During the years 1999-2003 the Canadian Physical Oceanographic community continued the type of comprehensive and varied studies of oceanographic processes that were outlined in previous reports in 1999 and 1995, by Michael Foreman and Yves Gratton.

The academic research community is spread across Canada from Memorial University in Newfoundland to the University of Victoria in British Columbia. Internet Links to these groups are listed below.

Université Laval

http://www.bibl.ulaval.ca/ress/oceano-a-laval.html

University of Alberta

http://easweb.eas.ualberta.ca/department/welcome.cfm

Université du Québec à Rimouski

http://wer.uqar.qc.ca/ismer/angl.htm

http://wer.uqar.qc.ca/ismer/fran.htm

McGill University

http://www.mcgill.ca/meteo/

Memorial University

http://www.physics.mun.ca/Ocean/physocean.htm/

University of British Columbia

http://www.eos.ubc.ca/

Dalhousie University

http://www.dal.ca/~wwwocean/index.html

Royal Military College

http://www.rmc.ca/academic/physics/index_e.html

Within Quebec the GIROQ organization also coordinates oceanographic research between Laval and McGill Universities.

http://www.bio.ulaval.ca/GIROQ/A_pres.htm

These academic organizations are spread among a wide variety of university departments, such as Physics, Earth and Ocean Sciences, and Meteorology & Oceanography. Only occasionally is Oceanography a department itself. This wide distribution is an indication of the varied affiliations that physical oceanographers have formed over the past century of research.

Fisheries and Oceans Canada hosts the main government physical oceanographic research centres in Canada, at locations spread across the country.

A general listing of all centres is at the following internet site

http://www.dfo-mpo.gc.ca/science/facilities-installations/facilities_e.htm

Listed below are Internet links to some of these labs with research activities in physical oceanography.

Institute of Ocean Sciences, BC

http://www-sci.pac.dfo-mpo.gc.ca/sci/facilities/ios_e.htm

Bedford Institute of Oceanography, NS

http://www.mar.dfo-mpo.gc.ca/science/ocean/home.html

Maurice Lamontagne Institute, PQ

http://www.qc.dfo-mpo.gc.ca/iml/en/intro.htm

Central and Arctic Region

http://www.dfo-mpo.gc.ca/regions/central/science/research-recherche/index_e.ht

A major change that has taken place in the way research is conducted since the last IAPSO review has been the increased access to real-time and archived oceanographic observations over the Internet. A rich collection of oceanographic data and products are provided by a number of research institutes and government agencies, in both English and French, specifically;

Marine Environmental Data Service (MEDS)

http://www.meds-sdmm.dfo-mpo.gc.ca/

Bedford Institute of Oceanography (BIO)

http://www.mar.dfo-mpo.gc.ca/science/ocean/home.html

http://www.iob.gc.ca/welcome-f.html

L'Observatoire du Saint-Laurent (OSL)

http://www.osl.gc.ca/

Institute of Ocean Sciences (IOS)

http://www-sci.pac.dfo-mpo.gc.ca/osap/data/default_e.htm

http://www-sci.pac.dfo-mpo.gc.ca/osap/data/default_f.htm

Topics covered by one of more of the locations cited include ocean contaminants, real time oceanographic observations and archival data from moored and drifting buoys, archived observations of ocean profiles and currents, tides and water levels, including tidal and drift prediction models, monitoring of oceanography, biology and ocean chemistry, ocean forecast models of sea ice, ocean currents, waves and water levels, thermosalinograph observations from research vessels and commercial shipping, coastal temperature from moored thermographs, and remote sensing of sea-surface temperature and ocean color.

We present a review of a selection of research activities and publications over the past four years, with primary focus on studies within the three oceans surrounding Canada, and a short section on global ocean research. Much of this material is based on input from Canadian oceanographers who responded to our request for research summaries for this report: Ken Denman, Yves Gratton, Richard Greatbatch, Doug Gregory, Guoqi Han, Charles Hannah, Greg Holloway, William Hsieh, Peter Jones, Rob Macdonald, Bill Merryfield, Paul Myers, Lawrence Mysak, Simon Prinsenberg, Frank Whitney, Francois Saucier and Dan Wright.

The summary is organized by oceans (Arctic, Atlantic, Pacific, Global) with a final section to describe a few process studies.

Polynyas are areas of open water or thin ice in the ocean that occur where one would expect to find thick sea ice. The North Water Polynya lies between Canada and Greenland at the extreme northern end of Baffin Bay. Exploited by Inuits for at least 5,000 years, probably known to Vikings in the 13^th century, and revealed to Europe by William Baffin in 1616, the North Water is considered by several specialists as the most productive ecosystem north of the Arctic circle.  Its ice-free and highly productive waters serve as feeding, mating, spawning and over-wintering grounds for huge populations of key species of birds and mammals. Scientists suspect that the North Water is a focal point for the intense production of the planktonic herbivores that ensure the transfer of the solar energy fixed by microalgae to Arctic cod, seals, polar bear and native man. 

The functioning of the North Water ecosystem, its role in the overall Arctic biota and its potential response to global warming are poorly understood. The NOW Research Network funded by the Natural Sciences and Engineering Research Council of Canada (NSERC) brought together Canadian and foreign expertise in Arctic oceanography to study and model the climatic and oceanographic mechanisms of formation of the North Water, the biological production taking place within and around its ice boundaries, and the fate of this production. (Louis Fourtier http://www.fsg.ulaval.ca/giroq/now/wel.htm)

Physical processes in this region are described in a special issue of ATMOSPHERE-OCEAN devoted to an intensive scientific study of this region. A summary of the preface to this issue (Barber et al. (2001a)) is presented below. Melling et al. (2001) show that the polynya is dominated by a strong southward flow of cold water and ice from the Arctic Ocean. A branch of the West Greenland Current provides a modest northward flow of warm water up the eastern side of Baffin Bay. This warm northward flow is diverted by the complex bathymetry near the Carey Islands, where it loses much of its heat through re-circulation and isopycnal mixing with the Arctic outflow. Upwelling near the Greenland coast, forced by Ekman transport, brings the warm water to the base of the turbulent surface layer of the polynya, where it is entrained. Clausi (2001) shows how various image-processing algorithms can be used to estimate the type of sea ice occurring within and around the North Water Polynya. Yankel et al. (2001) use RADARSAT data to estimate the thermodynamic state of first year sea-ice based on the seasonal evolution of the microwave scattering coefficient. Vincent and Marsden (2001) show how data from NOAA’s AVHRR sensor can also be used to understand the timing of the dissolution of the ice-bridge in Nares Strait, adjacent to this polynya. Hanesiak et al. (2001) compare various parametrization schemes for shortwave exchange over the polynya. Selection of optimal schemes is seen as a contribution to future thermodynamic sea-ice modelling and regional scale atmospheric modelling studies. Vincent et al. (2001) show how optical remote sensing data from the AVHRR sensor can also be used to estimate ice motion. Mundy and Barber (2001) show that a distinct "icescape" exists for this polynya. Statistical clustering of ice types showed that regionally and seasonally specific clusters of a mixture of ice types are created depending on dominance of latent or sensible heating in the creation and maintenance of the polynya. Barber et al. (2001b) use historical satellite microwave observations at weekly temporal resolution for the period 1978 to 1996 to determine typical conditions, and to examine modes of variability.

Significant scientific papers are also published in other journals. Deming, Fortier, and Fukuchi (2002) have edited an issue of Deep Sea Research II devoted to several polynya studies. In this issue, Ingram et al. (2002) present an overview of physical oceanographic processes. Miller et al. (2002) examine deep-sea fluxes of carbon in the North Water Polynya. Mei et al. (2002) describe physical control of spring-summer phytoplankton dynamics in the North Water Polynya. Bacle, Carmack, and Ingram (2002) discuss water column structure and circulation under the North Water Polynya during spring transition in April-July 1998.

In other publications, Belzile et al. (2000) examine ultraviolet attenuation by dissolved and particulate constituents of first-year ice during late spring in an Arctic polynya. Fisk, Hobson and Norstrom (2001) note the influence of chemical and biological factors on trophic transfer of persistent organic pollutants in the Northwater Polynya marine food web. Tremblay, et al. (2002) describe the impact of large-scale Arctic circulation and North Water Polyna on nutrient inventories in Baffin Bay

Bjornsson et al. (2001) simulate polynyas in a high-resolution dynamic-thermodynamic sea ice model and examine their parameterization using flux models.

Through the 1990s a series of papers described the water masses and their circulations within the Arctic Ocean (Jones, 2001). Some of the earlier work describing the circulation of the Atlantic Layer and deeper waters (~200 m to 1700 m) was amplified and clarified by an extensive survey of the Nansen, Amundsen, and Makarov basins (Schauer et al., 2002). A very recent contribution to this evolving picture was recognizing that the Barents Sea was the source of halocline water that circulated close to the Eurasian shelf, then expanded to cover most of the Canada Basin south of the Chukchi Cap, then proceeded along the North American coast to exit the Arctic Ocean through Nares and Fram straits (Rudels et al., 2003). This result has significance in the context of the potential redistribution of nuclear waste dumped in the Barents Sea by the former Soviet Union as well as the catastrophe that would occur if Lake Karachi in the Urals ever released its water into tributaries whose waters eventually reach the Barents Sea.

A simple box model exploiting CFC measurements was devised to describe the ventilation of the Arctic Ocean. One conclusion was that deep water formation via shelf-slope plumes within the Arctic Ocean boundaries could indirectly account for as much as 40% of the Denmark Strait Overflow (Anderson et al. 1999). Tracers that delineate river runoff in the Arctic Ocean showed that the runoff front (5% river runoff in surface waters) in the Eurasian Basin varied significantly during the period 1991-2001. The variation could be related to the NAO with a 5-year lag (Anderson et al., 2003). The excess evaporation over precipitation in the Atlantic Ocean largely falls as rain in the Pacific Ocean, resulting in Pacific surface waters that are fresher than those of the North Atlantic. The Arctic Ocean provides a pathway for the return flow of this freshwater to the North Atlantic. This return flow Pacific water has been traced to the deep convection regions east of Greenland as far south as Demark Strait and west of Greenland to the Labrador Sea as far south as the Tail of the Banks (Jones et al., 2003).

The Arctic Ocean has been considered to be a sink for atmospheric CO₂. A study to assess the "biological pump" shows this pathway to be extremely small compared to the global sink (Anderson et al., 2003). A second study showed that shelf regions do provide a significant sink at least on a regional scale (Fransson et al. 2001).

McLaughlin, et al. ( 2002) examine the effects of upstream events on southern Canada Basin waters using physical and geochemical data collected at one location between 1989 and 1995. These events included Atlantic layer warming, relocation of the Atlantic/Pacific water mass boundary, and increased ventilation of boundary current waters. Macdonald, Harner, and Fyfe (2003) review and discuss recent physical changes in the Arctic forced by sea level pressure drop over the pole (the Arctic Oscillation). They present the ramifications for contaminant pathways – in much of it, contaminants can be viewed as tracers of change.

Macdonald (2000) studies interaction of freshwater runoff and sea ice formation on Arctic shelves using the Mackenzie shelf of the Beaufort Sea as a model. Arctic shelves share the common physiognomy of strongly seasonal positive estuaries separated from the interior ocean during winter by flaw leads over the mid shelf where enhanced ice growth produces a negative estuary. In the Arctic, there are two sources of manipulation of freshwater: runoff and sea ice formation/melt. Buoyancy supplied by the positive estuary has the potential to stall convection in the negative estuary and thereby control whether brine is permanently separated from sea ice (penetrative convection) or whether it simply remains in the surface water to remix with ice melt in summer. This control can have a large effect on thermohaline circulation and the freshwater balance of the Arctic Ocean. The close juxtaposition of positive and negative estuaries in the Arctic provides a delicately timed balance that will be particularly sensitive to global change.

During project SHEBA (Surface Heat Balance in the Arctic), thin ice and freshening of the Arctic Ocean surface in the Beaufort Sea led to speculation that perennial sea ice was disappearing. Macdonald et al. (1999) collected salinity, d ¹⁸O and Ba profiles near the initial SHEBA site and, in 1997, ran a section out to SHEBA. Resolving fresh water into runoff and ice melt, they found a large background of Mackenzie River water with exceptional amounts in 1997 explaining much of the freshening at SHEBA. Macdonald, McLaughlin, and Carmack (2002) describe salinity and d ¹⁸O measurements made along this shelf-basin section out to the initial SHEBA site and along the SHEBA drift track to determine the sources and amounts of surface freshening observed in the Canada Basin. Abundant amounts of Mackenzie River runoff, up to 8 m of inventory in the top 40 m of the water column, were found in the interior ocean at the start of the SHEBA program. Toward the eastern edge of the Northwind Ridge, enhanced amounts of runoff were also observed, originating either from the Mackenzie River or from the Yukon River (Alaska Coastal Current). For sea-ice melt, the shelf-basin section revealed net ice formation (brine) from the continental margin out to 200 km, but net sea-ice melt under the permanent pack. The sea-ice melt inventory in the upper water column reflected the annual cycle of sea-ice formation and melting (~ 1 m). However, this seasonal variability was dwarfed by large spatial variation.

Macdonald, Carmack, and Paton (1999) distinguish estuaries that produce substantial amounts of ice from those that do not. They show how the oxygen isotope composition (d ¹⁸O) in landfast ice at the end of winter provides a record of surface water properties throughout winter. Two arctic estuaries are contrasted: the Mackenzie estuary which faces directly onto a broad, open shelf and the Husky Lakes estuary which comprises a series of basins that exchange with one another and the shelf through narrow channels. The method of converting records of d ¹⁸O in ice cores to surface water salinity as a function of time throughout winter is outlined. The Mackenzie estuary has a large winter inflow which spreads beneath the ice as a plume overwhelming brine production by sea-ice formation and thereby shutting down convection. Using only ice records, confident estimates can be made of the rate and direction of plume spreading. In contrast, the Husky Lakes estuary is supported only by a local, truly Arctic, drainage basin which becomes frozen in winter. The landfast ice in this system records a large, and relatively static, horizontal gradient in surface salinity. Small variations in d ¹⁸O with time, evident in all ice cores, are found to correlate between different sites, suggesting coherent displacements of surface water by 10 - 15 km as might be produced by seiching.

Melling (2002) describes ice coverage in the Canadian Arctic Archipelago based on historical and recent observations, including more than 100,000 holes drilled in the 1970s. He observed that the flux of ice through the Sverdrup Basin is intermittent. It is interrupted annually by the onset of land-fast conditions and interannually in late summer and autumn by factors that are not understood. Van der Baaren and Prinsenberg (2002) determine geostrophic transport estimates in the Canadian Arctic Archipelago. Hamilton, Prinsenberg and Malloch (2002) describe moored current meter and CTD observations from Barrow Strait, 1998 – 1999. Prinsenberg and Peterson (2002) examine variations in air-ice drag coefficient due to ice surface roughness. Prinsenberg, Holladay, and Lee (2002) measure ice thickness with a fixed-mounted helicopter electronic-laser system.

Li et al. (2002) examine a large data base for a -HCH together with multimedia models, showing that this chemical exhibits classical ‘cold condensation’ behaviour. The surface water of the Arctic Ocean became loaded between 1950 and 1990 because atmospheric transport of a -HCH from source regions to the Arctic was rapid and because a -HCH partitioned strongly into cold water there. Following emission reductions during the 1980s, a -HCH remained trapped under the permanent ice pack with the result that the highest oceanic concentrations in the early 1990s were to be found in surface waters of the Canada Basin. Gobeil et al. (2001a) find that contaminant Pb in sediments underlying boundary currents in the Arctic Ocean provides an image of current organization and stability during the past 50 years. The sediment distributions of Pb, stable Pb isotope ratios, and ²¹⁰Pb in the major Arctic Ocean basins reveal close coupling of the Eurasian Basin with the North Atlantic during the 20^th century. They indicate that the Atlantic water boundary current in the Eurasian Basin has been a prominent pathway. Gobeil et al. (2001b) have found evidence of recent large-scale change in redox conditions in Arctic Ocean basin sediments, in profiles of solid phase acid volatile sulfide (AVS), manganese, and rhenium. The most likely origin for such widespread change is the ice climate. Macdonald et al. (2000) summarise recent studies of contaminants under the Canadian Northern Contaminants Program (NCP). They highlight new information developed under the NCP on the sources, occurrence and pathways of contaminants (organochlorines, Hg, Pb and Cd, PAHs, artificial radionuclides).

Several scientists investigated deep ocean mixing and advection in the Arctic, to examine details of double diffusion (Merryfield, 2002) and the physics of intrusions (May and Kelly, 2001; Walsh and Carmack 2002).

A major new initiative is CASES (Canadian Arctic Shelf Exchange Study) program in the Beaufort Sea. The general objective is to determine the influence of climatic changes on polynyas opening and closing in the Canadian Polar waters, and on their carbon budget. The CASES field work is barely beginning with the main event being a one-year (2003-2004) wintering of a Canadian Iceabreaker in the Bathurst Polynya.

The Arctic Ocean is largely covered by sea ice, a feature that dominates its impact on our world, and our research into its physical oceanographic processes. The possible retreat of this ice in future years, as part of a global warming scenario has been considered by several scientific projects. Canadian scientists have contributed many key papers to this topic. Mysak, (2001a) and Mysak, ( 2001b) describe patterns of arctic circulation, and the role of sea ice in climate variability and change. Holloway and Sou (2002) have examined past interannual changes in sea ice coverage and thickness in the Arctic Ocean, through application of numerical models forced at ocean surface with past atmospheric conditions. They observe that previous observations of 40% loss in Arctic ice volume were incorrect. These previous results were based largely on measurements from submarines that did not include observations from the Canadian EEZ, a region into which the Holloway and Sou (2002) study determined that much of the "missing" ice advected. This paper is part of a growing contribution by Canadian oceanographers to numerical modelling studies of this ocean, including intercomparison of many Arctic Ocean climate models presently in use by the global scientific community (Proshutinsky, et al., 2001; Steele, 2001; Steiner et al, 2003).

Saenko, Flato and Weaver (2002) focus on two model parametrisations of sea-ice related processes. The first involves the ocean component in which they generalize a recently developed parametrization of brine rejection during sea ice formation for use in a multi-category sea ice model (i.e. one that resolve the thickness distribution function). The model employs explicit subsurface mixing of brine-enriched surface waters, resulting from sea ice growth. The second part of this paper focuses on the sea-ice component. They perform a series of stand-alone sea-ice model experiments comparing a recently developed multi-layer energy-conserving thermodynamic scheme with the simplified scheme used in many existing climate models. (Source: http://www.climate.uvic.ca/ people/weaver/documents/oleg)

Duffy, Eby and Weaver (2001) show that results of an atmosphere-ocean sea-ice model are sensitive to treatment of salt rejected during formation of sea-ice. They show improved results in a simulation that mixes rejected ice into the ocean down to a level dependent on local density gradients, rather than mixed through the upper layer of the ocean model.

Observations from sonar data have suggested enhanced melting of thick ridged ice relative to level ice. Schramm, Flato, and Curry (2000) apply 2-dimensional models to examine several physical processes that might account for this difference. In another paper on ice ridges, Bellchamber, Melling and Ingram (2002) model the evolution of draft distribution in the sea ice pack of the Beaufort Sea.

Kreyscher et al. (2000) compare a hierarchy of ice rheologies as part of the Sea Ice Model Intercomparison Project (SIMAP), finding that the viscous-plastic rheology yields the most realistic simulation, and its increased computational cost is minor compared with the atmospheric and oceanic model components in global climate simulations.

Arfeuille, Mysak, and Tremblay (2000) apply a thermodynamic-dynamic sea-ice model to study the interannual variability of the Arctic sea-ice cover during the 41-year period 1958-98. They focus on analyzing the variability of the sea-ice volume in the Arctic Basin and the subsequent changes in sea-ice export into the Greenland Sea via Fram Strait. Large sea-ice export anomalies are generally preceded by large volume anomalies formed along the East Siberian coast due to anomalous winds which occur when the Arctic High is centered closer than usual to this coastal area. The overall results from this study show that the Arctic Basin and its ice volume anomalies must be considered in order to fully understand the export through Fram Strait. This research was continued by Armstrong, Tremblay, and Mysak (2003). The mechanisms responsible for producing SIC anomalies were evaluated by studying the coupled variance (using the Singular Value Decomposition (SVD) method) between simulated SIC anomalies and the ice speed and air temperature anomalies. To execute this validation, a 49-year (1949-97) simulation (including a 9-year spin-up period) of the Arctic and peripheral sea-ice cover using daily varying winds and monthly mean air temperatures was produced. Results from the SVD analysis of this output show that the main sources of variability in the peripheral seas are associated with the variation in the strength of the Arctic High; in the East Siberian and Laptev seas, the strengthening and weakening of the Transpolar Drift Stream also play an important role. Over the entire Arctic domain, surface air temperature anomalies are negatively correlated with sea ice anomalies. Finally, the record minima in the East Siberian Sea is well produced in the simulation. This work was an outgrowth of the MSc thesis research (1999-01) of Anne Armstrong and was supported by the Arctic Node of the Canadian Institute for Climate Studies.

Merryfield, W. J., and G. Holloway, (2003) adapt an accurate advection algorithm originally developed for atmospheric chemistry applications to sea-ice modelling and tested in an Arctic ice-ocean model. The method was found to improve the representation of ice edges, and under climatological forcing yielded approximately 7% greater ice volume than the commonly used centered difference and upstream methods.

Several studies have linked changes in nutrient concentrations in the mid-Gulf of Alaska to temperature changes and El Nino. Both temperature and nutrient levels are considered important for survival of marine fish. Whitney and Freeland (1999) review the oceanographic properties of the Gulf of Alaska near Ocean Station Papa based on 42 years of observations. El Niño events influence this area by transporting heat northward. Warming persisted during the prolonged El Niño of the early 1990s, resulting in a reduced nutrient supply to surface waters during winter mixing. Whitney and Welch (2002) describe nutrient surveys of the Gulf of Alaska, from 1997 through 1999, showing that coastal waters of British Columbia and southern Alaska experienced nitrate depletion each spring and summer. Detailed sampling off the southwest coast of British Columbia revealed that 1998 nitrate levels were only half the average of that during the 1970s winter. However, conditions changed dramatically during the 1999 La Niña. Winter nutrient levels increased and summer upwelling returned.

Cummins and Lagerloef (2002) compare a numerical model of the Gulf of Alaska with observations of thermocline depth at Station Papa. The model is forced by monthly wind stress curl anomalies derived from the NCEP re-analyses for the period 1948-2000. The leading mode of their model’s response has the signature of the Pacific Decadal Oscillation. Variability in depth on interannual to inter-decadal time scales is largely an integrated response to local Ekman pumping. Gower (2002) examines temperature, wind and wave climatologies and trends from marine meteorological buoys in the NE Pacific. Bograd et al. (1999) describe the mid-gulf surface flow based on Lagrangian drifters.

Freeland (2002) investigates the northward heat flux along the Canadian continental margin, passing across Line-P. (Line-P extends from the western end of Juan de Fuca Strait to Ocean Station Papa). Canadian ships sample Line-P about three to four times a year on their way to OSP, and additional samples were taken during the 1997/98 El Niño. By combining these ship-based measurements with altimetry observations of sea surface height anomaly (SSHA) along this line, he is able to compute this northward heat flux in the upper 500 m across the ten Line-P stations closest to shore for the period 1997 to 1999, a time interval that includes the 1997/98 El Niño. Most of the temperature and velocity structure lies in the top 500 m; therefore little transport is expected at greater depths. For the period of November 1997 to February 1998, inclusive, he determines a maximum heat transport of 15 to 26 terawatts, (1 terawatt=10¹² watts). This heat transport is significantly greater than northward heat flux across this line due to surface Ekman flux, and much greater than observed during normal, non-ENSO winters.

Castro et al. (2002) present a description of North American coastal oceanographic events during the 1997/98 El Niño. Chavez et al. (2002) provide a introduction to a special issue of Progress in Oceanography on this topic. Mackas and Galbraith (2002) describe zooplankton community composition along the inner portion of Line P during the 1997-98 El Niño event. Wong et al. (2002) present the variability in the distribution of surface nutrients and dissolved inorganic carbon in the northern North Pacific as influenced by El Niño. Wong and Crawford (2002) show that fluxes of particulate inorganic and organic carbon to the deep subarctic Pacific correlate with El Niño. Subbotina and Thomson (2001) describe the spectral characteristics of sea level variability along the West Coast of North America during the 1982-83 and 1997-98 El Niño events.

William Hsieh and collaborators at the University of British Columbia have applied neural network techniques to prediction of El Niño and Southern Oscillation events in a series of papers (Hsieh, 2001; Tang and Hsieh, 2003a; Tang and Hsieh, 2003b; Wu and Hsieh, 2002). Hsieh, Tang and Garnett (1999) have discovered a link between Pacific sea surface temperatures and Canadian prairie wheat yield. Wu and Hsieh (1999) undertook a modelling study of the 1976 climate regime shift in the North Pacific Ocean.

Canadian scientists discovered several new ocean processes in the waters of the Gulf of Alaska, largely based on sea level measurements by radar sensor on several satellites. Tabata (1982) had previously defined anticyclonic eddies using ship-based and drifter observations. Thomson and Gower (1998) and Gower and Tabata (1993) determined their motion and geographical extent. But near-real time studies awaited the launch of TOPEX/POSEIDON altimetry satellite, and processing of its measurements.

Each eddy contains a core of anomalously low density water, and produces an upward doming of the sea surface detectable by satellite altimetry. Crawford and Whitney (1999) describe five years of almost daily altimeter measurements of these eddies, which acquired the names Sitka (for eddies formed along the Alexander Archipelago of Alaska) and Haida (for eddies formed along the Canadian Queen Charlotte Islands.) They determined that these eddies are almost always formed in winter along the eastern continental margin of the Gulf of Alaska, north of 51N, rotate anticyclonically, and drift mainly westward for several years. The largest of Haida Eddies carried up to 6,000 km³ of coastal water into mid-gulf. Both eddies suppress the pycnocline in their centre. Crawford, Cherniawsky, and Foreman (2000) examine TOPEX/POSEIDON satellite altimetry observations of eddies in the Alaskan Stream along the southern continental margin of the Aleutian Islands, after extracting tides using the Cherniawsky et al. (2001) method. They observe these eddies to drift between the Aleutian Trough and Alaskan shelf for periods of up to 3 years. At least one of these eddies was a Sitka Eddy that became trapped along this continental margin. All eddies are anticyclonic, and some reach the theoretical maximum height-to-diameter ratio for planetary anticyclonic eddies.

A series of cruises to study these eddies took place between 1999 and 2001 during the program by Canadian Scientists to measure water properties along Line-P in the Gulf of Alaska between Juan de Fuca Strait, and Station Papa at 50N, 145W, as well as on cruises devoted to study only Haida Eddies. All measurements were taken from the Canadian Coast Guard Ship John P. Tully, which was guided to the eddies by near-real-time altimetry maps generated by Robert Leben and colleagues at a University of Colorado Internet site. Whitney and Robert (2002) determine that Haida Eddies carry relatively high levels of macro-nutrients into the gulf, mainly in their central waters, and likely inject nutrients up into the euphotic zone from below. Mackas and Galbraith (2002) observe that the zooplankton community in the eddies is a mixture between shelf-slope species and subarctic ocean species. The former are transported from the near-shore formation region, whereas the latter colonize the eddy from the sides and below. Crawford (2002) compares satellite altimetry observations with ship-based studies to determine that Haida-1998, the largest Haida Eddy ever observed, was almost totally baroclinic, with very little barotropic rotation. All eddies examined are deep, suppressing isotherms in their centres at depths of 1000 metres or more. Crawford et al. (2002) gather observations of Haida Eddies as they form along the West Coast of the Queen Charlotte Islands, noting that most form at the southern tip of these islands in winter. Using detailed altimetry observations (de-tided with the Cherniawsky et al. (2001) method and the Foreman et al. (2000) model) and water property measurements they determine that the source waters are outflowing currents from Hecate Strait in winter, and the adjustment of this outflow as its enters the Gulf of Alaska might be the source of eddy rotation, through a process similar to that examined in this location by Thomson and Wilson (1987) and in a rotating tank by Cenedese and Whitehead (2000). Sitka Eddies are likely set up by baroclinic instability processes as modelled by Melsom et al. (1999).

Additional studies of coastal waters have been linked to satellite altimetry observations. Foreman et al. (2000) have developed a high-resolution, data-assimilating, numerical tidal model of the northeast Pacific Ocean, including the continental shelves. This model, when combined with the Cherniawsky et al. (2001) tidal analysis of altimetry observations, enables very accurate de-tiding of all satellite altimetry observations. Chernaiwsky et al. (in press) have applied this de-tiding to determine sea level anomalies along the West Coast of U.S.A. and Canada for the three winters between 1996/97 and 1998/99, which straddle the 1997/98 El Niño. They compiled winter observations of sea level from coastal tide gauges (pressure-adjusted using nearby air pressure sensors) and satellite altimeters. Their images show greatly increased sea level heights along the entire continental margin during the El Niño winter. Kang et al (2000) apply the Cherniawsky et al. (2001) method of tidal analysis to compute heights of the internal M₂ tide near Hawaii, to compare with their numerical model. Cummins, Cherniawsky and Foreman (2001) model North Pacific internal tides emanating from the Aleutian ridge, finding good agreement with tidal heights determined by the Cherniawsky et al. (2001) method. Kang et al (2002) also apply the Cherniawsky et al. (2001) technique to observations, which are compared to numerical models of seasonal variability of the M₂ tide in the Yellow and East China Seas. This work is based on a model developed by Lee et al. (2000).

Cummins, Masson, and M.G.G. Foreman (2000) have developed models to simulate the stratification and mean flow effects on diurnal tidal currents off Vancouver Island. These unusual tidal currents are due to rotational waves on the continental shelf, set up by tidal flows at the entrance to Juan de Fuca Strait, and their accurate simulation has challenged physical oceanographers for more than a decade. This paper describes a successful simulation of seasonal changes in this tidal current. Rabinovich and Thomson (2001) find evidence of similar diurnal shelf waves in satellite-tracked drifter trajectories off the Kuril Islands. Masson and Cummins (1999) simulate numerically the buoyancy-driven current adjacent to the West Coast of Vancouver Island.

Iron limitation has been considered to be the cause of the high-nutrient, low-chlorophyll waters near Ocean Station Papa (OSP) in the Gulf of Alaska, following the ship-board experiments of Martin and Fitzwater (1988). A major iron-injection study in the summer of 2002 at OSP tested this hypothesis, injecting about a metric tonne of iron sulphate heptahydrate into surface waters in early July, and a second smaller injection a week later (Wong and Johnson, 2003). Diatom growth rates increased, while iron and macro-nutrient levels declined in the following weeks. The phytoplankton patch stimulated by this iron injection was visible in SeaWiFS images of the region – clear proof of the Martin and Fitzwater (1988) hypothesis. D.W. Crawford et al. (2003) have published one paper on this project. A SERIES (!) of papers will describe this experiment in much greater detail in future months and years.

Foreman, Thomson, and Smith (2000) compare models and observations of seasonal currents on the western continental margin of Vancouver Island. Allen et al. (2001) compiled a synthesis of models and observations of physical and biological processes over a submarine canyon on the Vancouver Islands continental shelf during an upwelling event. Many of these studies (Allen, 2000; Faucher, Burrows, and Pandolfo, 1999; Ianson and Allen, 2002; Mackas, Thomson, and Galbraith, 2001; Peña et al. 1999) formed part of the Canadian West Coast GLOBEC program. A summary of the Canadian GLOBEC program is presented by Mackas and deYoung (2001).

Several programs have investigated the dynamics of Juan de Fuca and Haro Strait, two navigation channels connecting the Pacific Ocean with the Strait of Georgia and the British Columbia Inside Passage. Pawlowicz (2000, 2002) examines internal tides and two-layer flows in this system. Stansfield, Garrett, and Dewey (2001) develop Thorpe-scale techniques for studying turbulence in Juan de Fuca Strait. Li, Gargett, and Denman (1999) develop a box model of the exchange between the Strait of Georgia and the Pacific Ocean.

Scientists at the University of British Columbia and Institute of Ocean Sciences are presently collaborating on a project to examine physical and biological ocean processes in the Strait of Georgia. Masson (2002) has developed numerical models of the deep water exchange between this strait and the Pacific Ocean through Haro and Juan de Fuca Strait, and has completed several years of measurements of currents and water mass changes for model evaluation. During this program Masson and Cummins (2000) modelled the fortnightly modulation of this exchange by the spring-neap cycle of tidal currents.

Along the Inside Passage lie many deep fjords. A series of papers have described ocean processes of these basins. Tinis and Pond (2001) apply observations and models to tidal energy dissipation of Sechelt Inlet. Stacey, Pieters and Pond (2002) simulate the deep water exchange in Indian Arm. Isachsen and Pond examine the influence of spring-neap tidal modulation on currents and water masses of Burrard Inlet. These three studies rely for observations on the long time series of profiling current and temperature observations by Steven Pond of UBC. Gargett, Stucchi and Whitney (2003) examine the exchange between Saanich Inlet and Haro Strait, based on the Whitney and Stucchi time series of bi-daily measurements in these two channels.

The West Coast of Canada and the United States is prone to tsunamis set up by earthquakes and landslides. Richard Thomson and colleagues have completed a series of studies of landslide-generated tsunamis in coastal waters of these two countries (Rabinovich et al. 1999; Thomson et al., 2001; Fine et al., 2002). An ongoing study aims to determine the strength of tsunami currents in navigable channels following the next Cascadia Subduction Zone megathrust earthquake. This program is funded by the Canadian Coast Guard New Initiatives Program, and shared between the Canadian Hydrographic Service, and the Ocean Science and Productivity Section of Fisheries and Oceans Canada, Pacific Region.

A substantial amount of work regarding the development of numerical models has been done.

The semi-prognostic method is being used by researchers at Dalhousie University (Sheng et al., 2001). The semi-prognostic method is a technique for adjusting the momentum balance of an ocean model to correct for systematic error. Examples of systematic error are (i) the separation point of the Gulf Stream is too far to the north (ii) the separated Gulf Stream is too diffuse and (iii) poor representation of the Gulf Stream to the southeast of Newfoundland, all three of which lead to poor representation of the SST and surfaces fluxes (important for interaction with the atmosphere and for tracer uptake, e.g. the carbon cycle). The semi-prognostic method offers a cheap and effective way to correct these errors. The method also offers a new and exciting technique for transferring information between the different subcomponents of a nested-modelling system and as a diagnostic tool.

By coupling realistic models of the North Atlantic Ocean to a simple stochastic model for the atmosphere, researchers (Eden et al., 2002; Greatbatch, 2000; Lu and Greatbatch, 2002; Peterson et al., 2002; Lin et al., 2002; Peterson et al., 2003) have made progress in assessing the possible predictability of the North Atlantic atmosphere/ocean system on time scales out to decadal. The predictability arises from a damped, decadal oscillation associated with the adjustment of the thermohaline circulation to changing surface forcing. They have also made significant contributions towards understanding the North Atlantic Oscillation.

Mercer et al. (2002) have been able to explain some unusual and destructive sea level events along the east coast of Newfoundland in terms of the non-isostatic response to atmospheric pressure forcing associated with storms moving rapidly across the Grand Banks of Newfoundland.

Sheng et al. (2001), Sheng and Tang (2003) are developing regional models of the Northwest Atlantic and the Caribbean Sea. Higher resolution submodels are being nested inside the coarser resolution outer models. A high-resolution coastal circulation model (Sheng and Wang, in preparation) is being developed for Lunenburg Bay, Nova Scotia, as part of a multidiscipinary ocean prediction experiment.

New satellite-mounted gravity measuring instruments (e.g. GRACE) have led to a re-examination of the fundamental approximations used in ocean models. A series of papers (McDougall et al., 2002; Greatbatch et al., 2001; Greatbatch, 2001; Greatbatch and McDougall, 2003) have resolved the "Boussinesq Condundrum" posed by McDougall and Garrett(1992, Deep Sea Res.), suggested a way to easily modify existing Boussinesq model code to make it non-Boussinesq, and shown how to represent mesoscale eddies within a non-Boussinesq framework.

Over the last four years a regional model of the sub-polar North Atlantic (Myers, 2002a,b) has been designed for use in process and sensitivity studies. The regional model has been shown to do a good job in reproducing the circulation and hydrography in this region. Recent emphasis has been on examining the how freshwater is transported in this region (especially in to the Labrador Sea), the importance of correctly representing the bottom topography in the model, and the role of eddies (and their representation in the model) on setting up the circulation in the Labrador Sea.

Advances in continental shelf physics include: numerical modelling of the seasonal mean circulation on the Scotian Shelf and Gulf of Maine (Shore et al. 2000; Hannah et al. 2001); modelling of interannual variability of drift in the vicinity of Browns Bank (Hannah et al. 2000); retrospective analysis of the variable southward penetration of the Labrador Current along the Scotian Slope (Marsh et al. 1999; Loder et al. 2002, 2003); high resolution tidal modelling of the northwest Atlantic (Dupont et al. 2002); methodologies for data assimilation in circulation models (Lynch and Hannah 2001); modelling of benthic boundary layer dispersion (Tedford et al. 2002, 2003), and observations of cross-overs of Scotian shelf water directly onto the northeast peak of Georges Bank (Smith et al. 2003).

Development of desktop applications that allows the user to 1) interact directly with model output to extract tidal predictions for any place and time inside the available model domains (WebTide) and 2) track particles in flow fields created by combining the seasonal mean currents, the tidal currents and additional contributions due to wind forcing (WebDrogue). The applications can be downloaded from http://www.mar.dfo-mpo.gc.ca/science/ocean/home.html

Guoqi Han and his colleagues at Fisheries and Oceans Canada made significant advances in the application of altimetry to oceanography and in the understanding of the Labrador Sea circulation variability and Atlantic Canadian shelf circulation dynamics. A novel method was put forward for calculating transport, using the sea surface as the level of known motion instead of assuming a level of no motion in the traditional geostrophic method (Han and Tang, 1999a,b, 2001). The mean transport (Han and Tang, 1999b) and seasonal (Han and Tang, 1999a) and interannual (Han and Tang, 2001) variability were studied. The interannual transport variability was found positively correlated with the North Atlantic Oscillation (Han and Tang, 2001). Han et al. (2002) studied seasonal variations of sea level and currents over the Scotian Shelf and Slope from altimetry, hydrography and modelling. Han (2002) used the altimetry data to examine interannual sea level variability in the Scotia-Maine shelf seas. The work revealed substantial interannual variability of oceanographic origin in the 1990s and plausibly attributed it to the large-scale oceanic forcing. Han (2003) investigated currents and Gulf Stream ring variability over the Scotian Slope using altimetry and frontal analysis data.

Han et al. (1999) provided a comprehensive description of wind-, density-, and boundary-driven 3-D currents in the Gulf of St Lawrence and on the eastern Scotia Shelf and the southern Newfoundland Shelf from the perspective of a coupled gulf-shelf system. Savenkoff et al. (2001) used the above model circulation fields to constrain a box model for nutrient cycling in the Gulf of St. Lawrence.

Han (2000) developed a 3-D tidal model for the Grand Bank, presented a dynamical description of tidal currents and mixing supported by observations, and demonstrated stratification influences and model sensitivities. Han et al. (2000) implemented an error analysis package in a response analysis method and developed a novel assimilation scheme in a primitive-equation model, demonstrating the capability of producing robust tidal results from a regional hydrodynamic model by assimilating altimetric tides only. Han and Loder (2002) developed a fully non-linear, prognostic model for the eastern Scotian Shelf and provided dynamically consistent hydrography, circulation, tides and turbulence. The study demonstrated the important role of tidal mixing on seasonal circulation over shallow banks. Han et al. (2002) from high-resolution models and measurements examined seasonal-mean circulation, barotropic and internal tides, hydrography and mixing in the Sable Gully and its vicinity.

Operational oceanography is becoming increasingly important, and there is a significant research effort on the Atlantic Coast (e.g., Thompson et al., 2000, 2003).

Acoustical, and near-shore/sediment transport oceanography are active areas of study, the research at times involving all three aspects together (Smyth and Hay, 2002; Henderson and Bowen, 2001; Henderson et al., 2001; MacAulay et al., 1999; Foster et al., 1999; Amos et al., 1999; Zedel and Hay, 1999, 2002; Smyth et al., 2002; Irish et al., 1999, 2002; Zou and Hay, 2001; Crawford and Hay, 2001; Hay et al., 1999).

Considerable research continues to be conducted in the St. Lawrence. The region is of great importance not only because of the fisheries but also because it is part of a very important shipping route stretching into the heart of North America, to the Great Lakes. Operational oceanography is therefore an important part of the research and forecasts are provided of the currents and ice conditions on a routine basis (http://www.osl.gc.ca/previsions-oceaniques/jsp/en/menu-previsions.jsp).

Since 2000, Fisheries and Oceans Canada operates a Gulf of St. Lawrence ice-ocean model for daily short-term forecasting of the surface currents, water levels, and sea ice for the shipping industry, the national Search and Rescue program, civil protection agencies, and environmental protection agencies for oil spill trajectory modeling. The Gulf of St. Lawrence ice-ocean model of Saucier et al. (2003) was coupled to the Canadian weather forecast model (Pellerin et al., submitted to Monthly Weather Review). It was shown that the coupled atmosphere-ocean model improves the weather and sea ice predictions over eastern Canada. Using a numerical model of tidal motion in the St. Lawrence Estuary, a new Atlas of tidal currents was published in 1997 (Saucier et al, 1997; Saucier et al., 1999; Saucier and Chassé, 1998).

A new Atlantic Zone Monitoring Program was established by DFO in 1998 (Therriault et al., 1998). This program includes CTD instruments continuously operating onboard commercial ships sailing the Canadian East Coast (Galbraith et al., 2002). In addition, a network of fixed stations is occupied monthly for biological and physical sampling throughout the Gulf of St. Lawrence and the Labrador and Scotian Shelves.

In recent years, it has been possible to successfully reproduce detailed coupled ice-ocean seasonal cycles for the Gulf and Estuary of St. Lawrence (Saucier et al., 2003), and Hudson Bay and Strait (Saucier et al., submitted to Clim. Dyn.; Senneville et al., 2001; Saucier and Dionne, 1998). These models are fully prognostic and forced through tides, three-hourly atmospheric forcing, and observed open ocean conditions. Through the NSERC-CFCAS CLIVAR and other programs, these ice-ocean models are being coupled to the Canadian Regional Climate Model (e.g., Faucher et al., 2001; Gachon et al, 2001, 2002) for downscaling climate change scenarios to eastern Canada.

Other recent work on the St. Lawrence includes Lu et al. (2001) and Bobanovic and Thompson (2001).

Canadians have contributed to Project Argo, a global operational oceanography and research effort to provide real-time data from ocean profilers launched into all oceans. Each profiler drifts at 2000 metres depth in the ocean, rising to the surface every ten days, measuring temperature and salinity during this rise. Upon reaching the surface each profiler transmits these data to satellite. All observations from all Argo profilers are placed in public Internet directories for simultaneous access by all interested public as well as the research community (Freeland (2003)). Fisheries and Oceans Canada has provided considerable funding for purchase and launch of these drifters. Information on the present status is available at the following Internet sites.

http://argo.jcommops.org/graph_ref/last_status.gif

http://www.pac.dfo-mpo.gc.ca/sci/osap/projects/argo/info_e.htm

http://www.argo.ucsd.edu/

Large scale models are being used to investigate the role of the oceans in climate and paleoclimate. Variability and stability of the climate system and feedbacks associated with air-sea interaction are under active study. The influence of sub-grid-scale parameterizations has been investigated (Knutti et al., 2000; Weaver and Wiebe, 1999). The influence of ventilation of the North Atlantic Ocean on the Last Glacial Maximum (Meissner et al., 2003), the influence of heat, freshwater runoff and momentum fluxes on the ocean circulation (e.g., Saenko et al., 2002), the role of the thermohaline circulation on abrupt climate change (Clark et al., 2002), the influence of sea ice physics on simulations of climate change (Holland et al., 2001), ocean-atmosphere feedback (Topliss, 2002) are among topics that have been studied.

A model has been developed to increase our understanding of the fundamental behavior of the Mediterranean (Myers and Haines, 1999, 2000; Samuel et al., 1999). Besides reproducing the observed circulation, this work has concentrated on the variability within the basin. The role of using surface fluxes to force the model has been examined, with the highlights including the role of changing intermediate water pathways on explaining the changes presently occurring in the eastern basin, the existing of multiple equilibria and the suggestion that hydraulic control changes at the Strait of Gibraltar may play a role in setting up different circulation regimes.

Another sequence of work examines the sapropel S1, during the Holocene, and attempts to combine paleo-data with general circulation modelling to test hypotheses and validate paleo-scenarios (Stratford et al., 1999; Myers and Rohling, 2000; Rohling et al., 2000; Myers, 2002c). Based upon a set of significantly different salinity reconstructions, the modelling results show the circulation in the Mediterranean at the time of S1 was a weakened anti-estuarine mode. Extensions to this work have examined the role of air-sea fluxes in setting the paleo-circulation of the Mediterranean. Output from the general circulation model has also been used to force a set of offline nutrient and ecosystem models of S1.

Double diffusion continues to be focus of research for scientists at Dalhousie University (Kelley et al., 2003; Kelley, 2001; May and Kelley, 2000; Walsh and Ruddick, 2000).

Turbulence closure schemes have also received attention. Stacey (1999) estimated the roughness length on the sea-side of the air-sea interface using near-surface current and density observations from Knight Inlet, British Columbia. The roughness length was estimated by comparing the currents and density produced by a numerical model to the observations. The model used a Mellor-Yamada turbulence closure and a Charnock parameterization of the roughness length. It was found that the roughness length was much larger on the sea-side of the interface than on the air-side and may be of the order of the significant wave height in magnitude.

Stacey et al., (2002) and Stacey and Pond (2003) simulated the circulation in Burrard Inlet/Indian Arm, British Columbia. It was found that deep-water renewal events in Indian Arm, and the circulation in the lee of one of the major sills in Burrard Inlet could be noticeably better simulated by a numerical model if, in the Mellor-Yamada turbulence closure scheme, the influence of horizontal (as well as vertical) spatial variations on the production of turbulent kinetic energy was taken into account.

Research is also being directed to generic questions about how well we understand the physics we might wish models to represent. Traditional ocean modeling follows classical mechanics plus added-on mixing terms. Progress from classical to statistical mechanics shows ways in which models may be wrong and may also provide concrete steps to better physics (Holloway, 1999, 2002; Merryfield et al., 2001).

The influence of sub-grid-scale parameterizations has been a subject of investigation (e.g., Knutti et al., 2000; Weaver and Wiebe, 1999).

As it became recently possible to successfully reproduce ice-ocean conditions in the Canadian Inland Seas over several years using prognostic 3D models, efforts have significantly increased in our ability to drive ecosystem models. Those include a hierarchy of models for primary production (Tian et al., 2000; Le Fouest et al, 2002), and other applications for understanding the distribution of whales and fishes (e.g., Lavoie et al., 2000; Simard et al., 2002; Weise et al., 2002; Tian et al., 2001). New projects aim at the consistent modeling of the carbon cycle including studies of its sensitivity to climate variability and change.

Other work on the influence of physical processes on biology includes Drinkwater et al. (2000), de Young et al. (1999) and Rose et al. (1999).