OPERATIONAL OCEAN MODELING AT NCEP/NWS;
TOWARDS A GLOBAL CAPABILITY

A. Mehra1 .and H. L. Tolman1

1EMC/NCEP/NWS/NOAA, Camp Springs, MD, USA

Abstract

Real Time Ocean Forecast System (RTOFS) in the Atlantic Ocean is an National Centers for Environmental Prediction (NCEP ) operational ocean forecast system based on Hybrid Coordinate Ocean Model (HYCOM) on a 1200 x 1684 orthogonal grid with 26 vertical coordinates composed of 21- isopycnal and 5-isolevel coordinates. It is forced with winds, heat flux, evaporation and precipitation every 3 hour derived from NCEP's global atmospheric GFS (Global Forecast System). The model is run once a day. This operational high-resolution ocean forecast system was funded partly by National Ocean Partnership Program (NOPP), as part of the U. S. Global Ocean Data Assimilation Experiment (GODAE), to develop and evaluate a data-assimilative hybrid isopycnal-sigma-pressure (generalized) coordinate ocean model.

National Weather Service (NWS) and National Centers for Environmental Prediction (NCEP ) are uniquely positioned to provide a backbone global ocean modeling and data distribution framework within NOAA due to the operational orientation of, in particular, the NWS. The operational focus of the NWS includes: i) high-reliability computing (99% or higher data availability guarantee); ii) continuous support of computing and data dissemination; iii) liability coverage for official government issued guidance and forecast products; and iv) dissemination of data to the public through Weather Forecast Offices (WFOs). This existing capability could be expanded to provide ocean-oriented products to the public. Such an operational framework will be part of a national US backbone capability provided along with the US Navy and other partners. In this context, the main objective of operational ocean modeling at NCEP/NWS is to provide daily high-resolution global ocean modeling products to the public, possibly with increased resolution at the US coast. The potential applications for these products include (but are not limited to): i) providing boundary data for regional models; ii) providing global data for (downstream) ecosystems models; iii) providing a modeling backbone for ocean modeling as part of coupled weather/climate modeling within the NWS; and iv) providing a starting point for ocean re-analysis, Observing System Simulation Experiments (OSSE) etc.

An additional application for the operational ocean global forecast system is the National Environment Modeling System (NEMS) currently under development at NCEP. NEMS is envisioned as an operational framework for multiple earth science modeling components (atmosphere, air quality, land surface, ocean, wind waves, sea ice) using the Earth System Modeling Framework (ESMF) superstructure to facilitate coupling within these components. Tools and functionality to execute ensembles and nests of these components with common data assimilation modules are also being developed as part of this operational framework.

Number 143 - Session 3

ECCO2: HIGH RESOLUTION GLOBAL OCEAN AND SEA ICE DATA SYNTHESIS

Dimitris Menemenlis1, Jean-Michel Campin2, Patrick Heimbach2, Chris Hill2, Tong Lee1, An Nguyen1, Michael Schodlok1, and Hong Zhang1

1Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, U.S.A.
2Massachusetts Institute of Technology, Camgridge, MA, U.S.A.

Abstract

The Estimating the Circulation and Climate of the Ocean, Phase II (ECCO2) project aims to produce a best-possible, time-evolving synthesis of most available ocean and sea-ice data at a resolution that permits ocean eddies. ECCO2 analyses are obtained via least squares fit of a global, full-depth-ocean, and sea-ice configuration of the Massachusetts Institute of Technology general circulation model (MITgcm) to the available satellite and in-situ data. What sets apart ECCO2 analyses from operational high-resolution ocean data assimilation products is their physical consistency; the analyses do not contain discontinuities when and where data are ingested. ECCO2 analyses are intended to help quantify the role of the oceans in the global carbon cycle, to understand the recent evolution of the polar oceans, to monitor time-evolving term balances within and between different components of the Earth system, and for many other science applications.

A first ECCO2 analysis for the 1992-2007 period has been obtained using a Green's Function approach to estimate initial temperature and salinity conditions, surface boundary conditions, and several empirical model parameters. Data constraints include altimetry, gravity, drifter, hydrographic, and sea-ice data. A large complement of high-frequency and high-resolution diagnostics has been saved; these diagnostics are made available to the scientific community via ftp and OPeNDAP servers at http://ecco2.org. This poster presentation provides a brief overview of this first ECCO2 analysis, of the estimation methodology, of the solution characteristics, and of some early science applications.

Number 50 - Session 4

EVALUATION, VALIDATION AND TRANSITION OF THE 1/12° GLOBAL HYCOM/NCODA SYSTEM

E.J. Metzger1, H.E. Hurlburt1, A.J. Wallcraft1, O.M. Smedstad2, L.F. Smedstad1, J.F. Shriver1, P. Thoppil2, D.S. Franklin2,

1Naval Research Laboratory, Stennis Space Center, MS, USA
2QinetIQ North America / Planning Systems, Inc., Slidell, LA USA

Abstract

A global ocean nowcast/forecast system consisting of the 1/12° global HYbrid Coordinate Ocean Model (HYCOM) and the NRL Coupled Ocean Data Assimilation (NCODA) system is scheduled for transition to the U.S. Naval Oceanographic Office (NAVOCEANO) by the end of 2008. With ~7 km mid-latitude resolution (3-4 km near the poles), the system will depict the location of mesoscale features such as oceanic eddies and fronts, and provide the 3 dimensional ocean temperature, salinity and current structure. The system has been running in real-time since December 2006. On a daily basis, it produces a 5-day hindcast with an NCODA analysis each day up to the nowcast time and then a 5-day forecast. Results may be viewed at http://www.hycom.org.

A series of phased validations/evaluations for the system began in 2007 and continues this year and next. These will be performed relative to the existing operational global ocean nowcast/forecast system running at NAVOCEANO (1/8° Navy Coastal Ocean Model - 1/32° Navy Layered Ocean Model - 1/8° Modular Ocean Data Assimilation System). First year evaluations include examination of mixed layer depth, sonic layer depth, vertical profiles of temperature and salinity, eddy kinetic energy, sea surface temperature and sea surface height at tide gauge stations. Second year evaluations will look at HYCOM as a provider of boundary conditions to nested models, comparisons with drifting buoys, current cross-sections, eddy tracking and ice performance. Validations have been performed on both nowcast and forecast fields. A subset of these results will be presented.

Sea surface height variability (in cm) from Collecte, Localisation, Satellites (top) spanning the period October 1992 - May 2007 and from 1/12° global HYCOM/NCODA (bottom) spanning the period January 2004 - December 2006. Areal average of the satellite data is 7.8 cm compared to 8.4 cm for the simulated output, an indication the mesoscale eddy field is properly represented in HYCOM/NCODA.

Number 29 - Session 2

THE AUSTRALIAN INTEGRATED MARINE OBSERVING SYSTEM

G. Meyers1, IMOS Operators2

1 University of Tasmania, Hobart, Australia
2 See list below

Abstract

The Integrated Marine Observing System (IMOS) is a $92M project established with $50M from the National Collaborative Research Infrastructure Strategy (NCRIS) and co-investments from 10 operators including Universities and government agencies (see below). It is a nationally distributed set of equipment established and maintained at sea, oceanographic data and information services that collectively will contribute to meeting the needs of marine research in both open oceans and over the continental shelf around Australia. In particular, if sustained in the long term, it will permit identification and management of climate change in the marine environment, an area of research that is as yet almost a blank page, studies relevant to conservation of marine biodiversity and research on the role of the oceans in the climate system. While as an NCRIS project IMOS is intended to support research, the data streams are also useful for many societal, environmental and economic applications, such as management of offshore industries, safety at sea, management of marine ecosystems and fisheries and tourism. The infrastructure also contributes to Australia's commitments to international programs of ocean observing and international conventions, such as the 1982 Law of the Sea Convention that established the Australian Exclusive Economic Zone, the United Nations Framework Convention on Climate Change, the Global Ocean Observing System and the intergovernmental coordinating activity Global Earth Observation System of Systems.

IMOS is made up of nine national facilities that collect data, using different components of infrastructure and instruments, and two facilities that manage and provide access to data and enhanced data products, one for in situ data and a second for remotely sensed satellite data. The observing facilities include three for the open (bluewater) ocean (Argo Australia, Enhanced Ships of Opportunity and Southern Ocean Time Series), three facilities for coastal currents and water properties (Moorings, Ocean Gliders and HF Radar) and three for coastal ecosystems (Acoustic Tagging and Tracking, Autonomous Underwater Vehicle and a biophysical sensor network on the Great Barrier Reef). The value from this infrastructure investment lies in the coordinated deployment of a wide range of equipment aimed at deriving critical data sets that serve multiple applications. Additional information on IMOS is available at the website (http://www.imos.org.au).

²The IMOS Operators are Australian Institute of Marine Science, James Cook University, Sydney Institute of Marine Science, Geoscience Australia, Bureau of Meteorology, South Australia Research and Development Institute, University of Western Australia, Curtin University of Technology, CSIRO Marine and Atmospheric Research, University of Tasmania.

Number 94 - Session 4

WATER MASS VARIABILITY IN THE WESTERN NORTH PACIFIC DETECTED IN 15-YEAR EDDY RESOLVING OCEAN REANALYSIS

Y. Miyazawa, R. Zhang1, X. Guo1, 2, H. Tamura1, K. Komatsu3, T. Setou4, J.-S. Lee4, D. Ambe4, A. Okuno4, H. Yoshinari4

1 Frontier Research Center for Global Change/Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan
2Center for Marine Environmental Studies, Ehime University, Matsuyama, Japan
3Guraduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan
4 National Research Institute of Fisheries Science, Fishery Research Agency, Yokohama, Japan.

Abstract

We have created the reanalysis data of the western North Pacific with horizontal high resolution (1/12 ) to describe the meso-scale events related with the Kuroshio-Kuroshio Extension, the Oyashio, and the meso-scale eddies from 1993 to 2007. The products made by an eddy-resolving ocean model combined with the 3-dimensional variational data assimilation well simulated the water mass property, the Kuroshio-Kuroshio Extension, the Oyashio coastal branch, and the warm and cold core rings. From the reanalysis data, we found that both the intensity of the first crest of the Kuroshio Extension and southward intrusion of the Oyashio coastal branch were closely related with the horizontal distribution of both the Oyashio Water and North Pacific Intermediate Water with inter-annual time scale. We also found that the north-south migration of the Kuroshio Extension associated with its regime transitions affected the intensity of the Subtropical Mode Water in the recirculation region.

Number 172 - Session 4

SSS model error statistics for future satellite SSS data assimilation

B. Mourre, J. Ballabrera, A. Aretxabaleta, J. Gourrion, E. GarcíaLadona and J. Font

Institut de Ciències del Mar, CSIC, Pg. Marítim, 3749, 08003 Barcelona Spain

Abstract

Satellite observations of Sea Surface Height and Sea Surface Temperature have largely contributed to the successes of GODAE during the last decade. Future satellite Sea Surface Salinity (SSS) observations, which will soon strengthen this satellite ocean observing system, will hopefully participate to the success of post-GODAE initiatives. The Soil Moisture and Ocean Salinity mission (SMOS) from the European Space Agency, scheduled for launch in early 2009, will initiate the era of satellite SSS observations, quickly followed by the American Argentinean Aquarius mission.

In an ocean analysis and data assimilation context, one of the main challenges will consist in properly integrating these new observations into ocean models. This requires a better understanding of the sources of SSS model error. With that aim, this paper investigates ensemble-generated SSS error covariances integrating uncertainties in the atmospheric forcings.

The ocean model used in this study is a 1/3º regional configuration of the NEMOOPA model implemented over the eastern North Atlantic Ocean. The study period extends from 2000 to 2008. Model output differences due to perturbations in wind stress and precipitations are investigated using an ensemble method. We present the spatial and temporal variability of SSS ensemble variances, as well as the covariances between SSS on the one hand, and three-dimensional temperature and salinity fields on the other hand. The covariances are evaluated for different time windows: full length simulation, seasonal and weekly (comparable to usual assimilation intervals) time scales.

This work is part of the effort conducted at the SMOS Barcelona Expert Center (http://www.smosbec.icm.csic.es/) aiming at contributing to the ground segment of the SMOS mission.

Number 139 - Session 5

EASTERN TROPICAL PACIFIC DATA ASSIMILATION EXPERIMENTS USING ROMS

Á. G. Muñoz1,2, R. Martínez2, R. Pacheco1, J. Nieto2, K. Briones2

1Centro de Modelado Científico (CMC). La Universidad del Zulia. Maracaibo 4004. VENEZUELA
2Centro Internacional de Investigaciones del Fenómeno de El Niño (CIIFEN). Guayaquil. ECUADOR

Abstract

The methodology and results for two data assimilation experiments for the Eastern Tropical Pacific is presented. Using the Rutgers version of the Regional Oceanic Modeling System (ROMS), we have employed an objective analysis for data assimilation from Ecuadorian Navy Cruisers. We have compared the results of our retrospective runs with the corresponding assimilation of ARGOS data. We provide also a unified analysis using both kind of data sources, available via our ftp site. We discuss the use of this methodology for its use in an operational oceanic forecasting system for the western coast of South America.

Number 19 - Session 4

VARIATIONS OF DISSOLVED OXYGEN IN THE SURFACE LAYER OF THE MID-LATITUDE NORTH PACIFIC BASED ON ARGO FLOAT DATA

Y. Nakagawa1, T. Suga1,2, C. Sukigara1, K. Hanawa1, T. Kobayashi2 and N. Shikama2.

1 Graduate School of Science, Tohoku University, Sendai, Japan
2 JAMSTEC-IORGC, Yokosuka, Japan

Abstract

The concentration of oceanic dissolved oxygen and its temporal and spatial distribution depend on physical, chemical and biological processes and vary associated with changes in those processes. Therefore, it is important and useful to analyze the variability of dissolved oxygen for understanding of ocean circulation, biogeochemical fluxes and their variability. Garcia et al. (2005) quantitatively described seasonal variability of dissolved oxygen for global ocean in the upper 100m layer, in which gas exchange with atmosphere and biological activity are predominant, by integrating World Ocean Atlas 2001 data vertically. As a result, they conclude that the behavior of dissolved oxygen in the extra-tropics of the global ocean is determined by changes in oxygen saturation associated with changes in temperature. JAMSTEC and Tohoku University have deployed 16 Argo floats with oxygen sensor in the mid-latitude North Pacific (Kobayashi et al., 2006).

These floats provide vertical profile data of dissolved oxygen with considerable temporal and spatial coverage. Using these data, we examine seasonal variations of dissolved oxygen in the upper 100m layer in more details than Garcia et al. (2005) did.

Argo floats provide temperature, salinity and dissolved oxygen data at sea surface to 500-2000dbar every 3- 10 days. These data are interpolated onto each 1dbar depth by Akima method (Akima, 1970). Oxygen saturation is calculated by temperature, salinity and density, and then percent saturation is calculated by dissolved oxygen and oxygen saturation.

When dissolved oxygen and oxygen saturation are averaged over the upper 100m, seasonal change in dissolved oxygen is similar to that in oxygen saturation, which is essentially the same result as Garcia et al. (2005). On the other hand, when averaged over 0-50m depth and over 50-100m depth individually, it shows undersaturation in 50-100m layer while supersaturation appears in the 0-50m layer in any season. In summer, when oxygen saturation decreases, dissolved oxygen often increases. As a result, great supersatulation occurs in the 0-50m layer. In the 50-100m layer, dissolved oxygen decreases more rapidly than oxygen saturation and thus percent saturation decreases.As described above, changes in dissolved oxygen in the upper 100m layer are not only due to changes in oxygen saturation associated with changes in temperature but also due to combination of different processes occurring at each depth. According to Kobayashi et al (2006), oxygen sensors of Argo floats have negative biases of 0-10µ ?/? in deep layer. Assuming that the biases do not depend on depth and do not change with time, the present signals of seasonal variations and the magnitudes of supersaturation are meaningfully large. We are examining the accuracy of oxygen data taken by Argo floats by comparing them with WOD data; we will mention the result of this examination in the poster.

In the poster presentation, we will also discuss air-sea oxygen exchange, oxygen utilization rate and biological production.

Number 125 - Session 4

CYCLING THE REPRESENTER METHOD WITH NONLINEAR MODELS



Hans E Ngodock, Scott R. Smith and Gregg A. Jacobs

The Naval Research Laboratory, Stennis Space Center, Mississippi (USA)

Abstract

Variational data assimilation with nonlinear models requires tangent linearization, which may be sufficiently accurate only for relatively short time scales. However, for time intervals beyond the scales of nonlinear event development, the tangent linearization cannot be expected to be sufficiently accurate. The representer method would, therefore, not be able to yield a reliable and accurate assimilation solution. However, the method can be implemented for successive cycles in order to solve the entire nonlinear problem. By cycling the representer method, it is possible to reduce the assimilation problem into intervals in which the linear theory is able to perform accurately. For each cycle, the background needed for the tangent linearization is computed by propagating the nonlinear dynamics using the final solution to the linearized assimilation problem from the previous cycle as the initial conditions. This study demonstrates that by cycling the representer method, the tangent linearization is sufficiently accurate once adequate assimilation accuracy is achieved in the early cycles. The outer loops that are usually required to contend with the linear assimilation of a nonlinear problem are not required beyond the early cycles, because the tangent linear model is sufficiently accurate at this point. The combination of cycling the representer method and limiting the outer loops to one significantly lowers the cost of the overall assimilation problem. In addition, this study shows that weak constraint assimilation is capable of extending the assimilation period beyond the time range of the accuracy of the tangent linear model. That is, the weak constraint assimilation can correct the inaccuracies of the tangent linear model and clearly outperform the strong constraint method. Preliminary examples using the Lorenz attractor and a reduced gravity ocean model will be presented.

Number 22 - Session 4

Assimilation of Lagrangian data

Maelle Nodet

Universit_e de Grenoble / INRIA

Abstract

In the context of the GODA Experiment (Global Ocean Data Assimilation) an increasing number of ocean data are available. To make the most of the information contained in these data is a fundamental issue for oceanographers, in order to improve forecasts, models, climatology, etc. The ocean is mainly observed at the surface, thanks to satellite remote-sensing. However, the Argo Program provides valuable in-situ observations: temperature and salinity profiles. The Argo floats are drifting at depth (around 2000 meters). Every ten days they perform a vertical profile of temperature and salinity. Then they come to the surface to transmit their data by satellite and the GPS gives their positions.

So a new type a data is also available, namely the positions of the oats every ten days. A variational method has been implemented to address the assimilation of these Lagrangian data. The 4D-Var formalism allows us to assimilate directly the position thanks to a relevant observation operator, which is the non linear mapping between the Lagrangian observations and the states variables of the ocean (which are the horizontal velocity, the temperature and the salinity).

This method has been implemented (see [1]) into an idealised 3D Primitive Equation OGCM, namely OPA. We present various numerical results: validation, comparison to an Eulerian approach, sensitivity studies.

For example the figure presents horizontal section of the velocities u and v at the surface for the true state (used to simulate the observations), the background (without assimilation) and the assimilated state, for 1000 oats drifting at 1000 meters depth, whose positions are sampled every day.

Number 14 - Session 5

RENANALYSIS OF EXTREME OCEANIC EVENTS IN THE TASMAN SEA

P. R. Oke1, D. A. Griffin1

1Centre for Australian Weather and Climate Research: a partnership between CSIRO and the Bureau of Meteorology, Hobart, Australia
Abstract

During the Austral summer of 2006-07 an extreme cold-core eddy, with a sub-surface temperature anomaly in excess of 8°C at 200 m depth, developed off the New South Wales (NSW) coast in the Tasman Sea. The formation and intensification of the eddy followed a series of strong and sustained wind-driven upwelling events along the NSW coast. Based on analysis of a data assimilating, eddy-resolving ocean model, we suggest that the magnitude of the eddy may have resulted from a mixed barotropic-baroclinic instability resulting from the offshore advection of cold upwelled waters that originated in the vicinity of 32°S, where the East Australian Current (EAC) separates from the coast. We investigate the three-dimensional circulation around the eddy and find that to the north and east there is a strong downwelling flow; and to the south and west there is a strong upwelling flow. This three-dimensional circulation 1) is not consistent with common conceptual models of the vertical circulation associated with cold-core eddies; 2) may provide an efficient mechanism for the nutrient enrichment of the surface waters around the eddy and 3) may explain the anomalously high level of chlorophyll-a in satellite ocean colour images of the event.

Number 17 - Session 3

DATA ASSIMILATION FOR LIMITED-AREA OCEAN MODELS

P. R. Oke1, M. Herzfeld2

1Centre for Australian Weather and Climate Research: a partnership between CSIRO and the Bureau of Meteorology, Hobart, Australia
2CSIRO Marine and Atmospheric Research, Hobart, Australia

Abstract

The provision of high-resolution, short-range ocean forecasts often involves the integration of a limited-area ocean model. Such models typically derive initial conditions and boundary values from a coarser large-scale model. Sequential data assimilation in limited-area models poses the challenge of updating the model's initial conditions and boundary values in a consistent manner. A simple, computationally efficient approach to this problem is developed here. Suppose a 7-day forecast is to be performed for the time period T=0-7, following a short period of initialization for the period T = -4-0. The method proposed here simply uses a sub-domain of the large-scale model as the background field, using daily mean fields for each day in the period T<0. Each background field is combined with observations, yielding a series of regional analyses on the sub-domain for each model day for T<0. Quasi-analyses are also produced for T>0 by combining the daily mean forecasts from the large-scale model with some fraction of the analysis increment on day T=0 (e.g., 80% at T=1; 60% at T=2 etc.). Thus the quasi-analyses for T>0 "remember" some of the increments from the period T<0. The limited-area model is subsequently integrated, drawing on the regional analyses for initial conditions (at T=-4) and boundary values (T=-4-7). The limited area-model is nudged towards the regional analysis during the period of initialization (T<0) using a time-dependent relaxation time-scale that is shortest around T=0. Following the initialization, a forecast is integrated for the period T=0-7 with no nudging, using the updated quasi-analyses for boundary values.

The benefits of the data assimilation approach described above, for limited-area ocean models, are demonstrated through the application of the Bluelink Ocean Data Assimilation System (BODAS) to the ocean component of the Relocatable Ocean Atmosphere Model (ROAM). BODAS is an ensemble optimal interpolation system that underpins Australia's operational ocean forecast system. ROAM is a high-resolution, relocatable, coupled atmosphere-ocean model that is configured from a graphical user interface, allowing the user to specify the model region, resolution and period of interest. The ocean component of ROAM is nested within either, the Bluelink forecast or reanalysis system, that includes a global general circulation model with 1/10o resolution around Australia. The robustness of the system's performance is demonstrated through a series of case studies using different model domains, spanning a broad range of different oceanographic phenomena. The relative forecast skill of each modelling system, including the global system, the non-assimilating ROAM and the assimilating ROAM, is evaluated through comparisons with independent observations. In each case, the forecast skill of the assimilating limited-area model consistently exceeds that of the free-running limited-area model and the large-scale model.

Number 165 - Session 2

Monitoring the exchanges between the Atlantic and the Arctic across the Greenland-Scotland Ridge

Hansen Bogi1, Hjálmar Hátún1, Svein Østerhus2, Detlef Quadfasel3, Steingrímur Jónsson4, Héðinn Valdimarsson4, Bill Turrell5, Sarah Hughes5, Toby Sherwin6

1 Faroese Fisheries Laboratory, Tórshavn, Faroe Islands
2 Bjerknes Centre for Climate Research, University of Bergen, Bergen, Norway
3 Universität Hamburg, Zentrum für Meeres- und Klimaforschung, Hamburg, Germany
4 Marine Research Institute, Reykjavík, Iceland
5 Marine Laboratory, Fisheries Research Services, Aberdeen, UK
6 Scottish Association for Marine Science, Oban, UK

Abstract

The Arctic Mediterranean (the Arctic Ocean and the Nordic Seas) receives a continual input of warm water from the Atlantic Ocean. This Atlantic inflow supplies about 90% of the total inflow to the region and keeps large areas much warmer than they would otherwise have been and free of ice. In the Arctic Mediterranean, the imported Atlantic water is cooled and freshened. It returns to the Atlantic, partly in near-surface layers on both sides of Greenland, and partly as a deep overflow of cold, dense water that crosses the Greenland-Scotland Ridge. After crossing the ridge, the overflow water entrains ambient water and sinks into the depths of the Atlantic where it becomes the main contributor to the North Atlantic thermohaline circulation. Some climate models indicate that anthropogenic climate warming may induce a substantial weakening of these flows in the 21st century (IPCC, 2007). If this were to happen, large ecological and societal impacts may be expected (ACIA, 2005) and monitoring the characteristics and intensity of the exchanges is therefore a high priority task. To fulfill this task, a group of European marine research institutes have implemented a monitoring system, including regular research vessel cruises and quasi-permanent moorings with self-recording current meters and other instruments. This system has been developed over the last decade and has provided time-series of the characteristics and intensity of the exchanges. The field-work has mainly been carried out by national fisheries research institutes, but a number of other European institutes have been involved. Funding has been supplied by grants from national, Nordic, and European research funds, most recently within the ASOF (Arctic / Subarctic Ocean Fluxes) project, which has been funded by the European Framework Programme.