Reconstructing 30 million years of past climate change (CENTA 2024-OU2)

Climate is a critical driver for the evolution and extinction of species, and existing patterns of terrestrial biodiversity can only be fully explained through the effects of climate variations reaching back tens of millions of years. Over the last 30 million years, a principal driving factor for such

change has been the evolution of the Earth's crust, including the rise of the Himalayan-Tibetan and Andes ranges, the slow drift of continents and the opening and closing of ocean gateways between them. Superimposed on these tectonically induced variations are climatic oscillations on timescales of tens to hundreds of millennia driven by cyclical changes in the shape of the Earth's orbit and the

resulting growth and decay of the great polar ice caps and ice sheets. On even shorter, decadal to millennial timescales, nonlinear feedbacks internal to the ocean-climate dynamical system can sometimes drive almost equally large variations. At all timescales, the carbon cycle feeds back on these changes, sometimes amplifying and sometimes attenuating the variability by driving changes

in atmospheric CO2. We know a great deal about past temporal variations in climate from ocean and lake sediment cores, but the spatial variations in climate, from tropical rainforests to arid deserts and mountain and polar ice, are comparably dramatic, and these changes can only be fully described and explained

by computer models of the Earth system. Unfortunately, the huge complexity of the system means that projecting and analysing changes over the multi-million year timescales important for species evolution with the spatial detail required to assess the suitability of local habitats that drives that evolution, is effectively impossible using conventional modelling techniques.

This project will build on recently developed techniques that combine conventional Earth system model simulation with a range of spatio-temporal statistical simulation approaches (Holden et al. 2019, Thomson et al. 2021) to produce integrated reconstructions of Earth's climate change and decadal to millennial variability reaching back tens of millions of years, with the spatial resolution

required to understand the effects on biodiversity evolution