Although climate change is often characterised as “global warming”, the impact of climate change will vary greatly from region to region. Regional aspects of climate change are controlled by atmospheric circulation patterns, which moreover exhibit considerable chaotic variability.
Model predictions of the atmospheric circulation response to climate change are in many cases highly uncertain, presumably because of systematic errors in the climate models (e.g. the location of the jet stream). The fact that these errors have stubbornly persisted despite increases in spatial resolution suggests that they are somehow linked to unresolved processes, whose effects need to be parameterised in the models. Thus, improving climate models requires a better understanding of multi-scale interactions.
This talk will present some examples of these kinds of uncertainties and some potential ways forward, which involve a more comprehensive use of observations and methods from data assimilation.