In the era of overshoot,are climate risk assessments fit for purpose?

Abstract

Climate policy needs to respond to the realities of overshooting 1.5°C of warming. Geoengineering approaches – both large-scale carbon dioxide removal and solar radiation management – have been proposed as ways to both limit global temperature increase and return the climate to a stable and safe state. These provide policy makers with new options, but come with notable biophysical and socioeconomic risks. Current methodologies for assessing climate risk have potentially significant weaknesses when evaluating the risks generated by geoengineering. Given the potential for social and biophysical systems to produce complex and destabilising feedback loops, careful assessments of geoengineering must be conducted within holistic and integrative risk frameworks.

We are heading into climate overshoot. The twelve months of 2024 were both the warmest ever directly recorded by humans and also produced the greatest annual emissions of carbon dioxide (Friedlingstein et al., 2024). The available carbon budget for a 50% chance of limiting warming to 1.5°C will be exhausted within the next few years. This will come as no surprise to those who have been engaged with climate policy. The 2015 Paris Agreement was intended to limit warming to well below 2°C via a voluntary mechanism of increasing mitigation ambition. While much progress has been made in some nations, the global phase out of fossil fuels has failed to materialise. At the same time, our assessments of the risks that climate change presents to societies have changed. The 1.5°C threshold was once interpreted as being the amount of warming that would represent an existential risk to low-lying island nation states, which would be submerged under rising sea levels if warming passes 1.5°C since pre-industrial periods. The rapidly developing science of tipping points now indicates that warming in excess of 1.5°C carries increasing risks of large and effectively irreversible changes to the Earth system such as ice sheet and ocean current collapse. This would have global consequences that would also negatively impact rich, industrialised nations (Lenton et al., 2025). Arguably, the architects of the Paris Agreement were very much alive to the chances of exceeding 1.5°C, because they framed the treaty around end-of-century temperature targets that would allow for a period of overshoot. This has been progressed via the framing of climate policy in terms of net zero and the reliance on carbon dioxide removal (CDR).

The first use of CDR in climate policies and scenarios is to deal with residual emissions from so-called ‘hard-to-abate’ sectors, such as aviation, because there are currently no credible plans to rapidly decarbonise them. However, even with large-scale carbon removal, we would still be on course to exceed 1.5°C of long-term warming. Therefore, CDR is required to go beyond net zero to net negative in which the total amount of carbon dioxide removed each year is greater than that emitted. This would reduce cumulative emissions and so drag temperatures back down.

How could this be achieved? For some time, BECCS – Bioenergy with Carbon Capture and Storage – was proposed as the CDR technique of choice. This would involve the large-scale planting of rapidly growing trees that would be harvested and burnt in thermal power stations to generate electricity, with the carbon dioxide being captured from the chimney stacks. This would then be pumped into geological storage sites such as depleted oil and gas fields. This process would both generate electricity and reduce atmospheric concentrations of carbon dioxide because as they grew, the trees would draw down carbon dioxide from the atmosphere.

Unfortunately, it has since been established that BECCS would generate significant risks to food, water and energy security, plus biodiversity, given the vast amounts of land and water it would consume if it were deployed at gigaton scale (Creutzig et al., 2021). This research makes clear that we must carefully evaluate all the risks of potential climate solutions and policies so that we can make the most informed decisions on how to respond to what is rapidly becoming an emergency.

“A climate risk management typology: Integrating approaches to reduce risk”, hereafter referred to as Hyslop et al., is focused on this issue. The authors evaluate climate risks by expanding the ‘napkin’ diagram (Long and Shepherd, 2014) which was a sketch intended to show how different climate actions could affect total radiative forcing and therefore temperature increases, as shown in Figure 1.

Figure 1.

The ‘napkin diagram’ based on Long and Shepherd (2014). CDR – carbon dioxide removal; SRM – solar radiation management. Reproduced from Hyslop et al. (2025).

It is here that Hyslop et al.'s approach offers important new opportunities that could allow policy makers and wider society to begin the challenging process of critically evaluating both the risks of climate change and our responses to it. The IPCC has made great strides in assessing and communicating the first aspect - the risks that climate change presents to human societies. This is a challenging task. The situation becomes much more challenging when evaluating the second aspect - the risks that our responses to climate change might produce. Within current policy discourse, these are often bundled into a basket of ‘transition risks’. For example, the phase out of fossil fuels could generate risks of stranded assets, large-scale unemployment and social deprivation, along with geopolitical risks that emerge from changes to the access and exploitation of energy sources. But given the very wide range of responses and actions, such an approach may be too limited. Hyslop et al. convincingly argue that there must be at least an attempt at producing risk response functions for the different approaches. That begins with an understanding and agreement of which categories of risk are deemed important, especially within the context of breaching 1.5°C. The prospect of entering a period of overshoot is leading to global calls for stronger and more urgent responses. What those responses should be are a matter of some debate, encompassing a vast range of competing ideas, theories, beliefs and even value systems.

The dominant narrative, visually represented in the napkin diagram, is that rates of mitigation are insufficient to limit warming to well below 2°C. Consequently, we will need to undertake some sort of deliberate intervention in the Earth system in order to limit global temperature increases. That is to say, undertake geoengineering. The stroke of the pen indicating large-scale CDR is intended to reduce the positive radiative forcing humans have produced by reducing atmospheric concentrations of carbon dioxide. Beyond BECCS, it now encompasses a wide range of proposed approaches such as soil carbon sequestration, accelerating the natural weathering of rocks, and enhancing the alkalinity of the ocean. A great deal of funding, and hype, has also been directed towards engineered removal techniques such as Direct Air Capture (DAC). But the ability to permanently store any carbon captured is very limited. Currently, of the 2 billion tons of carbon dioxide currently removed from the atmosphere each year, over 99% comes from reforestation and afforestation (Smith et al., 2024). The carbon stored in biomass is vulnerable to climate disruptions such as high heat and drought that can kill trees and so release the carbon back into the air, or forest fires which do the same but over much faster timescales.

The SRM – Solar Radiation Management - curve on the diagram similarly represents a spectrum of methods. Their aim is to either reduce the total amount of the Sun's energy that reaches the surface of the Earth, or increase the escape of long-wave terrestrial radiation into space. Whilst some climate models show that the various approaches could have a large and near-term global cooling effect, the risks of deployment include changing the timing and duration of major regional weather systems such as the South Asian monsoon (Royal Society, 2025). An additional risk is termination shock. This emerges from the successful deployment of SRM, because if it were stopped then this could produce a very rapid rise in global temperatures if progress had not been made on mitigation and/or CDR.

Hyslop et al. are motivated to try to better assess how a particular climate intervention or approach generates risks. They do this by considering possible risk functions. For example, would increasing SRM deployment produce a monotonic increase in risk? Do risks saturate? In addition to answering these questions there must be a wider discussion about non-linear dynamics with respect to climate change risk. This is because our current understanding of the system of feedback loops between climate impacts and our responses to them is very limited. Perhaps dangerously so.

While the immediate or ‘first order’ impacts of significant climate change can be very serious – e.g., a lethal heatwave – it is the knock-on effects or ‘higher order’ impacts that can produce extremely serious risks. Consider the simultaneous failure of bread basket harvests. Food prices spike, governments put in place export controls that increase prices even further, food security for people around the world collapses as they cannot access food, social instability increases, governments respond with violent suppression, tensions increase at borders. This would not facilitate collective efforts to rapidly reduce greenhouse gas emissions. This sequence of events is an example of where the consequences of climate change could begin to impact our ability to deal with the drivers of climate change. Such a feedback loop has the potential to entirely derail efforts to progress sustainability transitions.

This class of risk has been dubbed ‘derailment risk’ (Laybourn et al., 2023). Derailment risks can stem from both the impacts of climate change and the consequences of climate action itself. Climate impacts—such as worsening disasters in low-income countries—can divert resources away from proactive adaptation and decarbonisation efforts, deepening vulnerability and leading to ever-costlier disasters. Likewise, the side effects of climate policies can be exploited or misrepresented to fuel political backlash, slowing progress on decarbonisation. This delay can in turn necessitate faster, more disruptive measures later on, which may again be politicised to further obstruct the required changes to socioeconomic and political systems for an equitable transition away from fossil fuels. Would a rush towards SRM and the governance challenges it would pose increase or decrease derailment risks?

Circling back to a consideration of which risks should be included in these assessments presents an opportunity to consider a key risk of geoengineering – that of mitigation deterrence. If policy makers believe that some of the dangerous impacts of climate change can be avoided in the future by deploying SRM and/or gigaton-scale CDR, then they may be less likely to implement mitigation policies today. This would be an example of a ‘burn now – pay later’ approach in which the continued use of fossil fuels is promoted along with promises of future technological solutions (Dyke et al., 2021). The potential for this sort of moral hazard represents a risk generated by the mere suggestion of geoengineering. A related risk is that geoengineering may divert attention and resources away from alternative responses to the challenge of climate change. The 12 leverage points framework developed by Donella Meadows can help provide insights here (Meadows, 2008). Figure 2 graphically represents the various leverage points and how interventions further to the right produce proportionately large effects on the system in question.

Figure 2.

Visual representation of the twelve leverage points proposed by Donella Meadows. The large circle represents the system of interest. Various intervention points in the system are distributed from left to right.

Intervening to reduce anthropogenic climate change can be most easily understood with leverage point 3 - stocks and flows (of carbon). However, with regard to efforts to precipitate the rapid phase out of fossil fuels, leverage points 7 and 8 - information flow structures, and systems rules - are key. Given that 80% of global energy supply is still satisfied by fossil fuels, the rapid phase out of coal, oil, and gas would require significant work at this level if climate policies are to be equitable. For example, the UNFCCC explicitly recognises that there are common but differentiated responsibilities with regard to emissions reductions and adaptation. Less industrialised nations that have very little responsibility for the climate crisis are often most exposed to climate change impacts, whilst having the least ability to undertake rapid low-carbon development and adaptation. Richer nations should be driving a flow of information and finance to help rapidly increase capacity. This may well demand a change in system rules with regard to debt because in the absence of debt forgiveness, many nations will not have sufficient fiscal space to finance the transition. There are also deeper considerations of justice in terms of the total flow of money and resources between the Global North and Global South. Unequal exchange theory concludes that economic growth in the Global North relies on the significant appropriation of resources and labour from the South, via price differences in international trade (Hickel et al., 2022). Addressing that requires considerations of system structure, and even leverage point 10, the goal of our globalised industrialised civilisation.

Unfortunately, frameworks such as the napkin diagram may be directing focus towards the least effective places to intervene in the climate system – constants and parameters, the size of buffers, and material stocks and flows. For example, SRM promises to reduce climate risks by modulating the flow of energy into and out of the climate system, while CDR is an approach specifically intended to reduce the stock of atmospheric carbon dioxide. In some respects these approaches are treating the symptoms and not the root cause. Moreover, they may be effectively excluding consideration of deeper and more powerful interventions. It is in that context that risk-risk assessments offer a narrow, binary worldview where geoengineering is offered as the only alternative to climate breakdown, justifying interventions which were previously viewed as unthinkable (McLaren, 2025). The risks of the various geoengineering approaches, both individually and collectively, are wide ranging and little understood (Lazard et al., 2025). It is therefore essential to explicitly consider the whole spectrum of possible climate mitigation pathways and the risks they might produce. This includes how biophysical and social systems may respond to not only a changing climate but also our attempts to further interfere with it. Hyslop et al.'s approach may facilitate this if it can help us dig deeper into the systemic drivers of climate change.