I. Introduction

All coastal States are entitled to maritime zones in the water columns surrounding their land and in the seabed extending from their coastlines, in accordance with the United Nations Convention on the Law of the Sea (UNCLOS).1 UNCLOS determines the maximum width of each maritime zone and when the entitlements of opposite or adjacent coastal States overlap, a maritime boundary must be established in accordance with Articles 15, 74 and 83 of UNCLOS. These three provisions require that an agreement on the basis of international law, as referred to in Article 38 of the Statute of the International Court of Justice, is reached between relevant neighbouring States. The territorial sea must be delimited by reference to a median or equidistance line (unless special circumstances or historic title justify application of a different method)2 and delimitation of the exclusive economic zone and continental shelf must always result in an equitable solution.3 If efforts to negotiate maritime boundaries fail, boundaries are delimited through the means provided for in UNCLOS Part XV,4 for example, through judicial settlement, unless relevant States have opted out of the mandatory settlement of boundary disputes as permitted under Article 298(1(b)) of UNCLOS. Consequently, the requirement to establish an agreed boundary can be met by either negotiation or submission of the dispute to a court or a tribunal.

Agreed maritime boundaries are final and binding, whether they are negotiated or judicially decided, as they are subject to the general principle pacta sunt servanda.5 Treaties generally rely on the assumption that certain circumstances, essential to the conclusion of the treaty, will remain unchanged. It is on the basis of those circumstances that parties reach an agreement specifying their shared expectations and pacta sunt servanda aims to safeguard these shared expectations.6 However, pacta sunt servanda does not provide that all agreements remain inviolable till the end of time.7 When circumstances leading to the conclusion of a treaty have changed and obligations under a treaty have become unduly burdensome, States can be freed from their contractual obligations through peaceful means, by virtue of a different principle: rebus sic stantibus.8 The doctrine of rebus sic stantibus allows for a treaty’s unilateral termination when an unforeseen fundamental change of circumstance has affected the essential basis of the treaty. This customary rule is codified in Article 62 of the Vienna Convention on the Law of Treaties (VCLT).9

VCLT Article 62(2)(a) explicitly excludes treaties establishing boundaries from its application. For that reason, one might assume that maritime boundaries could never be set aside due to a fundamental change of circumstances, such as the submergence of an island or the appearance of a new volcanic island. There are however, no indicators that the original customary rule excluded maritime boundary treaties. Further, when members of the International Law Commission (ILC) decided to exclude treaties establishing boundaries from the application of VCLT Article 62, they never discussed the possibility of also excluding maritime boundaries. The nature of maritime zones is different from that of land territory and the reason for excluding treaties establishing land boundaries, the need for stability, is not a necessary requirement of maritime frontiers, which generally fluctuate with changes to the coastal front.10

This article provides a doctrinal analysis of VCLT Article 62, with particular emphasis on the ‘boundary’ concept. The origin and travaux préparatoires of the provision are examined, as well as the provision’s application in relation to treaties establishing maritime boundaries. This analysis indicates that particular types of maritime boundaries may indeed be set aside by virtue of VCLT Article 62 when relevant coastal geography has undergone drastic and unforeseen changes, leading to the radical transformation of maritime entitlements under UNCLOS. The exclusion of treaties establishing boundaries only applies to boundaries delimiting sovereign territory and, consequently, does not cover boundaries delimiting the exclusive economic zone, exclusive fisheries zone, or the continental shelf.

II. Rebus Sic Stantibus: The Origin of VCLT Article 62

Gentilis, who wrote the first textbook in international law,11 introduced the maxim omnis conventio intelligitur rebus sic stantibus12 in the sixteenth century. He is believed to have been the first legal scholar to discuss the theory more generally known as rebus sic stantibus. Gentilis based his findings on the writings of civil lawyers but according to Pal of the ILC, the theory derived from canon law.13 Its existence was later affirmed by Suárez and Vattel but rejected by Grotius in the seventeenth century because he made a distinction between treaties and contracts and was reluctant to accept the termination of treaties due to changed circumstances because of derivative complications.14 Bynkershoek is also said to have rejected the theory of rebus sic stantibus in the eighteenth century, although he held that an impossibility of performance might allow for derogation from treaty obligations.15 Such a contention does not necessarily reject the doctrine of rebus sic stantibus since some members of the ILC did not distinguish between rebus sic stantibus and impossibility of performance while drafting the articles for the VCLT.16 Several members wanted to merge the two rules into one article while others suggested that even if dealt with under separate articles, they might apply in tandem.17

The existence of the customary rule rebus sic stantibus in public international law has been evidenced by State practice dating back to the eighteenth century.18 States have invoked the doctrine eo nomine or referenced a general principle allowing the termination or modification of a treaty due to changed circumstances.19 Russia raised the principle in 1870 in order to withdraw from the 1856 Treaty of Paris concerning the neutralisation of the Black Sea.20 The endeavour proved successful, despite some opposition. Russia was released from the treaty obligations at a London conference soon after declaring termination under rebus sic stantibus.21 Critics claimed that treaties could not be terminated without the acceptance of all parties but even if renegotiations led to Russia’s withdrawal, and not a unilateral denunciation, the negotiations were clearly a result of Russia invoking the principle of rebus sic stantibus.22 Some commentators maintain that the principle involves nothing more than the obligation to renegotiate23 and, if so, it served its purpose when invoked by Russia in 1870. However, rebus sic stantibus has developed through the course of time and the rule codified in VCLT Article 62 allows States to unilaterally terminate, withdraw from, or suspend the operation of a treaty.24

Another example of the invocation of rebus sic stantibus, preceding the VCLT, is Norway’s denunciation, in 1922, of its treaty with France, Germany, Great Britain and Russia.25 Representatives of the Norwegian government did not explicitly refer to the doctrine of rebus sic stantibus but they relied on altered circumstances prompted by changes in foreign politics and declared that the foundations of the treaty were lost.26 The French Government also invoked the principle of rebus sic stantibus in 1922 to terminate two treaties27 in the Nationality Decrees case,28 where it proclaimed that the principle was applicable to all treaties of indefinite duration.29 The Union of Soviet Socialist Republics (USSR) declared in 1924 that it wished to renegotiate and determine the status of old treaties with the United Kingdom (UK) since they were no longer enforceable according to the principle of rebus sic stantibus due to changed circumstances generated by World War I.30 The Soviet government again invoked the rule of rebus sic stantibus to denounce a treaty of neutrality with the Japanese in 1945. The reasoning provided was that circumstances had changed fundamentally after the German invasion of the USSR, Japan’s alignment with Germany, and Japan’s war with the United States (US) and UK.31

According to Elias, judicial decisions had never been based on the rule rebus sic stantibus in 1963 when the Commission was in the midst of codifying the principle as an objective rule of law, even though it was ‘regarded as one of the fundamental assumptions in public international law’.32 The International Court of Justice (ICJ) never explicitly rejected the existence of rebus sic stantibus before the implementation of VCLT Article 62, although it was, and still is, quite reluctant to apply the doctrine; the same is true for municipal courts.33 Vagts considers the doctrine of rebus sic stantibus to be an ‘odd corner’ of the law of treaties34 and, as such, it has been the object of scrutiny and controversy. It was quite controversial among the members of the ILC and State Representatives; their opinions on the principle ranged from complete denial of its existence to regarding it as a rule of jus cogens.35

Lukashuk, a Ukranian representative at the UN General Assembly, asserted that ‘the doctrine of rebus sic stantibus was one of the most controversial in the history of international law’ and that deciding whether or not rebus sic stantibus existed as a rule of law was a matter of great responsibility.36 Yasseen, of the ILC, derived the existence of the doctrine from the nature of written law. He said that ‘conventional rules could not be adapted ad infinitum’ and that rebus sic stantibus was a rule of jus cogens from which States could not derogate.37 Waldock, Special Rapporteur of the ILC, disagreed with Yasseen’s contentions and advised State Parties to take changing circumstances into account when negotiating, thus preventing the application of rebus sic stantibus under given circumstances.38 Another member of the ILC, Amado, called the doctrine of rebus sic stantibus the ‘serpent of the law’ as he promoted its implementation.39 According to him, the opposition to the doctrine was based on a belief in the sanctity of pacta sunt servanda, which was an outdated view, and therefore, he encouraged other members of the ILC to abandon their defensive attitudes.40 Paredes asserted that:

[n]o one could enter into an undertaking forever. That was why, under most systems of municipal law, contracts of service for life were prohibited. The same should apply even more strongly to States, because their life was much longer.41

On the opposite end of the spectrum, Gros was of the view that:

(. . .) anyone who cast doubt on the durability of treaties was contributing not to the progress of international law, but to its ruin. That theory could only lead to a regionalisation of international undertakings, not to the development of friendly relations between States.42

Despite this, Gros did not object to the implementation of what is now VCLT Article 62 and his only suggestion was that the words ‘territorial rights’ would be deleted from draft Article 22, as their meaning was unclear.43 Bartos agreed with Yasseen that rebus sic stantibus was a rule of jus cogens, although it was susceptible to criticism.44 He said that the principle of pacta sunt servanda could not apply without anomalies and that the rebus sicstantibus rule provided a necessary countermeasure to deal with instances where the literal application of pacta sunt servanda would lead to injustice and thus impair relations between States.45

Members of the ILC who objected to the implementation of a rebus sic stantibus clause did so because it represented a way to avoid a lawfully concluded treaty for the benefit of one of the contracting States. However, according to Tunkin, the existence of the rule in international law had been made evident by State practice and scholarly review.46 States generally acknowledged this fact when confronted with the principle, even if its application to the case at hand was contested.47 Special Rapporteur Waldock reached the conclusion that there was ample evidence of the principle’s general acceptance in international law that called for its codification and indicated ‘a general belief in the need for a safety-valve of this kind in the law of treaties’.48 The Commission ultimately concluded that the principle should be included in the law of treaties because treaty obligations might become too burdensome for a contracting State over time due to fundamentally changed circumstances. If such a party had no legal means of terminating a treaty without the consent of opposing parties, it might be driven to take action outside the law.49

Finally, after all this controversy, rebus sic stantibus was included in the 1969 VCLT and it has since existed both as a treaty obligation for parties to the Convention and as a customary rule of international law.50

III. Codification of VCLT Article 62

Deciding to include the principle of rebus sic stantibus in the law of treaties was only the beginning of an arduous process. Years of discussions and redrafting ensued before the ILC could reach a compromise regarding what exactly the provision should entail and how to articulate it. The article is not an exact replica of the original customary rule. Members of the Commission decided, after lengthy discussions and with regard to State commentary, to implement a modified version of the principle and this constituted, in the words of the Commission, ‘progressive development of customary international law’.51

There are two clear differences between the original customary rule and the rule codified in VCLT Article 62. First, the latter applies to treaties of both limited duration as well as perpetual treaties. Members of the Commission and other consultants however, agreed that the customary rule only applied to treaties that were inherently infinite. Second, the Commission decided to exclude treaties establishing boundaries from the scope of rebus sic stantibus to avoid dangerous friction in international relations, despite the fact that this exclusion had not existed in customary law. The Commission’s decision to exclude treaties establishing boundaries was heavily criticised throughout the codification process and it seems to have been the primary reason why twelve out of 105 State Representatives did not accept the Article when it was finally put to a vote on 13 May 1969.52

The ILC and State Representatives in delegations of the UN General Assembly discussed several versions of the Article before the final version was drafted and agreed upon. It reads as follows:

Article 62: Fundamental Change of Circumstances

  1. A fundamental change of circumstances which has occurred with regard to those existing at the time of the conclusion of a treaty, and which was not foreseen by the parties, may not be invoked as a ground for terminating or withdrawing from the treaty unless:
    1. The existence of those circumstances constituted an essential basis of the consent of the parties to be bound by the treaty; and
    2. The effect of the change is radically to transform the extent of obligations still to be performed under the treaty.
  2. A fundamental change of circumstances may not be invoked as a ground for terminating or withdrawing from a treaty:
    1. If the treaty establishes a boundary; or
    2. If the fundamental change is the result of a breach by the party invoking it either of an obligation under the treaty or of any other international obligation owed to any other party to the treaty.
  3. If, under the foregoing paragraphs, a party may invoke a fundamental change of circumstances as a ground for terminating or withdrawing from a treaty it may also invoke the change as a ground for suspending the operation of the treaty.

The rule was expressed in negative terms in order to emphasise the fact that legitimate treaties should generally remain in force despite a change of circumstances and to prevent contracting States from using it ‘as a pretext to escape from inconvenient obligations.’53 To this end, the rule was made subject to several procedural requirements.54 Some of these requirements are easily understood but others are more ambiguous and thus require further attention.

A. Fundamental Change of Circumstances

Before concluding whether VCLT Article 62 can be invoked to terminate, withdraw from, or suspend the operation of a treaty, one must first ascertain whether the circumstances have changed fundamentally from those existing at the time of the conclusion of the treaty. Special Rapporteur Fitzmaurice gave a description of what constituted an essential change55 in his second report on the law of treaties. He said that in order to be classified as an essential change, justifying the application of the principle rebus sic stantibus, a change should be objective and it should affect the factual circumstances surrounding the treaty rather than the parties’ attitudes towards the treaty.56 The change should relate to circumstances existing at the time of the conclusion of the treaty; the continuance of those circumstances should have been anticipated by both parties and have motivated them to consent to the treaty, or, at least, the specific obligations affected by the fundamental change. The change should either render the performance of treaty obligations impossible, or destroy or alter their foundation. Fitzmaurice noted two instances that would not suffice for the invocation of the principle: firstly, a change affecting the party’s will to enter into a treaty or to continue performing treaty obligations and, secondly, a change which had been foreseen.57

Reuter has pointed out that, in order to qualify as a fundamental change in circumstances under VCLT Article 62, a change must be of a qualitative and quantitative nature.58 Regarding the qualitative element, the change shall affect the facts that led both parties to give their consent to be bound by the treaty, as required by VCLT Article 62(2)(1)(a). The quantitative element is embodied in VCLT Article 62(2)(1)(b) and it provides that the change must be so extensive that it alters the conditions of the treaty and its raison d’être.59

The drying up of a river or the silting up of a harbour could qualify as a fundamental change of circumstances within the meaning of VCLT Article 62.60 Such changes would justify termination or withdrawal from a treaty, if other requirements for the application of VCLT Article 62 were also met. Waldron noted that a fundamental change in circumstances could be economic, giving examples of changes in the costs derived from treaty obligations and changes in the fiscal situation of a contracting party.61 Other examples of changes that could potentially qualify as a fundamental change of circumstance that were discussed during the codification process related to natural occurrences, such as floods or earthquakes.62 Accordingly, geographical changes, affecting the maritime entitlements of coastal States, could qualify as a fundamental change of circumstances within the meaning of VCLT Article 62.

The total extinction of one of the parties to a treaty, such as might result from the complete submergence of an island State, would not amount to a fundamental change of circumstances triggering the application of VCLT Article 62. Rather, it would be considered to constitute an impossibility of performance and consequently be subject to VCLT Article 61 or the rules on State succession if there was a change in the identity of one of the State Parties.63

B. Foreseeability

VCLT Article 62(1) provides that a fundamental change of circumstances may not be invoked as a ground for terminating, withdrawing from, or suspending the operation of a treaty unless the parties did not foresee it at the time of the treaty’s conclusion. According to Special Rapporteur Fitzmaurice this requirement meant that a fundamental change of circumstance should be one that the parties to a treaty could not have been able to anticipate with reasonable foresight.64 Consequently, a fundamental change of circumstances could not trigger the application of rebus sic stantibus if the treaty in question anticipated the change in any way, either expressly or by implied terms.65

Seemingly, Fitzmaurice used a subjective approach to determine foreseeable changes that State Parties might reference in their treaties. Indeed, the terminology of VCLT Article 62(1) does suggest that the foreseeability requirement entails a subjective, rather than an objective, test because it explicitly refers to circumstances which were ‘not foreseen by the parties’ instead of generally unforeseeable circumstances. Therefore, it seems that the ability for State Representatives, or the parties negotiating a treaty, to foresee a fundamental change of circumstances would not suffice to prevent the application of rebus sic stantibus. VCLT Article 62(1) seems to necessitate actual awareness that the fundamental change will occur, although the level of accuracy required is unclear.

The foresight requirement might actually be irrelevant under a literal interpretation of VCLT Article 62 because there is a minor flaw in the language of the provision. The paragraph expressly provides that ‘[a] fundamental change of circumstances which has occurred with regard to those existing at the time of the conclusion of a treaty, and which was not foreseen by the parties, may not be invoked as a ground for terminating or withdrawing from the treaty’ unless subsequent requirements of the article are met. Fleischhauer drew attention to the fact that the literal interpretation of the provision suggested that a fundamental change of circumstances could be invoked as a ground for terminating or withdrawing from a treaty, even if the changes were foreseen, if all other requirements were satisfied.66 Fleischhauer proposed an amendment to the article67 and so did Lissitzyn, who highlighted the same problem.68 Their efforts were to no avail because the paragraph was not amended. However, it seems clear that the ILC did not intend to allow the application of rebus sic stantibus in cases where parties had foreseen the fundamental change later invoked to terminate the treaty.69

VCLT Article 32 provides that supplementary means of interpretation, such as the preparatory work of the treaty, may be used to confirm the meaning of a treaty provision if the black letter interpretation ‘[l]eaves the meaning ambiguous or obscure’70 or ‘[l]eads to a result which is manifestly absurd or unreasonable.’71 The unclear or even unreasonable interpretation discussed above would surely justify recourse to the travaux préparatoires of the ILC and thus exclude application of rebus sic stantibus in relation to foreseen changes.

C. Essential Basis of the Treaty

A fundamental and unforeseen change of circumstances can potentially prompt unilateral termination, withdrawal from, or suspension of a treaty, only if the existence of those circumstances constituted an essential basis of the parties’ consent to be bound by the treaty.72 This is an ambiguous requirement as various factors can contribute to the creation of circumstances which lead to parties giving their mutual consent to be bound by a treaty. There is rarely a clearly defined ‘essential basis’ of a treaty and States may have different prerequisites for giving their consent. Preambles can be good indicators of what circumstances were of fundamental importance at the time of the conclusion of a treaty, but the search for an essential basis is simplified for treaties establishing maritime boundaries, as they invariably rely on coastal geography.

It seems clear that this requirement does not entail that the circumstances existing when a treaty was concluded were so essential that States would not have concluded any treaty if the circumstances had been fundamentally different. Rather, the requirement is that the fundamental change would have been taken into account when negotiating the treaty, had it been foreseen, and that different circumstances would have resulted in a substantively different treaty. Accordingly, this requirement provides that a fundamental change affecting the rights and obligations of States but relating to circumstances that were not essential at the time of the conclusion of the treaty, cannot be invoked to terminate that particular treaty.

In 1904, Bremen and Prussia concluded a treaty involving the transfer of territory, and one of the main incentives was Bremen’s desire to construct ports for the purpose of shipping.73 The treaty thus established new international boundaries between the then sovereign States and entailed on-going obligations as it stipulated that no activities relating to fisheries would be permitted in the territory ceded to Bremen. Bremen wished to set aside the prerequisite relating to fisheries due to a fundamental change of circumstance prompted by the World War, which minimised the shipping opportunities previously envisioned. Thus, the circumstances existing prior to the World War, which provided shipping opportunities for Bremen, seem to have constituted an essential basis for Bremen to be bound by the treaty. However, the German Staatsgerichtshof found, in its decision of 1925, that the principle of rebus sic stantibus could not be applied in the given case (although it was generally admissible) because the prohibition of fisheries activities on the territory ceded to Bremen was an essential requirement from the perspective of Prussia, and could therefore not be set aside without its consent.74 Consequently, what constituted an essential basis for one State was not a necessity for the other, rather an auxiliary part of the treaty, and, therefore, the change in those circumstances could not justify invocation of rebus sic stantibus. This decision indicates that a fundamental change must affect circumstances that constituted an essential basis for the consent of both or all parties to the treaty.

Regarding maritime boundaries, the circumstances forming an essential basis of an agreed maritime boundary will generally be the geographical circumstances giving rise to maritime entitlements for all relevant parties. States give their consent to honour a maritime boundary, after having established its location, by reference to geographical features in accordance with UNCLOS. A fundamental difference in coastal geography could surely affect the material content of a maritime boundary treaty and, thus, the basis for consent.

D. Radical Transformation of Obligations

VCLT Article 62(1)(b) also provides that the fundamental change of circumstance must radically transform the extent of obligations still to be performed under the treaty. This provision refers to obligations still to be performed under the treaty because the ILC wanted to clarify that the article was not applicable to treaties that had been fully executed.75 Nonetheless, the ILC did not explain what was meant by the condition that obligations were radically transformed. Arguably, it requires that obligations under the treaty have been affected to a large extent. This excludes fundamental changes that affect the essential basis of a treaty but only have an insignificant effect on treaty obligations. The obligations that are transformed by the change do not necessarily have to be primary obligations; the provision entails no requirement in that regard, but the extent of the obligations, which are still to be performed, must be radically transformed.

In order to terminate, withdraw from, or suspend the operation of a treaty establishing a maritime boundary, the fundamental change affecting the coastal geography of the relevant States must also transform the extent of the on-going obligations related to the boundary. Under UNCLOS States enjoy various rights and obligations in their maritime zones, and a maritime boundary determines the seawards extent of relevant maritime zones. However, a natural occurrence which alters a State’s maritime entitlements can increase or decrease the extent of a maritime zone, e.g. submergence of land and the appearance of new coastal features can change the ratio between land and sea in areas that have been subject to maritime boundary delimitation. States’ obligations may change as their maritime zones grow larger or smaller but their entitlements will also change in the opposite manner. This is because when maritime zones grow larger, maritime entitlements may decrease and vice versa. Thus, on-going State obligations under UNCLOS can undoubtedly be altered by changing circumstances but such changes may only be invoked as grounds for setting aside a maritime boundary if they radically transform the extent of the maritime zones (or the entitlement to maritime zones).

E. Exclusion of Boundary Treaties

According to Reuter ‘[t]here is hardly any theoretical reason why certain treaties should a priori escape a possible challenge due to a change of circumstances’.76 However, as previously mentioned, treaties establishing boundaries are excluded from this rule, according to VCLT Article 62(2)(a). A similar exception does not seem to have been attached to the customary rule, rebus sicstantibus, since the rule was invoked in relation to boundary treaties before the codification of VCLT Article 62. Further, courts did not reject its application by reference to the nature of the treaty.77 Moreover, several members of the ILC and State Representatives raised objections to the exclusion of treaties establishing boundaries, but this exclusion was implemented into the law of treaties to safeguard the stability of boundaries in order to promote peace and security in the international community.78

Special Rapporteur Waldock first proposed that an essential change of circumstances could not be invoked for the purpose of terminating treaties or provisions concerning territorial rights or boundaries in his second report on the law of treaties in 1963.79 ILC members heavily criticised this addition to Special Rapporteur Fitzmaurice’s 1957 earlier draft article.80 Elias wished to delete the entire paragraph so as not to exclude treaties establishing boundaries or transferring territory from the application of rebus sic stantibus. He noted that the Permanent Court of International Justice (PCIJ) had not put forth such a limitation when addressing the substantive elements of rebus sic stantibus in the 1932 case of the Free Zones of Upper Savoy.81

Elias recognised the importance of respect for territorial rights and frontiers but suggested that courts should be entrusted with decision making in individual cases rather than laying down a general rule in the Convention on the Law of Treaties.82 Bartos was also reluctant in accepting the Special Rapporteur’s draft article, in particular since recent examples showed that boundary lines might be drawn precisely because of circumstances existing at the time of the conclusion of the treaty, which then later changed.83 Ago did not wish to exclude these categories of treaties entirely, and recommended leaving some latitude for exceptional instances,84 although he admitted that certain dangers could ensue if boundary treaties were susceptible to termination under rebus sic stantibus.85 De Luna saw no reason to prevent the application of rebus sic stantibus to boundary treaties,86 and for that reason, he and Verdross jointly submitted a revised draft article where no mention was made of treaties relating to frontiers or territories.87 Castren doubted that the draft article proposed by De Luna and Verdross could prevent misuse of the rebus sic stantibus principle, therefore, he also proposed an alternative, simplified version of the article.88

El-Erian and the Commission’s Chairman, De Arechaga, agreed that the exclusion of treaties establishing boundaries or territorial rights could be omitted because the situation created by such treaties could not be reversed by termination.89 They maintained that the issue was adequately dealt with by Draft Article 28 which stated that the termination of a treaty could not affect the validity of any act performed, or of any right acquired, under the provisions of the treaty prior to its termination.90 Briggs also believed that executed treaty provisions could not fall within the scope of rebus sic stantibus but he still wished to explicitly exclude treaties establishing boundaries or territorial rights. This was presumably to emphasise the fact that rebus sic stantibus could not be invoked in order to invalidate already established boundaries or the completed transferral of territorial rights.91 This reasoning was based on the doctrine of executed treaties, which provides that the termination of a treaty does not undo a legal situation created through the rightful execution of treaty obligations. However, boundary treaties often entail on-going rights and obligations which would not be exempt from unilateral termination under the doctrine of executed treaties.92 Consequently, even if established boundaries are inviolable, boundary treaties cannot be excluded as a category on the grounds that they are all fully executed.

Special Rapporteur Waldock confronted and rejected the assertion that rebus sic stantibus would be inapplicable to boundary treaties because they were fully executed. He said that the purpose of paragraph 2 was not to exclude all executed treaties but only treaties fixing a boundary or affecting the transfer of territory and that the reason behind that exclusion was not that those treaties had ceased to exist, but that they belonged to a category of treaties that required stability.93 Consequently, treaties establishing boundaries would not be subject to unilateral termination or suspension, regardless of whether or not they had been fully executed.

Several members of the ILC argued that the exclusion of treaties establishing boundaries was too restrictive or incompatible with the principle of self-determination. Those arguing that it was contradictory to the principle of self-determination said that ‘[a]ny attempt to keep a treaty in force against the wishes of a people would involve a greater danger to peace than the application of the rebus sic stantibus doctrine.’94 State Representatives also noted that legal theorists, judicial or arbitral decisions, or State practice had never affirmed the exclusion of boundary treaties.95 Waldock acknowledged the widespread opposition to the exclusion of boundary treaties from State Representatives who commented on the draft article but he concluded that a majority of State Representatives endorsed this exclusion.96 He anticipated energetic attempts to invoke the principle of rebus sic stantibus in relation to boundary establishing treaties if they were not expressly excluded,97 and thus continued to promote the exclusion throughout the codification process.

The ILC never provided a clear definition of the term ‘boundary,’ but according to the 1966 Commission’s report to the UN General Assembly it encompassed treaties of cession and delimitation.98 But what exactly did the ILC mean by treaties of cession and treaties of delimitation?

Cession is the ‘transfer of sovereignty over state territory (. . .) from one state to another’.99 Consequently, treaties of cession are treaties concerning territory that is yielded or surrendered to another State and such treaties cannot be unilaterally terminated according to the ILC’s report referenced above.100 Identifying the precise meaning of the term ‘treaties of delimitation’ is more complex and more relevant for the purpose of this article. The term clearly refers to treaties delimiting land territory101 but the ILC did not clarify whether it encompasses treaties delimiting maritime zones, customs zones, airspace or any other type of zone subject to an agreed boundary.

States conclude treaties delimiting natural resources, such as migratory or trans-boundary fish stocks, and they establish exclusive fisheries zones, all of which could be categorised as treaties of delimitation. Treaties establishing boundaries to exclusive fisheries zones should be regarded as those establishing other maritime boundaries that do not delimit full sovereignty, that is, boundaries to the exclusive economic zone and continental shelf. On the other hand, rights to fish stocks can be delimited through quota allocation without establishing boundaries. Such treaties would surely fall outside the scope of the boundary exclusion and, thus, be subject to termination when all the requirements of VCLT Article 62(1) are met.

The ICJ determined whether agreed maritime limits to exclusive fisheries zones could be subject to unilateral termination due to a fundamental change of circumstances in United Kingdom v Iceland, (the Fisheries Jurisdiction case), when Iceland wished to extend its limits further seawards.102 The ICJ rejected the claim on the grounds that on-going obligations had not been radically transformed but the nature of the treaty, including, for example, the fact that it delimited natural resources, did not prevent the application of rebus sic stantibus.103 It should be noted that this case was not concerned with a boundary between the maritime zones of two States, but rather a maritime limit between Iceland’s maritime zone and the high seas. This maritime limit was subject to an agreement. The UK however, had a vested interest in the location of the maritime limit because its fishermen were exercising their right in the high seas just outside the Icelandic limits and therefore the limits were not unilaterally proclaimed by Iceland but subject to an agreement between Iceland and the UK. Consequently, the limits to Iceland’s exclusive fisheries zone were in fact much like maritime boundaries and they did delimit entitlement to natural resources but the argument relating to rebus sic stantibus was not rejected on those grounds.

France v Switzerland (the Free Zones case)104 also involved the application of rebus sic stantibus in relation to limits or boundaries, other than land boundaries. The case concerned a treaty establishing customs zones. France wished to terminate the treaty but the Swiss government argued that the principle of rebus sic stantibus could not apply to the treaty in question because of its territorial nature.105 This case was cited by the ILC in support of its eventual decision to exclude treaties establishing boundaries when it highlighted the fact that France never objected to the Swiss assertion relating to territorial treaties.106 However, the Free Zones case was also referenced by those members of the ILC who opposed the exclusion of boundary treaties because, while upholding other points on the substantive elements of the principle, the PCIJ did not confirm the assumption that boundary treaties were excluded from the application of rebus sic stantibus.107

The US delegation expressed its view, when discussing the VCLT draft articles, that it did not consider treaties of customs concessions excluded under paragraph 2(a).108 Rather, the US took the view that treaties establishing specific regimes within a territory should be excluded, and therefore, proposed an amendment to the article, also excluding treaties ‘establishing territorial status’.109 The delegation of Switzerland supported this proposal since it would mean that treaties concerning the joint utilisation of rivers forming boundaries, freedom of navigation in rivers, or rights of passage could not be terminated under rebus sic stantibus.110 It is interesting to note that the Swiss representative specifically talked about freedom of navigation in rivers but not in the sea which once again indicates that makers of the Article were thinking of land territory boundaries rather than maritime boundaries.

Waldock contemplated widening the scope of paragraph 2(a) in accordance with the proposal put forth by the US, but the Commission was adamant on keeping the final wording of the article because they did not want to place too many restrictions on the general rule of rebus sic stantibus.111 Thus, the proposal to widen the scope of paragraph 2(a) was not accepted which suggests that the types of treaties referred to by the delegations of the US and Switzerland do not fall within the scope of rebus sic stantibus.

The US delegation referred to the Antarctic Treaty112 as an example of a treaty that recognised a status quo or created a regime, a category of treaties that was not excluded under paragraph 2(a) but should be. The Australian delegation agreed with that contention and said that the Antarctic Treaty did not establish boundaries or a territorial status, and consequently, would not be excluded under paragraph 2(a). The treaty set up a system of scientific co-operation and demilitarisation, which, according to the Australian delegation, meant that it established a special regime for a particular delineated area, and was subject to termination upon fundamentally changed circumstances. The same description could be applied to treaties delimiting maritime zones; they set up special regimes within defined areas and should therefore, in accordance with the arguments advanced by the Australian delegation, not fall within the scope of paragraph 2(a).113

IV. Application of VCLT Article 62 in Relation to Maritime Boundaries

It seems obvious that VCLT Article 62 may not be invoked to withdraw from, terminate, or suspend the operation of treaties delimiting land boundaries, but the same does not necessarily apply to treaties delimiting maritime boundaries.

There is an inherent difference between boundaries delimiting land territory and those delimiting maritime zones. Marston has noted that maritime delimitations are different from land delimitations in several factual and juridical ways.114 One of those differences is that some maritime boundaries are merely lines of allocation and consequently not considered boundaries within the meaning of VCLT Article 62(2)(a).115 According to Marston ‘[t]he word “boundary” is appropriate to describe a line which delimits the contiguous land areas of states’ and Jennings, former president of the ICJ, makes a distinction between boundaries and frontiers which establish zones.116 These definitions imply that boundaries only delimit land territory and that so-called maritime boundaries are actually frontiers rather than boundaries.

Caflisch has also expressed his view that the term ‘boundary’ should be used only for land territory and other areas under full sovereignty and that the terms ‘limits’ or ‘lines of delimitation’ are more appropriate for maritime zones where States exercise functional powers and not full sovereignty.117 An arbitral tribunal reflected this distinction in the Guinea v Guinea-Bissau arbitration,118 and treaties establishing maritime boundaries often do the same, using terms like ‘sea frontier’ or ‘maritime frontier’.119 The sovereignty of States extends towards the outer limits of territorial or archipelagic waters.120 Therefore, lines delimiting such maritime zones could fall within the boundary concept and be subject to VCLT Article 62(2)(a). However, States only enjoy sovereign rights to explore and exploit specific resources in the exclusive economic zone and continental shelf, not full sovereignty, so this is one more reason why boundaries delimiting such zones could be subject to termination or suspension due to a fundamental change of circumstances.

Land and maritime boundaries also differ in the way they are demarcated and the principles they are founded on. The location of land boundaries can be completely arbitrary but once established, the significance of land boundaries is vested in their duration.121 After all, treaties establishing land boundaries belong to a category of treaties that demands stability.122 On the other hand, maritime boundaries are based on maritime entitlements guaranteed to all coastal States under UNCLOS and these entitlements are dependent upon the presence and exact location of geographical features.123 Maritime boundaries represent an equitable solution,124 rather than a stable one. In fact, maritime entitlements change regularly, in accordance with changes to coastal geography, and maritime limits are generally understood to be ambulatory.125 Therefore, the reasons for excluding treaties establishing boundaries from the application of VCLT Article 62 do not seem to justify a general exclusion of maritime boundaries.

Stoutenburg is among those arguing that maritime boundary treaties cannot be terminated under VCLT Article 62.126 She maintains that the ICJ gave a clear answer in Greece v Turkey (the Aegean Sea Continental Shelf case),127 confirming that all agreements delimiting the continental shelf fall within the exception provided for in VCLT Article 62(2)(a).128 The Court stated that:

[w]hether it is a land frontier or a boundary line in the continental shelf that is in question, the process is essentially the same, and inevitably involves the same element of stability and permanence, and is subject to the rule excluding boundary agreements from fundamental change of circumstances.129

It should be noted that this was an obiter dictum and no arguments relating to rebus sic stantibus were raised in the case. Moreover, the statement is not very convincing because the process of establishing land frontiers and maritime boundaries is actually quite different because maritime boundary delimitation is based on legal entitlement and governed by a clear set of rules whereas land frontiers are settled arbitrarily. The Court seems to indicate that similarities between the processes concerning land and maritime boundaries should lead to the same result in terms of stability and permanence. Even if there were a correlation between the process behind the construction of a boundary and the determination as to whether or not the resulting boundary would be subject to the exclusion of VCLT Article 62(2)(a), maritime boundaries should not be equated to land boundaries because the processes differ.

It is important to note that the Aegean Continental Shelf case only dealt with boundary lines in the continental shelf, not lines delimiting maritime zones in general. This holds considerable significance because there is an important distinction between lines delimiting the outer limits of the continental shelf and other maritime zones as the continental shelf is land territory, beneath the ocean, while other maritime zones delimit water columns. Thus, boundaries between the continental shelves of two States could fall within the ‘boundary’ concept, as well as territorial sea boundaries, although frontiers of other maritime zones would not. However, this seems unlikely, as the ICJ’s rationale does not change the fact that continental shelf boundaries do not delimit full sovereignty and are, as such, inherently different from land boundaries and territorial sea boundaries.

The tribunal in the Bangladesh v India (Bay of Bengal) arbitration expressed its view that maritime boundaries should be stable and definitive like land boundaries, and it referenced the ICJ’s decision in the Temple of Preah Vihear case, saying that ‘when two countries establish a frontier between them, one of the primary objects is to achieve stability and finality.’130 The tribunal asserted that this should apply equally to land and maritime boundaries, whether they were established through agreements or adjudication, and further declared that ‘neither the prospect of climate change nor its possible effects can jeopardize the large number of settled maritime boundaries throughout the world.’131 This recent decision cannot be overlooked since it clearly implies that changed circumstances cannot be invoked to terminate maritime boundary treaties. However, the tribunal here obviously assumed that the fundamental basis of maritime delimitation was the same as that of land boundaries: acquiring stability. This is contrary to the primary purpose of maritime delimitation however, because as previously explained, the primary purpose is the achievement of an equitable solution132 and not continued stability. Therefore, treaties delimiting maritime boundaries should not automatically be put in the same category as those establishing land boundaries.

It should also be noted that the tribunal’s reasoning in the 2014 Bay of Bengal decision was based largely on the fact that the exploration and exploitation of natural resources in the continental shelf generated great expense and that stability should be ensured to pave the way for further development and investment.133 The fact that the termination of a maritime boundary is costly and inconvenient for a State however, should not suffice to make VCLT Article 62(2)(a) applicable to all maritime boundaries. Similar arguments could be advanced in most instances where one State invokes rebus sic stantibus to terminate a treaty because the purpose of the rule is to force unilateral termination against the will of a State that would benefit from the on-going application of a treaty that has lost its essential basis.

Lisztwan has expressed doubts as to whether VCLT Article 62 can be applied to terminate or suspend the operation of maritime boundary treaties.134 She has raised four plausible arguments supporting her theory which will be addressed in the following discussion. Firstly, Lisztwan has noted that international courts and tribunals have shown some reluctance to employ the principle embodied in VCLT Article 62.135 Secondly, she maintains that VCLT Article 62(2)(a) covers treaties establishing maritime boundaries.136 Thirdly, she anticipates difficulty in proving that the locations of baselines were an essential basis for parties giving their consent to be bound by a maritime boundary treaty.137 Her final argument is that there would be great difficulty in proving that changes in coastal geography were not foreseen.138

There is very little judicial precedent for the application of VCLT Article 62. When drafting VCLT Article 62, Kabbaj, a Morrocan representative at the UN General Assembly, cautioned that the Article was encompassed by such stringent requirements that States had nothing to fear regarding its application.139 He may have been right and these stringent requirements are perhaps the reason for the apparent reluctance of courts and tribunals in basing their decisions on VCLT Article 62. However, despite of the provision’s strict conditions and the fact that it is phrased in a negative manner, it was, in the view of the ILC, a necessary tool that could extricate States from treaties that had become too burdensome and irrelevant due to a fundamental change of circumstances.140 The Article is still in force and its rare use, although possibly suggestive as to the willingness of courts and tribunals to apply it, should not affect the interpretation of its material content or its enforceability.

Lisztwan draws the assumption that the term ‘boundaries’ encompasses maritime boundaries, from the drafting history of the Article, although she admits that the drafting history of VCLT Article 62 provides little guidance regarding the scope of the boundary exception.141 Lisztwan notes that the Ukrainian delegation asserted that the exclusion of treaties establishing boundaries would cover island disputes.142 This however, did not explicitly relate to maritime boundaries surrounding islands. In fact it seems more likely that the reference to island disputes related to disputes concerning sovereignty over islands. Such disputes would affect land boundaries and consequently be excluded. Actually, members of the ILC and State Representatives never mentioned maritime boundaries in their lengthy discussions about VCLT Article 62 and when discussing islands, they only referred to the land territory that constituted the island, never the surrounding maritime zones.

In relation to the drafting history of VCLT Article 62, Lisztwan references the definition of boundaries put forth by Oppenheim but Kearney, the American representative, advanced the same definition, as he argued that the provision would not include treaties otherwise establishing territorial status.143 Oppenheim defined State boundaries as ‘the imaginary lines on the surface of the earth which separate the territory of one State from that of another, or from unappropriated territory, or from the open sea’.144 Oppenheim made a distinction between boundaries between the territories of two or more States, boundaries between a State and unappropriated territory, and boundaries where territory of a State meets the ocean. His definition implies that boundaries are either drawn between States’ territories or where territory ends, which would be at the baselines of coastal States or the outer limits of the territorial sea, depending on whether or not the territorial sea forms part of a State’s territory. It is unclear whether Oppenheim’s reference to ‘the open sea’ applies to specific maritime zones, the high seas beyond national jurisdiction, or the ocean in general, but if the term ‘territory’ were to cover maritime zones then the reference to ‘the open sea’ would be wholly unnecessary as it could be classified as ‘unappropriated territory’. Maritime boundaries would then fall within the first category (separating the territory of two States) and boundaries between maritime limits and the high seas would fall within the second category (separating territory from unappropriated territory). Similarly, if ‘the open sea’ referred only to the high seas, and the lines between all maritime zones fell within the boundary concept, then there would be a need for a fourth category (boundaries between different maritime zones). Thus, Oppenheim’s definition assumes that ‘territory’ is either: (a) land territory; or (b) land territory, internal waters, archipelagic waters, and the territorial sea. Further, ‘the open sea’ incorporates: (a) all maritime zones beyond baselines (which would mean that the territorial sea could not be territory); or (b) only the sea beyond the territorial sea. According to this maritime boundary agreements are not subject to the exclusion in paragraph 2(a).

Lisztwan further submits that the fact that States rarely explicitly refer to the essential basis of their decision to consent to a treaty makes it difficult to determine whether or not coastal geography has constituted an essential basis of the parties’ consent to be bound by a treaty establishing maritime boundaries.145 However, one can assume that natural catastrophes, relating to the material content of a treaty, would have a critical effect on the decision to consent to that treaty. As explained in section III.C above, coastal geography is an essential foundation for all treaties establishing maritime boundaries because the boundaries are determined on the basis of geographical features, which generate the maritime entitlements that require delimitation.

When discussing the codification of rebus sic stantibus, the Canadian government suggested a modification to paragraph 2(a) specifying that boundary treaties were excluded from the application of the Article, except for those based on natural features that were relocated by virtue of natural occurrences. The relocation of a thalweg of a river was mentioned as an example of what could potentially qualify as a fundamental change of circumstance in relation to a land boundary established by reference to said thalweg.146 Special Rapporteur Waldock acknowledged that a natural occurrence, such as a flood or an earthquake, could cause the relocation of a river channel or other natural feature. However, he rejected the proposal put forth by the Canadian delegation, and expressed his view that such changes might not qualify as fundamental changes leading to termination. Instead, he believed that natural occurrences altering the location of geographical features, might affect the interpretation of treaties based on such features.147 Lisztwan submits that this demonstrates a conscious decision to exclude termination of boundary agreements on the basis of changes to the geographical features they were dependent upon. She rightly notes that State Representatives must have been aware that such features could undergo changes and that boundary agreements could be formulated and interpreted in such a way that boundaries could fluctuate in accordance with the natural phenomena they were based upon.148 However, this does not prove that a fundamental change to coastal geography could not affect the essential basis of a maritime boundary in such a way as to justify its termination under VCLT Article 62.

Furthermore, the altered interpretation or application of a maritime boundary treaty would be subject to either the mutual consent of all relevant States or to judicial settlement. Thus, Waldock’s solution would not be able to bring about peaceful change unless a court or a tribunal had jurisdiction to settle a dispute relating to the interpretation or application of a maritime boundary treaty or both parties agreed to alter the treaty and alternatively submit the dispute to judicial settlement. Rebus sic stantibus does not rely on the willingness of all States to settle their disputes because it assumes that a State benefitting from a fundamental change of circumstances which affects the essential basis of a treaty will not wish to terminate that treaty.

Lisztwan’s fourth point, in support of her contention that rebus sic stantibus cannot be invoked to terminate maritime boundary agreements, relates specifically to whether States could be able to anticipate or foresee a fundamental change of circumstances.149 She submits that even if a maritime boundary were affected by changed circumstances, States would rarely be able to prove that they had not foreseen natural changes to the coastline. Lisztwan addresses climate-related changes in particular and asserts that the effects of climate change could have been foreseen after 1980 when knowledge about climate change became widespread.150 However, as discussed in section III.B, VCLT Article 62(1) entails a subjective test, which requires that parties to a treaty actually foresaw the change in question, not that it was objectively foreseeable. This article is not specifically concerned with climate-related changes but it should be noted that such changes could not have been anticipated with any precision years or decades ago because scientific predictions have only recently become available, and they are subject to continuing updates. The Intergovernmental Panel on Climate Change first published an assessment report in 1990, enabling States to foresee the possible consequences of climate change,151 and this assessment has changed in subsequent reports.152 Moreover, even if States take all the most recent climate-related predictions into account when establishing maritime boundaries they cannot foresee exactly how their coastal geography will be affected. For example, the effects of extreme weather events cannot be foreseen with any precision and the same goes for seismic and volcanic activity. States will not always be able to predict the severity or precise character of future changes and VCLT Article 62 should be available for situations where unforeseeable fundamental changes occur and radically affect the essential basis of maritime boundary treaties.

V. Conclusion

VCLT Article 62 represents an important mechanism in international law, freeing States from treaty obligations that they should not be held to, once a treaty’s foundations have been shattered. Treaties that demarcate boundaries to State sovereignty require absolute stability and are, as such, excluded from the application of this otherwise general rule.153 The purpose of the boundary exclusion is to safeguard the sanctity of boundaries and to promote peace and security in the international community.154 However, this exclusion could be a source of dangerous friction if it were to cover all treaties establishing maritime boundaries because they hardly ever have provisions relating to termination155 and, thus, can never be set-aside without the will of both parties or through the operation of law.156

The nature of land and maritime boundaries is inherently different as they have distinct functions and follow different principles. Land boundaries seek to achieve stability while maritime boundaries represent an equitable solution. The entitlements to land territory are arbitrary and founded on permanent occupation while maritime entitlements are subject to continuously applicable requirements of UNCLOS. Land boundaries delimit full sovereignty, unlike all maritime boundaries except those that delimit the territorial sea, and the terminology most commonly used to describe the lines delimiting maritime zones cannot, in itself, suffice to put maritime boundaries in the same category as land boundaries.

Climate-related changes are now altering coastal geography worldwide. Some of these changes may be anticipated but predictions change rapidly and the precise effects of sea level rise, coastal erosion, coastal accretion and land rise are difficult to predict. Furthermore, some changes occur with little warning, such as extreme weather events, floods, hurricanes, seismic activity and volcanic eruptions, and the often-devastating effects of such fundamental changes cannot be foreseen when establishing all maritime boundaries.

The eventual conclusion on whether a fundamental change of circumstances can justify the termination of a maritime boundary will always depend on the specifics of each case but, as argued by the French government in the Nationality Decrees case, a change of circumstance justifying the application of rebus sic stantibus is most manifest when the change deprives a regime of its raison d’être.157 A treaty affording rights and obligations to States on the basis of coastal features must lose its raison d’être and thus its enforceability when relevant coastal features disappear or otherwise cease to generate maritime entitlements. This is especially true when the coastal features that generated overlapping claims, and thereby the need for a maritime boundary, have been affected. The ILC once noted that:

[t]he fact that international law recognized no legal means of terminating or modifying the treaty otherwise than through a further agreement between the same parties might impose a serious strain on the relations between the States concerned; and the unsatisfied State might ultimately be driven to take action outside the law.158

This is precisely why rebus sic stantibus is a necessary safety-valve which should be applied to facilitate equitable changes to maritime boundaries in times of shifting coastlines.