Can Sea-Level Rise Meet the Risk Threshold for Non-Refoulement under International Law?
Introduction
International law has traditionally responded to sudden forms of danger such as persecution, armed conflict, torture, and political violence. It is less settled when harm develops gradually over time. Climate change increasingly forces legal systems to confront this difficulty. Among its many consequences, sea-level rise presents one of the clearest long-term threats to human habitation. Low-lying coastal regions and small island states are already experiencing flooding, erosion, saltwater intrusion, and damage to homes, livelihoods, and public infrastructure.
This creates an important legal question. If a person leaves because their home is progressively becoming unliveable, can international law prevent their return to unsafe conditions? Traditional refugee law was largely shaped by twentieth-century experiences of persecution and conflict. It offers only limited guidance where displacement is linked to environmental collapse. At the same time, the United Nations High Commissioner for Refugees has recognised that climate-related movement is an urgent protection concern.
One possible route through this gap lies in the principle of non-refoulement, which may provide protection where return would expose an individual to a serious risk of harm. McAdam's analysis of climate change-related displacement and complementary protection examines how existing international legal frameworks may offer protection to persons affected by environmental change (Jane McAdam, Climate Change, Forced Migration and International Law (Oxford University Press 2012) 22-30). This principle prohibits returning individuals to places where they face serious harm. This article argues that sea-level rise can, in appropriate circumstances, satisfy the legal risk threshold necessary to trigger non-refoulement obligations under international law.
Sea-Level Rise and the Erosion of Habitability
Sea-level rise should not be understood only as the physical disappearance of territory beneath water. Long before complete submersion occurs, rising seas may erode the conditions necessary for ordinary life. Saltwater intrusion can contaminate freshwater reserves. Repeated flooding may damage roads, homes, hospitals, and schools. Agricultural land may become less productive, undermining food security and livelihoods.
The World Bank Groundswell reports have highlighted the growing relationship between climate stress and human mobility. These concerns are particularly acute in states with limited land and constrained adaptation capacity. Countries such as Kiribati, Tuvalu, the Marshall Islands, and the Maldives are frequently cited in discussions on climate vulnerability.
The more precise legal issue, therefore, is habitability, a concept explored by Susin Park in her analysis of climate change and the risk of statelessness (Susin Park, Climate Change and the Risk of Statelessness (UNHCR Legal and Protection Policy Research Series 2011) 14-18). A place may continue to exist geographically while becoming increasingly unsafe socially and economically. If return would expose a person to conditions incompatible with dignity and survival, the absence of total territorial disappearance should not end the legal inquiry.
The Risk Threshold in Non-Refoulement Law: Three Distinct Doctrinal Sources
Non-refoulement is not a single, uniform rule. It arises from three separate treaty regimes, each with its own threshold, scope, and institutional guardian, and the differences between them matter directly to a sea-level rise claim.
The first and narrowest source is Article 33 of the 1951 Refugee Convention, which prohibits returning a refugee to a territory where life or freedom would be threatened. Critically, this protection is tied to the Convention's own definition of a refugee: it applies only where the threatened harm stems from persecution on the grounds of race, religion, nationality, membership of a particular social group, or political opinion. As Goodwin-Gill and McAdam have observed, this persecution-based architecture was built around targeted, human-driven harm, and does not naturally accommodate displacement caused by an impersonal environmental process (Guy S Goodwin-Gill and Jane McAdam, The Refugee in International Law (3rd edn, Oxford University Press 2007) 345-350). A person fleeing sea-level rise, without more, is unlikely to satisfy the persecution nexus and therefore falls outside Article 33 in most cases.
The second source, non-refoulement under the Convention Against Torture (CAT), is narrower in one respect and broader in another. It requires no persecution nexus at all, but Article 3 requires substantial grounds for believing that the individual would be in danger of being subjected to torture upon return. The CAT Committee assesses this risk as foreseeable, personal, present and real. Although the Convention separately addresses cruel, inhuman or degrading treatment under Article 16, the non-refoulement obligation under Article 3 is specifically concerned with the risk of torture. Slow-onset environmental degradation, even where severe, therefore sits uneasily within this framework because it is diffuse and does not ordinarily involve the intentional infliction of severe pain or suffering contemplated by the definition of torture.
The third source is Article 6 of the International Covenant on Civil and Political Rights, protecting the right to life, as interpreted by the UN Human Rights Committee. This is the route most relevant to sea-level rise, because it requires neither a persecution nexus nor proof of an intentional, targeted act. It asks only whether removal would expose the individual to a real risk of arbitrary deprivation of life, whatever the source of that risk. Because it is actor-neutral and outcome-focused, the ICCPR right-to-life framework is structurally best suited to environmental harm, and it is this framework, rather than the Refugee Convention or CAT, that carries the greatest doctrinal weight in the analysis that follows.
Despite these structural differences, all three frameworks share a common threshold inquiry: is there a real, personal, and foreseeable risk of serious harm upon return? Non-refoulement is not engaged by every hardship. Economic disadvantage or general instability may be insufficient. Yet the law does not require certainty of death or total collapse.
Comparatively, the European Court of Human Rights has also emphasised in its Article 3 non-refoulement jurisprudence that removal must be assessed against the actual conditions facing the individual concerned. While this regional framework is distinct from the three treaty regimes discussed above, its emphasis on individualised risk assessment reinforces the importance of examining the concrete circumstances of a climate-related claim. Once framed in this way, climate harms cannot be dismissed merely because they develop slowly or affect wider populations.
Can Sea-Level Rise Meet That Threshold?
In principle, yes, and most persuasively under the ICCPR right-to-life route identified above. The purpose of non-refoulement is protective rather than categorical. It seeks to prevent return to serious harm. If the effects of sea-level rise create conditions incompatible with safe and dignified life, the environmental source of that harm should not be decisive.
Consider a person returned to a place facing recurrent flooding, unsafe drinking water, destroyed housing, and no realistic means of livelihood. The danger in such a case is neither abstract nor remote. It concerns health, bodily security, and minimum subsistence. The UN Human Rights Committee has recognised that environmental degradation may, in certain circumstances, implicate the right to life, and has explicitly stated that state parties may be in violation of their obligations by removing individuals to conditions that expose them to such risk (Human Rights Committee, General Comment No 36 on Article 6 of the ICCPR, UN Doc CCPR/C/GC/36 (2018)).
There is also a fairness dimension. Communities most exposed to sea-level rise have often contributed least to historic greenhouse gas emissions. While non-refoulement is not a reparations mechanism, refusing even minimal protection in severe cases would deepen existing inequalities. This concern has become increasingly visible in negotiations under the UN Framework Convention on Climate Change, particularly in discussions on loss and damage.
States may nonetheless object that recognising climate-based claims could create indeterminate migration obligations. That concern should be taken seriously, but it rests on a slippery-slope premise that the doctrine itself does not bear out. The ICCPR right-to-life threshold, unlike a general humanitarian discretion, already builds in three limiting requirements that prevent open-ended liability: the risk must be real rather than speculative, personal to the individual rather than a description of general regional hardship, and reasonably foreseeable at the point of removal rather than a distant contingency. Economic disadvantage, discomfort, or the mere fact of living in a climate-vulnerable state will not, on their own, satisfy this standard; only conditions that concretely threaten survival will. This is precisely how the UN Human Rights Committee approached Teitiota's claim, and precisely why it was rejected on the facts even while the legal principle was accepted. The objection therefore misdescribes the doctrine: individualised assessment is not a weaker alternative to a rule-based bar on climate claims, it is the mechanism by which the risk threshold is already self-limiting.
Teitiota v New Zealand: A Doctrinal Turning Point
The most significant authority in this area remains Teitiota v New Zealand, decided under the ICCPR right-to-life framework rather than the Refugee Convention, which reinforces the doctrinal point made above. Ioane Teitiota, a national of Kiribati, argued that removal from New Zealand would expose him to climate-related dangers in Kiribati.
The UN Human Rights Committee rejected the claim on the specific facts, concluding that the immediate threshold had not yet been met. However, that is only part of the decision's significance. The Committee expressly accepted that environmental degradation and climate change can, in principle, trigger non-refoulement obligations under Article 6 where the right to life is seriously threatened (Teitiota v New Zealand CCPR/C/127/D/2728/2016 (7 January 2020) paras 9.11-9.12).
Accordingly, Teitiota should not be read as closing the door on climate claims. Rather, it establishes that such claims are legally cognisable under the ICCPR specifically, with outcomes depending on the quality of evidence and the severity of risk. Its significance lies not in the outcome for Mr Teitiota, but in the Committee's confirmation that the correct doctrinal home for sea-level rise claims is the right to life, not the persecution-based Refugee Convention or the intent-based Convention Against Torture.
Conclusion
Sea-level rise exposes a limitation in traditional refugee protection frameworks, but it also demonstrates the continued relevance of non-refoulement as a flexible, multi-source principle capable of responding to new forms of harm. Of its three doctrinal channels, the persecution-based Refugee Convention and the intent-based Convention Against Torture are poorly suited to slow-onset environmental harm, while the actor-neutral right to life under Article 6 of the ICCPR offers the most credible route to protection. Where rising seas create conditions that threaten access to water, shelter, health, or minimum subsistence, the required risk threshold under international law may be satisfied through this route.
Its gradual nature should not operate as an automatic barrier. Serious harm can emerge slowly, just as it can emerge suddenly. If international law is to remain credible in an era of climate disruption, it must be able to recognise that reality. Non-refoulement, correctly understood through its ICCPR channel, provides one principled avenue through which that evolution can occur, a trajectory Walter Kalin anticipated in his early conceptual work on climate-induced displacement (Walter Kalin, 'Conceptualising Climate-Induced Displacement' in Jane McAdam (ed), Climate Change and Displacement (Hart Publishing 2010) 81-103).
