1.5 at the ICJ

The ICJ’s climate change Advisory Opinion of July 2025 capped a year in which global temperatures averaged 1.53°C above pre-industrial levels. In this context, the Court’s strong endorsement of the Paris Agreement’s more stringent 1.5°C limit is remarkable. The Court made three broad statements on 1.5. First, it recast the Paris objective of ‘pursuing efforts to limit the temperature increase to 1.5°C above preindustrial levels’ as a ‘primary temperature goal’ for parties (para. 224). Second, it specified that states must orient themselves towards that global goal in their NDCs (para. 242). Third the Court quoted, with apparent normative intent, a 2023 observation of the Paris parties: ‘limiting global warming to 1.5°C with no or limited overshoot requires … reductions in greenhouse gas emissions of 43 per cent by 2030 and 60 per cent by 2035 relative to the 2019 level’ (para 243).

All to the good—but the ruling leaves much unsaid: the 1.5°C target is, in practice, neither transparent nor straightforward. The Court made no attempt to enter into the complicated science of ‘1.5°C’ itself, how it is to be conceived, whether it is in fact achievable—and what it would mean if it were not. The reference to ‘1.5°C with no or limited overshoot’ is particularly striking—interpolating an immensely complex term of art from the world of climate science, but without explaining, or even signalling, the complexities. Strikingly too, the Court says essentially nothing about the second (and, on its face, stronger) element of the Paris objective, that temperature increases should be ‘held’ ‘well below 2°C’.

In this post, I focus on the 1.5°C Paris goal, aiming to provide background that is absent from the ruling but relevant to assessing the implications of this ‘primary’ obligation. I will look first at the language articulating the Court’s decision to prioritise 1.5°C, and then at some concerns its endorsement might raise. 

Choosing 1.5

By the crude measure of word count, the ICJ clearly chose 1.5 over 2: in contrast to 21 iterations of ‘1.5°C’, the Court mentions 2°C on five occasions—all but one of which appear in quoted text. That one occasion is the same moment, in para. 224, in which the Court gives 1.5°C pre-eminence, in a statement significant enough that I hope I may be forgiven for laying it out in full (emphases in bold added):

224. As a general matter, the Court notes that while the Paris Agreement provides for limiting the global average temperature increase to well below 2°C above pre-industrial levels as a goal and 1.5°C as an additional effort, 1.5°C has become the scientifically based consensus target under the Paris Agreement. At the twenty-sixth COP, which was the third CMA to the Paris Agreement, parties “[r]ecognize[d] that the impacts of climate change will be much lower at the temperature increase of 1.5°C compared with 2°C and resolve[d] to pursue efforts to limit the temperature increase to 1.5°C” (see decision 1/CMA.3, Glasgow Climate Pact, 13 November 2021, UN doc. FCCC/PA/CMA/2021/10/Add.1, p. 4, para. 21 (emphasis added)). Subsequently, at the fifth CMA, parties to the Paris Agreement reiterated this resolve and encouraged all parties to submit

“ambitious, economy-wide emission reduction targets, covering all greenhouse gases, sectors and categories and aligned with limiting global warming to 1.5°C, as informed by the latest science, in the light of different national circumstances” (decision 1/CMA.5, Outcome of the first global stocktake, 13 December 2023, UN doc. FCCC/PA/CMA/2023/16/Add.1, p. 7, para. 39).

In the Court’s view, these decisions express the agreement in substance between the parties regarding the interpretation of Articles 2 and 4 of the Paris Agreement, and thus constitute subsequent agreements in relation to the interpretation of the Paris Agreement within the meaning of Article 31, paragraph 3 (a), of the Vienna Convention on the Law of Treaties (see paragraph 184 above). Accordingly, the Court considers the 1.5°C threshold to be the parties’ agreed primary temperature goal for limiting the global average temperature increase under the Paris Agreement. The Court adds that this interpretation is consistent with Article 4, paragraph 1, of the Paris Agreement, which requires that mitigation measures be based on the “best available science” (see paragraph 74 above).

There are several remarkable elements to this statement: I will pick out three. First it is not at all clear on what basis the Court concluded that the 1.5°C goal ‘has become the scientifically based consensus target’. The sources adduced by the Court cite statements by the parties to the Paris Agreement from COP26 and COP28 (to be precise, in both instances, from the third and fifth meetings of the CMA—the ‘Conference of the Parties serving as the meeting of the Parties to the Paris Agreement’).[1] But it is difficult to see how either of these statements could properly be referred to as reflecting ‘scientifically based consensus’. Both are clearly political, not to say aspirational, statements. By contrast, the IPCC, whom the court affirms as providing the ‘best available science’, plainly regards the 1.5°C target as a longshot (see further below).

Second—following the ILC—the Court interpreted these statements of the CMA as ‘subsequent agreements’ within the meaning of the Vienna Convention on the Law of Treaties art 31(3)(a). The latter provides as a ‘general rule of treaty interpretation’ that ‘any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions’ should be ‘taken into account’ in interpreting a treaty (in this case the Paris Agreement). This reading is entirely plausible insofar as the format of the CMA involves the parties to the Paris Agreement deciding on its implementation iteratively over time. Still, the Court’s reading potentially increases enormously the relevant text to be taken into account—as a matter of law—in interpreting the Paris, Kyoto, and UNFCCC treaties. The climate treaties are unusual, though not unique, in the structured format within which ‘subsequent agreements’ are concluded, on an essentially annual basis—at time of writing, the Decisions of the three bodies run cumulatively to over 900.

At the same time, the yearly revisitation of progress is, and has always been, an intensely political and dynamic process, which does not easily lend itself to consistency. To be sure, the Court narrows the potentially vast scope of these materials, clarifying that they ‘may constitute subsequent agreements … in so far as such decisions express agreement in substance’ regarding interpretation (para. 184, my emphasis). Nevertheless, as both climate science and climate impacts have progressed over time, the language in COP/CMA decisions has likewise evolved—even ‘agreements in substance’ are subject to revision (avoiding this is arguably the point of the looser ‘framework’ language of the treaties themselves). Several developments mark out the particular timing of the ICJ’s 2025 intervention. For one, the fifth CMA (CMA.5) of late 2023 marked the first ‘global stocktake’—a periodic quasi- accountability mechanism built into the Paris Agreement, providing a useful handle for the Court. Also, between the two cited CMA Decisions, the IPCC published volume 3 of its Sixth Assessment Report (AR6 WGIII)—this being ‘the latest science’ referred to in para 224. Relatedly, CMA.5 introduced a new term of art: ‘1.5°C with no or little overshoot’—to which I will turn in a moment—but it also ‘acknowledged’ with ‘concern’ that the carbon budget for 1.5°C was already almost depleted (para. 25). From this perspective, it is striking that CMA.6 is, across its 23 Decisions of late 2024, essentially silent on 1.5°C.

A third notable element in para. 224 of the Advisory Opinion is the reference to science. Here the Court harks back to para. 74, where it asserts that ‘it will rely [in its assessments] primarily on the IPCC reports, which participants agree constitute the best available science on the causes, nature and consequences of climate change’. It is, of course, unsurprising that the Court formally ratifies the IPCC’s pre-eminence in this way—in doing so it follows a lengthening list of courts, both national and international, dealing with climate science. The effect, however—since this is intended as a judicial statement on the interpretation of a multilateral treaty—is to embed the IPCC into the Paris Agreement in a manner that has heretofore been implicit. The clarity of naming the IPCC as repository of the ‘best available science’—a term of art in the treaty—brings that body directly into the Paris interpretative architecture in a dynamic manner, arguably bypassing state discretion.

The Court did not enter into any assessment of the IPCC’s competence to produce the ‘best available science’, nor did it examine any of the (by now very many) critical perspectives on the IPCC’s knowledge-producing work. These have been particularly salient with regard to the 1.5°C target itself, a point to which I turn next. In general, though, the IPCC is neither prescriptive nor predictive; it speaks in probabilities and spectra, providing a range of options in terms of their likely consequences. The IPCC is, in effect, multivocal, complicating efforts to glean a clear normative steer from its vast opus.

But what—and when—is 1.5?

The Court avoids the somewhat arcane history of the 1.5°C target—itself a contested IPCC artefact (a story well told in Andreas Malm’s Overshoot). The appeal of 1.5°C is self-evident—but a close reading of IPCC reports shows that ‘1.5°C’ and ‘1.5°C with no or limited overshoot’ are not coextensive terms and that both come with significant baggage. The latter notion was introduced in the IPCC’s 2018 Special Report on Global Warming of 1.5°C and played a key role in AR6 WG III. The term refers to the possibility, under a spectrum of future emissions pathways produced by what are called, in IPCC jargon, IAMs—‘Integrated Assessment Models’—that global warming will peak at 1.5°C (or, effectively, 1.6°C).[2] This is contrasted with pathways in which warming is ‘returned’ to 1.5°C ‘after a high overshoot’ (eg AR6 WG III, 330 (table 3.2)).[3] ‘Overshoot’ refers to a scenario in which temperatures go above the target but are then drawn back down to meet it, ideally within ‘several decades’. This is to be done through ‘negative emission technologies’ (NETs)—that is, burying CO₂ in the ground or oceans.  

The Court is hardly responsible for the timing of its Opinion nor for the convoluted scientific and policy underpinnings of 1.5°C—but both are likely to be key elements in any further interpretation of this finding. I will take each in turn.

First, on timing. Five years may not seem like much in the broader trajectory of international law and the progress of climate change—but in terms of climate goals, the years 2020-2025 have been pivotal. Its previous exhortations having failed, the IPCC is categorical in AR6 that, in order to achieve ‘1.5°C with no or limited overshoot’, global greenhouse gas emissions would need to peak ‘between 2020 and at the latest before 2025’ (AR6 WG III, 17, emphasis added). Even then, the chances of succeeding were, at best, >50% (rather than say 67% or 90%). By CMA.5 in late 2023, with emissions still not having peaked, ‘1.5°C with no or limited overshoot’ looked extremely unlikely. By the time of the Court’s ruling in July 2025 the window for ‘no or limited overshoot’ had essentially closed. The timing is excruciating. The UNFCCC has been in force since 1994. There has been ample—if steadily attenuating—time to ensure global temperatures would peak at or below 1.5°C above preindustrial temperatures. The Court is not to blame that its ruling appears at the very moment this goal slipped out of reach.[4] But it is also true that the closure of this window was long foreseeable.

This leads to my second point: the term ‘1.5°C’ now incorporates ‘after a high overshoot’—which is to say 1.5°C does not signify ‘1.5°C’ at all. Rather it means (optimistically) ‘up to 1.8°C’ for ‘up to several decades’ (AR6 WG III, 21 (Box SPM.1)). A few remarks. First, even at a 1.8°C peak (itself unlikely), the chances of ‘returning’ to 1.5°C are not good. The IPCC puts them at >50% at best—and that is with ‘immediate action’ in 2022 and global GHG emissions falling 23% by 2030 (AR6 WG III, 17). Global GHG emissions rose by 500 million tonnes in the year following the report. Second, needless to say, global temperatures have never actually been reduced through deliberate human intervention: ‘overshoot’ is entirely speculative, relying on unavailable-as-yet NETs—that is technologies that bury CO₂ permanently and at scale: CMA.5 explicitly calls for ‘accelerating … removal technologies such as carbon capture and utilization and storage’ (para. 28(e)). To be clear as to the meaning of ‘at scale’, with current annual emissions of c.38 Gt CO₂ (according to Our World in Data), at least half that amount would need to be buried each year to stabilise current temperatures (half, because the oceans currently absorb about half of annual CO₂ emissions). For a >50% chance of returning to 1.5°C, the IPCC reckons on CCS of 360 Gt CO₂ (385). But current CCS removal capacity is c. 0.05 Gt CO₂ per year, scheduled to double to 0.1 Gt CO₂ once ‘facilities [now] under construction begin operating’. By contrast, 1.5°C ‘with no or limited overshoot’ would have relied on less NET (23)—which comes on line too late to avoid the peak—and therefore much steeper GHG reductions.

One ironic effect of the disappearance of ‘1.5°C with no or limited overshoot’ is that many of the advantages of avoiding a 2°C rise are not in fact avoided: temperatures increase, severe climate impacts are experienced, some effects are irreversible: there is the danger of severe tipping points. A second effect is a general blurring of the parameters of future policy. How high can temperatures go before ‘returning’? If we are relying on CCS in any case, might this delay the phasing out of fossil fuels? Where will CCS be located (estimates reckon on land requirements equivalent to the territory of India and possibly twice as much)? How would domestic NDCs reflect these global-level calculations? The policy world of ‘overshoot’ is, unfortunately, essentially conjectural, and the IPCC’s IAMs—produced by large computers at a small coterie of institutions—are, as the IPCC itself notes, non-realistic: the assumptions for overshoot in particular result from abstracted modelling of industrial processes on a vast scale, with little reference to the political, social, practical or economic circumstances of their supposed implementation.

Another irony is that the alternative limit of 2°C—as viewed from within the IPCC’s IAMs, may still be within reach without overshoot, at a 50-67% chance—and though emissions would still need to peak before 2030 and fall steeply, this pathway requires even fewer NET removals (AR6 WG III, 23-4). To aim at 2°C rather than 1.5°C as a ‘primary temperature target’, then, might yet permit refusal of the dizzying politics of overshoot and permit clarity and definition on the available modes of action, the likely arc(s) of emissions, and the distributed responsibilities of states. Most IPCC 2°C pathways, however, do involve overshoot, and these too depend heavily on NETs, according to a long line of IAMs. Indeed, since meeting either Paris goal ‘with overshoot’ involves deep reliance on NETs, the concept of a ‘final’ temperature goal can appear somewhat arbitrary. Unsurprisingly the ICJ showed no inclination to enter this unrewarding terrain.

Nevertheless, it is curious that we now have a ‘primary temperature target’ ensconced in law that is quite certain to be transgressed in the immediate future. The policy tools that might ‘return’ us to this target, having transgressed it, are speculative, and the mechanisms for allocating responsibility for development and deployment of these—now indispensable—technologies, for ensuring ‘return’ to the target, and for the profound consequences of lengthy, or even permanent, ‘overshoot’ likewise remain unclear.

Was the Court wrong, then, to plump for 1.5°C? Hardly. The case for high ambition is hard to argue against. But a danger lies in deferring entirely to the IPCC on how to get there. Among the criticisms of IAMs are that (i) they extrapolate from past to future—they are incapable of imagining sharp historical discontinuities (of the kind 1.5°C seemingly requires); (ii) they operate under orthodox neoclassical economic assumptions, including significant future discounting; and (iii) they remain—for such a critical element of international climate policy—largely opaque and quite arcane. Scrutiny of the IPCC’s IAMs—the conditions of their production, their assumptions and orientation—is vital if they are to sit, as now seems, at the apex of climate law.

[1] The Court also uses the term CMA to refer to the relevant body in respect of the Kyoto Protocol (para. 184), whose formal acronym is in fact CMP—the arcane nature of climate initials tripping up even the ICJ. 

[2] ‘Limited overshoot refers to exceeding 1.5°C global warming by up to about 0.1°C and for up to several decades’. IPCC AR6 WG III, 21 (Box SPM.1).

[3] ‘High overshoot refers to temporarily exceeding 1.5°C global warming by 0.1°C–0.3°C for up to several decades’ IPCC AR6 WG III, 21 (Box SPM.1).

[4] For the IPCC, the current ‘overshoot’ of 1.53°C does not in itself indicate that the 1.5°C limit has been breached, since a determination of the ‘average global’ temperature at any given moment is based on an assessment across three decades. Rather, for the IPCC, future breach is no longer avoidable.