Can We Trust Climate Models?

12/21/20221 min read

All the global climate models considered in the Climate Model Inter-comparison Project are consistent in projecting, amongst others, a global and annual mean temperature rise of around 2 to 4°C over the 21st century (Hargreaves and Annan 2014: 435). This conclusion is credible not only because of the model consensus, but also because it is also supported by comparisons with proxy data on broad scales (Hargreaves and Annan 2014: 437). Present-day climate models simulate various attributes of behaviour of the climate system, with enhanced performance year on year testifying to their reliability (Hargreaves and Annan 2014: 437). Long-term projections of climate change are also, in principle, falsifiable (Hargreaves and Annan 2014: 437).

The primary limitations of modern climate models are, first, their weakness in producing reliable projections on finer spatial and temporal scales; and second, the insufficient falsifiability of the knowledge produced by those models.

Today’s climate models do not appear to generate dependable predictions on annual, multiannual, and decadal timescales (Hargreaves and Annan 2014: 437). In spatial terms, there is a lack of model agreement on the extent and pattern of regional climate phenomena between models and the data, and within ensemble members (Hargreaves and Annan 2014: 437).

According to Karl Popper’s theory of falsifiability, a hypothesis is falsifiable – and therefore scientifically valid - if it permits one to arrive at extrapolations which, when applied to evidence, could be shown to be false (Vignero and Wenmackers 2021: 11729). Climate model projections of climate changes for the next 100 years can only be falsified after observations of the actual changes during said period; however, this long wait is unfeasible in practice (Hargreaves and Annan 2014: 437). Alternative or competing models, which adopt different parameter values, possess too much common theory and code such as to provide inadequate contrast (Hargreaves and Annan 2014: 437).

References:

  • Hargreaves, J.C. and J.D. Annan. (2014). ‘Can we trust climate models?’ Wiley Interdisciplinary Reviews: Climate Change 5(4) 2014, pp.435–440.

  • Vignero, L. and S. Wenmackers. (2021). ‘Degrees of riskiness, falsifiability, and truthlikeness’. Synthese 199, 11729–11764 (2021). https://doi.org/10.1007/s11229-021-03310-5