There are a lot of factors that influence the climate system on Earth given the fact that we are dealing with three different levels - ocean, atmosphere and space. As a result, researchers are faced with ice-ocean, air-ice, and air-ocean interactions. Currents, wind as well as natural and human activities play a substantial role in processes such as emission, absorption and reflection. Next to this, scientists have to study the net solar short-wave and the net-terrestrial long-wave radiation, as the speaker explained.

Ulrich Cubasch first addressed the anthropogenic climate change which occurs through human activity. A new set of IPCC emissions or SRES scenarios for CO2 and SO2 were introduced taking into account the four parameters of rapid economic growth; the heterogeneity in local traditions and family values among the world population; a growing notion of dematerialisation and the introduction of clean technologies; and a new emphasis on local solutions to economic and environmental sustainability. In addition, a fifth "business as usual" scenario called IS 92a, was used by the speaker to show the impacts of both CO2 and SO2 emissions in the different scenario schedules.

The graphics demonstrate a more or less strong rise in CO2 emissions for the IS 92a, regional and environmental scenarios towards the year 2100. For the SO2 emissions, all scenarios show a strong reduction, especially between 2050 and 2100, except for the IS 92a one, which displays a spectacular growth. In addition, the global temperature rise, resulting from the SRES emission scenarios, has been simulated in 9 models and was calculated with a box diffusion model calibrated on various 3d models, according to Ulrich Cubasch.

Paleo-climatic research is performed to gain more insight in the climate changes that occurred during glacial and interglacial cycles. In the models the change of insolation can be varied by inserting orbital parameters. As such, it is possible to visualise the temperature evolution during different times in the past, as the speaker explained while showing a number of examples. Researchers also try to trace down the climate system from a geological perspective right from the beginning of Earth.

Scientific questions that occur are the following, according to Ulrich Cubasch:

• How will the climate of the future look like?
• How reliable are the projections?
• What are the important feedback mechanisms of the system?
• What role plays the human dimension? Can it be modelled?
• How can the climate change be monitored using inverse modelling?
• What are the important feedback mechanisms of the system?
• Why does the Milankovic theory not explain all the transitions in climate change? What causes ice ages to come and to go?
• What is the influence of the interstellar particle flux, solar irradiance variations and variations of Earth's magnetic field on climate?
• How did the climate look like, when the continents were shifted in Jura, Trias and other periods?
• Does evolution potentially have different pathways? Can we learn something about sustainability?

In order to predict future climate changes, we have to develop models that can integrate a complete biogeochemical Earth system. To this end, researchers need the right hardware and sufficient data storage, as well as high-end modelling software and pre- and post-processing programmes in an enhanced development environment, as the speaker noted.

In practice, this comes down to each time a 100 x increased computational demand to respectively expand the global resolution to 50km; to generate biology, chemistry and clouds; and to create multi-member ensembles out of chaos and quantify uncertainty. For the paleo-simulations, Ulrich Cubasch cited figures of 1000 x enhanced computational power to generate historical simulations; 100,000 x to simulate ice ages and factor 10** 9 to visualise climate evolutions.

Which tools are available right now? In Japan, the Earth Simulator has cost approximately $500 million and in Germany, about $50 million have been reserved over 10 years for the German Climate Computer Centre. According to the speaker there are at least 4 competitive ocean and 3 competitive atmosphere models available in Europe. However, this is only a strength in terms of science, but not in terms of infrastructure because many aspects are very similar in these models. Therefore, they should be unified for efficiency reasons.

Ulrich Cubasch stated that Europe needs to come up with a software framework that guarantees scientific diversity while at the same time increasing process efficiency in order to stay at the forefront of Earth system research. There are some important hurdles to overcome though because the different countries in the European federation not always share the same agenda and scientists are reluctant to move away from a centre where they feel at home so the "One-bit-centre-provides-everything" approach does not function.

From a user's perspective, the issue is the data storage. Research groups are structured around national centres which are connected through a high-speed network. The ideal solution would be to create a network of excellence. Given this context, Ulrich Cubasch proposed the opening of a virtual centre for competence in a metwork of excellence portal. This would allow scientists to plan and conduct common experiments and to share expertise. It gives access to data via semantic data storage and to visualisation and analysis, providing standard tools for easy comparison and possibilities to assemble individual solutions.

In addition, the portal should offer access to an Earth system modelling framework with standard configurations, to component models, couplers, an execution environment for experiment control, and to sufficient high-performance computing power, as the speaker explained. In this regard, such portal can be a means to exploit the potential of the European Earth system modelling infrastructure. Ulrich Cubasch concluded by stressing that at present Europe is missing an ambitious computing industry to create such a virtual European Earth system modelling facility.