Blog post by Paula Küppers *


Climate change is a significant driver of displacement across Africa, with an estimated 7.4 million people internally displaced in sub-Saharan Africa due to climate-related disasters in 2022. While most climate-related displacement remains internal, estimates for cross-border displacement are “significant” and presents a pressing challenge for international law. However, the international refugee protection regime, with the 1951 Refugee Convention at its core, remains largely inapplicable to climate change displacement scenarios due to a lack of (fear of) persecution.

The principle of non-refoulement under international human rights law has often been pointed to as way to fill this legal lacuna. Broadly speaking, this principle prohibits states from deporting individuals if their human rights will be violated in the receiving country. In its 2019 Teitiota decision, the UN Human Rights Committee recognized that climate change effects in the receiving country can potentially trigger non-refoulement obligations.

The present article explores the potential of non-refoulement protection from climate change under the European Convention on Human Rights (ECHR). To examine this possibility, the article considers a hypothetical scenario: a Somali national facing deportation from the UK claims that her return would expose her to severe drought conditions, violating her Convention rights. This thought experiment will explore how existing ECtHR case law might support such a claim and the potential implications for future climate change litigation involving African nationals. To give a theoretical background, I will first briefly expand on how non-refoulement is dogmatically grounded within the ECHR, namely by reading the obligation into Article 3. The goal for a potential climate-related non-refoulement suit is therefore to present climate effects as cruel or inhuman treatment as defined by Article 3. The challenge here is the lack of human agency in climate change. I will therefore present caselaw in which the Court has found non-refoulement obligations to arise despite a lack of this agency element. After a summary of the treatment hierarchy developed by the Court, I will identify two ways in which our potential claim could latch onto the reasoning used in this caselaw. The article concludes with an evaluation of the viability of this non-refoulement approach for the actual protection need of people displaced by climate change.

The non-refoulement principle before the ECtHR

Starting with the first judicial recognition of the principle in its 1989 Soering decision, the Court has continually held that Article 3 of the Convention, which prohibits torture and ill-treatment, implicitly contains non-refoulement obligations for the signatory states. If a person experiences ill-treatment as a direct consequence of the contracting state’s removal decision, the ill-treatment therefore falls within the removing states’ responsibility – even though it happens outside the removing state’s jurisdiction, or even beyond the ECHR’s territorial jurisdiction.

In assessing whether a removal violates the Convention, the Court requires that the person in question, if deported, will face a real risk of being subjected to treatment contrary Article 3 (i.e., ill-treatment) in the receiving country. For any treatment to violate Article 3, it must attain a minimum level of severity. This severity threshold is relative and depends on all the circumstances of the individual case.

The focal question of a climate-related return case is whether the adverse environmental conditions in the receiving country can amount to ‘ill-treatment’. Going off the literal meaning of ‘treatment’ alone, this poses a challenge: typically, an intentional human act is required for something to be considered (ill-)treatment.

The Court has nevertheless found for removal decisions to violate Article 3 in some situations “where the source of the risk of proscribed treatment in the receiving country stems from factors which cannot engage either directly or indirectly the responsibility of the public authorities in the receiving country”. In its jurisprudence on ‘medical cases’ and returns to socio-economic deprivation, a ‘source of harm approach’ has emerged over time, in which the Court considers the extent to which the harm upon return is attributable to the receiving state when determining the appropriate severity threshold. While it did not yet have the opportunity to decide whether non-refoulement obligations can arise from the environmental conditions in a receiving country alone, the caselaw warrants a closer look as to its potentials for extension to such a context.

Medical cases

In D. v. United Kingdom (1997), the Court found that the UK would violate the Convention by returning a terminally ill applicant to St. Kitts, where the palliative treatment for his late-stage HIV-infection and AIDS was unavailable. In its reasoning, the Court acknowledged that while the non-refoulement principle had so far only been applied in contexts where the “proscribed treatment” upon return emanated from either intentional acts by public authorities in the receiving country or their inability to afford protection against intentionally inflicted acts by non-state bodies. That said, it established that the fundamental importance of Article 3 means the Court needs to be flexible enough to extend the non-refoulement principle to scenarios which do not fall into these categories.

In N. v. United Kingdom (2008), the Court held that returning an HIV-infected applicant to Uganda, where the lack of adequate treatment would shorten her life expectancy from several decades to two years, would not violate Article 3. It reasoned that, when the source of ill-treatment is a “naturally occurring” illness (or, more precisely, the receiving state’s inability to deal with it), Article 3 is violated “only in a very exceptional case, where the humanitarian grounds against the removal are compelling”.

Returns to socio-economic deprivation

In M.S.S. v. Belgium and Greece (2011), Belgium was found to have violated the Convention by returning the applicant to Greece, where he lived in a state of the most extreme poverty, unable to cater for his most basic needs for months – i.e., the harm flowed from the living conditions in the receiving country. The Court used a lower threshold than the ‘very exceptional’ one, saying that the living conditions could be attributed to the Greek authorities for failing to implement an EU-directive for asylum reception conditions.

In Sufi and Elmi v. United Kingdom (2011), the Court held that a removal of two applicants to Somalia, where they would face dire humanitarian conditions due to, inter alia, a drought and the on-going civil war, would violate Article 3. The Court distinguished this case from its medical caselaw because the ‘predominant cause’ of the harm (i.e., humanitarian conditions) in the receiving country was the warfare actions of conflict parties. It therefore used the lower threshold from M.S.S. In an obiter dictum, it stated: “If the dire humanitarian conditions in Somalia were solely or even predominantly attributable to poverty or to the State’s lack of resources to deal with a naturally occurring phenomenon, such as a drought, the test in N. v. the United Kingdom may well have been considered to be the appropriate one”.

In S.H.H. v. United Kingdom (2013), the Court distinguished M.S.S. and Sufi and Elmi from a case in which an applicant, disabled by a rocket launch in Afghanistan, faced dire socio-economic living conditions upon return to Afghanistan due to the country’s lacking welfare system. The Court applied the ‘very exceptional’ threshold, arguing that, while the applicant’s disability was not a ‘naturally occurring phenomenon’, Afghanistan’s poor welfare system was predominantly attributable to the state’s lack of resources, not intentional actions as in Sufi and Elmi. It found that the conditions did not exceed the threshold and therefore the UK’s removal decisions did not violate Article 3.

The threshold hierarchy (link)

To summarize: A very high (‘very exceptional’) threshold applies when the harm upon return flows from a ‘naturally occurring phenomenon’, like illness (as in D.) or a drought (as mentioned in Sufi and Elmi), or a state’s general lack of resources (as in S.H.H.). The threshold is slightly lower when the harm lies in the socio-economic conditions, but the predominant cause of these conditions are states’ or non-state actors’ direct actions, such as a war (as in Sufi and Elmi), or lacking implementation of laws providing adequate conditions (as in M.S.S.).

In general terms, this caselaw shows that non-refoulement obligations can arise even where the harm upon return cannot be directly linked to the realm of state responsibility of the receiving state. It seems that with the ‘source of harm’ approach, the Court adjusts the threshold of severity depending on how closely the harm in the receiving is related to the realm of intentionality or human agency. If the state is unable, instead of unwilling, to provide appropriate living conditions for its inhabitants – regardless of whether the cause of these living conditions is a ‘naturally occurring phenomenon’– the ‘very exceptional’ threshold is used.

Relating back to our initial thought experiment, if an applicant claimed that the sole source of the harm is the environmental conditions in the receiving country, the Court’s obiter dictum in Sufi and Elmi indicates that it would apply the ‘very exceptional’ threshold. In practice, a drought in Somalia therefore would only very rarely give rise to non-refoulement obligations. However, there are two ‘points of entry’ in the current caselaw with which our application could argue for an extension of the doctrine.

‘Points of Entry’ for evolution of the caselaw

There are several doctrinal shortcomings in the Court’s approach to non-refoulement, specifically in relation to socio-economic rights, which have rightly been criticized. Most prominently, it is questioned why it should matter whether the harm originates from a benign or malign source, if the actual harm to the individual is the same. However, if the Court choses to stick with its approach despite this need for doctrinal change, there are several points of entry in which an applicant could argue for a lower threshold.

1. Cumulative state responsibility for climate change

It has been proposed that the anthropogenicity of climate change could warrant the application of the ‘predominant cause’ threshold. By establishing a causal link between 1) a specific environmental hazard in the receiving country and climate change, and 2) climate change and deliberate omissions of the global community as a whole to reduce greenhouse gases, a climate displacement scenario could be framed closer to the realm of intentionality required for the application of the ‘predominant cause’ threshold.

Firstly, extreme weather event attribution is one of the most rapidly expanding areas of climate science, so it is increasingly likely to establish a link between a singular disaster and climate change. Secondly, the IPCC report confirms that climate change is a collective global failure to reduce GHG emissions, and in that sense the international community is responsible for the related harm.

The ECtHR’s newest climate change litigation could be used to support this line of argument. In Verein KlimaSeniorinnen (2024), the Court confirmed that the adverse effects and risks for individuals or groups in specific locations result from global aggregate GHG emissions, but that emissions from any single jurisdiction contribute only partially to the harm. In Duarte Agostinho (2024), it acknowledged that “there is a certain causal relationship between public and private activities based on a State’s territories that produce GHG emissions and the ad-verse impact on the rights and well-being of people residing outside its borders […] Climate change is a global phenomenon, and each State bears its share of responsibility for the global challenges generated by climate change and has a role to play in finding appropriate solutions.”

Relating to these new climate judgements could support a non-refoulement application in linking climate impacts in, for example, Somalia, to the shared responsibility of the international community.

2. The sending state as the source of harm

A climate-related return case arguably faces the unique scenario that the sending state itself may be identified as the ‘predominant cause’ of the harm in the receiving country, as many Convention states are among the biggest GHG polluters. Scott (2014) argues that this scenario does not fit into pre-existing caselaw categories, and that the Court should thus conceive of a sui generis approach for assessing an Article 3 violation. “It would be anomalistic to find that a host state carries a heavier non-refoulement obligation in cases where only another country is responsible for the harm feared, than where the host state itself contributes to that harm.” 

His arguments may be extended upon by factoring in that not only are European states’ GHG emissions a key driver of climate change, but historic and ongoing forms of (European) colonialism have increased the vulnerability of specific people and places to environmental hazards. For example, a prohibition of indigenous farming methods imposed by French colonizers throughout equatorial Africa resulted in desertification, which in turn can exacerbate climate hazards like floods and droughts and reduces access to food. An applicant might argue that the specific sending country should have higher non-refoulement obligations due to its bigger role in creating the harm in the receiving country.

For this line of argument, the harm would need to be linked back to the actions or negligence of the sending state. Several climate change litigation cases on a domestic and international level (apart from KlimaSeniorinnen e.g. cases before the Dutch, German, and Colombian constitutional courts) have established such a legal link. However, these cases concerned climate harms within the concerned states jurisdiction. The harm in refoulement cases is felt outside of the sending state’s jurisdiction, placing it further outside the realm of state control. Additionally, only the cumulative emissions of many states, not those of the sending state alone, could have triggered climate change and the consequent environmental harm in the receiving state. It will therefore generally be difficult to substantiate a direct causality between a specific climate-related event triggering the displacement of a specific person, and the specific GHG emissions of the country of refuge.

On the other hand, as mentioned above, the Court already employs a vague concept of state responsibility in non-refoulement contexts. The inconsistencies in the ‘source of harm’ jurisprudence could prove beneficial in arguing for a sui generis approach: The Court uses a criterion that is ultimately unrelated to the extent of infringement to determine the severity threshold. Why should the responsibility of the sending state in creating the harm upon return not also factor into the Court’s considerations? Would it, then, even be necessary to establish a direct responsibility?

Summary and evaluation

The ECtHR’s non-refoulement caselaw so far does not clearly cover climate-related harms in the receiving country. However, jurisprudence on medical cases and on returns to socio-economic destitution has shown that the Court is open to extending protection to scenarios where the harm upon return is connected to the living conditions in the receiving state. The ‘source of harm’ approach, albeit critique-worthy in itself, offers argumentative opportunities in relation to the anthropogenicity of climate change. Framing the (climate-related) harm upon return as a preventable failure of the global community to reduce GHG emissions could put it closer to the realm of intentionality that warrants the application of the lower ‘predominant cause’ threshold. Alternatively, a sui generis approach might incorporate the sending state’s particular responsibility in causing the harm.

Of course, practical viability of an ECtHR non-refoulement claim is limited for the majority of people displaced by climate change on the African continent. Claims before human rights bodies are often a last resort, and the ECtHR’s territorial jurisdiction requires an applicant’s presence in Europe. However, for lack of a normative international protection of climate change displacement, non-refoulement obligations are a subsidiary mechanism that is worth exploring. While the actual number of people seeking climate protection in Europe might be low, Strasbourg could play a pivotal role in further developing an international legal standard of climate-related non-refoulement first set out in Teitiota.

* Paula Küppers is a freelance journalist and student research assistant at the Chair of German and International Public Law and Comparative Law (Prof Dr Michaela Hailbronner) at the University of Münster, Germany.

This was produced as part of the ‘Refocus: An Online Summer Forced Migration in Africa Workshop Series’ which ran in April 2024. Look out for the special collection of blogs from the series shortly. To listen back to all four sessions, follow this link to our Spotify channel.


The views expressed in this article belong to the author/s and do not necessarily reflect those of the Refugee Law Initiative. We welcome comments and contributions to this blog – please comment below and see here for contribution guidelines.