ORNL/CDIAC-114 NDP-067 THE INTERNATIONAL INTERCOMPARISON EXERCISE OF UNDERWAY fCO2 SYSTEMS DURING THE R/V METEOR CRUISE 36/1 IN THE NORTH ATLANTIC OCEAN Contributed by Arne Koertzinger, Ludger Mintrop, and Jan C. Duinker Department of Marine Chemistry Institute of Marine Research University of Kiel Kiel, Germany Other data contributors Kenneth M. Johnson, Craig Neill, Douglas W. R. Wallace, Bronte Tilbrook, Philip Towler, Hisayuki Inoue, Masao Ishii, Gary Shaffer, Rodrigo Torres, Eiji Ohtaki, Eiji Yamashita, Alain Poisson, Christian Brunet, Bernard Schauer, Catherine Goyet, and Greg Eischeid Prepared by Alexander Kozyr Carbon Dioxide Information Analysis Center Oak Ridge National Laboratory Oak Ridge, Tennessee, U.S.A. Environmental Sciences Division Publication No. 4844 Date Published: March 1999 Prepared for the Environmental Sciences Division Office of Biological and Environmental Research U.S. Department of Energy Budget Activity Numbers KP 12 04 01 0 and KP 12 02 03 0 Prepared by the Carbon Dioxide Information Analysis Center OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831-6335 managed by LOCKHEED MARTIN ENERGY RESEARCH CORP. for the U.S. DEPARTMENT OF ENERGY under contract DE-AC05-96OR22464 ABBREVIATIONS AT total alkalinity BMBF Bundesministerium f r Bildung, Wissenschaft, Forschung und Technologie (German Ministry for Education, Science, Research and Technology) BNL Brookhaven National Laboratory CMDL Climate Monitoring and Diagnostics Laboratory CSIRO Commonwealth Scientific and Industrial Research Organisation CT total dissolved inorganic carbon (synonyms: DIC, TCO2, ?CO2) CTD conductivity-temperature-depth probe DFG Deutsche Forschungsgemeinschaft (German Research Foundation) DOE U.S. Department of Energy fCO2 fugacity of CO2 GMT Greenwich mean time GPS global positioning system IfMK Institut f r Meereskunde Kiel (Institute of Marine Research at the University of Kiel) IOC Intergovernmental Oceanographic Council JGOFS Joint Global Ocean Flux Study MRI Meteorological Research Institute NBI Niels Bohr Institute for Astronomy, Physics and Geophysics NDIR nondispersive infrared NDP numeric data package NOAA National Oceanic and Atmospheric Administration OU Okayama University pCO2 partial pressure of CO2 pH pH value PSS Practical Salinity Scale RF Reedereigemeinschaft Forschungsschiffahrt GmbH R/V Research Vessel SCOR Scientific Committee on Oceanic Research SIO Scripps Institution of Oceanography SOMMA single-operator multiparameter metabolic analyzer UP&MC Universit Pierre et Marie Curie UTC universal time coordinated WHOI Woods Hole Oceanographic Institution WMO World Meteorological Organization XBT expendable bathythermograph xCO2 mole fraction of CO2 ACKNOWLEDGMENTS The authors would like to thank the following people for their cooperation and participation in this exercise. Their motivation and active involvement made this study possible. Our special thanks are addressed to those who joined the scientific party on board R/V Meteor for their willingness to cope with all the little problems such an exercise inevitably provides. * Douglas Wallace, Kenneth Johnson, and Craig Neill, Brookhaven National Laboratory, Department of Applied Science, Upton, Long Island, New York, U.S.A. * Bronte Tilbrook and Philip Towler, Commonwealth Scientific and Industrial Research Organisation, Division of Oceanography, Hobart, Tasmania, Australia * Joanna Waniek, Susanne Schweinsberg, and Frank Malien, Institut fuer Meereskunde Kiel (Institute of Marine Research), Kiel, Germany * Hisayuki Inoue and Masao Ishii, Meteorological Research Institute, Tsukuba, Japan * Gary Shaffer, Niels Bohr Institute for Astronomy, Physics and Geophysics, University of Copenhagen, Copenhagen, Denmark * Rodrigo Torres, Goteborg University and Chalmers University of Technology, Department of Analytical and Marine Chemistry, Goteborg, Sweden * Eiji Ohtaki and Eiji Yamashita, Okayama University, Okayama, Japan * Andrew Dickson and Justine Parks, Scripps Institution of Oceanography, Marine Physical Laboratory, La Jolla, California, U.S.A. * Alain Poisson, Christian Brunet, and Bernard Schauer, Universite Pierre et Marie Curie, Laboratoire de Physique et Chimie Marines, Paris, France * Catherine Goyet and Greg Eischeid, Woods Hole Oceanographic Institution, Department of Marine Chemistry and Geochemistry, Woods Hole, Massachusetts, U.S.A. In addition, we wish to thank the chief scientist of Meteor Cruise 36/1, Detlev Schulz-Bull, for perfect cooperation and for giving this exercise the high priority it needed. We appreciate the valuable comments and helpful advice from Andrew Dickson during the early stages of the exercise. Our thanks to Douglas Wallace and Andrew Watson for reviewing the manuscript and making helpful comments. We also express our appreciation to the Reedereigemeinschaft Forschungsschiffahrt GmbH and the crew of R/V Meteor for their extremely helpful technical support before and throughout the cruise. On behalf of the participants of this intercomparison exercise, we extend special thanks to all national funding agencies for providing the participants with the necessary funding. Last but not least, the authors thank the Bundesministerium f r Bildung, Wissenschaft, Forschung und Technologie (BMBF) for funding this exercise through German Joint Global Ocean Flux Study (JGOFS) funds as well as the Deutsche Forschungsgemeinschaft (DFG) for making it possible to invite the international participants on this cruise. ABSTRACT Koertzinger, A., L. Mintrop, J.C. Duinker, K. Johnson, C. Neill, D.W.R. Wallace, B. Tilbrook, P. Towler, H. Inoue, M. Ishii, G. Shaffer, R. Torres, E. Ohtaki, E. Yamashita, A. Poisson, C. Brunet, B. Schauer, C. Goyet, G. Eischeid, and A. Kozyr (ed.). 1999. The International Intercomparison Exercise of Underway fCO2 Systems During the R/V Meteor Cruise 36/1 in the North Atlantic Ocean. ORNL/CDIAC-114, NDP-067. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee, U.S.A. Measurements of the fugacity of carbon dioxide (fCO2) in surface seawater are an important part of studies of the global carbon cycle and its anthropogenic perturbation. An important step toward the thorough interpretation of the vast amount of available fCO2 data is the establishment of a database system that would make such measurements more widely available for use in understanding the basin- and global-scale distribution of fCO2 and its influence on the oceanic uptake of anthropogenic CO2. Such an effort, however, is based on knowledge of the comparability of data sets from different laboratories. Currently, however, there is not much known about this subject. In the light of the aforementioned situation, an International Intercomparison Exercise of Underway fCO2 Systems was proposed and carried out by the Institut fuer Meereskunde Kiel (IfMK) (Institute of Marine Research at the University of Kiel), Kiel, Germany, during the R/V Meteor Cruise 36/1 from Hamilton, Bermuda, to Las Palmas, Gran Canaria, Spain. Nine groups from six countries (Australia, Denmark, Germany, France, Japan, and the United States) participated in this ambitious exercise, bringing together 15 participants with 7 underway fCO2 systems, 1 discrete fCO2 system, and 2 underway pH systems, as well as discrete systems for alkalinity and total dissolved inorganic carbon. This report presents only the results of the underway measurements of fCO2. The main idea of the exercise was to compare surface seawater fCO2 synchronously measured by all participating instruments under identical conditions. This synchronicity was accomplished by providing the infrastructure during the exercise, such as a common seawater and calibration gas supply. Another important issue was checks of the performance of the calibration procedures for CO2 and of all equilibrator temperature sensors. Furthermore a common procedure for the calculation of final fCO2 was applied to all data sets. All these measures were taken in order to reduce the largest possible amount of controllable sources of error. In this report we will demonstrate that the results of three of the seven underway systems agreed to within 2 µatm throughout the cruise. This was not only the case for seawater fCO2 measurements but also for measurements of theµatmospheric mole fraction of CO2 (xCO2). One system was in good agreement (+/-2 µatm) for most of the time but showed a considerable positive offset of up to 9 µatm for about 40 h. However, it was found that significant offsets of up to 10 µatm occurred in underway fCO2 measurements for three systems under typical and identical field work conditions. Although at least in one case this may be a consequence of a technical failure, it is an indication of significant systematic differences in other cases. Finally, the discrete fCO2 system measurements agreed within its nominal accuracy of 1% with the three most consistent underway fCO2 systems data sets. On the basis of a detailed comparison and evaluation of this large intercomparison data set we try to come up with general conclusions and recommendations for underway fCO2 work. These may serve as background information for a successful preparation of a coherent database of surface ocean fCO2 values. The results of this exercise certainly underline the need to address carefully the important issue of the inter-laboratory comparability of fCO2 data. PART 1 OVERVIEW 1. INTRODUCTION 1.1 SCIENTIFIC BACKGROUND OF THE EXERCISE Currently marine scientists are applying different concepts to quantify the oceanic uptake of CO2. These efforts are being undertaken in the light of theµatmospheric CO2 perturbation and its possible impact on the earth's climate. One important concept is based on the determination of the partial pressure difference of CO2 (delta pCO2) between the surface seawater and the overlying air, which is the thermodynamic driving force for any net exchange of CO2. By means of a transfer coefficient, a measured (pCO2 can be converted into a momentary net flux of CO2 across the air-sea interface. Given the strong spatial and temporal variability of pCO2 in the ocean, this concept faces the challenge of coming up with representative mean delta pCO2 values on a global grid. If this concept is to be successful in pinning down the present oceanic uptake of CO2 reliably, the combined efforts of research groups all over the world are necessary. The Intergovernmental Oceanographic Council (IOC)/Scientific Committee on Oceanic Research (SCOR) Carbon Dioxide Advisory Panel recently established an international inventory of pCO2 measurements that have been identified so far (http://cdiac.esd.ornl.gov/oceans/pco2inv.html). One important requirement in this context is a good inter-laboratory comparability of the data sets, which were generated by quite different types of analytical systems. While the analytical precision of the various systems in use is mostly of the order of 1 µatm or better, not much is known presently about the comparability between different laboratories. As a first important step to assess the current state of this parameter, an international shore-based intercomparison exercise of underway fugacity of CO2 (fCO2) systems was carried out by Andrew Dickson in June 1994 at the Scripps Institution of Oceanography, Marine Physical Laboratory, La Jolla, California, U.S.A. (http://www-mpl.ucsd.edu/people/adickson/CO2_QC) on behalf of the Joint IOC/SCOR CO2 Advisory Panel. However, the general consensus in the scientific community was that a necessary second step would be an at-sea intercomparison under more typical and identical operation conditions. Such an exercise, to be carried out during the R/V Meteor cruise 36/1, was proposed by the Kiel CO2 group in June 1995 and received very positive feedback within the scientific community. For a number of reasons the proposed cruise leg was perfectly suited for such an exercise. Funding of the exercise came through the German Joint Global Ocean Flux Study (JGOFS) program. More than fifteen research groups, representing a fairly good geographical distribution, were contacted and invited to participate in the exercise, nine of which were finally able to do so (Koertzinger et al. 1996a). 1.2 THE PRINCIPAL DESIGN OF THE EXERCISE The basic idea of the exercise was to operate as many underway fCO2 systems simultaneously for as much time as possible. Combined with in situ salinity and temperature as well as navigational and meteorological data, this combined underway fCO2 data set is the mainstay of the exercise. Whereas shore-based intercomparison exercises allow researchers to devise special experiments that reflect extreme situations, ship-based exercises have to rely fully on the conditions that are provided by the ocean. The chosen cruise track reflects the attempt to include-within the limits of a single and comparatively short cruise-extreme oceanic regimes. Whereas the situation was very stable in the Eastern North Atlantic with not much variability in surface seawater temperatures and salinities and likewise fCO2, the North Atlantic Drift region off Newfoundland provided extreme variability with steep gradients. The overall temperature range during the exercise was from 6.0°C to 25.1°C, while the salinity varied between 32.3 and 37.0. In the western part our cruise track hit warm and cold ring features. Associated with these rings were steep frontal gradients with changes of up to 15 C and more than 3 in salinity over a few nautical miles. These different regimes provide different information about the performance and comparability of the participating systems. The stable situation during the second half of the exercise allows the detection of systematic offsets between the data sets, thus providing the basic information about the inter-laboratory comparability. In contrast to this, the strong gradient regime mimics to some extent the step experiments of shore-based intercomparison exercises. The fast change between two "batches" of seawater, which are characterized by different fCO2 values, reveals the different time constants of the analytical systems. Fast responding systems are able to follow the signal much more closely than the more slowly responding ones. So, even if there are no systematic differences between two systems, the systems may have quite different response times, which translates into different spatial resolution in underway work. Right from the beginning, it was regarded as high priority to measure as many parameters [i.e., pH, fCO2, total dissolved organic carbon (CT), and total alkalinity (AT)] of the marine CO2 system as possible rather than restricting the exercise to mere fCO2 measurements. For this purpose, we followed two different sampling strategies (i.e., underway sampling and discrete sampling). As all participating fCO2 systems (CSIRO, IfMK, MRI, NBI, OU, UP&MC, WHOI; see Sect. 2.1.2 for a list of participating institutions) were operated in an underway mode on the same seawater source, it was highly desirable to back up these fCO2 measurements with additional underway measurements of other CO2 parameters. This was accomplished by underway pH measurements with two different spectrophotometric systems (SIO, WHOI) as well as underway CT measurements (BNL/IfMK) with a newly modified single-operator multiparameter metabolic analyzer (SOMMA) coulometric titration system (Johnson et al. 1998), all of which were hooked up to the seawater pumping system. Discrete sampling was carried out for discrete measurements of fCO2 (BNL), CT (BNL/IfMK), AT (IfMK), and salinity (IfMK) as well as nutrients (IfMK) in samples taken regularly from the same seawater pumping system. By measuring more than two parameters of the CO2 system in seawater, the system is overdetermined, as all parameters can be calculated from any combination of two measured parameters and knowledge of the thermodynamic relationships involved. This was the case for both sampling strategies. Overdetermination will therefore allow for consistency checks on the data sets. It may also provide additional information in the question of the best set of thermodynamic constants for the CO2 system. The broad CO2 database furthermore serves as valuable background information and will strongly enhance further interpretation of the results. The exercise also included checks on ancillary measurements, such as temperature and barometric pressure, as performed by most of the analytical systems. All temperature sensors were compared against a calibrated Pt-100 reference thermometer. The barometric pressure readings were also referenced against a high-quality digital barometer. In many cases, these checks revealed offsets and miscalibrations, which, if not corrected for, would have led to significant biases of the final fCO2 values. These checks helped to identify the error contribution from these sources. They also allowed us to correct all fCO2 measurements for these effects to reveal any systematic differences that cannot be attributed to the quality of temperature and pressure measurements. Further checks were carried out with the calibration gases. The suite of calibration gases supplied by the organizer covered a range of CO2 concentrations between 250 and 500 ppmv with nominal values of 250, 300, 350, 400, 450, and 500 ppmv. While every group required one or more of these calibration gases for their calibration procedure, they measured all other concentrations as unknown samples on their systems. The results provide information on the quality and reliability of the calibration procedures over the whole range from 250 to 500 ppmv. As the infrared detectors used by all groups generally show nonlinear response functions, the calibration procedure is a crucial point. 2. DESCRIPTION OF THE EXERCISE 2.1 THE CRUISE 2.1.1 R/V Meteor, Technical Details and Brief History The R/V Meteor is owned by the Federal Republic of Germany, represented by the Ministry for Education, Science, Research and Technology (BMBF), which financed its construction. It is operated by the German Research Foundation (DFG), which provides about 70% of its operating funds (the remainder is supplied by the BMBF). The Senate Commission for Oceanography of the DFG plans expeditions from the scientific viewpoint and appoints cruise coordinators and chief scientists. The Operations Control Office of the University of Hamburg is responsible for management, logistics, execution, and supervision of ship operations. These functions are exercised by direct cooperation with expedition coordinators and the managing owner, the Reedereigemeinschaft Forschungsschiffahrt GmbH (RF). The latter is responsible for hiring, provisioning, and coordinating ship maintenance. Designed as a multipurpose vessel for living and nonliving resources and worldwide operation, the R/V Meteor routinely carries scientists from many different countries. The basic technical details are Port of registration Hamburg Call sign DBBH Classification GL + 100 A4 E2 + MC Auto Operator University of Hamburg, Institute for Marine Research Managing owner RF Reedereigemeinschaft Forschungsschiffahrt GmbH, Bremen Built 1985/86 at Schlichting Werft, Travem nde, Germany Basic dimensions: Gross registered tonnage 4280 t Net registered tonnage 1284 t Displacement 4780 t Length overall 97.50 m Beam 16.50 m Draught max. 5.60 m Service speed 12 kn Personnel Crew: 32, Scientists: 28, German Weather Service: 2 Main engine 4 ( Mak 6 M 332 = 4 ( 1000 kW at 750 rpm Propulsion Diesel-electrical, tandem motor = 2 ( 1150 kW Maneuvring propulsion devices: Special rudder with flap, type Becker FKSR Omnithruster-bowthruster 919 kW, 10 t thrust thwartships, Fuel consumption About 12 t IFO 80 per day at service speed Maximum cruise duration 60 days Nautical equipment Integrated navigation system with data transfer to position computer, echo sounder synchronization and supervision, data processing facility Science quarters 20 laboratories on the main deck with approximately 400 m2 working space for multidisciplinary research. Air chemistry lab above the wheelhouse. About 400m2 of free deck working area, mainly with timber planking. Very little vibration and noise achieved by special construction. Meteor (I) was built in 1915 in Danzig as a gunboat for the German navy. However, it never reached completion as such and remained in an unfinished state until 1925, when it was converted in Wilhelmshaven to the first German research and survey vessel of that name. The steel-hull ship Meteor (I) had a length overall of 71.15 m, a displacement of 1179 t, and carried a crew of 122 plus 11 scientists. One of its first expeditions was the German Atlantic Ocean Expedition of 1925-27, which was organized by the Institute for Marine Research in Berlin. Thereafter, the vessel was used until 1934 for German physical, chemical, and microbiological marine investigations and for navy surveying as well as fishery protection duties. Meteor (II) was carefully planned after the 1950s; it was jointly operated by the German Research Foundation (DFG) in Bonn and the German Hydrographic Institute (DHI) in Hamburg. With a length overall of 82.10 m and a displacement of 3054 t, the second Meteor carried 52 in crew and 24 scientists. Commissioned in 1964, Meteor (II) participated in the International Indian Ocean Expedition. During 73 voyages between 1964 and 1985, the Meteor (II) sailed a total distance of about 650,000 nm to all parts of the world's oceans. Meteor (III), used during the intercomparison exercise described here, was completed in 1986, replacing Meteor (II). Based in Hamburg, it is used for German marine research worldwide and for cooperative efforts with other nations in this field. The vessel serves scientists of all marine disciplines in all of the world's oceans. 2.1.2 R/V Meteor, Cruise 36/1 Information Ship name Meteor Cruise/leg 36/1 Location Hamilton, Bermuda, to Las Palmas, Gran Canaria, Spain Dates June 6-19, 1996 Chief scientist D. Schulz-Bull, Institute of Marine Research, Kiel Master M. Kull Institutions Participating in the Exercise BNL Brookhaven National Laboratory, Department of Applied Science, Upton, Long Island, New York, U.S.A. CSIRO Commonwealth Scientific and Industrial Research Organisation, Division of Oceanography, Hobart, Tasmania, Australia IfMK Institut fuer Meereskunde Kiel (Institute of Marine Research at the University of Kiel), Kiel, Germany MRI Meteorological Research Institute, Tsukuba, Japan NBI Niels Bohr Institute for Astronomy, Physics and Geophysics, University of Copenhagen, Copenhagen, Denmark OU Okayama University, Okayama, Japan SIO Scripps Institution of Oceanography, Marine Physical Laboratory, La Jolla, California, U.S.A. UP&MC Universit Pierre et Marie Curie, Laboratoire de Physique et Chimie Marines, Paris, France WHOI Woods Hole Oceanographic Institution, Department of Marine Chemistry and Geochemistry, Woods Hole, Massachusetts, U.S.A. Parameters measured Institution Principal investigators Conductivity-temperature-depth (CTD), IfMK J. Waniek salinity, expendable bathythermograph (XBT) Nutrients IfMK D. Schulz-Bull, A. Koertzinger Oxygen IfMK D. Schulz-Bull, A. Koertzinger Total dissolved inorganic carbon (CT) BNL, IfMK K. Johnson, A. Koertzinger Alkalinity (AT) IfMK L. Mintrop PH WHOI C. Goyet SIO A. Dickson BNL D. Wallace Fugacity of CO2 (fCO2) CSIRO B. Tilbrook IfMK A. Koertzinger MRI H. Inoue NBI R. Torres OU E. Ohtaki UP & MC A. Poisson WHOI C. Goyet 2.1.3 Brief Cruise Summary After completion of the previous cruise 35/4, the R/V Meteor reached Hamilton, Bermuda, on June 4, and Detlev Schulz-Bull (IfMK) relieved Dieter Meischner (University of Gottingen, Germany) as chief scientist. A reception for invited officials of governmental and scientific institutions as well as private companies was held on board the Meteor on June 4. The scientific party of cruise 36/1 embarked on June 5. Equipment setup began on the same day. The R/V Meteor departed Hamilton at 9:00 a.m. local time on June 6, 1996. The cruise track of cruise 36/1 (Fig. 1 in hard copy) ran on straight lines from Bermuda to the Flemish Cap off Newfoundland, Canada, and then to Gran Canaria, Spain. The turning point was located at 46° 40' N, 41° 54' W. All seven underway pCO2 systems were operated simultaneously for most of the time between June 7 and June 17. Small technical problems that occurred to some of the systems only caused short interruptions. Only one system suffered heavy damage in the infrared gas analyzer and it had to quit measurements by June 14. The two underway spectrophotometric pH systems were operated throughout the cruise. The newly modified coulometric SOMMA system for underway determination of CT was tested successfully at sea and contributed about 450 high-quality underway CT measurements along the cruise track (Johnson et al. 1998). Synchronized with the XBT survey, a total of 57 discrete samples were taken from the seawater supply and were analyzed for pH, CT, and AT. The discrete fCO2 measurements could not be carried out on the same schedule; samples were taken for this parameter only at about 17 stations. In addition to the various surface measurements (whether continuous or discrete), five hydrographic stations were occupied during the cruise. Samples were drawn for measurements of all four CO2 system parameters (pH, fCO2, CT, AT) thus yielding the highest possible overdetermination of the marine CO2 system. The R/V Meteor arrived at Las Palmas, Gran Canaria, Spain, on June 19, 1996. Weather and sea conditions had been excellent throughout the cruise allowing for uninterrupted scientific work. 2.2 TECHNICAL ASPECTS OF THE EXERCISE 2.2.1 The Underway Pumping System The intercomparison exercise was almost entirely based on continuous underway sampling of surface seawater. All participating groups operated their underway fCO2 systems simultaneously on the same seawater pumping system. Like most up-to-date research vessels, the R/V Meteor provides a special seawater pumping system for scientific purposes. However, from experience, it is known that the use of this kind of pumping system for measurements of dissolved gases may be hampered by a number of problems. Pump action may cause cavitation when underpressure is applied to the water flow, thus making undisturbed gas measurements nearly impossible. Because of the location of the seawater intake close to the bow on R/V Meteor, air bubbles are introduced into the water lines in a rough sea. This again possibly biases the concentration of dissolved gases or even makes seawater sampling technically impossible in such cases. Furthermore the unavoidable warming of seawater during its travel from the bow intake to the user may be quite significant. In the case of the fCO2 intercomparison exercise, it was desirable to keep the temperature change as small as possible. As a result of the sluggish exchange of CO2 between the gas phase and the water phase, sampling for CO2 measurements (e.g., fCO2, pH, CT) is less susceptible to biases caused by inadequate pumping techniques than is sampling of reactive gases like oxygen. Nevertheless a careful sampling technique was an important aspect of the exercise. For this reason, a simple and reliable underway pumping system (see also Koertzinger et al. 1996b) was designed for use in the "moon pool" of R/V Meteor. The system consisted of a small CTD probe (ECO type, ME Meerestechnik-Elektronik GmbH, Trappenkamp, Germany) for measuring in situ seawater temperature and salinity at the intake as well as a submersible pump, both of which were installed in the shell plating at the bottom of the "moon pool." Figure 2 in hard copy shows a schematic drawing of this underway pumping system. The system also includes a separate Global Positioning System (GPS) receiver (GPS 120, Garmin/Europe Ltd., Romsey, Hampshire, U.K.). Navigational data from the GPS system as well as CTD data were continuously logged on a computer. The moon pool of R/V Meteor is specially designed for sampling purposes so that no cooling or wastewaters are emitted ahead of it and even at full speed or in a very rough sea no air bubbles reach it. Seawater was pumped through the moon pool from below the ship by means of a large submersible pump (multivane impeller pump, type CS 3060, ITT Flygt Pumpen GmbH, Langenhagen, Germany) at a pumping rate of about 350 L/min-1 (pump head approx. 12 m). The CTD probe was installed next to the submersible pump. All underway fCO2 systems were assembled in the geology lab of R/V Meteor (see Sect. 2.2.2). Two seawater supply lines (port and starboard) were teed-off from the main bypass and laid through the lab. All underway systems were hooked-up to these supply lines which delivered the necessary flow rates of seawater to each system (approx. 1-15 L/min-1). The wastewater from the systems was collected in three 100-L carboys and from there was disposed of continuously through the floor drains of the geology lab. In case of (occasionally observed) clogging of the lab's floor drains as a result of rough sea conditions, small submersible pumps (multivane impeller pump, type GS 9565, ITT Flygt Pumpen GmbH, Langenhagen, Germany) were at hand to pump the wastewater actively out of the lab. These pumps did not have to be used during this cruise, however. 2.2.2 The Laboratory Setup During cruise 36/1 four labs were reserved for the intercomparison exercise (Fig. 3 in hard copy). All underway systems were assembled side by side in the geology lab (no. 16) of R/V Meteor. This lab is the largest lab on the R/V Meteor. It is located on the main deck, starboard side, with direct access to the working deck. Adjacent to the geology lab is the universal lab (no. 15, not shown in Fig. 3), where the dynamic transformer was installed. The SOMMA coulometric analyzer for CT and the alkalinity titration system were installed in the clean lab (no. 4, not shown in Fig. 3) on the port side of the ship. The moon pool is located in the hold (Lab 17) for the CTD rosette. This lab is very close to the main lab of the exercise (Lab 16) thus allowing for short water lines of the seawater pumping system. 2.2.3 Other Infrastructure of the Exercise In addition to the common seawater line (Sect. 2.2.1), a common supply of calibration gases was regarded a key requirement for the exercise, as otherwise systematic errors most likely would have been introduced. We therefore provided a whole suite of calibration gases. Fifteen cylinders with precisely known amounts of CO2 in natural dry air covering a nominal concentration range from 250 ppmv to 500 ppmv were purchased from the National Oceanic and Atmospheric Administration (NOAA) Climate Monitoring Diagnostics Laboratory (CMDL) in Boulder, Colorado, U.S.A. Before final filling, all aluminum standard cylinders (Scott Specialty Gases Inc., Plumsteadville, Pennsylvania, U.S.A.) undergo a conditioning period of at least one week with clean ambient air. To prepare the standards, the cylinders are filled with ambient air at Niwot Ridge, Colorado. The air is dried using magnesium perchlorate and either scrubbed with Ascarite or spiked with a 10% CO2-in-air mixture to obtain mixing ratios below or above ambient levels, respectively (Zhao et al. 1997). Six cylinders of this consistent suite of gases were used during the exercise by all groups for calibrating their instruments. Additionally nitrogen (purity 99.999%) was used by some groups for zeroing their gas analyzers. The mixing ratios of CO2 in the cylinders were calibrated in the NOAA/CMDL Carbon Cycle Group laboratories on three separate days over a period of 2-3 weeks. The results of these calibrations are summarized in Table 1. The CO2 mixing ratios are reported as micromoles per mole (µmol/mol-1 = ppmv) of dry air in the World Meteorological Organization (WMO) X85 mole fraction scale, traceable to primary standards at the Scripps Institution of Oceanography (SIO). The NOAA/CMDL calibrations are done by comparison on a nondispersive infrared CO2 analyzer against four tertiary standards with assigned mixing ratios traceable to SIO (Thoning et al. 1987; Zhao et al. 1997). The uncertainty of the assigned values for the tertiary standards is approximately 0.06 ppmv. The tertiary set of standards used ranges between 250 and 450 ppmv CO2. The repeatability of the NOAA/CMDL calibrations depends on the stability of the CO2 mixing ratio in the cylinder and the fit of the analyzer response to the known tertiary standards. For cylinders that are stable and within the range of standards, the repeatability is on the order of 0.01 ppmv. The overall uncertainty associated with precision is therefore about 0.06 ppmv. When calibrating cylinders at the extremes of the tertiary standards or extrapolated outside the range, the reproducibility decreases. For mixing ratios above 450 ppmv, the reproducibility is on the order of (0.3 ppmv and further decreases with the interpolation away from the tertiary standards. The absolute accuracy of the assigned mixing ratios is determined by the accuracy of the SIO standards (Keeling et al. 1986, and references therein). Table 1. Summary of calibration results for six cylinders with CO2 in natural dry air The measurements were carried out at the NOAA/CMDL Carbon Cycle Group Laboratory in Boulder, Colorado, U.S.A. These six cylinders constitute the suite of calibration gases used by all participating groups during the exercise. ------------------------------------------------------------------ Cylinder Date of Measured CO2 Average CO2 Standard # measurement concentration concentration deviation (MM/DD/YY) (ppmv) (ppmv) (ppmv) ------------------------------------------------------------------ 2178 03/06/96 252.42 2178 03/11/96 252.46 2178 03/14/96 252.45 252.44 0.02 1996 02/22/96 298.43 1996 03/07/96 298.47 1996 03/18/96 298.42 298.44 0.03 2172 02/05/96 349.53 2172 03/14/96 349.52 2172 03/18/96 349.51 349.52 0.01 1980 03/05/96 403.85 1980 03/11/96 403.84 1980 03/13/96 403.87 403.85 0.02 2186 02/27/96 450.69 2186 02/29/96 450.68 2186 03/13/96 450.73 450.70 0.03 2112 04/10/96 511.28 2112 04/15/96 511.61 2112 04/19/96 511.60 511.50 0.19 ------------------------------------------------------------------- According to the different power requirements of the analytical systems, the ship provided three different power sources, the standard 220V/50Hz system as well as two additional systems for 220V/50Hz (static transformer) and 220V/60Hz (dynamic transformer). 3. METHODS AND PROCEDURES 3.1 PRINCIPLE OF MEASUREMENT OF THE FUGACITY OF CO2 The principle of the measurement of the fugacity of CO2 (fCO2) in seawater is based on the determination of the CO2 mixing ratio in a gas phase that is in equilibrium with a seawater sample at known temperature and pressure. The CO2 mixing ratio can either be measured with a nondispersive infrared analyzer (NDIR) or with a gas chromatograph (GC) with flame ionization detector after catalytic conversion of the CO2 into methane. Whereas the GC approach has a few advantages (e.g., small sample volume and the ability to measure additional trace gases), the more rugged infrared technique has shown better suitability for use at sea and allows measurements in a truly continuous fashion. Depending on the sampling strategy (discrete or continuous), two different families of analytical systems have been developed. For the determination of the fCO2 in air that is in equilibrium with a discrete sample, a known amount of seawater is isolated in a closed system containing a small known volume of air with a known initial CO2 mixing ratio. For the determination of the fCO2 in air that is in equilibrium with a continuous flow of seawater, a fixed volume of air is equilibrated with seawater that flows continuously through an equilibrator. Continuous (or underway) fCO2 systems are more widely used in marine CO2 research. They provide important information about the saturation state of seawater at the air-sea interface when operated on board research vessels with a continuous flow of seawater usually obtained by means of a shipborne pumping system. 3.2 PARTICIPATING UNDERWAY fCO2 SYSTEMS Throughout this report we present technical details as well as the results of the participating systems in a semi-anonymous fashion. The main reason for this is the fact that the results of the exercise cannot easily be extrapolated to the performance of any participating system in general. Strictly they are only representative for this single cruise. To avoid the erroneous association in the scientific community of the performance of a particular system during this exercise with the general performance of this system, we choose to report in this semi-anonymous fashion. Seven underway fCO2 systems, all of which are based on NDIR detection of CO2, participated in this exercise. Most of these systems have received detailed descriptions in the literature, which can therefore be omitted here. Where such publications are available, they are reprinted at the end of hard copy report (Appendix B). For two systems, however, this is not the case. One is the fCO2 system of CSIRO which features a slightly smaller Weiss-type equilibrator and is otherwise quite similar to the other systems. The second one is a system that is manufactured commercially by a U.K. company (Challenger Oceanic, Haslemere, Surrey, U.K.). For details about the latter system, further information is available through the company's internet site http://www1.btwebworld.com/challengeroceanic/index.html Whereas most of the underway fCO2 systems are similar in the general design and principle of measurement, they are considerably different in detail. For quick reference, the main features of all underway fCO2 systems are summarized in Table 2. All different equilibrator design principles (i.e., showerhead, bubbler, and thin film type) were represented by at least one system, with the majority being of the showerhead type. In most systems (except "D" and "F") these equilibrators are vented to theµatmosphere and thus operated at ambient pressure. The volumes of water and air in the equilibrators cover a wide range from a few milliliters to 15 liters. This is also true for the flow rates of water (0-15 L/min) and air (0.17-0.8 L/min) through the equilibrators. Table 2. Summary of main features of the underway fCO2 systems "A" through "G" that participated in the exercise ------------------------------------------------------------------------------------------------ "A" "B" "C" "D" "E" "F" "G" ------------------------------------------------------------------------------------------------ Equilibrator Design Showerhead Bubbler Showerhead Thin film* Showerhead Bubbler Showerhead Total volume 1000 mL 1400 mL 13.1 L 119 mL 11.0 L 36 mL 1200 mL Water volume 500 mL 1000 mL 2.3 L 21 mL 10.0 L 18 mL ~75 mL Air volume 500 mL 400 mL 10.8 L 98 mL 1.0 L 18 mL 500 mL Water flow rate 4-6 L/min 2.0 L/min 8.0 L/min 2.0 L/min 10-15 L/min 0 L/min** 1.2 L/min Air flow rate 0.2 L/min 0.8 L/min 0.5 L/min 2.0 L/min 0.5 L/min 0.17 L/min 0.18 L/min Vented? Yes Yes Yes No*** Yes No Yes CO2 measurement Method NDIR NDIR NDIR NDIR NDIR NDIR NDIR Wet/dry? Wet Wet Wet Dry Dry Dry Wet Analyzer calibration No. of st. gases 2 2 2 2 4 2**** 2 Zero gas? No Yes No No No Yes No Measurement cycle Calibr. freq. 6-8 h 6 h 6 h 4-6 h 1.5 h 15 min 2 h Air meas. freq. 6-8 h 1 h 6 h 4-6 h 0.5 hr n/a 7 min Interrogat. interv. 6 sec 6 sec 1 sec 10 sec 0.1 sec 15 min 0.33 sec Averaging interv. 1; 3 min 1 min 4 min 5 min 1 min n/a 1 sec Data points/interv. 10; 30 10 240 33 600 1 3 ----------------------------------------------------------------------------------------------- *Film thickness approximately 0.75 mm. **Semicontinuous approach. ***Vented only every 20 min. ****Standard gas generator is initially calibrated using all six calibration gases; linearity checks are carried out for every sample with only two calibration gases. A further distinction can be made in whether the sample gas is measured dry or wet. The traditional procedure is based on NDIR measurement of the dried sample gas ("D," "E," and "F"). However, in four systems ("A," "B," "C," and "G") the sample gas is not dried prior to NDIR measurement. This is feasible on the basis of the LI-6262 CO2/H2O gas analyzer (Li-Cor Inc., Lincoln, Nebraska, U.S.A.) which is a dual-channel instrument that simultaneously measures CO2 and H2O mole fractions of the sample gas and provides internal algorithms for correction of the diluting and pressure-broadening effects of water vapor on the CO2 measurement (McDermitt et al. 1993). All NDIR instruments were calibrated with the NOAA/CMDL CO2 standards provided by the organizer (Table 1). Because of the individual calibration procedures, different numbers of gases (2 to 4) were required. Some systems also required a zero gas (nitrogen, purity 99.999%) for calibration purposes or as a reference gas. Whereas underway fCO2 systems "A" through "E" and "G" are similar in that fCO2 is calculated from the CO2 mixing ratio in a gas phase that is in equilibrium with a constantly renewed seawater phase, system "F" is of a principally different design. Here, for every fCO2 measurement, five aliquots of a discrete seawater sample (semicontinuous mode) are equilibrated with five different standard gases bracketing the observed range of seawater fCO2. For each equilibration run, changes with time in the standard gas CO2 concentration as a result of CO2 exchange with the sample aliquot are recorded in terms of positive or negative deviations from the standard's initial CO2 concentration. If flow conditions during these five equilibration runs are kept identical, the heights of the resulting deviation peaks are proportional to the concentration difference between the carrier gas and a gas that is in equilibrium with the sample. If peak heights are plotted versus the initial xCO2 of the standard gases, the equilibrium xCO2 can be found where a linear regression to the five data points intersects the x-axis. Participating groups were asked to operate their systems according to their typical operation profile (i.e., frequency of calibration and air measurements, interrogation, and averaging intervals, etc.). This strategy was chosen to ensure that all systems were operated in modes to which they have been optimized in the field and in which their operators have gained the highest confidence. The consequence, however, was quite different averaging and/or reporting intervals for the different groups. In particular, the averaging intervals between 1 and 5 minutes have certain implications that need to be taken into account when the data are being compared. This inherent discrepancy of the whole data set represents a certain limitation for the temporal resolution to which the interpretation can be extended. This is discussed in more detail in the results section. 3.3 PARTICIPATING DISCRETE fCO2 SYSTEM The only discrete fCO2 system ("H") involved in this intercomparison exercise is based on a batch-equilibration, static-headspace technique that requires a small sample volume of 60 mL and has an average analysis time of only 2 min per sample. It includes closed-system equilibration of a headspace in a shaking water bath, followed by analysis of the CO2 mole fraction in the water-saturated equilibrated headspace by gas chromatography with flame-ionization detection (GC-FID). The method has been described in detail by Neill et al. (1997). This paper is also reprinted in Appendix B at the end of hard copy report. It should be noted that this method is not specifically designed for work in surface seawater but for full-depth profiling. The equilibration temperature (i.e., the water bath temperature) was changed two times during the exercise, from 17°C (June 10) to 20°C (June 11-13, first sample) and, finally, to 25°C (from June 13, second sample). The magnitude of the correction of fCO2 from the temperature of equilibration to the in situ temperature was 44-133 µatm (mean: 89 µatm) for the samples presented here. 3.4 CHECKS AND CALCULATION ROUTINES The main idea of the exercise was to compare the surface seawater fCO2 data as measured by all participating instruments under identical conditions. This was to some extent accomplished by providing the infrastructure during the exercise, such as a common seawater and calibration gas supply (Sect. 2.2). The same care, however, that had been taken on the side of the logistical infrastructure was also advisable with respect to ancillary measurements as well as the calculation procedures involved in the computation of final fCO2 values. This issue was addressed in different ways. In the following sections we describe the results of two different experiments: (1) a check of the performance of the calibration procedures for CO2 (Sect. 3.4.1), and (2) a check of all temperature sensors that were used to measure the seawater temperature in the equilibrators (Sect. 3.4.2). We also describe the common procedure of the calculation of final fCO2 values (Sect. 3.4.3) and of the synchronization of the final fCO2 profiles (Sect. 3.4.4). 3.4.1 Check of CO2 Calibration Performance In order to check the performance of the individual calibration procedures, every group measured between one and four NOAA/CMDL CO2 standards in the nominal concentration range of 250-500 ppmv as "unknown samples." Depending on the individual calibration procedure, different CO2 standards were measured. Figure 4 in hard copy shows the results of this exercise. It should be pointed out that this check was carried out on the last day of the exercise (June 17). Therefore no data are available for system "A," which had to prematurely quit the exercise on June 13 because of major technical problems. For system "C" only one standard could be measured because the measurement range had been fixed to an upper limit of 400 ppmv, which was slightly exceeded by the relevant NOAA/CMDL standard (403.85 ppmv). System "F" required all six standards for initial calibration, which could therefore not be measured as "unknown samples." The results show that essentially all checked NDIR instruments were calibrated to an accuracy on the order of 1 ppmv or better over the whole concentration range of 250 to 500 ppmv. Only system "G" shows deviations of 3 to 6 ppmv. This is indicative of a systematic problem associated with the calibration of the CO2 analyzer or with the measurement itself. Such deviations are clearly not tolerable and need to be addressed thoroughly. System "B" makes use of the factory calibration of the LI-COR LI-6262 instrument, which only requires the adjustment of "zero" (with a CO2-free gas) and "span" (with a single CO2 standard). It appears that some accuracy is lost by this somewhat crude calibration technique,1 and the "classical" approach using at least two CO2 standards spanning the range of anticipated CO2 mixing ratios is preferred. 3.4.2 Check of Equilibrator Temperature Sensors In marine applications, fCO2 results are generally reported at in situ seawater temperature (T in situ). As the seawater temperature in the equilibrator during measurement (T eq) usually deviates from T in situ, a temperature correction needs to be applied. The size of this correction obviously depends on the choice of the parameterization (Sect. 3.4.3) and the size of the temperature deviation itself. As the fCO2 strongly varies with temperature, measurements of T in situ and T eq have to be made rather accurately. An error of 0.1°C in the resulting temperature deviation (Teq-Tin situ) is equivalent to an error of about 0.4 % or 1.5 µatm in fCO2 (at 350 µatm). During the exercise, T in situ was measured with a CTD installed at the seawater intake in the bottom plate of the "moon pool". The CTD had been calibrated just before the exercise to an accuracy of ±0.05°C. These in situ temperature readings were used in the calculation of all fCO2 data. In order to exclude possible errors contributed by inaccurate measurements of T eq, all groups had their equilibrator temperature probes referenced against a recently calibrated platinum resistance thermometer (Pt-100) provided by WHOI. For every comparison, equilibrator probe and reference probe were kept together in a water bath until the readings had stabilized. In most cases this was done at three temperatures between 0°C and up to 30°C. Based on the deviation from the reference, an individual linear correction was calculated for every system and applied to its measurements of T eq. In one case (lab "F"), the temperature probe could not be removed easily from the water bath surrounding the equilibrator and the reference probe had to be installed next to it in the bath thus yielding only a single measurement which was then treated as a uniform offset. Figure 5 in hard copy shows measured deviations and the resulting correction lines of only those temperature sensors which were used in the calculation of final fCO2 values (some systems feature up to three equilibrator temperature sensors). The observed deviations are roughly between -0.5°C and +0.1°C with a clear tendency towards negative values and a negative slope of the linear correction line. If this inconsistency of the temperature measurements is not accounted for, differences of up to 2% or about 7 µatm (at 350 µatm) in the final fCO2 values are caused as an artifact entirely the result of inaccurate temperature measurements. Even though the CTD as well as the WHOI reference thermometer may themselves have been affected by some degree of miscalibration, the present procedure of referencing all measurements to these two temperature sources removes the incompatibility of all temperature readings to better than 0.1°C or 1.5 µatm. It should be pointed out that the observed deviations of up to 0.5°C are clearly above a tolerable level. Temperature readings have to be carried out with an accuracy of at least 0.1°C. Ideally they should be checked for consistency with the temperature probe used to measure in situ temperature. 3.4.3 Calculation of fCO2 Results The calculation of final fCO2 values from the raw voltage readings of an NDIR analyzer involves a number of steps that are only briefly described here. More detail of the calculation procedure can be found in Appendix A and in the reprints of the pertinent literature section (Appendix B). The NDIR detector signal depends on the number of CO2 molecules in the optical path which, in turn, is mainly a function of pressure and temperature for a given CO2 mixing ratio. The calculation procedure, therefore, requires temperature and pressure corrections to account for any fluctuations in these parameters as well as a calibration function. First the raw voltage readings are corrected to a standard pressure of one atmosphere (p^0) to account for fluctuations of the NDIR cell pressure. This requires continuous monitoring of the pressure in the cell. It has been found empirically (Welles and Eckles 1991) that pressure p affects the voltage signal v of NDIR analyzers in a linear fashion: v'=v(p^0/p) NDIR instruments are calibrated using standard gases with known CO2 mixing ratios in dry air. The mixing ratio of a component gas (like CO2) in a mixture of gases (like air) is equivalent to its mole fraction (xCO2), assuming ideal behavior. The CO2 mixing ratio of the standard gases should closely bracket the expected range of the sample xCO2. Although the response of NDIR analyzers is considerably nonlinear, the use of a simple linear calibration function is generally justified over a small concentration range of 100-200 ppmv. The error incurred by this approximation is typically on the order of a few tenths of a ppmv. Furthermore any deviation of the NDIR cell temperature T from the calibration temperature T^0 has to be accounted for. Welles and Eckles (1991) have shown that the mole fraction xCO2* is scaled linearly with the inverse of the absolute temperature: xCO2=xCO2*(T/T^0) The resulting CO2 mole fraction xCO2 in dry air is temperature and pressure corrected. The latter because the sample gas is either measured dry (i.e., after full removal of water vapor) or has been arithmetically corrected for the diluting and pressure-broadening effects of water vapor based on simultaneous wet xCO2 and xH2O measurements. As the air at the air-sea interface can be assumed to be at 100% humidity, a correction has to be applied to account for the increase of the CO2 mole fraction that is the result of the (actual or arithmetical) removal of water vapor prior to the infrared measurement. Here the saturation water vapor pressure of seawater at equilibrator temperature was calculated using an equation by Weiss and Price (1980), which is valid over the temperature range 273-313 K and the salinity range 0-40 (see Appendix A, Part 1). For very accurate interpretations the non-ideal behavior of CO2 should be taken into account (i.e., fugacity has to be used instead of partial pressure). As the results are to be used later for consistency checks, together with other parameters of the CO2 system in seawater, we decided to use fCO2. The calculation of the fCO2 at equilibrator temperature from the measured mole fraction (xCO2) in dry air is described in detail in Appendix A (Part 2). The fugacity coefficient (i.e., the ratio between fugacity and partial pressure of CO2), is on the order 0.996 to 0.997 under typical conditions (p = 1µatm, T = 270-300 K, pCO2 = 350 µatm). Barometric pressure readings from the shipborne meteorological sensor were used for all calculations of final fCO2 data. Because the fCO2 in seawater strongly varies with temperature, the final step in the calculation of fCO2 (in situ) requires a correction to compensate for any difference between the equilibration temperature and the in situ seawater temperature. Different equations have been proposed for the temperature dependence of CO2 partial pressure/fugacity in seawater (e.g., Gordon and Jones 1973; Weiss et al. 1982; Copin-Montegut 1988, 1989; Goyet et al. 1993; Takahashi et al. 1993). Because temperature deviations were typically on the order of a few tenths of a degree for all systems during the exercise, the correction is rather small and the choice from the above suite of equations is not critical. We have chosen the equation based on temperature and salinity given by Weiss et al. (1982), which is valid for ranges of 0 to 36°C in temperature, 30 to 38 in salinity, and 80 to 2000 µatm in fCO2 (see Appendix A, Part 3.). All temperature corrections of the fCO2 measurements during this exercise are based on this equation. 3.4.4 Synchronization of Surface Measurements Profiles of in situ temperature and salinity in surface seawater, as measured by the CTD probe at the seawater intake, and the different seawater fCO2 profiles had to be matched and synchronized. Given the strong gradients in surface seawater temperature encountered during some periods of the cruise, this was very important for a reliable estimation of the differences between equilibrator and in situ temperatures. At the beginning of the exercise, all systems were switched to Universal Time Coordinated (UTC) time. The UTC time readings of all measurements are therefore the primary criterion for matching the data sets. However, UTC time alone would not produce a proper synchronization of the profiles for two reasons. The first is the different time the water travels from the seawater intake (where its temperature and salinity are being measured) to a given equilibrator (where the equilibration temperature is determined). This depends mainly on the individual flow rate of water and to some extent also on the location of the equilibrator in the supply line. By running two separate supply lines (port and starboard side line), which were kept at roughly the same total flow rate, we tried to make the supply flow characteristics comparable for all systems. With a single supply line, the ratio of water consumed for analyses to the water bypassing through this supply line would have changed more strongly en route with unknown implications for the water characteristics (such as the temperature deviation). With the chosen setup (see also Fig. 3), we tried to make the supply similar for all systems. The second, rather trivial, reason for an insufficient synchronization of the profiles based on UTC readings alone is that there are errors in the UTC readings themselves, which in some cases appear to account for 1 to 2 min during the course of the exercise. The final matching of the profiles is based on the assumption that the profiles of in situ and equilibrator temperature should be connected by a fixed daily temperature offset. This is a first-order approximation, because the offset certainly depends on the stability of the water flow rate and the difference between seawater and ambient temperature. Flow rates were usually kept constant during the course of the exercise. The change in seawater temperature was significant, but its effect was minimized by matching the profiles on a daily basis. The matching procedure involved correcting every fCO2 profile with daily time lags in 1-min steps until the standard deviation of the difference between in situ and equilibrator temperature reached a minimum. This could always be achieved by time lags of < or = 3 min. In other words the fCO2 profiles were shifted minute-wise backwards in time against the CTD readings until the two temperature profiles showed the best match with the smallest standard deviation of the resulting offset. This procedure proved very necessary. Even a mismatch of 2 min could cause a bias in the calculated temperature difference of up to 1°C and more (i.e., > or = 10 µatm) in the strong gradient regime. In the more stable regime, toward the end of the exercise the effect of this synchronization procedure is less pronounced or even negligible. On the other hand, the profiles could not be synchronized to better than 1 min, which still allows errors of the order of several µatm in some cases only because of temporal mismatch. This is an important aspect which restricts the interpretability of the results during passage of the very strong gradients. Even after correction of all equilibrator temperature readings and after this synchronization procedure, the remaining uncertainty is on the order of 2 µatm for the largest portion of the cruise. To put it the other way round, any differences of < or = 2 µatm between the final fCO2 profiles are not significant under the circumstances of this exercise. During passage of the strongest gradients, the overall uncertainty is definitely higher than 2 µatm, at least for short periods, and may account for a mismatch of up to 5 µatm. 4. RESULTS 4.1 SURFACE TEMPERATURE AND SALINITY As said before, the cruise track of the R/V Meteor during the exercise (Fig. 1) was chosen in order to provide the largest possible range of surface temperatures and salinities in the whole area of the North Atlantic Ocean accessible during this rather short cruise. It was assumed that this track should likely provide more stable conditions in the eastern part as well as a highly variably situation at the northern turning point near the Flemish Cap off Newfoundland. This assumption was later verified to a full extent by the encountered ranges of surface temperature and salinity. Figure 6 shows the large observed ranges in surface temperature and salinity as measured during the course of the exercise: Surface seawater temperatures ranged from 6.2°C to 25.1°C while surface salinities covered a range from 32.6 to 37.0. This is equivalent to a span of 19°C in temperature as well as 4.4 in salinity. It should be pointed out that the observed temperature and salinity drop around 51° W was as large as 2.9°C/min-1 and 0.4/min-1 for salinity, which is equivalent to 4.2°C/km-1 and 0.9/km-1, respectively. Such gradients can be regarded as extreme situations that represent a "worst case scenario" for any kind of intercomparison rather than a typical open ocean situation. Toward the eastern part of the cruise track, a more typical regime was found that represents the standard case for at-sea operation. In order to provide the hydrographic background for the fCO2 data, measurements of surface temperature and salinity are given as 1-min averages in daily figures in hard copy (Figs. 7-9) for the period June 8-16. 4.2 COMPARISON OF ATMOSPHERIC xCO2 DATA Measurements of theµatmospheric xCO2 were carried out by all underway fCO2 systems except system "F" (see Table 2 in Sect. 3.2). As will be shown, the atmospheric xCO2 data-while not immediate focus of this exercise-may still provide additional information for identifying likely sources of error in the surface fCO2 profiles. All xCO2 data are given (in ppmv) for dry air and shown in Fig. 10 in hard copy. Of the six data sets, four show good agreement to within ±1 ppmv throughout the exercise: Profiles "C" and "D" show virtually identical values (except for a few data points), whereas profile "E" tends to values that are lower by -0.5 to -1 ppmv. Profile "B" is characterized by a somewhat variable behavior: For most of the time, "B" is in very good agreement with "E." However, from June 10, 08:30 UTC, to June 11, 14:30 UTC, "B" shows a positive offset of about 1 ppmv from "E," hereby agreeing perfectly with "C" and "D." In contrast, "B" deviates by -0.5 to -1.0 ppmv from "E" during the period from June 14, 14:30 UTC, until June 13, 18:00 UTC, which is equivalent to an offset of -1.0 to -1.5 ppmv with respect to profiles "C" and "D." We have calculated a mean xCO2 (air) of 366.21 ± 0.72 ppmv for the period of the exercise where data from all six systems are available from the means of profiles "B" through "E" which is shown in Fig. 10 (red line). Only results from this restricted period are used in the following comparison. Two profiles ("A" and "G") are characterized by a much larger scatter which obscures the pattern of theµatmospheric CO2 contained in the other four profiles. This scatter is not a real property of the sampled air, as proved by profiles "B" through "E," and thus indicates an analytical problem associated with these systems. All air intakes were located on the same spot above the wheelhouse of R/V Meteor approximately 20 m above sea level, thus making differences in the properties of the sampled air very unlikely. The majority of measurements of "G" show a positive deviation of up to 8 ppmv which is consistent with the rather large positive offset of 3 to 6 ppmv determined during the checks of the CO2 calibration performance (Sect. 3.4.1). The mean of "G" (368.27 ppmv) is 2.06 ppmv higher than the combined mean of "B" through "E" (366.21 ppmv). With a mean value of 362.89 ppmv, the xCO2 measurements of "A" are clearly marked by a negative offset of 3.32 ppmv with respect to the mean of "B" through "E." Figure 11 in hard copy shows the individual mean and standard deviation of each data set as well as an overall mean calculated from the mean of profiles "B" through "E," all for the restricted period of time only. The individual standard deviations reflect the averaging interval in the case of laboratories "B" through "E," where the smaller scatter is associated with the longer averaging intervals of 4 to 5 min (laboratories "C" and "D") and the somewhat larger scatter reflects averaging intervals of 1 min (laboratories "B" and "E"). In the case of laboratories "A" and "G" the scatter is no obvious function of the averaging interval but an expression of an analytical problem. In the comparison of surface fCO2 data in Sect. 4.3, these results, sometimes referred to as the general trends of agreement or disagreement between the xCO2 (air) data sets, will largely be retained in the fCO2 data. The combination of both results provides much of the argument for the discussion of the overall results. We will demonstrate that the three laboratories "C," "D," and "E" show the same high degree of agreement in surface fCO2 data as they do in xCO2, and a strong case will be made that these systems represent the "best" values of xCO2 (air) and fCO2. 4.3 COMPARISON OF SURFACE fCO2 DATA As described in Sect. 3.4, the following main steps in the calculation of final fCO2 values constitute the general procedure that was applied identically to all underway fCO2 data sets: - Calculation of xCO2 in dry sample air (final data product received from every group) - Synchronization of daily CTD and equilibrator profiles based on standard deviation of the temperature offset - Calculation of fCO2 in equilibrator (at T eq, 100% humidity) - Correction of fCO2 to in situ seawater temperature based on corrected temperature readings In order to gain better interpretability of any differences in the final fCO2 data sets, we tried to exclude as many controllable sources of error as possible. This was accomplished by carefully addressing the following points: - Temperature readings can be a significant source of error as shown in Sect. 3.4.2. However, on the basis of the checks of the temperature probes against a reference probe we were able to remove this error and assure consistent temperature measurements. - The choice of the parameterizations for calculating the saturation water vapor pressure and for the temperature correction of fCO2 also introduces some kind of uncertainty, which, however, in our case seems to be rather small compared with the errors of the temperature measurements. Again, the common calculation procedure (Sect. 3.4.3) excludes inconsistencies based on the use of different equations. - Finally, the common infrastructure (i.e., the seawater and calibration gas supply) assured a physically identical background for all systems (Sect. 2.2). It should be emphasized that none of these consistent conditions are usually present in typical fCO2 measurements in the field (i.e., temperature probes are sometimes used uncalibrated or at least not calibrated to the same standard; calculation procedures vary; calibration gases are of different origin and likely quality, too; the seawater sources may be quite different or even inadequate for gas measurements etc.). In the interpretation of the results, any differences of >2 µatm (up to >5 µatm in the highly variable regime) in the final fCO2 data can be attributed either to differences in the equilibration process itself and/or to differences in the subsequent measurement of CO2. A tool to separate these two possible sources of error is measurements of the atmospheric xCO2, which were discussed in some detail in Sect. 4.2. Unlike the calibration gases, atmospheric air is comparable to the seawater equilibrated air in that it has a wet sample matrix. Thus atmospheric air undergoes the same procedure of (physical or arithmetical) drying. If, for example, differences between seawater fCO2 data from two systems were also present in the atmospheric xCO2 data, this is indicative of problems associated with the infrared CO2 measurement and/or the drying procedure. If, in contrast, the atmospheric xCO2 data turned out to be identical while seawater fCO2 was different, the source of error must be attributed to the equilibration process and/or the way of handling of the seawater equilibrated air. Whereas these reasons add to the interpretability of the results, it should be pointed out that any observed differences cannot per se be attributed to a particular data set or system. As a superior reference method was not available, the "true" fCO2 values are simply not known. Given the still remaining uncertainty about the valid set of dissociation constants of carbonic acid in seawater, even consistency checks based on the other three parameters of the CO2 system in seawater (i.e., CT, AT, pH)-although to be carried out later on-will not provide an unambiguous means of finding "true" fCO2 values. However, we found three data sets (systems "C," "D," and "E") to be very close in seawater fCO2 and atmospheric xCO2 values throughout the cruise, whereas the other data sets show variable offsets to these three profiles and some of them are also associated with significantly larger scatter. Because the general design of the three systems "C," "D," and "E" is significantly different (showerhead equilibrator-thin film equilibrator, small equilibrator volume-large equilibrator volume, small flow rates-large flow rates, equilibrator vented-equilibrator not vented, wet CO2 measurement-dry CO2 measurement, etc.), this agreement cannot simply be attributed to an essentially identical design. This is by no means a sufficient argument to regard the three consistent fCO2 profiles as the "truth," although we feel that this marked agreement is at least a strong indication of this. However, with the lack of a superior method, this sort of discussion is to some extent futile and cannot be solved here. When preparing the following figures, we wanted to discuss the fCO2 results not only as absolute numbers but also as deviations from a reference. Because a superior method was not available and the choice of a single "true" fCO2 profile was not feasible, we decided to calculate the deviation of every single fCO2 data point from an 11-min running mean calculated from the three most consistent profiles (labs "C," "D," and "E"). In the light of the arguments given in the foregoing discussion, this choice remains arbitrary, but it nevertheless seems to be the most reasonable choice. However, it should be kept in mind that these deviations are, of course, dependent on the choice of the reference and are therefore not independent results. We are fully aware that this is a somewhat critical step in the interpretation, which seems only justified by the better visualization of the differences and the enhanced interpretability of the data set. A further problem associated with calculating deviations from an 11-min running mean stems from the fact that this reference represents a strongly smoothed profile, whereas the original fCO2 data represent significantly smaller averaging intervals (minimum 1 min). Thus, all temporal variability on the minute-scale as contained in the fCO2 data with higher temporal resolution (e.g., profiles "B" and "E") translates into a larger scatter than that of the deviations from the smoothed reference profile. This artifact has to be kept in mind, because it is of a different magnitude for the various fCO2 data sets. This effect is more strongly obvious in the strong gradient regime (e.g., June 10). The main message of these deviation figures therefore has to be the general offset rather than the scatter of a profile. Table 3 provides an overview of the minima, maxima, and differences of measured in situ temperature, salinity, and fCO2 (11-min running mean from profiles "C," "D," and "E") on a daily basis. As intended with the choice of the cruise track and as already documented in the daily profiles of temperature and salinity (Sect. 4.1), the encountered conditions of the surface waters along the cruise track varied between a smooth regime with low variability during the second half of the cruise and a strong gradient regime with much higher variability in the area close to the northern turning point off Newfoundland (marked in yellow) during the first half. Table 3. Overview of minimum, maximum, and difference of measured values of temperature T (°C), salinity S, and the fugacity of CO2 (fCO2, 11-min running mean from profiles "C," "D," and "E"). The strong gradient regime is shaded. -------------------------------------------------------------------------------------- June 8 June 9 June 10 June 11 June 12 June 13 June 14 June 15 June 16 -------------------------------------------------------------------------------------- Tmin 20.3 6.0 6.4 10.1 12.9 15.3 16.8 18.4 19.6 Tmax 25.1 24.5 16.4 16.6 16.9 17.1 19.8 20.6 21.0 del T 4.8 18.5 10.0 6.5 4.0 1.8 3.0 2.2 1.4 Smin 36.38 32.57 32.90 33.88 34.66 35.98 35.97 36.11 36.35 Smax 36.81 36.61 36.24 36.32 36.29 36.25 36.25 36.46 36.96 del S 0.43 4.04 3.34 2.44 1.63 0.27 0.28 0.35 0.61 fCO2min 315.6 270.2 264.6 281.5 276.9 303.8 306.7 332.2 338.6 fCO2max 340.7 339.2 321.0 327.9 327.3 326.4 346.4 359.3 355.7 del fCO2 25.1 69.0 56.4 46.4 50.4 22.6 39.7 27.1 17.1 -------------------------------------------------------------------------------------- 4.3.1 Underway Profiles Figures 12-20 in hard copy show the final underway fCO2 profiles "A" through "G" as well as the discrete fCO2 data of laboratory "H" (top) and the deviations of all fCO2 data from the 11-min running mean calculated from profiles "C," "D," and "E" (bottom). It should be noted that the top figures show variable scaling of the y-axis (see 5 µatm bar indicator), while the bottom figure is always at the same scale. The latter also includes two horizontal lines, one at +2 and one at -2 µatm deviation which is about the limit of interpretation. 4.3.2 Discussion of Profiles Following is a brief day-by-day description of the major features contained in Figs. 12-20. We tried to identify the most important results and to point to some major trends and changes. Again we would like to emphasize that the scatter of the bottom figures is mainly an artifact of the referencing procedure. This can be readily observed in Fig. 13 (June 9, 1996): Between 08:00 and 19:00 UTC the seawater exhibits low variability resulting in very little scatter in the bottom figure. Immediately before and after this period the seawater was much more variable which translates into the high scatter of the deviation figure. The observed offsets discussed here can therefore only be identified in the trends and have to be regarded as rough approximations. June 8, 1996 Missing data - "A": 04:00 to 10:00 UTC, "B": before 18:30 UTC (start delayed because of sample gas leakage), "F": 03:00 to 10:30 UTC, "H": no samples measured. Agreement to within ±2 µatm - "B," "C," "D," and "E". Positive offset - "F": 3 µatm. Negative offset - "G": 6 µatm. Variable offset - "A": -3 to +3 µatm. Comment - "G" starts with a marked negative offset, which turns slowly into a positive offset during the next days and then disappears toward the end of the exercise. However, the large scatter of "G" seen in theµatmospheric xCO2 readings is not visible here, which points toward problems with the handling of atmospheric air within this system (e.g., leakage in air pump, valves, or tubing). June 9, 1996 Missing data - "A": 10:00 to 12:00 UTC, "C": 13:00 to 24:00 UTC, "H": no samples measured. Agreement to within ±2 µatm - "B," "C," "D," and "E." Negative offset - "G": 6 µatm. Variable offset - "A": within 2 µatm (before 07:00 UTC), -5 to -8 µatm after 12:00 UTC; "F": within ±2 µatm (09:00 to 19:00 UTC), +5 to +10 µatm (before 09:00 UTC and after 19:00 UTC). Comment - "A" shows a sudden change around 12:00 UTC from good agreement to a negative offset of the order of 5 µatm. This offset remained until the end of the exercise. The scatter of profile "A" (3-min intervals) is significantly larger than in the 1-min averages of "B" as can be seen in the smooth period (13:00 to 17:00 UTC). This is contradictory to what one would expect and may be related to the rather large scatter observed in the atmospheric xCO2 readings of "A." Interestingly, the offset of the latter showed up from the beginning of the exercise (i.e., before June 9, 12:00 UTC when it suddenly appeared in seawater fCO2 readings). This is indicative of different reasons for the offsets observed inµatmospheric xCO2 and seawater fCO2 readings of "A." June 10, 1996 Missing data - "A": after 12:00 UTC, "C": before 13:00 UTC, "G": 12:00 to 23:00 UTC. Agreement to within ±2 µatm - "C," "D," and "E." Positive offset - "B": 3 to 9 µatm, "F": 4 to 10 µatm. Negative offset - "A": 5 µatm, "G": 5 µatm. Variable offset - "H": within ±2 µatm at 13:22 and 21:09 UTC, +7 µatm at 06:12 UTC. Comments - "B" immediately started to develop a positive offset which more or less remained until about 18:00 UTC of the following day. This offset is not seen inµatmospheric xCO2 measurements of "B." "F" also lost its good agreement and started to develop a positive offset which stabilized toward the end of the exercise. Interestingly these offsets of "B" and "F" show up at about the same time and with a very similar pattern over the 2-day period. Furthermore, the 06:12 UTC data point of "H" has the same positive offset as "B" and "F." Whether this is pure coincidence or an expression of something real is not known. These three systems, however, are very different in their principle of measurement and the location in the seawater supply line so that a common systematic error can be ruled out. Also their common offset seems to be inversely correlated with seawater temperature (see Fig. 7). On the basis of this observation it also has to be questioned whether a systematic offset may be present in the "reference" profiles "C," "D," and "E." This puzzle, however, cannot be solved here. June 11, 1996 Missing data - "A": before 05:30 UTC, "B": 17:00 to 19:00 UTC, "C": 12:00 to 16:00 UTC, "F": 20:30 to 23:30 UTC. Agreement to within ±2 µatm - "C," "D," "E," and "H." Positive offset - "F": 2 to 8 µatm, "G": 2 to 9 µatm. Negative offset - "A" 3 to 8 µatm. Variable offset - "B": within ±2 µatm (after 18:00 UTC), +3 to +9 µatm (00:00 to 17:00 UTC). Comment - Positive offsets of "B," "F," and "G" are essentially parallel throughout the day (see also comment of the previous day). Between 12:00 and 19:00 UTC the positive offset of "B" slowly disappears while at the same time a negative offset in theµatmospheric xCO2 measurements of "B" develops. June 12, 1996 Missing data - "C": before 16:00 UTC. Agreement to within ±2 µatm - "C," "D," and "E." Positive offset - "F": 3 to 9 µatm. Negative offset - "A": 3 to 6 µatm. Variable offset - "B": 0 to +3 µatm, "G" -2 to +8 µatm, "H": +1 to -6 µatm. Comments - The top figure shows nice parallel patterns of all fCO2 profiles even in this strongly variable environment. The bottom figure heavily suffers from the artificial scatter but nevertheless reveals the general offsets and trends. The negative offset of "H" at 17:29 UTC is likely due to this artifact because "H" is in very good agreement with "B" which does not show a general offset here. June 13, 1996 Missing data - "A": after 18:00 UTC, "B": after 22:00 UTC, "F": 16:00 to 20:00 UTC, "G": 03:00 to 05:00 UTC. Agreement to within ±2 µatm - "B," "C," "D," "E," and "G." Positive offset - "F": 6 to 9 µatm. Negative offset - "A": 5 to 6 µatm. Variable offset - "H": -2 and -5 µatm. Comment - This is about the beginning of the "smooth regime" with comparatively low variability in surface water which persisted for the rest of the exercise. The kind of agreement seen in this figure continues to exist in the following figures with very little alteration. In contrast to the highly variable situation encountered earlier this cruise, this situation is probably more representative of typical oceanic conditions in underway fCO2 field work. Again, the negative offset of "H" at 00:31 UTC is likely an artifact as it follows the profile of "B" which itself is in good agreement with "C," "D," and "E." System "A" had to quit the exercise at about 18:00 UTC because of a technical problem associated with the NDIR instrument, and no more data from this system are available beyond this point. June 14, 1996 Missing data - "A": no data available, "B": 17:00 to 22:00 UTC, "G": before 09:00 UTC. Agreement to within ±2 µatm - "B," "C," "D," "E," "G," and "H." Positive offset - "F": 7 to 10 µatm. Comment - Whereas the general agreement of all profiles except "F" is rather good, even among them slight trends toward positive ("H") or negative ("B" and "G") offsets can be identified that persist for the rest of the exercise. June 15, 1996 Missing data - "A": no data available. Agreement to within ±2 µatm - "C," "D," and "E." Positive offset - "F": 5 to 10 µatm. Negative offset - "B": 3 µatm, "G": 4 µatm. Variable offset - "H": +1.5 to +4.5 µatm. Comment - See comment for previous day. June 16, 1996 Missing data - "A": no data available, "F": 17:00 to 21:00 UTC. Agreement to within ±2 µatm - "C," "D," "E," and "H." Positive offset - "F": 5 to 8 µatm. Negative offset - "B": 2 to 3 µatm, "G": 2 to 4 µatm. Comment - As the hydrographic conditions have become much less variable, the overall picture of agreement among the various systems is very consistent for the last three days of the exercise. Overview The overall picture of agreement is characterized by a very good agreement of profiles "C," "D," and "E" essentially throughout the cruise. While also in good agreement for most of the time, profile "B" shows a 2-day period with a marked positive offset. Two profiles show a more or less constant sign of deviation, which is positive in the case of "F" and negative in the case of "A." The reason for this could not be identified easily. However, for system "A" we know of an instance of severe damage in the NDIR instrument toward the end of the exercise, which may well have started biasing the measurements in an early stage of the exercise. With respect to system "F," which is of a principally different design (see Sect. 3.2, Table 2), the question of whether the different principle of measurement could be the reason for the rather large observed offset should be addressed carefully. Finally, system "G" shows anything from large negative offsets over periods of good agreement to rather strong positive offsets. These problems were also apparent inµatmospheric xCO2 measurements and checks of the CO2 calibration performance probably because of an improper calibration technique. The calibration of the system appears to lack-at least during this exercise-the necessary reproducibility (i.e., it may be good in one case and bad in another one). This obvious problem of system "G" also needs careful checks. In addition to the daily figures (Figs. 12 to 20) representing the full data set, we present three figures in hard copy (Figs. 21 to 23) with enlarged views of shorter periods. These were chosen because they reveal more detail than is available in the daily figures. Furthermore they also cover the whole range of situations, from smooth to highly variable. Figure 21 shows a 3-hour period of measurements on June 9 that was characterized by very low variability in the surface seawater fCO2 (Fig. 13) as well as temperature and salinity (Fig. 7). The total change in fCO2 values during this period of time is about 6 µatm. This is uniformly seen in all profiles, which are almost perfectly parallel. Profiles "B," "D," "E," and "F" agree to within 1 µatm, while profiles "A" and "G" are characterized by a negative offset of about 8 µatm and 5 µatm, respectively. The scatter is smallest in profile "D" (averaging interval 5 min; large time constant, as shown) and highest in profile "A" (averaging interval 3 min; short time constant). The comparatively small scatter in profile "B" with 1-min averaging intervals shows that much of the scatter in profile "A" (also seen in theµatmospheric xCO2 data of "A") is not real and may thus indicate again the existence of a technical problem. In contrast, Fig. 22 shows the much more variable situation of a 3.5-hour period of measurements during June 12. The total range of fCO2 values covered during this period is about 35 µatm with gradients of up to 3 µatm/min-1. Again, the agreement is good in profiles "B," "D," and "E." Profiles "F" and "G" show positive offsets, while profile "A" has a negative offset of a few µatm. Individual time constants involved in the equilibration process going on in every system can be estimated rather precisely with step experiments carried out under well-defined conditions in a shore-based laboratory (Copin-Montegut 1988; Koertzinger et al. 1996b). This is definitely not the case in the present intercomparison exercise. We therefore do not try make any estimates of individual time constants. Nevertheless, in addition to the examination of offsets we do try to gain insight into the apparent time constants (i.e., we want to see whether there is any indication of differences in kinetic aspects of the equilibration processes). Because most time constants are on the order of a few minutes, this analysis is only feasible where fCO2 was measured at rather short intervals of (5 min (only profiles "A," "B," "C," and "D"), but even in these cases this is not a sound approach. We have marked approximate relative minima and maxima observed in the enlarged periods shown in Figs. 22 and 23. The pattern of vertical lines observed in these groups is highly consistent: Extrema always occur first and simultaneously in profiles "A" and "B," while profiles "C" and "D" lag behind by 5 to 8 min and 2 to 5 min, respectively. The range is mainly a consequence of the different averaging intervals. These time lags cannot be attributed to a temporal mismatch of the profiles (see Sect. 3.4.4). They are, however, clearly related to differences in the general design of these systems. Systems "A" and "B" are similar with respect to volumes and flow rates of water and air. For example, the total air volume of the equilibrator is exchanged every 2.5 min and 0.5 min, respectively, hence the similar equilibration times. In system "C" the large volume of air in the equilibrator is only exchanged every 20 min, which explains the more sluggish response seen in Fig. 23. System "D" is of the thin film type (i.e., unlike in the other system no turbulent mixing occurs in the equilibrator). It is known that this equilibration concept is characterized by somewhat larger time constants. We would like to point out that different time constants are no quality criterion per se but rather must be seen in the context of the application. A detailed process study would certainly require high spatial and temporal resolution and hence an fCO2 system with rather small time constants to resolve small-scale features. This is not equally the case in a basin-wide assessment of the fCO2 in surface seawater, where the large-scale averaging would eliminate the effect of different time constants. The main point here is simply to show that these different characteristics are clearly reflected in the fCO2 data set. 5. CONCLUSIONS We have provided a common infrastructure to all participating groups in this exercise. We also have carried out several checks to exclude possible sources of error. Furthermore all raw data were run through the same calculation procedure. All these measures were taken in order to reduce as much as possible controllable sources of error. In this respect the exercise was technically a full success, and as summarized in the following discussion we also think that the exercise was a success scientifically. We have demonstrated that the results of three out of seven underway systems agree to within about ±2 µatm throughout the cruise. This is not only the case for underway seawater fCO2 measurements but also for measurements ofµatmospheric xCO2. Interestingly, these three systems represent differences in such aspects as the design principle of the equilibrator, the volumes and flow rates of water and air involved, and the choice of wet or dry NDIR measurements. Thus this perfect agreement shows that-at least for NDIR instruments-the variety of designs used in the scientific community does not necessarily give rise to comparability problems or, to put it the other way round, systems of different design can produce reliable and consistent results. We have also demonstrated that significant offsets of up to 10 µatm can be found in underway fCO2 measurements under typical and identical field work conditions. Although in at least one case this may be a consequence of a technical failure, it is an indication of significant systematic differences in other cases. We certainly cannot claim that the observed differences are representative for these fCO2 systems in general. They may also be typical only for the specific conditions of this particular cruise. There is, however, no indication that this cruise provided in any way untypical circumstances that could be made responsible for some of the observed deviations. Finally we were able to demonstrate that discrete fCO2 measurements agree with the results of the three most consistent underway fCO2 systems. Therefore, measurements with these quite different approaches can be made with sufficient consistency, and the horizontal and vertical fCO2 profiles generated from these different techniques can be expected to match in surface waters. In conclusion, therefore, three main messages can be derived from this exercise: - Underway measurements of the CO2 fugacity in surface seawater and overlying air can be done to a high degree of agreement (±1 µatm) with a variety of possible equilibrator and system designs. - Even well-designed systems, which are operated without any obvious sign of malfunction, can show significant differences of up to 10 µatm. - The discrete fCO2 measurements are in good agreement with the three most consistent underway fCO2 data sets, at least to within its nominal accuracy of 1%. These results pose the important question of how fCO2 data sets acquired from different groups can be combined into a common database in light of possible incompatibilities of up to 10 µatm. Although the results of this exercise do not solve this problem, they underline the importance of this aspect which must be taken into account in the construction of a consistent global fCO2 database. Contributing to this dilemma is the fact that, in contrast to this exercise, other sources of error (temperature and pressure measurements, calibration gases etc.) further contribute to this uncertainty in field data. In addition to this more general outcome, some of the results in more concrete terms follow. These may also serve as recommendations for future fCO2 work in the ocean. - The exercise shows no "best choice" for the type of the equilibrator (i.e., "showerhead," "bubbler," or "thin film") nor specifics on its dimensions and flow rates of seawater and air in regard to the achievable accuracy of the fCO2 system. - In contrast, the equilibrator type and its flow rates of seawater and air are important aspects with respect to the time constant of the equilibration process. - Wet measurements can be done on the basis of the LI-6262 CO2/H2O gas analyzer (LI-COR Inc., U.S.A.) without necessary loss of accuracy when compared with traditional dry measurements. - The factory calibration of the LI-COR LI-6262 CO2/H2O gas analyzer, which only requires the user to adjust "zero" and "span" of the instrument, seems to result in a loss of accuracy, which can easily be avoided by establishing an individual calibration curve on the basis of measurements of standard gases. - The importance of rather accurate measurements of in situ and equilibrator temperature does not seem to be addressed adequately in the community. The observed differences between temperature measurements are clearly above a tolerable level and contribute-if representative and usually left unaccounted for-inconsistencies of several µatm (up to about 7 µatm in the present exercise). - Calibration gases are an important issue. Even with the provided suite of consistent calibration gases, the NDIR analyzers could only be calibrated to an accuracy of 0.5 to 1.0 ppmv. We feel that this is about the tolerable limit. So any further error contribution from the calibrated standard concentrations worsens the situation. Use of calibration gases that are traceable to the same primary standards, such as the WMO primary standards maintained at SIO, would be desirable. 6. DATA CHECKS AND PROCESSING PERFORMED BY CDIAC An important part of the numeric data package (NDP) process at the Carbon Dioxide Information Analysis Center (CDIAC) involves the quality assurance (QA) of data before distribution. Data received at CDIAC are rarely in a condition that would permit immediate distribution, regardless of the source. To guarantee data of the highest possible quality, CDIAC conducts extensive QA reviews that involve examining the data for completeness, reasonableness, and accuracy. Although they have common objectives, these reviews are tailored to each data set and often require extensive programming efforts. In short, the QA process is a critical component in the value-added concept of supplying accurate, usable data for researchers. The following information summarizes the data-processing and QA checks performed by CDIAC on the underway data obtained during the R/V Meteor Cruise 36/1 in the North Atlantic Ocean. 1. All data files were provided to CDIAC as 10 comma-separated files (9 for surface seawater underway measurements of fCO2 and 1 for all marine air xCO2 measurements) by Dr. Arne Koertzinger of IfMK. A FORTRAN 77 retrieval program was written and used to reformat the original files into uniform ASCII-formatted "water" and "air" data files. 2. All underway "water" data are presented as 9 daily data files. These files can be merged into a single data file by request. 3. All data were plotted to check for obvious outliers. 4. Dates and times were checked for bogus values (e.g., values of MONTH < or >6, DAY <8 or >16, YEAR < or >1996, TIME <0000 or >2400. 5. The cruise track was plotted using the coordinates presented in data files and compared with the maps and cruise information supplied by A. Koertzinger. 7. HOW TO OBTAIN THE DATA AND DOCUMENTATION This database (NDP-067) is available free of charge from CDIAC. The data are available from CDIAC's anonymous file transfer protocol (FTP) area via the Internet. Please note: your computer needs to have FTP software loaded on it (this is built into most newer operating systems). Commands used to obtain the database are >ftp cdiac.esd.ornl.gov or >ftp 128.219.24.36 Login: "anonymous" or "ftp" Password: YOU@your internet address Guest login ok, access restrictions apply. ftp> cd pub/ndp067/ ftp> dir ftp> mget (files) ftp> quit The complete documentation and data may also be obtained from CDIAC's Web site at the following URL: http://cdiac.esd.ornl.gov/oceans/doc.html For non-FTP data acquisitions (e.g., floppy diskette, 8-mm tape, CD-ROM, etc.), users may order through CDIAC's online ordering system (http://cdiac.esd.ornl.gov/pns/ how_order.html) or contact CDIAC directly to request the data and choice of media. For additional information, contact CDIAC. Address: Carbon Dioxide Information Analysis Center Oak Ridge National Laboratory P.O. Box 2008 Oak Ridge, Tennessee 37831-6335 U.S.A. Telephone: (423) 574-3645 (Voice) (423) 574-2232 (Fax) Electronic mail: cdiac@ornl.gov URL: http://cdiac.esd.ornl.gov/ 8. REFERENCES Copin-Montegut, C. 1988. A method for the continuous determination of the partial pressure of carbon dioxide in the upper ocean. Mar. Chem. 17:13-21. Copin-Montegut, C. 1989. A new formula for the effect of temperature on the partial pressure of CO2 in seawater. Corrigendum. Mar. Chem. 27:143-44. Gordon, L. I., and L. B. Jones. 1973. The effect of temperature on carbon dioxide partial pressure in seawater. Mar. Chem. 1:317-22. Goyet, C., F. J. Millero, A. Poisson, and D. K. Shafer. 1993. Temperature dependence of CO2 fugacity in seawater. Mar. Chem. 44:205-19. Goyet, C., and E. Peltzer. 1994. Comparison of the August-September 1991 and 1979 surface partical pressure of CO2 distribution in the Equatorial Pacific Ocean near 150 W. Mar. Chem. 45:257-66. Guggenheim, E. A. 1967. pp. 175-77 In: Thermodynamics, North-Holland, Amsterdam. Johnson, K. M., A. Koertzinger, L. Mintrop, J. C. Duinker, and D. W. R. Wallace. 1998. Coulometric total carbon dioxide analysis for marine studies: Measurement and internal consistency of underway surface TCO2 concentrations. Mar. Chem. (in press). Keeling, C. D., P. R. Guenther, and D. J. Moss. 1986. Scripps Reference Gas Calibration System for Carbon Dioxide-in-Air Standards: Revision of 1985. WMO Environmental Pollution Monitoring and Research Programme No. 42, Technical Document WMO/TD-125, World Meteorological Organization, Geneva, Switzerland. Koertzinger, A., L. Mintrop, and J. C. Duinker. 1996a. The floating dripstone cave: CO2 teams compare underway systems. U.S. JGOFS News 7(4):14-15. Koertzinger, A., H. Thomas, B. Schneider, N. Gronau, L. Mintrop, and J. C. Duinker. 1996b. At-sea intercomparison of two newly designed underway pCO2 systems: Encouraging results. Mar. Chem. 52:133-45. McDermitt, D. K., J. M. Welles, and R. D. Eckles. 1993. Effects of Temperature, Pressure, and Water Vapor on Gas Phase Infrared Absorption by CO2. Li-Cor Inc., Lincoln, Nebraska, U.S.A. Neill, C., K. M. Johnson, E. Lewis, and D. W. R. Wallace. 1997. Accurate headspace analysis of fCO2 in discrete water samples using batch equilibration. Limnol. Oceanogr. 42:1774-83. Ohtaki, E., E. Yamashita, and F. Fujiwara. 1993. Carbon dioxide in surface seawaters of the Seto Inland Sea, Japan. J. Oceanogr. 49:295-304. Poisson, A., N. Metzl, C. Brunet, B. Schauer, B. Bres, D. Ruiz-Pino, and F. Louanchi. 1995. Variability of sources and sinks of CO2 in the Western Indian and Southern Oceans during the year 1991. J. Geophys. Res. 98(C12):22759-778. Robertson, J. E., and A. J. Watson. 1992. Thermal skin effect of the surface ocean and its implications for CO2 uptake. Nature 358:738-40. Schluessel, P., W. J. Emery, H. Grassl, and T. Mammen. 1990. On the bulk-skin temperature difference and its impact on satellite remote sensing of sea surface temperature. J. Geophys. Res. 95:13341-356. Takahashi, T., J. Olafsson, J. G. Goddard, D. W. Chipman, and S. C. Sutherland. 1993. Seasonal variations of CO2 and nutrient salts in the high latitude oceans: A comparative study. Global Biogeochem. Cycles 7:843-848. Thoning, K. W., P. Tans, T. J. Conway, and L. S. Waterman. 1987. NOAA/GMCC calibrations of CO2-in-air reference gases: 1979-1985, NOAA Technical Memorandum ERL/ARL-150 Environmental Research Laboratory, Boulder, Colorado. Welles, J. M., and R. D. Eckles. 1991. Li-Cor 6262 CO2/H2O Analyzer Operating and Service Manual. Publication No. 9003-59. Li-COR Inc., Lincoln, Nebraska, U.S.A. Weiss, R. F. 1974. Carbon dioxide in water and seawater: The solubility of a non-ideal gas. Mar. Chem. 2:203-15. Weiss, R. F., and B. A. Price. 1980. Nitrous oxide solubility in water and seawater. Mar. Chem. 8:347-59. Weiss, R. F., R. A. Jahnke, and C. D. Keeling. 1982. Seasonal effects of temperature and salinity on the partial pressure of CO2 in seawater. Nature 300:511-13. Zhao, C. L., P. P. Tans, and K. W. Thoning. 1997. A high precision manometric system for absolute calibrations of CO2 in dry air. J. Geophys. Res. 102:5885-94. PART 2 CONTENT AND FORMAT OF DATA FILES 9. FILE DESCRIPTIONS This section describes the content and format of each of the 13 files that make up this NDP (see Table 4). Because CDIAC distributes the data set in several ways (e.g., via anonymous FTP and on floppy diskette), each of the 13 files is referenced by both a file number and an ASCII file name, which is given in lowercase, bold faced type (e.g., ndp067.doc). The remainder of this section describes (or lists, where appropriate) the contents of each file. Table 4. Content, size, and format of data files -------------------------------------------------------------------------------- File number, name, and description | Logical records | File size in bytes -------------------------------------------------------------------------------- 1. ndp067.txt: 3,059 122,515 A detailed description of the cruise network, the two FORTRAN 77 data-retrieval routines, and the ten oceanographic data files 2. fco2wat.for: 54 2,005 A FORTRAN 77 data-retrieval routine to read and print all *w.txt files (Files 4-12) 3. xco2air.for: 42 1,264 A FORTRAN 77 data-retrieval routine to read and print xco2air.txt (File 13) 4. 080696w.txt: 1,444 218,336 Underway measurements of surface seawater fCO2 and hydrographic parameters during 8 June 1996 5. 090696w.txt: 1,442 218,032 Underway measurements of surface seawater fCO2 and hydrographic parameters during 9 June 1996 6. 100696w.txt: 1,442 218,032 Underway measurements of surface seawater fCO2 and hydrographic parameters during 10 June 1996 7. 110696w.txt: 1,442 218,183 Underway measurements of surface seawater fCO2 and hydrographic parameters during 11 June 1996 8. 120696w.txt: 1,442 218,032 Underway measurements of surface seawater fCO2 and hydrographic parameters during 12 June 1996 9. 130696w.txt: 1,442 218,032 Underway measurements of surface seawater fCO2 and hydrographic parameters during 13 June 1996 10. 140696w.txt: 1,443 218,184 Underway measurements of surface seawater fCO2 and hydrographic parameters during 14 June 1996 11. 150696w.txt: 1,443 218,184 Underway measurements of surface seawater fCO2 and hydrographic parameters during 15 June 1996 12. 160696w.txt: 1,444 218,336 Underway measurements of surface seawater fCO2 and hydrographic parameters during 16 June 1996 13. xco2air.txt: 1,744 104,604 Underway measurements of air xCO2 during the entire expedition ---------------------------------------------------------------------------- Total 17,883 2,193,739 ---------------------------------------------------------------------------- 9.1 ndp067.txt (FILE 1) This file contains a detailed description of the data set, the two FORTRAN 77 data-retrieval routines, and the ten oceanographic data files. It exists primarily for the benefit of individuals who acquire this database as machine-readable data files from CDIAC. 9.2 fco2wat.for (FILE 2) This file contains a FORTRAN 77 data-retrieval routine to read and print all *w.txt files. The following is a listing of this program. For additional information regarding variable definitions, variable lengths, variable types, units, and codes, please see the description for *w.txt files in Sect. 9.4. c**************************************************************** c* FORTRAN 77 data retrieval routine to read and print the files* c* named "*w.txt" (Files 4-12) * c**************************************************************** c*Defines variables* CHARACTER date*10, time*8 INTEGER course REAL latdcm, londcm, speed, temp, salt, press, fco2a, fco2b REAL fco2c, fco2d, fco2e, fco2f, fco2g, fco2h OPEN (unit=1, file='input.txt') OPEN (unit=2, file='output.txt') write (2, 5) c*Writes out column labels* 5 format (5X,'DATE',9X,'TIME',5X,'LATITUDE',3X,'LONGITUDE', 1 3X,'SPEED',1X,'COURSE',4X,'TEMP',2X,'SALNTY',4X, 2 'PRESS',4X,8('fCO2',4X),/,5X,'GMT',10X,'GMT',8X, 3 'DEG N',6X,'DEG E',7X,'KN',3X,'DEG',5X,'DEG C', 4 2X,'PSS-78',5X,'hPA',5X,8('µatm',4X),/, 5 2X,'DD.MM.YYYY',4X,'HH:MM:SS',66X,'LAB A',3X, 6 'LAB B',3X,'LAB C',3X,'LAB D',3X,'LAB E',3X, 7 'LAB F',3X,'LAB G',3X,'LAB H') c*Sets up a loop to read and format all the data in the file* read (1, 6) 6 format (////////////) 7 CONTINUE read (1, 10, end=999) date, time, latdcm, londcm, speed, 1 course, temp, salt, press, fco2a, fco2b, fco2c, fco2d, 2 fco2e, fco2f, fco2g, fco2h 10 format (2X, A10, 4X, A8, 4X, F7.4, 4X, F8.4, 4X, F4.1, 1 2X, I3, 5X, F5.2, 3X, F5.2, 3X, F6.1, 1X, F7.2, 1X, F7.2, 2 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2) write (2, 20) date, time, latdcm, londcm, speed, 1 course, temp, salt, press, fco2a, fco2b, fco2c, fco2d, 2 fco2e, fco2f, fco2g, fco2h 20 format (2X, A10, 4X, A8, 4X, F7.4, 4X, F8.4, 4X, F4.1, 1 2X, I3, 5X, F5.2, 3X, F5.2, 3X, F6.1, 1X, F7.2, 1X, F7.2, 2 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2) GOTO 7 999 close(unit=1) close(unit=2) stop end 9.3 xco2air.for (FILE 3) This file contains a FORTRAN 77 data-retrieval routine to read and print xco2air.txt file. The following is a listing of this program. For additional information regarding variable definitions, variable lengths, variable types, units, and codes, please see the description for xco2air.txt file in Sect. 9.5. c**************************************************************** c* FORTRAN 77 data retrieval routine to read and print the files* c* named "xco2air.txt" (File 13) * c**************************************************************** c*Defines variables* CHARACTER date*10, time*8, lab*1 REAL latdcm, londcm, xco2 OPEN (unit=1, file='xco2air.txt') OPEN (unit=2, file='xco2air.dat') write (2, 5) c*Writes out column labels* 5 format (5X,'DATE',9X,'TIME',4X,'LATITUDE',2X,'LONGITUDE', 1 2X,'xCO2_AIR',1X,'LAB',/,5X,'GMT',10X,'GMT',7X, 3 'DEG N',6X,'DEG E',5X,'PPMV',4X,'NO',/, 5 2X,'DD.MM.YYYY',4X,'HH:MM:SS') c*Sets up a loop to read and format all the data in the file* read (1, 6) 6 format (////////////) 7 CONTINUE read (1, 10, end=999) date, time, latdcm, londcm, xco2, lab 10 format (2X, A10, 4X, A8, 4X, F6.3, 4X, F7.3, 4X, F6.2, 1 3X, A1) write (2, 20) date, time, latdcm, londcm, xco2, lab 20 format (2X, A10, 4X, A8, 4X, F6.3, 4X, F7.3, 4X, F6.2, 1 3X, A1) GOTO 7 999 close(unit=1) close(unit=2) stop end 9.4 *w.txt (FILES 4-12) These 9 data files contain the underway measurements of surface seawater fCO2 and hydrographic parameters made by participants in the systems intercomparison exercise during the R/V Meteor Cruise 36/1 in the North Atlantic Ocean. All files have the same ASCII format and can be read by using the following FORTRAN 77 code [contained in fco2wat.for (File 2)]: CHARACTER date*10, time*8 INTEGER course REAL latdcm, londcm, speed, temp, salt, press, fco2a, fco2b REAL fco2c, fco2d, fco2e, fco2f, fco2g, fco2h read (1, 10, end=999) date, time, latdcm, londcm, speed, 1 course, temp, salt, press, fco2a, fco2b, fco2c, fco2d, 2 fco2e, fco2f, fco2g, fco2h 10 format (2X, A10, 4X, A8, 4X, F7.4, 4X, F8.4, 4X, F4.1, 1 2X, I3, 5X, F5.2, 3X, F5.2, 3X, F6.1, 1X, F7.2, 1X, F7.2, 2 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2, 1X, F7.2) Stated in tabular form, the contents include the following: ---------------------------------------------- Variable Variable Variable Starting Ending type width column column ---------------------------------------------- date Character 10 3 12 time Character 8 17 24 latdcm Numeric 7 29 35 londcm Numeric 8 40 47 speed Numeric 4 52 55 course Numeric 3 58 60 temp Numeric 5 66 70 salt Numeric 5 74 78 press Numeric 6 82 87 fco2a Numeric 7 89 95 fco2b Numeric 7 97 103 fco2c Numeric 7 105 111 fco2d Numeric 7 113 119 fco2e Numeric 7 121 127 fco2f Numeric 7 129 135 fco2g Numeric 7 137 143 fco2h Numeric 7 145 151 --------------------------------------------- The variables are defined as follows: date is the sampling date (day/month/year); time is the sampling time [Greenwich mean time (GMT)]; latdcm is the latitude of the sampling location (decimal degrees; negative values indicate the Southern Hemisphere); londcm is the longitude of the sampling location (decimal degrees; negative values indicate the Western Hemisphere); speed is the speed of the ship during the measurements (kn); course is the course of the ship during the measurements (degrees); temp is the sea-surface temperature ((C); salt is the sea-surface salinity [on the Practical Salinity Scale (PSS)]; press is theµatmospheric pressure (hPA); fco2a is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory A; fco2b is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory B; fco2c is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory C; fco2d is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory D; fco2e is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory E; fco2f is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory F; fco2g is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory G; fco2h is the underway fugacity of CO2 in surface seawater (µatm) measured by the laboratory H. 9.5 xco2air.txt (FILE 13) This data file contains the underway measurements ofµatmospheric xCO2 made during R/V Meteor Cruise 36/1 in the North Atlantic Ocean. The data are presented in ASCII format and can be read by using the following FORTRAN 77 code [contained in xco2air.for (File 3)] in Sect. 9.3: CHARACTER date*10, time*8, lab*1 REAL latdcm, longdcm, xco2 read (1, 10, end=999) date, time, latdcm, londcm, xco2, lab 10 format (2X, A10, 4X, A8, 4X, F6.3, 4X, F7.3, 4X, F6.2, 1 3X, A1) Stated in tabular form, the contents include the following: ----------------------------------------------- Variable Variable Variable Starting Ending type width column column ----------------------------------------------- date Character 10 3 12 time Character 8 17 24 latdcm Numeric 6 29 34 londcm Numeric 7 39 45 xco2 Numeric 6 50 55 lab Character 1 59 59 ----------------------------------------------- The variables are defined as follows: date is the sampling date (day/month/year); time is the sampling time (GMT); latdcm is the latitude of the sampling location (decimal degrees; negative values indicate the Southern Hemisphere); londcm is the longitude of the sampling location (decimal degrees; negative values indicate the Western Hemisphere); xco2 is the mole fraction ofµatmospheric CO2 (ppmv) measured in dry air; lab is the laboratory identifier.