There is no serious doubt that the presence of islands may constitute a relevant circumstance in maritime delimitation. However, State practice is so diverse that it is difficult to specify a general rule with respect to the legal effect given to islands. Thus, international courts and tribunals are to decide the effect given to islands within the framework of equitable principles. In broad terms, four modes of effect given to islands can be identified in case law.
The first mode is to give full effect to an island. In the Qatar/Bahrain case, for instance, the ICJ gave full effect to the Hawar Islands and Janan Island when drawing an equidistance boundary line. In the Nicaragua/Honduras case, Honduran islands in the Caribbean Sea – Bobel Cay, Port Royal Cay, Savanna Cay and South Cay – and Nicaragua’s Edinburgh Cay were given full effect. In relation to this, it is of particular interest to note that the Arbitral Tribunal, in the Eritrea/Yemen case (Second Phase), presented the ‘integrity test’ as a criterion for determining the effect given to islands. According to this approach, where relevant islands constitute an integral part of a mainland coast, full effect may be given to those islands. This criterion was expressed in relation to the Dahlaks, a tightly knit group of islands and islets belonging to Eritrea. The Arbitral Tribunal gave full effect to the Dahlaks because it ‘is a typical example of a group of islands that forms an integral part of the general coastal configuration’. For the same reason, full effect was given to the Yemeni islands of Kamaran, Uqban and Kutama. To a certain extent, the integrity test would seem to present a useful criterion for determining the effect to be given to islands.
The second mode is to give no effect to an island. For instance, the ICJ, in the Tunisia/Libya case, neglected the island of Jerba, which is separated from the mainland by a very narrow strait, in drawing a boundary. In the Guinea/Guinea-Bissau case, the Arbitral Tribunal gave no effect to coastal islands, the Bijagos Islands and Southern Islands. In the Qatar/Bahrain case, the Court gave almost no effect to Qit’at Jaradah, by drawing a delimitation line passing immediately to the east of Qit’at Jaradah. Similarly, the Court decided that Fasht al Jarim should have no effect on the boundary line in the northern sector. By referring to the integrity text expressed in the 1999 Eritrea/Yemen award, the ICJ, in the Black Sea case, gave no effect to Serpents’ Island belonging to Ukraine.
The third mode involves the enclave solution. This method was adopted by the Court of Arbitration in the Anglo-French Continental Shelf case with regard to the Channel Islands. The Channel Islands, which are under British sovereignty, lie off the French coasts of Normandy and Brittany. The Court of Arbitration took the view that giving full effect to the Channel Islands would constitute a circumstance that would entail inequity. Thus the Court adopted a twofold solution. First, as the primary boundary, the Court drew a median line between the mainlands of the two States. Second, it created a 12-mile enclave to the north and west of the Channel Islands. While a precedent for the enclave solution could be found in the delimitation of lakes, there was no precedent in the context of maritime delimitation before this award. It seems, therefore, that the enclave solution is a novel creation of the Court of Arbitration.
Subsequently, an enclave with a 12-nautical-mile arc of the territorial sea was adopted by the ICJ. In the Black Sea case, the Court attributed only a 12-nautical-mile arc of the territorial sea to Serpents’ Island. Although the Court did not decide the legal status of Serpents’ Island, the Court’s solution seems to produce the same effect as if Serpents’ Island were treated as a rock. Likewise, the Court, in the 2007 Nicaragua/Honduras case, ruled that the Honduran islands of Bobel Cay, Savanna Cay, Port Royal Cay and South Cay should be accorded a territorial sea of 12 nautical miles. It thus traced the 12-nautical-mile arc of the territorial sea around the south of Bobel Cay until it reached the median line in the overlapping territorial seas of Bobel Cay, Port Royal Cay and South Cay (Honduras) and Edinburgh Cay (Nicaragua). Also the Court traced the arc of the outer limit of the 12 nautical-mile territorial sea of South Cay round to the north until it connected with the bisector line.
It would seem to follow that the Court’s solution produced the same effect that these cays are treated as rocks under Article 121(3) of the LOSC. In the 2012 Nicaragua/Colombia case, the effect given to Quitasueño and Serrana was at issue. Quitasueño is a large bank approximately 57 kilometres long and 20 kilometres wide. Among multiple marine features at Quitasueño, only one feature named QS 32, which is barely one square metre in area, is above water at high tide only by some 0.7 metres. The Court regarded Quitasueño as a rock under Article 121(3) of the LOSC and established a 12-nautical-mile envelope of arcs around Quitasueño. Likewise it established only a 12 nautical-mile envelope of arcs measured from Serrana Cay and other cays in its vicinity in light of its small size, remoteness and other characteristics. The ICJ jurisprudence seems to suggest that all maritime features which remain above water at high tide may be attributed at least 12 nautical miles of the territorial sea, regardless of their legal status.
The fourth mode is to give only partial effect, such as half effect, to islands in drawing a maritime boundary.
In the Anglo-French Continental Shelf case, the Court of Arbitration took the view that the projection westwards of the Scilly Isles constituted a special circumstance. The Court thus determined to give the Scilly Isles half effect. The distance between the Scilly Isles and the mainland of the United Kingdom is twice that separating Ushant from the French mainland. For the Court, this was an indication of the suitability of the half-effect method. Accordingly, the Court drew, first, an equidistance line without using offshore islands as a base point and, next, an equidistance line using them as a base point. A boundary line was then drawn midway between those two equidistance lines.
The half-effect solution was also adopted in the Tunisia/Libya judgment. A line drawn from the most westerly point of the Gulf of Gabes along the seaward coast of the Kerkennah Islands would run at a bearing of approximately 62± to the meridian. However, the Court considered that the line of 62± to the meridian, which runs parallel to the coastline of the islands, would give excessive weight to the Kerkennahs. For that reason, the Court decided to attribute half effect to the Kerkennah Islands. It did so by drawing a line bisecting the angle between the line of the Tunisian coast (42±) and the tangent of the seaward coast of the Kerkennah Islands (62±). Consequently, a line of 52± to the meridian was to be the boundary of the continental shelf in this area.
In the Gulf of Maine case, the Chamber of the ICJ determined to give the Canadian territory of Seal Island half effect. Specifically, the Chamber drew Seal Island back to half its real distance from the mainland. The distance between Seal Island and Chebogue Point in Nova Scotia is 14,234 metres. Dividing this distance by two, a position of 7,117 metresfrom Chebogue Point would represent a notional half-effect position for Seal Island. More recently, only half effect was given to the Corn Islands in the Costa Rica/Nicaragua case.
Overall, as ITLOS pointedly observed, ‘neither case law nor State practice indicates that there is a general rule concerning the effect to be given to islands in maritime delimitation’.
According to the Tribunal:
The effect to be given to an island in the delimitation of the maritime boundary in the exclusive economic zone and the continental shelf depends on the geographic realities and the circumstances of the specific case. There is no general rule in this respect. Each case is unique and requires specific treatment, the ultimate goal being to reach a solution that is equitable.
