The Interdisciplinary Science Research Lectures

Can models simulate climates of the past?

Due to anthropogenic emissions of greenhouse gasses global temperatures will rise during this century by an amount not observed for millions of years. Numerical climate models are the only tool we have to predict the outcomes of humankind’s grand climate experiment. Models however are by definition imperfect tools. How do we know that their predictions for the future are as good as they need to be? Model predictions for current climate can be compared to observations that provide strong evidence of their performance. However, how can we be sure that model can successfully predict climates that are radically different than the present and perhaps towards what we are expect by the end of this century?  For much of Earth's climate history our planet has been warmer than present, and the geological record provides a rich and diverse archive of information on past climate states. If this information can be gathered together and synthesised for a particular interval or intervals of time then model predictions of radically different climate states in the past can be evaluated. If models predict the past with skill we can have confidence in their predictions of the future. Although this eloquent approach has yielded many interesting outcomes the signal or error between data and models is often obscured by uncertainties in palaeoclimate and palaeoenvironmental reconstruction. Do to this kind of research work requires modellers and geologists to work together and to share openly the uncertainties and weaknesses in their models and climate archives.  In this talk we will review some of the examples of best and worst practice in joint studies that aim to examine past climate evolution and how well models are able to reproduce it.  This will include examples from climate extremes such as the Last Glacial Maximum when the UK was largely buried under a major ice sheet, from the mid Cretaceous and mid Eocene thermal maximum when no ice was present at either pole, and finally from the Pliocene epoch for which climate archives indicate a number f surprising parallels to model predictions of future climate.