Stabilising Society: Why Earth Systems Stabilisation Needs a Human-Centred Sibling

Image: Sock_AnggalihPs / Shutterstock

Last month I joined more than 250 people from six continents at the Harpa Centre in Reykjavík for the first Global Risk & Stabilization Summit. In his reflection on the week, “After Reykjavík“, summit convenor Joshua Elliott made a clear case for why a growing community now argues we must prepare options to intervene directly in the Earth system, stabilizing the Arctic, the ice sheets, the biosphere, Earth’s reflectivity, because mitigation and adaptation alone may no longer keep catastrophic risk to an acceptable level.

I want to be clear about where I stand: I am not writing as an advocate for or against Earth systems interventions. My scepticism is directed elsewhere: at a discourse that risks building a sophisticated technical and governance architecture for the Earth system while treating human and socioeconomic dimensions as an afterthought. Whatever is eventually decided about stabilisation, that decision cannot be void of those considerations. At Reykjavík, I left with the sense that it currently is. Elliott himself notes that “underneath all of it are the human systems that absorb the shocks.” That line deserves to be more than a closing clause. It is the thread I want to pull on here.

Socioeconomic catastrophe does not wait for Earth-system thresholds

The stabilisation discourse in Reykjavík was organised almost entirely around biophysical tipping points—AMOC collapse (Atlantic Meridional Overturning Circulation), ice sheet loss, degrees of warming—framed as triggers for catastrophic downstream harm to societies. Catastrophic socioeconomic risk does not wait for those triggers to be pulled. It has its own thresholds, often lower and faster-arriving than any physical one.

In 2015, adelphi led the G7-commissioned report A New Climate for Peace, which identified seven compound climate-fragility risks: pathways through which climate stress interacts with weak governance, inequality, and resource competition to produce state fragility and conflict. None of these seven risks requires an Earth-system tipping point. Shoring Up Stability, an assessment I led of the Lake Chad basin, documented one of the world’s largest humanitarian crises, driven by chronic poverty, weak governance, and climate stress compounding one another—at roughly 1.3 degrees of warming. If we are looking for catastrophic socioeconomic risk at low levels of Earth-system change, we do not need a model. We have a live case study.

Part of why this risk sits outside the stabilisation conversation is that it is harder to instrument—there is no equivalent of an ice-mass balance measurement for social cohesion or governance legitimacy. But it is not intractable, and a sibling stabilisation field, with its own decision-support infrastructure built around governance and livelihood indicators, is both necessary and achievable.

The architecture has to differ, not just scale up

The biophysical stabilisation community can offer real discipline on early warning and pre-agreed triggers. But there is no equivalent of a physical intervention lever for socioeconomic risk. The intervention is pre-emptive investment in governance capacity and social contracts: slower, more political, and far harder to attribute causally. The politics is the intervention.

This matters because any interventions proposed in the biophysical space are not neutral with respect to this human layer—whatever their technical merits, their governance and deployment can worsen existing injustice if designed without the people who would carry the risk. Solar radiation management illustrates this best. I am not taking a position on whether Solar Radiation Modification (SRM) research or deployment should proceed here; what concerns me is that the governance question it raises is, at root, about control rather than climate science alone. Only around ten states currently have the aerospace capability to deliver stratospheric aerosol injection at scale, meaning a handful of actors would effectively hold the global thermostat regardless of the eventual scientific consensus. Deployment would likely require coordinated dual-hemisphere injection, an extraordinary trust requirement among a small club of capable states. SRM has also increasingly been framed as a security issue, including by the United States, which should prompt the question of what that framing protects: genuine concern about destabilisation, or an interest in controlling who gets to use the capability at all.

The African Group of Negotiators has translated these concerns into concrete, repeated positions: prevention through an international non-use agreement rather than governance that only activates once deployment is imminent; ensuring Africa is not used as a testing ground; a moratorium against irreversible harm, drawing on the Convention on Biological Diversity’s precedent; avoiding an SRM arms race; and keeping a firm line between adaptation, mitigation, and SRM, so SRM is never classed as mitigation or eligible for climate finance. This position is formal and sustained: from the African Ministerial Conference on the Environment (AMCEN)’s 2023 call for a non-use governance mechanism, through the African Group’s leadership—led by Djibouti—in blocking a scientific-expert-panel resolution at UNEA-6 in 2024, to the Tripoli Declaration’s 2025 reaffirmation of an “unequivocal rejection” of solar geoengineering on environmental, ethical, and geopolitical grounds. That sustained mistrust is not a communications problem to be managed; it is evidence that governance built without addressing who controls deployment, and who bears the risk, will struggle to command legitimacy.

A feedback loop, not a competing priority

Socioeconomic risk is not a lesser or separate category to the biophysical risks Reykjavík was built around; the two are in a feedback loop. Conflict driven by climate-fragility produces its own, largely unaccounted-for emissions: the global military sector’s full lifecycle footprint is estimated at roughly 2.75 billion tonnes of CO2 equivalent a year, around 5.5 percent of global emissions—if the world’s militaries were a country, they would be the fourth-largest emitter on Earth, ahead of Russia, though this figure comes from a research consortium rather than an official inventory and should be treated only as indicative. These emissions sit outside the accounting frameworks underpinning mitigation pledges, and conflict itself strips away the governance capacity mitigation and adaptation depend on, as we have seen directly in Lake Chad and Somalia. This is not a risk category to be ranked against AMOC collapse or ice sheet loss. It is a mechanism that actively erodes our collective capacity to prevent those outcomes in the first place.

Human systems cannot be a closing clause

I am not in a position to say whether Earth systems stabilisation should or should not proceed. What I am able to say is that whatever is decided, it cannot be decided, designed, or governed without the human and socioeconomic layer as a co-equal consideration from the outset, not a line added once the physical science and engineering questions are settled.

Lake Chad, the sustained mistrust over SRM governance, and the uncounted emissions of climate-fragility conflict are not arguments for or against intervention. They are evidence that the human systems absorbing climate shocks already have their own catastrophic thresholds, their own governance gaps, and their own legitimate claim to a seat at the table. The sibling category to Earth systems stabilisation is not hypothetical. It is underway, under-resourced, and inseparable from the risks this field already treats as catastrophic. Whichever way the stabilisation debate goes, the two communities should be learning from each other now, not building in parallel.

Related articles:

Fragile Progress: How COP31 Can Advance Climate Security (3-minute read)

Climate Risk Is Forcing Multilateralism to Evolve (3-minute read)

When Metrics Miss the Point: Why Climate Adaptation Metrics Must Reflect Plural Worldviews (3-minute read)

The Author