CArbon VAriability Studies by Ships Of Opportunity

Transcription

1 1 CArbon VAriability Studies by Ships Of Opportunity (CAVASSOO) Contract: EVK FIRST ANNUAL REPORT Editor: Nathalie Lefèvre TABLE OF CONTENTS SUMMARY...2 Highlights...2 Objectives...2 Context of the work...3 Outline of methodology...4 Progress to date...4 Preliminary conclusions...5 EXECUTIVE SUMMARY...5 INTRODUCTION...6 Objectives...6 Context...8 Brief presentation of the consortium...9 WORK ACHIEVED TO DATE...9 Overview...9 Milestones reached...9 Progress by work package...10 Results...15 Collection of new data...15 Database and time-series stations...17 OCMIP models-data compared with seasonal climatology over the North Atlantic...20 Inverse modelling to quantify the North Atlantic carbon balance...24 PRELIMINARY CONCLUSIONS...26 Progress towards project objectives...26 Implications for next reporting period...27 REFERENCES...27 ANNEXES...27 Specific deliverables...27 Reports on workshops...33 Published material...36 Publicity material...36 Presentation of project web page...43

2 2 SUMMARY Highlights Autonomous pco 2 system installed on the R/V Hesperides, on the MV Falstaff, on the MV Santa Lucia New instrumentation has been rigorously and successfully tested at sea and will be installed on board the R/V Nuka Arctica in early December First CO 2 data of the network collected along the meridional transect of the R/V Hesperides New pco 2 data collected from the northern North Atlantic Additional CO 2 data from a transatlantic research cruise along 48 o N Historical CO 2 data gathered in a relational database Key time-series stations identified in the North Atlantic Ocean 10 different ocean carbon model compared with a CO 2 climatology North Atlantic uptake estimated at 0.7 ± 0.26/0.58 GtC yr -1 from 27 atmospheric inversions Objectives The overall objective is to provide reliable estimates of the uptake of CO 2 by the North Atlantic, and how this varies from season to season and year to year. These will in turn assist in constraining estimates of European and North American terrestrial (vegetation) sinks, using atmospheric inverse modelling techniques. The overall objective is broken into subsidiary objectives below: Objective 1: Establish a basic North Atlantic surface pco 2 observing system The first year was devoted to the set-up of the surface pco 2 network. Several routes covering the major physical and biogeochemical provinces of the North Atlantic have been identified. The routes and the ships of the network are: Spain to Antarctica by the R/V Hesperides Denmark to Greenland by the R/V Nuka Arctica Goteborg to Charleston by the M/V Falstaff UK to Windward Islands (Caribbean) by the MV Santa Lucia This network will resolve, for the first time across the whole region, the annual seasonal cycle of CO 2 and its interannual variability. Some of the ships (Nuka Arctica, Hesperides) already record hydrographic properties along the route so that complementary measurements (e.g. upper ocean thermal structure) will also be obtained. The implementation of the CO 2 network in the North Atlantic lays the foundation for long-term monitoring of CO 2 uptake in the region to enable CO 2 fluxes to be directly assessed year-by year. Objective 2: Produce and make available a North Atlantic pco 2 database Extensive data for pco 2 and associated variables already exists for the North Atlantic, but it has been collected sporadically over the past 30 years. Some of the data are publicly available, for example at CDIAC (Carbon dioxide Information and Analysis Center, Oak Ridge National Laboratory, USA). Much however is only available from individual investigators around the world. A CO 2 database for the North Atlantic will be constructed with the public data and as much non-public data as these investigators will allow, and will be used to construct new historical climatologies under various assumptions. The database will be made available initially to project members and those others who have contributed to it. Near the end of the three year project, with our new measurements included, it will be made available via a subcontract to the UK data centre BODC. The data will be used to test

3 with independent data the assumptions made in combining many years of data into climatologies for a single year. These assumptions very significantly affect the results obtained for the climatologies. Objective 3: Assess errors and uncertainties in existing pco 2 climatologies Climatologies derived from the existing database will be compared directly with our new measurements, and the assumptions made in deriving those climatologies tested. Information on pco 2 variability will form the basis for long-term optimization of a carbon observing system in the North Atlantic. The new data will be used to assess the CO 2 variability at different time and space scales, and to investigate the causes of that variability. Objective 4: Estimate seasonal air-sea CO 2 fluxes for the North Atlantic Updated quarterly gridded (1 o latitude by 1 o longitude) pco 2 fields will be obtained by independent methods, for example (a) Direct from the new data, using SST and a surface advection-diffusion scheme for the region to interpolate the data. (b) By "stretching" the existing pco 2 climatological fields to fit the new data using correlation length scales derived from the pco 2 data. The uncertainty in the fields will be estimated statistically within each method, and by comparison of different methods. Best estimates of the pco 2 fields will be made, and seasonal estimates of the CO 2 flux derived using a modern parameterization of the gas exchange rate. Objective 5: Evaluate Ocean carbon models with regard to air-sea flux variability Observations will be compared with the output of an ocean carbon model. The international project OCMIP (Ocean Carbon Model Intercomparison Project) already provides results of 13 models which can be used for models-data comparisons. One ocean carbon model will be used in the project to provide information on inter-annual variability of the fluxes, as controlled by individual processes (e.g. ocean circulation, marine biology, solubility). Objective 6: Improve atmospheric inversion estimates of carbon sources and sinks The better-constrained North Atlantic carbon fluxes calculated in objective 4 will be used in atmospheric inversion models to improve the estimates of the European and North American fluxes. We will also use the ships of opportunity to take selected atmospheric measurements for CO 2 and isotopes. The lack of samples for atmospheric concentrations over the North Atlantic is an important source of error in the inversions. These new atmospheric measurements, as well as the oceanographic data collected, will be integrated within a global atmospheric transport model to calculate a "Grand Inversion" of North Atlantic fluxes. Uncertainties linked to the inverse technique, to sparse and uneven atmospheric sampling and to the a priori information introduced in the model will be quantified. 3 Context of the work The objectives of the project are particularly relevant to international policy in relation to the Kyoto agreement. Member states of the EU are committed to jointly decreasing their net CO 2 emissions to 92% of 1990 levels by the years However, this political imperative is at present running well ahead of the science required to deliver it. In order to monitor changes in both land and ocean CO 2 sinks, and verify the efficiency of policies aimed at sequestering carbon, it will be necessary to have in place a means of observing the changes in CO 2 sources and sinks. The atmospheric inverse technique may be able to provide this, but not without an improved, independent quantification of the North Atlantic sink. Atmospheric inverse modelling is reasonably robust when estimating net CO 2 sinks as a function of latitude. However, because transport in the atmosphere is much faster in an east-

4 west direction than a north-south direction, the technique is much less successful at quantifying sinks in the same latitude bands. Thus at present the European, North Atlantic and North American sinks are subject to large errors when determined by inversions. Reducing uncertainty over one region, in our case the North Atlantic, directly reduces uncertainty in adjacent regions. Accurate independent estimates of the North Atlantic net sink will greatly improve the diagnostic of carbon fluxes in the Northern Hemisphere. Our project will thus contribute a critical missing element to a global carbon dioxide observing strategy, now beginning to be discussed at national and international level (for example within the IGBP). This strategy will aim to enable sources and sinks of CO 2 in the Northern hemisphere to be assigned to land or ocean, and the land sinks to be better divided between North America and Europe/Asia. It also promotes the development of systematic observation systems and data archives to reduce uncertainties related to the climate system. The project sets up the foundation of a European network, which will provide an observing system to monitor, on longer time scales, changes of the oceanic uptake of greenhouse gases. The work planned here is directly linked to the environmental policies arising from the Kyoto protocol and post-kyoto process, and will inform EU governments on how best to preserve the global environment. CAVASSOO will also complement the European project AEROCARB, dealing with terrestrial CO 2 fluxes over Europe, by quantifying CO 2 fluxes in the North Atlantic Ocean. A link between the programs is planned within WP 5 of CAVASSOO. The European project NOCES dealing with carbon ocean modelling includes a link with CAVASSOO for modeldata comparisons. The European project ANIMATE plans CO 2 measurements in the North Atlantic Ocean and the data should be made available to CAVASSOO members. A UK consortium bid has been recently approved for funding and more CO 2 measurements will be made on a meridional transect from UK to Antarctica. In the context of international CLIVAR CO 2 data will be collected in June-July 2002 between Portugal and Greenland. 4 Outline of methodology The methodology consists of using three main tools: an observing network, a database and ocean and atmospheric carbon models. The project involves the building of a CO 2 network in the North Atlantic Ocean. Four ships have been equipped with an autonomous pco 2 system to collect data across the different biogeochemical provinces of the North Atlantic Ocean, on a monthly basis for most of the ships. Atmospheric and oceanic pco 2 are measured along the track using a nondispersive infrared detector. All shipping companies involved in the network have agreed to allow someone on board when necessary. Additional data (nutrients, chlorophyll) are then collected when a scientist sails with the ship. The amount of data collected by this network is expected to be large and will complement existing CO 2 measurements. A relational database has been developed to easily extract the relevant information from a huge dataset. The data will provide information on the seasonal variability of the air-sea CO 2 flux and a better estimate of the mean flux in the North Atlantic Ocean, which will be used to validate ocean carbon model and better constraint atmospheric inversion models. Progress to date The progress is good and the project has been developed according to the description of work. All milestones have been reached or will be reached within a month of their anticipated

5 delivery date. There have been some delays to establish the network attributable to ship breakdowns and changes of routes. However the network is expected to be fully operational within the next few months as all instruments have been built and problems with ships solved. The database has been created and populated with existing data and key time-series stations have been identified. Modelling runs have been performed according to the project plan. Preliminary conclusions The pco 2 instruments are built and after negotiations with shipping companies they have been installed on board ships sailing along the routes proposed in the project. The data collection has started on board some of the ships. New data from the network have been recorded along the meridional transect on board the R/V Hesperides in October-November 2000, March-April 2001 and October-November Regular data from the other ships are expected within the next few months. Links with other project and opportunistic research cruises have provided and will provide additional data. The database currently includes 160 cruises and more than 450,000 seawater pco 2 data measured underway. Time-series stations have been identified mainly along 20 o W in the different biogeochemical provinces (North Atlantic gyre, tropics, subarctic Atlantic). CO 2 climatologies assume that oceanic pco 2 follows the atmospheric increase except in cold regions (i.e. North Atlantic >45 o N) where oceanic pco 2 reflects the characteristics of the deep water because of the convective mixing. In these regions oceanic pco 2 does not increase like atmospheric pco 2 or increases at a lower rate. The evolution of pco 2 (difference between seawater pco 2 and atmospheric pco 2 ) will be studied at stations located at 60 o N, 20 o W, 50 o N, 20 o W, 25 o N, 65 o W, 40 o N, 20 o W, and 10 o N, 22 o W. 10 different ocean carbon models have been compared with CO 2 climatology. Overall the models tend to capture the mean north-south pattern of the observed annual mean air-sea CO 2 flux but the longitudinal variability is overestimated by many of the models. A quantitative assessment of overall model performance has been performed using commonly used summary statistics. Atmospheric inversions with different settings have been performed using 46 atmospheric sites available from the NOAA/CMDL network. Three different atmospheric transport models with varying spatial resolution have been used. The North Atlantic uptake has been estimated at 0.7 ± 0.26/0.58 GtC yr -1 from 27 atmospheric inversions. The inverse estimate gives a larger sink than the one given by CO 2 climatology and the oceanic/atmospheric measurements from CAVASSOO would help to resolve the disagreement with inverse results. 5 EXECUTIVE SUMMARY The North Atlantic Ocean is a strong sink of CO 2 for the atmosphere and that sink may be expected to vary seasonally and over time. We have climatologies of pco 2 (the chemical gradient driving the air-sea flux) for the North Atlantic, resolved seasonally but some of the assumptions used in combining the data have yet to be verified. If we knew with precision the flux of CO 2 into the North Atlantic and its variability, the uncertainties in calculating with models the size of the natural land sinks in Europe and North America would be substantially reduced. CAVASSOO aims at providing reliable estimates of the uptake of CO 2 by the North Atlantic and how this uptake varies from season to season and year to year. These will in turn

6 assist in constraining estimates of European and North American terrestrial (vegetation) sinks using atmospheric inverse modelling techniques. To obtain whole-basin estimates of CO 2 we set up an observing network using ships of opportunity equipped with automatic and unattended systems to record the partial pressure of CO 2 (pco 2 ) underway. The network consists of the following routes: Spain to Antarctica by the R/V Hesperides Denmark to Greenland by the R/V Nuka Arctica Goteborg to Charleston by the M/V Falstaff UK to Windward Islands (Caribbean) by the MV Santa Lucia The network will resolve for the first time across the North Atlantic basin the seasonal cycle of CO 2 and its year to year variability and will lay the foundation for long-term monitoring of CO 2 uptake in this region. The large amount of data collected by these ships will be gathered in a central database that has been built as part of the project. The database includes existing data from previous cruises made in the North Atlantic. Locations where we have several years of measurements have been identified in different provinces of the North Atlantic Ocean. At these locations we will be able to test if the oceanic pco 2 follows the atmospheric CO 2 increase as previously assumed for regions <45 o N. A better quantification of the oceanic pco 2 increase should be achieved. The database and its new data will also provide refined CO 2 climatologies. Ocean carbon models have been evaluated based on their ability to simulate the seasonal climatology of air-sea CO 2 fluxes in the North Atlantic Ocean. We used output from ten different models, standard simulations made during the second phase of the EC funded Ocean Carbon-Cycle Model Intercomparison Project (OCMIP, ). Overall, the models tend to capture the mean north-south pattern of the observed annual mean sea-air CO 2 flux, with large CO 2 uptake in the high latitudes and CO 2 release in the tropics. Yet, longitudinal variability is overestimated by many of the models. The best performing model is often the mean of all OCMIP models. Atmospheric samples will be taken on board the ships of the network. The lack of samples for atmospheric concentrations over the North Atlantic is an important source of error in atmospheric inversions. The principle of atmospheric inversions is to infer surface fluxes from atmospheric measurements, via the use of transport model. In such an exercise, an optimal set of fluxes is sought, which minimises a distance between the transport model results and the observations. Recent inverse studies have given conflicting results for the mean carbon balance, but it is difficult to analyse the causes of such discrepancies because published inversions are all performed with different settings. Atmospheric inversions with different settings have been performed and the North Atlantic uptake has been estimated at 0.7 ± 0.26/0.58 GtC yr -1 from 27 atmospheric inversions. The better-constrained North Atlantic carbon fluxes calculated from the new data will be used to improve the estimates of the European and North American fluxes estimated from atmospheric inversion models. 6 INTRODUCTION Objectives The overall objective is to provide reliable estimates of the uptake of CO 2 by the North Atlantic, and how this varies from season to season and year to year. These will in turn assist in constraining estimates of European and North American terrestrial (vegetation) sinks,

7 using atmospheric inverse modelling techniques. The overall objective is broken into subsidiary objectives below: Objective 1: Establish a basic North Atlantic surface pco 2 observing system The first year was devoted to the set-up of the surface pco 2 network. Several routes covering the major physical and biogeochemical provinces of the North Atlantic have been identified. The routes and the ships of the network are: Spain to Antarctica by the R/V Hesperides Denmark to Greenland by the R/V Nuka Arctica Goteborg to Charleston by the M/V Falstaff UK to Windward Islands (Caribbean) by the MV Santa Lucia This network will resolve, for the first time across the whole region, the annual seasonal cycle of CO 2 and its interannual variability. Some of the ships (Nuka Arctica, Hesperides) already record hydrographic properties along the route so that complementary measurements (e.g. upper ocean thermal structure) will also be obtained. The implementation of the CO 2 network in the North Atlantic lays the foundation for long-term monitoring of CO 2 uptake in the region to enable CO 2 fluxes to be directly assessed year-by year. Objective 2: Produce and make available a North Atlantic pco 2 database Extensive data for pco 2 and associated variables already exists for the North Atlantic, but it has been collected sporadically over the past 30 years. Some of the data are publicly available, for example at CDIAC (Carbon dioxide Information and Analysis Center, Oak Ridge National Laboratory, USA). Much however is only available from individual investigators around the world. A CO 2 database for the North Atlantic will be constructed with the public data and as much non-public data as these investigators will allow, and will be used to construct new historical climatologies under various assumptions. The database will be made available initially to project members and those others who have contributed to it. Near the end of the three year project, with our new measurements included, it will be made available via a subcontract to the UK data centre BODC. The data will be used to test with independent data the assumptions made in combining many years of data into climatologies for a single year. These assumptions very significantly affect the results obtained for the climatologies. Objective 3: Assess errors and uncertainties in existing pco 2 climatologies Climatologies derived from the existing database will be compared directly with our new measurements, and the assumptions made in deriving those climatologies tested. Information on pco 2 variability will form the basis for long-term optimization of a carbon observing system in the North Atlantic. The new data will be used to assess the CO 2 variability at different time and space scales, and to investigate the causes of that variability. Objective 4: Estimate seasonal air-sea CO 2 fluxes for the North Atlantic Updated quarterly gridded (1 o latitude by 1 o longitude) pco 2 fields will be obtained by independent methods, for example (a) Direct from the new data, using SST and a surface advection-diffusion scheme for the region to interpolate the data. (b) By "stretching" the existing pco 2 climatological fields to fit the new data using correlation length scales derived from the pco 2 data. The uncertainty in the fields will be estimated statistically within each method, and by comparison of different methods. Best estimates of the pco 2 fields will be made, and seasonal estimates of the CO 2 flux derived using a modern parameterization of the gas exchange rate. Objective 5: Evaluate Ocean carbon models with regard to air-sea flux variability 7

8 Observations will be compared with the output of an ocean carbon model. The international project OCMIP (Ocean Carbon Model Intercomparison Project) already provides results of 13 models which can be used for models-data comparisons. One ocean carbon model will be used in the project to provide information on inter-annual variability of the fluxes, as controlled by individual processes (e.g. ocean circulation, marine biology, solubility). Objective 6: Improve atmospheric inversion estimates of carbon sources and sinks The better-constrained North Atlantic carbon fluxes calculated in objective 4 will be used in atmospheric inversion models to improve the estimates of the European and North American fluxes. We will also use the ships of opportunity to take selected atmospheric measurements for CO 2 and isotopes. The lack of samples for atmospheric concentrations over the North Atlantic is an important source of error in the inversions. These new atmospheric measurements, as well as the oceanographic data collected, will be integrated within a global atmospheric transport model to calculate a "Grand Inversion" of North Atlantic fluxes. Uncertainties linked to the inverse technique, to sparse and uneven atmospheric sampling and to the a priori information introduced in the model will be quantified. 8 Context The objectives of the project are particularly relevant to international policy in relation to the Kyoto agreement. Member states of the EU are committed to jointly decreasing their net CO 2 emissions to 92% of 1990 levels by the years However, this political imperative is at present running well ahead of the science required to deliver it. In order to monitor changes in both land and ocean CO 2 sinks, and verify the efficiency of policies aimed at sequestering carbon, it will be necessary to have in place a means of observing the changes in CO 2 sources and sinks. The atmospheric inverse technique may be able to provide this, but not without an improved, independent quantification of the North Atlantic sink. Atmospheric inverse modelling is reasonably robust when estimating net CO 2 sinks as a function of latitude. However, because transport in the atmosphere is much faster in an eastwest direction than a north-south direction, the technique is much less successful at quantifying sinks in the same latitude bands. Thus at present the European, North Atlantic and North American sinks are subject to large errors when determined by inversions. Reducing uncertainty over one region, in our case the North Atlantic, directly reduces uncertainty in adjacent regions. Accurate independent estimates of the North Atlantic net sink will greatly improve the diagnostic of carbon fluxes in the Northern Hemisphere. Our project will thus contribute a critical missing element to a global carbon dioxide observing strategy, now beginning to be discussed at national and international level (for example within the IGBP). This strategy will aim to enable sources and sinks of CO 2 in the Northern hemisphere to be assigned to land or ocean, and the land sinks to be better divided between North America and Europe/Asia. It also promotes the development of systematic observation systems and data archives to reduce uncertainties related to the climate system. The project sets up the foundation of a European network, which will provide an observing system to monitor, on longer time scales, changes of the oceanic uptake of greenhouse gases. The work planned here is directly linked to the environmental policies arising from the Kyoto protocol and post-kyoto process, and will inform EU governments on how best to preserve the global environment.

9 CAVASSOO will also complement the European project AEROCARB, dealing with terrestrial CO 2 fluxes over Europe, by quantifying CO 2 fluxes in the North Atlantic Ocean. A link between the programs is planned within WP 5 of CAVASSOO. The European project NOCES dealing with carbon ocean modelling includes a link with CAVASSOO for modeldata comparisons. The European project ANIMATE plans CO 2 measurements in the North Atlantic Ocean and the data should be made available to CAVASSOO members. A UK consortium bid has been recently approved for funding and more CO 2 measurements will be made on a meridional transect from UK to Antarctica. In the context of international CLIVAR CO 2 data will be collected in June-July 2002 between Portugal and Greenland. 9 Brief presentation of the consortium The consortium includes groups with strong experience in CO 2 measurements, ocean carbon modelling and atmospheric modelling. Partners 1, 3, 4 and 5 are responsible for one line of the pco 2 network. Partner 2 focuses on the modelling aspect of the project. Institute Summary of role in the project 1 UEA, UK Coordinator Construct CO 2 database UK-Caribbean line CO 2 climatology and variability 2 CEA-DSM, France Ocean and atmospheric carbon modelling High precision atmospheric sampling 3 CSIC, Spain Meridional CO 2 line CO 2 variability 4 IFM Kiel, Germany Coordinate pco 2 network Germany-USA line CO 2 variability 5 UIB, Norway Nuka-Arctica CO 2 line CO 2 variability WORK ACHIEVED TO DATE Overview The partners involved in the set up of the network have established the collaboration with the shipping companies. The pco 2 instruments are built, have been tested in the laboratory and have been installed on board the ships. For some of them the data collection has already started. The relational database has been constructed with existing data and the first data of the network have been included. This allowed the identification of time-series stations in the North Atlantic Ocean. 10 different ocean carbon models have been compared with an existing pco 2 climatology in the North Atlantic Ocean. Overall the models tend to capture the mean north-south pattern of the observed annual mean air-sea CO 2 flux but the longitudinal variability is overestimated by many of the models. 27 atmospheric inversions have been performed and the North Atlantic uptake has been estimated at -0.7 ± 0.26/0.58 GtCy -1. Milestones reached All the milestones planned for year 1 and given in the technical annex have been reached, or will be within a month of their anticipated delivery date:

10 An autonomous pco 2 system has been installed on the 4 ships identified for the network (R/V Nuka Arctica due mid December). The North Atlantic database has been created and populated with existing data and the first data of the network Key time series stations have been identified in the North Atlantic from the database The North Atlantic carbon balance and its rate of year-to-year variability over have been estimated The second management report has been sent. Progress by work package WP1. Establishment of an initial North Atlantic surface pco 2 observing system This work package involves partners 1, 3, 4 and 5. Equip ships with autonomous pco 2 system A number of maritime trading companies (the ships charteres) and shipping lines (the ships owners) trading along UEA s route (between the UK and the Caribbean), were approached to enquire about their interest in supporting our project and about the length of the ships charters on this particular route. Several ships were then visited to establish their suitability for the installation of our equipment. Once a ship was found suitable, discussions were started with the ship s owners, charterers, captain and chief engineer, concerning the details of installation and maintenance of the equipment. When two suitable ships experienced engine failure, new shipping companies and ships had to be found. Eventually, a successful and very supportive collaboration was established with Geest Line, Southampton, UK, and the officers and crew of the MV Santa Lucia, and modifications on the ship have been done to accommodate the instrument. A new system was built at UEA, designed for the installation on a commercial vessel. The new design included aspects for the ease of maintenance and repair by non-specifically trained scientific personell, and a variaty of fail-safe mechanisms to protect the safety of the ship in possible emergencies during long-term unattended measurements. Once built, the instrument was put through a large number of laboratory tests, to establish measurements accuracy and precision. The main important accuracies of the detector (within ± 0.05 ppm CO 2 in dry gas, dew point 0 C) and temperature sensors (within ± 0.05 ºC) are well within necessary specifications. However, certain tests can only be done in the environment on board ship. Therefore, one member of our group, Ute Schuster, will accompany the instrument during its first voyage. The building of pco 2 measuring equipment at CSIC (partner 3) called GASPAR (see annex)- is finished. The installation of GASPAR on board the R/V Hespérides was definitively done between 8 to 11 October The work developed to construct the pco 2 -measuring equipment GASPAR has been presented at the University of Vigo in November 2001, as final career plan of a Technical Engineer to obtain the degree of Engineer. The IFM (partner 4) received a fully tested and working pco 2 system middle of October. The system is designed to be installed on board of merchant ships and it runs autonomously. This instrument is also commercially available and a twin system is in continuous use by a Japanese research institute. It measures the pco 2 of seawater and air and monitors surface properties such as temperature and pressure. Also, to continuously monitor salinity and temperature of the surface waters, a thermosalinograph (SeaBird SBE 21) will be installed upstream of the pco 2 system. This instrument is owned by the Federal Maritime and Hydrographic Agency of Germany (BSH) who were willing to lend it to us. Every six months 10

11 the thermosalinograph will be replaced by a newly calibrated instrument. Maintenance and calibration will be performed by the BSH. The IfM has started a collaboration with the shipping company Wallenius-Wilhelmsen. They have agreed to our installation of such a monitoring system on the MV Falstaff, which is a car-carrier covering the following route on a monthly basis: Goteborg-Bremerhaven- Zeebrugge-Southampton-Baltimore-Charleston. The route is convenient for us in Kiel as it will be possible to conduct maintenance on the section Bremerhaven-Goteborg and then return economically to Kiel on the Goteborg-Kiel ferry. The installation started at the end of November and we expect to be collecting data on a regular basis by the middle of January. The shipping company has also kindly agreed to allow a scientist to sail with the ship in order to perform equipment maintenance as necessary and in order to sample for additional carbonrelated properties en route. The University of Bergen measures pco 2 using the colorimetric method used by the SAMI-CO 2 instrument. In summary, the method uses the colour changes of a ph sensitive indicator solution after equilibration with the seawater pco 2. The minimum response time is 5 minutes but the instrument is generally operated at much slower measurement frequencies when operated in stand-alone mode to conserve battery power (12 months at 2 measurements per hour, 2 months at 12 measurements per hour). To optimise the system further, we have revaluated the design of the flowcell supplied by the manufacturers and adopted a radically different design: 11 Plastic tubing Water flow during on-line SAMI-CO2 measurements Flowcell Water escapes through hole at base Temperature probe (a) Original US flowcell and seawater configuration. Note the completely independent water flows across the equilibrator and temperature sensor. (b) the new Bergen set-up where the seawater intake encompasses both the equilibrator and the temperature sensor The original flowcell had two water flows; one each to the temperature sensor and another to the silicon equilibration coil. We believed that this was a potential source of error as the pco2 calculation from the absorbance measurements uses the measured temperature. The new design ensures that the temperature of the indicator solution at the time of absorbance measurements is accurately known. It has been imperative that we have confidence in the measurements made by the instrument. An agreement has been finalised with the shipping company Royal Arctic Lines to install the pco 2 sensor on the merchant vessel Nuka Arctica between Aalborg, Denmark and Nuka, Greenland. The instrument has now been operating successfully in a fully stand-alone mode for 4 months. Collection of new data Partner 1: UEA Data collection will start in December 2001 on board the MV Santa Lucia and the robustness of the equipment will also be tested during this first voyage.

12 Partner 3: CSIC Surface pco 2 and relevant related data were collected along the routes to the Antarctic of the R/V Hespérides during the cruises FICARAM 1 and FICARAM 2 carried out in October- November 2000 and March-April 2001, respectively. The FICARAM 3 cruise has started in Cartagena (Spain) the 27 October and finished the 9 November 2001 in Recife (Brasil). A collaboration between CSIC-IIM and LPO (Laboratoire de Physique des Océans, France) in the frame of CLIVAR was established to participate in the OVIDE cruise to carry out in June-July 2002 between Lisbon (Portugal) and Greenland. 12 Fig. 1. Track of the OVIDE cruise framed in the international program CLIVAR. Partner 4: IFM The route of MV Falstaff covers approximately the 48 N (WOCE A2-) hydrographic line which has been and will be sampled on an annual to biennial basis for carbon and other parameters from research vessels. This year (2001), data for pco 2, chlorophyll and additional samples (e.g. dissolved organic/inorganic carbon, alkalinity, ph) were sampled on the RV Meteor (cruise M50) covering this route in spring and summer. The IfM will send a scientist to collect chlorophyll and nutrient data along the shipping route up to 4 times per year, starting during the wintertime (January, 2002). Especially the chlorophyll data will be used to evaluate the impact of biological parameters on the pco 2. Partner 5: UIB The new colorometric pco 2 sensor has been evaluated during an experiment in the Greenland Sea where the concurrent measurement of TCO 2 and total alkalinity has allowed over-determination of the CO 2 system. Thus, it was possible to compare the derived pco 2 with the direct measurements. WP2. Production and reporting of North Atlantic pco 2 database The database design and management is under the responsibility of UEA (partner 1). Partners 3, 4 and 5 submit their data to partner 1 to develop the database. The relational database has been constructed, using open source software under UNIX environment, and populated with available data obtained from PIs, public and restricted access web sites. The data structure consists of different tables containing the data, information about the cruises (dates, track, vessel) and the origin of the data (PIs, source) with their status. In addition to the CO 2 parameters (pco 2, alkalinity, TCO 2, ph), temperature, salinity, wind speed, nutrients, oxygen, chlorophyll and CFCs have been incorporated in the database.

13 It currently includes 160 cruises and more than 450,000 seawater pco 2 data measured underway. The FICARAM data collected by partner 3 have already been included in the database and further data will be incorporated as they arrive. WP3. Assessment and reduction of uncertainty in pco 2 climatologies and seasonal airsea CO 2 fluxes for the North Atlantic. This workpackage will start at year 2. WP4. Ocean model evaluation of air-sea flux variability During the first year of CAVASSOO, we evaluated ocean carbon cycle models based on their ability to simulate the seasonal climatology of air-sea CO 2 fluxes in the North Atlantic Ocean. We used output from ten different models, standard simulations made during the second phase of the Ocean Carbon-Cycle Model Intercomparison Project (OCMIP, ). European contributions to OCMIP were funded by the EU s Environment and Climate Programme (Project GOSAC, ENV4-CT ). The data reference for this exercise is the recently revised climatology from Takahashi et al. (2001), which is based on interpolation 940,000 measurements to a global 4 x 5 grid. For consistency, we calculated the data-based flux by taking the product of Takahashi et al. s (2001) fields of pco 2 and the standard gas transfer coefficient computed for the standard OCMIP simulations (see WP5. Atmospheric transport modelling of carbon sources and sinks This work package is under the responsibility of partner 2 (CEA). Flask air sampling of CO 2 over the North Atlantic An air sample will be taken approximately every 1 of latitude on board two of the ships of opportunity for high precision (±0.1 ppm) measurements of CO 2 and stable isotopes. The installation of the first air sampling unit will be soon on board of the MV Santa Lucia (Geest Bananas Ltd., Southampton, UK.) and sampling will start December The design of the flask sampling systems is shown in the schematics below. Dekabon tubing will be used for the air intake lines on board. Whole air samples will be collected into pre-conditioned 1-litre cylindrical flasks made of Pyrex glass with PFA o-ring valves at both ends. Flasks were flushed for more than 10 minutes at a flow rate of ca. 5 litre per min, and pressurised to 1 atmosphere above ambient pressure at final filling. It was initially planed to dry the air on board via magnesium perchlorate. But it turned out that it would be difficult to maintain necessary manual exchange of magnesium perchlorate by the crew of the ship. Therefore we are going to redesign this part of the device to make it more user-friendly and less time-consuming. We started to test a Nafion drying unit and were optimistic that it would solve this problem. Air intake 13 Mast Pneurop connectors 3-way toggle valve Relief valve Drying agent Pump manometer Flowmeter Cajon Ultra-Torr connectors

14 14 Schematic of the flask-sampling unit. Inverse modelling to quantify the North Atlantic carbon balance The principle of atmospheric inversions is to infer surface fluxes from atmospheric measurements, via the use of transport model. In such an exercise, an optimal set of fluxes is sought, which minimises a distance between the transport model results and the observations. Recent inverse studies (Fan et al., 1998, Bousquet et al., 2001, Rayner et al., 1999, Kaminski et al., 1999, Ciais et al., 2000) have given conflicting results for the mean carbon balance, but it is difficult to analyse the causes of such discrepancies because published inversions are all performed with different settings. We realized an ensemble of atmospheric inversions, where different settings are varied systematically, for the time discretization of measurements and fluxes, the spatial resolution of the fluxes, and the atmospheric transport model. This systematic sensitivity analysis of the inverse results to the inversion settings allows to estimate an usually quoted Gaussian errors calculated by the inversion assuming Gaussian a priori error on the observations, but also systematic errors addressed through the spread of inversion results where different settings are used. The methodology that we followed is a Bayesian inversion with priors (Takahashi et al., 1999 for prior sea-to-air fluxes). The period of study is for which we should keep in mind that abnormally low CO 2 growth rates occurred, indicating enhanced sinks over the globe. A total of 46 atmospheric sites available from the NOAA/CMDL network over this period have been retained in our set of inversions, which form a coherent ensemble. The temporal discretization of our inversions consists in inverting either monthly atmospheric observations or annual ones, either with an annual adjustment of the sources (with a priori specified seasonality) or with a monthly adjustment of the sources. The spatial discretization of our inversions is varied by solving the fluxes either over 7, 12 or 17 regions (Fig 1). Finally, we used three different global atmospheric transport models (TM2, TM3, GCTM) with varying spatial resolution, both horizontal and vertical. Such models generate varying rectifier effects as produced by the covariance of seasonal fluxes and transport, which in turn produce a systematic bias in the inverse results. Overall, we tested systematically 3 transport models, 3 spatial and 3 temporal discretizations to produce a total of 27 inversions. Map of different regions used in the inversion: 3 land regions in gray and 4 oceanic regions separate by solid lines for the 7 regions cases. The dotted lines figure the division for the 12 regions case. The location and acronyms of the stations used in this study are also shown.

15 15 Results Collection of new data The pco 2 instruments have been built, tested, installed on board the ships sailing along the proposed routes. For some of the ships the data collection has started and for the other it is about to start. During the cruises FICARAM-1 and FICARAM 2 pco 2 measurements both in air and sea as well as related data were collected along the route shown in figure FICARAM 1 50 FICARAM Figure 1. Tracks of FICARAM and FICARAM 2 cruises surveyed respectively in October-November 2000 and March-April The meridional evolution of pco 2 (= pco 2sea pco 2air ) along 28ºW obtained during FICARAM-1 and FICARAM 2 in the North Atlantic is shown on figure 2a. The longitude evolution of pco 2 along the latitude band ºN during FICARAM 2 is depicted on figure 2b. Figure 2. Course of pco 2 along a) longitude 28ºW during FICARAM 1 and FICARAM 2 and b) along the latitude band ºN during FICARAM 2.

16 In general we can observe four geographical zones that are related with different current systems. The Equatorial Current Systems (ECS) from 3ºS to 4ºN, the North Equatorial CounterCurrent (NECC) flowing eastward from 4ºN to 8ºN, the North Equatorial Current system (NEC) flowing to the west between 8ºN and 18ºN, and the Subtropical Gyre (STG) spreading to the north of 18ºN. During both cruises FICARAM 1 and FICARAM 2, the ECS domain acts as a source of pco 2 to the atmosphere. The NECC zone acts as a source of pco 2 in autumn and as a sink in spring. The NEC system is a zone of CO 2 entry in the ocean in both spring and autumn. From 5º to 7ºN. Whereas the STG in spring behaves as a sink of CO 2 and in autumn as a source. The variation of pco 2 in surface seawater along the section matches well with the thermohaline distributions, indicating that the variations of pco 2 in the surface seawater are determined by the physical and chemical characteristics of the surface water bodies. In May and August 2001 IFM (partner 4) collected pco 2 and additional data on a cruise along the 48 N route. The results are shown in figure 3 and will be shared on the CAVASSOO database. 16 f CO 2 [µatm] 355 spring summer f CO Lon E SST [ C] Chlorophyll [µg/l] SST Chl a Lon E Figure 3. Results of cruise M50 of the research vessel Meteor. The spring data were recorded in May, the summer data in August, The fco 2 and the SST (sea surface temperature) data were continously measured whereas the chlorophyll data represent discrete samples. Here, rather fco 2 than pco 2 data are shown which relates to the non-ideal behaviour of CO 2 as a gas. The difference fco 2 to pco 2 is ca. 1%. The University of Bergen has collected data on a recent survey in the northern North Atlantic. The results in Figure 4 show the excellent agreement between the measured and calculated pco 2 particularly considering that the sampling depths between the ship intake and the CTD varied considerably during the study.

17 at 270 pc m) O2 250 ( CTDStation# Figure 4. Comparison of measured pco 2 (pink) and calculated pco 2 from TCO 2 and alkalinity (blue) from a recent survey in the northern North Atlantic. Database and time-series stations The monthly maps (Figure 5) show the distribution of oceanic pco 2 recorded underway (lines) and analysed from discrete samples in the surface layer (dots).

18 18

19 19 Figure 5. Monthly distribution of pco 2 recorded underway (lines) and analysed from discrete samples in the surface layer (dots). Locations where several years of CO 2 data are available can be identified to estimate the oceanic pco 2 evolution. CO 2 climatologies assume that it follows the atmospheric increase except in cold regions (i.e. North Atlantic >45 o N) where oceanic pco 2 reflects the characteristics of the deep water because of the convective mixing. In these regions oceanic pco 2 does not increase like atmospheric pco 2 or increases at a lower rate. In the Atlantic Ocean the 20 o W meridian was a location visited many times and this is where most of the time-series stations can be found. Ideally, a span of about 10 years would be required to address the increase of oceanic pco 2. This should be achieved for some of the stations listed (Table 1) once the CAVASSOO network is fully operational. Table 1. Times-series stations 60 o N, 20 o W 50 o N, 20 o W Cruise Date Cruise Date BOFS DI182 24/5/89 BOFS DI182 20/5/89 NABE 1989 ATLANTIS II 6/6/89 NABE 1989 ATLANTIS II 2/6/89 Poseidon /6/91 BOFS DI190 26/4/90 Meteor Cruise 21 29/5/92 Poseidon /6/91 NATL 93 28/8/93 Meteor Cruise 21 23/5/92 Meteor /7/96 NATL 93 25/8/93 WOCE A24 Knorr /6/97 POS /6/95 Meteor /7/96

20 20 Meteor 1997 Cruise 39/2 5/6/97 25 o N, 65 o W 40 o N, 20 o W TTO-NAS Leg 2 23/4/81 BOFS CD053 26/9/90 POS /6/94 Vivaldi Cruise 6/5/91 POS /7/94 Meteor Cruise 21 14/5/92 POS /8/94 NATL 93 21/8/93 POS /9/94 POS /2/95 POS /10/94 AMT-2 18/5/96 POS /12/94 Meteor /6/96 POS /2/95 AMT-3 26/9/96 POS /4/95 GASEX 98 Leg 1 18/5/98 WOCE A22 Knorr /8/97 GASEX 98 Leg 2 24/6/98 24N98 Leg 1 12/1/98 Ficaram 2 12/4/01 24N98 Leg 2 14/2/98 10 o N, 22 o W FOCAL 0 21/7/82 FOCAL 2 13/1/83 FOCAL 4 4/7/83 FOCAL 6 13/1/84 AMT-1 6/10/95 AMT-2 11/5/96 AMT-3 3/10/96 AMT-7 2/10/98 OCMIP models-data compared with seasonal climatology over the North Atlantic Figure 6 shows the resulting annual mean sea-to-air flux from the data-based estimates and the ten models. As a measure of uncertainty, the same figure also shows the sea-air CO 2 flux computed using an earlier climatology of pco 2 from the same authors (Takahashi et al., 1999, interpolation of 550,000 measurements).

21 21 Figure 6. Map of the annual mean sea-to-air CO2 flux in 1995 in the North Atlantic from the observed climatologies of Takahashi et al. (1999), Takahashi et al. (2001), the OCMIP-2 model mean, and ten OCMIP-2 models. The OCMIP model results are sum of CO2 fluxes from the Biotic simulation + the Abiotic Historical run (1995) the Abiotic equilibrium run (see OCMIP HOWTO documents on Overall, the models tend to capture the mean north-south pattern of the observed annual mean sea-air CO 2 flux, with large CO 2 uptake in the high latitudes and CO 2 release in the tropics. Yet, longitudinal variability is overestimated by many of the models. For example between 30 N and 50 N, there is no east-west difference in the observed sea-air CO 2 flux, whereas simulated differences reach at least to 4 mol C m -2 yr -1 in half of the models (AWI, CSIRO, IGCR, IPSL, SOC). These annual mean maps clearly show that no two models have the same local variability and that they all differ from the observed distribution. However, it is difficult to go beyond this qualitative assessment and to evaluate model-data coherence in a concise manner for the full spatio-temporal distribution.

22 22 To provide a concise quantitative assessment of overall model performance and to include evaluation of seasonal variability, we have computed several commonly used summary statistics. These include the standard deviation of the model σ f and that of the data σ r, the correlation coefficient R, the centered pattern RMS error E, and the overall bias Ē. To simplify presentation, we display all five of these summary statistics in one diagram. This approach was recently developed to compare atmospheric models (Taylor, 2001), and its utility was demonstrated for coupled climate models in the recent Third Assessment Report (TAR) report of the IPCC (McAvaney et al., 2001). In short, the "Taylor" diagram is based on the geometric relationship between the correlation coefficient, the simulated and observed variances, and the centered pattern RMS difference (Figure 7). More specifically, it is based on the recognition that R, E, σ f, and σ r are related through the Law of Cosines: E 2 = σ f 2 + σ r 2-2 σ f σ r R Figure 7. Geometric relationship between the correlation coefficient, R, the pattern RMS error E, and the standard deviations for the model and the reference fields. Figure 8 shows the Taylor diagram for the air-sea CO 2 fluxes in the North Atlantic (20 N to 70 N). This figure includes this analysis for the overall spatio-temporal distribution in the North Atlantic (colored filled ovals). Additionally, it also breaks down the different spatial and temporal components separately to show their contribution to the overall error (asterisks colored according to component). The relative position of the different reference points indicates the relative importance of each component to the overall variance. In the North Atlantic, seasonal variability dominates, and latitudinal variability is larger than longitudinal variability. The models generally perform well in regards to the zonal mean, with the majority of the models showing a correlation of 0.9 or better. Conversely, the models do poorly when it comes to simulating the observed longitudinal variations in the sea-air CO 2 flux (correlations between 0.2 and 0.4 for the seasonal models). The skill of the models in reproducing the seasonal cycle in the North Atlantic is only slightly better (R from 0.3 to 0.5), but not by Overall, models explain 40 to 70% of the spatio-temporal variability in the North Atlantic (R=0.4 to 0.7). In such comparisons, a common question is often asked. Which model is the best? This question is somewhat naive because in such comparisons, there is seldom a model, which outperforms all the others. However, the modeling community does need to be able to identify when a particular model distinguishes itself, in order that significant advances are identified and can be incorporated by other groups. The Taylor diagram offers one means to identify such advances. For the total air-sea CO 2 flux climatology, the UL model shows the best representation of the observed zonal mean, but it is a relatively poor performer in terms of its ability to capture seasonal variability. The CSIRO and NCAR models offer the best overall performance, often yielding results, which are similar to those of the mean. Interestingly, if a model outperforms the mean, that is a generally a good performance indicator. That is, the best performing model is often the mean of all OCMIP-2 models. Preliminary tests indicate that the median provides even somewhat better results because

23 outliers play less of role. These results clearly show the utility of making simulations with an ensemble of models such as carried out during OCMIP. 23 Figure 8. Taylor Diagram of the seasonal sea-air CO 2 flux in the North Atlantic (22 N to 70 N). One reads a Taylor diagram (an r-θ or polar plot) by first comparing a model s standard deviation (the radius r) to the standard deviation of the data. If a model exhibits less overall variance than that of the data, its colored oval will have an r which is less than the black reference curve. In this low-variance category fall the LLNL, IGCR, CSIRO models, the Mean of all the ten OCMIP-2 models, and the older data-based climatology (Takahashi et al., 1999) relative to the reference, i.e., the more recent climatology (Takahashi et al., 2001). Second, the angle θ of the Taylor diagram indicates the correlation coefficient between the model and the data reference. If a model were perfect, it would lie along the X-axis, right on top of the reference point. The reference point is marked Seas. Map for the full space-time distribution (black). The other reference points represent analogous analysis but for different components of the space-time distribution: the Annual Zonal Mean (green), the spatial deviation from the Annual Zonal mean (Annual Map Annual Zonal Mean, red), and the Time deviations from the annual map (Seasonal Map Annual Map, blue). Third, the distance from the appropriate reference point to a given model represents that model s central pattern root mean square (r.m.s) difference between the two fields E. Finally, the overall bias Ē (a scalar indicating the difference between the spatio-temporal mean of each of the two fields) is indicated by the color which fills each of the ovals, as indicated by the color key to the right of the plot.. Note that units for Ē, E, σ f, and σ r are identical (mol C m -2 yr -1 ) and that the overall r.m.s. error E is the quadratic sum of Ē and E (i.e., E 2 = Ē 2 + E 2 ). This analysis of the seasonal sea-air CO 2 flux does not provide by itself, an indication of how good a model may be to study future uptake of CO 2. However, when combined with

24 information from similar analysis involving other anthropogenic tracers (e.g., CFC-11) and other timescales (e.g. interannual to decadal variability) this should lead to greater confidences than at present. The present analysis indicates that for the North Atlantic, the difference between a typical model simulation and a baseline set of observations is larger than the difference between different sets of observations. However, the two sets of observations are not independent; the actual uncertainty may be larger. This could mean then that ocean carbon cycle models are about as accurate as observational uncertainty currently allows them to be in terms of the quantities highlighted by our summary statistics. Independent estimates of the seasonal air-sea CO 2 flux in the North Atlantic from CAVASSOO, will improve estimates of the seasonal cycle of the air-sea CO 2 flux and will improve our ability to assess uncertainty, which is important for such analysis. 24 Inverse modelling to quantify the North Atlantic carbon balance The 27 inversion results are discussed in greater detail in Peylin et al We recall them briefly here and discuss specifically the North Atlantic results (Figure 9). The estimated fluxes are stable over broad zonal bands (Southern Hemisphere; Northern Hemisphere; Tropics), but the land versus ocean apportionment is not stable within one latitude band. In the Northern Hemisphere, the sum of all land and ocean regions is also remarkably robust (sink of 3.2 ± 0.3 GtC y -1 ). Figure 9. Annual land ocean and total fluxes from our 27 inversions reported for three latitude bands and the global total. The gray bars give the ±1σ range of the results, with the mean lying in the middle of the bar. Dashed horizontal lines indicates the ocean uptake estimate by Takahashi et al. (1999) A large uptake being required in the North to match the observed north-south gradient in CO 2 observations, while the sum North + Tropics + South must match the global trend. It is interesting to note that the tropics as a whole (ocean + land) is only a net source of 0.2 GtC y -1 across all inversions, suggesting that deforestation emissions could be significantly offset by forest regrowth over tropical continents. Another striking -and remarkably robust- result is that all inversions place over the Southern Ocean a smaller sink (-0.6 GtCy -1 ± 0.2) than the one (-1.3 GtCy -1 ) given by Takahashi et al. (1999) in their 1995 compilation of pco 2 measurements. While the latitudinal distribution of the fluxes is rather robust, the longitudinal distribution within the Northern Hemisphere is not for the four regions considered (North America, Eurasia, North Pacific and North Atlantic) as shown in Figure 10. A larger sink is obtained over Eurasia. Sinks over North America and over the North Atlantic are of similar magnitudes, while the North Pacific is a slight sink. The northern ocean sink is of 0.8 ± 0.6 GtC y-1 on average, within 1σ of the value given in Takahashi et al. (1999). The North

25 Atlantic average sink from the 27 inversions is of 0.7 GtC y-1. Individually, many of the 27 inversion also differ from the Takahashi et al. (1999) estimate Figure 10. Same Figure as before, but for the four regions of the northern hemisphere. The error on the North Atlantic flux that incurs from the Gaussian error in the observation is 0.26 GtC y -1. On the other hand, the confidence on the inverse results, given from the spread of the 27 inversions is of 0.58 GtC y -1. The North Pacific flux though proves more robust within 0.2 GtCy-1 than the North Atlantic flux, probably due to the presence of several atmospheric stations over the North Pacific Ocean. Accounting for Gaussian errors / confidence, we estimate across the 27 inversions a North Atlantic uptake of 0.7 ± 0.26 / 0.58 GtCy -1. Much of the systematic uncertainty in the North Atlantic inversion result is due to large anticorrelation between the fluxes for North America and the North Atlantic for some settings (Figure 11). This indicates that the stations are not able to fully separate North America and North Atlantic mean fluxes, when using an annual adjustment of monthly data. Removing outliers where a correlation between North America and the North Atlantic is obtained in the case of annual adjustment of monthly data, yields a best estimate of 0.9 ± 0.3 / 0.3 GtCy-1 for the North Atlantic sink north of 15 N. This inverse estimate gives a larger sink within 2σ than the one given in Takahashi et al. of 0.55 GtC y-1. Although the North Atlantic is one of the places where Takahashi s fluxes are most tightly constrained by measurements, more oceanic /atmospheric measurements there would help to resolve the disagreement with inverse results. 1 Note that one of our inversion with a similar setting as the one of Fan et al. gives a larger North Atlantic uptake (-1.0 GtC y-1 instead of GtCy-1 in Fan et al.), whereas the North American uptake decreases (-0.9 GtC y-1 instead of 1.6 GtCy-1 in Fan et al.).

26 26 Figure 11. Annual CO 2 fluxes for North America versus North Atlantic, as estimated in the 27 inversion cases. The 7, 12, and 17 regions inversions show the same symbols. PRELIMINARY CONCLUSIONS Progress towards project objectives The milestones of year 1 have been reached, with only slight (less than one month) delay on one ship installation. The pco 2 instruments are built and after negotiations with shipping companies they have been installed on board ships sailing along the routes proposed in the project. The data collection has started on board some of the ships. New data from the network have been recorded along the meridional transect on board the R/V Hesperides in October-November 2000, March-April 2001 and October-November The network is set up and should be fully operational within the next few months. Opportunistic research cruises and links with other projects will provide additional data and complement the network. The database currently includes 160 cruises and more than 450,000 seawater pco 2 data measured underway. It is ready for loading new data. Time-series stations have been identified mainly along 20 o W in the different biogeochemical provinces (North Atlantic gyre, tropics, subarctic Atlantic). CO 2 climatologies assume that oceanic pco 2 follows the atmospheric increase except in cold regions (i.e. North Atlantic >45 o N) where oceanic pco 2 reflects the characteristics of the deep water because of the convective mixing. In these regions oceanic pco 2 does not increase like atmospheric pco 2 or increases at a lower rate. The evolution of pco 2 (difference between seawater pco 2 and atmospheric pco 2 ) will be studied at stations located at 60 o N, 20 o W, 50 o N, 20 o W, 25 o N, 65 o W, 40 o N, 20 o W, and 10 o N, 22 o W. The model-data seasonal comparison has started with existing climatologies. 10 different ocean carbon models have been compared with CO 2 climatology. Overall the models tend to capture the mean north-south pattern of the observed annual mean air-sea CO 2 flux but the longitudinal variability is overestimated by many of the models. A quantitative assessment of overall model performance has been performed using commonly used summary statistics. The model-data comparison should be refined with update climatologies of the North Atlantic Ocean.

27 27 Atmospheric inversions with different settings have been performed using 46 atmospheric sites available from the NOAA/CMDL network. Three different atmospheric transport models with varying spatial resolution have been used. The North Atlantic uptake has been estimated at 0.7 ± 0.26/0.58 GtC yr -1 from 27 atmospheric inversions. The inverse estimate gives a larger sink than the one given by CO 2 climatology and the oceanic/atmospheric measurements from CAVASSOO would help to resolve the disagreement with inverse results. Implications for next reporting period At the end of year 2 we should have many new data and a better knowledge of seasonal/ interannual variability from these new measurements and the existing data available in the database. Assumptions relative to climatologies should start being tested. REFERENCES Bousquet, P., Peylin,P., Ciais, P., Rayner, P. Friedlingstein, P., Lequere, C. and.tans, P.P., Interannual CO2 sources and sinks as deduced by inversion of atmospheric CO 2 data. Science, 290, , Ciais et al., 2000 Fan, S., Gloor, M., Mahlman, J., Pacala, S., Sarmiento, J., Takahashi, T. and Tans, P.P., A large terrestrial carbon sink in North America implied by atmospheric and oceanic carbon dioxide data and models, Science, 282, , Kamininski et al., 1999 McAvaney, A.J.,Covey, C., Joussaume, S., Kattsov, V, Kitoh, A., Ogana, W., Pitman, A.J., Weaver, A.J., Wood, R.A., Zhao, Z.-C.,Chapter 8. Model Evaluation, IPCC Third Assessment Report, Working Group I, Cambridge Univ. Press, Peylin, P., Baker, D., Sarmiento, J., Ciais, P., and Bousquet, P., Influence of transport uncertainty on annual mean and seasonal inversions of atmospheric CO 2 data. Submitted Rayner, P.J., I.G. Enting, R.J. Francy, and R. Langenfelds. Reconstructing the recent carbon cycle from atmospheric CO2, d13c and O2/N2 observations. Tellus, 51b, ,1999. Takahashi, T. Wanninkhof, R.H., Feely, R.A., Weiss, R.F., Chipman, D.W., Bates, N., Olafsson, J., Sabine, C. and Sutherland, S.C., Net sea-air CO 2 flux over the global oceans: An improved estimate based on the sea air pco 2 difference. In 2 nd International Symposium, CO2 in the Oceans, extended abstracts, Center Global Env. Res., Tsukuba, Japan, Takahashi, T., Sutherland, S.C, Sweeney, C., Poisson, A., Metzl. N., Tilbrook, B., Bates, N., Wanninkhof, R., Feely, R.A., Sabine, C., Olafsson, J., And Noriji, Y., Global sea-air CO 2 flux based on climatological surface ocean pco2 and seasonal biological and temperature effects. Deep-Sea Research, Brest Meeting Volume, accepted, Taylor, K.E., Summarizing multiple aspects of model performance in a single diagram, J. Geophys. Res. 106, D7, , 2001 ANNEXES Specific deliverables The pco 2 instrument by partner 3 is called GASPAR and a picture of GASPAR with some information is enclosed.

28 28 GASPAR WAS BORN Aida F. Ríos His birth had to be by Caesarean operation because time is over and he risks losing something important. He has an older brother and some cousins in German, England and Norway. Through their tygon veins ands stainless steel arteries pass fluids pumped by a heart formed by three Fürgut pumps working at 24 V. One of them makes to function the peripheral system of closed-circuit. The other two, working in a connected way, drive its right and left ventricles that make run the main circulation system communicating his lung also call gas equilibrator- with his liver, named LiCor that is a measured system of gasses by infrared. Like his brother, he has a congenital illness and he needs have attended respiration. An Air-Cadet machine provides him comfort and air at atmospherically pressure. His brain is a last generation Pentium with 2000 Professional Millennium software. His preceptors and his parents have put inside their knowledge with the aim that his behaviour and cerebral orders will be the best for a long life and a good performance. Thanks to this brain he can move his arms and legs the valves of 2 and 3 ways- give him mobility to calibrate and check the levels of partial pressure of carbon dioxide in the atmosphere and in the sea water. He never gets lost. A GPS provides him the needed coordinates to know always where he is. He has very clear his destiny: to measure continuously partial pressure of carbon dioxide both in the air and in the seawater after their liquid and gas phases were equilibrated. For this reason, he was baptized GAS-PAR. If he were not born in time, he would have lost a cruise along the Atlantic Ocean on board R/V Hespérides towards he leaves today. There, I am convincing that the fresh air agrees with him, and I think also that he will feel useful.

29 Pictures of GASPAR 29

30 30 LICOR SAMPLE IN SAMPLE OUT REFERNCE OUT REFERNCE IN CIRCUITO GASES CALIBRACION CIRCUITO AIRE EQUILIBRADOR CIRCUITO AIRE ATMOSFERICO 2 3 PURGA H L C 1 HST LST CERO NUPRO AIR CADET EQU GASES CALIBRACION LICOR LICOR LICOR EQUILIBRADOR AIRE pco 2 ENTRADA AGUA DE MAR INTERCAMBIADOR DE CALOR AGUA DE MAR SALIDA AGUA DE MAR Diagrama de funcionamiento en modo CALIBRACION.

31 31 LICOR SAMPLE IN SAMPLE OUT REFERNCE OUT REFERNCE IN CIRCUITO GASES CALIBRACION CIRCUITO AIRE EQUILIBRADOR CIRCUITO AIRE ATMOSFERICO 2 3 PURGA H L C 1 HST LST CERO NUPRO AIR CADET EQU GASES CALIBRACION LICOR LICOR LICOR EQUILIBRADOR AIRE pco 2 ENTRADA AGUA DE MAR INTERCAMBIADOR DE CALOR AGUA DE MAR SALIDA AGUA DE MAR Diagrama de funcionamiento en modo ATMOSFERA.

32 LICOR 32 SAMPLE IN SAMPLE OUT REFERNCE OUT REFERNCE IN CIRCUITO GASES CALIBRACION CIRCUITO AIRE EQUILIBRADOR CIRCUITO AIRE ATMOSFERICO 2 3 H L C 1 HST LST CERO NUPRO AIR CADET EQU GASES CALIBRACION LICOR LICOR LICOR EQUILIBRADOR AIRE pco2 ENTRADA AGUA DE MAR INTERCAMBIADOR DE CALOR AGUA DE MAR SALIDA AGUA DE MAR Diagrama de funcionamiento en modo EQUILIBRADOR

33 Reports on workshops A workshop was organised in UK to discuss the progress of the pco 2 network. 33

34 34 Report on CAVASSOO project meeting, 24 th September 2001, Gt Dunmow, UK Agenda: Status of the pco 2 instruments Identification of ships for the network Database Others Provisional date for next meeting Participants: Partners involved in the set up of the network UEA Andy Watson Nathalie Lefèvre Ute Schuster CSIC Aida Rios IFM Arne Körtzinger Heike Lueger UIB Truls Johannessen Richard Bellerby Status of the pco 2 instruments/ Identification of ships for the network Each partner presented the status of the pco 2 instrument and the negotiations with the shipping companies. UEA New instrument built, being tested Agreement reached with Geest Bananas about the MV Santa Maria or MV Santa Lucia Ships, installation & maintenance are known Uncertainty about exact ship to be on the route by when; plan to go out on 5 November (risk of change of ship in December). Therefore, instrument is at sea on 1 December, but maybe not measuring routinely IFM Instrument on loan from Japan, fully tested Agreement reached with Willenius Wilhelmsen about the MV Falstaff Discussion about installation & maintenance: end of November, Bremerhaven to Göteborg Installation start end of November, start of regular data collection mid January 2002 Therefore, will be 1 to 2 months behind schedule UIB New SAMI just delivered Agreement reached with Royal Arctic Line about the MV Nuka Arctica and replacement winter ship No air pco 2 data for first six months Change onto winter ship at Christmas, three weeks gap Discussion about installation & maintenance: early November in Aalborg Start of data collection in December Therefore, up to 1 month behind schedule CSIC

35 35 Already started measuring onboard the RV Hepérides Next cruise end of October to March 2002 Therefore, on schedule Database The structure of the database was presented. The list of tables and parameters was distributed to each partner for comments (addition of other variables etc ). The database was set up on a UNIX workstation at UEA and data were being loaded. The work was going on schedule and time-series stations could be identified by the 1 st of December. Others The partners wished to discuss the CO 2 analysis and compared their analytical protocol. Principle of measurements in wet or dry gas: UEA and IfM measure dry gas CSIC measures wet gas, heating tubing, needs to do humidity calibrations UoB measures wet gas, calibration with Li-Cor dew point generator The drying of gas is not easy: IfM use Peltier drying and Nafion counter flow and Mg(ClO 4 ) 2 UEA wants to use counter flow, with or without downstream chemical dryer IfM should pass information on Nafion counter flow and Peltier dryer on to UEA CO 2 calibrations with gas standards: UEA, IfM, and UoB calibrate 3 primary NOAA gas standards each CSIC gets standards calibrated externally CSIC needs to check whether gas standards calibrations can be traced back to NOAA standards; if no, such calibration needs to be established; for example via CAVASSOO partners? Seawater temperature measurements All: Schedule for regular temperature calibrations All: Calibration to ± 0.1 o C CAVASSOO should be part of the world wide network JCOMM. A CAVASSOO member should go to the JCOMM meeting in Goa in February The CAVASSOO web site is on a temporary location due to the breakdown of the workstation "tracer". A re-direction should be established to keep the same URL address. Provisional date for next meeting A meeting with all the project members is planned for next year in Vigo.

36 36 Published material Bellerby, R.G.J., Olsen A., Johannessen T. and Croot P., A high precision spectrophotometric method for on-line shipboard seawater ph measurements: The Automated Marine ph Sensor (AMpS), pp. 22, Talanta, in press. Körtzinger, A., J.I. Hedges, and P.D. Quay. Redfield ratios revisited: Removing the biasing effect of anthropogenic CO 2. Limnology and Oceanography, 46, , Körtzinger, A., W. Koeve, P. Kähler, and L. Mintrop. C:N ratios in the mixed layer during the productive season in the Northeast Atlantic Ocean. Deep-Sea Research I, 48, , Pérez, F.F., L. Mintrop, O. Llinás, M. Glez-Dávila, C.G. Castro, M. Alvarez, A. Körtzinger, M. Santana-Casiano, M.J. Rueda, and A.F. Ríos. Mixing analysis of nutrients, oxygen and inorganic carbon in the Canary Islands region. Journal of Marine Systems, 28, , Lefèvre, N. and Taylor A.H., Estimating pco 2 from sea surface temperatures in the Atlantic gyres, Deep-Sea Research, in press. Wallace, D.W.R., Prentice I.C., and Schimel D., The Global Carbon Cycle. In Contributions to Global Change Research, Report of the German National Committee on Global Change Research, Bonn, 2001, pp Wallace, D.W.R., Radiatively active gases and the role of the oceans, in Marine Science Frontiers for Europe, Proceedings of a Hanse Conference, February 2001, in review. Watson, A.J. and Orr, J.C., Carbon dioxide fluxes in the global ocean. Chapter 5 in Ocean Biogeochemistry: a JGOFS synthesis Eds: Fasham, M. Field, J. Platt, T. & B. Zeitzschel, in review. Publicity material Press summary of the X Antarctic cruise of Hespérides. This document was sent to al communication media during a press conference held in Cartagena the 17 April 2001 at the end of the Antarctic expedition.

37 37 RESUMEN DE PRENSA DE LA X CAMPAÑA ANTÁRTICA DEL HESPÉRIDES (A-33). 25 de octubre de de abril de 2001 La campaña Antártica marcó un hito en la investigación científica española en el Continente Blanco, pues por primera vez dos buques de la Armada estvieron apoyando y posibilitando la acción de la comunidad científica. Para el Buque de Investigación Oceanográfica (BIO) "Hespérides" (A-33) fue su X Campaña Antártica, con presencia en los veranos australes de forma ininterrumpida desde su entrada en servicio y alta en la lista Oficial de Buques de la Armada el 16 de mayo de 1991, y la segunda campaña antártica para su actual Comandante, el capitán de fragata Carlos Cordón Scharfhausen. Este año participó como buque de apoyo logístico a las Bases Antárticas Españolas (BAE s) Juan Carlos I (en isla Livingston) y Gabriel de Castilla (en isla Decepción) y a la Base Antártica Búlgara (BAB) San Clemente de Orhida el buque de la Armada "Las Palmas" (A-52) mandado por el capitán de corbeta Juan Sobrino Pérez-Crespo y bajo el mando táctico del Comandante del Hespérides. Para el "Las Palmas" fue su IV Campaña Antártica (efectuó las de los veranos australes de los años 88/89, 89/90 y 90/91 hasta ser relevado por el "Hespérides"). El disponer de dos barcos de la Armada permitió que el "Hespérides" se dedicase más de lleno a las campañas de investigación oceanográficas, y se obtuviese una gran flexibilidad y mayores posibilidades en el apoyo a las Bases Antárticas y a los investigadores que trabajan en ellas. Nº DE DÍAS %DEL TOTAL DÍAS TOTALES DE CAMPAÑA % SINGALDURAS (DÍAS DE MAR) % DÍAS DE MAR (COMPLETOS) % DÍAS DE PUERTO % ESCALAS DESCANSO DOTACIÓN % ESCALAS LOGÍSTICAS % La X Campaña Antártica del "Hespérides" comenzó el pasado día 25 de octubre de 2000 saliendo de su base, el Arsenal Militar de Cartagena, en tránsito hacia Las Palmas de Gran Canaria llevando embarcados profesores y alumnos de las Facultades de Ciencias del Mar (Alicante, Cádiz, Las Palmas y Vigo) haciendo cometidos de buque-escuela, y también llevando investigadores de la Universidad Politécnica de Cataluña para un estudio del control de emisiones y de interferencias electromagnéticas, EMI (HE 069). La llegada al Arsenal Militar de Las Palmas fue el día 28 de octubre.

38 Al día siguiente, día 29, salió a la mar iniciando tránsito por el Atlántico Central efectuando la campaña FICARAM 1 (HE 069) del doctor Fiz Fernández Pérez del Instituto de Investigaciones Marinas (IIM) del CSIC en Vigo, como parte de un proyecto de la Comunidad Europea que investiga el ciclo del carbono y el efecto "invernadero", pero en el que los sensores y el equipamiento científico son de diseño español, por lo que es probable que la toma de datos se repita en tránsitos de próximos años. Los objetivos específicos de esta campaña fueron evaluar el intercambio aire-mar del dióxido de carbono y estimar la incorporación de carbono antropogénico. Durante la FICARAM 1 se efectuaron ocho mediciones con la roseta CTD y botellas a profundidades cercanas a los 3000 mts., finalizando la misma a la llegada a Salvador de Bahía (Brasil) el 10 de noviembre, con escala hasta el día 13. La siguiente escala fue en Montevideo (República Oriental del Uruguay) del 19 al 21 de noviembre y se aprovechó para consolidar las relaciones con la Armada uruguaya y el programa que el Instituto Antártico Uruguayo desarrolla en su Base Antártica Artigas (isla Rey Jorge). La última escala del tránsito fue Ushuaia (Argentina) del 26 al 28 de noviembre embarcando al personal científico y técnico de la campaña SCAN 2001 y de las Bases Antárticas Juan Carlos I y San Clemente de Orhida. El día 30 de noviembre se procedió a la apertura de la BAE Juan Carlos I en bahía Sur de la isla Livingston de las Shetland del Sur. También se desembarcó personal y material en la misma bahía para la BAB San Clemente de Orhida. Una vez completada la apertura de las Bases de bahía Sur, el "Hespérides" se trasladó a isla Decepción para un proyecto de geología (HE 070) del Dr. Luis Somoza del Instituto Geominero de España (ITGE) para la actualización de la base de datos geoquímicos de los procesos hidrotermales de la isla Decepción. Se efectuaron tres dragados, otros tantos lanzamientos de batitermógrafos, líneas de batimetría con la ecosonda multihaz y sedimentos con la ecosonda paramétrica en el contorno interior de la isla Decepción y se recogieron muestras petrológicas del fondo de Caleta Péndulo por medio de buzos de la dotación. También se aprovisionó de víveres a la BAE del Ejército de Tierra Gabriel de Castilla. Desde el 5 de diciembre del 2000 hasta el día 8 de enero del 2001 tuvo lugar la campaña científica SCAN 2001 (HE 071) del Dr. Andrés Maldonado del Instituto Andaluz de Ciencias de la Tierra, organismo mixto del CSIC y la Universidad de Granada sobre Geodinámica del borde de placas tectónicas SCotia/ANtártica, paleoceanografía de la denominada Agua Antártica Profunda, y el desarrollo de las cuencas oceánicas antárticas. Esta campaña fue una continuación de cuatro campañas efectuadas por el mismo equipo en la Antártida en el "Hespérides", denominadas ANT 92, HESANT 92/93, SCAN 97, y ANTPAC 97/98. Presentó un gran reto y dificultad por desarrollarse en la zona nororiental del Mar de Weddell y áreas próximas al denominado microcontinente de las islas Orcadas del Sur, zona que presentó unas condiciones glaciológicas (de hielos) y meteorológicas muy difíciles, incluso para las consideradas como normales en el mes de diciembre, lo que muchas veces imposibilitó el acceso y navegación con sistemas de sísmica multicanal (remolcando un conjunto hidrofónico de 2,5 kilómetros de longitud), lo que ha llevado a que esta zona haya sido poco investigada en comparación con otros sectores antárticos. Se navegó por el Mar del Scotia y se utilizaron ecosondas multihaz, paramétrica (de sedimentos), magnetómetro, gravímetro y dragas de arrastre. En la campaña participaron dos investigadores brasileños (de la empresa PETROBRAS y de la Universidad de Niteroi) y una doctora de geofísica de Sofía (Bulgaria), así como investigadores españoles de las Universidades de Granada, Barcelona y Cádiz (Facultad de Ciencias del Mar). 38

39 En el transcurso de dicha campaña, del 30 de diciembre de 2000 al 1 de enero de 2001, recibimos la visita del Secretario de Estado de Defensa Excmo. Sr. Fernando Díez Moreno quien compartió las fiestas de Nochevieja y Año Nuevo con las dotaciones del Hespérides, Las Palmas y las Bases Antárticas Españolas Gabriel de Castilla y Juan Carlos I. Ha sido la primera vez que un alto cargo de Defensa ha visitado la Antártida. Durante la campaña SCAN-2001 se realizaron 2199 millas (4090 Km) de sísmica multicanal y 3623 millas (6730 Km) de batimetría, gravimetría y magnetometría. Según palabras del propio Jefe Científico, Dr. Maldonado, los resultados obtenidos se pueden considerar, espectaculares desde el punto de vista técnico, así como por el interés científico. Del 8 al 15 de enero del 2001 se efectuó una escala en Ushuaia para descanso de la dotación y actividades logísticas de relevos de personal investigador de las campañas y de las Bases Antárticas. Del 21 de enero al 12 de febrero tuvo lugar la campaña ICEFISH (HE 072) de la Dra. Beatriz Morales del Instituto Mediterráneo de Estudios Avanzados (IMEDEA) del CSIC. El nombre de la campaña viene de que por primera vez se pretende estudiar las adaptaciones fisiológicas de los PECES DE HIELO (denominados así por su color blanco acerado y por ser los únicos vertebrados que carecen de hemoglobina en su sangre, pero que contiene un péptido como anticongelante natural). Este estudio sobre las larvas de los peces de hielo podría abrir una nueva generación de conocimientos que tendrían aplicación en procesos patológicos humanos (asma, insuficiencias respiratorias, hipoxias, etc.,) del área de la salud. Por ello, participó en esta campaña la Unidad de Apoyo a la Investigación (UAI) del Hospital Universitario de Son Dureta (Palma de Mallorca). Los resultados de esta campaña cuyo objetivo principal era la captura de larvas de peces antárticos, en especial del pez de hielo para hacer un estudio del consumo de oxígeno, incluso sometido a presión dentro de la cámara hiperbárica, también se pueden considerar como extraordinario, dado que inicialmente ya se hubiese considerado un éxito obtener 20 larvas de este pez debido a su irregular y escasa distribución, y el resultado final fue 800 larvas de peces antárticos, de las cuales 135 eran de pez de hielo, algunas de ellas de gran tamaño (varios centímetros) y en buen estado. Para obtener estos resultados se efectuaron 189 maniobras científicas que consistieron en: 50 muestreos con roseta CTD y botellas hasta una profundidad entre 200 y 500 mts, 53 pescas con la red múltiple de plancton BIONESS y 86 pescas con la red pelágica de plancton IKMT (Isaacs Kid Midwater Trawl). Durante esta Campaña, el día 1 de febrero el Hespérides recibió la llamada del Comando Naval Austral y MRCC (Centro de Coordinación de Rescate Marítimo) de Ushuaia para colaborar en la operación de rescate del súbdito australiano Peter J. Bland accidentado con politraumatismo (entre otros fractura en la base del cráneo) en una grieta del glaciar McNeile a 1200 mts de altitud, en las proximidades de bahía Carcot de la Península Antártica y procedente del yate Tooluka, encontrándonos 30 millas al norte efectuando estaciones y pescas en las proximidades de isla Torre e isla Astrolabio. Se establece enlace y se coordinan los apoyos aéreos de rescate con la Base Antártica Chilena Presidente Frei en isla Rey Jorge que envía un helicóptero que intenta el escate pero sin conseguirlo por las malas condiciones de visibilidad y glaciología de la zona del accidentado. El día 2, después del informe meteorológico favorable proporcionado por el Hespérides que alistó la cubierta de vuelo y el equipo médido y coordinó las comunicaciones con el equipo de rescate en tierra y el yate Tooluka el herido es recatado por el helicóptero chileno siendo trasladado a la base chilena Frei para su posterior aeroevacución a Punta Arenas (Chile). Del 12 al 17 de febrero se efectuó una escala en Ushuaia para efectuar el desembarco de los investigadores de la campaña ICEFISH y el embarco de periodistas de distintos Medios de Comunicación Social (MCS) españoles. El Almirante Kenny, Comandante del 39

40 Área Naval, Austral hizo entrega del Diploma de Honor al Hespérides, En reconocimiento a su valiosa colaboración en las tareas de rescate de un herido en la Antártida. Desde la salida de Ushuaia, el día 17 de febrero, hasta el regreso, el día 3 de marzo, tuvo lugar una campaña divulgativa para los MCS constituidos por las televisiones de TELE 5 (Informativos), TVE 1 (Informe Semanal), CANAL 9 (Dossier), TV3 (2000EF), el diario EL PAÍS, las cadenas de radio COPE y ONDA CERO, la agencia COLPISA y redactores y fotógrafos de La Revista Española de Defensa y el Boletín Inforemativo de Tierra. En esta campaña ha tenido una amplia repercusión en la opinión pública española dándose así a conocer las actividades desarrolladas por el Hespérides, Las Palmas y las bases antárticas españolas Gabriel de Castilla y Juan Carlos I en el Continente Helado. El 28 de febrero se procedió al cierre de la BAE Gabriel de Castilla y el 1 de marzo se cerraron las BAE Juan Carlos I y la BAB San Clemente de Orhida siendo testigos privilegiados los MCS embarcados finalizando el Hespérides la actividad logística y científica en la Antártida e iniciando tránsito a Ushuaia donde se entra el día 3 de marzo. En el tránsito de regreso de la campaña antártica desde la salida de Ushuaia el día 5 de marzo hasta la llegada a Cartagenail, el "Hespérides" llevó a cabo la campaña FICARAM 2 (HE 073) de la doctora Aída Fernández Ríos del IIM del CSIC de Vigo.A diferencia de la FICARAM 1, esta campaña constaba de dos partes; la primera se desarrolló en el obrode oeste del Atlántico Sur desde las Malvinas hasta el Ecuador, siguiéndose el transecto efectuado, en el año 1994, por el buque norteamericano Maurice Ewing dentro del programa internacional WOCE (World Ocean Circulation Experiment). La segunda parte de la campaña se desarrolló en el Atlántico Norte y está encuadrada en el proyecto europeo CAVASSOO (CArbon VAriability Studies by Ships Of Opportunity). Esta campaña es parte de un programa de varios años de duración, cuyo objetivo es determinar la captación de carbono de origen antropogénico como consecuencia de la combustión de derivados del petróleo, así como su evolución internacual del Atlántico Norte. Para obtener estas estimaciones de CO 2 en la cubeta del Atlántico Norte se instalarán equipos automáticos de medida de pco 2 (presión parcial de CO 2 ) tanto de agua de mar como en la atmósfera, a bordo de buques de oportunidad. El proyecto CAVASSOO está coordinado por la Universidad de Norwich (Reino Unido) y en él participan, entre otros, el Institut für Meereskunde de la Universidad de Kiel (Alemania) y el Geophysical Institute de la Universidad de Bergen (Noruega). El proyecto tiene un gran interés científico dentro del marco del cambio climático global y la contribución española del Hespérides es esencial para el seguimiento de la variación latitudinal. Durante esta campaña el Hespérides efetuó 30 estaciones CDT hasta 3000 mts de profundidad y recorrió 8826 millas midiendo en continuo. Se efectuaron escalas en Río de Janeiro (Brasil) del 15 al 20 de marzo y en Punta Delgada de la isla de San Miguel de las Azores (Portugal) del 9 al 12 de abril, donde embarcaron profesores y alumnos de las Facultades de Ciencias del Mar hasta nuestra llegada a Cartagena el 17 de abril, donde rendió la X Campaña Antártica del Hespérides

41 News about the CAVASSOO project and first results appeared in the local journal Faro de Vigo 28 October

42 42 Press release from UEA. Banana Trade to Help with Climate Research A refrigerated cargo vessel in the banana trade is set to contribute to research about climate change, in a new collaboration between the University of East Anglia (UEA) and The Geest Line, a commercial shipping line based in Southampton. As part of a unique Europe-wide project, a commercial vessel, involved in shipping bananas from the Caribbean to this country, will carry scientific equipment on board to take measurements of carbon dioxide in the air and surface seawater. The M/V Santa Lucia will set sail from Southampton on Wednesday 12 December, carrying new computer-controlled instrumentation. "The scientific equipment on board needs to be able to operate without human supervision, for many weeks at a time, and be robust enough to cope with the difficult conditions at sea," said Project Manager, Dr Ute Schuster, of UEA's School of Environmental Sciences and responsible for data collection on the Santa Lucia. "It is exciting that we are now using commercial vessels for the collection of scientific data, since we will be able to get information from the same route over many seasons and years. With the measurements we are able to monitor the changes of carbon dioxide in the marine environment over years to come, and to produce better models of global climate change" she continued. "We have weekly sailings to the Caribbean on a 35 day round voyage, and are pleased to be able to help with the project" said Captain Peter Dixon, shipping director for Geest Bananas. Research partners at the University of Bergen in Norway, the Consenjo Superior de Investigaciones Científicas in Spain and the University of Kiel in Germany will also be installing measuring equipment onto ships that travel other shipping routes, all of them together covering the North Atlantic. The Laboratoire des Sciences du Climat et de l'environnement in France will be involved in the climate modelling.

43 Presentation of project web page The web site of the project can be found at 43