Conferences - Seminars

  Monday 16 April 2018 10:30 - 11:30 PPB 019

Climate Change: From the Greenhouse Effect to High-Resolution Climate Modeling

By Prof. Christoph Schär, ETH-Zürich, Atmospheric and Climate Science

Climate change is one of the most pressing social and economic issues, but also one of the most complex scientific challenges. The basic concepts of anthropogenic climate change are very old, and first estimates of the climate effects of increased atmospheric carbon dioxide concentrations date back to the work of Svante Arrhenius in 1896. Despite this long history, uncertainties in climate change projections have remained very large. More specifically, the key uncertainty is phrased in terms of the equilibrium climate sensitivity (ECS), which is the equilibrium global-mean surface warming from a doubling of atmospheric CO2 concentrations. In the influential 1979 Charney report, the ECS was estimated to be between 1.5 and 4.5 K, while estimates of current climate models are scattered between 2 and 5 K. Thus the wide range of ECS estimates has neither shifted nor narrowed substantially. Projections of sea-level rise and regional climate change typically have even larger uncertainties, thereby challenging appropriate response strategies for adaptation and mitigation purposes.
The main cause behind the slow progress is the representation of clouds in climate models, especially convective clouds (i.e. thunderstorms, rain showers, shallow convective cloud layers). If clouds reflect more sunlight as the climate warms they reduce the warming, and vice versa. Convective clouds challenge current modeling strategies. They are characterized by small horizontal scales of O(1 km), while global climate models operate at horizontal resolutions of typically 50-150 km, implying the use of semi-empirical parameterizations. With the advent of high-resolution climate models, there are now promising prospects, as it becomes feasible to base the models on a set of equations that is much closer to first principles.
The presentation will report about current approaches to increase the horizontal resolutions of climate models to O(1 km). Examples from limited-area simulations demonstrate how some of the key biases can be overcome and how the simulation of clouds may dramatically be improved. This development raises a number of technical issues, covering a wide range of areas including hardware and software design, as well as the optimal choice of numerical and analysis methodologies. The presentation will address simulations over Europe and the tropical Atlantic and discuss recent approaches to better exploit next generation supercomputing and mass-storage systems.

Organization Prof. P. Ricci  

Contact Prof. P. Ricci

Accessibility Informed public

Admittance Free