The American geographer Stephen Jones in his still-influential guidebook on international boundary-making gave the traditional descriptions of accretion and avulsion. However, he was clearly uncomfortable with them and concluded: the doctrine that the boundary follows only accretionary changes is not always a good one. A meandering river on a flood plain will, given time, shift avulsively in every part of its course, unless restrained by engineering works. If the boundary remains unchanged, ultimately it will lose all relationship to the actual river.
Even in the two cases relating directly to river boundary definition, Botswana / Namibia and Benin / Niger, the International Court of Justice (ICJ) made no direct mention of accretion or avulsion in its decisions. Since the Chamizal arbitration, the closest any international court or tribunal has come to voicing an opinion on accretion and avulsion was the Chamber of the ICJ’s 1992 and 2003 decisions in relation to the Land and Maritime boundary case between El Salvador and Honduras. Concerning the western extremity of the boundary, El Salvador originally claimed that the boundary had been unaffected by an avulsion event on the Goascoran river in 1821 and therefore remained fixed in an older bed of the river. In its initial 1992 decision, the Chamber found that El Salvador had not produced evidence to specify the exact date and event when the avulsion took place and avoided drawing any conclusions about the legal validity of accretion and avulsion by using ambiguous language: were the Chamber satisfied that the river’s course was earlier so radically different from its present one, then an avulsion might reasonably be inferred. While the area is low and swampy, so that different channels might well receive different proportions of the total run-off at different times, there does not seem to be a possibility of the change having occurred slowly by erosion and accretion, to which, as El Salvador concedes, different legal rules may apply.
The Chamber also adjudged that as the alleged avulsion took place in 1821 (prior to both states’ independence) its effect on the boundary would be governed by Spanish colonial law rather than international law. Again, the Chamber deftly avoided any comment on the validity of avulsion in international law:
On this basis, what international law may have to say, on the question of the shifting of rivers which form frontiers, becomes irrelevant: the problem is mainly one of Spanish colonial law. In fact the alleged rule originated in Roman law as a rule applicable to private property, not as a rule relating to rivers as boundaries of jurisdiction and administration. Furthermore, whatever its status in international law – a matter to be determined, if necessary, by the Chamber, on the basis of the principle of jura novit curia – its possible application to the boundaries of Spanish colonial provinces would require to be proved.
Ten years later, El Salvador asked the Chamber to revise its decision, claiming to possess new evidence that pinpointed the exact date of avulsion. Discounting the application for revision, the Chamber again dodged the issue of avulsion by rejecting the new evidence and indicating that its original decision had not been made on the avulsion argument in the first place: “Even if avulsion were now proved, and even if its legal consequences were those inferred by El Salvador, findings to that effect would provide no basis for calling into question the decision taken by the Chamber in 1992 on wholly different grounds”. The careful wording of both decisions indicates that the ICJ will deal with the possible legal validity of avulsion on a case-by-case basis thereby eroding any notion that avulsion is an accepted and mandatory principle of international law. Instead it is a rule that may or may not be applied by respective courts or tribunals at their discretion.
However, without being defined in international convention or consistently used in state practice ( the majority of international boundary treaties (particularly older and often colonial boundary treaties) are silent on what should happen to the boundary when a river moves), accretion and avulsion clearly have not achieved the high threshold for becoming mandatory rules of international law.