data of WOCE cruise A11 examines the following two CTD data sets: individual 2 decibar down-profile data (a total of 91 station files) and the subset of the up-profile CTD observations stored in the bottle file together with the water sample oxygens and salinities. The cruise report (IOS Report # 234) covers the CTD calibration and processing methods including the the laboratory and in situ calibrations. The need to adjust the CTD salinity on a station by station to match the bottle salinities is contrary to my experience with the Neil Brown Mark III CTD. I did notice a few differences with how we correct conductivty at WHOI. At WHOI the CTD conductivity model expands the cell geometry corrections around a deep water value (2.8 and 3000 dbars) which tends to force the fit to match in the deep water independent of mismatches in the conductivity cell geometry effects (alpha=-6.5E-6 & beta = 1.5 E-8). We also allow another term in our fit, the conductivity bias. That said, both the CTD salinity and oxygen data in the bottle file (A11.HYD) and the individual 2- decibar down-profiles for WOCE cruise A11 are found to be well matched to water sample data with the exception of the CTD oxygen data in the bottle file which appears to have a systematic bias of about 6 Umol/kg. To assess the CTD quality of the CTD data following data checks were carried out: o Calibration checks: CTD and water sample Salinity and Oxygens Checks involve both the individual 2 decibar profiles and the bottle file CTD subset. The calibration checks are divided into an assessments at all depths and then only the deeper levels (defined as pressures greater than 1000 decibars). The calibration checks of salinity and oxygen involved looking at the differences of the CTD minus the water sample values. Both the down and up- profile CTD salinity and oxygen data were examined against bottle values. The salinity differences presented are formed using the bottle file CTD data while the oxygen differences presented are created by interpolating the down-profile 2-decibar profiles CTD oxygens at the bottle depths. o Check for spurious salinity and oxygen values deep: An evaluation of the CTD salinity and oxygen noise levels with checks for spurious data values. To check for spurious salinity and oxygen observations in the 2 decibar CTD data the standard deviation (RMS) of the high-pass filtered oxygen and salinity with wavelengths between 4 and 25 decibars is summarized in the deep water depth ranges to the cast bottom. The RMS scatter value is plotted versus station for several depth intervals from the bottom to the surface. Stations with a large scatter compared to the cruise average are plotted versus pressure with suspect data values (values greater than 5 standard deviations) identified on the plots. o Vertical stability check. A check for density inversions provides additional information about spurious salinity and/or temperature values particularly in the near surface region where this method provides more a sensitive test than looking at the high wave number salinity variability. The vertical gradient of potential density (first difference) is calculated and checked for decreases in density with depth exceeding one of two thresholds : (-0.0075 and -0.01 kg/m3). Salinity calibration The bottle file salinity differences are plotted versus station number, first at all pressures (figure 3a) and then the subset below 1000 decibars with a station average value indicated by the solid line in figure 3b. The third panel, figure 3c, is a plot of salinity differences versus pressure from 500 decibars to the bottom. Figure 3c begins at 500 decibars to permit an expanded salinity range and indicates that the CTD salinity is well calibrated in the vertical. Both plots versus station (3a and 3b) show the CTD salinity (conductivity) to be well matched to the water sample salts, the only evidence of a station off-set in figure 3b is for stations 12254, 12270 and 12319. A look at the deep potential temperature- salinity for these and neighboring stations (not shown) does not reinforce these stations to be miscalibrated. A histogram of salinity differences is shown for all pressures in figures 6c and below 1000 dbars in figure 6d. The standard deviation for all salinity differences is 0.0047 psu. The standard deviation of the salinity differences below 1000 decibars is 0.0014 psu which is a very tight scatter indicative of careful water sample salinity sampling and analysis. Oxygen calibration Figures 4 a, b, c shows the interpolated down-profile oxygen differences versus station, overall and deep, and versus pressure. The average oxygen difference below 1000 decibars in figure 4b shows that the 2 decibar oxygens are well matched to the water sample oxygens across the entire cruise. The CTD oxygens below 1000 decibars for stations 12271-12273 and 12305-12307 may be from 1-2 Umols/kg high and are checked further. The oxygen differences versus pressure in figure 4c indicates that the CTD oxygen is overestimated from 4500 decibars to the bottom by an amount of up to 5 Umol/kg at 6000 dbars. Similar plots of the up-profile oxygen differences from the bottle file, shown in figures 5 a-b, indicate a systematic difference between the bottle file CTD and water sample oxygens with the CTD oxygens an average of about 6.0 Umol/kg greater than the water samples. As noted earlier, all CTD oxygens in the bottle file are flagged as questionable in the second quality word. A histogram of oxygen differences for all pressure levels figure 6a and below 1000 dbars in figure 6b. The standard deviation using all of the good interpolated down-profile CTD oxygen differences is 3.31 Umol/kg (using the up-profile CTD oxygens yields a standard deviation of 2.99 umol/kg). The oxygen differences below 1000 dbars are normally distributed with a standard deviation of 2.05 Umol/kg. A series of waterfall plots consisting of down-profile CTD oxygen minus up water sample differences Dox= ( OXctd_dwn - WS) Umol/kg versus station are shown encompassing the 12273-12274 (figures 7a) and 12305-12307 (figure 7b). There is no systematic depth off-set to either stations 12273-12274 or 12305-12307. On the other hand, the deepest oxygen differences (greater than 4500 dbars) of stations 12260-12265 do show the CTD oxygen to be high. Spurious salinities and Oxygens The standard deviation of the high-pass filtered salinity (between vertical wavelengths of 4 and 25 decibars) from 3201 decibars to the bottom is shown in figure 8a. The bottom pressure is plotted versus station number in figure 8c. The average RMS CTD salinity scatter over the cruise of 0.00033 psu becomes as low as 0.0002 psu (stations 12268-12272). The deep water salinity scatter is higher than the salinity noise level found on other cruises examined which have been observed to be as low as .00013 psu. Figure 8a indicates that stations 12292-12293 and stations 12306-12308 have elevated deep water noise levels. These stations are examined and contrasted with some better behaved profiles of salinity later. The station averaged RMS oxygen scatter (noise level) for wavelengths between 4- 25 dbars is over twice as large as the best cruises examined (~0.1 Umol/kg). This may, in part, be due to a larger oxygen current quantizing although this can't be verified. Stations 12286-12288, 12291 and 12313-12315 have abnormally large RMS oxygen scatters which carry over to the depth interval from 1199-3201 dbars shown in figure 9b. The stability of all 2 decibar CTD data is checked by looking at potential density differences that exceed one of two thresholds. A plot of the pressure levels at which these instabilities occur (table I) is shown in figure 10 with potential density differences exceeding -0.0075 kg/m3/dbar marked with an (x) and the subset of these data less than -0.01 kg/m3/dbar marked with a (*). A tabular listing of these 73 points with negative density gradients exceeding -.0075 kg/m3/dbar is given below. The data set has 33 levels exceeding -.01 kg/m3/dbar. For the most part, instabilities are in the shallow depths regions less than 500 decibars where the largest temperature and salinity gradients occur. Some comments on individual or groups of stations 1: The salinities of stations 12291-12294 are overplotted and 12292-12294 show an elevated deep water noise level as figure 11 indicates when contrasted with figure 13. In addition there are spurious questionable salinity observations (x's) in stations 12292, 12293, & 12294. None appear to flagged in the quality word of the 2-dbars data files (see the quality word for the salt spikes of station 12292 at 3971-3973 dbars or sta. 12294 at 3461 dbars, all marked good). 2: The salinities of stations 12306-12308 are overplotted and show an elevated deep water noise level as figure 12 indicates when contrasted with figure 13. In addition there are spurious bad observations (x's) in stations 12306 & 12308. None appear to flagged in the quality word of the 2-dbars data files (see the quality word for station 12306 at 3493 dbars, marked good). 3: The salinities of stations 12269-12272 are overplotted in figure 13 as a control for deep water salinity variations for this data set. 4: The oxygens of stations 12286-12287 are overplotted and show an elevated deep water noise level as figure 14 indicates when contrasted with figure 18. There are bursts of noisy oxygens particularly for stations 12286 & 12287. Stations 12285 seems free of excessive noise and 12288 also show fewer problems pressure levels. 5: The oxygens of stations 12288-12291 are overplotted and show an elevated deep water noise level as figure 15 indicates when contrasted again with figure 18. There are bursts of noisy oxygens in station 12291 while variations of 12290 seems reasonable. 6: The oxygens of stations 12311-12315 are overplotted and all show bursts of noisy oxygens as figure 16 indicates when contrasted with figure 18. 7: The oxygens of station 12317 are overplotted with stations 12316-12319 and shows spikes of noisy oxygens as figure 17 indicates. 8: The oxygens of stations 12269-12272 are overplotted in figure 18 as a control for deep water oxygens variations for this data set. Table I dsg/dp > -.0075 kg/m3/dbar dsg/dp station # Prs dbars salinity -1.6825525e-002 1.2252000e+004 1.5650000e+003 3.4766700e+001 -1.4934576e-002 1.2252000e+004 1.5670000e+003 3.4716000e+001 -7.5215734e-003 1.2252000e+004 1.5710000e+003 3.4707500e+001 -8.9730034e-003 1.2254000e+004 1.9090000e+003 3.4810300e+001 -7.5567868e-003 1.2255000e+004 2.3450000e+003 3.4783100e+001 -9.8730225e-003 1.2256000e+004 2.5610000e+003 3.4772600e+001 -1.5949690e-002 1.2258000e+004 9.3000000e+001 3.4857600e+001 -1.2130733e-002 1.2258000e+004 1.3300000e+002 3.4938300e+001 -1.2609297e-002 1.2258000e+004 1.3900000e+002 3.4924600e+001 -8.8043772e-003 1.2258000e+004 1.8690000e+003 3.4663300e+001 -1.0486886e-002 1.2258000e+004 1.8790000e+003 3.4645100e+001 -2.4276438e-002 1.2259000e+004 1.1500000e+002 3.5164100e+001 -1.7380894e-002 1.2259000e+004 1.2300000e+002 3.5334300e+001 -8.4313404e-003 1.2259000e+004 1.6700000e+002 3.5076100e+001 -8.7635833e-003 1.2259000e+004 2.3300000e+002 3.4497100e+001 -9.4563893e-003 1.2259000e+004 2.3700000e+002 3.4442500e+001 -9.9873559e-003 1.2261000e+004 2.2900000e+002 3.4284300e+001 -1.9777770e-002 1.2261000e+004 2.3300000e+002 3.4254300e+001 -2.1413379e-002 1.2262000e+004 4.9000000e+001 3.4114500e+001 -2.8243981e-002 1.2262000e+004 5.3000000e+001 3.4137700e+001 -7.6608833e-003 1.2262000e+004 6.5000000e+001 3.4126100e+001 -9.4308111e-003 1.2262000e+004 7.3000000e+001 3.4149800e+001 -9.2692607e-003 1.2262000e+004 9.0900000e+002 3.4415100e+001 -9.2732815e-003 1.2262000e+004 9.1100000e+002 3.4395200e+001 -7.7542689e-003 1.2262000e+004 2.0010000e+003 3.4761900e+001 -1.6234888e-002 1.2262000e+004 2.0770000e+003 3.4758700e+001 -8.1060643e-003 1.2265000e+004 5.1000000e+001 3.4784700e+001 -9.6157972e-003 1.2272000e+004 2.2900000e+002 3.4240200e+001 -1.0727370e-002 1.2275000e+004 8.8300000e+002 3.4440400e+001 -2.4690527e-002 1.2277000e+004 8.7000000e+001 3.4541400e+001 -8.7019074e-003 1.2277000e+004 1.4500000e+002 3.4654500e+001 -8.1801374e-003 1.2278000e+004 4.9000000e+001 3.4259100e+001 -2.1828669e-002 1.2278000e+004 1.3090000e+003 3.4552400e+001 -1.1476531e-002 1.2279000e+004 8.5000000e+001 3.4497700e+001 -1.7991091e-002 1.2280000e+004 9.7000000e+001 3.3998300e+001 -8.9597959e-003 1.2282000e+004 1.8630000e+003 3.4758100e+001 -2.1340326e-002 1.2286000e+004 7.1000000e+001 3.4276500e+001 -1.0336119e-002 1.2286000e+004 1.3700000e+002 3.4369500e+001 -9.2833570e-003 1.2286000e+004 1.6300000e+002 3.4295200e+001 -1.7413396e-002 1.2292000e+004 6.3500000e+002 3.4177000e+001 -2.6613984e-002 1.2294000e+004 1.8500000e+002 3.4405500e+001 -8.3040160e-003 1.2294000e+004 1.9100000e+002 3.4450400e+001 -8.4622237e-003 1.2294000e+004 3.8300000e+002 3.4178600e+001 -8.5588970e-003 1.2296000e+004 2.1100000e+002 3.4354600e+001 -3.2241057e-002 1.2298000e+004 1.0900000e+002 3.4019300e+001 -2.7949113e-002 1.2298000e+004 1.1300000e+002 3.4033200e+001 -9.3840991e-003 1.2298000e+004 1.4500000e+002 3.4153700e+001 -1.8078311e-002 1.2302000e+004 9.9700000e+002 3.4266500e+001 -8.5959918e-003 1.2305000e+004 4.5300000e+002 3.4358800e+001 -9.0837387e-003 1.2307000e+004 1.1230000e+003 3.4290100e+001 -1.5554406e-002 1.2308000e+004 8.5000000e+001 3.4682400e+001 -8.0072034e-003 1.2308000e+004 3.2900000e+002 3.4603000e+001 -9.0148257e-003 1.2310000e+004 2.8900000e+002 3.4445000e+001 -7.5515126e-003 1.2311000e+004 4.9500000e+002 3.4301700e+001 -1.4240604e-002 1.2312000e+004 8.9000000e+001 3.4771000e+001 -9.2638822e-003 1.2312000e+004 1.2900000e+002 3.4742700e+001 -8.9953691e-003 1.2312000e+004 2.8500000e+002 3.4653900e+001 -9.2608014e-003 1.2312000e+004 3.5100000e+002 3.4539100e+001 -7.8618847e-003 1.2314000e+004 4.4500000e+002 3.4546500e+001 -9.7132557e-003 1.2315000e+004 1.0300000e+002 3.4959000e+001 -9.2969453e-003 1.2316000e+004 6.9000000e+001 3.4942200e+001 -9.8689572e-003 1.2316000e+004 1.4500000e+002 3.4835500e+001 -1.2335594e-002 1.2316000e+004 2.0500000e+002 3.4847000e+001 -1.2643058e-002 1.2316000e+004 2.2900000e+002 3.4805400e+001 -9.7481726e-003 1.2316000e+004 2.8900000e+002 3.4829000e+001 -1.0933894e-002 1.2323000e+004 1.0970000e+003 3.4302100e+001 -8.3410129e-003 1.2325000e+004 3.0000000e+000 3.5632200e+001 -8.4431894e-003 1.2325000e+004 8.5000000e+001 3.5562300e+001 -2.9400095e-002 1.2325000e+004 8.3100000e+002 3.4416500e+001 -1.3439053e-002 1.2325000e+004 8.5300000e+002 3.4402400e+001 -1.6726372e-002 1.2325000e+004 8.8300000e+002 3.4316600e+001 -8.5529223e-003 1.2325000e+004 1.1050000e+003 3.4426400e+001 -1.0191833e-002 1.2326000e+004 8.5500000e+002 3.4393600e+001 Subset of above that exceed dsg/dp > -.01 kg/m3/dbar dsg/dp station # Prs dbars salinity -1.6825525e-002 1.2252000e+004 1.5650000e+003 3.4766700e+001 -1.4934576e-002 1.2252000e+004 1.5670000e+003 3.4716000e+001 -1.5949690e-002 1.2258000e+004 9.3000000e+001 3.4857600e+001 -1.2130733e-002 1.2258000e+004 1.3300000e+002 3.4938300e+001 -1.2609297e-002 1.2258000e+004 1.3900000e+002 3.4924600e+001 -1.0486886e-002 1.2258000e+004 1.8790000e+003 3.4645100e+001 -2.4276438e-002 1.2259000e+004 1.1500000e+002 3.5164100e+001 -1.7380894e-002 1.2259000e+004 1.2300000e+002 3.5334300e+001 -1.9777770e-002 1.2261000e+004 2.3300000e+002 3.4254300e+001 -2.1413379e-002 1.2262000e+004 4.9000000e+001 3.4114500e+001 -2.8243981e-002 1.2262000e+004 5.3000000e+001 3.4137700e+001 -1.6234888e-002 1.2262000e+004 2.0770000e+003 3.4758700e+001 -1.0727370e-002 1.2275000e+004 8.8300000e+002 3.4440400e+001 -2.4690527e-002 1.2277000e+004 8.7000000e+001 3.4541400e+001 -2.1828669e-002 1.2278000e+004 1.3090000e+003 3.4552400e+001 -1.1476531e-002 1.2279000e+004 8.5000000e+001 3.4497700e+001 -1.7991091e-002 1.2280000e+004 9.7000000e+001 3.3998300e+001 -2.1340326e-002 1.2286000e+004 7.1000000e+001 3.4276500e+001 -1.0336119e-002 1.2286000e+004 1.3700000e+002 3.4369500e+001 -1.7413396e-002 1.2292000e+004 6.3500000e+002 3.4177000e+001 -2.6613984e-002 1.2294000e+004 1.8500000e+002 3.4405500e+001 -3.2241057e-002 1.2298000e+004 1.0900000e+002 3.4019300e+001 -2.7949113e-002 1.2298000e+004 1.1300000e+002 3.4033200e+001 -1.8078311e-002 1.2302000e+004 9.9700000e+002 3.4266500e+001 -1.5554406e-002 1.2308000e+004 8.5000000e+001 3.4682400e+001 -1.4240604e-002 1.2312000e+004 8.9000000e+001 3.4771000e+001 -1.2335594e-002 1.2316000e+004 2.0500000e+002 3.4847000e+001 -1.2643058e-002 1.2316000e+004 2.2900000e+002 3.4805400e+001 -1.0933894e-002 1.2323000e+004 1.0970000e+003 3.4302100e+001 -2.9400095e-002 1.2325000e+004 8.3100000e+002 3.4416500e+001 -1.3439053e-002 1.2325000e+004 8.5300000e+002 3.4402400e+001 -1.6726372e-002 1.2325000e+004 8.8300000e+002 3.4316600e+001 -1.0191833e-002 1.2326000e+004 8.5500000e+002 3.4393600e+001 APPENDIX II Cruise A11 changes to Quality word of A1.hyd file Below is a list of the bottles that have had a CTD or water sample salinity or oxygen flag changed. Only the first 5 field of the quality flags Qual1 and Qual2 (DQE) are given as these were the only ones modified. Note that all CTD oxygens have been flagged as questionable "3" as the CTD oxygens in the bottle file are systematically higher than the water samples by an average of 6.0 Umol/kg across the cruise. On the other hand, the CTD oxygens in the individual 2 decibar CTD files do not show a systematic error with water sample oxygens. Stations 12251 through 12255 CTD salts flagged questionable but the magnitude of the CTD water sample salinity difference (Ds) for the most part are small (less than 3 standard deviations) and don't substantiate flagging as questionable. The first two observations of 12251 below have ctd salt flagged missing when CTD O2 is the missing parameter. Sta. 12254 CTD up profile bottle data is systematically fresh except in deep water. Station 12325 CTD salts are flagged "3" in the upper 1200 dbars when bottle differences are consistent with vertical structure. St.No. Prs. S_ws Ox_ws Ds(ctd-ws) Dox(ctd-ws) Qual1 Qual2 12251 3.0 -9.0000 271.9000 43.0880 -280.9000 29299 23999 ctd o2=9 12251 87.0 -9.0000 300.0000 43.1260 -309.0000 29299 23999 ctd o2=9 12251 504.6 34.2346 235.5000 0.0000 6.9000 23222 22322 12251 762.4 34.4159 186.8000 -0.0014 3.6000 23222 22322 12251 1011.6 34.5563 170.3000 -0.0006 4.7000 23222 22322 12251 1267.8 34.6339 167.3000 0.0008 5.5000 23222 22322 12251 1525.2 34.7052 175.1000 0.0005 5.4000 23222 22322 12251 1730.2 34.7571 189.6000 0.0017 4.8000 23222 22322 12251 1890.2 34.7321 182.4000 0.0025 4.7000 23222 22322 12252 10.0 34.0907 280.9000 -0.0005 4.8000 23222 22322 12252 55.6 34.1215 299.6000 0.0005 16.2000 23222 22322 12252 105.1 34.1246 300.8000 0.0037 3.8000 23222 22322 12252 154.9 34.1360 280.2000 -0.0011 10.4000 23222 22322 12252 204.3 34.1391 278.2000 0.0002 2.3000 23222 22322 12252 253.9 34.1503 268.7000 -0.0040 5.2000 23222 22322 12252 353.8 34.1483 266.4000 0.0000 4.8000 23222 22322 12252 498.9 34.2350 235.9000 0.0026 10.8000 23222 22322 12252 757.6 34.4173 193.3000 0.0008 6.4000 23222 22322 12252 1017.6 34.5653 172.0000 -0.0008 6.4000 23222 22322 12252 1267.2 34.6370 169.7000 -0.0010 3.7000 23222 22322 12252 1512.9 34.6965 175.3000 -0.0029 6.2000 23222 22322 12252 1768.4 34.7526 190.2000 0.0022 2.3000 23222 22322 12252 2029.3 34.7628 192.7000 -0.0011 2.3000 23222 22322 12252 2291.5 34.7455 192.2000 -0.0003 5.6000 23222 22322 12252 2547.6 34.7459 191.9000 0.0008 5.6000 23222 22322 12252 2611.8 34.7465 192.6000 0.0007 5.8000 23222 22322 12252 2611.8 34.7471 191.5000 0.0001 6.9000 23222 22322 12253 15.6 34.0895 285.6000 -0.0021 4.9000 23223 22323 12253 105.2 34.1221 304.1000 -0.0008 5.5000 23223 22323 12253 155.6 34.1350 289.0000 -0.0006 6.7000 23223 22323 12253 206.9 34.1370 264.2000 -0.0029 20.3000 23223 22323 12253 257.2 34.1348 333.4000 0.0021 -53.5000 23234 22324 12253 356.3 34.1452 272.1000 -0.0024 -0.3000 23223 22323 12253 504.8 34.2330 223.6000 0.0037 12.7000 23243 22323 12253 760.8 34.4286 192.7000 0.0005 -9.0000 23223 22323 12253 1015.5 34.5679 158.2000 -0.0004 7.1000 23223 22323 12253 1271.5 34.6502 163.4000 -0.0005 4.0000 23223 22323 12253 1526.6 34.7203 180.3000 -0.0007 0.6000 23223 22323 12253 2035.9 34.7355 176.3000 0.0006 14.1000 23223 22323 12253 2289.4 34.7384 186.9000 -0.0038 9.7000 23223 22323 12253 2543.9 34.7398 199.9000 -0.0019 4.7000 23223 22323 12253 3052.8 34.7260 197.1000 0.0005 9.3000 23223 22323 12253 3136.0 34.7258 199.2000 0.0041 2.8000 23223 22323 12253 3136.0 34.7272 199.3000 0.0027 2.7000 23223 22323 12254 3453.7 34.7089 202.7000 0.0009 4.2000 23223 22323 12254 3453.7 34.7085 197.3000 0.0013 9.6000 23223 22323 12255 9.8 34.0770 292.2000 0.0000 7.1000 23223 22323 12255 55.8 34.1199 307.0000 0.0022 15.5000 23223 22323 12255 106.2 34.1391 296.7000 -0.0013 -1.6000 23223 22323 12255 155.7 34.1456 294.7000 -0.0017 8.0000 23223 22323 12255 206.0 34.1432 287.9000 0.0002 6.0000 23223 22323 12255 256.4 34.1470 282.4000 -0.0005 9.5000 23223 22323 12255 355.7 34.1489 274.6000 0.0022 9.0000 23223 22323 12255 507.4 34.2501 238.0000 0.0025 4.8000 23223 22323 12255 763.8 34.3957 194.0000 -0.0009 10.2000 23223 22323 12255 1018.5 34.5243 178.0000 -0.0009 4.8000 23223 22323 12255 1273.3 34.6743 183.9000 -0.0010 3.6000 23223 22323 12255 1527.2 34.7510 194.8000 -0.0008 3.6000 23223 22323 12255 1781.0 34.7748 199.2000 -0.0014 10.4000 23223 22323 12255 2037.6 34.8212 214.6000 -0.0036 0.7000 23223 22323 12255 2548.5 34.7583 198.6000 -0.0028 5.0000 23223 22323 12255 3058.0 34.7368 201.9000 0.0000 6.0000 23223 22323 12255 3571.4 34.7072 207.4000 0.0007 5.2000 23223 22323 12255 4047.9 34.6789 217.7000 0.0006 7.3000 23223 22323 12263 1525.8 34.6011 177.0000 -0.0021 7.5000 22232 22322 12263 5590.2 34.6696 220.0000 0.0006 8.0000 22232 22322 12263 5889.8 34.6690 218.3000 0.0001 10.2000 22232 22322 12268 510.0 34.1905 261.1000 0.0011 6.1000 22232 22322 12271 1271.1 34.5404 179.7000 -0.0012 6.3000 22232 22322 12271 2029.7 34.7901 203.0000 -0.0015 6.2000 22232 22322 12288 357.7 34.1860 266.1000 0.0011 9.1000 22232 22322 12288 2281.4 34.7811 199.2000 -0.0019 6.4000 22232 22322 12318 358.1 34.7762 221.8000 0.0078 3.6000 22232 22322 12325 8.2 35.6719 221.1000 -0.0007 6.4000 23222 22322 12325 53.5 35.5839 221.2000 -0.0054 12.8000 23222 22322 12325 103.7 35.5527 200.1000 0.0009 0.5000 23222 22322 12325 154.0 35.5267 199.8000 0.0081 0.8000 23222 22322 12325 203.7 35.4352 197.0000 0.0055 4.0000 23222 22322 12325 353.8 35.1547 215.8000 0.0063 6.8000 23222 22322 12325 753.3 34.4864 198.9000 0.0043 12.2000 23222 22322 12325 1002.8 34.3861 196.6000 -0.0056 8.1000 23222 22322 12325 1239.0 34.4520 182.8000 0.0050 6.8000 23222 22322 12334 53.9 35.5565 224.1000 -0.0059 9.3000 22232 22322 *Figures shown in PDF file. --------------------------------------------------------------------------------- 16 August 1996 All DQE notes: Overall, the nutrient data appear to be of very good quality. Most of the data points which were outside of regional nutrient/theta trends had been flagged by the data originator. Specific bottles which had problems noted by either the data originator or the WOCE DQE evaluator are listed below. STATION BOTTLE PROBLEM Q1 Q2 12257 25701 Low P 222 223 12257 25702 Low P 222 223 12257 25703 Low P 222 223 12257 25704 Low P 222 223 12257 25705 Low P 222 223 12259 25920 High Sil 333 333 12261 26112 N & P a bit high 222 233 12261 26119 P high 222 223 12262 26204 N high 222 232 12264 26424 High Sil 444 333 12284 28405 Low Sil 444 333 12283 28324 All nuts high 444 333 12288 28802 N low 222 232 12288 28804 N low 222 232 12288 28805 N low 222 232 12293 29320 N and P high 333 333 12302 30211 High P 222 223 12306 30619 Low P 223 223 12306 30618 Low P 223 223 12319 31906 High P 222 223 12319 31914 High P 222 223 12322 32224 Q1 flagged, theta high, could be real 444 333 12323 32324 Sil and P a bit high 444 333 12329 32905 Sil high, O2 low 333 333 12331 33109 Sil high 333 333 INPUT FILE: A11.JCJ THE DATE TODAY IS: 21-AUG-96 STNNBR CASTNO SAMPNO CTDPRS SILCAT NO2+NO3 PHSPHT QUALT1 QUALT2 12257 1 25705 3528.0 2.03 ~~2 ~~3 12257 1 25704 4043.8 2.07 ~~2 ~~3 12257 1 25703 4558.6 2.10 ~~2 ~~3 12257 1 25701 5397.3 2.10 ~~2 ~~3 12261 1 26112 2553.3 28.96 1.92 ~22 ~33 12262 1 26204 5099.3 33.92 ~2~ ~3~ 12264 1 26424 11.7 2.75 10.66 0.66 444 333 12283 1 28324 16.3 14.72 21.09 1.38 444 333 12284 1 28405 3018.2 76.66 4~~ 3~~ 12302 1 30211 760.9 2.12 ~~2 ~~3 12319 1 31906 3045.5 1.78 ~~2 ~~3 12322 1 32224 11.5 3.35 4.03 0.39 444 333 12323 1 32324 10.6 2.46 3.64 0.37 444 333 ---------------------------------------------------------------------------------- iodide concentration of iodide, nmoles per liter stdv_I- standard deviation of iodide iodate DPP concentration of iodate, nmoles per liter stdv_IO3- standard deviation of iodate spectro_IO3 spectrophotometric concentration of iodate, nmoles per liter stdv_spec_iodate standard deviation of iodate sum -Ired concentration of total iodine, nmoles per liter stdv_sum-Ired standard deviation of total iodine COLLECTION Samples were collected from routine hydrocasts. Care was taken to draw samples after the dissolved oxygen reagents were removed from the hydrolab to avoid any potential sources of contamination during sampling. ANALYSES Iodide and iodate concentrations were determined using polarographic and voltammetric methods. Iodide (I-) was measured using cathodic stripping square wave voltammetry (CSSWV) [Luther et. al., 1988]. Iodate (IO3-) was measured using differential pulse polarography (DPP) [Herring and Liss, 1974]. Total iodine (_Ired) was measured using CSSWV [Campos (1997)]. The instrument minimum detection limits in seawater for I-, IO3-, _Iox and _.Ired using polarography are 0.2, 20, and 5 nM respectively. For detailed methods please consult (Farrenkopf, 1997 -- Dissertation University of Delaware). Precision for iodide based upon triplicate measurements of individual samples is within 5-10% in samples greater than 200 nM and within 1-2% for iodide concentrations less than 200 nM. Method precisions in 3.5% NaCl were ± 1%. Precisions for the total methods tend to vary significantly from sample to sample and so reported errors "stdev Tot_I" reflect the standard deviation of at least three replicates with three distinct standard addition curves (n>3). Spectro_IO3, spectrophotometric concentrations of iodate were determined by the spectrophotometric method as modified by Chapman and Liss (1977). EQUIPMENT Electrochemical measurements were made in 10 mL glass polarographic cells. EG & G Princeton Applied Research model 384 B polarographic analyzers equipped with 303A hanging mercury drop working electrode (HDME) stands were used throughout. Potentials were measured vs. a saturated calomel reference electrode (SCE). A platinum counter electrode was used for current measurements in a standard three electrode voltammetric arrangement. Iodide gives rise to a peak at a potential of -0.306 V, and iodate has a peak potential of -1.08 V. The concentrations of iodine species were determined by the method of standard addition. A minimum of three standard additions were made for each determination. Absorbance measurements were obtained at a wavelength of 285 nm using a Milton-Roy Spectronic 601 spectrophotometer equipped with 10 cm quartz cuvettes. Concentrations were determined based upon comparison with an external standard curve generated in 3.5% NaCl solution. REFERENCES: Campos, M.L.A. (1997) New approach to evaluating dissolved iodine speciation in natural waters using cathodic stripping voltammetry. Marine Chemistry Chapman, P. and P.S. Liss (1977) The effect of nitrite on the spectrophotometric determination of iodate in seawater. Marine Chemistry 5: 243-249. Luther, G. W., III, C. Branson Swartz and W.J. Ullman (1988) Direct determination of iodide in seawater by cathodic stripping square wave voltammetry. Analytical Chemistry. 60: 1721-1724. Luther, G.W., III (1991) Sulfur and iodine speciation in the water column of the Black Sea, in Black Sea Oceanography, E. Izdar and J. W. Murray, Editors. Kluwer Publishers: Netherlands. p. 187- 204. Herring, J.R. and P.S. Liss (1974) A new method for the determination of iodine species in seawater. Deep-Sea Research I. 21: 777-783. Truesdale, V.W. and C.P. Spencer (1977) Studies on the determination of inorganic iodine in seawater. Marine Chemistry 2: 33-47