Most of the forecasts of global warming and climate change you read about are produced through massive 'General Circulation Simulation Models' (GCMs). But there are alternative statistical methods of doing this which are far easier and more flexible, yet answer some of the same questions. Emeritus Professor Peter Young, an associate of the Centre for Marketing Analytics and Forecasting for many years, has shown, in a recent paper, how such 'data-based mechanistic models', in addition to producing much better forecasts than the climate models, can give insight into the forces that lead to warming. For example, one of the model parameters is a multiplier, referred to by climate scientists as the ’equilibrium climate sensitivity’ (ECS), that shows the impact CO2 emissions have on temperature. It is interesting to note that, in another paper published more recently in the journal Nature by climate scientists Cox et al. (2018), a new estimate of the ECS is produced that is a very similar, but significantly less accurate, than Young’s estimate.
Is this another case where simpler models do better - or at least should be used in combination with the established physical models?Back to News