Climate change policy-making is nothing without climate models. At global, national, and local levels, climate models of different kinds—from General Circulation Models (GCMs) of future climate change to Integrated Assessment Models (IAMs) that seek to balance social, environmental and financial costs and benefits—frame and are framed by policy concerns and requirements that shape the contours of what climate change is and what its implications will be.

In 2021, at the end of the Glasgow 26^th Conference of the Parties (COP26) to the UN Framework Convention on Climate Change (UNFCCC), the final text was published including the statement that the Committee of the Parties:

[…] recognizes that limiting global warming to 1.5 °C requires rapid, deep and sustained reductions in global greenhouse gas emissions, including reducing global carbon dioxide emissions by 45 per cent by 2030 relative to the 2010 level and to net zero around mid-century, as well as deep reductions in other greenhouse gases.

It was climate models that backed this claim. When, in the United Kingdom’s 2019 update to the 2008 Climate Change Act, the government committed the United Kingdom to net-zero carbon emissions by 2050, it was also climate models that were invoked as the evidential basis for this policy intervention. And when my own university, University College London (UCL), recently put in place a new travel policy that sought to reduce carbon emissions from flying, once again climate models were the basis that backed up this policy change.

I first became aware of the work done by climate models when researching climate change mitigation in the city of Manchester in the United Kingdom in the early 2010s. Concerned to respond to a national requirement to reduce carbon emissions in the city, Manchester City Council turned to the Tyndall Centre for Climate Research for help in making sure that their plans were “science-based.”

The first thing to become clear from this work was that the models themselves did not straightforwardly determine the response that policymakers developed. Models are probabilistic techniques that serve to create scenarios which provide options to policymakers as to how to best respond to the challenge in hand. In the case of the city of Manchester this required aligning the categorizations of problem areas that appeared in the climate mitigation models—“services, electricity, residential, industry, and transportation”—with organizational categories that the municipality already used to govern the city—“buildings, energy and transport” (Knox 2020). Climate modeling then was a form of description which, whilst being focused on constructing a scientific framework for understanding the likelihood of future events, was at the same time flexible enough to be massaged and molded so as to be inserted into already existing policy contexts.

A second point to highlight about climate models is that even those who work closely with them recognize that they are likely to be wrong - in that they are not simple representations of the world as it is. During my own discussions with climate scientists and policy makers in 2019, I was pointed in the direction of a paper that had caused something of a stir in the modeling community, entitled “Escape from Model Land” (Thompson and Smith 2019). In this paper, written by economic modelers, the authors warn against assuming that the models are faithful representations of the worlds that they describe and predict. Another climate scientist I met, who was intrigued by the cultural assumptions of modelers themselves, pointed me in the direction of early ethnographic research on energy modeling conducted in the early 1980s which also showed how wrong models often were, and explored just why it was that either modelers or policymakers became so enraptured by the magic of the models and their persuasiveness of their outputs (Keepin 1984).

Another issue with climate models, is that they are not just wrong, but they are also generative of fantastical thinking. Serious concerns exist among many climate scientists about the assumptions that are being built into the integrated assessment models which inform policy on climate change. One example of this has been the use of Bioenergy Carbon Capture and Storage (BECCS), which has been inserted into Integrated Assessment Models to fill a gap between government commitments to make concrete policy changes and the actual projected geophysical effects of those changes. Put another way, as climate scientists have signaled the alarm that radical and immediate reductions in greenhouse gas emissions are necessary, and policy-makers have responded with conservative and long-term emissions reductions plans and policies, BECCS has emerged as a technological silver bullet capable of filling the gap between political and physical realities. The problem is, however, that bioenergy carbon capture and storage does not actually yet exist—or at least not as a viable and existing technology operational at scale. Climate models create the grounds then, for technological fantasies, from which all kinds of investments, speculations, new enclosures and accumulations, and counter-political experiments are being seeded.

When climate models are invoked by policy-makers we must not then assume that this simply means that policy-makers are following a singular version of scientific truth. “The science” in the case of climate modeling is provisional, probabilistic, and based on scenarios from which policy-makers select those that best fit their understanding of their field of action and spheres of influence. Climate models have a power to seem like premonitions, but as at least one climate scientists warned me, they should not be read as predictions at all but as contingent materially grounded suggestions as to possible future worlds that may or may not appear depending on action in the present. My work has focused mostly on the United Kingdom and Europe, but one climate modeler I spoke to told me that US climate policy prior to the Trump era had a profoundly different relationship to climate models. Rather than calculating the costs of intervention, US policy-makers would calculate the cost of inaction, and then use this knowledge to set the levels of tax on carbon emissions—a policy which, if true, likely had very different effects to the ones I have described here. Whilst as anthropologists we are well used to tracing the effects of state based bureaucratic practice on local populations, there is still more to be done to understand the struggles that are taking place in relation to model land, about what models are in different times and places, how they are used, and the diverse futures they are bringing into being.