Article 76 provides a definition of the juridical continental shelf by incorporating both legal and scientific terms (Fig. 1). The Article also describes the entitlement of coastal states to delineate the extended continental shelf. The continental shelf is defined as: … the natural prolongation of its land territory to the outer edge of the continental margin, or to a distance of 200 nautical mile …(paragraph 1). This definition uses the continental margin as a ‘yardstick’, which is defined as: …the submerged prolongation of the land mass of the coastal State, and consists of the seabed and subsoil of the shelf, the slope and the rise. It does not include the deep ocean floor with its oceanic ridges or the subsoil thereof (paragraph 3).

It is important to realize the differences between the scientific definition of the continental shelf (gently sloping submerged marginal zone of the continents extending from shore to an abrupt increase in bottom inclination) and the above legal definition, which includes the continental margin and can include parts of the deep ocean floor (Fig. 1). To avoid confusion, we will use in this paper the term ‘shelf’ when we indicate the scientific continental shelf. The term ‘continental shelf’ will be used for the legal continental shelf, as defined above.

If the outer edge of the continental margin extends beyond 200 nm, a coastal state must demonstrate that it meets the criteria of Article 76 and collect the geological and geomorpho-logical data to define its extended continental shelf. Article 76 provides formulae for measuring the continental shelf seaward (Fig. 1B). The first step is to identify the ‘foot of the continental slope’ (FOS), defined as the point of maximum change in gradient at its base. Measuring from this point, the outer limit is established either at a distance of 60 nm from the FOS, or the distance to a point where the thickness of the sedimentary layer is at least 1% of the shortest distance from such point to the FOS. A line connecting these points is known as the Gardiner line. It is often overlooked that establishing a Gardiner line requires a sediment thickness that is 1% of the distance to the FOS. Therefore, the seismic records need to be depth converted, using sediment velocities. Establishing those velocities is not always easy, especially in the Arctic (see below).

Article 76 also imposes constraints on the maximum extent of the continental shelf (Fig. 1C). Again, it provides two options: the outer limit shall not exceed 350 nm from the baselines of the coastal state, or extend beyond 100 nm from the 2500 m iso-bath. A coastal state can use a combination of the above criteria to determine the area of its continental shelf. The outer limit of the continental shelf beyond 200 nm is delineated by straight lines not exceeding 60 nm in length connecting the derived points.

The coastal state must gather and analyze scientific data and then prepare and file a submission to the Commission on the Limits of the Continental Shelf (CLCS). The role of the CLCS is to review a state’s submission in light of Article 76 criteria and make recommendations to the state. The members of the CLCS are elected by state parties to the Convention and are to be experts in the fields of geology, geophysics, or hydrography. They serve a five-year term and are eligible for reelection. To assist coastal states, the CLCS produced a set of technical guidelines outlining the information to be submitted (United Nations 1999). Only the coastal state can establish the outer limits of its shelf and if it does so based on recommendations of the CLCS, they are final and binding (Article 76, paragraph 8). It is important to note that the CLCS has no mandate to resolve disputes between states and that the actions of the CLCS are without prejudice to the delimitation of boundaries between coastal states. Disputes must be resolved by the states involved through negotiation or a dispute settlement process.

States have ten years from the date they became party to the Convention to make their submission to the CLCS, although by virtue of decisions taken at meetings of state parties to UNCLOS, this deadline can now be satisfied through the filing of preliminary information about an intended submission. Since states became party to the Convention at different times, they have different deadlines. Canada became a party to the Convention in December 2003 so its deadline is December 2013.

The Scotian margin developed after Pangea rifted and North America began to separate from the African continent by the Early to Middle Jurassic. The margin includes a steep narrow slope that extends from 200 to 4000 m water depth onto the Sohm Abyssal Plain (Fig. 4). Underlying the slope region is the Scotian Basin, a prominent sedimentary basin of Early Jurassic age. Basin fill thicknesses exceed 16 km in parts of the basin, and sedimentary units thin gradually seaward to the continental rise and adjacent Sohm Abyssal Plain. Jurassic salt deposition results in salt diapirism (Fig. 4), which impacts the morphology of the lower slope and rise even today (Shimeld 2004).

The Sohm Abyssal Plain generally has a thick sediment cover; compilations of sediment thickness estimate that the thickness is over 2 km in the area beyond 200 nm (GSC 1994). Because a sediment thickness of more than 1 km will result in an outer limit of the continental shelf that is farther seaward than the FOS + 60 nm, it was decided that in this area the Gardiner formula would be appropriate to apply, and additional seismic data were collected for this purpose.

In 2007, a total of 6900 km of seismic data were collected over the Sohm Abyssal Plain (see Fig. 3 for the survey lines). The survey was designed to cover the area beyond 200 nm off-shore, and the lines extended to 350 nm (the maximum possible outer limit). The processed seismic data show layers of sediment covering the plain (Fig. 4). The analysis is ongoing and we are using stacking or processing velocities for depth conversion to locate the 1% sediment thickness points (using existing bathymetry profiles to define the FOS). However, locating the FOS along this margin is complicated by the presence of slope failures and other downslope processes.

The Grand Banks is a broad platform, generally shallower than 100 m, under-lain by Appalachian continental crust and extending in some places beyond Canada’s 200 nm limit. The southwestern Grand Banks is a transform margin associated with the rifting event that separated Nova Scotia from Africa. The eastern Grand Banks margin formed as a result of the rifting and subsequent seafloor spreading that separated it from the Iberian margin. The northern segment of the Grand Banks was formed as a result of the separation from Goban Spur (Keen et al. 1990).

Much of the slope around the Grand Banks is similar to the Scotian Slope, with canyons and gullies providing conduits for downslope transport of sediment. The southwestern Grand Banks slope shows evidence of mass transport processes moving sediment to the lower slope, rise and abyssal plain (Giles et al. 2010). Orphan Basin, on the north flank of the margin, largely consists of stacked mass transport deposits and interbedded turbidites (Tripsanas et al. 2008), providing the sedimentary infill between the Grand Banks margin and Orphan Knoll (Fig. 6).

For a significant part of the Grand Banks margin, the most seaward outer constraint under Article 76 is formed by the 2500 m (isobath) +100 nm limit. Moreover, for a large part of the margin, the FOS + 60 nm points are located either near or beyond that constraint. Therefore, it is important that around the Grand Banks we have a good definition of the FOS and also of the location of the 2500 m isobath. For this reason, in 2006 we acquired 18 500 line-km of multi-beam bathymetry profiles (Fig. 6).

Around the Grand Banks, there are several features that might complicate the determination of the outer limit, or offer an opportunity to further extend the limit (Fig. 6). For example, Flemish Cap is a submarine knoll consisting of a central core of Neoproterozoic rocks associated with the Appalachian Avalon Zone (King et al. 1985). Orphan Knoll is located farther north and its continental origin was confirmed by drilling (Laughton et al. 1972). Orphan Basin links Orphan Knoll to the Grand Banks and this basin is underlain by thinned continental crust (Chian et al. 2001). To consider these features part of the Grand Banks, it must be demonstrated that they all fall inside the envelope of the FOS around the Grand Banks. A bathymetric profile across the Orphan Basin and Orphan Knoll (Fig. 6, see profile bottom of figure) indicates that this is the case: the deep ocean floor and the FOS, as defined under Article 76, are located on the seaward side of Orphan Knoll.

The Labrador margin was formed by rifting of the North Atlantic Precambrian craton (Keen et al. 1994). This rifting started in the Early Cretaceous and the subsequent seafloor spreading formed the Labrador Sea. The exact timing of the onset of seafloor spreading is debated, highlighting the complication of the location and nature of the ocean–continent boundary. It is generally agreed that seafloor spreading began during chron 27 (Paleocene) and stopped in the Early Oligocene (just before chron 13). The Labrador Sea remains a small ocean basin that is significantly shallower than the standard deep ocean (Roest et al. 1992).

The upper slope of the Labrador margin appears almost entirely eroded, though some buttes and ridges exist as remnants. Mid-slope gullies coalesce and form channels that transect the rise and abyssal plane to eventually merge with the Northwest Atlantic Mid-Ocean Channel (NAMOC) in the central Labrador Sea. Large mass transport deposits were also imaged on the slope and rise (Deptuck et al. 2007). Another dominant process on the Labrador margin is contour-parallel currents entraining, transporting and depositing sediment. The strong Labrador and North Atlantic deep-water currents form abundant sediment waves along the slope and rise, and have constructed several large drift deposits such as Hamilton Spur (Goss 2006).

The initial analysis of the Labrador margin and possible locations of the foot of the slope concluded that using the FOS+60 nm option would define an outer limit of the continental shelf that is likely to be mostly inside 200 nm (GSC 1994). Therefore, we reviewed existing seismic data to define sediment thickness. Figure 7 shows the location of existing seismic profiles and illustrates that several tracks cross the entire Labrador Sea, covering both the Labrador and Greenland margins (Hinz et al. 1979; Keen et al. 1994). Based on that review, it was decided to collect additional multi-channel seismic data (Fig. 7). Several Ocean Drilling Program well sites provide groundtruth to these seismic data.

Denmark is also in the process of defining the outer limit of the Greenland continental shelf. For that purpose, they have acquired new multi-channel seismic data off the Greenland margin (Fig. 7). During the last few years, Canada collaborated with Denmark (acting for Greenland, in this and subsequent contexts) in sharing and jointly analyzing the seismic data in the Labrador Sea. This approach is beneficial to both countries: since the interpretation of the data from each side of the Labrador Sea is consistent and in agreement, each country’s submission is strengthened.

The Arctic Ocean coastal states (Canada, Russia, Norway, Denmark and the US) are all in the process of defining their extended continental shelves, although they are at different stages. Since data acquisition in the Arctic is a challenging and expensive undertaking, Canada has explored and implemented international partnerships to collect data of mutual benefit. This cooperation mitigates the risks, and reduces costs and environmental impacts; at the same time, a common data interpretation by neighbouring countries enhances the probability of a successful submission.

Over the past five years, Canada has conducted seven joint surveys with Denmark, and is moving forward on joint interpretation and publication of results. With the United States, Canada has conducted three joint surveys in the western Arctic (2008–2010) using the Canadian icebreaker CCGS Louis S. St-Laurent and the US ice-breaker USCGC Healy; a fourth survey is planned for 2011. Since 2007, annual meetings have taken place between Canadian, Danish and Russian officials to discuss the ongoing programs and results; these meetings recently also included officials from the US (from 2009) and Norway (2010).

From the Canadian perspective, the eastern Arctic continental margin is dominated by the Lomonosov and Alpha ridges (Fig. 8). It is important to demonstrate that these features are natural prolongations of the continent; for the Lomonosov Ridge, this is of interest to both Canada and Denmark, whereas Alpha Ridge is important only to Canada. The bathymetry in that region shows a trough north of Ellesmere Island that seemingly separates the ridges from the mainland (Fig. 9). It was therefore necessary to image the crustal structure below the seafloor by measuring the crustal seismic velocities of the ridges, to compare them with those on the adjacent continent.

The 2006 LORITA project (LOmonosov RIdge Test of Appurtenance) was a joint project between Canada and Denmark that conducted seismic and bathymetric surveys on the Lomonosov Ridge. For this expedition, Canadian Forces Station Alert was used as a main base and we established a small ice camp about 100 km offshore. Despite losing 65–70% of the days to bad weather, the primary objective of collecting seismic refraction data was achieved along a 440 km-long north–south profile and a 110 km-long profile along the bathymetric trough (Fig. 9). The data were interpreted jointly with Denmark and the conclusion is that there is continuity of the continental crust from the coast across the trough and onto the Lomonosov Ridge. The velocity structure in the trough suggests that no oceanic crust occurs there (Jackson et al. 2010).

In the spring of 2008, a similar project (ARTA) was undertaken on Alpha Ridge, north of Ellesmere Island (Fig. 9). For this project, a large ice camp was constructed near the mouth of Nansen Sound as well as a small camp about 100 km farther off-shore. The seismic refraction experiment consisted of a 350 km-long line perpendicular to the margin and a 175-km-long cross line on the bathymetric trough (Fig. 9). Initial results of this experiment were reported by Funck et al. (2010). Together with previously acquired seismic refraction data and recently acquired Russian data on the Mendeleev and Lomonosov ridges, a more complete model of the deeper crustal structure will be developed to assess that these features are natural components of the Canadian continental margin.

A 2009 joint project with Denmark collected bathymetric data to measure the shape of the seafloor in the area between the Alpha and Lomonosov Ridges. These three surveys were successful and collected a significant amount of new data. However, the construction of large ice camps in remote areas (10–15 large tents, with a population of 30–40 people; see Fig. 10), including flattening the area for runways for supply aircraft, was time consuming and expensive. Moreover, unpredictable weather and ice conditions became a main concern. Open leads generated ice fog, preventing helicopters from flying, and ice floes were breaking up, leading to dangerous situations (requiring an emergency evacuation of a small off-shore camp in 2009). Moving farther offshore, as was planned for the last phase of the program, was deemed very risky, and alternatives were investigated. In 2008, a joint project was initiated with DRDC (Defence Research and Development Canada) to use autonomous underwater vehicles (AUV) to go under rather than on or through the ice to collect the requisite data.

To define the outer limits of Canada’s extended continental shelf in the Canada Basin, the thickness of the sediment cover is needed to be able to define the points where this thickness is at least 1% of the distance to the foot of the slope. Therefore, the program in the western Arctic consists mainly of seismic surveys. Prior to 2006, there were only a few seismic lines crossing the basin (Fig. 8), constituting a serious knowledge gap in producing a sediment thickness map of the Arctic (Jackson et al, 1990).

The collection of seismic data in the Arctic given prevailing ice conditions is extremely difficult. The potentially heavy ice conditions require that surveying be done with an icebreaker and the noise of its powerful engines can interfere with the seismic sound source. Moreover, the standard equipment for seismic data collection needs to be strengthened because of fragments of ice behind the icebreaker, which may interfere with the towing of the seismic sound sources and receivers. A modified seismic system was developed (e.g. Mosher et al. 2009), consisting of an 1150 cubic inch (18.8 litre) G-gun array and a 16 channel digital hydrophone streamer. To prevent interference of the source array with ice, it is towed at 11.5 m depth, below the keel of most sea ice. The hydrophone streamer is towed from the rear of the source array, also to keep it deep beneath the ice. The length of the streamer (100 m) is short for operating in deep water, providing little move-out for velocity analysis. Instead, velocity information is derived from wide angle reflection and refraction data received on expendable sonobuoys, deployed at regular intervals.

After testing in 2006, on board the CCGS Louis S. St-Laurent, a six-week seismic and bathymetric survey was conducted in September 2007 in the southern part of Canada Basin. Ice conditions varied; to the north and east, near the Canadian Archipelago, heavier ice was generally encountered, and it was concluded that to collect seismic data in these regions, two icebreakers would be required. In 2008, a joint survey with the US was organized using both the CCGS Louis S. St-Laurent and the US icebreaker the USCGC Healy. In areas where seismic data are important, the USCGC Healy breaks ice and the CCGS Louis S. St-Laurent follows with its seismic system; in areas where the shape of the seafloor is important, the Canadian vessel breaks ice and the US vessel follows with its high resolution multi-beam bathymetry system. The advantage of this configuration is that since the following icebreaker does not have to break heavy ice, the quality of the data collected is significantly improved.

Four major surveys have now been completed in the western Arctic (3 joint with the US) and over 13 500 km of high quality seismic data were collected (Fig. 11). During these same surveys, in excess of 18 000 km of bathymetry data were also collected. In addition, velocity information was collected with 146 sonobuoy deployments across the entire Canada Basin. After processing, these data will provide a model for the velocity structure of the sediments to be used in seismic depth conversion.

These recent icebreaker surveys covered most of the area in which data were needed for ECS purposes (Fig. 11). The data coverage has more than tripled as a result of these surveys and these new data will significantly improve our knowledge of the region. Figure 12 shows flat lying sediments with thicknesses generally decreasing northward, and pinching out against the topographic highs associated with the Alpha Ridge (Mosher et al. 2011).

Initial interpretations of the data were presented at scientific conferences (Shimeld et al. 2010; Mosher et al. 2011) and publications are in preparation. In the context of defining the outer limits of Canada’s continental shelf, the main finding is that the Canada Basin is generally covered by sediment more than 4 km thick, thinning to the north and to the west. These data will support Canada’s definition of a significant extended continental shelf in this region.