'Climate change models: what are they evidence for?
Sponsored by ESRC funded Centre for Climate Change Economics and Policy|
Investigators: N. Cartwright; Roman Frigg
Short description:
State-of-the-art, physics-based computer models in climate science are currently unable to represent some processes at all (hurricanes are just too small) and they make first order errors in the representation of other processes which are well observed (the timing of tropical rainfall, the gradient of sea surface temperatures across the tropical Atlantic, the frequency of blocking phenomena that impact European weather. They are the most complicated members of a hierarchy of science-based models, and they support the general conclusions suggested by simpler "back of the envelope" models. As only a naïve realist view would argue that these models are picture-perfect mirrors of the real world, how could then their outputs be used as 'evidence' for policy decision?
The development of a coherent approach for interpreting state-of-the-art climate models as evidence for decision support would be a major contribution to philosophy, physical science and decision-theory with immediate practical applications.