Tuning climate models with Palaeoclimate Data

Abstract:

It has been common to perform comparisons between palaeoclimate model simulations with palaeoclimate reconstructions to better understand the mechanisms of climate change, and the accuracy of data reconstructions. However, it is surprisingly rare that palaeoclimate data has actually resulted in a change of the models. This is particularly true for the state-of-the-art GCMs. The talk will focus on two examples of using model parameter perturbation methodogies to effectivily tune the model to palaeoclimate data. The main example will use the Last Glacial Maximum and focus on the stability of the Atlantic Meridional Overturning Circulation. We will also discuss older time periods such as the earlly Eocene. In both cases we will use versions of the Hadley Centre climate model. It will be shown that the model initially appears to be too stable to forcings, particularly input of fresh water. By tuning the model, we can get enhanced sensitivity of the model. The results show that this can give significant differences in the response of the model at regional scale, although the overall climate sensitivity of the model to future clinate change is relatively unaltered.