2.1 Governing Law

Maritime boundaries delimit overlapping maritime entitlements and the rules governing the delimitation process differ somewhat depending on the maritime zones in question. UNCLOS Article 15 provides that the default boundary line delimiting the territorial sea between opposite or adjacent states is ‘the median line every point of which is equidistant from the nearest points on the baselines from which the breadth of the territorial seas of each of the two States is measured’. This makes the equidistance method the default method for delimitation of the territorial sea and the reference to baselines means that the location of the equidistance line can change with ambulatory baselines. An equidistance or median line can be modified ‘where it is necessary by reason of historic title or other special circumstances to delimit the territorial seas of the two States in a way which is at variance therewith’.^{Footnote 20}

UNCLOS Articles 74(1) and 83(1) govern the delimitation of exclusive economic zone and continental shelf boundaries. Both provisions are identical, except that where the former refers to the exclusive economic zone, the latter relates to the continental shelf. Delimitation of the exclusive fisheries zone is not explicitly governed by UNCLOS but the International Court of Justice (ICJ or the Court) has held that the customary law enshrined in UNCLOS Article 74 also applies to delimitation of an exclusive fisheries zone.^{Footnote 21} These provisions make no reference to the equidistance method and merely instruct that delimitations ‘shall be effected by agreement on the basis of international law, as referred to in Article 38 of the Statute of the [ICJ], in order to achieve an equitable solution’. Relevant international law has evolved into a relatively transparent three-step rule, hereinafter referred to as the equidistance/relevant circumstances method of delimitation. This method involves the drawing of a provisional equidistance line, its subsequent adjustment by reference to relevant circumstances and, finally, a disproportionality test to ensure that the end-result is equitable.^{Footnote 22} The equidistance/relevant circumstances method is generally preferred by courts and tribunals but special or relevant circumstances can justify derogation from the equidistance method altogether, even for delimitation of the territorial sea. The ICJ confirmed this in 2007 when delimiting a maritime boundary between Nicaragua and Honduras.^{Footnote 23}

A distinction can be made between the equidistance/special circumstances method under UNCLOS Article 15 and, the more elaborate, equidistance/relevant circumstances method described above. Yet, the methods are closely related and the difference not always reflected in the literature.^{Footnote 24} The former method involves the drawing of a provisional equidistance line and subsequent adjustment in light of relevant circumstances.^{Footnote 25} As such, it includes the first two steps in the equidistance/relevant circumstances approach but not the third step, the disproportionality test. Furthermore, unlike Articles 74 and 83, Article 15 refers to a line that automatically marks the limit to entitlements between states with overlapping claims in the absence of a delimitation agreement. These limits to territorial sea claims might be characterized as provisional boundaries and they arguably fluctuate with changing coastlines, unless and until otherwise agreed. Indeed, provisional arrangements may be necessary for overlapping territorial seas because all coastal states have a territorial sea without any proclamation to that end;^{Footnote 26} the possession of a territorial sea may even be compulsory.^{Footnote 27} It follows that a provisional territorial sea boundary may be necessary until a definitive boundary is established. It may be controversial to refer to such limits as ‘boundaries’ but the word is used here to distinguish bilateral boundaries from unilateral limits. After all, the word boundary can be broadly construed for different purposes. The ILC has explained that the fact that a line qualifies as a boundary in some respects, does not automatically make it a boundary as referred to in the law of treaties.^{Footnote 28}

2.2 Provisional boundaries

UNCLOS Article 15 provides that states with overlapping territorial sea entitlements cannot, failing agreement, extend their territorial sea beyond the equidistance line measured from relevant baselines, except where justified by reference to special circumstances or historic title. This provision establishes a bilateral limit to the territorial sea and explains ‘how States must act failing an agreement on delimitation’.^{Footnote 29} Thus, an equidistance line can function as a provisional arrangement until either state initiates the delimitation of a more definitive boundary. States can also take the view that they have a lawful claim to an area beyond the equidistance or median line on the grounds of special circumstances or historic title. Compelling evidence of historic title or special circumstances might justify the exercise of territorial sea rights beyond the equidistance line, particularly if no objections are raised. However, in lieu of an agreement to the contrary, the default provisional boundary would be the equidistance or median line, midway between neighbouring states with overlapping territorial seas. Such a line would assumedly fluctuate in accordance with ambulatory baselines to continuously reflect relevant coastlines, unless and until otherwise agreed.

UNCLOS Article 15 is identical to Article 12(1) of the 1958 Convention on the Territorial Sea and the Contiguous Zone and Article 6 of the Convention on the Continental Shelf, except for minor editorial changes.^{Footnote 30} Therefore, Article 6 of the 1958 Convention on the Continental Shelf arguably establishes a provisional equidistance boundary to the continental shelf. This is of course the predecessor of UNCLOS Article 83(1) and might be relevant for informing the content of that provision. However, UNCLOS Articles 74 and 83 were intentionally made less explicit than the delimitation rules enshrined in the Geneva Conventions because of competing views among the drafters of UNCLOS concerning the most suitable method for delimiting the exclusive economic zone and continental shelf.^{Footnote 31} For the foregoing reasons (and without ruling out the possibility that equidistance lines can represent provisional boundaries to other maritime zones unless and until otherwise agreed)^{Footnote 32} this section focuses on provisional boundaries under UNCLOS Article 15, i.e., territorial sea boundaries.

Most commentators accept the theory of ambulatory baselines and agree that changes to coastal geography have an impact on maritime entitlements and zonal limits.^{Footnote 33} In the words of the ILA Baselines Committee, ‘the normal baseline is ambulatory, moving seaward to reflect changes to the coast caused by accretion, land rise, and the construction of humanmade structures … and also landward to reflect changes caused by erosion and sea level rise’.^{Footnote 34} This has implications for provisional equidistance boundaries under UNCLOS Article 15 since they are measured from points on relevant baselines. Caron alluded to this in 1990, noting that:

Caron may have been referring to outer maritime limits as ‘boundaries’ but the principle applies equally to outer limits (when relevant basepoints are affected) and maritime boundaries that are contingent upon parts of the baselines.

The 2007 maritime boundary arbitration between Guyana and Suriname dealt with the legal effects of a provisional equidistance boundary established under UNCLOS Article 15. The tribunal was tasked with the delimitation of territorial sea entitlements extending from a highly unstable coastline. The adjacent states had previously agreed on a fixed location for a territorial sea boundary reaching three nautical miles (nm) seawards but no location was agreed upon when both states extended their territorial sea limits to 12 nm in 1977 and 1978.^{Footnote 36} The tribunal found that a three-nm territorial sea boundary had been established by reference to the N10°E line^{Footnote 37} but no explicit arrangement was concluded beyond that point.^{Footnote 38} According to Guyana, both states relied on the ‘historical equidistance line’ along the N34°E azimuth, beyond the established three nm limit.^{Footnote 39} However, the equidistance line had shifted because relevant coasts were made up of mud banks ‘constantly subjected to erosion and accretion’.^{Footnote 40}

Suriname sought to rely on the settled practice of the states in relation to the 10° line, allegedly amounting to a ‘perfected instrument’, but the tribunal held that no legal obligations had been created beyond the three nm limit.^{Footnote 41} Thus, the provisional arrangement concerning delimitation beyond three nm had not resulted in a permanent boundary. The tribunal concluded that the territorial sea boundary should follow the N10°E line from relevant coasts seawards to three nm and, thereafter, intersect the equidistance line as it was at the time of delimitation^{Footnote 42} – not the ‘historical equidistance line’. Therefore, the boundary had shifted in accordance with the relevant coastal geography. This case demonstrates that the location of a lawful equidistance line, satisfying the conditions of UNCLOS Article 15, can fluctuate from the time when a provisional arrangement is established until a permanent boundary is delimited through adjudication or agreement. However, it also indicates that, in other circumstances, the practice of states could amount to a ‘perfected instrument’, creating a stable boundary that would be unaffected by changing baselines.

4.1 Changing coastal geography

Maritime boundary delimitation usually results in geographically stable boundaries. However, as rightly noted by Bangladesh in the maritime boundary dispute between Bangladesh and India, ‘it would be against equity and common sense to draw a permanent boundary line using basepoints on an unstable coast’.^{Footnote 70} Doing so would distort the correlation between relevant maritime entitlements under UNCLOS and derived maritime boundaries. The solution proposed by Bangladesh was to disregard submerging features and establish a geographically stable boundary by reference to more permanent basepoints.^{Footnote 71} Another solution would be to construct a fluctuating boundary by reference to the fluctuating coastline or make revisions in case of subsequent changes.

Coastal instability has influenced the delimitation process in judicial decisions and led to the adjustment of geographically stable delimitation lines. Courts and tribunals have confirmed, on numerous occasions, that they can give small, submerging or otherwise insignificant features limited or no effect to avoid a disproportionate effect on the resulting boundary.^{Footnote 72} This was the case in Bangladesh v. India when the arbitral tribunal found that the submerging and barely visible South Talpatty/New Moore Island was ‘not a suitable geographical feature for the location of a base point’.^{Footnote 73} Furthermore, the ICJ decided to give the tiny, uninhabited, and unstable feature, Qit’at Jaradah, no effect in the boundary dispute between Qatar and Bahrain, even after classifying it as an island.^{Footnote 74} The Court’s decision does not explicitly refer to coastal instability as a determining factor in having the boundary line pass immediately east of Qit’at Jaradah. However, Judge Vereshchetin declared that the feature should not be considered an island due to its constantly changing physical condition.^{Footnote 75} These examples show that fluctuations of relevant coastlines may already influence the thinking of courts and tribunals in some maritime boundary delimitation cases. Yet, the end result is usually a geographically stable boundary – with two notable exceptions: the Territorial and Maritime Dispute between Nicaragua and Honduras in the Caribbean Sea and the Maritime Delimitation in the Caribbean Sea and the Pacific Ocean (Costa Rica v. Nicaragua).

In the first case, the ICJ constructed an angle-bisector boundary^{Footnote 76} because continued coastal accretion at the river delta made it practically impossible to identify suitable basepoints.^{Footnote 77} Furthermore, the Court established a fluctuating boundary for the first segment of the boundary, from the land boundary terminus and seaward. The first three nm of the boundary were not fixed at geographic co-ordinates. The purpose was to allow for adjustments to the land boundary terminus, the fluctuating Coco River.^{Footnote 78} Nicaragua and Honduras both proposed, in their written submissions, that, due to the changing nature of the coastline, the ‘starting-point for the boundary line between them should be located some distance from the mainland coast, but disagreed on exactly where’.^{Footnote 79} During the oral proceedings, Nicaragua advocated for a fluctuating boundary, which would connect the geographically-fixed maritime boundary by a straight line to the land boundary terminus, ‘wherever it [might] be located on any given day’.^{Footnote 80} However, Honduras urged that a diplomatic solution be found for the first three nm of the boundary.^{Footnote 81} The Court ultimately concluded that the Parties should:

Thus, the first segment was left to fluctuate unless and until otherwise agreed by the parties.

The solution presented by Nicaragua in the oral proceedings went even further as that would have permanently established a fluctuating boundary. However, the Court favoured the approach originally advanced in written proceedings by both parties, i.e., to fix no boundary for the first segment.^{Footnote 83} In so doing, the Court could rely on judicial precedents confirming that maritime delimitation might ‘begin at some distance out at sea’, particularly where there was ‘an uncertain land boundary terminus’.^{Footnote 84}

The ICJ recently took the next step and explicitly established a fluctuating maritime boundary in a dispute between Costa Rica and Nicaragua. In this case, the Court constructed a ‘mobile line’ connecting a point on solid land to the median line two nm offshore due to the ‘high instability and narrowness of the sandspit near the mouth of the San Juan River which constitutes a barrier between the Caribbean Sea and a sizeable territory appertaining to Nicaragua’.^{Footnote 85} The Court found it more appropriate to create a fluctuating boundary-segment connecting the delimitation line to a fixed point on solid land rather than placing basepoints on the unstable sandspit. To this end, a court-appointed expert identified the point on solid land (on the Costa Rican side) which was closest to the mouth of the San Juan River. The Court explained where this point would be ‘[f]or the present’ while also noting that there might ‘be geomorphological changes over time’.^{Footnote 86}

The Costa Rica v. Nicaragua case demonstrates a clear effort to create a stable legal outcome while also acknowledging the effects of unforeseeable environmental changes. The Court used salient basepoints on land because of ‘a relatively higher stability than points placed on sandy features’^{Footnote 87} but it also established a ‘mobile line’ for the most unstable segment and even indicated that the point anchoring this segment might be changed in the future.

4.2 Changing ecological circumstances

The foregoing suggests that changing coastal geography may sometimes lead to fluctuations in derived maritime boundaries. This section explores the possibility of fluctuations, if maritime boundaries are established by reference to changing ecological circumstances. Two factors make this perspective relevant for the present article. First, the fact that non-geographic circumstances affecting maritime boundaries often relate to ecological circumstances, i.e., the exploitation of living resources.^{Footnote 88} Second, the fact that ecological circumstances usually provide limited geographic stability and have, perhaps for that reason, often been disregarded in the past.

The ICJ confirmed the relevance of ecological changes in the Jan Mayen case when it considered the potential effect of seasonal changes on access to fish stocks.^{Footnote 89} The Court had to determine whether the presence of ice qualified as relevant circumstances for the purpose of maritime delimitation, due to its effect on the exploitation of capelin, the only commercially exploitable fish resource in the area. It concluded that ice sheets should not affect the location of the boundary because exploitable capelin could only be expected to appear at times and in areas where the ice had melted.^{Footnote 90} The reasoning of the Court suggests that similar seasonal changes to relevant ecology could affect the delimitation process if access to natural resources was affected. The ICJ’s decision in the Jan Mayen case sheds no light on what the potential effect would be on a resulting boundary. One solution would be to establish a boundary that fluctuates with ice sheets along relevant coastlines or the migration patterns of relevant fish stocks, although this might not be practically desirable for states or mariners.

A chamber of the ICJ also had to consider the relevance of changing ecological circumstances in the Gulf of Maine case in 1984. Moreover, this case dealt specifically with the relevance of a fluctuating limit in the natural environment. The United States and Canada acknowledged the existence of discernible oceano-biological boundaries separating distinct ecological communities, but they disagreed on their location.^{Footnote 91} Canada identified an oceano-biological boundary at the extreme western limit of the delimitation area, east of the Great South Channel.^{Footnote 92} Meanwhile the United States found three ‘different oceanographic and ecological’ boundaries in the area and submitted that the most important one, separating most of the commercially exploitable fish stocks, ran along the Northeast Channel.^{Footnote 93}

The chamber confirmed that the ‘great mass of water is taken into consideration not as some inert mass, but as the habitat of an exceptionally extensive wealth of fauna and flora’.^{Footnote 94} It was ‘not however convinced of the possibility of discerning any genuine, sure, and stable “natural boundaries” in so fluctuating an environment as the waters of the ocean, their flora and fauna’^{Footnote 95} or of the relevance of such circumstances for the seabed.^{Footnote 96} Therefore, it seems that these boundaries were rejected for two reasons. First, because they were not sufficiently ‘sure and stable’. Second, because the parties had requested a single maritime boundary and the ecological boundary was not relevant for delimiting the continental shelf. These two reasons will be addressed in turn.

First, the Chamber of the ICJ ‘reached the conviction that it would be vain to seek, in data derived from the biogeography of the waters covering certain areas of sea-bed, any element sufficient to confer the property of a stable natural boundary’.^{Footnote 97} It is not entirely clear whether this reference to stability relates to geographic stability, certainty of the boundary, or both. The parties had requested a geographically fixed boundary^{Footnote 98} but the natural boundary that the United States relied upon as a basis for delimitation was bound to change.^{Footnote 99} Therefore, the Chamber was arguably faced with an impossible task: to delimit a fixed boundary by reference to a fluctuating limit. It proceeded to establish a geographically stable boundary without specifically addressing the possibility of making it ambulatory or the apparent contradiction in aforementioned claims. In fact, the Chamber declared that in this case, there were ‘no geological, geomorphological, ecological or other factors sufficiently important, evident and conclusive to represent a single, incontrovertible natural boundary’.^{Footnote 100}

Second, the Chamber rightly concluded that an ecological boundary could not be relevant for establishing a maritime boundary ‘serving a double purpose’,^{Footnote 101} also known as a single purpose boundary. The construction of single purpose boundaries has been generally accepted since first employed in the Gulf of Maine case.^{Footnote 102} Yet, this approach is not always suitable because the application of international law can have different results for different maritime zones, resulting in different boundaries.^{Footnote 103} Indeed, circumstances that are relevant for the exclusive economic zone or exclusive fisheries zone may be irrelevant for delimitation of the continental shelf, and vice versa. The Chamber found, in the Gulf of Maine decision, that in delimiting two maritime zones by a single boundary it had to apply neutral criteria, equally suitable to both zones.^{Footnote 104} This made the oceano-biological boundary inappropriate. One can wonder if the decision might have been different if the parties had requested two distinct boundaries, one for the exclusive fisheries zone (or exclusive economic zone) and another one for the continental shelf,^{Footnote 105} i.e., if the natural boundary had been sufficiently important, evident, and conclusive.